Advances in Cancer Research, Volume 91

Imatinib as a Paradigm of Targeted Therapies Brian J. Druker Howard Hughes Medical Institute, Cancer Institute, Oregon ...

Author: George F. Vande Woude (Editor) | George Klein (Editor)

25 downloads 893 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Imatinib as a Paradigm of Targeted Therapies Brian J. Druker Howard Hughes Medical Institute, Cancer Institute, Oregon Health and Science University, Portland, Oregon 97239

I. Introduction II. Identification of the BCR-ABL Tyrosine Kinase as a Therapeutic Target A. Chronic Myeloid Leukemia: Clinical Features B. Genetic Changes in Cancer Cells and the Philadelphia Chromosome C. Oncogenes and Mapping of the ABL Gene D. BCR-ABL, Tyrosine Kinase Activity, and CML E. BCR-ABL and Leukemia F. Animal Models of CML G. BCR-ABL Signaling and CML Pathogenesis H. BCR-ABL as a Therapeutic Target III. Development of an ABL-Specific Tyrosine Kinase Inhibitor A. Chemistry B. Preclinical Studies IV. Clinical Trials in CML A. Phase I Clinical Trials B. Phase II Studies C. Phase III Randomized Comparison of Imatinib with IFN- Plus Ara-C V. Mechanisms of Relapse VI. Structural Basis of ABL Inhibition by Imatinib VII. Activity of Imatinib in Other Indications A. Gastrointestinal Stromal Tumors B. Other Malignancies VIII. Lessons Learned from the Clinical Trials of Imatinib A. Patient Selection B. Dose Selection IX. Translating the Success of Imatinib to Other Malignancies References

Imatinib (Gleevec) exemplifies the successful development of a rationally designed, molecularly targeted therapy for the treatment of a specific cancer. This article reviews the identification of the BCR-ABL tyrosine kinase as a therapeutic target in chronic myeloid leukemia and the steps in the development of an agent to specifically inactivate this abnormality. The clinical trials results are reviewed along with a description of resistance mechanisms. As imatinib also inhibits the tyrosine kinase activity of KIT and the platelet-derived growth factor receptors, the extension of imatinib to malignancies Advances in CANCER RESEARCH 0065-230X/04 $35.00

1

Copyright 2004, Elsevier Inc. All rights reserved

2

Brian J. Druker driven by these kinases will be described. Issues related to clinical trials of molecularly targeted agents are discussed, including patient and dose selection. Last, the translation of this paradigm to other malignancies is explored. ß 2004 Elsevier Inc.

I. INTRODUCTION A wealth of knowledge has emerged regarding the molecular events involved in human cancer. A major priority is the translation of this growing body of knowledge into therapeutics targeted specifically to these pathways. Perhaps the best example in which an understanding of the molecular pathogenesis of a human malignancy has been translated into clinical reality is imatinib for chronic myeloid leukemia (CML). The success of imatinib has also exemplified how the confluence of several fields in cancer research can result in significant advances. These include the investigations of chromosomal or genetic changes in cancer, the study of transforming retroviruses leading to the discovery of oncogenes, the availability of molecular techniques to map genes, and the biochemical evaluation of protein phosphorylation leading to the discovery of tyrosine kinases. All of these have contributed to the identification of the BCR-ABL oncogene as the causative molecular abnormality of CML and the development of a drug designed to inactivate this enzyme. This review traces these discoveries and discusses some of the lessons learned in the clinical development of a molecularly targeted agent and the implications for translating this success to other malignancies.

II. IDENTIFICATION OF THE BCR-ABL TYROSINE KINASE AS A THERAPEUTIC TARGET A. Chronic Myeloid Leukemia: Clinical Features CML is a clonal hematopoietic stem cell disorder with an annual incidence of 1 or 2 cases per 100,000 per year. The first description of CML was by two pathologists, Rudolf Virchow and John Hughes Bennett, in 1845 (Bennett, 1845; Virchow, 1845). Although a debate ensued as to whose description was first, Virchow acknowledged that Bennett’s case report had predated his (Geary, 2000). Of note, these first accounts of CML occurred before staining methods for blood, which were not developed until the late 1800s. The chronic, or stable, phase of CML is characterized by excess numbers of myeloid cells that differentiate and function normally. Approximately 90% of patients will be diagnosed in this phase of the disease; however, within an average of 4 to 6 years, the disease transforms through an

Imatinib and Cancer Therapy

3

‘‘accelerated phase’’ to an invariably fatal acute leukemia, also known as blast crisis. Disease progression is likely due to the accumulation of molecular abnormalities that lead to a progressive loss of the capacity for terminal differentiation of the leukemic clone (Faderl et al., 1999; Sawyers, 1999). Treatment choices for patients with CML, before the introduction of imatinib, included stem cell transplantation, hydroxyurea or interferon (IFN-)-based regimens, with allogeneic stem cell transplantation being the only proven curative therapy. As the average age of onset of CML is greater than 50 years, this factor, plus the inability to identify suitably matched donors, limits allogeneic stem cell transplantation to a minority of patients. Thus, less than 20% of CML patients are cured with these treatment options (Faderl et al., 1999; Sawyers, 1999). Patients in the blast phase of the disease are highly refractory to therapy. Response rates to standard chemotherapy are 20% or less and median survival is 2 to 3 months (Kantarjian et al., 1987; Sacchi et al., 1999).

B. Genetic Changes in Cancer Cells and the Philadelphia Chromosome In 1960, Peter Nowell and David Hungerford described a consistent chromosomal abnormality in CML patients, an acrocentric chromosome that was thought to be a chromosomal deletion (Nowell and Hungerford, 1960). This was the first example of a chromosomal abnormality linked to a specific malignancy. In this seminal article, the authors stated, ‘‘the findings suggest a causal relationship between the chromosomal abnormality observed and chronic granulocytic leukemia’’ (Nowell and Hungerford, 1960). This prescient statement was met with skepticism as it was felt that the chromosome abnormality was an associated rather than causative phenomenon. As chromosomal banding techniques improved, it became apparent that the chromosome abnormality was a shortened chromosome 22. Then, in 1973, Janet Rowley determined that the shortened chromosome 22, the so-called Philadelphia (Ph) chromosome, was the product of a reciprocal translocation between the long arms of chromosomes 9 and 22, t(9;22)(q34;q11) (Rowley, 1973) (Fig. 1A).

C. Oncogenes and Mapping of the ABL Gene The study of transforming retroviruses led to the recognition that mutations in normal cellular genes could be oncogenic. One such retrovirus, the Abelson murine leukemia virus, was initially described in 1970 (Abelson and Rabstein, 1970). Studies of this retrovirus led to the identification of its

4

Brian J. Druker

Fig. 1 (A) Schematic diagram of the translocation that creates the Philadelphia chromosome. The ABL and BCR genes reside on the long arms of chromosomes 9 and 22, respectively. As a result of the translocation, a BCR-ABL gene is formed on the derivative chromosome 22 (Philadelphia chromosome). (B) Structures of the BCR and ABL genes, showing locations of the breakpoints and various mRNAs created. The arrows indicate possible breakpoints within ABL. Regardless of the specific breakpoint in ABL, fusion mRNAs are produced that fuse BCR sequences to the second ABL exon. Fusions between BCR exon e13 (previously b2) or exon e14 (previously b3) and ABL exon a2 produce p210BCR-ABL that is characteristic of CML, whereas


5

transforming gene, v-ABL, and to the cloning of its normal cellular homolog, c-ABL (Rosenberg and Witte, 1988). By mapping oncogenes to specific chromosomal locations, it was recognized that c-ABL, which normally resides on the long arm of chromosome 9, had been translocated to chromosome 22 in CML patients (de Klein et al., 1982). As the breakpoints on chromosome 22 clustered in a relatively small region that spanned 5.3 kilobases (kb), this region was named the breakpoint cluster region or BCR (Deininger et al., 2000; Groffen et al., 1984).

D. BCR-ABL, Tyrosine Kinase Activity, and CML The discovery of the translocation of c-ABL to chromosome 22 in CML patients prompted investigators to perform Northern blots with ABL probes. These studies demonstrated a larger than normal ABL mRNA in CML patients (Collins et al., 1984; Gale and Canaani, 1984). This was subsequently shown to be a chimeric mRNA that was a fusion of BCR and ABL sequences (Shtivelman et al., 1985). Similarly, a larger than normal ABL protein with tyrosine kinase activity was detected in CML cells and was identified as the product of the BCR-ABL mRNA (Ben-Neriah et al., 1986; Davis et al., 1985; Mes-Masson et al., 1986). These discoveries provided an important link between the study of oncogenes and the biochemistry of protein kinases, as it had previously been recognized that v-ABL possessed a novel kinase activity, the ability to phosphorylate tyrosine residues (Hunter and Cooper, 1985). BCR-ABL is now known to have elevated tyrosine kinase activity as compared with c-ABL, and the kinase activity of BCR-ABL is essential for its ability to transform cells (Konopka et al., 1984; Lugo et al., 1990).

E. BCR-ABL and Leukemia BCR-ABL can be detected in more than 95% of patients with CML. Most of the breakpoints in the ABL gene occur in a 200-kB region between two alternate first exons. Despite this, the first exon of ABL is not included in the BCR-ABL chimeric mRNA, presumably due to alternative splicing. Two slightly different chimeric BCR-ABL genes are present in patients with CML, depending on the precise location of the breakpoint in the BCR gene. Breaks can occur between exons b2 (also known as e13) and b3 (e14),

fusions between BCR exon e1 and ABL exon a2 give rise to p190BCR-ABL (found in two-thirds of patients with Ph-positive acute lymphoblastic leukemia). Rare CML patients have a breakpoint in the so-called micro-breakpoint cluster region (-BCR) and produce a 230-kDa BCR-ABL fusion protein (not shown).

6

Brian J. Druker

yielding a b2a2 fusion mRNA, whereas a break occurring between exons b3 and b4 produces a b3a2 fusion mRNA (Deininger et al., 2000) (Fig. 1B). Although the b3a2 mRNA encodes a BCR-ABL protein that is 25 amino acids larger than that encoded by the b2a2 transcript, both are referred to as p210BCR-ABL and have similar prognoses. The Ph chromosome and the BCR-ABL fusion gene are also found in 25–50% of adult patients with acute lymphoblastic leukemia (ALL) and rare cases of acute myeloid leukemia. In adults with Ph chromosome-positive ALL, one-third have BCR-ABL transcripts indistinguishable from those found in CML. In two-thirds, the genomic breakpoint on chromosome 22 occurs in the first intron of the BCR gene (between e1 and e2), resulting in a protein of 190 kDa (p190), also referred to as p185 (Fig. 1B) (Chan et al., 1987; Clark et al., 1987, 1988; Hermans et al., 1987). Approximately 5% of children with ALL are Ph chromosome positive and 95% of these patients have the p190 form of BCR-ABL. Other types of fusions have been observed in rare cases (Deininger et al., 2000).

F. Animal Models of CML In 1990, two experimental approaches demonstrated the ability of BCRABL, as the sole oncogenic abnormality, to cause leukemia. In one set of experiments, transgenic mice that express BCR-ABL were shown to develop a rapidly fatal acute leukemia (Heisterkamp et al., 1990). Using a different approach, a BCR-ABL-expressing retrovirus was used to infect murine bone marrow. These BCR-ABL-expressing marrow cells were used to repopulate irradiated mice. The transplanted mice developed a variety of myeloproliferative disorders, including a CML syndrome (Daley et al., 1990; Kelliher et al., 1990). Both of these approaches clearly demonstrated the leukemogenic potential of BCR-ABL; however, even in these models, it was still possible that secondary changes were required for leukemia to develop. More recently, Huettner et al. placed BCR-ABL under the control of a tetracycline-repressible promoter (Huettner et al., 2000). In these experiments, when tetracycline was withdrawn, the mice developed pre-B cell leukemia that completely remitted on administration of tetracycline. These experiments even more clearly demonstrate the leukemogenic potential of BCR-ABL as a sole oncogenic abnormality.

G. BCR-ABL Signaling and CML Pathogenesis Significant advances have been made in determining the signaling pathways that are impacted by BCR-ABL tyrosine kinase activity (Fig. 2). Numerous substrates and binding partners have been identified and current


7

Fig. 2 Signaling pathways impacted by BCR-ABL expression. BAD, BCL-2 antagonist of cell death; BCL-XL, long form of BCL-X protein (BCL, B cell lymphoma); CBL, Casitas B-lineage lymphoma protein; CRKL, CRK-like; DOK, docking protein; ERK, extracellular signal-related protein kinase; GAP, GTPase-activating protein; GRB-2, growth factor receptor-bound protein 2; MEK, MAPK (mitogen-activated protein kinase)/ERK kinase; MYK, PI3K, phosphatidylinositol 3-kinase; SAPK, stress-activated protein kinase; SHC, SRC homology 2 domain-containing protein; SOS, ‘‘son of sevenless’’ (guanine nucleotide exchange factor); STAT, signal transducer and activator of transcription; AKT, CRK, MYC, RAS, and RAF are products of viral oncogene homologs, and their abbreviations are derived from the viral oncogene names.

efforts are directed at linking these pathways to the specific pathologic defects that characterize CML (Deininger et al., 2000). These defects include the following: increased proliferation or decreased apoptosis of a hematopoietic stem cell or progenitor cell, leading to a massive increase in myeloid cell numbers; premature release of immature myeloid cells into the circulation, postulated to be due to a defect in adherence of myeloid progenitors to marrow stroma; and genetic instability resulting in disease progression. An example of a cellular pathway that links to an increased proliferative rate is activation of the RAS pathway. Protection from programmed cell death may be mediated in part through STAT-5 upregulation of the antiapoptotic molecule BCL-XL and phosphorylation of and inactivation of the proapoptotic molecule BAD by AKT (Deininger et al., 2000). CML cells also exhibit reduced adhesion to fibronectin, possibly as a downstream effect of CRKL phosphorylation (Deininger et al., 2000).

8

Brian J. Druker

Despite the seemingly endless expansion of the list of pathways activated by BCR-ABL and the increasing complexity that is being revealed in these pathways, all of the transforming functions of BCR-ABL are dependent on its tyrosine kinase activity (Lugo et al., 1990).

H. BCR-ABL as a Therapeutic Target BCR-ABL possesses many characteristics of an ideal therapeutic target. It is expressed in the majority of patients with CML and it has been shown to be the cause of CML. BCR-ABL functions as a constitutively activated tyrosine kinase and mutagenic analysis has shown that this activity is essential for the transforming function of the protein. Thus, an inhibitor of the BCR-ABL kinase would be predicted to be an effective and selective therapeutic agent for CML. This recognition has relied on numerous fields of study and their convergence (Fig. 3). There is the field of tumor virology, from which oncogenes such as v-ABL were identified. The field of chromosome banding and gene mapping led to the recognition of BCR-ABL arising from the (9;22) chromosomal translocation. In addition, there is the entire field of protein phosphorylation, from serine/threonine kinases to tyrosine kinases, linking to oncogenes, leading to an understanding that BCR-ABL functions as a tyrosine kinase. This understanding generated interest in developing a specific inhibitor of this protein tyrosine kinase.

Fig. 3 Summary of the significant events leading to an understanding of the molecular pathogenesis of CML and to specific therapy for this disease. TK, tyrosine kinase.


9

III. DEVELOPMENT OF AN ABL-SPECIFIC TYROSINE KINASE INHIBITOR A. Chemistry Given the success of imatinib and the enormous interest in protein kinase inhibitors, it is easy to forget the degree of skepticism that kinase inhibitors faced from the scientific community and the pharmaceutical industry in the 1980s and 1990s. Much of this skepticism was due to the prevailing thought that inhibitors of ATP binding would lack sufficient target specificity to be clinically useful. However, in 1988, Yaish et al. published a series of compounds, known as tyrphostins that demonstrated that specific tyrosine kinase inhibitors could be developed (Yaish et al., 1988). Working independently, scientists at Ciba-Geigy (now Novartis) under the direction of N. Lydon and A. Matter, performed high-throughput screens of chemical libraries, searching for compounds with kinase-inhibitory activity. From this time-consuming approach, a lead compound of the 2-phenylaminopyrimidine class was identified. The activity of the 2-phenylaminopyrimidine series was optimized for various kinases by synthesizing chemically related compounds and analyzing the relationship between their structure and activity (Fig. 4) (Buchdunger et al., 1996; Zimmermann et al., 1996, 1997). A key finding was that substitutions at the 6-position of the anilino phenyl ring led to loss of serine/threonine kinase inhibition, whereas the introduction of a methyl group at this position retained or enhanced activity against tyrosine kinases (Fig. 4C). The activity against the platelet-derived growth factor receptor (PDGFR) tyrosine kinase was further enhanced by the introduction of a benzamide group at the phenyl ring (Fig. 4B). These compounds were also found to possess inhibitory activity toward ABL, with CGP57148 (STI571, now imatinib mesylate, Gleevec, Glivec) emerging as the lead compound for clinical development. Introduction of N-methylpiperazine as a polar side chain greatly improved water solubility and oral bioavailability (Fig. 4D). Imatinib emerged from these efforts as the lead compound for preclinical development on the basis of its selectivity against CML cells in vitro and its drug-like attributes, including pharmacokinetic and formulation properties.

B. Preclinical Studies Imatinib showed inhibitory activity against ABL and its activated derivatives v-ABL, BCR-ABL, and TEL-ABL (Buchdunger et al., 1996; Carroll et al., 1996; Druker et al., 1996). Fifty percent inhibitory concentration

10

Brian J. Druker

Fig. 4 Optimization of the phenylaminopyrimidine lead structure and synthesis of imatinib. The initial 2-phenylaminopyrimidine lead inhibitor is shown (A), with steps (B) and (C) leading to imatinib (D). See text for details.

(IC50) values are in the range of 0.025 M, using in vitro kinase assays with immunoprecipitated or purified proteins. Activity against the PDGFR and KIT are in a similar range. In contrast, the IC50 values for a large number of other tyrosine and serine/threonine kinases were generally at least 100-fold higher, demonstrating that imatinib exhibits a high level of selectivity (Buchdunger et al., 2001). Cellular studies showed that imatinib specifically inhibited the proliferation of myeloid cell lines that express BCR-ABL or CML blast crisis cell lines (Deininger et al., 1997; Druker et al., 1996; Gambacorti-Passerini et al., 1997). The IC50 for BCR-ABL phosphorylation in intact cells is between 0.25 and 0.5 M (Buchdunger et al., 2001). Complete inhibition of proliferation with cell death through apoptotic mechanisms occurs between 0.5 and 1 M concentrations of imatinib. Further experiments showed that p185- and p210-expressing cells were equally sensitive to imatinib (Beran et al., 1998; Carroll et al., 1997). Concentrations of up to 10 M imatinib did not affect the growth of BCR-ABL-negative cell lines (Druker et al., 1996). To assess the effects of imatinib on committed hematopoietic progenitors, mononuclear cells from patients with CML and normal individuals were


11

studied in assays of colony formation. These studies showed a marked decrease (92–98%) in the number of BCR-ABL-positive colonies formed with no inhibition of normal colony formation, using 1 M imatinib (Deininger et al., 1997; Druker et al., 1996). The differential sensitivity of CML versus normal cells to imatinib was also confirmed with long-term culture-initiating cells, which represent early hematopoietic progenitor cells (Kasper et al., 1999; Marley et al., 2000). Dose-dependent inhibition of tumor growth was seen in animals inoculated with BCR-ABL-expressing cells and treated daily with imatinib, with no effects against v-SRC-expressing tumors. Using a once-per-day injection schedule of up to 50 mg/kg, tumor growth was inhibited, but not eradicated (Druker et al., 1996). The reason for this modest in vivo activity became apparent from the murine pharmacokinetic profile of imatinib. This profiling revealed a short drug half-life in mice, which was not seen in other species (rat, dog, human). Thus, in nude mice a single dose of imatinib inhibited BCR-ABL kinase activity for only 2 to 5 h. A three times-per-day dosing schedule led to a continual block of BCR-ABL kinase activity, resulting in eradication of tumors in 87% of imatinib-treated mice (le Coutre et al., 1999). On the basis of these data, it was considered likely that continuous exposure to imatinib would be required for optimal antileukemic effects.

IV. CLINICAL TRIALS IN CML A. Phase I Clinical Trials A standard dose-escalation, phase I study of imatinib began in June 1998. The study population consisted of CML patients in chronic phase, refractory or resistant to IFN--based therapy or intolerant of this drug (Druker et al., 2001b). At later stages of the study, patients with CML in blast crisis and patients with Ph chromosome-positive ALL were also enrolled (Druker et al., 2001a). Imatinib was well tolerated, with the most common side effects including occasional nausea, periorbital edema, and muscle cramps. Despite dose escalation from 25 to 1000 mg in 14 cohorts of patients, a maximally tolerated dose could not be defined. Imatinib was administered once daily and pharmacokinetics showed a half-life of 13–16 h (Druker et al., 2001b). At doses of 300 mg and above, significant therapeutic benefits were observed. In chronic phase patients who had failed therapy with IFN-, 53 of 54 (98%) patients treated at 300 mg and above achieved a complete hematologic response, and with 1 year of follow-up, only 1 of these patients relapsed (Druker et al., 2001b). In myeloid blast crisis

12

Brian J. Druker

patients, 21 of 38 (55%) patients treated at doses at or exceeding 300 mg/ day responded, with 18% having responses lasting beyond 1 year (Druker et al., 2001a).

B. Phase II Studies These remarkable phase I data led to rapidly accruing phase II clinical trials. These clinical trials confirmed the results seen in the phase I studies and led to Food and Drug Administration approval of imatinib in May 2001 (Fig. 5). In chronic phase patients who had failed IFN- therapy, 95% of patients achieved a complete hematologic response and 60% a major cytogenetic response, defined as a reduction in the percentage of Ph chromosome-positive metaphases to less than 35%. With a median follow-up of 29 months, only 13% of these patients have relapsed (Kantarjian et al., 2002). In accelerated phase and blast crisis patients, the response rates were also quite high, but relapses have been much more common, with the majority of blast crisis patients relapsing during the first year of therapy (Sawyers et al., 2002; Talpaz et al., 2002).

Fig. 5 Phase II clinical trials of imatinib for CML. The results of the phase II studies are shown for chronic phase patients who failed interferon therapy, and for accelerated phase and blast crisis patients. Results shown are with a median follow-up of up to 30 months, and the percent of patients with disease progression is at 24 months (Kantarjian et al., 2002; Sawyers et al., 2002; Talpaz et al., 2002). CR (complete response) and PR (partial response) for cytogenetic responses include patients with Ph chromosome-positive metaphases of less than or equal to 35%.

13


C. Phase III Randomized Comparison of Imatinib with IFN-a Plus Ara-C The next clinical trial performed was a phase III study of newly diagnosed chronic phase patients, comparing imatinib with the standard therapy of IFN- plus cytarabine (Ara-C). This study accrued more than 1000 patients in a 7-month period. Five hundred and fifty-three patients were randomized to each of the two treatments, imatinib at 400 mg/day or IFN- plus Ara-C. There were no significant differences in prognostic features between the two arms. With a median follow-up of 18 months, patients randomized to imatinib had significantly better results than did patients treated with IFN- plus Ara-C in all parameters measured, including rates of complete hematologic response (97 versus 69%, p < 0.001), major and complete cytogenetic responses (87 and 76% versus 35 and 14%, p < 0.001), discontinuation of assigned therapy due to intolerance (3 versus 31%), and progression to accelerated phase or blast crisis (3 versus 8%, p < 0.001) (Table I) (O’Brien et al., 2003). Responses to imatinib are rapid, with most patients taking imatinib achieving complete hematologic responses within the first 4 to 6 weeks of therapy. In addition, more than 50% of patients obtain a complete cytogenetic response in 3 months. Despite the fact that 76% of patients randomized to imatinib achieve a complete cytogenetic response, the majority of these patients have detectable leukemia as analyzed by reverse transcriptase-polymerase chain reaction (RT-PCR) for BCR-ABL (Hughes et al., 2003). When BCR-ABL transcript levels are analyzed by log reduction, 39% of patients achieved

Table I

Phase III Results of Imatinib versus IFN- Plus Cytarabine for Newly Diagnosed Chronic Phase Patients with CMLa Imatinib percentage responding at 400 mg/day (n ¼ 553)

CHR MCR CCR Intolerance Progressive disease

97 87 76 3 3

IFN- plus Ara-C (n ¼ 553) 69 35 14 31 8.5

aFrom O’Brien et al. (2003). Results are with a median follow-up of 18 months. Abbreviations: CHR, complete hematologic response; MCR, major cytogenetic response; CCR, complete cytogenetic response; intolerance, leading to discontinuation of first-line therapy; progressive disease, progressing to accelerated phase or blast crisis. All of these differences are highly significant, with p < 0.001.

14

Brian J. Druker

at least a 3-log reduction in levels, but only 13 and 3% achieved a 4- and 5-log reduction, respectively. Thus, most patients treated with imatinib have persistent disease at the molecular level. Remaining questions for chronic phase patients include the durability of responses and how to integrate imatinib therapy with allogeneic stem cell transplantation. The mechanisms of disease persistence and whether it will be possible to completely eradicate the disease are currently being investigated. However, for advanced phase patients, the more pressing question is, why do they relapse?

V. MECHANISMS OF RELAPSE Response rates to imatinib in chronic phase patients are quite high and thus far, responses have been durable. Response rates are also quite high in patients with advanced phase disease, but relapses, despite continued therapy with imatinib, have been common. This high response rate in advanced phase patients is encouraging as imatinib targets an early molecular change in a malignancy that presumably contains multiple molecular abnormalities. In all patients who have relapsed, the BCR-ABL kinase remains present. One of the most useful categorizations of relapse mechanisms has been to separate patients into those with persistent inhibition of the BCRABL kinase and those with reactivation of the BCR-ABL kinase (Fig. 6). Patients with persistent inhibition of the BCR-ABL kinase would be predicted to have additional molecular abnormalities besides BCR-ABL driving the growth and survival of the malignant clone. In contrast, patients with persistent BCR-ABL kinase activity or reactivation of the kinase would be postulated to have resistance mechanisms that either prevent imatinib from reaching the target or render the target insensitive to imatinib. In the former category are mechanisms such as drug efflux or protein binding of imatinib. In the latter category would be mutations of the BCR-ABL kinase that render BCR-ABL insensitive to imatinib or amplification of the BCR-ABL protein. To examine BCR-ABL kinase activity, an assay has been developed that looks at the major tyrosine-phosphorylated protein in CML patient samples, CRKL (Druker et al., 2001b; Oda et al., 1994). Using this assay, it has been determined that the majority of patients who respond to imatinib and then relapse have reactivation of the BCR-ABL tyrosine kinase (Gorre et al., 2001). In these studies, greater than 50% and perhaps as many as 90% of patients have BCR-ABL point mutations encoding at least 18 different amino acids scattered throughout the ABL kinase domain (Fig. 7) (Nardi et al., 2004; Shah and Sawyers, 2003). Other patients have amplification of


15

Fig. 6 Distinguishing between potential mechanisms of relapse among advanced phase CML patients.

Fig. 7 Schematic of point mutations in the ABL kinase domain. The ABL kinase domain, from amino acid 240 to 500, is shown with the ATP-binding domain (P), the catalytic domain (C), and the activation loop (A). The numbers below the kinase domain indicate amino acids that are mutated in patients who relapsed on therapy with imatinib. The columns above the kinase domain indicate the number of times each amino acid has been found to be mutated as compiled from the literature (Al-Ali et al., 2004; Barthe et al., 2002; Branford et al., 2002, 2003; Gorre et al., 2001; Hochhaus et al., 2002; Hofmann et al., 2001; Roche-Lestienne et al., 2002; Shah et al., 2002; von Bubnoff et al., 2002). Adapted from Shah et al. (2002).

16

Brian J. Druker

Fig. 8 (A) Structure of the activation loop (in black) in the kinase domains of ABL, HCK, and IRK in their inactive states. Note that, in ABL, the activation loop is folded back over the catalytic center of the molecule and occludes the mouth of the kinase. In the inactive state, the activation loops have distinct configurations, which forms the basis for specific inhibition.


17

BCR-ABL at the genomic or transcript level (Hochhaus et al., 2002). In contrast, in patients with primary resistance, that is, patients who do not respond to imatinib therapy, BCR-ABL-independent mechanisms are most common (Hochhaus et al., 2002). Whether these mechanisms of relapse will apply to the situation of disease persistence in chronic phase patients treated with imatinib is being investigated. Analysis of the ability of imatinib to inhibit the kinase activity of the BCR-ABL kinase domain mutations found in relapsed patients has shown that some might be sensitive to dose escalation, but that most are highly resistant to imatinib (Corbin et al., 2002, 2003). ABL kinase inhibitors with specificity that differs from that of imatinib have already been synthesized and one of these compounds, PD180970, is capable of inhibiting some, but not all, of the common BCR-ABL kinase mutations (La Rosee et al., 2002). These data suggest that it may be possible to treat patients with several different ABL kinase inhibitors to circumvent resistance.

VI. STRUCTURAL BASIS OF ABL INHIBITION BY IMATINIB The catalytic domains of kinases include the ATP-binding lobe (P-loop), the catalytic site, and the activation loop (A-loop). In active kinases, the A-loop is in an ‘‘open’’ conformation, as it swings away from the catalytic center of the kinase (Fig. 8A). Whereas the conformation of the A-loop is structurally similar in kinases when they are in the active, open conformation, there are considerable differences between their inactive (closed) conformations. The crystal structure of the catalytic domain of the ABL kinase in complex with an imatinib analog and with imatinib has been solved (Nagar et al., 2002; Schindler et al., 2000). The most important finding of these studies is that the compound binds to the inactive conformation of ABL, In contrast, tyrosine kinase domains are similar in their active states, when the activation loop is folded away from the catalytic center as in active LCK. (B) Structure of imatinib bound to the kinase domain of ABL. The arrow indicates the P-loop, and the arrowhead points to helix C. The activation loop is shown in blue, with the conserved DFG motif in gold. Imatinib penetrates the center of the kinase, stabilizing the inactive conformation of the activation loop. Binding to the activation loop occurs without major steric clashes, whereas binding to the P-loop involves an induced fit mechanism. Imatinib would be unable to bind to the active configuration of the kinase (A). (C) Amino acids contacting imatinib. Imatinib: carbon is shown in green, nitrogen in blue, and oxygen in red. Carbons of the protein backbone are shown in orange. Hydrogen bonds are indicated as dashed lines. Residues that form hydrophobic interactions are circled. Imatinib contacts 21 amino acids within the ABL kinase domain (adapted from Nagar et al., 2002).

18

Brian J. Druker

contacting 21 amino acid residues. By exploiting the distinct inactive conformation of the A-loop of ABL, imatinib is able to achieve its high specificity. No major structural rearrangements are required for imatinib to bind to the A-loop. In contrast, there are significant structural alterations induced by imatinib binding in the P-loop, which is glycine rich and highly flexible (Fig. 8B). In examining imatinib-resistant mutations, they fall into several categories. One obvious set of mutations includes those at contact points between the ABL kinase domain and imatinib. Examples of this include T315I and E255K. Another set includes those that alter the conformation of the binding pocket such that imatinib can no longer bind or those that prevent the alterations in the P-loop required for imatinib binding. A related class includes mutations that stabilize the open, active conformation of ABL, such that imatinib can no longer bind. Interestingly, ABL inhibitors with activity against these mutants have been shown to bind to the open, active conformation of ABL (Nagar et al., 2002). These inhibitors are also inactive against mutations that impair their binding (La Rosee et al., 2002); thus to date, no inhibitor of T315I has been identified.

VII. ACTIVITY OF IMATINIB IN OTHER INDICATIONS A. Gastrointestinal Stromal Tumors In addition to inhibiting the ABL tyrosine kinase, imatinib inhibits the PDGFR and KIT tyrosine kinases. There are now several other cancers in which imatinib has shown clinical benefits that are based on the profile of kinases inhibited by imatinib and an understanding of the genetic defects causing various malignancies (Table II). One such disease is gastrointestinal stromal tumor (GIST). GISTs are mesenchymal neoplasms that can arise from any organ in the gastrointestinal tract or from the mesentery or omentum. There are approximately 5000 new cases per year in the United States. Although GISTs morphologically resemble leiomyosarcomas and nerve sheath tumors, they are a distinct entity (Fletcher et al., 2002). Published data suggest that the response rate of GISTs to single- or multiagent chemotherapy is less than 5%. More than 90% of GISTs express KIT and biochemical evidence of KIT activation can be found in almost all GISTs (Hirota et al., 1998; Rubin et al., 2001). In approximately 90% of cases, this activation is linked to somatic mutations of KIT, usually involving exon 9 or 11 (Rubin et al., 2001). Several lines of evidence indicate that KIT mutations are an early pathogenetic event in GISTs: (1) a similar frequency and spectrum of mutations is seen in histologically benign versus malignant tumors and early/localized

19


Table II Target ABL KIT PDGFRA PDGFRB

Diseases Against Which Imatinib Has Shown Activity Disease

Mechanism of activation

CML ALL GIST GIST HES CMML DFSP

Chromosomal translocation t(9;22): BCR-ABL Chromosomal translocation t(9;22): BCR-ABL Point mutation Point mutation Intrachromosomal deletion: FIP1L1-PDGFRA Chromosomal translocation t(5;12): EVT6-PDGFRB Chromosomal translocation t(17;22): Col1A1-PDGFB

Abbreviations: ALL, acute lymphoblastic leukemia; CML, chronic myeloid leukemia; CMML, chronic myelomonocytic leukemia; DFSP, dermatofibrosarcoma protuberans; GIST, gastrointestinal stromal tumor; HES, hypereosinophilic syndrome.

versus late/metastatic tumors; (2) familial syndromes of GIST are associated with germ line mutations of KIT; and (3) molecular studies suggest that KIT mutations are acquired before the development of cytogenetic abnormalities (Corless et al., 2002; Heinrich et al., 2002; Rubin et al., 2001). The concept that patients with GISTs might benefit from treatment with imatinib was based in part on two experimental observations: (1) treatment of GIST cell lines with imatinib inhibited proliferation and induced apoptosis, and (2) several GIST-associated mutant KIT isoforms were potently inhibited by imatinib in vitro at concentrations similar to wild-type KIT (Heinrich et al., 2000a; Tuveson et al., 2001). On the basis of the identification of mutated KIT as a therapeutic target in GIST and the lack of an effective conventional medical therapy, a patient with chemotherapy-resistant gastric GIST metastatic to omentum and liver was started on imatinib at 400 mg/day in March 2000 (Joensuu et al., 2001). The tumor in this patient expressed an exon 11 mutant KIT isoform and the patient responded dramatically to imatinib therapy. Spurred by this clinical success, larger clinical trials for this disease indication were performed. In these clinical trials, the objective response rate to imatinib as a single agent in patients with advanced GIST was 53 to 65%, with another 19 to 36% of patients having disease stabilization (Demetri et al., 2002; van Oosterom et al., 2001). These clinical trials have served as the basis for further exploration of the utility of imatinib in GISTs, including the adjuvant and neoadjuvant, because the recurrence rate of GISTs after surgery is quite high.

B. Other Malignancies Imatinib has also shown significant activity in patients with ALL who are BCR-ABL positive (Ottmann et al., 2002), but responses are generally transient. Translocations involving the PDGFRB gene have been identified

20

Brian J. Druker

in several myeloproliferative and myelodysplastic syndromes. The most common of these translocations, t(5;12)(q33;p13), is seen in a subset of patients with chronic myelomonocytic leukemia (CMML) and results in fusion of the EVT6 (TEL) and PDGFRB genes (Golub et al., 1994). Several patients with CMML containing the (5;12) translocation have been treated with imatinib and all achieved complete hematologic remissions (Apperley et al., 2002; Magnusson et al., 2002). The PDGFRB pathway is also a target in dermatofibrosarcoma protuberans (DFSP), a low-grade sarcoma of the dermis that often recurs after surgical excision. These tumors are characterized by a (17;22) translocation involving the COL1A1 and PDGFB genes, which results in overproduction of PDGF-BB and consequent hyperactivation of PDGFRB (Simon et al., 1997). It has been shown that imatinib inhibits the growth of DFSP cells both in culture and in immunodeficient mice (Sjoblom et al., 2001) and preliminary results in patients look promising (Maki et al., 2002; Rubin et al., 2002). Another disease caused by KIT mutations is systemic mastocytosis. The majority of tumors have a mutation of Asp-816 to valine (D816V) in the kinase domain of KIT, resulting in activation of KIT. Unfortunately, the kinase activity of the D816V mutant isoform is resistant to imatinib (Heinrich et al., 2000b; Ma et al., 2002), probably because this mutation induces conformational changes in the activation loop of KIT that prevent the binding of imatinib. Thus, imatinib is unlikely to be useful in this disorder, but an inhibitor of this KIT mutation would be predicted to be an effective therapy for this disorder. As KIT and the PDGFRs are expressed in many common tumors and are reported to be activated by both autocrine and paracrine mechanisms, there has been much interest in using imatinib broadly. As no other tumor has genetic evidence of involvement of these receptor systems, clinical trials with imatinib in these indications would be considered empiric clinical trials. Thus far, single-agent activity of imatinib in tumors such as smallcell lung cancer and breast, prostate, melanoma, and non-GIST sarcomas has been minimal (Johnson et al., 2003; Modi et al., 2003; Rao et al., 2003; Verweij et al., 2003; Wyman et al., 2003). One example in which empiric clinical trials of imatinib have been successful is hypereosinophilic syndrome (HES) (Gleich et al., 2002). The dramatic empiric results of imatinib in HES prompted investigations of the molecular basis for imatinib activity in this disease. Two groups independently arrived at the conclusion that an intrachromosomal deletion on chromosome 4 resulted in a fusion between a gene of unknown function, FIP1L1, and a truncated PDGFRA in a large percentage of patients with this disorder (Cools et al., 2003; Griffin et al., 2003). The resulting FIP1L1-PDGFRA fusion protein is a constitutively activated tyrosine kinase that is imatinib sensitive, thus accounting for the responsiveness of this disease to imatinib.


21

VIII. LESSONS LEARNED FROM THE CLINICAL TRIALS OF IMATINIB A. Patient Selection One of the goals of cancer research is to develop the ability to match the right patient with the right drug, based on specific knowledge of the genetic abnormalities of a tumor. The impact of this was clear in the clinical trials of imatinib. As an ABL inhibitor was being tested, enrollment was limited to patients with activated ABL driving their cancer, and these patients could be identified easily as they had the Ph chromosome or BCR-ABL. In GIST, the situation was more complicated. In these clinical trials, virtually all patient tumors expressed KIT. The majority of tumors had activating KIT mutations in exon 11 and these patients had a partial response rate of close to 80%. In contrast, patients whose tumors expressed wild-type KIT, with no mutation, had a response rate of only 18% (Heinrich et al., 2003a). Thus, KIT mutational status correlated with response and in GIST, expression of KIT is not sufficient to predict responses; rather, a mutation in KIT is necessary for responses to be observed. It has also been instructive to determine why 18% of patients with wildtype KIT expression respond to imatinib. Examination of tumors from patients with wild-type KIT expression showed that one-third of these tumors had activating mutations of the PDGFRA gene (Heinrich et al., 2003b). These mutations occurred in two different exons. One set of mutations was imatinib sensitive and this accounted for responses observed in patients whose tumors expressed wild-type KIT (Heinrich et al., 2003b). Thus, careful evaluations of subsets of responding patients can yield important insights into disease pathogenesis and the mechanism of response to an agent.

B. Dose Selection In the phase I clinical trials of imatinib, a maximally tolerated dose was never reached. Even among those closely associated with the phase I studies there was a lack of consensus about when to discontinue dose escalation. There were some investigators who felt that no cap on the dose should be considered except the maximally tolerated dose, particularly because studies of solid tumors were planned and because penetration of imatinib into solid tumors might require higher doses. Others believed that alternative end points, such as optimal therapeutic response and/or pharmacokinetic end points, could be used. In evaluating pharmacokinetic end points,

22

Brian J. Druker

it was known from preclinical studies that continuous exposure of cells to imatinib doses of 1 M or higher resulted in maximal cell killing (Druker and Lydon, 2000). In the phase I clinical trials, a trough level of 1 M was reached at a dose level of 300 mg, which corresponds to a threshold for significant therapeutic benefits (Peng et al., 2004). Thus, pharmacokinetic parameters could have been used to predict therapeutic responses. With solid tumors, pharmacokinetic analyses of tumor samples might be necessary to address the issue of tumor penetrance. When using molecularly targeted agents, it would seem more reasonable to consider maximal modulation of the target as the therapeutic end point. In the case of CML and a BCR-ABL kinase inhibitor, the obvious choice would be to assess for maximal inhibition of the BCR-ABL kinase activity. Direct assays of BCR-ABL tyrosine phosphorylation in patient samples have been difficult; therefore, assays of CRKL phosphorylation have been used. CRKL, a 39-kDa SH2, SH3-containing adaptor protein, is the major novel tyrosine phosphorylated protein in CML neutrophils and is a direct substrate of BCR-ABL (Oda et al., 1994). Assays of CRKL phosphorylation, based on its migration in sodium dodecyl sulfate–polyacrylamide gels, has shown that the BCR-ABL kinase is inhibited by imatinib in vivo (Druker et al., 2001b). This assay has also been useful in evaluating relapse mechanisms (Gorre et al., 2001; Hochhaus et al., 2002). However, as the amount of inhibition is not easily quantitated, this assay has been less useful at defining an optimal dose of therapy.

IX. TRANSLATING THE SUCCESS OF IMATINIB TO OTHER MALIGNANCIES The clinical trials with imatinib are a dramatic demonstration of the potential of targeting molecular pathogenetic events in a malignancy. As this paradigm is applied to other malignancies, it is worth remembering that BCR-ABL and CML have several features that were critical to the success of this agent. One of these is that BCR-ABL tyrosine kinase activity has clearly been demonstrated to be critical to the pathogenesis of CML. Thus, not only was the target of imatinib known, but also the target is a critical factor required for the development of CML. Another important feature is that as with most malignancies, treatment earlier in the course of the disease yields better results. Specifically, the response rate and durability of responses have been greater in chronic phase patients as opposed to blast phase patients. Thus, for maximal utility as a single agent, the identification of crucial, early events in malignant progression is the first step in reproducing the success with imatinib in other malignancies. An equally important issue is the


23

selection of patients for clinical trials on the basis of the presence of an appropriate target. Again, in the CML experience, patients with activation of BCR-ABL were easily identifiable by the presence of the Ph chromosome. In this regard, as reagents to analyze molecular end points are developed, these same reagents should be useful in identifying appropriate candidates for treatment with a specific agent. When all of these elements are put together, a critical pathogenetic target that is easily identifiable early in the course of the disease, remarkable results with an agent that targets this abnormality can be achieved. The obvious goal is to identify these early pathogenetic events in each malignancy and to develop agents that specifically target these abnormalities.

REFERENCES Abelson, H. T., and Rabstein, L. S. (1970). Lymphosarcoma: Virus-induced thymic-independent disease in mice. Cancer Res. 30, 2213–2222. Al-Ali, H.-K., Heinrich, M. C., Lange, T., Krahl, R., Mueller, M., Mueller, C., Niederwieser, D., Druker, B. J., and Deininger, M. W. (2004). High incidence of BCR-ABL kinase domain mutations in CML patients with secondary resistance to imatinib. Hematol. J. 5, 55–60. Apperley, J. F., Gardembas, M., Melo, J. V., Russell-Jones, R., Bain, B. J., Baxter, E. J., Chase, A., Chessells, J. M., Colombat, M., Dearden, C. E., Dimitrijevic, S., Mahon, F. X., Marin, D., Nikolova, Z., Olavarria, E., Silberman, S., Schultheis, B., Cross, N. C., and Goldman, J. M. (2002). Response to imatinib mesylate in patients with chronic myeloproliferative diseases with rearrangements of the platelet-derived growth factor receptor . N. Engl. J. Med. 347, 481–487. Barthe, C., Gharbi, M. J., Lagarde, V., Chollet, C., Cony-Makhoul, P., Reiffers, J., Goldman, J. M., Melo, J. V., and Mahon, F. X. (2002). Mutation in the ATP-binding site of BCR-ABL in a patient with chronic myeloid leukaemia with increasing resistance to STI571. Br. J. Haematol. 119, 109–111. Ben-Neriah, Y., Daley, G. Q., Mes-Masson, A.-M., Witte, O. N., and Baltimore, D. (1986). The chronic myelogenous leukemia-specific P210 protein is the product of the bcr/abl hybrid gene. Science 233, 212–214. Bennett, J. H. (1845). Case of hypertrophy of the spleen and liver in which death took place from suppuration of the blood. Edinb. Med. Surg. J. 64, 413–423. Beran, M., Cao, X., Estrov, Z., Jeha, S., Jin, G., O’Brien, S., Talpaz, M., Arlinghaus, R. B., Lydon, N. B., and Kantarjian, H. (1998). Selective inhibition of cell proliferation and BCR-ABL phosphorylation in acute lymphoblastic leukemia cells expressing Mr 190,000 BCR-ABL protein by a tyrosine kinase inhibitor (CGP-57148). Clin. Cancer Res. 4, 1661–1672. Branford, S., Rudzki, Z., Walsh, S., Grigg, A., Arthur, C., Taylor, K., Herrmann, R., Lynch, K. P., and Hughes, T. P. (2002). High frequency of point mutations clustered within the adenosine triphosphate-binding region of BCR/ABL in patients with chronic myeloid leukemia or Ph-positive acute lymphoblastic leukemia who develop imatinib (STI571) resistance. Blood 99, 3472–3475. Branford, S., Rudzki, Z., Walsh, S., Parkinson, I., Grigg, A., Szer, J., Taylor, K., Herrmann, R., Seymour, J. F., Arthur, C., Joske, D., Lynch, K., and Hughes, T. (2003). Detection of

24

Brian J. Druker

BCR-ABL mutations in patients with CML treated with imatinib is virtually always accompanied by clinical resistance, and mutations in the ATP phosphate-binding loop (P-loop) are associated with a poor prognosis. Blood 102, 276–283. Buchdunger, E., Zimmermann, J., Mett, H., Meyer, T., Mu¨ ller, M., Druker, B. J., and Lydon, N. B. (1996). Inhibition of the Abl protein-tyrosine kinase in vitro and in vivo by a 2-phenylaminopyrimidine derivative. Cancer Res. 56, 100–104. Buchdunger, E., Matter, A., and Druker, B. J. (2001). Bcr-Abl inhibition as a modality of CML therapeutics. Biochim. Biophys. Acta 1551, M11–M18. Carroll, M., Tomasson, M. H., Barker, G. F., Golub, T. R., and Gilliland, D. G. (1996). The Tel platelet-derived growth factor receptor (PDGF--R) fusion in chronic myelomonocytic leukemia is a transforming protein that self-associates and activates PDGF--R kinase-dependent signaling pathways. Proc. Natl. Acad. Sci. USA 93, 14845–14850. Carroll, M., Ohno-Jones, S., Tamura, S., Buchdunger, E., Zimmermann, J., Lydon, N. B., Gilliland, D. G., and Druker, B. J. (1997). CGP 57148, a tyrosine kinase inhibitor, inhibits the growth of cells expressing BCR-ABL, TEL–ABL and TEL–PDGFR fusion proteins. Blood 90, 4947–4952. Chan, L. C., Karhi, K. K., Rayter, S. I., Heisterkamp, N., Eridani, S., Powles, R., Lawler, S. D., Groffen, J., Foulkes, J. G., Greaves, M. F. et al. (1987). A novel Abl protein expressed in Philadelphia chromosome positive acute lymphoblastic leukaemia. Nature 325, 635–637. Clark, S. S., McLaughlin, J., Crist, W. M., Champlin, R., and Witte, O. N. (1987). Unique forms of the Abl tyrosine kinase distinguish Ph1-positive CML from Ph1-positive ALL. Science 235, 85–88. Clark, S. S., McLaughlin, J., Timmons, M., Pendergast, A. M., Ben-Neriah, Y., Dow, L. W., Crist, W., Rovera, G., Smith, S. D., and Witte, O. N. (1988). Expression of a distinctive BCR-ABL oncogene in Ph1-positive acute lymphocytic leukemia (ALL). Science 239, 775–777. Collins, S. J., Kubonishi, I., Miyoshi, I., and Groudine, M. T. (1984). Altered transcription of the c-abl oncogene in K-562 and other chronic myelogenous leukemia cells. Science 225, 72–74. Cools, J., DeAngelo, D. J., Gotlib, J., Stover, E. H., Legare, R. D., Cortes, J., Kutok, J., Clark, J., Galinsky, I., Griffin, J. D., Cross, N. C., Tefferi, A., Malone, J., Alam, R., Schrier, S. L., Schmid, J., Rose, M., Vandenberghe, P., Verhoef, G., Boogaerts, M., Wlodarska, I., Kantarjian, H., Marynen, P., Coutre, S. E., Stone, R., and Gilliland, D. G. (2003). A tyrosine kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. N. Engl. J. Med. 348, 1201–1214. Corbin, A. S., Buchdunger, E., Pascal, F., and Druker, B. J. (2002). Analysis of the structural basis of specificity of inhibition of the Abl kinase by STI571. J. Biol. Chem. 277, 32214–32219. Corbin, A. S., La Rose´ e, P., Stoffregen, E., Druker, B. J., and Deininger, M. W. (2003). Several Bcr-Abl kinase domain mutants associated with imatinib mesylate resistance remain sensitive to imatinib. Blood 101, 4611–4614. Corless, C. L., McGreevey, L., Haley, A., Town, A., and Heinrich, M. C. (2002). KIT mutations are common in incidental gastrointestinal stromal tumors one centimeter or less in size. Am. J. Pathol. 160, 1567–1572. Daley, G. Q., Van Etten, R. A., and Baltimore, D. (1990). Induction of chronic myelogenous leukemia in mice by the P210bcr/abl gene of the Philadelphia chromosome. Science 247, 824–830. Davis, R. L., Konopka, J. B., and Witte, O. N. (1985). Activation of the c-abl oncogene by viral transduction or chromosomal translocation generates altered c-Abl proteins with similar in vitro kinase properties. Mol. Cell. Biol. 5, 204–213. Deininger, M. W., Goldman, J. M., Lydon, N., and Melo, J. V. (1997). The tyrosine kinase inhibitor CGP57148B selectively inhibits the growth of BCR-ABL-positive cells. Blood 90, 3691–3698.


25

Deininger, M. W., Goldman, J. M., and Melo, J. V. (2000). The molecular biology of chronic myeloid leukemia. Blood 96, 3343–3356. de Klein, A., van Kessel, A. G., Grosveld, G., Bartram, C. R., Hagemeijer, A., Bootsma, D., Spurr, N. K., Heisterkamp, N., Groffen, J., and Stephenson, J. R. (1982). A cellular oncogene is translocated to the Philadelphia chromosome in chronic myelocytic leukemia. Nature 300, 765–767. Demetri, G. D., von Mehren, M., Blanke, C. D., Van den Abbeele, A. D., Eisenberg, B., Roberts, P. J., Heinrich, M. C., Tuveson, D. A., Singer, S., Janicek, M., Fletcher, J. A., Silverman, S. G., Silberman, S. L., Capdeville, R., Kiese, B., Peng, B., Dimitrijevic, S., Druker, B. J., Corless, C., Fletcher, C. D., and Joensuu, H. (2002). Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors. N. Engl. J. Med. 347, 472–480. Druker, B. J., and Lydon, N. B. (2000). Lessons learned from the development of an Abl tyrosine kinase inhibitor for chronic myelogenous leukemia. J. Clin. Invest. 105, 3–7. Druker, B. J., Tamura, S., Buchdunger, E., Ohno, S., Segal, G. M., Fanning, S., Zimmermann, J., and Lydon, N. B. (1996). Effects of a selective inhibitor of the ABL tyrosine kinase on the growth of BCR-ABL positive cells. Nat. Med. 2, 561–566. Druker, B. J., Sawyers, C. L., Kantarjian, H., Resta, D. J., Reese, S. F., Ford, J. M., Capdeville, R., and Talpaz, M. (2001a). Activity of a specific inhibitor of the BCR-ABL tyrosine kinase in the blast crisis of chronic myeloid leukemia and acute lymphoblastic leukemia with the Philadelphia chromosome. N. Engl. J. Med. 344, 1038–1042. Druker, B. J., Talpaz, M., Resta, D. J., Peng, B., Buchdunger, E., Ford, J. M., Lydon, N. B., Kantarjian, H., Capdeville, R., Ohno-Jones, S., and Sawyers, C. L. (2001b). Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N. Engl. J. Med. 344, 1031–1037. Faderl, S., Talpaz, M., Estrov, Z., and Kantarjian, H. M. (1999). Chronic myelogenous leukemia: Biology and therapy. Ann. Intern. Med. 131, 207–219. Fletcher, C. D., Berman, J. J., Corless, C., Gorstein, F., Lasota, J., Longley, B. J., Miettinen, M., O’Leary, T. J., Remotti, H., Rubin, B. P., Shmookler, B., Sobin, L. H., and Weiss, S. W. (2002). Diagnosis of gastrointestinal stromal tumors: A consensus approach. Hum. Pathol. 33, 459–465. Gale, R. P., and Canaani, E. (1984). An 8-kilobase abl RNA transcript in chronic myelogenous leukemia. Proc. Natl. Acad. Sci. USA 81, 5648–5652. Gambacorti-Passerini, C., le Coutre, P., Mologni, L., Fanelli, M., Bertazzoli, C., Marchesi, E., Di Nicola, M., Biondi, A., Corneo, G. M., Belotti, D., Pogliani, E., and Lydon, N. B. (1997). Inhibition of the ABL kinase activity blocks the proliferation of BCR/ABLþ leukemic cells and induces apoptosis. Blood Cells Mol. Dis. 23, 380–394. Geary, C. G. (2000). The story of chronic myeloid leukaemia. Br. J. Haematol. 10, 2–11. Gleich, G. J., Leiferman, K. M., Pardanani, A., Tefferi, A., and Butterfield, J. H. (2002). Treatment of hypereosinophilic syndrome with imatinib mesilate. Lancet 359, 1577–1578. Golub, T. R., Barker, G. F., Lovett, M., and Gilliland, D. G. (1994). Fusion of PDGF receptor to a novel ets-like gene, tel, in chronic myelomonocytic leukemia with t(5;12) chromosomal translocation. Cell 77, 307–316. Gorre, M. E., Mohammed, M., Ellwood, K., Hsu, N., Paquette, R., Rao, P. N., and Sawyers, C. L. (2001). Clinical resistance to STI-571 cancer therapy caused by BCR-ABL gene mutation or amplification. Science 293, 876–880. Griffin, J. H., Leung, J., Bruner, R. J., Caligiuri, M. A., and Briesewitz, R. (2003). Discovery of a fusion kinase in EOL-1 cells and idiopathic hypereosinophilic syndrome. Proc. Natl. Acad. Sci. USA 100, 7830–7835. Groffen, J., Stephenson, J. R., Heisterkamp, N., de Klein, A., Bartram, C. R., and Grosveld, G. (1984). Philadelphia chromosome breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 36, 93–99.

26

Brian J. Druker

Heinrich, M. C., Griffith, D. J., Druker, B. J., Wait, C. L., Ott, K. A., and Zigler, A. J. (2000a). Inhibition of c-Kit receptor tyrosine kinase activity by STI 571, a selective tyrosine kinase inhibitor. Blood 96, 925–932. Heinrich, M. C., Wait, C. L., Yee, K. W. H., and Griffith, D. J. (2000b). STI571 inhibits the kinase activity of wild type and juxtamembrane c-kit mutants but not the exon 17 D816V mutation associated with mastocytosis. Blood 96, 173b. Heinrich, M. C., Rubin, B. P., Longley, B. J., and Fletcher, J. A. (2002). Biology and genetic aspects of gastrointestinal stromal tumors: KIT activation and cytogenetic alterations. Hum. Pathol. 33, 484–495. Heinrich, M. C., Corless, C. L., Demetri, G. D., Blanke, C. D., von Mehren, M., Joensuu, H., McGreevey, L. S., Chen, C. J., Van den Abbeele, A. D., Druker, B. J., Kiese, B., Eisenberg, B., Roberts, P. J., Singer, S., Fletcher, C. D., Silberman, S., Dimitrijevic, S., and Fletcher, J. A. (2003a). Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J. Clin. Oncol. 21, 4342–4349. Heinrich, M. C., Corless, C. L., Duensing, A., McGreevey, L., Chen, C. J., Joseph, N., Singer, S., Griffith, D. J., Haley, A., Town, A., Demetri, G. D., Fletcher, C. D., and Fletcher, J. A. (2003b). PDGFRA activating mutations in gastrointestinal stromal tumors. Science 299, 708–710. Heisterkamp, N., Jenster, G., ten Hoeve, J., Zovich, D., Pattengale, P. K., and Groffen, J. (1990). Acute leukaemia in bcr/abl transgenic mice. Nature 344, 251–253. Hermans, A., Heisterkamp, N., von Lindern, M., van Baal, S., Meijer, D., van der Plas, D., Wiedemann, L. M., Groffen, J., Bootsma, D., and Grosveld, G. (1987). Unique fusion of bcr and c-abl genes in Philadelphia chromosome positive acute lymphoblastic leukemia. Cell 51, 33–40. Hirota, S., Isozaki, K., Moriyama, Y., Hashimoto, K., Nishida, T., Ishiguro, S., Kawano, K., Hanada, M., Kurata, A., Takeda, M., Muhammad Tunio, G., Matsuzawa, Y., Kanakura, Y., Shinomura, Y., and Kitamura, Y. (1998). Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 279, 577–580. Hochhaus, A., Kreil, S., Corbin, A. S., La Rose´ e, P., Mu¨ ller, M. C., Lahaye, T., Hanfstein, B., Schoch, C., Cross, N. C. P., Berger, U., Gschaidmeier, H., Druker, B. J., and Hehlmann, R. (2002). Molecular and chromosomal mechanisms of resistance to imatinib (STI571) therapy. Leukemia 16, 2190–2196. Hofmann, W. K., Jones, L. C., Lemp, N. A., de Vos, S., Gschaidmeier, H., Hoelzer, D., Ottmann, O. G., and Koeffler, H. P. (2001). Phþ acute lymphoblastic leukemia resistant to the tyrosine kinase inhibitor STI571 has a unique BCR-ABL gene mutation. Blood 99, 1860–1862. Huettner, C. S., Zhang, P., Van Etten, R. A., and Tenen, D. G. (2000). Reversibility of acute B-cell leukaemia induced by BCR-ABL1. Nat. Genet. 24, 57–60. Hughes, T. P., Kaeda, J., Branford, S., Rudzki, Z., Hochhaus, A., Hensley, M. L., Gathmann, I., Bolton, A. E., van Hoomissen, I. C., Goldman, J. M., and Radich, J. P. (2003). Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N. Engl. J. Med. 349, 1423–1432. Hunter, T., and Cooper, J. A. (1985). Protein-tyrosine kinases. Annu. Rev. Biochem. 54, 897–930. Joensuu, H., Roberts, P. J., Sarlomo-Rikala, M., Andersson, L. C., Tervahartiala, P., Tuveson, D., Silberman, S., Capdeville, R., Dimitrijevic, S., Druker, B., and D. D. G. (2001). Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N. Engl. J. Med. 344, 1052–1056. Johnson, B. E., Fischer, T., Fischer, B., Dunlop, D., Rischin, D., Silberman, S., Kowalski, M. O., Sayles, D., Dimitrijevic, S., Fletcher, C., Hornick, J., Salgia, R., and Le Chevalier, T. (2003). Phase II study of imatinib in patients with small cell lung cancer. Clin. Cancer Res. 9, 5880–5887.


27

Kantarjian, H., Sawyers, C., Hochhaus, A., Guilhot, F., Schiffer, C., Gambacorti-Passerini, C., Niederwieser, D., Resta, D., Capdeville, R., Zoellner, U., Talpaz, M., and Druker, B. (2002). Hematologic and cytogenetic responses to imatinib mesylate in chronic myelogenous leukemia. N. Engl. J. Med. 346, 645–652. Kantarjian, H. M., Keating, M. J., Talpaz, M., Walters, R. S., Smith, T. L., Cork, A., McCredie, K. B., and Freireich, E. J. (1987). Chronic myelogenous leukemia in blast crisis: Analysis of 242 patients. Am. J. Med. 83, 445–454. Kasper, B., Fruehauf, S., Schiedlmeier, B., Buchdunger, E., Ho, A. D., and Zeller, W. J. (1999). Favorable therapeutic index of a p210(BCR-ABL)-specific tyrosine kinase inhibitor: Activity on lineage-committed and primitive chronic myelogenous leukemia progenitors. Cancer Chemother. Pharmacol. 44, 433–438. Kelliher, M. A., McLaughlin, J., Witte, O. N., and Rosenberg, N. (1990). Induction of a chronic myelogenous leukemia-like syndrome in mice with v-abl and BCR/ABL. Proc. Natl. Acad. Sci. USA 87, 6649–6653. Konopka, J. B., Watanabe, S. M., and Witte, O. N. (1984). An alteration of the human c-Abl protein in K562 leukemia cells unmasks associated tyrosine kinase activity. Cell 37, 1035–1042. La Rosee, P., Corbin, A. S., Stoffregen, E. P., Deininger, M. W., and Druker, B. J. (2002). Activity of the Bcr-Abl kinase inhibitor PD180970 against clinically relevant Bcr-Abl isoforms that cause resistance to imatinib mesylate (Gleevec, STI571). Cancer Res. 62, 7149–7153. le Coutre, P., Mologni, L., Cleris, L., Marchesi, E., Buchdunger, E., Giardini, R., Formelli, F., and Gambacorti-Passerini, C. (1999). In vivo eradication of human BCR/ABL-positive leukemia cells with an ABL kinase inhibitor. J. Natl. Cancer Inst. 91, 163–168. Lugo, T. G., Pendergast, A. M., Muller, A. J., and Witte, O. N. (1990). Tyrosine kinase activity and transformation potency of bcr-abl oncogene products. Science 247, 1079–1082. Ma, Y., Zeng, S., Metcalfe, D. D., Akin, C., Dimitrijevic, S., Butterfield, J. H., McMahon, G., and Longley, B. J. (2002). The c-KIT mutation causing human mastocytosis is resistant to STI571 and other KIT kinase inhibitors: Kinases with enzymatic site mutations show different inhibitor sensitivity profiles than wild-type kinases and those with regulatory-type mutations. Blood 99, 1741–1744. Magnusson, M. K., Meade, K. E., Nakamura, R., Barrett, J., and Dunbar, C. E. (2002). Activity of STI571 in chronic myelomonocytic leukemia with a platelet-derived growth factor receptor fusion oncogene. Blood 100, 1088–1091. Maki, R. G., Awan, R. A., Dixon, R. H., Jhanwar, S., and Antonescu, C. R. (2002). Differential sensitivity to imatinib of 2 patients with metastatic sarcoma arising from dermatofibrosarcoma protuberans. Int. J. Cancer 100, 623–626. Marley, S. B., Deininger, M. W., Davidson, R. J., Goldman, J. M., and Gordon, M. Y. (2000). The tyrosine kinase inhibitor STI571, like interferon-, preferentially reduces the capacity for amplification of granulocyte-macrophage progenitors from patients with chronic myeloid leukemia. Exp. Hematol. 28, 551–557. Mes-Masson, A.-M., McLaughlin, J., Daley, G. Q., Paskind, M., and Witte, O. N. (1986). Overlapping cDNA clones define the complete coding region for the P210c-abl gene product associated with chronic myelogenous leukemia cells containing the Philadelphia chromosome. Proc. Natl. Acad. Sci. USA 83, 9768–9772. Modi, S., Seidman, A., Dickler, M., Moasser, M., D’Andrea, G., Moynahah, M. E., Panageas, K. S., Tan, L., Norton, L., and Hudis, C. (2003). A phase II trial of STI571 in patients with metastatic breast cancer. Proc. Am. Soc. Clin. Oncol. 22, 18a. Nagar, B., Bornmann, W. G., Pellicena, P., Schindler, T., Veach, D. R., Miller, W. T., Clarkson, B., and Kuriyan, J. (2002). Crystal structures of the kinase domain of c-Abl in complex with the small molecule inhibitors PD173955 and imatinib (STI-571). Cancer Res. 62, 4236–4243.

28

Brian J. Druker

Nardi, V., Azam, M., and Daley, G. Q. (2004). Mechanisms and implications of imatinib resistance mutations in BCR-ABL. Curr. Opin. Hematol. 11, 35–43. Nowell, P. C., and Hungerford, D. A. (1960). A minute chromosome in human chronic granulocytic leukemia. Science 132, 1497. O’Brien, S. G., Guilhot, F., Larson, R. A., Gathmann, I., Baccarani, M., Cervantes, C., Cornelissen, J., Fischer, T., Hochhaus, A., Hughes, T., Kantarjian, H., Lechner, K., Nielsen, J. L., Rousselot, P., Reiffers, J., Saglio, G., Shepherd, J., Simonsson, S., Gratwohl, A., Goldman, J. M., Taylor, K., Verhoef, G., Bolton, A. E., Capdeville, R., and Druker, B. J. (2003). Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N. Engl. J. Med. 348, 994–1004. Oda, T., Heaney, C., Hagopian, J., Okuda, K., Griffin, J. D., and Druker, B. J. (1994). CRKL is the major tyrosine phosphorylated protein in neutrophils from patients with chronic myelogenous leukemia. J. Biol. Chem. 269, 22925–22928. Ottmann, O. G., Druker, B. J., Sawyers, C. L., Goldman, J. M., Reiffers, J., Silver, R. T., Tura, S., Fischer, T., Deininger, M. W., Schiffer, C. A., Baccarani, M., Gratwohl, A., Hochhaus, A., Hoelzer, D., Fernandes-Reese, S., Gathmann, I., Capdeville, R., and O’Brien, S. G. (2002). A phase 2 study of imatinib in patients with relapsed or refractory Philadelphia chromosomepositive acute lymphoid leukemias. Blood 100, 1965–1971. Peng, B., Hayes, M., Resta, D., Racine-Poon, A., Druker, B. J., Talpaz, M., Sawyers, C. L., Rosamilia, M., Ford, J., Lloyd, P., and Capdeville, R. (2004). Clinical investigations of the pharmacokinetics and pharmacodynamics of imatinib in a phase 1 trial with chronic myeloid leukemia patients. J. Clin. Oncol. 22, 935–942. Rao, K. V., Goodin, S., Capanna, T., Doyle-Lindrud, S., and Dipaola, R. S. (2003). A phase II trial of imatinib mesylate in patients with PSA progression after local therapy for prostate cancer. Proc. Am. Soc. Clin. Oncol. 22, 409a. Roche-Lestienne, C., Soenen-Cornu, V., Grardel-Duflos, N., Lai, J. L., Philippe, N., Facon, T., Fenaux, P., and Preudhomme, C. (2002). Several types of mutations of the Abl gene can be found in chronic myeloid leukemia patients resistant to STI571, and they can pre-exist to the onset of treatment. Blood 100, 1014–1018. Rosenberg, N., and Witte, O. N. (1988). The viral and cellular forms of the Abelson (abl) oncogene. Adv. Virus Res. 35, 39–81. Rowley, J. D. (1973). A new consistent abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and giemsa staining. Nature 243, 290–293. Rubin, B. P., Singer, S., Tsao, C., Duensing, A., Lux, M. L., Ruiz, R., Hibbard, M. K., Chen, C. J., Xiao, S., Tuveson, D. A., Demetri, G. D., Fletcher, C. D. M., and Fletcher, J. A. (2001). KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res. 61, 8118–8121. Rubin, B. P., Schuetze, S. M., Eary, J. F., Norwood, T. H., Mirza, S., Conrad, E. U., and Bruckner, J. D. (2002). Molecular targeting of platelet-derived growth factor B by imatinib mesylate in a patient with metastatic dermatofibrosarcoma protuberans. J. Clin. Oncol. 20, 3586–3591. Sacchi, S., Kantarjian, H. M., O’Brien, S., Cortes, J., Rios, M. B., Giles, F. J., Beran, M., Koller, C. A., Keating, M. J., and Talpaz, M. (1999). Chronic myelogenous leukemia in nonlymphoid blastic phase: Analysis of the results of first salvage therapy with three different treatment approaches for 162 patients. Cancer 86, 2632–2641. Sawyers, C. L. (1999). Chronic myeloid leukemia. N. Engl. J. Med. 340, 1330–1340. Sawyers, C. L., Hochhaus, A., Feldman, E., Goldman, J. M., Miller, C. B., Ottmann, O. G., Schiffer, C. A., Talpaz, M., Guilhot, F., Deininger, M. W. N., Fischer, T., O’Brien, S. G., Stone, R. M., Gambacorti-Passerini, C. B., Russell, N. H., Reiffers, J. J., Shea, T. C., Chapuis, B., Coutre, S., Tura, S., Morra, E., Larson, R. A., Saven, A., Peschel, C., Gratwohl, A., Mandelli, F., Ben-Am, M., Gathmann, I., Capdeville, R., Paquette, R. L.,


29

and Druker, B. J. (2002). Imatinib induces hematologic and cytogenetic responses in patients with chronic myeloid leukemia in myeloid blast crisis: Results of a phase II study. Blood 99, 3530–3539. Schindler, T., Bornmann, W., Pellicena, P., Miller, W. T., Clarkson, B., and Kuriyan, J. (2000). Structural mechanism for STI-571 inhibition of abelson tyrosine kinase. Science 289, 1938–1942. Shah, N. P., and Sawyers, C. L. (2003). Mechanisms of resistance to STI571 in Philadelphia chromosome-associated leukemias. Oncogene 22, 7389–7395. Shah, N. P., Nicoll, J. M., Nagar, B., Gorre, M. E., Paquette, R. L., Kuriyan, J., and Sawyers, C. L. (2002). Multiple BCR-ABL kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib (STI571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer Cell 2, 117–125. Shtivelman, E., Lifshitz, B., Gale, R. P., and Canaani, E. (1985). Fused transcript of abl and bcr genes in chronic myelogenous leukaemia. Nature 315, 550–554. Simon, M. P., Pedeutour, F., Sirvent, N., Grosgeorge, J., Minoletti, F., Coindre, J. M., TerrierLacombe, M. J., Mandahl, N., Craver, R. D., Blin, N., Sozzi, G., Turc-Carel, C., O’Brien, K. P., Kedra, D., Fransson, I., Guilbaud, C., and Dumanski, J. P. (1997). Deregulation of the platelet-derived growth factor B-chain gene via fusion with collagen gene COL1A1 in dermatofibrosarcoma protuberans and giant-cell fibroblastoma. Nat. Genet. 15, 95–98. Sjoblom, T., Shimizu, A., O’Brien, K. P., Pietras, K., Dal Cin, P., Buchdunger, E., Dumanski, J. P., Ostman, A., and Heldin, C. H. (2001). Growth inhibition of dermatofibrosarcoma protuberans tumors by the platelet-derived growth factor receptor antagonist STI571 through induction of apoptosis. Cancer Res. 61, 5778–5783. Talpaz, M., Silver, R. T., Druker, B. J., Goldman, J. M., Gambacorti-Passerini, C., Guilhot, F., Schiffer, C. A., Fischer, T., Deininger, M. W. N., Lennard, A. L., Hochhaus, A., Ottmann, O. G., Gratwohl, A., Baccarani, M., Stone, R., Tura, S., Mahon, F.-X., Fernandes-Reese, S., Gathmann, I., Capdeville, R., Kantarjian, H. M., and Sawyers, C. L. (2002). Imatinib induces durable hematologic and cytogenetic responses in patients with accelerated phase chronic myeloid leukemia: Results of a phase 2 study. Blood 99, 1928–1937. Tuveson, D. A., Willis, N. A., Jacks, T., Griffin, J. D., Singer, S., Fletcher, C. D., Fletcher, J. A., and Demetri, G. D. (2001). STI571 inactivation of the gastrointestinal stromal tumor c-KIT oncoprotein: Biological and clinical implications. Oncogene 20, 5054–5058. van Oosterom, A. T., Judson, I., Verweij, J., Stroobants, S., Donato di Paola, E., Dimitrijevic, S., Martens, M., Webb, A., Sciot, R., Van Glabbeke, M., Silberman, S., and Nielsen, O. S. (2001). Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumours: A phase I study. Lancet 358, 1421–1423. Verweij, J., van Oosterom, A., Blay, J. Y., Judson, I., Rodenhuis, S., van der Graaf, W., Radford, J., Le Cesne, A., Hogendoorn, P. C., di Paola, E. D., Brown, M., and Nielsen, O. S. (2003). Imatinib mesylate (STI-571 Glivec, Gleevec) is an active agent for gastrointestinal stromal tumours, but does not yield responses in other soft-tissue sarcomas that are unselected for a molecular target. Results from an EORTC Soft Tissue and Bone Sarcoma Group phase II study. Eur. J. Cancer 39, 2006–2011. Virchow, R. (1845). Weisses Blut. Frorieps Notizen 36, 151–156. von Bubnoff, N., Schneller, F., Peschel, C., and Duyster, J. (2002). BCR-ABL gene mutations in relation to clinical resistance of Philadelphia-chromosome-positive leukaemia to STI571: A prospective study. Lancet 359, 487–491. Wyman, K., Atkins, M. B., Hubbard, F., McDermott, D., Kelley, M., Ko, Y.-L., Viar, V., Mier, J., Roberts, S., and Sosman, J. (2003). A phase II trial of imatinib mesylate at 800 mg daily in metastatic melanoma: Lack of clinical efficacy with significant toxicity. Proc. Am. Soc. Clin. Onc. 22, 713a.

30

Brian J. Druker

Yaish, P., Gazit, A., Gilon, C., and Levitzki, A. (1988). Blocking of EGF-dependent cell proliferation by EGF receptor kinase inhibitors. Science 242, 933–935. Zimmermann, J., Buchdunger, E., Mett, H., Meyer, T., Lydon, N. B., and Traxler, P. (1996). Phenylamino-pyrimidine (PAP) derivatives: A new class of potent and highly selective PDGF-receptor autophosphorylation inhibitors. Bioorg. Med. Chem. Lett. 6, 1221–1226. Zimmermann, J., Buchdunger, E., Mett, H., Meyer, T., and Lydon, N. B. (1997). Potent and selective inhibitors of the Abl-kinase: Phenylamino-pyrimidine (PAP) derivatives. Bioorg. Med. Chem. Lett. 7, 187–192.

Prostate Cancer and the Met Hepatocyte Growth Factor Receptor Beatrice S. Knudsen* and Magnus Edlund{ *Division of Public Health Sciences, Fred Hutchinson Cancer Research { Center, Seattle, Washington 98125; and Department of Urology Molecular and Therapeutics Program, Winship Cancer Institute, Emory University School of Medicine, Atlanta, Georgia 30322

I. Hepatocyte Growth Factor/Scatter Factor–Met and Stromal–Epithelial Interactions in the Prostate A. Introduction B. HGF/SF Expression in Prostate Stroma C. Met Expression in Normal Developing and Adult Prostate Epithelium II. Met in Localized Prostate Cancer A. Immunohistochemical Studies of Met Expression in Prostate Cancer B. Met and Cell–Matrix Adhesion in Prostate Cancer C. Met and Cell Surface Receptor-Linked Proteoglycans in Prostate Cancer D. Met and Cell–Cell Adhesion in Prostate Cancer III. Met in Prostate Cancer Metastasis A. Immunohistochemical Studies of Met Expression in Prostate Cancer Metastasis B. Met in Metastatic Prostate Cancer Cell Lines and Xenografts C. Converging Signaling Pathways between Met and the Androgen Receptor in Metastatic Prostate Cancer Cells IV. Therapeutic Implications of Met Expression in Prostate Cancer A. HGF/SF-Targeted Therapeutic and Imaging Agents B. Met-Targeted Therapeutic and Imaging Agents References

The hepatocyte growth factor/scatter factor (HGF/SF) and its receptor, the Met protein tyrosine kinase, form a classic ligand–receptor system for epithelial– mesenchymal communications in the normal and cancerous prostate. This review illustrates the expression and activities of HGF/SF and Met during prostate development, homeostasis, and carcinogenesis. The participation of HGF/SF in the morphogenetic program of rodent prostate development, the role of Met in normal human prostate epithelium, and underlying mechanisms of deregulated Met expression in localized and metastatic prostate cancer are discussed. On the basis of the commonly observed overexpression of Met in metastatic prostate cancer, HGF/SF–Met-targeted imaging and therapeutic agents can now be applied toward diagnosis and treatment. ß 2004 Elsevier Inc.

Advances in CANCER RESEARCH 0065-230X/04 $35.00

31


32

Beatrice S. Knudsen and Magnus Edlund

I. HEPATOCYTE GROWTH FACTOR /SCATTER FACTOR–MET AND STROMAL–EPITHELIAL INTERACTIONS IN THE PROSTATE A. Introduction Excessive and uncontrolled branching morphogenesis is a hallmark of carcinogenesis in the prostate and a diagnostic feature of prostate cancer (Bonkhoff, 2001; McNeal, 1965). The hepatocyte growth factor/scatter factor (HGF/SF)–Met ligand–receptor system constitutes a classic mechanism for stromal–epithelial cross-talk and has been implicated in the formation of an epithelial ductal system in the kidney, mammary gland, liver, pancreas, and lung (Brinkmann et al., 1995; reviewed in Birchmeier and Gherardi, 1998; Birchmeier et al., 2003; Rosario and Birchmeier, 2003; Rosen et al., 1994). The Met receptor kinase was discovered in the mid1980s by Vande Woude and colleagues through its oncogenic activity in NIH 3T3 cells (Cooper et al., 1984; Park et al., 1987) and later identified as the receptor for HGF/SF (Bottaro et al., 1991; Gherardi et al., 1993; Weidner et al., 1993). HGF/SF is confined to secretion by stromal cells, and stimulates tubulogenesis during organ regeneration of kidney and liver (Bell et al., 1999; Kawaida et al., 1994). Neutralizing HGF/SF antibodies inhibit differentiation, induce apoptosis, and perturb branching morphogenesis in an organ culture system of nephrogenesis (Woolf et al., 1995). In addition to its role during development, Met has assumed an important position in our understanding of carcinogenesis, invasion, and metastasis (Behrens et al., 1991; Rong et al., 1992, 1994). Autocrine or paracrine HGF/SF–Met causes tumor metastasis in animals (Jeffers et al., 1998, 1996c) and high levels of Met and Met-activating mutations are detected on a variety of human cancers in advanced or metastatic stages (Di Renzo et al., 2000; Lee et al., 2000a; reviewed in Maulik et al., 2002). One example consists of the Met-activating mutations in familial papillary renal cell carcinoma (Schmidt et al., 1999). The signal transduction pathways that govern the activity of the Met receptor are described in detail in reviews (Furge et al., 2000; Rosario and Birchmeier, 2003; Zhang and Vande Woude, 2003). In this review, we focus on the activity of the HGF/SF–Met system, which through stromal–epithelial interactions may mediate morphogenesis and homeostasis in the normal prostate and cell invasion, proliferation, and metastasis in cancerous prostate.

B. HGF/SF Expression in Prostate Stroma The most plausible mode of Met activation in normal and cancerous prostate epithelial cells is through a paracrine mechanism. Accordingly, immunohistochemical analysis demonstrates the presence of HGF/SF

Prostate Cancer and the Met Hepatocyte Growth Factor Receptor

33

protein throughout the prostate stroma, which is consistent with a paracrine effect on the adjacent epithelium, where the Met receptor is expressed (Gmyrek et al., 2001). Thus, HGF/SF is secreted by stromal cells and immobilized within stromal extracellular matrix, where HGF/SF is likely bound to heparan sulfate proteoglycans (Hartmann et al., 1998; Lamszus et al., 1996; Lyon et al., 1994; Rosen et al., 1989; Weidner et al., 1990). In addition, stromal smooth muscle exhibits weak Met immunoreactivity, implicating autocrine stimulation by HGF/SF. It is not clear which cells in the prostate stroma secrete HGF/SF in vivo. Likely candidates include prostate smooth muscle cells, myofibroblasts, and fibroblasts; vascular wall smooth muscle cells; and nerve sheath cells. In vitro-cultured prostate stromal cells, which consist of smooth muscle actin- and vimentin-positive myofibroblastic cells, secrete HGF/SF (Gmyrek et al., 2001; Nakashiro et al., 2000; Nishimura et al., 1999). These cells could well be a source of HGF/SF in vivo because of their proximity to the epithelium. Vimentin-positive myofibroblastic cells form a circumferential layer around normal epithelial glands, project into stalks of hyperplastic epithelial papillae, and are more abundant in areas of cancer (Tuxhorn et al., 2002). Could HGF/SF be an androgen-regulated stromal morphogenetic factor? Circumstantial evidence connects regulation of HGF/SF mRNA synthesis and bioactivity to the presence of steroid hormones (Blanquaert et al., 2000; Matsumoto et al., 1992; Skrtic and Ohlsson, 2000). Transcription of the HGF/SF gene is directly regulated by glucocorticoid and estrogen receptors, through consensus DNA-binding sequences in the HGF/SF promoter or within the first intron (Plaschke-Schlutter et al., 1995); however, direct regulation by the androgen receptor has so far not been reported. It is possible that HGF/SF synthesis is indirectly regulated by androgen through interleukin 6 (IL-6) and consensus motifs for the NF-IL-6 (C/EBP- ) transcription factor in the HGF/SF promoter (Akakura et al., 2003; Liu et al., 1994b; Ohira et al., 1996; Zarnegar, 1995). IL-6 synthesis increases under conditions of low androgen (Gornstein et al., 1999; Hofbauer et al., 1999; Zhang et al., 1998). In addition, the increase in circulating IL-6 levels causes a generalized cachexia in patients with advanced prostate cancer (Adler et al., 1999; Drachenberg et al., 1999; Twillie et al., 1995), and may directly stimulate the growth of prostate cancer cells, as is observed after IL-6 stimulation in vitro (Borsellino et al., 1999; Culig et al., 2002; Degeorges et al., 1996; Lee et al., 2003; Okamoto et al., 1997; Ritchie et al., 1997). IL6 also augments the sensitivity of the androgen receptor to androgen (Chen et al., 2000; Deeble et al., 2001; Hobisch et al., 1998; Lin et al., 2001; Lou et al., 2000; Ueda et al., 2002b). The increase in IL-6 under conditions of androgen deprivation could adversely result in elevated HGF/SF synthesis and secretion by the prostate or bone stroma, which may promote tumor growth and metastasis. It is, however, possible that HGF/SF is not regulated

34


by androgen, but requires the concerted activity of androgen-dependent factors to promote branching morphogenesis of benign or cancerous prostate epithelium in a temporally and spatially controlled fashion. The regulation of cell type-specific HGF/SF expression lies within an element of the mouse hgf/sf promoter that confers expression in mesenchymal, but not in epithelial, cells (Liu et al., 1994a). The murine and human hgf/sf promoters show 89.5% sequence identity in 453 nucleotides before the transcriptional start site (Liu et al., 1994b). In this region lie conserved regulatory elements for the C/EBP family of transcription factors and consensus sequences for AP-1 and SP-1 transcription factors, which are activated by many growth factors and cytokines, including IL-6 ( Jiang and Zarnegar, 1997; Jiang et al., 1997). In addition, there is a TGF- negative regulatory element, raising the possibility that HGF/SF activity is suppressed by TGF- (Liu et al., 1994b). Inhibition of HGF/SF synthesis by TGF- has been validated in cell lines and animals (Joseph et al., 1999; Ohira et al., 1996). Investigations in Chung Lee’s laboratory provided evidence that the effects of TGF- on ductal epithelial cells vary with their location within the ductal system of rodent prostate (Lee et al., 1999; Nemeth et al., 1997). In the proximal prostatic ducts, a thick muscular ring surrounds the epithelium, secreting TGF- . This epithelium shows frequent apoptotic cells. In the midportion of the duct, a single continuous layer of TGF- -positive stromal cells abuts a proliferating epithelium. In the distal portion of ducts, which undergo branching morphogenesis, sparse TGF- -positive stromal cells are adjacent to epithelial ducts lined by basal and secretory, nonproliferating epithelial cells. It is possible that the regional activity of TGF- is in part mediated through suppression or modulation of HGF/SF synthesis or activity (Santos and Nigam, 1993), which acts as a survival and differentiation factor for prostate epithelial cells. Thus, in the proximal prostate, copious amounts of TGF- prevent HGF/SF synthesis and thereby cause cell death, whereas in the distal prostate, TGF- is low and HGF/SF high, contributing to cell survival and branching morphogenesis (Lee et al., 1999). Another illustration of reciprocal regulation of TGF- and HGF/SF exists in inflamed prostatic tissues. Here, inflammation increases local concentrations of TGF- and stromal scarring and the epithelium becomes atrophic. Atrophic epithelial cells express abundant Met receptor, but are unable to differentiate (van Leenders et al., 2003). If HGF/SF is inhibited by TGF- the absence of terminal differentiation in atrophic epithelium could be attributed to the lack of stromal HGF/SF. HGF/SF is synthesized as a prepro single-chain polypeptide and is activated through cleavage by serine proteases. Matriptase and urokinase have emerged as two of the HGF/SF-activating serine proteases (Lee et al., 2000b; Mars et al., 1993; Naldini et al., 1992), matriptase also being an activator of urokinase. Hepatocyte growth factor activator inhibitor 1


35

(HAI-1) and HAI-2 are inhibitors of matriptase (Kataoka et al., 2003; Lin et al., 1999), and at least HAI-1 secretion is indirectly regulated by androgen (Martin et al., 2004) (Fig. 1). Therefore, androgen stimulation of HAI-1 secretion could lead to the inhibition of HGF/SF activation, whereas androgen deprivation would enhance the generation of active HGF/SF. In this new paradigm, androgen regulates the activation and not the synthesis of a stromal growth factor.

C. Met Expression in Normal Developing and Adult Prostate Epithelium 1. PROSTATE DEVELOPMENT, BRANCHING MORPHOGENESIS, AND HGF/SF–MET Branching morphogenesis during development of the ductal system involves invasion of basal epithelial cells into the prostate stroma. It is generally believed that regulatory components of stromal invasion, active during normal prostate development, become deregulated during oncogenesis and that this leads to tumor invasion and metastasis. The human fetal prostate consists of solid cords of basal epithelial cells embedded in an immature mesenchyme (Leav et al., 1999; Wang et al., 2001). Branching occurs during puberty, triggered by the rise in androgen and the expression of androgen receptors in the prostate stroma. Ductal canalization and the formation of a differentiated secretory cell layer are initiated in proximal ducts and proceed distally (Marker et al., 2003). This spatially organizes the growing ducts (Fig. 1). The ductal tips harbor basal epithelial cells, even in adulthood (Xue et al., 2001), which express the Met receptor and are highly responsive to induction by the prostate stroma. Met likely participates in the stromal invasion at the tip of prostatic buds by inhibiting proliferation, initiating an epithelial-to-mesenchymal transition and inducing the expression of proteolytic activity in invading basal epithelial cells (Tsarfaty et al., 1992). These changes are mediated by direct Met-triggered cytoskeletal reorganizations (Ridley et al., 1995), stimulation of urokinase secretion and expression of urokinase receptors ( Jeffers et al., 1996a; Nishimura et al., 2003; Vande Woude et al., 1997), and synthesis of matrix metalloproteinases (MMPs) (McCawley et al., 1998; Rosenthal et al., 1998). In human and dog prostate, we observe expression of the Met receptor along the entire length of prostatic ducts. Although it is likely that the HGF/ SF–Met system is part of a complex array of factors that regulate epithelial proliferation, survival, differentiation, and branching morphogenesis, it is difficult to assess the contribution of HGF/SF–Met to prostatic ductal development in vivo. While the induction of androgen-regulated genes in


37

rodents begins on embryonic day 14.5 (E14.5), budding of prostatic ducts only commences on E17.5 (Corpechot et al., 1981), which is after the death of HGF/SF or Met knockout mice between E13 and E16 (Schmidt et al., 1995). A conditional knockout of Met in the prostate, to overcome this problem, has not been generated. In a conditional knockout using a basal cell promoter, Met could be eliminated before shrinking the prostate through androgen withdrawal to the basal/stem cell compartment. Ensuing androgen replenishment reconstitutes prostates that appear architecturally normal. In such a system the role of Met in branching morphogenesis could be analyzed in vivo, in conjunction with growth factors and cytokines that have a documented role in prostate development. Several growth factors and cytokines influence prostatic development in vivo. In the insulin-like growth factor I (IGF-I) knockout, branching morphogenesis is reduced in the prostate (Ruan et al., 1999; van Leenders et al., 2003). Furthermore, overexpression of prolactin causes increased ductal branching, despite normal androgen levels (Kindblom et al., 2003). Although the regional distributions of IGF-I and prolactin along the prostatic ducts are not known, the morphogenetic sonic hedgehog (Shh) localizes to the distal ductal epithelium during prostate development (Barnett et al., 2002) (Fig. 1). Expression of Shh in prostate epithelial cells, testosterone surge, and branching morphogenesis all coincide with expression of Nkx3.1, a specific marker of prostate epithelial differentiation (Schneider et al., 2000). However, the roles of Shh during prostate development are still controversial. Shh is not necessary for induction of prostate buds, but is required for complete branching morphogenesis of the neonatal ventral prostate (Freestone et al., 2003). Inhibitory antibodies against Shh reduced branching morphogenesis, when urogenital sinus epithelium was implanted underneath the renal Fig. 1 Mechanisms regulating branching morphogenesis in the prostate. Branching morphogenesis involves spatially and timely regulated proliferation, invasion, and differentiation of prostate epithelial cells. Met is expressed in basal epithelial cells along the entire length of the prostatic ducts. Met is also present at the tips of the ducts, which extend into the prostate stroma during ductal growth. Activation of Met by stromal HGF/SF triggers expression of urokinase and urokinase receptor, increases integrin expression and avidity, and causes remodeling of the cytoskeleton necessary for cell motility (epithelial–mesenchymal transition). HGF/SF is synthesized in a latent proform and requires activation by proteolytic cleavage. This is accomplished by the serine proteases urokinase and matriptase. Secretion of the matriptase inhibitor HAI-1 by prostate epithelial cells is regulated by androgen. Met is also expressed in stromal smooth muscle cells, which are generated through TGF- -induced differentiation of myofibroblastic cells. During ductal morphogenesis, Met-triggered signal transduction events may converge with signals from other growth factor and adhesion receptors that are expressed on the same cells. These can include growth factor receptors for fibroblast growth factors (FGF), insulin-like growth factors (IGF), TGF- /activin, and bone morphogenetic proteins (BMPs). In addition, epithelial–stromal interactions that are mediated by sonic hedgehog/ patched can potentially modulate HGF/SF–Met activities.

38


capsule of adult male mice for differentiation (Podlasek et al., 1999). In the presence of testosterone, inhibition of Shh-signaling resulted in aberrant differentiation as assessed by expression of the basal cell markers p63 and K14, and a histologic phenotype resembling cribriform prostatic intraepithelial neoplasia (PIN) (Freestone et al., 2003). The Shh receptor, patched, is expressed in the stroma, implicating a reverse stromal–epithelial interaction that may regulate ductal branching morphogenesis (Lamm et al., 2002). Consequently, Shh-induced stromal growth (Freestone et al., 2003) and secretion of TGF- and activin A by prostate stromal cells inhibited proliferation and stimulated differentiation of epithelial cells in the proximal portion of the duct (Wang et al., 2003). Together, these studies suggest that the reciprocal activity of Shh–patched ligand–receptor system is highly regulated in a spatial and temporal fashion during prostate development and in the adult prostate. Urokinase and the urokinase receptor, which are induced by HGF/SF (Jeffers et al., 1996b), have been implicated in prostatic development, because peptides that disrupt the interaction between urokinase and its receptor cause apoptosis and prevent differentiation of epithelial cells (Elfman et al., 2001). The hyaluronan receptor CD44 also interacts with Met and plays an essential role in ductal morphogenesis in the prostate (Gakunga et al., 1997). Although the full spectrum of activity of the HGF/ SF–Met system remains obscure in vivo, in vitro studies with basal prostate epithelial cells clearly demonstrate the migration-inducing function of HGF/ SF–Met. In Boyden chamber assays, migration of basal prostate epithelial cells is strongly induced by the conditioned media of primary prostate stromal cells; these media contain concentrations of HGF/SF between 6 and 13 ng/ml (Gmyrek et al., 2001; Nishimura et al., 1998, 1999, 2003). Stimulation of migration depends on the concerted activities of cell adhesion proteins and HGF/SF (You et al., 2003). Whereas immobilized proteins from the conditioned medium mediate adhesion, spreading, and planar polarization of epithelial cells, HGF/SF is required for triggering cell migration. Consequently, neutralization of HGF/SF activity with a combination of three monoclonal antibodies (Cao et al., 2001) strongly inhibited cell migration, and decreased activation of the phosphatidylinositol 3-kinase (PI3-kinase)– Akt pathway (You et al., 2003).

2. HGF/SF–MET SIGNALING IN STROMAL–EPITHELIAL INTERACTIONS Activation of the Met receptor through dimerization recruits numerous cytoplasmic signal transduction proteins and thereby stimulates cell proliferation, survival, reorganization of the actin cytoskeleton, and proteolysis. Central to this activation and to the ability of Met to induce branching morphogenesis are two phosphorylated tyrosine residues (Y1349 and


39

Y1356) in the C terminus of Met that bind Grb2, p85/PI3-kinase, and Src as well as the association of Met with the scaffolding protein Gab1, a member of the insulin receptor substrate 1 family of proteins (Birchmeier et al., 2003; Gu and Neel, 2003; Ponzetto et al., 1994; Sachs et al., 1996; Weidner et al., 1996). A third tyrosine residue in the juxtamembrane domain binds c-Cbl, which possesses ubiquitin ligase activity and targets Met for degradation via the ubiquitination machinery (Fixman et al., 1997; Fournier et al., 2000). Mutations that induce constitutive tyrosine kinase activity or prevent c-Cbl binding are oncogenic (Lee et al., 2000a; Peschard et al., 2001; Schmidt et al., 1997). In Madin-Darby canine kidney cells, a commonly utilized tissue culture model for HGF/SF-induced branching morphogenesis, a complex between the Grb2 SH2 domain bound to Met-Y1356 and the Grb SH3 domain associated with Gab1 stabilizes Met–Gab1 interactions (Nguyen et al., 1997). PI3-kinase activity is initially activated through binding of p85 to Met-Y1356, thereby generating the necessary phosphoinositol 3,4,5-triphosphate for membrane attachment of Gab1 via its pleckstrin homology domain and further p85/PI3-kinase–Gab1 complexes (Kamikura et al., 2000). In addition, phosphorylation of Gab1 generates binding sites for SH2 or PTB domain-containing proteins, such as the SHP2 tyrosine phosphatase, phospholipase C , Crk/CrkL, and Shc. SHP2 primarily connects Met–Gab1 to the Ras–mitogen-activated protein kinase (MAPK) pathway (Kamikura et al., 2000) and Crk/CrkL link Met–Gab1 to the cytoskeletonremodeling GTPases Rac and Rap1 (Garcia-Guzman et al., 2000; Lamorte et al., 2000). These pathways also provide converging points for signals induced by cell–cell or cell–matrix adhesion, such as those triggered by the morphogenetic fibroblast growth factor receptors and by the TGF- and bone morphogenetic protein/activin receptors (Bowes et al., 1999; Ohmichi et al., 1998; Pollard, 2001; Sakurai and Nigam, 1997; Soriano et al., 1998; Zeng et al., 2001). The cross-talk with other cell surface receptors regulates the temporal and spatial organization of the Met signal inside Met-expressing cells and may contribute to the sustained activation of signal transduction pathways, such as the Grb2–Ras–MAPK pathway, that is necessary for branching morphogenesis (Ishibe et al., 2003; Maroun et al., 2000; Ueland et al., 2004). Met-induced signal transduction pathways have been reviewed in great detail (Birchmeier et al., 2003; Furge et al., 2000; Rosario and Birchmeier, 2003; Zhang and Vande Woude, 2003).

3. MET AND EPITHELIAL DIFFERENTIATION IN THE PROSTATE In the prostate epithelium, three compartments are defined by cytokeratin expression (Okada et al., 1992). High molecular weight (HMW) cytokeratin (K)-expressing basal epithelial cells form a continuous cell layer

40


adherent to the basement membrane. Differentiated prostate epithelial cells compose the lumenal cell layer and express low molecular weight (LMW) cytokeratins. A third compartment consists of intermediate/transiently proliferating epithelial cells expressing both HMW and LMW cytokeratins and contributing to both the basal and lumenal cell layers. These three compartments are distinguished by their localization within the epithelium, cellular morphology, immunophenotype, and proliferative capacities. It is generally believed that stromal factors stimulate differentiation of basal into secretory cells, transitioning through the intermediate/transiently proliferating compartment (Fry et al., 2000; Lang et al., 2001; Robinson et al., 1998). This review primarily focuses on the regulation of Met expression during differentiation of prostate epithelial cells. A detailed description of events associated with epithelial differentiation is provided in a review by De Marzo et al. (1998). Met is expressed in basal and intermediate prostate epithelial cells, whereas secretory cells are largely negative for Met expression (van Leenders et al., 2002). Basal cells are identified by their expression of p63, HMW cytokeratins K5 and K14, in combination with low levels of LMW K18 (K5þþ/K14þþ/K18þ) (Garraway et al., 2003; Signoretti et al., 2000; van Leenders et al., 2000; Weinstein et al., 2002). In vitro studies demonstrate that basal cells (K5þþ/K14þþ/K18þ) are progenitors of intermediate and secretory cells, thus supporting a hierarchical differentiation model proposed by Isaacs and Coffey (1989). According to this model, a basal cell divides asymmetrically, giving rise to a new basal cell and an intermediate cell. Intermediate cells do not express K14, amplify rapidly, and terminally differentiate into secretory cells. Basal cells express no or low levels of androgen receptor (AR) protein. On androgen deprivation they paradoxically proliferate, further demonstrating their lack of androgen dependence. In addition, growth factor receptors and cell–matrix adhesion molecules decorate the basal cell surface, whereas they are largely absent from secretory cells. Although Met is strongly expressed in the basal epithelial layer, interestingly, not all cells stain positively for Met (Fig. 2) (van Leenders et al., 2002). The exact nature of the Met-negative basal cells remains to be investigated, and it is conceivable that they may represent neuroendocrine cells, or else belong to the small population of prostate stem cells that is likely to exist within the basal cell layer. De Marzo and van Leenders identified K5þþ/K18þ intermediate/transiently proliferating cells in the basal, as well as in the lumenal cell layers, and named them ‘‘basal intermediate’’ and ‘‘secretory intermediate’’ cells (van Leenders et al., 2003). Intermediate cells are enriched in the lumenal cell layer in proliferative inflammatory atrophy (PIA) (De Marzo et al., 1999b), express less AR protein than do secretory cells, and are negative for the cyclin-dependent kinase inhibitor p27. Thus cell proliferation is greatly increased in the lumenal cell layer in PIA and restricted to K5þþ/


41

Fig. 2 Met expression in normal prostate epithelium. (A) Formalin-fixed, paraffin-embedded tissue section. Met protein expression in basal and intermediate prostate epithelial cells is visualized by dark brown staining. Met is expressed in the majority of basal epithelial cells and in intermediate epithelial cells. Note a Met-negative basal epithelial cell (arrowhead). In this section, Met is not expressed in secretory epithelial cells. The myofibroblastic cells in the prostate stroma do not express Met. (B) Schematic representation of epithelial differentiation in the prostate. Properties and major phenotypes of cells in the basal, intermediate, and secretory compartments are shown.

K18þ intermediate cells. Met is expressed in intermediate cells in the prostate epithelium and marks intermediate epithelial cells in between secretory epithelial cells (Fig. 2). Secretory cells are readily distinguished from basal cells by the absence of detectable K5/K14 expression and high expression of K8 and K18. They are not adherent to the basement membrane and do not proliferate. Expression of Met in secretory cells is controversial, possibly because Met immunoreactivity is sensitive to tissue fixation. In Met-positive secretory cells, the

42


staining appears at the basolateral cell surface, whereas Met staining in basal cells is distributed throughout the cytoplasm. In 18% of benign prostatic hyperplasia cases, secretory cell expression of Met is limited to the central zone (Pisters et al., 1995). However, other studies observing Met in secretory cells do not describe regional variability of expression (Humphrey et al., 1995; Watanabe et al., 1999).

II. MET IN LOCALIZED PROSTATE CANCER A. Immunohistochemical Studies of Met Expression in Prostate Cancer Because cancer cells display a strong phenotypic resemblance to secretory epithelial cells (which express LMW cytokeratins and AR), and at the same time express the basal/intermediate epithelial markers such as Met, syndecan-1, and CD44, the cellular origin of malignant transformation in the prostate remains unclear. In three of four immunohistochemical studies (Table I), Met was expressed in approximately 50% of localized prostate cancers and in almost all prostate cancer metastases (Humphrey et al., 1995; Knudsen et al., 2002; Watanabe et al., 1999). In a fourth study, Met was found in 84% of 43 cases of localized prostate cancers (Pisters

Table I

Immunohistochemical Studies of Met Expression in Prostate Cancer

Met expression Benign epithelium Basal cells Lumenal cells BPH Atrophy Cancer PIN Locally invasive carcinoma: % Met positive (no. of cases) Correlation with grade Metastasis: % Met positive (no. of cases)

Humphrey et al. (1995)

Positive Positive

Pisters et al. (1995)

Positive

Positive Central zone positive 18% positive 100% positive

33% 45% (108)

84% (36)

No 75% (20)

Yes 93% (30)

Watanabe et al. (1999)

Positive Positive

Knudsen et al. (2002)

Positive Negative

Positive 50% 33% latent (15); 71% clinical (14) Yes 100% (7)

Abbreviations: BPH, benign prostatic hyperplasia; PIN, prostatic intraepithelial neoplasia.

51% (90)

No 100% (86)


43

et al., 1995). In a study of 108 cases by Humphrey et al., 45% of primary prostate cancers expressed Met, but there was no correlation with Gleason grade (Humphrey et al., 1995). There was 33% concordance in Met expression between PIN and invasive carcinoma. Of 20 cases of metastatic prostate cancer, 15 expressed Met. These authors observed cytoplasmic and cell surface Met in human secretory cells. In parallel studies in rats, Met, epidermal growth factor, and transforming growth factor- receptors were all increased when animals were androgen deprived. Watanabe et al. analyzed a cohort of 36 patients. These investigators reported Met expression in 33% of latent and in 71% of clinically significant prostate cancers (Watanabe et al., 1999). In this cohort, 38% of low-grade and 80% of high-grade cancers were Met positive. In addition, seven of seven metastases expressed the Met receptor. In our cohort of 90 cases limited to Gleason Sum 6 and 7 prostate cancers, 50% of cancers expressed the Met receptor (Fig. 3A) (Knudsen et al., 2002). However, there was no correlation between Met expression and disease progression, as measured by prostatespecific antigen (PSA) recurrence within a 5-year follow-up period. The analysis also included 86 prostate cancer metastases from individual patients, all of which expressed the Met receptor. Met expression was greater in bone than in lymph node metastasis. Because the treatment status of these patients was not known, it is questionable whether the higher Met expression in bone was a result of androgen deprivation. Although we did not find a correlation between Met expression and PSA recurrence in the Gleason sum 6 or 7 tumors, it would be worthwhile to test a correlation with tumor progression in high-grade prostate cancer with a Gleason sum of 8 or more.

B. Met and Cell–Matrix Adhesion in Prostate Cancer Overexpression of receptor tyrosine kinases, such as Met, may constitutively activate the kinase, as could cross-talk with adhesion receptors or other growth factor receptors. This could enhance the HGF/SF signal, but may also have HGF/SF-independent effects. One example of such a mechanism was identified when the human Met gene was expressed in hepatocytes, under the control of tetracycline (Wang et al., 2001). Induction of high levels of human Met, which is not activated by murine HGF/SF, led to the development of hepatocellular carcinomas. Subsequent inactivation of the transgene resulted in regression of the tumors. Interestingly, the overexpressed, active Met was refractory to stimulation by HGF/SF. Adhesiondependent activation of the Met receptor kinase has also been observed in cell lines, where it depends on high Met expression levels (Wang et al., 1996). This mechanism of Met activation may be important in human prostate cancer metastasis, where high levels of Met are present.

44


Fig. 3 Met expression in invasive prostate cancer and prostate cancer metastasis to bone. (A) Invasive prostate cancer. The tumor glands consist of a single-cell layer of Met-positive cells, whereas neighboring normal glands show Met expression only in the basal cell layer. (B) Effects of androgen ablation on HGF/SF secretion from bone stromal cells. Bone marrow stromal cells and osteoblastic cells secrete HGF/SF. HGF/SF is transcriptionally regulated by glucocorticoid hormones and IL-6. Therefore, under conditions of androgen ablation, HGF/SF synthesis could be stimulated as a result of increased IL-6 concentrations in the tumor cell environment. The sources of IL-6 are either bone stromal cells or metastatic cancer cells and IL6 secretion is increased on androgen deficiency. (C) Prostate cancer metastasis to bone. Tumor cells as well as bone marrow stromal cells and osteoblasts demonstrate Met positivity. (D) Schematic representation of several converging signal transduction pathways and the androgen receptor in metastatic prostate cancer cells. Prostate cancer cells express multiple growth factor and cell adhesion receptors, including Met, ErbB2/Her2/Neu (Her2), IL-6 receptor (IL-6/gp130), and the insulin growth factor receptor (IGFR). These receptors, directly and indirectly through cytoplasmic tyrosine kinases, the SHP-2 phosphatase, and STAT-3, activate cytoplasmic serine/threonine kinase cascades, such as MAPK and PKA/C, that phosphorylate the AR, AR coactivators, and transcription factors. In addition, the transcriptional activity of the AR is modulated through active Rho-GTPase and the Rho-activated serine/ threonine kinases PRK1/2. The effects of growth and adhesion factor signaling pathways reduce the androgen requirement of the androgen receptor and stimulate the progression of prostate cancer cells to androgen independence.


45

Many adhesion activities are altered in cells during cancerous progression, and are likely to modulate Met activity. Alterations in integrin expression levels, molecular affinities, and subunit heterodimerization choices are likely behind the ability of HGF/SF to trigger cell motility (Beviglia and Kramer, 1999; Matsumoto et al., 1994; Qiao et al., 2000; Trusolino et al., 1998, 2000; Weimar et al., 1997). In three progressively tumorigenic breast cell lines an increase in Met expression and phosphorylation occurred on integrin activation by osteopontin (Tuck et al., 2000). The relevance of shifts in integrin subunit usage during cancerous progression in vitro (Edlund et al., 2001; Zheng et al., 2000) awaits further studies in animal models, where roles in Met-induced carcinogenesis may be revealed. Because cytoplasmic signal transduction pathways emerging from the Met receptor may converge with downstream pathways of integrins (Trusolino et al., 2001), it is likely that the Met signal differs between basal prostate epithelial cells expressing integrins 2, 3, 6, 1, and 4 and locally invasive prostate cancer cells expressing either integrins 6 1 (68% of prostate cancers) (10% of prostate cancers) (Cress et al., 1995; Davis et al., 2001; Nagle et al., 1995; Schmelz et al., 2002; Witkowski et al., 1993). We find that in primary basal epithelial cells, HGF/SF stimulates growth arrest and prolonged MAPK activation and differentiation, whereas in DU145 prostate cancer cells, HGF/SF triggers a transient MAPK signal and cell proliferation. Met activation may therefore switch from a growthinhibitory to a growth-stimulatory signal during prostate carcinogenesis or during progression to metastatic prostate cancer (Gmyrek et al., 2001). HGF/SF treatment of C4-2 prostate cancer cells (androgen-independent, bone metastatic, LNCaP lineage cells) results in both increased adhesion and altered matrix-specific cell migration, such that cells decrease motility on laminin I but not on fibronectin (our unpublished observation). HGF/SF treatment also results in rapid and transient change of focal adhesion kinase (FAK) phosphorylation state in attached C4-2 cells, which has been reported in other in vitro systems (Beviglia and Kramer, 1999; Chen et al., 1998; Lai et al., 2000; Matsumoto et al., 1994). Reciprocally, HGF/SF stimulation can alter integrin expression. In Madin-Darby canine kidney (MDCK) cells, HGF activation of MAPK is followed by increased expression of 2 integrin (Liang and Chen, 2001). Likewise, C4-2 cells of the LNCaP prostate cancer progression model, express 2 1 (a collagen I- and laminin I-binding integrin) as they assume metastatic phenotypes (Edlund et al., 2001), as do metastatic carcinomas in vivo (Bonkhoff, 1998; Lang et al., 1997). This HGF/SF-stimulated increase in expression of 2, however, cannot solely account for all HGF/SFinduced cell scattering and motility, as attempts to block the integrin 2 with specific antibodies reduce cell scatter by only 40% (Liang and Chen, 2001). About 1% of basal epithelial cells express high levels of integrin 2. Could

46


these represent a stem cell population in the basal cell compartment (Collins et al., 2001)?

C. Met and Cell Surface Receptor-Linked Proteoglycans in Prostate Cancer Cell surface-expressed heparan sulfate proteoglycans are predominantly linked to CD44, syndecan, glypican, or versican protein cores (Jackson, 1997). HGF/SF binds heparan sulfate proteoglycans with its second kringle domains and the N-terminal hairpin loop (Chirgadze et al., 1999; Mizuno et al., 1994). In contrast to basic fibroblast growth factor (bFGF), which critically depends on heparan sulfate proteoglycans for binding to the FGF receptor, the association between HGF/SF and heparan sulfate does not appear to increase the affinity for Met (Harmer et al., 2003). In in vitro experiments, HGF/SF immobilized by heparan sulfate possesses increased strength for receptor phosphorylation and mitogenicity (Zioncheck et al., 1995). In addition, heparan sulfate may modify the HGF/SF signal by altering HGF/SF presentation to the Met receptor (Hartmann et al., 1998). However, in vivo, heparan sulfates appear to merely serve as low-affinity receptors for HGF/SF, sequestering HGF/SF, and thereby increasing its local concentration. Membrane targeting of the oncogenic tpr–Met fusion protein, through which the Met receptor kinase was originally cloned (Cooper et al., 1984; Park et al., 1986), induced the synthesis and secretion of hyaluronic acid (HA) and, in addition, the HA receptor CD44. These events could be inhibited by a pharmacological inhibitor of PI3-kinase. Furthermore, the multisubstrate adapter protein Gab1, which connects the Met receptor to PI3-kinase, enhanced Met receptor-dependent HA synthesis. This study points to a role for HA and CD44 in Met receptor-mediated oncogenesis and implicates PI3-kinase in these events (Kamikura et al., 2000). In Namalwa Burkitt’s lymphoma cells, CD44v3 bound HGF/SF and was necessary for Met activation by HGF/SF (van der Voort et al., 1999) and in a colon cancer cell line, CD44 interacted with Met by inducing clustering of the Met receptor and possible Met autophosphorylation (Wielenga et al., 2000). Several immunohistochemical studies have analyzed expression of overall CD44, CD44s, and CD44v3 expression in prostate cancer. Early prostate cancer research established a reverse correlation for total CD44 expression and poor prognosis (Kallakury et al., 1996; Nagabhushan et al., 1996; Noordzij et al., 1997), which was contradicted by later analyses of CD44 isoforms, associating CD44v3 and adverse outcome (Aaltomaa et al., 2000). In a multiple myeloma cell line, syndecan-1 expression prolonged HGF/SFinduced MAPK phosphorylation (Derksen et al., 2002). Syndecan-1 and -4 are two members of the syndecan family that are primarily expressed on


47

epithelial cells. Syndecan-1 expression has been correlated with PSA progression and cancer-specific survival in a cohort of 550 patients (Zellweger et al., 2003). In a smaller study, we found that syndecan-1 expression was predictive of progression in Gleason sum 7 prostate cancer (Chen et al., 2004). This is surprising, because in several other cancers, syndecan-1 expression carries a favorable prognosis (Anttonen et al., 1999, 2001; Inki et al., 1994a,b; Matsumoto et al., 1997; Nackaerts et al., 1997; Toyoshima et al., 2001; Wiksten et al., 2000). However, in our cohort, syndecan-1 and Met expression were not correlated, and syndecan-1 was expressed in only one-third of prostate cancer metastases, whereas Met was expressed in all.

D. Met and Cell–Cell Adhesion in Prostate Cancer The active Met receptor causes the breakdown of cell–cell adhesions, which is necessary for cells to escape and travel from the tumor mass to distant sites (Birchmeier et al., 1997; Meiners et al., 1998; Nishimura et al., 1998, 1999). Met can directly regulate E-cadherin expression and bind -catenin (reviewed in Birchmeier et al., 1996). Specifically, Met immunoprecipitated with E-cadherin and - and -catenins, components of the cell– cell adhesion system, whose loss accompanies tumor progression (Hiscox and Jiang, 1999). Met binding to -catenin results in -catenin phosphorylation and nuclear translocation, which is considered a prooncogenic mechanism (Monga et al., 2002). HGF/SF also triggers release of MMP-7 (matrilysin), which cleaves E-cadherin (Davies et al., 2001). In prostate cancer, decreased E-cadherin expression correlates with the presence of concurrent lymph node metastases, as well as with tumor progression, as determined by PSA recurrence (Cheng et al., 1996a; De Marzo et al., 1999a; Kallakury et al., 2001; Koksal et al., 2002; Loric et al., 2001; Ross et al., 1994; Rubin et al., 2001; Ruijter et al., 1998). It is therefore possible that Met facilitates prostate cancer metastasis by downregulating E-cadherin and by activating -catenin.

III. MET IN PROSTATE CANCER METASTASIS A. Immunohistochemical Studies of Met Expression in Prostate Cancer Metastasis In four separate immunohistochemical studies, Met expression occurred in 75–100% of prostate cancer metastasis (Table I) (Humphrey et al., 1995; Knudsen et al., 2002; Pisters et al., 1995; Watanabe et al., 1999). Using a

48


tissue microarray (TMA) of prostate cancer metastases from 86 individuals, in which all samples are displayed and stained on the same slide for comparison of relative Met expression levels, we found higher Met expression in metastases to bone than in lymph node metastases (Knudsen et al., 2002). Met was expressed on the surface and in the cytoplasm of metastatic prostate cancer cells (Fig. 3C). Although the biologic consequences of differences in Met expression are not known, the presence of Met in most metastasis sites provides an incentive for the development of Met-targeted therapies. However, a better assessment of the amounts of Met cell surface expression and Met activation in vivo is desirable.

B. Met in Metastatic Prostate Cancer Cell Lines and Xenografts Among the commonly used tissue culture cell lines derived from metastatic prostate carcinoma, both DU145 and PC3 express Met at high levels (Table II) (Gmyrek et al., 2001; Humphrey et al., 1995; Nagakawa et al., 2000; Nakashiro et al., 2000; Nishimura et al., 1998, 1999, 2003; Qadan et al., 2000; van Leenders et al., 2002), but only the DU145 cell line shows a concentration-dependent response to HGF treatment, in the form of increased cell motility in scatter and invasion assays (Humphrey et al., 1995). We observed that PC3 cells respond to HGF/SF only weakly, in terms Table II

Expression Pattern of Met in Commonly Used Prostate Cancer Cell Lines and

Xenografts Cell line or xenograft

Androgen responsive

Origin

LNCaP LuCaP LAPC-4 CWR22 C4-2 DU145 PC3 ALVA-31/41 PC3-M ARCaP

Yes Yes Yes Yes No No No No No Yes

Lymph node Lymph node Prostate Prostate LNCaP Brain Bone PC3 PC3 Ascites fluid

HGF Met responsive mRNA No

Yes Yes Yes Yes

No/low No No No Low Yes Yes Yes Yes

Met PSA protein producing No No No No No Yes Yes Yes Yes Yes

Yes Yes Yes Yes Yes No No No No Yes

Ref.a 1–3 2, 4, 5 2, 4, 6 2, 4, 7, 8 9, 10 1, 4, 11 1, 12 13–15 16 17

Abbreviations: HGF, hepatocyte growth factor; PSA, prostate-specific antigen. aReferences: (1) Humphrey et al. (1995); (2) Knudsen et al. (2002); (3) Horoszewicz et al. (1980); (4) Gmyrek et al. (2001); (5) Ellis et al. (1996); (6) Klein et al. (1997); (7) Wainstein et al. (1994); (8) Cheng et al. (1996b); (9) Wu et al. (1994); (10) Edlund et al. (2004); (11) Stone et al. (1978); (12) Kaighn et al. (1979); (13) Loop et al. (1993); (14) Qadan et al. (2000); (15) Nakhla and Rosner (1994); (16) Dong et al. (1999); (17) Zhau et al. (1996).


49

of migration, probably because signaling pathways such as the PI3-kinase– Akt pathway are efficiently activated by cell–substrate adhesion and cannot be further enhanced by HGF/SF stimulation. This is not the case in DU145 cells, in which HGF/SF considerably enhances Akt phosphorylation. Although we found that HGF/SF stimulated proliferation of DU145 cells, Qadan et al. reported growth inhibition in a growth factor environment that was significantly different from ours (Qadan et al., 2000). In a PC3 xenograft model, injection of peritumoral HGF/SF significantly increased tumor size (Nakashiro et al., 2000). An increase in tumor size was also achieved by concurrent implantation of prostate stromal cells; also, HGF/SF secretion from stromal cells was necessary for PC3 growth in a collagen gel. Contradictory reports of anti- and proapoptotic responses to HGF treatment, in different cell lines, have led to arguments that cell lines vary in their downstream signaling repertoires, with some lines lacking necessary components or releasing different amounts of matrix-degrading enzymes (Nishimura et al., 1998). To further complicate matters, both PC3 and C4-2 prostate cancer cell lines respond to HGF in a concentrationdependent matter, as assayed by cell spreading, but response of C4-2 cells occurs in the complete absence of detectable levels of Met receptor (Edlund et al., 2004). The adhesion-stimulating activity of HGF/SF in C4-2 cells might suggest a Met-independent mechanism. Met is undetectable in C4-2 cells at both protein and RNA levels, and the MAPK and Akt pathways are not activated in C4-2 cells on HGF/SF stimulation. Finally, matrix specificity appears to play a role in this complex mechanism promoting cellular adhesion. In contrast to PC3 and DU145 cells, other commonly used cancer cell lines (Table II), which express the androgen receptor, are negative for Met on Western blots, including LNCaP, LuCaP, LAPC-4, and CWR22 (Knudsen et al., 2002). Thus, there appears to be an inverse relationship between Met and androgen receptor expression in normal prostate epithelium, which may be maintained in most ex vivo xenografts and prostate cancer cell lines derived from advanced prostate cancers (Table II). In contrast, locally invasive and metastatic prostate cancers express both Met and the AR, and signal transduction pathways from the Met receptor may affect the transcriptional activity of the AR.

C. Converging Signaling Pathways between Met and the Androgen Receptor in Metastatic Prostate Cancer Cells Prostate cancer cells express Met and other growth factor receptors as well as the androgen receptor (AR). Several growth factor receptors have been shown to hyperactivate the androgen receptor in the presence

50


of low levels of androgen. These include the Her2 and IL-6 receptors (Chen et al., 2000; Craft et al., 1999; Ueda et al., 2002a; Yoshiura et al., 1995). Because AR transcription is modulated through signal transduction pathways from these cell surface receptors, it is conceivable that the Met receptor, which also activates these pathways, could also affect AR transcription (Fig. 3D). The N-terminal trans-activation domain of the AR contains several serine and threonine amino acid residues that are targets for cytoplasmic serine/threonine kinases. One of these kinases is the mitogen-activated protein kinase (MAPK), which is activated by many growth factor receptors, including the Met receptor, the gp130 subunit of the IL-6 receptor, and the TGF- receptor, as well as cell adhesion, in some cell lines (Culig et al., 2002; Heinlein and Chang, 2002; Ueda et al., 2002a). Not only does expression of phospho-MAPK correlate with disease progression to androgen independence, but phosphorylation of the AR by MAPK also reduces the androgen requirement (Gioeli et al., 1999; Ueda et al., 2002b). In contrast to MAPK, protein kinase A (PKA) phosphorylates the AR on Ser-650, in the hinge region between the DNA- and hormonebinding domains (Gioeli et al., 2002). Efficient phosphorylation of the AR requires androgen either to induce a conformational change in the AR or to activate serine/threonine kinases (through AR-independent mechanisms) that subsequently phosphorylate the AR (Gioeli et al., 2002; Heinlein and Chang, 2002). The active serine/threonine kinase Akt raises expression of AR protein and both activated Akt and AR independently inhibit the proapoptotic forkhead transcription factor (FKHR), thereby promoting cell survival (Li et al., 2003; Manin et al., 2002). Therefore, when activated by growth factor receptors, PI3-kinase and Akt promoted the survival of prostate cancer cells under conditions in which the AR is inactive (Lin et al., 2001; Wen et al., 2000) and when activated by Ras, these kinases prevented apoptosis of the androgen-dependent LNCaP xenograft in castrated mice (Bakin et al., 2003). In addition to directly phosphorylating the AR, serine/threonine kinases activate transcription factors and AR coactivators through phosphorylation, which enhance the transcriptional activity of the AR. Under certain conditions, this may be the primary mechanism by which growth factor and adhesion receptors interact with the androgen receptor to reduce its dependence on androgen, enhance its survival and proliferation-stimulating activity, and promote progression of prostate cancers to androgen independence (Bakin et al., 2003; Ueda et al., 2002b; Wang et al., 2002). The Rho-GTPase is activated through Met and cell adhesion receptors and has been shown to modulate the activity of the AR (Miao et al., 2003; Ridley et al., 1995; Royal et al., 2000). GTPbound Rho activates the serine/threonine kinases PRK-1 and -2. PRK-1 directly binds to the AR, forms a functional complex with the AR and the AR coactivator TIF-II, hyperactivates the AR in the presence of adrenal


51

androgen, and is overexpressed in prostate cancer as demonstrated by immunohistochemistry (Metzger et al., 2003). In addition, the AR coactivator FHL2 is activated downstream of Rho, and effects AR-mediated transcriptional activity (Muller et al., 2002).

IV. THERAPEUTIC IMPLICATIONS OF MET EXPRESSION IN PROSTATE CANCER Because Met is expressed on metastatic prostate cancer cells, it represents an ideal target for both therapeutic intervention and image analysis of prostate cancer metastases (Hay et al., 2002a). Several approaches have been used to inhibit the HGF/SF–Met system, such as HGF/SF antagonists, monoclonal antibodies against HGF/SF or Met, agents that reduce expression of Met, Met kinase inhibitors, and inhibitors of Met-induced signaling pathways.

A. HGF/SF-Targeted Therapeutic and Imaging Agents Monoclonal antibodies that neutralize human HGF/SF have been developed and shown to significantly inhibit tumor growth in a murine animal model (Cao et al., 2001). However, no single monoclonal antibody blocked the activity of HGF/SF in MDCK scatter or branching morphogenesis assays. HGF/SF possesses at least three regions that mediate its binding to Met. In a glioblastoma multiforme xenograft mouse model, a combination of three neutralizing antibodies inhibited the growth of subcutaneously implanted tumors, which are dependent on an autocrine Met–HGF/SF loop. A combination of anti-HGF/SF and anti-Met antibodies was labeled with iodine-125 and administered to mice bearing tumors autocrine for human HGF/SF and human Met. These demonstrated significantly more rapid uptake and more rapid clearance of the iodine-125-monoclonal antibody mixture than control tumors expressing murine HGF/SF–Met (Hay et al., 2002b). Two naturally occurring HGF/SF antagonists, NK2 (an alternatively spliced HGF/SF consisting of the first two kringle domains) and NK4 (a noncleavable pro-HGF/SF), inhibit Met activation (Montesano et al., 1998; Tomioka et al., 2001). NK4 caused a decrease in the invasion, growth, and angiogenic activity of the PC-3 prostate cancer xenograft (Davies et al., 2003).

52


B. Met-Targeted Therapeutic and Imaging Agents The anisomycin derivative 17-allylamino-17-demethoxygeldanamycin (17-AAG) binds heat shock protein 90 (Hsp90) and thereby causes proteosomal degradation of proteins that require this chaperone for folding and stabilized expression. This is likely mediated by preventing the dissociation of Hsp90–protein complexes and by targeting of the inactive complex to degradation by the ubiquitin system (Schneider et al., 1996). Proteins that depend on Hsp90 include the growth factor receptor kinases Met, insulinlike growth factor receptor, and Her2; the AR and estrogen receptor; and the signal transduction proteins Akt and Raf (reviewed in Solit et al., 2003b). The full spectrum of geldanamycin-sensitive proteins is not known and may include additional prosurvival or proliferation-stimulating proteins. Geldanamycin downregulates Met expression and inhibits Metinduced plasminogen activation (Webb et al., 2000). The inhibition of AR expression in androgen-insensitive prostate cancer may be particularly beneficial, because tumor cells express high amounts of hyperactive AR that cannot be inhibited by antiandrogen therapy. 17-AAG has entered phase I clinical trials as a single agent and in combination with Taxol (Solit et al., 2003a). A single 125-iodine-conjugated antibody against the Met receptor was used to visualize a PC3 prostate cancer xenograft in a mouse (Hay et al., 2003). In xenograft mouse models, this antibody imaged autocrine Metexpressing mesenchymal and epithelial tumors, including the PC3 prostate carcinoma. A direct correlation between levels of anti-Met uptake by xenografts and quantities of Met expressed by the respective cultured tumor cell lines was observed. To utilize their full therapeutic potential, Met antibodies are tested for their inhibition of Met kinase activity and coupled with drugs targeting Met-expressing cancer cells. A small molecular, active site inhibitor of the catalytic activity of the Met kinase has been described (Christensen et al., 2003). In cell lines this inhibitor profoundly reduced Met phosphorylation and the activation of Met-triggered signal transduction pathways, and in xenograft models a Met-driven gastric carcinoma was significantly growth inhibited. The further development of Met-targeted therapies and image analysis tools constitutes the greatest challenge to future Met-related research.

ACKNOWLEDGMENTS The authors thank Drs. Collin Pritchard, Chung Lee, and Leland Chung for helpful suggestions.


53

REFERENCES Aaltomaa, S., Lipponen, P., Viitanen, J., Kankkunen, J. P., Ala-Opas, M., and Kosma, V. M. (2000). Prognostic value of CD44 standard, variant isoforms 3 and 6 and -catenin expression in local prostate cancer treated by radical prostatectomy. Eur. Urol. 38, 555–562. Adler, H. L., McCurdy, M. A., Kattan, M. W., Timme, T. L., Scardino, P. T., and Thompson, T. C. (1999). Elevated levels of circulating interleukin-6 and transforming growth factor- 1 in patients with metastatic prostatic carcinoma. J. Urol. 161, 182–187. Akakura, K., Suzuki, H., Ueda, T., Komiya, A., Ichikawa, T., Igarashi, T., and Ito, H. (2003). Possible mechanism of dexamethasone therapy for prostate cancer: Suppression of circulating level of interleukin-6. Prostate 56, 106–109. Anttonen, A., Kajanti, M., Heikkila, P., Jalkanen, M., and Joensuu, H. (1999). Syndecan-1 expression has prognostic significance in head and neck carcinoma. Br. J. Cancer 79, 558–564. Anttonen, A., Heikkila, P., Kajanti, M., Jalkanen, M., and Joensuu, H. (2001). High syndecan1 expression is associated with favourable outcome in squamous cell lung carcinoma treated with radical surgery. Lung Cancer 32, 297–305. Bakin, R. E., Gioeli, D., Sikes, R. A., Bissonette, E. A., and Weber, M. J. (2003). Constitutive activation of the Ras/mitogen-activated protein kinase signaling pathway promotes androgen hypersensitivity in LNCaP prostate cancer cells. Cancer Res. 63, 1981–1989. Barnett, D. H., Huang, H. Y., Wu, X. R., Laciak, R., Shapiro, E., and Bushman, W. (2002). The human prostate expresses sonic hedgehog during fetal development. J. Urol. 168, 2206–2210. Behrens, J., Weidner, K. M., Frixen, U. H., Schipper, J. H., Sachs, M., Arakaki, N., Daikuhara, Y., and Birchmeier, W. (1991). The role of E-cadherin and scatter factor in tumor invasion and cell motility. EXS 59, 109–126. Bell, A., Chen, Q., DeFrances, M. C., Michalopoulos, G. K., and Zarnegar, R. (1999). The five amino acid-deleted isoform of hepatocyte growth factor promotes carcinogenesis in transgenic mice. Oncogene 18, 887–895. Beviglia, L., and Kramer, R. H. (1999). HGF induces FAK activation and integrin-mediated adhesion in MTLn3 breast carcinoma cells. Int. J. Cancer 83, 640–649. Birchmeier, C., and Gherardi, E. (1998). Developmental roles of HGF/SF and its receptor, the c-Met tyrosine kinase. Trends Cell Biol. 8, 404–410. Birchmeier, C., Birchmeier, W., and Brand-Saberi, B. (1996). Epithelial–mesenchymal transitions in cancer progression. Acta Anat. (Basel) 156, 217–226. Birchmeier, C., Birchmeier, W., Gherardi, E., and Vande Woude, G. F. (2003). Met, metastasis, motility and more. Nat. Rev. Mol. Cell. Biol. 4, 915–925. Birchmeier, W., Brinkmann, V., Niemann, C., Meiners, S., DiCesare, S., Naundorf, H., and Sachs, M. (1997). Role of HGF/SF and c-Met in morphogenesis and metastasis of epithelial cells. Ciba Found. Symp. 212, 230–240. Blanquaert, F., Pereira, R. C., and Canalis, E. (2000). Cortisol inhibits hepatocyte growth factor/scatter factor expression and induces c-met transcripts in osteoblasts. Am. J. Physiol. Endocrinol. Metab. 278, E509–E515. Bonkhoff, H. (1998). Neuroendocrine cells in benign and malignant prostate tissue: Morphogenesis, proliferation, and androgen receptor status. Prostate Suppl. 8, 18–22. Bonkhoff, H. (2001). Morphogenesis of prostate cancer. Eur. Urol. 39(Suppl. 4), 5–7. Borsellino, N., Bonavida, B., Ciliberto, G., Toniatti, C., Travali, S., and D’Alessandro, N. (1999). Blocking signaling through the gp130 receptor chain by interleukin-6 and oncostatin M inhibits PC-3 cell growth and sensitizes the tumor cells to etoposide and cisplatinmediated cytotoxicity. Cancer 85, 134–144.

54


Bottaro, D. P., Rubin, J. S., Faletto, D. L., Chan, A. M., Kmiecik, T. E., Vande Woude, G. F., and Aaronson, S. A. (1991). Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product. Science 251, 802–804. Bowes, R. C., Lightfoot, R. T., III, Van De Water, B., and Stevens, J. L. (1999). Hepatocyte growth factor induces tubulogenesis of primary renal proximal tubular epithelial cells. J. Cell. Physiol. 180, 81–90. Brinkmann, V., Foroutan, H., Sachs, M., Weidner, K. M., and Birchmeier, W. (1995). Hepatocyte growth factor/scatter factor induces a variety of tissue-specific morphogenic programs in epithelial cells. J. Cell Biol. 131, 1573–1586. Cao, B., Su, Y., Oskarsson, M., Zhao, P., Kort, E. J., Fisher, R. J., Wang, L. M., and Vande Woude, G. F. (2001). Neutralizing monoclonal antibodies to hepatocyte growth factor/ scatter factor (HGF/SF) display antitumor activity in animal models. Proc. Natl. Acad. Sci. USA 98, 7443–7448. Chen, D., Adenekan, B., Chen, L., Vaughan, E. D., Gerald, W., Feng, Z., and Knudsen, B. S. (2004). Syndecan-1 expression in locally invasive and metastatic prostate cancer. Urology 63, 402–407. Chen, H. C., Chan, P. C., Tang, M. J., Cheng, C. H., and Chang, T. J. (1998). Tyrosine phosphorylation of focal adhesion kinase stimulated by hepatocyte growth factor leads to mitogen-activated protein kinase activation. J. Biol. Chem. 273, 25777–25782. Chen, T., Wang, L. H., and Farrar, W. L. (2000). Interleukin 6 activates androgen receptormediated gene expression through a signal transducer and activator of transcription 3-dependent pathway in LNCaP prostate cancer cells. Cancer Res. 60, 2132–2135. Cheng, L., Nagabhushan, M., Pretlow, T. P., Amini, S. B., and Pretlow, T. G. (1996a). Expression of E-cadherin in primary and metastatic prostate cancer. Am. J. Pathol. 148, 1375–1380. Cheng, L., Sun, J., Pretlow, T. G., Culp, J., and Yang, N. S. (1996b). CWR22 xenograft as an ex vivo human tumor model for prostate cancer gene therapy. J. Natl. Cancer Inst. 88, 607–611. Chirgadze, D. Y., Hepple, J. P., Zhou, H., Byrd, R. A., Blundell, T. L., and Gherardi, E. (1999). Crystal structure of the NK1 fragment of HGF/SF suggests a novel mode for growth factor dimerization and receptor binding. Nat. Struct. Biol. 6, 72–79. Christensen, J. G., Schreck, R., Burrows, J., Kuruganti, P., Chan, E., Le, P., Chen, J., Wang, X., Ruslim, L., Blake, R., Lipson, K. E., Ramphal, J., Do, S., Cui, J. J., Cherrington, J. M., and Mendel, D. B. (2003). A selective small molecule inhibitor of c-Met kinase inhibits c-Metdependent phenotypes in vitro and exhibits cytoreductive antitumor activity in vivo. Cancer Res. 63, 7345–7355. Collins, A. T., Habib, F. K., Maitland, N. J., and Neal, D. E. (2001). Identification and isolation of human prostate epithelial stem cells based on 2 1-integrin expression. J. Cell Sci. 114, 3865–3872. Cooper, C. S., Park, M., Blair, D. G., Tainsky, M. A., Huebner, K., Croce, C. M., and Vande Woude, G. F. (1984). Molecular cloning of a new transforming gene from a chemically transformed human cell line. Nature 311, 29–33. Corpechot, C., Baulieu, E. E., and Robel, P. (1981). Testosterone, dihydrotestosterone and androstanediols in plasma, testes and prostates of rats during development. Acta Endocrinol. (Copenh.) 96, 127–135. Craft, N., Shostak, Y., Carey, M., and Sawyers, C. L. (1999). A mechanism for hormoneindependent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase. Nat. Med. 5, 280–285. Cress, A. E., Rabinovitz, I., Zhu, W., and Nagle, R. B. (1995). The 6 1 and 6 4 integrins in human prostate cancer progression. Cancer Metastasis Rev. 14, 219–228.


55

Culig, Z., Bartsch, G., and Hobisch, A. (2002). Interleukin-6 regulates androgen receptor activity and prostate cancer cell growth. Mol. Cell. Endocrinol. 197, 231–238. Davies, G., Jiang, W. G., and Mason, M. D. (2001). Matrilysin mediates extracellular cleavage of E-cadherin from prostate cancer cells: A key mechanism in hepatocyte growth factor/ scatter factor-induced cell–cell dissociation and in vitro invasion. Clin. Cancer Res. 7, 3289–3297. Davies, G., Mason, M. D., Martin, T. A., Parr, C., Watkins, G., Lane, J., Matsumoto, K., Nakamura, T., and Jiang, W. G. (2003). The HGF/SF antagonist NK4 reverses fibroblastand HGF-induced prostate tumor growth and angiogenesis in vivo. Int. J. Cancer 106, 348–354. Davis, T. L., Cress, A. E., Dalkin, B. L., and Nagle, R. B. (2001). Unique expression pattern of the 6 4 integrin and laminin-5 in human prostate carcinoma. Prostate 46, 240–248. Deeble, P. D., Murphy, D. J., Parsons, S. J., and Cox, M. E. (2001). Interleukin-6- and cyclic AMP-mediated signaling potentiates neuroendocrine differentiation of LNCaP prostate tumor cells. Mol. Cell. Biol. 21, 8471–8482. Degeorges, A., Tatoud, R., Fauvel-Lafeve, F., Podgorniak, M. P., Millot, G., de Cremoux, P., and Calvo, F. (1996). Stromal cells from human benign prostate hyperplasia produce a growth-inhibitory factor for LNCaP prostate cancer cells, identified as interleukin-6. Int. J. Cancer 68, 207–214. De Marzo, A. M., Bradshaw, C., Sauvageot, J., Epstein, J. I., and Miller, G. J. (1998). CD44 and CD44v6 downregulation in clinical prostatic carcinoma: Relation to Gleason grade and cytoarchitecture. Prostate 34, 162–168. De Marzo, A. M., Knudsen, B., Chan-Tack, K., and Epstein, J. I. (1999a). E-cadherin expression as a marker of tumor aggressiveness in routinely processed radical prostatectomy specimens. Urology 53, 707–713. De Marzo, A. M., Marchi, V. L., Epstein, J. I., and Nelson, W. G. (1999b). Proliferative inflammatory atrophy of the prostate: Implications for prostatic carcinogenesis. Am. J. Pathol. 155, 1985–1992. Derksen, P. W., Keehnen, R. M., Evers, L. M., van Oers, M. H., Spaargaren, M., and Pals, S. T. (2002). Cell surface proteoglycan syndecan-1 mediates hepatocyte growth factor binding and promotes Met signaling in multiple myeloma. Blood 99, 1405–1410. Di Renzo, M. F., Olivero, M., Martone, T., Maffe, A., Maggiora, P., Stefani, A. D., Valente, G., Giordano, S., Cortesina, G., and Comoglio, P. M. (2000). Somatic mutations of the MET oncogene are selected during metastatic spread of human HNSC carcinomas. Oncogene 19, 1547–1555. Dong, Z., Greene, G., Pettaway, C., Dinney, C. P., Eue, I., Lu, W., Bucana, C. D., Balbay, M. D., Bielenberg, D., and Fidler, I. J. (1999). Suppression of angiogenesis, tumorigenicity, and metastasis by human prostate cancer cells engineered to produce interferon- . Cancer Res. 59, 872–879. Drachenberg, D. E., Elgamal, A. A., Rowbotham, R., Peterson, M., and Murphy, G. P. (1999). Circulating levels of interleukin-6 in patients with hormone refractory prostate cancer. Prostate 41, 127–133. Edlund, M., Miyamoto, T., Sikes, R. A., Ogle, R., Laurie, G. W., Farach-Carson, M. C., Otey, C. A., Zhau, H. E., and Chung, L. W. (2001). Integrin expression and usage by prostate cancer cell lines on laminin substrata. Cell Growth Differ. 12, 99–107. Elfman, F., Bok, R., Conn, M., Shuman, M., and Cunha, G. (2001). Urokinase plasminogen activator amino-terminal peptides inhibit development of the rat ventral prostate. Differentiation 69, 108–120. Ellis, W. J., Vessella, R. L., Buhler, K. R., Bladou, F., True, L. D., Bigler, S. A., Curtis, D., and Lange, P. H. (1996). Characterization of a novel androgen-sensitive, prostate-specific

56


antigen-producing prostatic carcinoma xenograft: LuCaP 23. Clin. Cancer Res. 2, 1039–1048. Fixman, E. D., Holgado-Madruga, M., Nguyen, L., Kamikura, D. M., Fournier, T. M., Wong, A. J., and Park, M. (1997). Efficient cellular transformation by the Met oncoprotein requires a functional Grb2 binding site and correlates with phosphorylation of the Grb2-associated proteins, Cbl and Gab1. J. Biol. Chem. 272, 20167–20172. Fournier, T. M., Lamorte, L., Maroun, C. R., Lupher, M., Band, H., Langdon, W., and Park, M. (2000). Cbl-transforming variants trigger a cascade of molecular alterations that lead to epithelial mesenchymal conversion. Mol. Biol. Cell 11, 3397–3410. Freestone, S. H., Marker, P., Grace, O. C., Tomlinson, D. C., Cunha, G. R., Harnden, P., and Thomson, A. A. (2003). Sonic hedgehog regulates prostatic growth and epithelial differentiation. Dev. Biol. 264, 352–362. Fry, P. M., Hudson, D. L., O’Hare, M. J., and Masters, J. R. (2000). Comparison of marker protein expression in benign prostatic hyperplasia in vivo and in vitro. BJU Int. 85, 504–513. Furge, K. A., Zhang, Y. W., and Vande Woude, G. F. (2000). Met receptor tyrosine kinase: Enhanced signaling through adapter proteins. Oncogene 19, 5582–5589. Gakunga, P., Frost, G., Shuster, S., Cunha, G., Formby, B., and Stern, R. (1997). Hyaluronan is a prerequisite for ductal branching morphogenesis. Development 124, 3987–3997. Garcia-Guzman, M., Larsen, E., and Vuori, K. (2000). The proto-oncogene c-Cbl is a positive regulator of Met-induced MAP kinase activation: A role for the adaptor protein Crk. Oncogene 19, 4058–4065. Garraway, L. A., Lin, D., Signoretti, S., Waltregny, D., Dilks, J., Bhattacharya, N., and Loda, M. (2003). Intermediate basal cells of the prostate: In vitro and in vivo characterization. Prostate 55, 206–218. Gherardi, E., Sharpe, M., Lane, K., Sirulnik, A., and Stoker, M. (1993). Hepatocyte growth factor/scatter factor (HGF/SF), the c-Met receptor and the behaviour of epithelial cells. Symp. Soc. Exp. Biol. 47, 163–181. Gioeli, D., Mandell, J. W., Petroni, G. R., Frierson, H. F., Jr., and Weber, M. J. (1999). Activation of mitogen-activated protein kinase associated with prostate cancer progression. Cancer Res. 59, 279–284. Gioeli, D., Ficarro, S. B., Kwiek, J. J., Aaronson, D., Hancock, M., Catling, A. D., White, F. M., Christian, R. E., Settlage, R. E., Shabanowitz, J., Hunt, D. F., and Weber, M. J. (2002). Androgen receptor phosphorylation. Regulation and identification of the phosphorylation sites. J. Biol. Chem. 277, 29304–29314. Gmyrek, G. A., Walburg, M., Webb, C. P., Yu, H. M., You, X., Vaughan, E. D., Vande Woude, G. F., and Knudsen, B. S. (2001). Normal and malignant prostate epithelial cells differ in their response to hepatocyte growth factor/scatter factor. Am. J. Pathol. 159, 579–590. Gornstein, R. A., Lapp, C. A., Bustos-Valdes, S. M., and Zamorano, P. (1999). Androgens modulate interleukin-6 production by gingival fibroblasts in vitro. J. Periodontol. 70, 604–609. Gu, H., and Neel, B. G. (2003). The ‘‘Gab’’ in signal transduction. Trends Cell Biol. 13, 122–130. Harmer, N. J., Chirgadze, D., Hyun Kim, K., Pellegrini, L., and Blundell, T. L. (2003). The structural biology of growth factor receptor activation. Biophys. Chem. 100, 545–553. Hartmann, G., Prospero, T., Brinkmann, V., Ozcelik, C., Winter, G., Hepple, J., Batley, S., Bladt, F., Sachs, M., Birchmeier, C., Birchmeier, W., Gherardi, E., and Ozcelik, O. (1998). Engineered mutants of HGF/SF with reduced binding to heparan sulphate proteoglycans, decreased clearance and enhanced activity in vivo. Curr. Biol. 8, 125–134.


57

Hay, R., Cao, B., Tsarfaty, I., Tsarfaty, G., Resau, J., and Woude, G. V. (2002a). Grappling with metastatic risk: Bringing molecular imaging of Met expression toward clinical use. J. Cell. Biochem. Suppl. 39, 184–193. Hay, R. V., Cao, B., Skinner, R. S., Wang, L. M., Su, Y., Resau, J. H., Vande Woude, G. F., and Gross, M. D. (2002b). Radioimmunoscintigraphy of tumors autocrine for human Met and hepatocyte growth factor/scatter factor. Mol. Imaging 1, 56–62. Hay, R. V., Cao, B., Skinner, R. S., Wang, L. M., Su, Y., Resau, J. H., Knudsen, B. S., Gustafson, M. F., Koo, H. M., Woude, G. F., and Gross, M. D. (2003). Radioimmunoscintigraphy of human met-expressing tumor xenografts using met3, a new monoclonal antibody. Clin. Cancer Res. 9, 3839S–3844S. Heinlein, C. A., and Chang, C. (2002). The roles of androgen receptors and androgen-binding proteins in nongenomic androgen actions. Mol. Endocrinol. 16, 2181–2187. Hiscox, S., and Jiang, W. G. (1999). Association of the HGF/SF receptor, c-Met, with the cellsurface adhesion molecule, E-cadherin, and catenins in human tumor cells. Biochem. Biophys. Res. Commun. 261, 406–411. Hobisch, A., Eder, I. E., Putz, T., Horninger, W., Bartsch, G., Klocker, H., and Culig, Z. (1998). Interleukin-6 regulates prostate-specific protein expression in prostate carcinoma cells by activation of the androgen receptor. Cancer Res. 58, 4640–4645. Hofbauer, L. C., Ten, R. M., and Khosla, S. (1999). The anti-androgen hydroxyflutamide and androgens inhibit interleukin-6 production by an androgen-responsive human osteoblastic cell line. J. Bone Miner. Res. 14, 1330–1337. Horoszewicz, J. S., Leong, S. S., Chu, T. M., Wajsman, Z. L., Friedman, M., Papsidero, L., Kim, U., Chai, L. S., Kakati, S., Arya, S. K., and Sandberg, A. A. (1980). The LNCaP cell line: A new model for studies on human prostatic carcinoma. Prog. Clin. Biol. Res. 37, 115–132. Humphrey, P. A., Zhu, X., Zarnegar, R., Swanson, P. E., Ratliff, T. L., Vollmer, R. T., and Day, M. L. (1995). Hepatocyte growth factor and its receptor (c-MET) in prostatic carcinoma. Am. J. Pathol. 147, 386–396. Inki, P., Joensuu, H., Grenman, R., Klemi, P., and Jalkanen, M. (1994a). Association between syndecan-1 expression and clinical outcome in squamous cell carcinoma of the head and neck. Br. J. Cancer 70, 319–323. Inki, P., Larjava, H., Haapasalmi, K., Miettinen, H. M., Grenman, R., and Jalkanen, M. (1994b). Expression of syndecan-1 is induced by differentiation and suppressed by malignant transformation of human keratinocytes. Eur. J. Cell Biol. 63, 43–51. Isaacs, J. T., and Coffey, D. S. (1989). Etiology and disease process of benign prostatic hyperplasia. Prostate Suppl. 2, 33–50. Ishibe, S., Joly, D., Zhu, X., and Cantley, L. G. (2003). Phosphorylation-dependent paxillin– ERK association mediates hepatocyte growth factor-stimulated epithelial morphogenesis. Mol. Cell 12, 1275–1285. Jackson, D. G. (1997). Human leucocyte heparan sulphate proteoglycans and their roles in inflammation. Biochem. Soc. Trans. 25, 220–224. Jeffers, M., Rong, S., Anver, M., and Vande Woude, G. F. (1996a). Autocrine hepatocyte growth factor/scatter factor–Met signaling induces transformation and the invasive/metastastic phenotype in C127 cells. Oncogene 13, 853–856. Jeffers, M., Rong, S., and Vande Woude, G. F. (1996b). Enhanced tumorigenicity and invasion– metastasis by hepatocyte growth factor/scatter factor–Met signalling in human cells concomitant with induction of the urokinase proteolysis network. Mol. Cell. Biol. 16, 1115–1125. Jeffers, M., Rong, S., and Woude, G. F. (1996c). Hepatocyte growth factor/scatter factor–Met signaling in tumorigenicity and invasion/metastasis. J. Mol. Med. 74, 505–513. Jeffers, M., Fiscella, M., Webb, C. P., Anver, M., Koochekpour, S., and Vande Woude, G. F. (1998). The mutationally activated Met receptor mediates motility and metastasis. Proc. Natl. Acad. Sci. USA 95, 14417–14422.

58


Jiang, J. G., and Zarnegar, R. (1997). A novel transcriptional regulatory region within the core promoter of the hepatocyte growth factor gene is responsible for its inducibility by cytokines via the C/EBP family of transcription factors. Mol. Cell. Biol. 17, 5758–5770. Jiang, J. G., Chen, Q., Bell, A., and Zarnegar, R. (1997). Transcriptional regulation of the hepatocyte growth factor (HGF) gene by the Sp family of transcription factors. Oncogene 14, 3039–3049. Joseph, H., Gorska, A. E., Sohn, P., Moses, H. L., and Serra, R. (1999). Overexpression of a kinase-deficient transforming growth factor- type II receptor in mouse mammary stroma results in increased epithelial branching. Mol. Biol. Cell 10, 1221–1234. Kaighn, M. E., Narayan, K. S., Ohnuki, Y., Lechner, J. F., and Jones, L. W. (1979). Establishment and characterization of a human prostatic carcinoma cell line (PC-3). Invest. Urol. 17, 16–23. Kallakury, B. V., Yang, F., Figge, J., Smith, K. E., Kausik, S. J., Tacy, N. J., Fisher, H. A., Kaufman, R., Figge, H., and Ross, J. S. (1996). Decreased levels of CD44 protein and mRNA in prostate carcinoma: Correlation with tumor grade and ploidy. Cancer 78, 1461–1469. Kallakury, B. V., Sheehan, C. E., Winn-Deen, E., Oliver, J., Fisher, H. A., Kaufman, R. P., Jr., and Ross, J. S. (2001). Decreased expression of catenins ( and ), p120 CTN, and Ecadherin cell adhesion proteins and E-cadherin gene promoter methylation in prostatic adenocarcinomas. Cancer 92, 2786–2795. Kamikura, D. M., Khoury, H., Maroun, C., Naujokas, M. A., and Park, M. (2000). Enhanced transformation by a plasma membrane-associated met oncoprotein: Activation of a phosphoinositide 30 -kinase-dependent autocrine loop involving hyaluronic acid and CD44. Mol. Cell. Biol. 20, 3482–3496. Kataoka, H., Miyata, S., Uchinokura, S., and Itoh, H. (2003). Roles of hepatocyte growth factor (HGF) activator and HGF activator inhibitor in the pericellular activation of HGF/ scatter factor. Cancer Metastasis Rev. 22, 223–236. Kawaida, K., Matsumoto, K., Shimazu, H., and Nakamura, T. (1994). Hepatocyte growth factor prevents acute renal failure and accelerates renal regeneration in mice. Proc. Natl. Acad. Sci. USA 91, 4357–4361. Kindblom, J., Dillner, K., Sahlin, L., Robertson, F., Ormandy, C., Tornell, J., and Wennbo, H. (2003). Prostate hyperplasia in a transgenic mouse with prostate-specific expression of prolactin. Endocrinology 144, 2269–2278. Klein, K. A., Reiter, R. E., Redula, J., Moradi, H., Zhu, X. L., Brothman, A. R., Lamb, D. J., Marcelli, M., Belldegrun, A., Witte, O. N., and Sawyers, C. L. (1997). Progression of metastatic human prostate cancer to androgen independence in immunodeficient SCID mice. Nat. Med. 3, 402–408. Knudsen, B. S., Gmyrek, G. A., Inra, J., Scherr, D. S., Vaughan, E. D., Nanus, D. M., Kattan, M. W., Gerald, W. L., and Vande Woude, G. F. (2002). High expression of the Met receptor in prostate cancer metastasis to bone. Urology 60, 1113–1117. Koksal, I. T., Ozcan, F., Kilicaslan, I., and Tefekli, A. (2002). Expression of E-cadherin in prostate cancer in formalin-fixed, paraffin-embedded tissues: Correlation with pathological features. Pathology 34, 233–238. Lai, J. F., Kao, S. C., Jiang, S. T., Tang, M. J., Chan, P. C., and Chen, H. C. (2000). Involvement of focal adhesion kinase in hepatocyte growth factor-induced scatter of Madin-Darby canine kidney cells. J. Biol. Chem. 275, 7474–7480. Lamm, M. L., Catbagan, W. S., Laciak, R. J., Barnett, D. H., Hebner, C. M., Gaffield, W., Walterhouse, D., Iannaccone, P., and Bushman, W. (2002). Sonic hedgehog activates mesenchymal Gli1 expression during prostate ductal bud formation. Dev. Biol. 249, 349–366. Lamorte, L., Kamikura, D. M., and Park, M. (2000). A switch from p130Cas/Crk to Gab1/Crk signaling correlates with anchorage independent growth and JNK activation in cells transformed by the Met receptor oncoprotein. Oncogene 19, 5973–5981.


59

Lamszus, K., Joseph, A., Jin, L., Yao, Y., Chowdhury, S., Fuchs, A., Polverini, P. J., Goldberg, I. D., and Rosen, E. M. (1996). Scatter factor binds to thrombospondin and other extracellular matrix components. Am. J. Pathol. 149, 805–819. Lang, S. H., Clarke, N. W., George, N. J., and Testa, N. G. (1997). Primary prostatic epithelial cell binding to human bone marrow stroma and the role of 2 1 integrin. Clin. Exp. Metastasis 15, 218–227. Lang, S. H., Stark, M., Collins, A., Paul, A. B., Stower, M. J., and Maitland, N. J. (2001). Experimental prostate epithelial morphogenesis in response to stroma and three-dimensional matrigel culture. Cell Growth Differ. 12, 631–640. Leav, I., Merk, F. B., Lee, K. F., Loda, M., Mandoki, M., McNeal, J. E., and Ho, S. M. (1999). Prolactin receptor expression in the developing human prostate and in hyperplastic, dysplastic, and neoplastic lesions. Am. J. Pathol. 154, 863–870. Lee, C., Sintich, S. M., Mathews, E. P., Shah, A. H., Kundu, S. D., Perry, K. T., Cho, J. S., Ilio, K. Y., Cronauer, M. V., Janulis, L., and Sensibar, J. A. (1999). Transforming growth factor- in benign and malignant prostate. Prostate 39, 285–290. Lee, J. H., Han, S. U., Cho, H., Jennings, B., Gerrard, B., Dean, M., Schmidt, L., Zbar, B., and Vande Woude, G. F. (2000a). A novel germ line juxtamembrane Met mutation in human gastric cancer. Oncogene 19, 4947–4953. Lee, S. L., Dickson, R. B., and Lin, C. Y. (2000b). Activation of hepatocyte growth factor and urokinase/plasminogen activator by matriptase, an epithelial membrane serine protease. J. Biol. Chem. 275, 36720–36725. Lee, S. O., Lou, W., Hou, M., de Miguel, F., Gerber, L., and Gao, A. C. (2003). Interleukin-6 promotes androgen-independent growth in LNCaP human prostate cancer cells. Clin. Cancer Res. 9, 370–376. Li, P., Lee, H., Guo, S., Unterman, T. G., Jenster, G., and Bai, W. (2003). AKT-independent protection of prostate cancer cells from apoptosis mediated through complex formation between the androgen receptor and FKHR. Mol. Cell. Biol. 23, 104–118. Liang, C. C., and Chen, H. C. (2001). Sustained activation of extracellular signal-regulated kinase stimulated by hepatocyte growth factor leads to integrin 2 expression that is involved in cell scattering. J. Biol. Chem. 276, 21146–21152. Lin, C. Y., Anders, J., Johnson, M., and Dickson, R. B. (1999). Purification and characterization of a complex containing matriptase and a Kunitz-type serine protease inhibitor from human milk. J. Biol. Chem. 274, 18237–18242. Lin, D. L., Whitney, M. C., Yao, Z., and Keller, E. T. (2001). Interleukin-6 induces androgen responsiveness in prostate cancer cells through up-regulation of androgen receptor expression. Clin. Cancer Res. 7, 1773–1781. Lin, H. K., Yeh, S., Kang, H. Y., and Chang, C. (2001). Akt suppresses androgen-induced apoptosis by phosphorylating and inhibiting androgen receptor. Proc. Natl. Acad. Sci. USA 98, 7200–7205. Liu, Y., Beedle, A. B., Lin, L., Bell, A. W., and Zarnegar, R. (1994a). Identification of a celltype-specific transcriptional repressor in the promoter region of the mouse hepatocyte growth factor gene. Mol. Cell. Biol. 14, 7046–7058. Liu, Y., Michalopoulos, G. K., and Zarnegar, R. (1994b). Structural and functional characterization of the mouse hepatocyte growth factor gene promoter. J. Biol. Chem. 269, 4152–4160. Loop, S. M., Rozanski, T. A., and Ostenson, R. C. (1993). Human primary prostate tumor cell line, ALVA-31: A new model for studying the hormonal regulation of prostate tumor cell growth. Prostate 22, 93–108. Loric, S., Paradis, V., Gala, J. L., Berteau, P., Bedossa, P., Benoit, G., and Eschwege, P. (2001). Abnormal E-cadherin expression and prostate cell blood dissemination as markers of biological recurrence in cancer. Eur. J. Cancer 37, 1475–1481.

60


Lou, W., Ni, Z., Dyer, K., Tweardy, D. J., and Gao, A. C. (2000). Interleukin-6 induces prostate cancer cell growth accompanied by activation of STAT3 signaling pathway. Prostate 42, 239–242. Lyon, M., Deakin, J. A., Mizuno, K., Nakamura, T., and Gallagher, J. T. (1994). Interaction of hepatocyte growth factor with heparan sulfate: Elucidation of the major heparan sulfate structural determinants. J. Biol. Chem. 269, 11216–11223. Manin, M., Baron, S., Goossens, K., Beaudoin, C., Jean, C., Veyssiere, G., Verhoeven, G., and Morel, L. (2002). Androgen receptor expression is regulated by the phosphoinositide 3-kinase/Akt pathway in normal and tumoral epithelial cells. Biochem. J. 366, 729–736. Marker, P. C., Donjacour, A. A., Dahiya, R., and Cunha, G. R. (2003). Hormonal, cellular, and molecular control of prostatic development. Dev. Biol. 253, 165–174. Maroun, C. R., Naujokas, M. A., Holgado-Madruga, M., Wong, A. J., and Park, M. (2000). The tyrosine phosphatase SHP-2 is required for sustained activation of extracellular signalregulated kinase and epithelial morphogenesis downstream from the Met receptor tyrosine kinase. Mol. Cell. Biol. 20, 8513–8525. Mars, W. M., Zarnegar, R., and Michalopoulos, G. K. (1993). Activation of hepatocyte growth factor by the plasminogen activators uPA and tPA. Am. J. Pathol. 143, 949–958. Martin, D. B., Gifford, D. R., Wright, M. E., Keller, A., Yi, E., Goodlet, D. R., Aebersold, R., and Nelson, P. S. (2004). Quantitative proteomic analysis of proteins released by neoplastic prostate epithelium. Quantitative proteomic analysis of proteins released by neoplastic prostate epithelium. Cancer Res. 64, 347–355. Matsumoto, A., Ono, M., Fujimoto, Y., Gallo, R. L., Bernfield, M., and Kohgo, Y. (1997). Reduced expression of syndecan-1 in human hepatocellular carcinoma with high metastatic potential. Int. J. Cancer 74, 482–491. Matsumoto, K., Tajima, H., Okazaki, H., and Nakamura, T. (1992). Negative regulation of hepatocyte growth factor gene expression in human lung fibroblasts and leukemic cells by transforming growth factor- 1 and glucocorticoids. J. Biol. Chem. 267, 24917–24920. Matsumoto, K., Nakamura, T., and Kramer, R. H. (1994). Hepatocyte growth factor/scatter factor induces tyrosine phosphorylation of focal adhesion kinase (p125FAK) and promotes migration and invasion by oral squamous cell carcinoma cells. J. Biol. Chem. 269, 31807–31813. Maulik, G., Shrikhande, A., Kijima, T., Ma, P. C., Morrison, P. T., and Salgia, R. (2002). Role of the hepatocyte growth factor receptor, c-Met, in oncogenesis and potential for therapeutic inhibition. Cytokine Growth Factor Rev. 13, 41–59. McCawley, L. J., O’Brien, P., and Hudson, L. G. (1998). Epidermal growth factor (EGF)- and scatter factor/hepatocyte growth factor (SF/HGF)-mediated keratinocyte migration is coincident with induction of matrix metalloproteinase (MMP)-9. J. Cell. Physiol. 176, 255–265. McNeal, J. E. (1965). Morphogenesis of prostatic carcinoma. Cancer 18, 1659–1666. Meiners, S., Brinkmann, V., Naundorf, H., and Birchmeier, W. (1998). Role of morphogenetic factors in metastasis of mammary carcinoma cells. Oncogene 16, 9–20. Metzger, E., Muller, J. M., Ferrari, S., Buettner, R., and Schule, R. (2003). A novel inducible transactivation domain in the androgen receptor: Implications for PRK in prostate cancer. EMBO J. 22, 270–280. Miao, H., Nickel, C. H., Cantley, L. G., Bruggeman, L. A., Bennardo, L. N., and Wang, B. (2003). EphA kinase activation regulates HGF-induced epithelial branching morphogenesis. J. Cell Biol. 162, 1281–1292. Mizuno, K., Inoue, H., Hagiya, M., Shimizu, S., Nose, T., Shimohigashi, Y., and Nakamura, T. (1994). Hairpin loop and second kringle domain are essential sites for heparin binding and biological activity of hepatocyte growth factor. J. Biol. Chem. 269, 1131–1136. Monga, S. P., Mars, W. M., Pediaditakis, P., Bell, A., Mule, K., Bowen, W. C., Wang, X., Zarnegar, R., and Michalopoulos, G. K. (2002). Hepatocyte growth factor induces


61

Wnt-independent nuclear translocation of -catenin after Met– -catenin dissociation in hepatocytes. Cancer Res. 62, 2064–2071. Montesano, R., Soriano, J. V., Malinda, K. M., Ponce, M. L., Bafico, A., Kleinman, H. K., Bottaro, D. P., and Aaronson, S. A. (1998). Differential effects of hepatocyte growth factor isoforms on epithelial and endothelial tubulogenesis. Cell Growth Differ. 9, 355–365. Muller, J. M., Metzger, E., Greschik, H., Bosserhoff, A. K., Mercep, L., Buettner, R., and Schule, R. (2002). The transcriptional coactivator FHL2 transmits Rho signals from the cell membrane into the nucleus. EMBO J. 21, 736–748. Nackaerts, K., Verbeken, E., Deneffe, G., Vanderschueren, B., Demedts, M., and David, G. (1997). Heparan sulfate proteoglycan expression in human lung-cancer cells. Int. J. Cancer 74, 335–345. Nagabhushan, M., Pretlow, T. G., Guo, Y. J., Amini, S. B., Pretlow, T. P., and Sy, M. S. (1996). Altered expression of CD44 in human prostate cancer during progression. Am. J. Clin. Pathol. 106, 647–651. Nagakawa, O., Murakami, K., Yamaura, T., Fujiuchi, Y., Murata, J., Fuse, H., and Saiki, I. (2000). Expression of membrane-type 1 matrix metalloproteinase (MT1-MMP) on prostate cancer cell lines. Cancer Lett. 155, 173–179. Nagle, R. B., Hao, J., Knox, J. D., Dalkin, B. L., Clark, V., and Cress, A. E. (1995). Expression of hemidesmosomal and extracellular matrix proteins by normal and malignant human prostate tissue. Am. J. Pathol. 146, 1498–1507. Nakashiro, K., Okamoto, M., Hayashi, Y., and Oyasu, R. (2000). Hepatocyte growth factor secreted by prostate-derived stromal cells stimulates growth of androgen-independent human prostatic carcinoma cells. Am. J. Pathol. 157, 795–803. Nakhla, A. M., and Rosner, W. (1994). Characterization of ALVA-41 cells, a new human prostatic cancer cell line. Steroids 59, 586–589. Naldini, L., Tamagnone, L., Vigna, E., Sachs, M., Hartmann, G., Birchmeier, W., Daikuhara, Y., Tsubouchi, H., Blasi, F., and Comoglio, P. M. (1992). Extracellular proteolytic cleavage by urokinase is required for activation of hepatocyte growth factor/scatter factor. EMBO J. 11, 4825–4833. Nemeth, J. A., Sensibar, J. A., White, R. R., Zelner, D. J., Kim, I. Y., and Lee, C. (1997). Prostatic ductal system in rats: Tissue-specific expression and regional variation in stromal distribution of transforming growth factor- 1. Prostate 33, 64–71. Nguyen, L., Holgado-Madruga, M., Maroun, C., Fixman, E. D., Kamikura, D., Fournier, T., Charest, A., Tremblay, M. L., Wong, A. J., and Park, M. (1997). Association of the multisubstrate docking protein Gab1 with the hepatocyte growth factor receptor requires a functional Grb2 binding site involving tyrosine 1356. J. Biol. Chem. 272, 20811–20819. Nishimura, K., Kitamura, M., Takada, S., Nonomura, N., Tsujimura, A., Matsumiya, K., Miki, T., Matsumoto, K., and Okuyama, A. (1998). Regulation of invasive potential of human prostate cancer cell lines by hepatocyte growth factor. Int. J. Urol. 5, 276–281. Nishimura, K., Kitamura, M., Miura, H., Nonomura, N., Takada, S., Takahara, S., Matsumoto, K., Nakamura, T., and Matsumiya, K. (1999). Prostate stromal cell-derived hepatocyte growth factor induces invasion of prostate cancer cell line DU145 through tumor–stromal interaction. Prostate 41, 145–153. Nishimura, K., Matsumiya, K., Miura, H., Tsujimura, A., Nonomura, N., Matsumoto, K., Nakamura, T., and Okuyama, A. (2003). Effects of hepatocyte growth factor on urokinasetype plasminogen activator (uPA) and uPA receptor in DU145 prostate cancer cells. Int. J. Androl. 26, 175–179. Noordzij, M. A., van Steenbrugge, G. J., Verkaik, N. S., Schroder, F. H., and van der Kwast, T. H. (1997). The prognostic value of CD44 isoforms in prostate cancer patients treated by radical prostatectomy. Clin. Cancer Res. 3, 805–815.

62


Ohira, H., Miyata, M., Kuroda, M., Takagi, T., Tojo, J., Ochiai, H., Kokubun, M., Nishimaki, T., Kasukawa, R., and Obara, K. (1996). Interleukin-6 induces proliferation of rat hepatocytes in vivo. J. Hepatol. 25, 941–947. Ohmichi, H., Koshimizu, U., Matsumoto, K., and Nakamura, T. (1998). Hepatocyte growth factor (HGF) acts as a mesenchyme-derived morphogenic factor during fetal lung development. Development 125, 1315–1324. Okada, H., Tsubura, A., Okamura, A., Senzaki, H., Naka, Y., Komatz, Y., and Morii, S. (1992). Keratin profiles in normal/hyperplastic prostates and prostate carcinoma. Virchows Arch. A Pathol. Anat. Histopathol. 421, 157–161. Okamoto, M., Lee, C., and Oyasu, R. (1997). Interleukin-6 as a paracrine and autocrine growth factor in human prostatic carcinoma cells in vitro. Cancer Res. 57, 141–146. Park, M., Dean, M., Cooper, C. S., Schmidt, M., O’Brien, S. J., Blair, D. G., and Vande Woude, G. F. (1986). Mechanism of met oncogene activation. Cell 45, 895–904. Park, M., Dean, M., Kaul, K., Braun, M. J., Gonda, M. A., and Vande Woude, G. (1987). Sequence of MET protooncogene cDNA has features characteristic of the tyrosine kinase family of growth-factor receptors. Proc. Natl. Acad. Sci. USA 84, 6379–6383. Peschard, P., Fournier, T. M., Lamorte, L., Naujokas, M. A., Band, H., Langdon, W. Y., and Park, M. (2001). Mutation of the c-Cbl TKB domain binding site on the Met receptor tyrosine kinase converts it into a transforming protein. Mol. Cell 8, 995–1004. Pisters, L. L., Troncoso, P., Zhau, H. E., Li, W., von Eschenbach, A. C., and Chung, L. W. (1995). c-met proto-oncogene expression in benign and malignant human prostate tissues. J. Urol. 154, 293–298. Plaschke-Schlutter, A., Behrens, J., Gherardi, E., and Birchmeier, W. (1995). Characterization of the scatter factor/hepatocyte growth factor gene promoter: Positive and negative regulatory elements direct gene expression to mesenchymal cells. J. Biol. Chem. 270, 830–836. Podlasek, C. A., Barnett, D. H., Clemens, J. Q., Bak, P. M., and Bushman, W. (1999). Prostate development requires Sonic hedgehog expressed by the urogenital sinus epithelium. Dev. Biol. 209, 28–39. Pollard, J. W. (2001). Tumour–stromal interactions: Transforming growth factor- isoforms and hepatocyte growth factor/scatter factor in mammary gland ductal morphogenesis. Breast Cancer Res. 3, 230–237. Ponzetto, C., Bardelli, A., Zhen, Z., Maina, F., dalla Zonca, P., Giordano, S., Graziani, A., Panayotou, G., and Comoglio, P. M. (1994). A multifunctional docking site mediates signaling and transformation by the hepatocyte growth factor/scatter factor receptor family. Cell 77, 261–271. Qadan, L. R., Perez-Stable, C. M., Schwall, R. H., Burnstein, K. L., Ostenson, R. C., Howard, G. A., and Roos, B. A. (2000). Hepatocyte growth factor and vitamin D cooperatively inhibit androgen-unresponsive prostate cancer cell lines. Endocrinology 141, 2567–2573. Qiao, H., Saulnier, R., Patryzkat, A., Rahimi, N., Raptis, L., Rossiter, J., Tremblay, E., and Elliott, B. (2000). Cooperative effect of hepatocyte growth factor and fibronectin in anchorage-independent survival of mammary carcinoma cells: Requirement for phosphatidylinositol 3-kinase activity. Cell Growth Differ. 11, 123–133. Ridley, A. J., Comoglio, P. M., and Hall, A. (1995). Regulation of scatter factor/hepatocyte growth factor responses by Ras, Rac, and Rho in MDCK cells. Mol. Cell. Biol. 15, 1110–1122. Ritchie, C. K., Andrews, L. R., Thomas, K. G., Tindall, D. J., and Fitzpatrick, L. A. (1997). The effects of growth factors associated with osteoblasts on prostate carcinoma proliferation and chemotaxis: Implications for the development of metastatic disease. Endocrinology 138, 1145–1150.


63

Robinson, E. J., Neal, D. E., and Collins, A. T. (1998). Basal cells are progenitors of luminal cells in primary cultures of differentiating human prostatic epithelium. Prostate 37, 149–160. Rong, S., Bodescot, M., Blair, D., Dunn, J., Nakamura, T., Mizuno, K., Park, M., Chan, A., Aaronson, S., and Vande Woude, G. F. (1992). Tumorigenicity of the met proto-oncogene and the gene for hepatocyte growth factor. Mol. Cell. Biol. 12, 5152–5158. Rong, S., Segal, S., Anver, M., Resau, J. H., and Vande Woude, G. F. (1994). Invasiveness and metastasis of NIH 3T3 cells induced by Met–hepatocyte growth factor/scatter factor autocrine stimulation. Proc. Natl. Acad. Sci. USA 91, 4731–4735. Rosario, M., and Birchmeier, W. (2003). How to make tubes: Signaling by the Met receptor tyrosine kinase. Trends Cell Biol. 13, 328–335. Rosen, E. M., Goldberg, I. D., Kacinski, B. M., Buckholz, T., and Vinter, D. W. (1989). Smooth muscle releases an epithelial cell scatter factor which binds to heparin. In Vitro Cell. Dev. Biol. 25, 163–173. Rosen, E. M., Nigam, S. K., and Goldberg, I. D. (1994). Scatter factor and the c-Met receptor: A paradigm for mesenchymal/epithelial interaction. J. Cell Biol. 127, 1783–1787. Rosenthal, E. L., Johnson, T. M., Allen, E. D., Apel, I. J., Punturieri, A., and Weiss, S. J. (1998). Role of the plasminogen activator and matrix metalloproteinase systems in epidermal growth factor- and scatter factor-stimulated invasion of carcinoma cells. Cancer Res. 58, 5221–5230. Ross, J. S., Figge, H. L., Bui, H. X., del Rosario, A. D., Fisher, H. A., Nazeer, T., Jennings, T. A., Ingle, R., and Kim, D. N. (1994). E-cadherin expression in prostatic carcinoma biopsies: Correlation with tumor grade, DNA content, pathologic stage, and clinical outcome. Mod. Pathol. 7, 835–841. Royal, I., Lamarche-Vane, N., Lamorte, L., Kaibuchi, K., and Park, M. (2000). Activation of cdc42, rac, PAK, and rho-kinase in response to hepatocyte growth factor differentially regulates epithelial cell colony spreading and dissociation. Mol. Biol. Cell 11, 1709–1725. Ruan, W., Powell-Braxton, L., Kopchick, J. J., and Kleinberg, D. L. (1999). Evidence that insulin-like growth factor I and growth hormone are required for prostate gland development. Endocrinology 140, 1984–1989. Rubin, M. A., Mucci, N. R., Figurski, J., Fecko, A., Pienta, K. J., and Day, M. L. (2001). E-cadherin expression in prostate cancer: A broad survey using high-density tissue microarray technology. Hum. Pathol. 32, 690–697. Ruijter, E., van de Kaa, C., Aalders, T., Ruiter, D., Miller, G., Debruyne, F., and Schalken, J. (1998). Heterogeneous expression of E-cadherin and p53 in prostate cancer: Clinical implications. BIOMED-II Markers for Prostate Cancer Study Group. Mod. Pathol. 11, 276–281. Sachs, M., Weidner, K. M., Brinkmann, V., Walther, I., Obermeier, A., Ullrich, A., and Birchmeier, W. (1996). Motogenic and morphogenic activity of epithelial receptor tyrosine kinases. J. Cell Biol. 133, 1095–1107. Sakurai, H., and Nigam, S. K. (1997). Transforming growth factor- selectively inhibits branching morphogenesis but not tubulogenesis. Am. J. Physiol. 272, F139–F146. Santos, O. F., and Nigam, S. K. (1993). HGF-induced tubulogenesis and branching of epithelial cells is modulated by extracellular matrix and TGF- . Dev. Biol. 160, 293–302. Schmelz, M., Cress, A. E., Scott, K. M., Burger, F., Cui, H., Sallam, K., McDaniel, K. M., Dalkin, B. L., and Nagle, R. B. (2002). Different phenotypes in human prostate cancer: 6 or 3 integrin in cell–extracellular adhesion sites. Neoplasia 4, 243–254. Schmidt, C., Bladt, F., Goedecke, S., Brinkmann, V., Zschiesche, W., Sharpe, M., Gherardi, E., and Birchmeier, C. (1995). Scatter factor/hepatocyte growth factor is essential for liver development. Nature 373, 699–702.

64


Schmidt, L., Duh, F. M., Chen, F., Kishida, T., Glenn, G., Choyke, P., Scherer, S. W., Zhuang, Z., Lubensky, I., Dean, M., Allikmets, R., Chidambaram, A., Bergerheim, U. R., Feltis, J. T., Casadevall, C., Zamarron, A., Bernues, M., Richard, S., Lips, C. J., Walther, M. M., Tsui, L. C., Geil, L., Orcutt, M. L., Stackhouse, T., Zbar, B. et al. (1997). Germline and somatic mutations in the tyrosine kinase domain of the MET proto-oncogene in papillary renal carcinomas. Nat. Genet. 16, 68–73. Schmidt, L., Junker, K., Nakaigawa, N., Kinjerski, T., Weirich, G., Miller, M., Lubensky, I., Neumann, H. P., Brauch, H., Decker, J., Vocke, C., Brown, J. A., Jenkins, R., Richard, S., Bergerheim, U., Gerrard, B., Dean, M., Linehan, W. M., and Zbar, B. (1999). Novel mutations of the MET proto-oncogene in papillary renal carcinomas. Oncogene 18, 2343–2350. Schneider, A., Brand, T., Zweigerdt, R., and Arnold, H. (2000). Targeted disruption of the Nkx3.1 gene in mice results in morphogenetic defects of minor salivary glands: Parallels to glandular duct morphogenesis in prostate. Mech. Dev. 95, 163–174. Schneider, C., Sepp-Lorenzino, L., Nimmesgern, E., Ouerfelli, O., Danishefsky, S., Rosen, N., and Hartl, F. U. (1996). Pharmacologic shifting of a balance between protein refolding and degradation mediated by Hsp90. Proc. Natl. Acad. Sci. USA 93, 14536–14541. Signoretti, S., Waltregny, D., Dilks, J., Isaac, B., Lin, D., Garraway, L., Yang, A., Montironi, R., McKeon, F., and Loda, M. (2000). p63 is a prostate basal cell marker and is required for prostate development. Am. J. Pathol. 157, 1769–1775. Skrtic, S., and Ohlsson, C. (2000). Cortisol decreases hepatocyte growth factor levels in human osteoblast-like cells. Calcif. Tissue Int. 66, 108–112. Solit, D. B., Basso, A. D., Olshen, A. B., Scher, H. I., and Rosen, N. (2003a). Inhibition of heat shock protein 90 function down-regulates Akt kinase and sensitizes tumors to Taxol. Cancer Res. 63, 2139–2144. Solit, D. B., Scher, H. I., and Rosen, N. (2003b). Hsp90 as a therapeutic target in prostate cancer. Semin. Oncol. 30, 709–716. Soriano, J. V., Pepper, M. S., Orci, L., and Montesano, R. (1998). Roles of hepatocyte growth factor/scatter factor and transforming growth factor- 1 in mammary gland ductal morphogenesis. J. Mammary Gland Biol. Neoplasia 3, 133–150. Stone, K. R., Mickey, D. D., Wunderli, H., Mickey, G. H., and Paulson, D. F. (1978). Isolation of a human prostate carcinoma cell line (DU 145). Int. J. Cancer 21, 274–281. Tomioka, D., Maehara, N., Kuba, K., Mizumoto, K., Tanaka, M., Matsumoto, K., and Nakamura, T. (2001). Inhibition of growth, invasion, and metastasis of human pancreatic carcinoma cells by NK4 in an orthotopic mouse model. Cancer Res. 61, 7518–7524. Toyoshima, E., Ohsaki, Y., Nishigaki, Y., Fujimoto, Y., Kohgo, Y., and Kikuchi, K. (2001). Expression of syndecan-1 is common in human lung cancers independent of expression of epidermal growth factor receptor. Lung Cancer 31, 193–202. Trusolino, L., Serini, G., Cecchini, G., Besati, C., Ambesi-Impiombato, F. S., Marchisio, P. C., and De Filippi, R. (1998). Growth factor-dependent activation of v 3 integrin in normal epithelial cells: Implications for tumor invasion. J. Cell Biol. 142, 1145–1156. Trusolino, L., Cavassa, S., Angelini, P., Ando, M., Bertotti, A., Comoglio, P. M., and Boccaccio, C. (2000). HGF/scatter factor selectively promotes cell invasion by increasing integrin avidity. FASEB J. 14, 1629–1640. Trusolino, L., Bertotti, A., and Comoglio, P. M. (2001). A signaling adapter function for 6 4 integrin in the control of HGF-dependent invasive growth. Cell 107, 643–654. Tsarfaty, I., Resau, J. H., Rulong, S., Keydar, I., Faletto, D. L., and Vande Woude, G. F. (1992). The met proto-oncogene receptor and lumen formation. Science 257, 1258–1261. Tuck, A. B., Elliott, B. E., Hota, C., Tremblay, E., and Chambers, A. F. (2000). Osteopontininduced, integrin-dependent migration of human mammary epithelial cells involves activation of the hepatocyte growth factor receptor (Met). J. Cell. Biochem. 78, 465–475.


65

Tuxhorn, J. A., Ayala, G. E., Smith, M. J., Smith, V. C., Dang, T. D., and Rowley, D. R. (2002). Reactive stroma in human prostate cancer: Induction of myofibroblast phenotype and extracellular matrix remodeling. Clin. Cancer Res. 8, 2912–2923. Twillie, D. A., Eisenberger, M. A., Carducci, M. A., Hseih, W. S., Kim, W. Y., and Simons, J. W. (1995). Interleukin-6: A candidate mediator of human prostate cancer morbidity. Urology 45, 542–549. Ueda, T., Bruchovsky, N., and Sadar, M. D. (2002a). Activation of the androgen receptor N-terminal domain by interleukin-6 via MAPK and STAT3 signal transduction pathways. J. Biol. Chem. 277, 7076–7085. Ueda, T., Mawji, N. R., Bruchovsky, N., and Sadar, M. D. (2002b). Ligand-independent activation of the androgen receptor by interleukin-6 and the role of steroid receptor coactivator-1 in prostate cancer cells. J. Biol. Chem. 277, 38087–38094. Ueland, J. M., Gwira, J., Liu, Z. X., and Cantley, L. G. (2004). The chemokine KC regulates HGF-stimulated epithelial cell morphogenesis. Am. J. Physiol. Renal Physiol. 286, F581–F589. van der Voort, R., Taher, T. E., Wielenga, V. J., Spaargaren, M., Prevo, R., Smit, L., David, G., Hartmann, G., Gherardi, E., and Pals, S. T. (1999). Heparan sulfate-modified CD44 promotes hepatocyte growth factor/scatter factor-induced signal transduction through the receptor tyrosine kinase c-Met. J. Biol. Chem. 274, 6499–6506. Vande Woude, G. F., Jeffers, M., Cortner, J., Alvord, G., Tsarfaty, I., and Resau, J. (1997). MetHGF/SF: Tumorigenesis, invasion and metastasis. Ciba Found. Symp. 212, 119–130. van Leenders, G., Dijkman, H., Hulsbergen-van de Kaa, C., Ruiter, D., and Schalken, J. (2000). Demonstration of intermediate cells during human prostate epithelial differentiation in situ and in vitro using triple-staining confocal scanning microscopy. Lab. Invest. 80, 1251–1258. van Leenders, G., van Balken, B., Aalders, T., Hulsbergen-van de Kaa, C., Ruiter, D., and Schalken, J. (2002). Intermediate cells in normal and malignant prostate epithelium express c-MET: Implications for prostate cancer invasion. Prostate 51, 98–107. van Leenders, G. J., Gage, W. R., Hicks, J. L., van Balken, B., Aalders, T. W., Schalken, J. A., and De Marzo, A. M. (2003). Intermediate cells in human prostate epithelium are enriched in proliferative inflammatory atrophy. Am. J. Pathol. 162, 1529–1537. Wainstein, M. A., He, F., Robinson, D., Kung, H. J., Schwartz, S., Giaconia, J. M., Edgehouse, N. L., Pretlow, T. P., Bodner, D. R., Kursh, E. D. et al. (1994). CWR22: Androgen-dependent xenograft model derived from a primary human prostatic carcinoma. Cancer Res. 54, 6049–6052. Wang, B. E., Shou, J., Ross, S., Koeppen, H., De Sauvage, F. J., and Gao, W. Q. (2003). Inhibition of epithelial ductal branching in the prostate by sonic hedgehog is indirectly mediated by stromal cells. J. Biol. Chem. 278, 18506–18513. Wang, R., Kobayashi, R., and Bishop, J. M. (1996). Cellular adherence elicits ligand-independent activation of the Met cell-surface receptor. Proc. Natl. Acad. Sci. USA 93, 8425–8430. Wang, R., Ferrell, L. D., Faouzi, S., Maher, J. J., and Bishop, J. M. (2001). Activation of the Met receptor by cell attachment induces and sustains hepatocellular carcinomas in transgenic mice. J. Cell Biol. 153, 1023–1034. Wang, X., Yang, Y., Guo, X., Sampson, E. R., Hsu, C. L., Tsai, M. Y., Yeh, S., Wu, G., Guo, Y., and Chang, C. (2002). Suppression of androgen receptor transactivation by Pyk2 via interaction and phosphorylation of the ARA55 coregulator. J. Biol. Chem. 277, 15426–15431. Wang, Y., Hayward, S., Cao, M., Thayer, K., and Cunha, G. (2001). Cell differentiation lineage in the prostate. Differentiation 68, 270–279. Watanabe, M., Fukutome, K., Kato, H., Murata, M., Kawamura, J., Shiraishi, T., and Yatani, R. (1999). Progression-linked overexpression of c-Met in prostatic intraepithelial neoplasia and latent as well as clinical prostate cancers. Cancer Lett. 141, 173–178.

66


Webb, C. P., Hose, C. D., Koochekpour, S., Jeffers, M., Oskarsson, M., Sausville, E., Monks, A., and Vande Woude, G. F. (2000). The geldanamycins are potent inhibitors of the hepatocyte growth factor/scatter factor–Met–urokinase plasminogen activator–plasmin proteolytic network. Cancer Res. 60, 342–349. Weidner, K. M., Behrens, J., Vandekerckhove, J., and Birchmeier, W. (1990). Scatter factor: Molecular characteristics and effect on the invasiveness of epithelial cells. J. Cell Biol. 111, 2097–2108. Weidner, K. M., Sachs, M., and Birchmeier, W. (1993). The Met receptor tyrosine kinase transduces motility, proliferation, and morphogenic signals of scatter factor/hepatocyte growth factor in epithelial cells. J. Cell Biol. 121, 145–154. Weidner, K. M., Di Cesare, S., Sachs, M., Brinkmann, V., Behrens, J., and Birchmeier, W. (1996). Interaction between Gab1 and the c-Met receptor tyrosine kinase is responsible for epithelial morphogenesis. Nature 384, 173–176. Weimar, I. S., de Jong, D., Muller, E. J., Nakamura, T., van Gorp, J. M., de Gast, G. C., and Gerritsen, W. R. (1997). Hepatocyte growth factor/scatter factor promotes adhesion of lymphoma cells to extracellular matrix molecules via 4 1 and 5 1 integrins. Blood 89, 990–1000. Weinstein, M. H., Signoretti, S., and Loda, M. (2002). Diagnostic utility of immunohistochemical staining for p63, a sensitive marker of prostatic basal cells. Mod. Pathol. 15, 1302–1308. Wen, Y., Hu, M. C., Makino, K., Spohn, B., Bartholomeusz, G., Yan, D. H., and Hung, M. C. (2000). HER-2/neu promotes androgen-independent survival and growth of prostate cancer cells through the Akt pathway. Cancer Res. 60, 6841–6845. Wielenga, V. J., van der Voort, R., Taher, T. E., Smit, L., Beuling, E. A., van Krimpen, C., Spaargaren, M., and Pals, S. T. (2000). Expression of c-Met and heparan-sulfate proteoglycan forms of CD44 in colorectal cancer. Am. J. Pathol. 157, 1563–1573. Wiksten, J. P., Lundin, J., Nordling, S., Kokkola, A., and Haglund, C. (2000). A prognostic value of syndecan-1 in gastric cancer. Anticancer Res. 20, 4905–4907. Witkowski, C. M., Rabinovitz, I., Nagle, R. B., Affinito, K. S., and Cress, A. E. (1993). Characterization of integrin subunits, cellular adhesion and tumorgenicity of four human prostate cell lines. J. Cancer Res. Clin. Oncol. 119, 637–644. Woolf, A. S., Kolatsi-Joannou, M., Hardman, P., Andermarcher, E., Moorby, C., Fine, L. G., Jat, P. S., Noble, M. D., and Gherardi, E. (1995). Roles of hepatocyte growth factor/scatter factor and the Met receptor in the early development of the metanephros. J. Cell Biol. 128, 171–184. Wu, H. C., Hsieh, J. T., Gleave, M. E., Brown, N. M., Pathak, S., and Chung, L. W. (1994). Derivation of androgen-independent human LNCaP prostatic cancer cell sublines: Role of bone stromal cells. Int. J. Cancer 57, 406–412. Xue, Y., Smedts, F., Ruijter, E. T., Debruyne, F. M., de la Rosette, J. J., and Schalken, J. A. (2001). Branching activity in the human prostate: A closer look at the structure of small glandular buds. Eur. Urol. 39, 222–231. Yoshiura, K., Kanai, Y., Ochiai, A., Shimoyama, Y., Sugimura, T., and Hirohashi, S. (1995). Silencing of the E-cadherin invasion-suppressor gene by CpG methylation in human carcinomas. Proc. Natl. Acad. Sci. USA 92, 7416–7419. You, X., Yu, H. M., Cohen-Gould, L., Cao, B., Symons, M., Vande Woude, G. F., and Knudsen, B. S. (2003). Regulation of migration of primary prostate epithelial cells by secreted factors from prostate stromal cells. Exp. Cell Res. 288, 246–256. Zarnegar, R. (1995). Regulation of HGF and HGFR gene expression. EXS 74, 33–49. Zellweger, T., Ninck, C., Mirlacher, M., Annefeld, M., Glass, A. G., Gasser, T. C., Mihatsch, M. J., Gelmann, E. P., and Bubendorf, L. (2003). Tissue microarray analysis reveals prognostic significance of syndecan-1 expression in prostate cancer. Prostate 55, 20–29.


67

Zeng, X., Gray, M., Stahlman, M. T., and Whitsett, J. A. (2001). TGF- 1 perturbs vascular development and inhibits epithelial differentiation in fetal lung in vivo. Dev. Dyn. 221, 289–301. Zhang, J., Pugh, T. D., Stebler, B., Ershler, W. B., and Keller, E. T. (1998). Orchiectomy increases bone marrow interleukin-6 levels in mice. Calcif. Tissue Int. 62, 219–226. Zhang, Y. W., and Vande Woude, G. F. (2003). HGF/SF-met signaling in the control of branching morphogenesis and invasion. J. Cell. Biochem. 88, 408–417. Zhau, H. Y., Chang, S. M., Chen, B. Q., Wang, Y., Zhang, H., Kao, C., Sang, Q. A., Pathak, S. J., and Chung, L. W. (1996). Androgen-repressed phenotype in human prostate cancer. Proc. Natl. Acad. Sci. USA 93, 15152–15157. Zheng, D. Q., Woodard, A. S., Tallini, G., and Languino, L. R. (2000). Substrate specificity of v 3 integrin-mediated cell migration and phosphatidylinositol 3-kinase/AKT pathway activation. J. Biol. Chem. 275, 24565–24574. Zioncheck, T. F., Richardson, L., Liu, J., Chang, L., King, K. L., Bennett, G. L., Fugedi, P., Chamow, S. M., Schwall, R. H., and Stack, R. J. (1995). Sulfated oligosaccharides promote hepatocyte growth factor association and govern its mitogenic activity. J. Biol. Chem. 270, 16871–16878.

Keratinocyte Growth Factor/Fibroblast Growth Factor 7, a Homeostatic Factor with Therapeutic Potential for Epithelial Protection and Repair Paul W. Finch* and Jeffrey S. Rubin{ {

*Croton-on-Hudson, New York 10520; and Laboratory of Cellular and Molecular Biology, National Cancer Institute, Bethesda, Maryland 20892

I. Introduction II. Molecular Biology of KGF A. Protein B. Expression C. Gene Structure and Promoter Analysis III. Cellular and Molecular Responses to KGF A. Mitogenicity B. Motility C. Differentiation D. Branching Morphogenesis E. Antiapoptotic Effects F. Cytoprotection IV. Signal Transduction V. Transgenic and Knockout Models VI. Regenerative and Protective Effects A. Skin B. Lung C. Bladder D. Gastrointestinal Tract E. Graft-versus-Host Disease and Thymus F. Physiological Mechanisms of Action VII. Clinical Trials: Amelioration of Severe Oral Mucositis A. Autologous Peripheral Blood Progenitor Cell Transplantation for Hematologic Malignancies B. Solid Tumors C. KGF and Cancer VIII. Concluding Remarks References

Advances in CANCER RESEARCH 0065-230X/04 $35.00

69


70

Paul W. Finch and Jeffrey S. Rubin

Keratinocyte growth factor (KGF) is a paracrine-acting, epithelial mitogen produced by cells of mesenchymal origin. It is a member of the fibroblast growth factor (FGF) family, and acts exclusively through a subset of FGF receptor isoforms (FGFR2b) expressed predominantly by epithelial cells. The upregulation of KGF after epithelial injury suggested it had an important role in tissue repair. This hypothesis was reinforced by evidence that intestinal damage was worse and healing impaired in KGF null mice. Preclinical data from several animal models demonstrated that recombinant human KGF could enhance the regenerative capacity of epithelial tissues and protect them from a variety of toxic exposures. These beneficial effects are attributed to multiple mechanisms that collectively act to strengthen the integrity of the epithelial barrier, and include the stimulation of cell proliferation, migration, differentiation, survival, DNA repair, and induction of enzymes involved in the detoxification of reactive oxygen species. KGF is currently being evaluated in clinical trials to test its ability to ameliorate severe oral mucositis (OM) that results from cancer chemoradiotherapy. In a phase 3 trial involving patients who were treated with myeloablative chemoradiotherapy before autologous peripheral blood progenitor cell transplantation for hematologic malignancies, KGF significantly reduced both the incidence and duration of severe OM. Similar investigations are underway in patients being treated for solid tumors. On the basis of its success in ameliorating chemoradiotherapy-induced OM in humans and tissue damage in a variety of animal models, additional clinical applications of KGF are worthy of investigation. ß 2004 Elsevier Inc.

I. INTRODUCTION Keratinocyte growth factor (KGF) was first isolated as an epithelial cell mitogen from the conditioned medium of the human embryonic lung fibroblast cell line, M426 (Rubin et al., 1989). Initially, KGF was shown to stimulate DNA synthesis in BALB/MK mouse keratinocytes and subsequently exhibited mitogenic activity for a wide variety of epithelial cells (Rubin et al., 1989, 1995). In contrast, no activity was seen on any nonepithelial cell types, such as fibroblasts, saphenous vein endothelial cells, melanocytes, or myoblasts (Halaban et al., 1991; Ron et al., 1993b; Rubin et al., 1989). In addition to human embryonic lung fibroblasts, stromal cells from a variety of sources expressed KGF in culture. These included fibroblasts from human adult lung, skin, mammary gland, stomach, bladder, and prostate (Rubin et al., 1995), as well as microvascular endothelial cells (Smola et al., 1993) and smooth muscle cells (Koji et al., 1994; Winkles et al., 1997). It had long been postulated that epithelial cell proliferation during development as well as in adult organs was mediated by diffusible substances released from the underlying mesenchymal tissue (Cunha et al., 1983; Sawyer and Fallow, 1983; Schor and Schor, 1987). The combination of KGF expression by stromal cells and activity specifically on epithelial cells supported the hypothesis that KGF functioned as just such a paracrine mediator of mesenchymal–epithelial communication. Database analysis revealed that KGF was the seventh member of the fibroblast growth factor (FGF) family of structurally related signaling molecules

KGF: Agent for Epithelial Protection

71

to be identified, and so it is also known as FGF7 (Finch et al., 1989). In vertebrates, there are currently 22 identified FGFs that range in molecular mass from 17 to 34 kDa, and share 13–71% amino acid sequence identity. FGFs can be classified into several subfamilies, according to sequence homology within a conserved 120-amino acid core (Kim, 2001), as well as biochemical and/or developmental properties. On the basis of amino acid sequence comparison, KGF has been placed in an FGF subfamily that includes FGF10 and FGF22. Human KGF has 54% sequence identity with human FGF10 (Emoto et al., 1997; Igarashi et al., 1998) and 40% with human FGF22 (Nakatake et al., 2001) within their conserved regions. Like KGF, FGF10 is synthesized predominantly by mesenchymal cells (Beer et al., 1997), and appears to act primarily on epithelial cells (Emoto et al., 1997; Igarashi et al., 1998). Thus in addition to their sequence homology, KGF and FGF10 share a number of biological properties. An initial report indicated that FGF22 has a more limited pattern of expression than either KGF or FGF10, and in the skin was expressed by epidermal keratinocytes rather than mesenchymal cells (Beyer et al., 2003; Nakatake et al., 2001). FGF activity on responsive cells is mediated by a family of high-affinity tyrosine kinase receptors (FGFRs) that are encoded by four structurally related genes (FGFR1–4) (Johnson and Williams, 1993; McKeehan et al., 1998). Further heterogeneity among the FGFRs is generated by alternative splicing of transcripts, resulting in transmembrane protein tyrosine kinases with either two or three immunoglobulin (Ig)-like domains, with or without a highly acidic region also in the extracellular region (Johnson and Williams, 1993; Powers et al., 2000). Specificity of FGF–FGFR binding is determined in part by alternative exons corresponding to the carboxyterminal half of the third Ig domain and an adjacent 20 residues of downstream sequence in FGFRs 1, 2, and 3. These alternative exons, designated IIIa, IIIb, and IIIc, generate receptor variants with different ligand-binding properties (Johnson and Williams, 1993; Powers et al., 2000). Expression cloning of the KGF receptor revealed that it was encoded by IIIb variants of the BEK/FGFR2 gene (FGFR2b) (Miki et al., 1991). Binding studies demonstrated that KGF did not interact with any other FGFR variant (Miki et al., 1991; Ornitz et al., 1996). FGF10 bound preferentially to the FGFR2b receptor variant (Igarashi et al., 1998), although it also associated with the IIIb splice variant of FGFR1, exhibiting approximately 10-fold lower affinity for it than for FGFR2b (Beer et al., 2000; Lu et al., 1999). At the time of writing, the target cell and receptor-binding specificity of FGF22 had not been reported. FGFR2b isoforms are found primarily in epithelial cells whereas FGFR2c isoforms have been observed in cells of mesenchymal origin. Thus, FGFR2b and FGFR2c proteins are expressed in a mutually exclusive manner in these cell lineages, through a positively regulated splicing mechanism that involves intron sequences adjacent to the isoform-specific

72


exons (Carstens et al., 1998, 2000; Del Gatto et al., 1997; Gilbert et al., 1993). The restricted pattern of FGFR2b distribution and remarkable specificity of KGF for FGFR2b isoforms account for the predominant epithelial activity of KGF. One of the defining features of the FGF family is a strong affinity for heparin and heparan sulfate proteoglycans (HSPGs), the latter present on cell surfaces and in the extracellular matrix (Burgess and Maciag, 1989). FGF proteins contain HSPG-binding domains that are topologically defined by specific loop regions in the conserved FGF secondary structure (Ornitz and Itoh, 2001). The spatial arrangement of basic amino acid residues within these loops as well as distinctive structural features in the HSPG molecules have a critical role in determining individual FGF–HSPG interactions (Raman et al., 2003). Although heparin binding initially facilitated the purification of FGFs (Burgess and Maciag, 1989), including KGF (Rubin et al., 1989), subsequent studies established that HSPGs bind both FGFs and FGFRs and have a critical role in strengthening the affinity of their mutual interactions (Pellegrini et al., 2000; Schlessinger et al., 2000). However, the function of HSPGs in FGF signaling is complex, and dependent both on the specific FGF–FGFR interaction and the ability of FGFs to bind different HSPGs (Ostrovsky et al., 2002). Furthermore, in some situations, HSPGs inhibit FGF activity (Aviezer et al., 1994; Bonneh-Barkay et al., 1997). Although KGF originally was suspected to have important activities during embryonic development, we now believe that its primary function is that of a homeostatic factor in the mature organism. In particular, KGF serves to maintain the barrier function of epithelial tissues. Several in vitro and in vivo studies have demonstrated that KGF has potent cytoprotective and regenerative activities in a variety of epithelial contexts. For this reason, efforts have been underway to identify clinical applications for KGF in which the integrity of epithelial surfaces is at risk, and preservation or rapid restoration of these tissues would be of benefit. Specifically, KGF is currently being evaluated in clinical trials sponsored by Amgen (Thousand Oaks, CA) to test its ability to ameliorate severe oral mucositis (OM) that results from cancer chemoradiotherapy.* In the following pages, we review the molecular properties of KGF, the cellular responses it elicits, and pertinent signaling mechanisms. After summarizing observations from transgenic and knockout models, we describe the salient findings from many in vivo injury models that provided the foundation for a series of ongoing clinical trials. The design and results of these clinical studies are presented, along with a concise view of future directions for KGF clinical development.

*Palifermin is the name for recombinant human KGF that will be used in the clinical setting. Note that a truncated derivative of FGF10 was given the name Repifermin.


73

II. MOLECULAR BIOLOGY OF KGF A. Protein 1. STRUCTURE KGF was purified as a monomeric polypeptide with an apparent molecular mass of 26–28 kDa (Rubin et al., 1989). The minor heterogeneity observed in molecular weight presumably was due to differences in glycosylation or partial proteolysis. The KGF cDNA encoded a 194-amino acid protein, including a classic signal peptide for secretion and one potential N (asparagine)-linked glycosylation site (Fig. 1) (Finch et al., 1989). When expressed in bacteria, a biologically active protein was obtained with an apparent molecular size of 21 kDa and a specific activity approximately 10-fold greater than that of the native protein isolated from fibroblastconditioned medium (Ron et al., 1993a). A likely explanation for the difference in size was the absence of glycosylation from the bacterially expressed recombinant protein, suggesting that this posttranslational modification may in some way hinder KGF stability or interactions with its receptor (Ron et al., 1993a). Analysis of recombinant KGF amino (N)-terminal truncation mutants demonstrated that the first 23 amino acid residues of KGF downstream from the signal sequence could be removed without decreasing mitogenic activity, whereas sequential deletion of the next 6 residues dramatically reduced biological activity (Nybo et al., 1997; Osslund et al., 1998; Ron et al., 1993a). Similarly, heparin-binding properties were preserved with deletion of up to 28 amino acid residues of the mature molecule, but lost with the removal of an additional 10 residues (Ron et al., 1993a). The crystal structure of KGF was determined at a resolution of 1.6 A˚ (Osslund et al., 1998), and featured a -trefoil motif similar to that of other FGF family members whose structures have been solved, as well as interleukin 1 (IL-1). This motif consists of 12 antiparallel strands in which three pairs of the strands form a -barrel structure, and the other three pairs cap the barrel with hairpin triplets, forming a triangular array (Ago et al., 1991; Blaber et al., 1996; Eriksson et al., 1991; Graves et al., 1990; Zhang et al., 1991; Zhu et al., 1991). The KGF structure has 10 well-defined strands that form five double-stranded antiparallel sheets (Fig. 2A and B), and a poorly defined sixth -strand pair identified by a single -sheet hydrogen bond between residues 168 and 172 (Osslund et al., 1998). A comparison of the three-dimensional structure of KGF with that of FGF10 revealed a high degree of homology (Osslund et al., 1998; Yeh et al., 2003). Furthermore, there are several distinctive traits unique to KGF subfamily members. For example, their 1 strands are longer than those

74


Fig. 1 The amino acid sequence of recombinant human KGF. The N-terminal sequence of the native protein purified from fibroblast culture fluid began at residue C32; the presumptive signal sequence is shown to the right. Biological activity remained intact after removal of the first 23 residues, starting with C32, but further truncation resulted in a loss of activity. A derivative lacking the first 23 residues and containing an R175Q substitution was used to obtain the crystal structure of KGF. A single putative N-linked glycosylation site occurs at N45. Disulfide bonds linking C32 to C46 and C133 to C137 are indicated by solid lines. Site-directed mutagenesis demonstrated that the residues highlighted in red, blue, and yellow, which are present in the N terminus, 4/5 loop, and 8 strand, respectively, are important for biological activity (Sher et al., 2000, 2003). The shaded residues S153–M163 comprise a loop contributing to the putative secondary binding site for FGFR2b; alanine scanning indicated that W156 was particularly important for biological activity (Osslund et al., 1998; Sher et al., 1999). (Adapted from Osslund et al., 1998, Fig. 1, p. 1682; numbering of residues has been changed from system used in that article to account for residues in the signal sequence.)

of FGF1 and FGF2, their 1/2 and 9/10 loops adopt different conformations because their lengths differ from the corresponding structures in FGF1 and FGF2 and, as mentioned previously, their 10/11-strand pair was defined by a single hydrogen bond between the two strands. The N termini of FGF1 and FGF2 are considerably disordered, perhaps as a result of a cis/trans isomerization of prolines, such that the N-terminal segments were not defined in the electron density maps. In contrast, the KGF and


75

Fig. 2 KGF secondary structure. (A) The KGF structure has 10 well-defined strands that form five double-stranded antiparallel sheets. In addition, the strand 4 is the central strand in a triple-stranded sheet. This triple-stranded sheet is not present in FGF1 or FGF2, as the strand 1 is significantly longer in KGF. A sixth poorly defined strand pair is in the loop between residues 164 and 175, and is defined by only a single hydrogen bond between the two strands (this corresponds to the similarly defined 10/11 pair in FGF10). (B) Pertinent information about the boundaries and connections of the strands in KGF. (Adapted from Osslund et al., 1998, Fig. 4, p. 1686, with numbering of residues modified as noted in the legend to Fig. 1.)

76


FGF10 N termini are substantially more ordered, in part because they do not exhibit this proline isomerization (note, however, that the crystal structure of KGF was determined with a truncated derivative that lacked a proline residue upstream of the conserved FGF core) (Osslund et al., 1998). In addition, a four-amino acid N-terminal extension of the 1 strand, and a common glycine turn that causes the first five ordered amino acids to fold back toward the core structure, serve to provide additional stabilization (Osslund et al., 1998; Yeh et al., 2003). KGF also possesses a hydrogen bond between the backbone nitrogen of Y56 and the carbonyl of C133, which further contributes stability to the N terminus (Osslund et al., 1998). Sequence comparison suggests that many of these features also may be present in FGF22 (Nakatake et al., 2001). It has been hypothesized that these structural characteristics play a role in determining the receptor specificity of the KGF subfamily (Yeh et al., 2003).

2. FGFR2B INTERACTIONS By utilizing domain swapping and site-directed mutagenesis, the N-terminal portion and the 4/5 loop of KGF were found to contribute to high-affinity binding to FGFR2b (Reich-Slotky et al., 1995; Sher et al., 2000). Point mutations in the loop connecting 9/10 strands did not alter receptorbinding affinity (Sher et al., 1999), but caused a decrease in mitogenic potency (Osslund et al., 1998; Sher et al., 1999) and receptor-mediated phosphorylation events (Sher et al., 1999). W156 appeared to be particularly crucial, as its replacement by an alanine residue almost completely abolished mitogenic activity (Osslund et al., 1998; Sher et al., 1999). These results were similar to results previously reported for FGF2, in which the loop connecting 9/10 strands was shown to form a secondary, low-affinity receptor-binding site that was required for receptor activation (Springer et al., 1994). Although a KGF–FGFR2b crystal structure has not been reported, the structure of a complex consisting of FGF10 and the ligand-binding portion of FGFR2b has been determined (Yeh et al., 2003). This study identified unique contacts between FGF10 and two loops in the third Ig domain of FGFR2b (D3), designated B0 -C and C0 -E, that provide the basis for ligand/receptor-binding specificity. The majority of contacts occur between FGF10 and a wide cleft in D3. One side of the cleft consists of the B0 strand of the B0 -C loop; this strand is also present in the FGFR2c isoform. The other side of the cleft is composed of the C0 -E loop, which is located in the second half of D3, and therefore is unique to FGFR2b. The N terminus, 1 strand, 4 strand, and 7/8 loop of FGF10 made specific contacts with both sides of this cleft. These segments exhibit significant sequence diversity among the FGF family members. Mutations within the N terminus and 7/8 loop of FGF10 resulted in molecules with reduced capacity to induce


77

DNA synthesis in BALB/MK mouse keratinocytes, underscoring the importance of these regions for receptor binding (Yeh et al., 2003). The only FGF10–D3 contact outside of the D3 cleft occurred between 8 of FGF10 and the FGFR2b-specific F-G loop of D3. Because key determinants of FGF10–FGFR2b binding involve structural features that typify the KGF subfamily, it is likely that these properties also are relevant to KGF–FGFR2b interactions. For example, the primary role of D3 in FGF10–FGFR2b binding was consistent with an earlier observation that a D3 derivative bound KGF with high affinity (Cheon et al., 1994). Furthermore, the portion of the FGF10 N terminus involved in receptor binding corresponded to KGF residues previously reported to be required for potent biological activity (Osslund et al., 1998; Ron et al., 1993a). As described previously, the 4 strand and N terminus of KGF also have been implicated in FGFR2b binding (Osslund et al., 1998; Reich-Slotky et al., 1995; Ron et al., 1993a; Sher et al., 2000). A mutation in the 8 loop of KGF reduced both receptor-binding affinity and mitogenic activity (Sher et al., 2003), like similar mutations introduced into FGF10 (Yeh et al., 2003). Mutagenesis and peptide-binding studies indicated that the FGFR2b-specific C0 -E loop, which comprised part of the D3 cleft involved in FGF10 binding, participated directly in KGF binding (Bottaro et al., 1993; Gray et al., 1995; Wang et al., 1999a). Finally, mutations within the FGFR2bspecific, F-G loop of D3, a region that makes contact with FGF10 outside of the D3 cleft (Yeh et al., 2003), have been shown to reduce KGF binding (Gray et al., 1995). Besides its interactions with D3, FGF10 induced a rotation of the second Ig loop of FGFR2 (D2), resulting in specific contacts with FGF10 (Yeh et al., 2003). Thus, although FGF10 and presumably KGF interactions with D3 are of primary importance, binding to D2 occurred as well. Interactions with both D2 and D3 also have been documented for FGF2 binding to FGFR1 (Plotnikov et al., 1999; Venkataraman et al., 1999) and for FGF1 binding to FGFR2 (Stauber et al., 2000). The FGF10–FGFR2b crystal structure did not identify receptor contacts with the 9/10 loop of FGF10. As noted previously, this loop is thought to function in KGF as a secondary receptor-binding site. Because the FGF10– FGFR2b complex was generated as a binary unit in the absence of heparin, it is possible that such interactions would be detected in a higher order complex consisting of FGF10 or KGF, FGFR2b, and heparin oligosaccharide. On the basis of prior work with other FGF–FGFR combinations, two ligand molecules would be expected to associate with two FGFR2b molecules. In such complexes involving FGF1 or FGF2 and FGFRs, individual FGF molecules form contacts with two receptor molecules via primary and secondary binding sites (Pellegrini et al., 2000; Plotnikov et al., 1999; Schlessinger et al., 2000; Stauber et al., 2000; Venkataraman et al., 1999).

78


We surmise that in similar complexes of KGF or FGF10 and FGFR2b, the 9/10 loop would be in contact with a second FGFR2b molecule.

3. HSPG BINDING AND ACTIVITY Heparin/HSPG binding has long been recognized to have a major impact on FGF activity. Many early studies indicated that HSPGs stimulated the activity of FGF1 and FGF2, presumably by increasing their local concentration at the cell surface, or by inducing conformational changes in the FGFs that enhanced their stability and/or FGFR interactions (Aviezer et al., 1994; Klagsbrun and Baird, 1991; Mansukhani et al., 1992; Rapraeger et al., 1991). More recent crystallographic work showed that HSPG molecules stabilize the formation of FGF–FGFR complexes by binding both to ligand and receptor molecules (Pellegrini et al., 2000; Schlessinger et al., 2000). HSPG enhanced the binding of individual FGF and FGFR molecules to each other, and facilitated the association of higher order FGF–FGFR complexes. In contrast to FGF1 and FGF2, initial reports demonstrated that heparin inhibited KGF activity (Reich-Slotky et al., 1994; Ron et al., 1993a; Strain et al., 1994). Moreover, a reduction in the amount of endogenous proteoglycan on the surface of BALB/MK mouse keratinocytes or rat myoblast– FGFR2b transfectants decreased FGF1 binding and mitogenicity, whereas corresponding KGF activities increased (Reich-Slotky et al., 1994). Subsequently, glypican-1 was identified as an HSPG that stimulated FGF1 signaling, but inhibited KGF activity in various assays (Berman et al., 1999; Bonneh-Barkay et al., 1997). In cells lacking HSPG, heparin exhibited a biphasic effect on KGF activity: at physiological levels KGF binding to FGFR2b was enhanced, whereas higher heparin concentrations were inhibitory (Berman et al., 1999; Hsu et al., 1999; LaRochelle et al., 1999). Presumably, when heparin is added to cells that express ample quantities of HSPG, the overall amount of heparin/HSPG at the cell surface may be sufficiently high to inhibit KGF activity. Cell-free analysis revealed that HSPG was required for KGF binding to soluble monomeric FGFR2b (Hsu et al., 1999), but not necessary for binding to preformed FGFR2b dimers (LaRochelle et al., 1999). Moreover, the KGF–FGFR2b complex formed when KGF and monomeric FGFR2b were incubated with heparin contained two FGFR2b molecules. These results implied that HSPG stimulated KGF binding to FGFR2b by promoting receptor dimerization and, consequently, the simultaneous binding of a single KGF molecule with two FGFR2b molecules. Other studies demonstrated that HSPG facilitated monomeric interactions between KGF and FGFR2b (Berman et al., 1999; Ostrovsky et al., 2002). Combined with the evidence that HSPG binds both to KGF and FGFR2b (Hsu et al., 1999; LaRochelle et al., 1999), the general pattern of HSPG effects on KGF–FGFR2b


79

interactions appears to resemble that documented for other FGF–FGFR combinations. Nonetheless, as noted at the outset of this section, there are differences in the way HSPGs regulate KGF compared with other FGFs. Differences even exist between KGF and FGF10: heparin stimulated FGF10 mitogenic activity on BALB/MK cells, whereas KGF activity was reduced (Igarashi et al., 1998). Contrasting effects of HSPG on FGF function can be attributed in large measure to the distinct activities, tissue distributions, and FGF–FGFR specificities of differentially modified HSPG molecules (Allen and Rapraeger, 2003; Chang et al., 2000; Friedl et al., 1997; Kreuger et al., 2001; Ostrovsky et al., 2002; Powell et al., 2002; Pye et al., 2000; Ye et al., 2001). For instance, 2-O- and 6-O-desulfated heparin activated FGF1 signaling via FGFR2b but had no effect on KGF signaling (Ostrovsky et al., 2002). Alternatively, heparin oligosaccharides rich in 3-O-sulfate were effective at protecting KGF from protease digestion, implying that this sulfated HSPG interacted with KGF (Ye et al., 2001). Interestingly, experiments with selectively sulfated heparin molecules indicated that O-sulfated groups were critical for FGF10 activity during lung bud formation. Furthermore, the effect of FGF10 in branching morphogenesis was in part determined by regional distribution of O-sulfated HSPGs (Izvolsky et al., 2003a,b). Taken together, these data indicate that the presence of specific patterns of HSPG modification represents a critical determinant for FGF binding, and may provide an important mechanism whereby ligands with similar FGFR-binding properties, such as KGF and FGF10, can elicit different biological responses in vivo. A number of studies have provided valuable insights about the HSPGbinding domain of KGF. An analysis with peptides spanning structural motifs in the KGF protein identified regions in the N terminus (residues 64–103) and C terminus (residues 148–194) that contributed to heparin binding (Kim et al., 1998). Superimposition of the C trace of the KGF -trefoil scaffold with that of FGF2 revealed that there was significant structural homology between the putative KGF HSPG-binding domain and that of FGF2 (Faham et al., 1996; Raman et al., 2003). However, the spatial arrangement of the basic amino acid residues and their side-chain conformations within the pocket were significantly different (Osslund et al., 1998; Raman et al., 2003; Ye et al., 2001). In particular, the nonpolar residue V143 of KGF, which superimposed with K126 of FGF2, could not participate in HSPG binding, although T154 may replace some of the binding energy lost due to the valine substitution (Osslund et al., 1998). Heparin oligosaccharides of six to eight residues were sufficient for FGF1and FGF2-induced FGFR dimerization and activation (Aviezer et al., 1994; Guimond et al., 1993; Ornitz et al., 1992). In contrast, only relatively long oligosaccharides stimulated receptor binding and activation by KGF (Hsu et al., 1999; Ostrovsky et al., 2002).

80


Furthermore, protection of KGF from protease digestion required longer heparin oligosaccharides compared with those required to protect FGF1 and FGF2 (Ye et al., 2001). These data indicated that the positive charge of the HSPG-binding site of KGF was weaker and more dispersed than that of FGF2. Such differences may account for the contrasting HSPG-binding specificities of individual FGFs (Raman et al., 2003). Understanding the HSPGbinding properties of KGF not only enhances our knowledge about the molecular basis of its activities, but also may provide insight about mechanisms that govern the bioavailability and pharmacokinetics of recombinant KGF used in a clinical setting (see Section VII). In addition to HSPG, KGF also binds dermatan sulfate, the predominant glycosaminoglycan in skin, and its activity is potentiated by this interaction (Trowbridge et al., 2002). Furthermore, besides associating with the heparan moieties of HSPG, KGF interacts with specific domains in the protein core of perlecan, an HSPG expressed primarily at cell surfaces and in basement membranes (Mongiat et al., 2000), and this interaction might contribute to the stimulatory effect of perlecan on KGF activity (Ghiselli et al., 2001). KGF also binds to other protein components of the extracellular matrix, in particular, collagens I, III, and IV, using the consensus sequence glycine-proline-hydroxyproline (Gly-Pro-Hyp) as the binding motif (Ruehl et al., 2002). Association with these collagens may further determine the spatial distribution of KGF.

B. Expression 1. MESENCHYMAL–EPITHELIAL PARADIGM KGF and FGFR2b expression in cell culture, as well as KGF target cell specificity, suggested a role for this pathway in mediating mesenchymal– epithelial interactions known to be important both during organogenesis and in the adult. To investigate this possibility, in situ hybridization (ISH) expression studies were performed with mouse embryos. FGFR2b transcripts were detected during gastrulation, but their spatial distribution was diffuse and overlapped with that of the FGFR2c transcript (Orr-Urtreger et al., 1993). However, as development progressed, FGFR2b transcripts became prominent in the surface ectoderm and its derivative structures, including the mammary gland, as well as in epithelial cells of the respiratory, gastrointestinal (GI), and urogenital systems (Finch et al., 1995a; Orr-Urtreger et al., 1993). KGF transcripts were detected in mesenchymal cells adjacent to epithelia that expressed FGFR2b (Finch et al., 1995a; Mason et al., 1994). Although KGF and FGFR2b typically were coexpressed in developing organs by


81

midgestation (Finch et al., 1995a), expression of FGFR2b often preceded that of KGF during organogenesis (Mason et al., 1994; Orr-Urtreger et al., 1993). For instance, FGFR2b transcripts were evident in the surface epithelium as early as 10 days postcoitus (dpc), whereas KGF expression was first detected in the dermis between 15.5 and 16.5 dpc, corresponding to the time when the epidermis changes from a simple to a stratified epithelium (Mason et al., 1994). In contrast, the temporal expression of FGF10 and FGFR2b was similar in developing organs. In lung, KGF transcripts were not present until 14.5 dpc (Bellusci et al., 1997), and were distributed in a diffuse manner throughout the entire lung mesenchyme (Finch et al., 1995a; Mason et al., 1994). However, FGF10 transcripts were expressed in the mesenchyme adjacent to distal buds from the earliest stages of lung development (11.5 dpc) (Bellusci et al., 1997; Park et al., 1998). A similar distinction in the temporal and spatial distribution of KGF and FGF10 transcripts was observed in the ventral prostate (Thomson and Cunha, 1999). These results suggested that KGF contributed to mesenchymal–epithelial communication, particularly in the later stages of organogenesis, but that other factors such as FGF10 probably had a more critical role in organ development. This interpretation was reinforced by the phenotypes of pertinent knockout mouse models (see Section V). KGF and FGFR2b transcripts were also widely expressed in adult tissues, including skin (Werner, 1998), lung (Ulich et al., 1994), breast (Pedchenko and Imagawa, 2000b), and organs of the GI system (Housley et al., 1994) and reproductive tracts (Koji et al., 1994; Thomson et al., 1997). ISH indicated that the spatial distribution of transcripts closely followed the paradigm established during development, with KGF expressed in the stroma and FGFR2b in the epithelium. These data suggested that KGF was a potential homeostatic factor for epithelia in the adult.

2. OTHER PATTERNS ISH analysis revealed that KGF was expressed during development in other locations not associated with mesenchymal–epithelial communication. Expression was first seen in the heart at 8.5 dpc and persisted until 11 dpc (Mason et al., 1994). Transcripts were detected throughout the myocardium in both trabeculated and compact regions, although signal was stronger in the atrium than the ventricle. Both KGF and FGFR2b transcripts were detected in the perichondrium and cartilage of limbs, ribs, pelvic bones, larynx, and trachea (Finch et al., 1995a). Furthermore, KGF and FGFR2b transcripts were jointly expressed in developing skeletal muscle (Finch et al., 1995a; Mason et al., 1994), as well as in smooth muscle of the small intestine (Finch et al., 1995a). KGF mRNA was also transiently expressed in the developing forebrain around 14.5 dpc in three separate regions of the

82


ventricular zone, where the progenitor cells of neurons and glia proliferate (Mason et al., 1994). These data suggested that KGF may have diverse roles during development in tissues not governed by mesenchymal–epithelial interactions.

a. Epithelial Cells Although numerous studies of KGF expression in vitro and in vivo indicated that KGF was primarily expressed by stromal cells from a variety of organs, there also have been reports of KGF expression by normal epithelial cells. KGF expression was observed by ISH in the endometrial epithelium of porcine uterus, and was especially prominent during days 12 and 15 of the estrous cycle and pregnancy (Ka et al., 2000). KGF expression was upregulated by 17-estradiol, and postulated to stimulate the proliferation and differentiation of the conceptus trophectoderm (Ka et al., 2000). KGF transcripts were detected in cultured ovarian surface epithelial cells derived from bovine ovaries, and KGF stimulated the growth of these cells in vitro (Parrott et al., 2000). Ovarian epithelial cells are of mesodermal origin, arising from the mesothelial lining of the abdomen, and express a mixture of classically epithelial and mesenchymal markers (Parrott et al., 2000). Conceivably, this mixed lineage may account for the coexpression of KGF and receptor. In addition, KGF transcript was detected in primary lens epithelial cells (Weng et al., 1997). Taken together, these findings suggest that certain normal epithelial cells express KGF and may be able to respond to it, although further investigation is required to establish the physiologic significance of these observations.

b. Vascular Cells KGF expression has been detected in a variety of vascular smooth muscle cells. For instance, ISH analysis identified KGF mRNA in smooth muscle cells from the spiral arteries of monkey endometrium (Koji et al., 1994). Cultured smooth muscle cells from human saphenous vein and iliac artery expressed and secreted mitogenically active KGF. KGF transcripts were detected in normal and atherosclerotic human arteries, although FGFR2b transcripts were not observed (Winkles et al., 1997). However, others described FGFR2b expression by cultured rat vascular smooth muscle cells. Furthermore, KGF was shown to stimulate the growth of these cells by approximately 10% (Onda et al., 2003). Even though this represented a small increase, it appeared to be statistically significant. Although KGF lacks activity on cultured endothelial cells derived from large vessels, there is one report that it stimulated neovascularization in the cornea and proliferation of endothelial cells derived from small vessels (Gillis et al., 1999). Because FGFR2b was not detected in these cells, it was hypothesized that KGF may act on microvascular endothelial cells


83

through an as yet undiscovered high-affinity receptor. The significance of this observation awaits further confirmatory studies.

c. Lymphocytes Epithelial tissues contain T cells that express the T cell receptor (TCR). T cells have evolved to recognize antigen in a different manner and perform a broader set of functions than T cells expressing TCRs. Activated T cells from the skin, intestine, and vagina express KGF, whereas intraepithelial T cells, as well as all lymphoid and T cells, did not produce measurable quantities of KGF (Boismenu and Havran, 1994; Rakasz et al., 1996). It was postulated that T cells recognized antigens expressed by injured epithelial cells, and this triggered the synthesis of KGF to minimize damage to epithelial surfaces, and hasten wound repair. T cells isolated from the thymus also have been reported to express KGF in a developmentally regulated manner (perhaps negative results mentioned previously were due to use of a less sensitive assay). KGF expression was undetectable in CD348 thymocytes, but was readily observed in mature CD4þ and CD8þ thymocytes (Erickson et al., 2002). Exposure of thymocyte-depleted fetal thymic lobes to KGF resulted in decreased thymic epithelial expression of class II major histocompatibility complex, and stimulated expression of IL-6. In intact fetal thymic organ cultures, KGF inhibited the generation of CD4þ thymocytes (Erickson et al., 2002). These findings implied that KGF–FGFR2b signaling participated in the development and function of thymic epithelium. As lymphocytes are derived from mesoderm, it is reasonable to consider KGF release by these cells as a special example of its paracrine mode of action on epithelial cells.

3. UPREGULATION AFTER INJURY KGF expression was dramatically upregulated after cutaneous injury in mouse and human full-thickness excisional wounds (Marchese et al., 1995; Werner et al., 1992). Increased KGF expression was confined to the dermal compartment. KGF transcripts also were elevated after tissue damage in models of surgical bladder injury (Baskin et al., 1997), chemically induced kidney injury (Ichimura et al., 1996), exposure of neonatal rabbit lungs to hyperoxia (Charafeddine et al., 1999), and acute lung injury resulting from bleomycin injection (Adamson and Bakowska, 1999). In an experimental model of fasting-induced gut atrophy, KGF was increased in the ileum, suggesting this may be an adaptive response to limit the extent of mucosal wasting (Estivariz et al., 2000). Furthermore, upregulation of KGF was documented in some human inflammatory diseases, including psoriasis (Finch et al., 1997) and inflammatory bowel disease (IBD) (Bajaj-Elliott et al., 1997; Brauchle et al., 1996; Finch et al., 1996). In both these

84


disorders, FGFR2b expression was observed in the epithelial cells, while KGF was synthesized in the underlying stroma. These studies suggested that KGF participates in a wide variety of epithelial preservation and/or repair processes.

4. CYTOKINE AND HORMONAL REGULATION The many examples of KGF induction during acute or chronic epithelial injury and repair prompted studies to determine how KGF expression is regulated in these situations. IL-1 and IL-6, proinflammatory cytokines expressed by macrophages and polymorphonuclear leukocytes, stimulated either a large (IL-1) or moderate (IL-6) increase in KGF mRNA and protein synthesis in fibroblasts from multiple sources (Brauchle et al., 1994; Chedid et al., 1994; Tang and Gilchrest, 1996). KGF expression also was induced by serum, or various purified serum growth factors, including plateletderived growth factor BB (PDGF BB) and transforming growth factor (TGF-) (Brauchle et al., 1994; Chedid et al., 1994). These results support the hypothesis that KGF upregulation after tissue injury is initiated by growth factors, such as PDGF BB and TGF-, released from platelets. However, subsequent induction of KGF mRNA presumably is due to the release of proinflammatory cytokines, such as IL-1 and IL-6, from macrophages and polymorphonuclear leukocytes, which infiltrate the wound within 24 h of injury (Werner, 1998). The fact that KGF expression also was increased by phorbol esters and nonhydrolyzable cAMP analogs implied that induction was mediated by at least two different pathways, involving protein kinase C and cAMP-dependent protein kinases (Brauchle et al., 1994). Sex steroid hormones are thought to affect epithelial cells via an indirect mechanism involving an initial interaction with nearby stromal cells. KGF has been implicated as a potential paracrine mediator of steroid hormone action on epithelia in organs of the male and female reproductive tracts, including seminal vesicle, prostate, and endometrium (Alarid et al., 1994; Koji et al., 1994; Sugimura et al., 1996). Androgen stimulated KGF expression in adult rat prostate stromal cells (Yan et al., 1992), as well as human fetal prostate stromal cells in vitro (Levine et al., 1998). FGF10 also was upregulated by androgen in normal prostate stromal cells (Lu et al., 1999). However, the significance of these findings is unclear as subsequent studies indicated that KGF and FGF10 might not be direct targets of androgen activity in vivo (Nemeth et al., 1998; Thomson et al., 1997). Progesterone treatment of rhesus monkeys resulted in a marked elevation in KGF transcript levels in stromal cells of the endometrium (Koji et al., 1994), whereas estrogen was reported to induce KGF expression in mammary stromal cells (Pedchenko and Imagawa, 2000a,b).


85

KGF expression was negatively regulated by glucocorticoids (Brauchle et al., 1995; Chedid et al., 1996; Tang and Gilchrest, 1996). Dexamethasone decreased the transcriptional rate and destabilized the KGF transcript (Chedid et al., 1996). In contrast, nonsteroidal antiinflammatory agents had no effect on KGF synthesis.

C. Gene Structure and Promoter Analysis The FGF genes are highly conserved in their overall structure, consisting of three exons that encode the primary translation product and intervening introns situated in the same locations relative to the protein-coding sequences (Ornitz and Itoh, 2001). The KGF gene conforms to this structure, with introns between the three coding exons positioned after nucleotides 731 and 835 in the cDNA sequence (Kelley et al., 1992). Both the human and rat KGF genes contain an intron in the 50 untranslated region. The intron in the human gene is approximately 650 bp long and located between nucleotides 179 and 180 (Finch et al., 1995b), whereas the intron in the rat gene is 585 bp long and located in a similar position (Fasciana et al., 1996). The locus of the human KGF gene is chromosome 15q13-q22 (Zimonjic et al., 1997). However, a portion of the gene consisting of exon 2, exon 3, the intron between them, and an element encoding 30 noncoding sequence of the KGF transcript was amplified during evolution to approximately 16 copies and dispersed to multiple chromosomes. These KGF-like sequences are transcriptionally active, differentially regulated in various tissues, and comprise three distinct classes of coding sequence that are 5% divergent from each other and from the true KGF gene product (Kelley et al., 1992). However, the functional significance of these sequences is not clear, as it is not known whether they direct protein synthesis. Interestingly, such dispersion of KGF-like genes was evident to a varying extent in the genomic DNAs of chimpanzee, gorilla, and orangutan, but not in gibbon, old world monkeys, mouse, or chicken (Kelley et al., 1992; Zimonjic et al., 1997), reflecting stepwise changes that have occurred in the genomes of the great apes and humans during evolution (Zimonjic et al., 1997). The 50 flanking regions of the human and rat KGF genes have been cloned and the transcription initiation sites determined (Fasciana et al., 1996; Finch et al., 1995b). The human promoter contains two transcription start sites, whereas only one was identified in the rat promoter region. The location of the first (and major) start site in the human promoter is identical to that present in the rat. Putative promoter sequences, TATTTA and CCAAT, were identified 31 and 50 bp upstream from the first mRNA start site. For the human 50 -flanking region, transient transfection analysis identified a basal promoter region located between bases 225 and þ190 (Finch

86


et al., 1995b). Within this region, the human and rat sequences are 78% homologous, but upstream of this point the sequences diverge. The basal promoter from the human gene directed transcription of a chloramphenicol acetyltransferase (CAT) reporter gene in fibroblasts and myoblasts, but no activity was observed when it was transfected into epithelial or lymphoid cells or macrophages. Consistent with cytokine regulation of KGF expression, transcription from the basal KGF promoter was induced by IL-1, IL-6, and forskolin. These results indicated the presence of cis-acting elements that are responsible for selective activation of the KGF promoter only in cells that normally express KGF mRNA (Finch et al., 1995b). Experiments performed with the rat KGF promoter region showed that the segment between bp 1200 and 1900 was involved in upregulation by the synthetic androgen R1881 after transient transfection into LNCaP prostatic cells (Fasciana et al., 1996). Furthermore, a longer construct containing the region up to bp 4700 had significantly higher activity than did the 1900 bp construct, indicating the presence of additional activating sequences. The inducibility of the human KGF promoter region by androgens has not been examined. An in vitro study of the human KGF promoter region focused on determining the role of a novel regulatory element, TGAGGTCAG, in mediating the induction of KGF transcription (Zhou and Finch, 1999). This element is homologous to binding sites for both the ATF/CREB (TGACGTCA) and C/EBP (TGNNGNAAG) families of transcription factors, both of which have been implicated in mediating induction of gene expression in response to extracellular stimuli. This element conferred sensitivity to induction by forskolin when cloned in front of a heterologous simian virus 40 (SV40) promoter. Gel mobility supershift assays indicated that two members of the ATF family, ATF1 and ATF2, were present in the nuclear protein complex bound to this region. Furthermore, purified ATF2 protein bound to the TGAGGTCAG sequence. There was no evidence of C/EBP transcription factors in the complex. Further information concerning the regulation of KGF gene expression came from an in vitro skin model consisting of human keratinocytes cocultured with immortalized fibroblasts derived from either wild-type, c-jun/, or junB/ mouse embryos (Szabowski et al., 2000). The epithelium resulting from coculture with c-jun/ fibroblasts contained a smaller number of cell layers and the number of proliferating cells in the basal layer was markedly reduced compared with skin in which wild-type fibroblasts had been used. Moreover, KGF expression was not detected in the c-jun/ fibroblasts. In contrast, junB/ fibroblasts contained high basal levels of KGF, and they stimulated hyperproliferation of the keratinocytes with an increased number of epithelial cell layers (Szabowski et al., 2000). Previous work had shown that keratinocytes secrete IL-1, which induces KGF


87

expression in fibroblasts, and KGF in turn stimulates keratinocytes (MaasSzabowski et al., 1999, 2000; Smola et al., 1993). Because c-jun/ fibroblasts did not express KGF, even after treatment with IL-1, it appears that c-Jun is not only critical for basal KGF expression, but also for IL-1 induction of KGF transcription (Szabowski et al., 2000). In contrast, the presence of high levels of KGF in junB/ fibroblasts indicated that JunB was a negative regulator of KGF transcription. Two potential AP-1-binding sites have been identified in the KGF promoter region (Finch et al., 1995b), raising the possibility that c-Jun directly activates KGF transcription by binding to these sites.

III. CELLULAR AND MOLECULAR RESPONSES TO KGF Expression studies indicated that KGF might play a role in the regulation of epithelial homeostasis in adult organs, particularly during epithelial repair. The following account summarizes responses to KGF that contribute to homeostatic and repair processes.

A. Mitogenicity Initially, KGF was shown to stimulate DNA synthesis in BALB/MK mouse keratinocytes, B5/589 human mammary epithelial cells, and CCL208 rhesus monkey bronchial epithelial cells (Rubin et al., 1989). Subsequently, additional responsive cells were identified: human keratinocytes, rat and human prostatic epithelial cells, rat hepatocytes, type II alveolar cells, corneal epithelial cells, and bovine granulosa cells (Rubin et al., 1995). No activity was seen on fibroblasts, saphenous vein endothelial cells, melanocytes, or myoblasts (Halaban et al., 1991; Ron et al., 1993b; Rubin et al., 1989). At least in the case of BALB/MK cells, proliferation in serum-free medium required a combination of KGF and insulin or insulin-like growth factors (IGFs) (Rubin et al., 1989). Besides stimulating mitogenic activity in a wide variety of epithelial cell types in vitro, several studies have confirmed that KGF is a powerful mitogen for epithelial cells in many different organ systems in vivo. These latter results are described in detail in Sections V and VI.

B. Motility During wound healing, keratinocyte migration is evident immediately after injury and is initially confined to cells at the wound edge. KGF enhanced the migration of normal human keratinocytes in a dose-dependent

88


manner that was inhibited by incubation with KGF-neutralizing monoclonal antibody (Tsuboi et al., 1993). KGF-treated keratinocytes displayed increased attachment to collagen and fibronectin, and KGF preferentially induced keratinocyte migration on these substrates (Putnins et al., 1999). KGF was subsequently shown to promote type II alveolar motility in a variety of in vitro models (Atabai et al., 2002; Galiacy et al., 2003; Isakson et al., 2001; Waters and Savla, 1999). Keratinocyte migration during reepithelialization is preceded by detachment from the underlying basement membrane and accompanied by successive degradation of the wound clot, dermal material, and granulation tissue. These remodeling processes involve matrix metalloproteinases (MMPs), which degrade essentially all extracellular matrix components. KGF has been reported to enhance the heparindependent synthesis of MMP-1 (collagenase) by keratinocytes (Putnins et al., 1996), and to directly induce the synthesis of MMP-9 (gelatinase) (Putnins et al., 1995), MMP-10 (stromelysin-2) (Madlener et al., 1996), MMP-13 (collagenase-3) (Uitto et al., 1998), and urokinase-type plasminogen activator activity (Putnins et al., 1995; Tsuboi et al., 1993; Zheng et al., 1996). These data suggest that KGF regulates the production of proteases that are active in tissue remodeling during wound healing.

C. Differentiation In a skin equivalent model, KGF induced expression of integrin 5 1, and delayed expression of keratin 10 and transglutaminase, markers of terminal keratinocyte differentiation (Andreadis et al., 2001). KGF stimulated the expression of differentiation-specific markers in human keratinocytes in response to an increase in extracellular calcium, consistent with the idea that KGF was involved in the initiation of the early stages of differentiation (Marchese et al., 1990, 1997). KGF regulates the expression of the gene encoding the estrogen-responsive B box protein (EBBP) in keratinocytes (Beer et al., 2002). In vivo, EBBP is expressed at high levels in basal keratinocytes. Stable overexpression of EBBP in HaCaT keratinocytes enhanced the early differentiation process, and may mediate the differentiation-stimulating activity of KGF under permissive conditions (Marchese et al., 1990). KGF also influences epithelial differentiation in the lung. Type I alveolar cells are flat, relatively metabolically inactive cells that are primarily involved in gas exchange. Type II cells are cuboidal, metabolically active cells that synthesize and secrete surfactant protein. After lung injury the type II cells proliferate, line the alveolar septae, and then differentiate into type I cells to facilitate the reconstitution of normal alveolar parenchymal


89

architecture. Type II cells are also prominent in the fetal lung, where they are thought to play a role in lung morphogenesis. In a cell culture system in which type II alveolar cells develop characteristics of type I cells over time, incubating cells in the presence of KGF maintained the type II phenotype (Borok et al., 1998b; Isakson et al., 2001; Rice et al., 2002). Addition of KGF-neutralizing antibody to isolated lung epithelium decreased fibroblast conditioned medium-stimulated surfactant synthesis by 50% (Chelly et al., 2001). Furthermore, administration of KGF to fetal rat type II cells in various mesenchymefree culture systems induced distal epithelial differentiation and expression of a number of genes encoding surfactant proteins (Cardoso et al., 1997; Chelly et al., 1999; Deterding et al., 1997; Mason et al., 2002; Sugahara et al., 1995; Xu et al., 1998b). Expression of the surfactant-A (SP-A) and surfactant-D (SP-D) genes is regulated in part by the C/EBP transcription factor (He and Crouch, 2002; Li et al., 1995). KGF increased the expression of C/EBP in type II alveolar cells (Mason et al., 2003). In addition to the surfactant proteins, surfactant also contains phospholipid. KGF stimulated lipogenesis in type II cells and the subsequent conversion of the newly synthesized fatty acids into phospholipids (Mason et al., 2003). The induction of the genes for fatty acid synthase, stearoyl-CoA desaturase-1, and epidermal fatty acidbinding protein, mediated in part by increased expression of C/EBP, as well as the transcription factors C/EBP and SREBP-1, likely contributed to the increased fatty acid synthesis in KGF-treated type II cells (Mason et al., 2003). Administration of recombinant KGF to adult rats caused a marked and selective induction of mucin-producing goblet cells throughout the GI tract (Housley et al., 1994). Members of the intestinal trefoil factor (ITF) family of proteins are selectively expressed in intestinal goblet cells and their expression correlates with intestinal goblet differentiation. In a study of the H2 subclone of the human HT29 colonic epithelial cell line, KGF promoted H2 differentiation into goblet cells as reflected by increased ITF expression (Iwakiri and Podolsky, 2001). Furthermore, KGF regulated mouse ITF transcription through the goblet cell silencer inhibiter, which is essential for goblet cell-specific expression of ITF (Iwakiri and Podolsky, 2001). In vivo, KGF enhanced expression of ITF2 and ITF3 throughout the intestine (Fernandez-Estivariz et al., 2003). KGF stimulated the proliferation of pancreatic exocrine cells and cytodifferentiation to a ductal epithelial phenotype (Miralles et al., 1999; Yi et al., 1994b), while repressing the formation of endocrine cells (Elghazi et al., 2002; Yi et al., 1994b). After KGF was removed from the culture medium, pancreatic precursor cells differentiated into endocrine cells (Elghazi et al., 2002). Thus, KGF may be useful in expanding the population of precursor cells that subsequently can be induced to form insulin-expressing cells.

90


D. Branching Morphogenesis Epithelial branching morphogenesis is a highly coordinated process involving cell proliferation, cell–cell and cell–matrix interactions, and remodeling of the basement membrane. A number of studies have demonstrated that KGF can stimulate growth and branching morphogenesis in various fetal explants including seminal vesicle (Alarid et al., 1994), prostate (Sugimura et al., 1996), and pancreas (Miralles et al., 1999). However, as discussed elsewhere in this review, FGF10 rather than KGF is the predominant mesenchymal mediator of epithelial branching morphogenesis during organogenesis. Nonetheless, KGF may provide a proliferative signal that contributes to the formation of branches during development (Park et al., 1998; Post et al., 1996).

E. Antiapoptotic Effects In postconfluent cultures of normal human keratinocytes, KGF promoted tight packing of cells characterized by a small basal cell morphology, suggesting that it could prevent terminal differentiation and/or apoptosis (Hines and Allen-Hoffmann, 1996). In support of this hypothesis, these cultures produced fewer cross-linked cell envelopes, and exhibited less membrane-associated transglutaminase activity and nucleosomal fragmentation compared with untreated cultures (Hines and Allen-Hoffmann, 1996). KGF-treated hepatocytes exhibited decreased apoptosis in response to actinomycin D and tumor necrosis factor (TNF) than did control cells, or cells treated with hepatocyte growth factor or epidermal growth factor (Senaldi et al., 1998). In a mouse model of total parenteral nutrition, which is associated with high levels of intestinal epithelial cell apoptosis, administration of KGF decreased apoptosis and increased expression of antiapoptotic Bcl-2 proteins (Wildhaber et al., 2003). In the lung, KGF administration induced type II alveolar cell proliferation. On withdrawal of KGF the hyperplastic type II cells underwent apoptosis, and a normal alveolar epithelium consisting primarily of type I cells was restored (Fehrenbach et al., 1999). Therefore, KGF not only stimulated the proliferation of type II cells, but maintained their phenotype by preventing both apoptosis and terminal differentiation into type I cells. KGF also inhibited hyperoxia-induced apoptosis of alveolar epithelial cells, probably by repressing the expression of apoptotic mediators such as p53, p21, Bax, and Bcl-x (Barazzone et al., 1999; Buckley et al., 1998). In a mouse model of oxidant-induced lung injury, expression of KGF from a tetracycline-inducible, lung-specific transgenic promoter protected the lung epithelium from the hyperoxic insult. Furthermore, KGF induced the


91

antiapoptotic Akt pathway, and inhibition of this activation by expression of a dominant-negative Akt mutant blocked KGF-mediated protection of the epithelium (Ray et al., 2003). In another study, FGFR2b was shown to interact with p21-activated protein kinase 4 (PAK4), a newly identified member of the PAK family of proteins that are regulated by the Rho family GTPases Rac and Cdc42. A dominant-negative PAK4 mutant blocked KGF inhibition of caspase-3-dependent apoptosis in epithelial cells subjected to oxidant stress (Lu et al., 2003). Thus, Akt and PAK4 appear to be important mediators of the antiapoptotic activity of KGF. As discussed previously, KGF was upregulated by progesterone in primate endometrium during the menstrual cycle (Koji et al., 1994). At the end of the cycle, withdrawal of progesterone induces the luteal–follicular transition (LFT), which is marked by menstrual sloughing, apoptotic regression of the basalis zone, and a 9-fold decrease in the abundance of KGF transcripts. Addition of exogenous KGF during the LFT inhibited apoptosis in the basalis zone, and had a marked trophic effect on the spiral arteries (Slayden et al., 2000). Therefore, the progesterone-dependent increase in KGF expression may serve to inhibit glandular apoptosis during the nonfertile menstrual cycle.

F. Cytoprotection The marked induction of KGF after injury raised the possibility that in addition to participating in the physical repair of wounded tissue, KGF might also have cytoprotective activities that would limit cellular damage from external insults and the associated inflammatory response. For example, hyperoxic injury to the lung results in alveolar hemorrhages, exudates, and inflammatory infiltrates (Panos et al., 1995). However, administration of exogenous KGF prevented or attenuated much of the injury caused by this toxic exposure (see Section VI.B). DNA strand break is one of the earliest abnormalities that occurs in cells exposed to oxidative stress and can lead to cell death by apoptotic and necrotic pathways. In vitro, KGF diminished the formation of strand breaks in A549 alveolar epithelial cells exposed to radiation (Takeoka et al., 1997) or hydrogen peroxide (Wu et al., 1998). The protective effect of KGF on these cells was blocked by inhibitors of DNA polymerases , , and " (Takeoka et al., 1997; Wu et al., 1998), and inhibitors of protein kinase C and tyrosine kinases (Wu et al., 1998). Addition of KGF to cultures of type II cells isolated from hyperoxic rats resulted in a significantly decreased number of oxygen-induced DNA strand breaks compared with control cells (Buckley et al., 1998). KGF also had cytoprotective effects on epithelial cells in the GI tract of whole animals, as discussed in Section VI.D.

92


The molecular mechanisms of the cytoprotective effects of KGF include the induction of genes encoding enzymes that limit oxidative stress. For instance, in HaCaT keratinocytes KGF upregulated the expression of a nonselenium glutathione peroxidase, known as peroxiredoxin VI, which plays a role in the detoxification of hydrogen peroxide and organic peroxides (Frank et al., 1997). Furthermore, in mice treated with KGF, increased expression of peroxiredoxin VI and glutathione-S-transferase (GST) was observed in the oral squamous epithelium and the intestine, indicating that KGF can regulate expression of these genes in different tissues in vivo (Farrell et al., 2002; Jonas et al., 2000). GST also inactivates reactive oxygen species and is a target gene for Nrf2, a transcription factor that interacts with leucine zipper proteins to bind cis-acting, antioxidant response elements in the promoters of genes that encode cytoprotective enzymes (Mulcahy et al., 1997; Rushmore et al., 1990; Venugopal and Jaiswal, 1996, 1998). Nrf2, itself, is a target of KGF action (Braun et al., 2002). Whereas Nrf2 was expressed at high levels in basal keratinocytes of hyperproliferative wound epithelium, Nrf2 knockout mice exhibited no delays in wound healing (Braun et al., 2002). This may be explained by upregulation of the related Nrf3 transcription factor, which also appeared to be induced by KGF (Braun et al., 2002).

IV. SIGNAL TRANSDUCTION Whereas significant progress has been made in delineating FGF-mediated signal transduction pathways, there have been relatively few studies that have specifically examined KGF signaling. However, the data currently available suggest that KGF signaling involves pathways common to other FGFs. In this section we provide a brief overview of FGF signaling pathways, and then review data that specifically pertain to KGF. FGF signaling is initiated when high-affinity binding of an FGF ligand and its cognate receptor induces receptor dimerization, activation of its intrinsic tyrosine kinase activity, and autophosphorylation of the cytoplasmic domains of the receptor (Powers et al., 2000). Autophosphorylation sites located within the catalytic core are involved in regulation of kinase activity, whereas autophosphorylation in other regions of the cytoplasmic domain increases the affinity of binding sites for Src homology 2 (SH2) or phosphotyrosinebinding (PTB) domains of effector proteins. On activation, tyrosinephosphorylated FGFRs function as platforms for the assembly of a variety of signaling proteins, including phospholipase C (Mohammadi et al., 1991) and Crk, an SH2/SH3-containing adaptor protein that may link FGFR to the downstream signaling molecules Shc, C3G, and Cas (Larsson et al., 1999).


93

Recruitment of signaling proteins in response to receptor stimulation is also facilitated by indirect mechanisms involving membrane-linked docking proteins. FGF stimulation leads to tyrosine stimulation of two highly homologous 90-kDa proteins identified as SNT-1 or FRS2 (Kouhara et al., 1997; Wang et al., 1996) and FRS2 (Ong et al., 2000). The FRS2/SNT proteins are targeted to the plasma membrane by myristoylation at the N terminus and contain a PTB domain that mediates interaction with FGF or nerve growth factor receptors (Dhalluin et al., 2000; Meakin et al., 1999; Ong et al., 2000; Xu et al., 1998a). The C-terminal region of FRS2 contains multiple tyrosine residues that are phosphorylated by active FGFRs, and serve as recognition motifs for SH2 domains of Grb2 and Shp2 (Hadari et al., 1998; Kouhara et al., 1997). The formation of the FRS2–Grb2 complex results in the recruitment of Ras to the plasma membrane, and its activation by Sos-dependent exchange of GDP for GTP. This in turn activates a signaling cascade consisting of the serine/threonine kinase Raf, the dual-specificity mitogen-activated protein kinase (MAPK) kinase MEK, and the MAPK isoforms Erk1 and Erk2. Alternatively, the assembly of a complex containing FRS2, Grb2, and Gab1 induced by FGF stimulation results in activation of phosphatidylinositol 3-kinase (PI-3K) and downstream effector proteins such as the serine/threonine kinase Akt, whose cellular localization and activation are regulated by products of PI-3K enzymatic activity (Ong et al., 2000). A group of four mammalian orthologs of the Drosophila Sprouty (Spry) gene have been identified as ligand-induced, feedback inhibitors of FGFRmediated, MAPK signaling (Casci et al., 1999; Gross et al., 2001; Hacohen et al., 1998; Kramer et al., 1999; Reich et al., 1999; Yusoff et al., 2002). Sprys translocate to the plasma membrane in response to FGF signaling, and have many different binding partners, including various effectors of the MAPK pathway (Impagnatiello et al., 2001; Leeksma et al., 2002; Lim et al., 2000; Sasaki et al., 2003; Wong et al., 2001, 2002). Sprys require tyrosine phosphorylation to interact with their binding partners (Fong et al., 2003; Hanafusa et al., 2002). Mouse Spry2 inhibited FGF10-activated MAPK signaling by differentially binding to upstream target proteins (Tefft et al., 2002). Sprys are abundantly expressed in the epithelium during the early stages of branching morphogenesis (Hashimoto et al., 2002), and mSpry2 functions as a negative regulator of embryonic lung branching morphogenesis and growth (Mailleux et al., 2001). KGF treatment of BALB/MK cells elicited the rapid tyrosine phosphorylation of a p90 protein that in all likelihood corresponded to the SNT/FRS2 proteins (Bottaro et al., 1990). Consistent with other FGFs, KGF administration caused a transient activation of MAPK in corneal epithelial cells (Liang et al., 1998), human epidermal keratinocytes (Zeigler et al., 1999), prostate DU145 cells (Mehta et al., 2001), and human endometrial

94


carcinoma cells (Taniguchi et al., 2003). In keratinocytes the transient nature of the MAPK activation did not correlate with the sustained activation required for cell movement and MMP9 production, cellular responses that are integral to the process of invasion (Zeigler et al., 1999). However, in human endometrial carcinoma cells, blocking the MAPK pathway completely neutralized the proliferative effect of KGF (Taniguchi et al., 2003). KGF also stimulated a 2- to 3-fold increase in PI-3K activity in corneal epithelial cells, and phosphorylation of its downstream target, p70 S6K, with a corresponding increase in its activity (Chandrasekher et al., 2001). As mentioned previously, Akt and PAK4 were required for KGF inhibition of apoptosis in different models of oxidant-induced cell injury (Lu et al., 2003; Ray et al., 2003).

V. TRANSGENIC AND KNOCKOUT MODELS A number of studies have examined the consequences of targeting KGF expression to various epithelial cell populations, using tissue-specific promoters in transgenic models. When KGF was targeted to the epidermis, using the keratin-14 promoter, the mice were weak and frail, and exhibited grossly wrinkled skin. There was an increase in epidermal thickness accompanied by alterations in epidermal growth and differentiation (Guo et al., 1993). Furthermore, hair follicle morphogenesis was suppressed in these animals, as was adipogenesis. With age, gross transformations in the epidermis and tongue epithelium developed, and animals exhibited increased salivation and altered differentiation of salivary glands. Similarly, when KGF was overexpressed in the embryonic liver of transgenic mice, using an apolipoprotein E promoter, there was marked epidermal papillomatous acanthosis and hyperkeratosis in the skin, with a notable decrease in the number of developing hair follicles. These animals also were characterized by marked hyperplasia and cystic dilation of the cortical and medullary kidney collecting duct system, a phenotype resembling infantile polycystic kidney disease in humans (Nguyen et al., 1996). Overexpression of KGF in the mouse lung epithelium either constitutively (Simonet et al., 1995) or conditionally (Tichelaar et al., 2000) caused a pulmonary malformation that resembled pulmonary cystadenoma in humans. The embryonic lungs had dilated saccules lined with columnar epithelial cells and normal alveolar structure. Embryos constitutively expressing KGF in the lung epithelium died before reaching term (Simonet et al., 1995). Ectopic expression of KGF in pancreatic beta cells resulted in significant intraislet duct cell proliferation, and the appearance of hepatocytes within the islets of


95

Langerhans (Krakowski et al., 1999). In transgenic mice with KGF overexpressed in the eye, hyperproliferation was observed in embryonic corneal epithelial cells that subsequently differentiated to form functional lacrimal gland-like structures (Lovicu et al., 1999). The interpretation of these transgenic studies with regard to the normal function of KGF is unclear, because endogenous KGF expression is under regulatory control and usually does not occur in responsive epithelial cells. Perhaps the major conclusion from these investigations is that KGF has strong activity in a variety of epithelial cells in vivo. Persistent production of KGF in transgenic animals caused permanent histological abnormalities in targeted tissues. An extreme example was observed in mice that expressed KGF as a transgene under the control of the mouse mammary tumor virus long terminal repeat (MMTV LTR). These animals were characterized by mammary epithelial cell hyperplasia and, after multiple pregnancies that presumably induced high levels of transgenic KGF expression, they developed mammary adenocarcinomas (Kitsberg and Leder, 1996). In contrast, when recombinant KGF was given to animals, the histological changes seen in multiple epithelial cell lineages were rapidly reversed after cessation of KGF treatment (Housley et al., 1994; Ulich et al., 1994; Yi et al., 1994a,b, 1995). This was consistent with the idea that physiologic mechanisms exist to mediate KGF responses and to resolve them in accordance with homeostasis. KGF knockout mice were viable and appeared to be essentially normal (Guo et al., 1996). However, further examination revealed subtle phenotypes involving the hair, kidneys, and bladder. Over time, the fur of mice lacking KGF developed a matted and greasy appearance, similar to the rough mouse, whose recessive mutation maps at or near the KGF locus on mouse chromosome 2. This defect appeared to be restricted to the cells giving rise to the hair shaft (Guo et al., 1996). The developing ureteric bud and mature renal collecting system of KGF knockout mice were markedly smaller than in wild-type kidneys (Qiao et al., 1999), consistent with an earlier study showing the responsiveness of urothelium to recombinant KGF (Yi et al., 1995). Furthermore, mature kidneys in the knockout mice had 30% fewer nephrons than wild-type kidneys (Qiao et al., 1999). In vitro experiments demonstrated that KGF augmented ureteric bud growth and, through associated inductive mechanisms, increased the number of nephrons that formed in rodent metanephric kidney organ cultures (Qiao et al., 1999). These results indicated that KGF modulated the extent of ureteric bud growth during development and the number of nephrons that eventually formed in the kidney. In KGF knockout mice the bladder urothelium was markedly thinner than that of wild-type mice and lacked the intermediate cell layers present in wild-type animals (Tash et al., 2001). Primary

96


urothelial cell cultures maintained without KGF stopped dividing, and expressed markers associated with terminally differentiated umbrella cells (Tash et al., 2001). Thus, KGF may be required for normal bladder urothelial stratification, and delay the differentiation of urothelial cells into post-mitotic umbrella cells. The importance of FGFR signaling for epidermal wound healing was demonstrated in mice expressing a dominant-negative FGFR2b derivative in basal keratinocytes of the epidermis. These animals displayed a severely reduced rate of keratinocyte proliferation at wound edges, resulting in delayed reepithelialization (Werner et al., 1994). In contrast, KGF knockout mice exhibited no abnormalities in epidermal wound healing (Guo et al., 1996), suggesting that despite its strong induction after injury, other FGFR2b ligands could compensate for its absence. In this regard, FGF10 was detected in the wound epidermis of KGF knockout mice, suggesting that it might substitute for KGF in the healing of skin wounds (Jameson et al., 2002). However, it should be noted that only full-thickness incisional wound healing was examined in KGF null mice (Guo et al., 1996). Excisional wound repair, which requires a far greater extent of reepithelialization than incisional wounds, was not studied in these mice (Werner and Grose, 2003). Interestingly, FGF22 was strongly expressed in the wound epidermis of wild-type mice during the final stages of healing (Beyer et al., 2003). This raises the intriguing possibility that FGF22 may be an FGFR2b antagonist whose function is to inhibit the proliferative effects of KGF and FGF10 once tissue repair is near completion. However, verification of this hypothesis will require more information about the receptor-binding and biological activity of FGF22. In contrast to results in the skin injury model, KGF knockout mice exhibited more severe inflammation in the colon and a delay in tissue repair compared with wild-type mice after treatment with dextran sodium sulfate, a model for IBD (Chen et al., 2002). This implied that KGF had a specific, nonredundant role in limiting damage to the intestine. The absence of gross developmental abnormalities in the KGF knockout mouse demonstrated that KGF did not have an important role in organogenesis. However, expression of a soluble dominant-negative FGFR2b receptor under the control of a metallothionein promoter resulted in markedly impaired development of many epithelial organs including the kidney, lung, various cutaneous structures, and exocrine and endocrine glands (Celli et al., 1998). The mouse embryos also displayed severe defects in craniofacial and limb formation. This implied that FGF ligands that could bind FGFR2b had a critical role in these developmental processes. Strikingly, the phenotype of FGF10 knockout mice exhibited many of these traits (Ohuchi et al., 2000; Sekine et al., 1999). FGF10 knockout mice died at birth as a result of a lack of lung development. Although the trachea was normal, subsequent


97

pulmonary branching was minimal (Sekine et al., 1999). Limb bud development was markedly impaired because of the failure to form an apical ectodermal ridge. When expression of FGFR2b isoforms was specifically disrupted in mice, defects were seen in many organs (De Moerlooze et al., 2000; Petiot et al., 2003; Revest et al., 2001). These studies established that FGF10–FGFR2b signaling has an essential role in organogenesis and limb formation, whereas KGF appears to have a more significant role in epithelial homeostasis later in life.

VI. REGENERATIVE AND PROTECTIVE EFFECTS As reviewed in the previous sections, much information has been obtained about KGF activity from a combination of in vitro and in vivo investigations. Taken together, these studies validated the concept that epithelial cells comprise the major targets of KGF action. Moreover, the multiplicity of KGF effects, including proliferation, migration, differentiation, cytoprotection, and inhibition of apoptosis, provides a mechanistic basis for its putative role as a homeostatic factor whose primary function is to maintain the integrity of epithelial tissues. The fact that KGF was upregulated after epithelial injury further suggested that it participates in normal tissue repair. This role distinguishes KGF from most other cytokines and provided a focus for efforts to identify clinical applications. It was hypothesized that exogenous KGF would augment the effects of the endogenous protein when epithelial tissues were at risk, and preservation or rapid restoration would be of clinical benefit. It was further surmised that, given its specificity for epithelial cells, this could be achieved without stimulating adverse inflammatory responses, fibrosis, or angiogenesis. This section summarizes the major findings from several animal models that have been used to examine the protective and regenerative properties of KGF.

A. Skin The first experiments performed with KGF in animals involved its topical application to skin with the aim of determining its efficacy in stimulating epidermal wound repair. KGF increased the rate of reepithelialization in partial-thickness wounds of the porcine epidermis (Staiano-Coico et al., 1993). In both partial- and full-thickness wounds there was a marked increase in the thickness of the epidermis after KGF treatment. In fullthickness wounds, this was associated with a deep rete ridge pattern, an increase in the number of serrated basal cells and increased deposition of

98


collagen fibers in the superficial dermis. Electron microscopy indicated that the serrated cells had better developed hemidesmosomes and thicker bundles of tonofilaments. These structural features suggested that KGF would enhance the strength and durability of healed wounds (Staiano-Coico et al., 1993). A similar stimulation of reepithelialization and epidermal thickness was observed in wounds extending through the cartilage of rabbit ear (Pierce et al., 1994). In addition, KGF increased the proliferation and differentiation of progenitor cells of hair follicles and sebaceous glands in the wound site. A subsequent study documented the expression of FGFR2b in basal keratinocytes and throughout the developing hair follicles, while KGF was detected in follicular dermal papillae of rat embryos and neonates (Danilenko et al., 1995b). Moreover, subcutaneous or intraperitoneal administration of KGF in nu/nu athymic nude mice stimulated hair growth and sebaceous gland hypertrophy. In addition, KGF treatment 1 day before cytosine arabinoside reduced the extent of chemotherapy-induced alopecia in neonatal rats by 50% (Danilenko et al., 1995b). When KGF was tested in porcine skin models of full- and partial-thickness burns, increased epidermal thickness and follicular proliferation were again observed (Danilenko et al., 1995a). Although the rate of reepithelialization also was increased, the effect was not marked. While these experiments established that recombinant KGF had significant biological effects on skin, the clinical benefit of these effects on wound healing remains uncertain. However, a report that local injection of liposomes containing KGF cDNA caused significant improvements in epidermal regeneration suggests another possible therapeutic approach with KGF to enhance wound healing (Jeschke et al., 2002). KGF was shown to increase hair follicle survival and regeneration after irradiation (Booth and Potten, 2000). KGF pretreatment increased the expression of p21, a protein that interrupts cell cycle progression. Thus, KGF may induce a more rapid onset or prolonged duration of cell cycle arrest that could allow for more efficient DNA repair and/or a decline in apoptosis within the hair follicle.

B. Lung Because KGF was originally purified from the conditioned medium of a human embryonic lung fibroblast cell line (Rubin et al., 1989) and was subsequently shown to be a mitogen for type II alveolar cells in vitro (Panos et al., 1993), it was reasoned that the lung epithelium might represent a therapeutic target for KGF action. Normal adult lung expresses both KGF and FGFR2b mRNAs, suggesting that endogenous KGF is involved in the normal growth of type II alveolar cells (Ulich et al., 1994). Intratracheal administration of recombinant KGF to lungs of adult rats resulted


99

in a dose-dependent proliferation of type II alveolar cells. This appeared histologically as a micropapillary epithelial cell hyperplasia in the form of monolayers of cuboidal epithelial cells lining the alveolar septae. The hyperplastic cells contained immunoreactive surfactant protein B and lamellar inclusions characteristic of surfactant-producing type II alveolar cells (Ulich et al., 1994). A series of studies using rodent models established that KGF has remarkable cytoprotective effects in response to acute lung injury. Intratracheal administration of KGF was shown to provide significant protection from a variety of toxic exposures, including hyperoxia (Barazzone et al., 1999; Guo et al., 1998; Panos et al., 1995), acid instillation (Yano et al., 1996), -napthylthiourea (ANTU, model of increased permeability pulmonary edema) (Guery et al., 1997; Mason et al., 1996), radiation (Yi et al., 1996), and bleomycin (Sugahara et al., 1998; Yi et al., 1996, 1998). Similar results were observed when KGF was administered intravenously, rather than intratracheally, in studies of bleomycin- and hyperoxia-induced lung injury (Guo et al., 1998) as well as ventilator-induced lung injury (Welsh et al., 2000). In these models, KGF-treated animals exhibited a dramatic, dose-dependent decrease in mortality and morbidity, as measured by criteria such as body weight and pulmonary function tests. In addition, KGF reduced the incidence or severity of histological changes associated with injury, including fibrosis, and physiological indices of lung injury, such as formation of pulmonary edema. Somewhat unexpectedly, KGF also reduced endothelial cell injury associated with ANTU (Guery et al., 1997; Mason et al., 1996) and hyperoxia (Barazzone et al., 1999). This beneficial effect on the endothelium may have been due to indirect mechanisms, as maintenance of the epithelium markedly limited the inflammatory response after toxic exposure. The results from preclinical models have prompted the suggestion that KGF should be tested in clinical trials involving lung injury (Ware and Matthay, 2002). The efficacy of intravenous KGF in some of the animal models was noteworthy, as this would provide a more convenient route of administration than intratracheal delivery. However, the beneficial effects observed in lung injury models typically required pretreatment with KGF. Because the development of acute lung injury in humans is often unpredictable, KGF treatment before the initial insult usually would not be feasible. However, before dismissing the use of KGF in these clinical settings, it is important to acknowledge that in most of the mouse injury models, because of the small size of the lungs, tissue damage occurred throughout the organs. In the corresponding human conditions, damage often begins in localized regions of the lung. Whether prompt KGF administration would protect the surrounding tissue and thereby limit the extent of tissue damage remains a theoretical possibility that warrants further consideration.

100


C. Bladder Systemic administration of KGF to rats and rhesus monkeys caused rapid and striking proliferation of the urothelium (Yi et al., 1995). Subsequently, KGF was shown to ameliorate cyclophosphamide (CP)-induced ulcerative hemorrhagic cystitis in rats. In this model, a single intravenous injection of KGF administered 24 h before treatment with CP almost completely prevented CP-induced ulcerative hemorrhagic cystitis (Fig. 3) (Ulich et al., 1997). The protective effect probably was due to the rapid proliferation of urothelial cells during and after chemotoxic injury, or to cytoprotective mechanisms (Ulich et al., 1997). Hemorrhagic cystitis occurs sporadically in patients after radiation therapy, but is most commonly observed after treatment with CP and its more urotoxic derivative, ifosfamide. They are thought to induce urothelial injury by multiple mechanisms that result from the accumulation of high urinary concentrations of these drugs and their highly reactive metabolites (Wagner, 1994). The incidence of CP-induced hemorrhagic cystitis varies from 2 to 40% depending on dose and duration of therapy (Foad and Hess, 1976). This study provided a clear demonstration that KGF could be of benefit in ameliorating the toxic effects associated with cancer treatment regimens, and as such foreshadowed the development of this application as the primary clinical indication for KGF. Importantly, the opportunity to give KGF before the chemotherapy regimen maximized its protective effects.

D. Gastrointestinal Tract KGF and FGFR2b mRNAs are expressed throughout the entire GI tract, indicating that the gut can both synthesize and respond to KGF (Housley et al., 1994). Administration of recombinant KGF to adult rats caused a marked increase in the proliferation of epithelial cells from the foregut to the colon, including hepatocytes. There also was a selective induction of mucin-producing goblet cells throughout the GI tract (Housley et al., 1994). The proliferative changes in the crypts of the small intestine that resulted from prolonged exposure to KGF appeared to be related to an increase in stem cell numbers and/or increases in the number of stem cells in the S phase of the cell cycle (Potten et al., 2001). Reports that KGF was upregulated in IBD implied that endogenous KGF was at least partially responsible for the intestinal crypt hyperplasia seen in the regenerative response to this chronic condition (Bajaj-Elliott et al., 1997; Brauchle et al., 1996; Finch et al., 1996). Subsequent work demonstrated that exogenous KGF ameliorated mucosal injury in several models of colitis (Byrne et al., 2002; Egger et al., 1999; Zeeh et al., 1996). As previously noted, KGF null mice were more susceptible than their wild-type


101

Fig. 3 Representative bladder H&E cross-sections from saline- or KGF-pretreated cyclophosphamide-challenged rats. Specimens from saline-pretreated rats (left) were characterized by severe hemorrhagic lesions (arrows) and extensive submucosal edema (*). The bladders of rats pretreated with KGF (5 mg/kg) 24 h before cyclophosphamide administration were largely free of ulcerations and edema (right). (From Ulich et al., 1997, Fig. 1, p. 473.)

102


Fig. 4 KGF increased survival of mice treated with high doses of chemotherapy and/or irradiation. BDF1 mice received (A) 5-fluorouracil (5-FU, 50 mg/kg per day, intraperitoneal) for 4 days (days 1–4; n ¼ 20/group); (B) methotrexate (MTX, 150 or 300 mg/kg, intraperitoneal) 1 h before receiving 6 Gy of radiation (day 1; n ¼ 20/group); or (C) 12 Gy of radiation from a cesium source (day 1) followed by a bone marrow transplant (BMT; n ¼ 20/group), and


103

counterparts to dextran sulfate-dependent colonic injury, and healing was delayed in the absence of KGF (Chen et al., 2002). This implied that KGF has a specific role in mucosal protection and repair, not redundant with FGF10 or other FGFs. Additional reports provided evidence that KGF treatment enhanced the repair of colonic anastomoses in rats (Egger et al., 1998, 2001) and reduced the extent of intestinal atrophy in animals maintained on total parenteral nutrition (Yang et al., 2002). Cytoablative doses of cancer chemotherapy and irradiation compromise the absorptive and barrier functions of the GI tract by killing rapidly dividing cells of the mucosa, thereby impairing normal cell renewal. Such treatments are often associated with mucositis, which is characterized by atrophy, ulceration, loss of barrier function, and infection. The beneficial effects of KGF in reducing chemotherapy and radiation-induced GI injury and mortality were first documented in a study that involved several mouse models (Farrell et al., 1998). Treatment with KGF (5 mg/kg per day for 3 days) before four daily doses of 5-fluorouracil (5-FU, 50 mg/kg per day) resulted in a striking improvement in survival (87 versus 27%, p < 0.006; Fig. 4A) and significant reduction in weight loss (p < 0.0001; Fig. 5A). However, KGF was not effective if given after 5-FU. The protective effect of KGF was not limited to a particular chemotherapeutic agent, as KGF pretreatment also reduced the weight loss associated with a single dose of carboplatin (Fig. 5B). In addition, KGF caused a marked improvement in survival of mice treated with a combination of total body irradiation (6 Gy) and either low (150 mg/kg) or high (300 mg/kg) doses of methotrexate (Fig. 4B). A decrease in weight loss of KGF-treated mice also was noted in this experiment (Fig. 5C). The most dramatic results were observed when mice were exposed to a lethal dose of total body irradiation (12 Gy) and supplemented with a bone marrow transplant. Whereas all the mice in the vehicle control group died within 8 days, 90% of the mice pretreated with KGF survived this toxic regimen (Fig. 4C). Additional studies have focused on the ability of KGF to palliate the effects of aggressive cancer treatment regimens in the upper aerodigestive tract. Cells in the basal germinal layer of this tissue proliferate at a high rate. The reproductive capacity of the basal cells is impaired after radiation and chemotherapy, resulting in desquamation, ulceration, and subsequent infection as the epithelial barrier is lost. When mice were treated with either a single dose of 12 Gy or four daily doses of 4 Gy, there was a 30–40% decrease in average epithelial thickness of the tongue, esophagus, and

were pretreated with KGF (5 mg/kg per day) or vehicle for 3 days. Statistical differences assessed by Kaplan–Mayer survival analysis proved that in all of these experiments, pretreatment with KGF significantly reduced mortality. (From Farrell et al., 1998, Fig. 1, p. 934.)

104


Fig. 5 KGF reduced weight loss in mice treated with chemotherapy and/or irradiation. BDF1 mice received (A) 5-fluorouracil (5-FU, 50 mg/kg per day, intraperitoneal) for 4 days (days 1–4; n ¼ 15/group); (B) carboplatin (125 mg/kg, intraperitoneal) (day 1; n ¼ 10/group); or (C) methotrexate (MTX, 150 or 300 mg, intraperitoneal) 1 h before receiving 6 Gy of radiation


105

buccal mucosa 4 days after irradiation (Farrell et al., 1999). This atrophy was reversed by KGF. Notably, in this model of tissue irradiation KGF was most effective when given after or during the toxic exposure, in contrast to the lung and the intestine where pretreatment was necessary for protection. Further analysis was performed with an experimental model in which a 3 3 mm2 area on the ventral surface of mouse tongue (strain C3H/Neu) was exposed to graded doses of X-rays, and the radiation dose resulting in ulcer formation in 50% of the mice (ED50) was determined. The ED50 for single dose irradiation was 10.9 0.7 Gy without KGF treatment. In contrast, when KGF was given subcutaneously at 5 mg/kg per day for 3–6 days the ED50 increased by a factor of 1.5 to 2.3. The maximum protective effect was observed when KGF was administered either from days 3 to þ1 or from days 0 to þ2, with an ED50 of 23.1 5.7 or 24.9 4.6 Gy, respectively (p < 0.0001 for both) (Do¨ rr et al., 2001). Similar beneficial effects were observed when KGF was administered to mice receiving five daily doses of 3 Gy (Do¨ rr et al., 2002a). Whereas mice became more radioresistant when treated with KGF before irradiation, maximal protection was obtained when KGF was given either during or both during and after irradiation, with the ED50 again more than doubling. Moreover, a single KGF injection administered either shortly before (day 1) or during (þ4) the irradiation period (days 0–4) was similarly (5 mg/kg) or more (15 mg/kg) effective than multiple daily KGF doses. KGF treatment was not beneficial if initiated after ulceration already was present (typical latency period for ulcer formation in this model was 10 days) (Do¨ rr et al., 2002b). Nonetheless, these data provided compelling evidence that KGF can reduce the damaging effects of X-rays in the oral cavity, and indicated that, in contrast to most other toxic exposures, prior treatment with KGF was not required to obtain mucoprotection.

E. Graft-versus-Host Disease and Thymus Allogeneic bone marrow transplantation (BMT) is used in the treatment of a number of malignancies. Graft-versus-host disease (GVHD) represents a major transplant-related complication initiated when alloreactive T cells

from a cesium source (day 1; n ¼ 20/group), and were pretreated with KGF (5 mg/kg per day) or vehicle for 3 days. The changes in weight in (A) and (C) were analyzed by calculation of the percentage of change from baseline, using a linear mixed model for repeated measures. It includes the fixed effects of treatments, days of measurements, and the interaction treatment by day. Bars represent the standard error. In (B), the mean weights at the nadir were compared by Student t test. The data are expressed as means SE. In all of these experiments, pretreatment with KGF significantly reduced weight loss induced by cytoablative exposure. (From Farrell et al., 1998, Fig. 2, p. 935.)

106


recognize histocompatibility antigens of the host tissue. Activation of inflammatory effector cells and secretion of cytopathic molecules result in tissue damage to the skin, GI tract, liver, lung, and immune system. Common prophylaxis includes the elimination of donor T cells from the host, or immunosuppression of the host that can result in the loss of graft-versusleukemia activity and a high risk of relapse. Damage to the GI tract is thought to have a pivotal role in the pathophysiology of GVHD, augmenting the release of cytokines that promote inflammation and cytotoxic immune cell activities (Hill and Ferrara, 2000; Krenger et al., 1997). By reducing injury to the GI tract, particularly damage resulting from the chemoradiotherapyconditioning regimens that precede BMT, it has been proposed that KGF might decrease the morbidity and mortality associated with GVHD (Hill and Ferrara, 2000; MacDonald and Hill, 2002; Wirth and Mertelsmann, 2002). The ability of KGF to limit GVHD has been investigated in mouse allogeneic BMT models. Recipient mice were conditioned with total body irradiation before receiving bone marrow and spleen cells as the source of GVHD-inducing T cells. An increased survival of transplant recipients was observed when KGF was administered before total body irradiation, either alone or in conjunction with chemotherapy. KGF also ameliorated GVHDrelated pathologic changes in liver, lung, and skin, but not in spleen, colon, or ileum (Panoskaltsis-Mortari et al., 1998). In a similar study, mice treated with KGF before total body irradiation exhibited increased leukemia-free survival, while GVHD in the GI tract was decreased (Krijanovski et al., 1999). Serum levels of lipopolysaccharide were decreased compared with controls, consistent with preservation of the GI epithelial barrier. There was a corresponding reduction in serum concentrations of TNF- that presumably contributed to the diminished inflammatory activity. A subsequent investigation using T cell transfer into severe combined immunodeficient (SCID) mice revealed that KGF reduced GVHD even when mice were not subjected to a mucotoxic conditioning regimen (Panoskaltsis-Mortari et al., 2000b). KGF treatment was associated with an increase in the helper T cell type 2 (Th2) cytokine IL-13 and decreases in the inflammatory cytokines TNF- and interferon-. This implied that, independent of any effect on limiting GVHD by reducing GI tract injury, KGF also has an immunomodulatory mechanism of action. Allogeneic BMT also can result in lung injury not necessarily related to GVHD. It often manifests itself as a noninfectious, idiopathic pneumonia syndrome (IPS). KGF pretreatment hastened the repair of lung damage in a mouse allogeneic BMT model by stimulating type II alveolar cell proliferation, decreasing expression of the cytotoxic molecule granzyme B and the costimulatory ligand B7 that enhances formation of cytotoxic T lymphocytes, and increasing the production of Th2 cytokines (IL-4, IL-6, and


107

IL-13) (Panoskaltsis-Mortari et al., 2000a). A KGF-dependent increase in the expression of surfactant-A also contributes to the antiinflammatory effects of KGF in the lung after allogeneic BMT (Haddad et al., 2003; Yang et al., 2000). Aside from the harm done to various organs as a consequence of GVHD, a major clinical concern after BMT is the development of satisfactory thymocyte function from the population of donor precursor cells. The success of the transplant is compromised in part by the deleterious effects of both the chemoradiotherapy-conditioning regimen and cytotoxic immunologic activity on the thymic epithelial cells (TECs). These cells have a critical role in promoting the differentiation of immature thymocytes by releasing IL-7. A surprising result in the SCID mouse experiment mentioned previously was the improvement in T cell alloengraftment that was associated with KGF treatment (Panoskaltsis-Mortari et al., 2000b). Later work established that KGF enhanced thymopoiesis by preserving the various TEC populations in the thymic cortex and medulla as well as TEC function (Min et al., 2002; Rossi et al., 2002). Moreover, the thymopoietic effects of KGF required TEC–thymocyte cross-talk mediated by IL-7 signaling (Min et al., 2002). This supported the view that improvement in thymic and peripheral T cell reconstitution after BMT was due to TEC cytoprotection by KGF (Min et al., 2002).

F. Physiological Mechanisms of Action In Section III, we described many responses of epithelial cells to KGF. Presumably, the regenerative and protective effects of KGF in whole animals are attributable to a combination of these cellular responses: proliferation, migration, differentiation, antiapoptosis, and induction of enzymes that reduce oxidative stress. KGF stimulates the expression of a variety of additional factors that might also contribute to these processes, such as syndecan-1, a heparan sulfate proteoglycan (Maatta et al., 1999), caveolin-1, a gene involved in the regulation of various signal transduction pathways (Gassmann and Werner, 2000), the mitogen-regulated protein Mrp3 (or proliferin) (Fassett and Nilsen-Hamilton, 2001), as well as a number of enzymes involved in nucleotide biosynthesis (Gassmann et al., 1999) and a putative DNA helicase that conceivably could function in DNA repair (Frank and Werner, 1996). The key effect of KGF in all the experimental injury models is preservation of the epithelial barrier. In the upper aerodigestive tract, this effect is manifested by a marked thickening of the squamous epithelium that results both from increased proliferation in the basal layer and reduced cell loss in the upper layers (Fig. 6). Desquamation is inhibited as a result of

108


Fig. 6 Semithin sections of the ventral surface of the mouse tongue. Comparison of toluidine blue-stained sections from normal, control mice and mice that were treated with KGF (5 mg/kg, subcutaneous) for 3 days showed a KGF-dependent increase in overall epithelial thickness. Greater cellularity and an increase in the number and size of keratohyalin granules were evident. (From Farrell et al., 2002, Fig. 1, p. 79.)

stimulation of cell–cell adhesion that is mediated by an increased number of desmosomes, which also are better developed (Farrell et al., 1999). The mechanical barrier is further reinforced by a KGF-dependent increase in the number and size of keratohyalin granules (Farrell et al., 1999) that are thought to strengthen the lipid permeability barrier (Squier and Kremer, 2001). In the intestines, KGF stimulates the formation of goblet cells (Housley et al., 1994) that enhance the lumenal barrier by producing mucins and ITFs (Corfield et al., 2000; Playford et al., 1996). Presumably, KGF-dependent proliferation of intestinal crypt stem cells (Khan et al., 1997; Potten et al., 2001) augments repair after damage and thereby limits the period when the barrier might be breached. The proliferative response of type II alveolar cells to KGF has a similar effect in lung, but these cells also perform other valuable tasks related to barrier function. Surfactant from the type II cells reduces surface tension and thereby keeps the airways patent, facilitating the mechanics of respiration and decreasing the likelihood of infection (as noted previously, surfactant-A also has antiinflammatory activity). KGF sustains the permeability barrier of the respiratory epithelium by promoting alveolar fluid and ion transport via increased expression of Naþ,Kþ-ATPase in type II cells (Borok et al., 1998a; Guery et al., 1997; Wang et al., 1999b) and maintaining intercellular junctions (Barazzone et al., 1999; Savla and Waters, 1998; Waters et al., 1997; Welsh et al., 2000). KGF downregulation of many interferon-induced genes in type II cells may reduce tissue damage resulting from interferon-dependent mechanisms (Dixon et al., 2000; Prince et al., 2001).


109

Preservation of the epithelial barrier has a multitude of beneficial consequences. The possibility of infection is reduced, and therefore inflammation and fibrosis become less likely. An intact oral epithelium enables the patient to continue eating, while a fully functional intestinal epithelium absorbs water and nutrients to maintain optimal nutritional status. When airways are free of edema they function well in gas exchange. Respiration is further enhanced by the production of surfactant that facilitates ventilation by allowing the airways to remain patent as the lungs expand and contract. These global effects of KGF treatment all derive from the maintenance of barrier and associated epithelial functions. They minimize the adverse sequelae of injury and ensure better health.

VII. CLINICAL TRIALS: AMELIORATION OF SEVERE ORAL MUCOSITIS Preclinical data from rodent experiments demonstrated a potent cytoprotective effect of KGF after a variety of toxic exposures to the epithelia of lung, bladder, and GI tract. To test the clinical significance of these findings, the initial focus has been to determine whether KGF can reduce the duration of severe oral mucositis (OM) in patients receiving highly mucotoxic cancer therapies. Such patients commonly develop ulceration in their oral cavities and pain with swallowing that limits food intake and impairs speaking and sleeping. Loss of the epithelial barrier predisposes the patients to infection. These clinical complications often result in delays and/ or reductions in chemotherapy or radiation that compromise the outcome of treatment regimens. Although existing palliative care is of some benefit, there are currently no effective remedies for OM and it remains an unmet clinical need (Duncan and Grant, 2003; Sonis, 1998; Squier and Kremer, 2001; Symonds, 1998). The first KGF clinical trial was performed in normal human volunteers to establish that it was safe when taken in a dose range that had a clear biological effect on the oral epithelium (Serdar et al., 1997). Prior work with rhesus macaques had defined a range likely to include biologically active doses, given the similarity of anatomy between macaques and humans (Danilenko, 1999). Volunteers were divided into groups that received KGF or placebo (3:1) administered either as a single intravenous injection (n ¼ 4), or a daily injection for three consecutive days (n ¼ 8), with the size of the dose for each group varying from 0.2 to 20 g/kg per day. Buccal biopsies were obtained from each of the trial participants for Ki67 immunostaining and counting of mitotic figures for evidence of epithelial proliferation. Three days of systemically administered KGF induced statistically

110


significant increases in proliferation with doses of 5–20 g/kg per day. KGF administration was determined to be safe and well tolerated, with no significant adverse effects. On the basis of these results, trials were initiated in cancer patients. To assess the potential efficacy of KGF in ameliorating oral OM, it was used in conjunction with cancer treatment regimens associated with a high degree of oral mucosal injury.

A. Autologous Peripheral Blood Progenitor Cell Transplantation for Hematologic Malignancies Myeloablative conditioning regimens associated with BMT or peripheral blood progenitor cell transplantation (PBPCT) typically have a substantial incidence of severe OM. A phase 1 trial was conducted in patients with Hodgkin’s disease and non-Hodgkin’s lymphoma who were receiving BEAM (BCNU, etoposide, Ara-C, and melphalan) chemotherapy before autologous PBPCT (Durrant et al., 1999). This was a randomized, placebo-controlled study in which patients received three daily doses of KGF intravenously, either before the start of chemotherapy or both before and after chemotherapy. Patients were divided into groups that received one of the following doses: 5, 20, 40, 60, or 80 g/kg per day. Side effects of KGF consisted of mild–moderate erythema with or without edema, and an asymptomatic transient elevation of serum amylase and lipase. These results were consistent with those seen in normal volunteers, and generally more evident at higher doses of KGF. The BEAM-conditioning regimen was characterized by a lower incidence of severe OM than most conditioning regimens for PBPCT. Using the World Health Organization (WHO) scale for OM (Table I), only 22% of the placebo patients experienced severe OM (grades 3 or 4) whereas 51% had ulcerative OM (grades 2–4). Although the numbers of patients in the KGF treatment groups were small and precluded definitive statistical analysis, it Table I

World Health Organization Grading Scale for Oral Mucositisa

Grade

Symptoms

0 1 2 3 4

None Soreness/erythema Erythema/ulcers/can eat solids Ulcers/extensive erythema/requires liquid diet Alimentation not possible

aUlcerative oral mucositis consists of grades 2–4, while severe oral mucositis corresponds to grades 3 and 4. (Adapted from World Health Organization, 1979.)


111

was noteworthy that only 13% of patients receiving 60 g/kg per day before BEAM had ulcerative OM. Moreover, the duration (mean SE) of ulcerative OM in this group was 0.8 0.5 days, compared with 4.6 1.8 days for the placebo group. Patients also reported a decrease in mouth and throat soreness as well as reduced use of analgesia. In view of these results, KGF was used at a dose of 60 g/kg per day in subsequent clinical trials involving PBPCT. A phase 2, double-blinded study was performed in patients treated for hematologic malignancies using a myeloablative regimen associated with more severe OM (Spielberger et al., 2001). The conditioning regimen consisted of 12 Gy of total body irradiation, etoposide (60 mg/kg) and cyclophosphamide (75 or 100 mg/kg). One hundred and twenty-nine patients were randomly assigned to groups that received placebo or three daily 60 g/kg intravenous doses of KGF before (pre) or both before and after (pre/post) conditioning. KGF treatment resulted in a highly significant decrease in the duration of severe OM, as the mean duration in the placebo group was 7.7 days but only 4.0 days in the pre/post KGF group (p < 0.001) and 5.0 days in the pre KGF group (p < 0.04). Patients reported a corresponding decrease in mouth/throat soreness and difficulty in swallowing, drinking, eating, talking, and sleeping. Moreover, those receiving KGF required less intravenous opioid analgesics and supplementation with total parenteral nutrition. KGF itself was well tolerated, with the same set of side effects previously described in the phase 1 trials. A phase 3, double-blinded study involving the same conditioning regimen has been completed (Spielberger et al., 2003). In this trial, 212 patients were randomly distributed (1:1) into a placebo group or a group receiving KGF intravenously at 60 g/kg for 3 days both before and after the conditioning regimen. As in the phase 2 trial, there was a marked decrease in the mean duration of severe OM associated with KGF use: 10.4 days in the placebo group versus 3.7 days in the KGF group (p < 0.001). Moreover, the incidence of severe OM was significantly reduced (98% in the placebo group versus 63% in the KGF group, p < 0.001), as was the incidence of the most severe form, grade 4 (62% in the placebo group versus 20% in the KGF group, p < 0.001). The decline in morbidity was reflected in a reduced use of intravenous opioid analgesics (median milligrams of morphine equivalent in the placebo group was 512 mg versus 212 mg in the KGF group) and total parenteral nutrition (40% in the placebo group versus 11% in the KGF group). Assessment by the patients of their condition again correlated well with the clinical data. These results provided strong evidence that KGF was safe and effective in reducing the incidence and duration of severe OM in patients receiving high-dose chemoradiotherapy before autologous PBPCT for hematologic malignancies.

112


B. Solid Tumors Various treatments for solid tumors, particularly those involving a combination of chemotherapy and radiation, also are characterized by a high incidence of severe OM. A phase 1 trial was performed with patients receiving 5-FU and leukovorin for advanced colorectal cancer (Meropol et al., 2003). They were divided into groups treated with various doses of KGF similar to the routine employed in the phase 1 PBPCT trial involving patients with hematologic malignancies (Durrant et al., 1999). As before, KGF was well tolerated with the same set of observed side effects. Combining the results from patients who had received KGF intravenously at >10 g/kg for 3 days before chemotherapy, the data suggested KGF was effective in reducing the incidence of ulcerative OM (67% in the placebo group versus 43% in the KGF group, p < 0.06). To better evaluate the therapeutic potential of KGF in this setting, additional patients were randomly assigned to a placebo group or to a group receiving KGF intravenously at 40 g/kg on days 1–3, followed by 5-FU (425 mg/m2 per day) and leukovorin (20 mg/m2 per day) on days 4–8 of a 28-day cycle (Clarke et al., 2001). The entire treatment routine was repeated in a second cycle. The mean duration of ulcerative OM was markedly reduced by KGF (10.2 days in the placebo group versus 3.4 days in the KGF group, p < 0.001), as was its incidence (78% in the placebo group versus 32% in the KGF group, p < 0.001). Again, the side effects were similar to those previously mentioned. Considering the possibility that colorectal tumor cells might express FGFR2b and consequently could be affected by exogenous KGF, it is worth noting that there was no significant difference in the median survival of patients in the two treatment groups. This study was not continued because the incidence of severe, as opposed to ulcerative, OM in patients receiving this treatment regimen was relatively low. Furthermore, the use of 5-FU for colorectal carcinoma has been superseded by a new generation of chemotherapy agents that cause less OM. Nevertheless, these results demonstrated the potential utility of KGF in ameliorating OM in a solid tumor setting. Another phase 1 trial of KGF involved a more mucotoxic regimen for patients with head/neck carcinoma (Brizel et al., 2001). This entailed a combination of hyperfractionated radiation (1.25 Gy twice a day for a total of 72.5 Gy over

Advances in Cancer Research, Volume 91

Advances in Cancer Research

Advances in Cancer Research Volume 51

Advances in Cancer Research Volume 43

Advances in Cancer Research Volume 42

Advances in Cancer Research Volume 41

Advances in Cancer Research Volume 78

Advances in Cancer Research, Volume 106

Advances in Cancer Research, Volume 84

Advances in Cancer Research, Volume 20

Advances in Cancer Research Volume 76

Advances in Cancer Research Volume 65

Advances in Cancer Research Volume 40

Advances in Cancer Research Volume 63

Advances in Cancer Research Volume 35

Advances in Cancer Research Volume 75

Advances in Cancer Research Volume 49

Advances in Cancer Research Volume 32

Advances in Cancer Research Volume 82

Advances in Cancer Research, Volume 57

Advances in Cancer Research, Volume 5

Advances in Cancer Research Volume 70

Advances in Cancer Research Volume 60

Advances in Cancer Research Volume 47

Advances in Cancer Research, Volume 30

Advances in Cancer Research Volume 52

Advances in Cancer Research Volume 54

Advances in Cancer Research Volume 99

Advances in Cancer Research, Volume 101

Advances in Cancer Research, Volume 109

Advances in Agronomy, Volume 91

Advances in Cancer Research, Volume 91