Differential Requirement for Focal Adhesion Kinase Signaling in Cancer Progression in the Transgenic Adenocarcinoma of Mouse Prostate Model

Abstract

Increasing evidence indicates that adhesion signaling plays an important role in the tumor microenvironment, contributing to cancer progression, invasion and metastasis. Focal adhesion kinase (FAK) is a non-receptor protein tyrosine kinase that regulates adhesion-dependent cell signaling and has been implicated in mediating steps in cancer progression and metastasis in many human cancers, including prostate. We have investigated the role of FAK in the appearance of adenocarcinoma (atypical epithelial hyperplasia of T antigen) and neuroendocrine carcinomas in the transgenic adenocarcinoma of mouse prostate (TRAMP) model employing either Cre--mediated recombination to genetically ablate FAK expression or pharmacological inhibition of FAK activity with the small molecule inhibitor, PF-562,271. We provide evidence that loss of FAK or its inhibition with PF-562,271 does not alter the progression to adenocarcinoma. However, continued FAK expression (and activity) is essential for the androgen-independent formation of neuroendocrine carcinoma. These data indicate that integrin signaling through FAK is an important component of cancer progression in the TRAMP model and suggest that treatment modalities targeting FAK may be an appropriate strategy for patients with castrate-resistant cancer.

Introduction

The tumor microenvironment plays a critical role in cancer invasion and metastasis (1–3). In particular, the loss of adhesion-dependent cellular regulation is a hallmark of cancer, and factors that regulate cell-ECM interactions and signaling have been implicated in tumorigenesis and metastatic progression. Focal adhesion kinase (FAK) is a non-receptor protein tyrosine kinase that regulates adhesion-dependent cell signaling (4). There is abundant evidence that FAK plays a role in cancer progression and metastasis including that for prostate cancer (4). First, FAK expression is increased in prostate cancer cell lines (5), and increased expression correlates with enhanced motility and tumorigenicity (6). In human prostate cancer, FAK is present in high grade PIN, primary adenocarcinomas and metastatic lesions (7). The inhibition of FAK activity by dominant interfering mutants or small interfering RNAs decreases prostate cancer cell migration (8) and inhibits growth of prostate tumor cells in immunocompromised mice. Finally, small molecule inhibitors, developed specifically for FAK family members (FAK and PYK2), inhibit prostate cancer cell migration, anchorage independent growth, and growth of prostate cancer xenografts (9, 10).

Transgenic adenocarcinoma of mouse prostate (TRAMP) is a well-studied mouse model of prostate cancer in which SV40 T-antigen is expressed in prostate secretory epithelial cells under the control of the androgen-responsive minimal rat probasin promoter (11, 12). In this model, mice develop progressive, multifocal, and heterogenous disease, characterized by atypical epithelial hyperplasia of Tag (referred to here after as adenocarcinoma) and a pronounced shift to neuroendocrine carcinomas in late stage disease (13). Castration of animals at 12–15 weeks of age results in the appearance of androgen-independent neuroendocrine carcinomas (14). Transplantation studies indicate that the neuroendocrine carcinomas develop from bipotential progenitor cells during an early stage of SV40 Tag-driven tumorigenesis (13).

We have investigated the role of FAK in the appearance of adenocarcinoma and neuroendocrine carcinomas in the TRAMP model employing either Cre--mediated recombination to genetically ablate FAK expression or pharmacological inhibition of FAK activity with the small molecule inhibitor, PF-562,271 (7, 8). We provide evidence that loss of FAK or its inhibition with PF-562,271 does not alter the progression to adenocarcinoma. However, we observe that continued FAK expression (and activity) is essential for the androgen-independent formation of neuroendocrine carcinoma. These data indicate that integrin signaling through FAK is an important component of cancer progression in the TRAMP model and suggest that treatment modalities targeting FAK may be an appropriate strategy for patients with castrate-resistant cancer.

Materials and Methods

Mice

Mice expressing SV40 T antigen from the probasin promoter (TRAMP), (12) (Jackson Laboratories), Cre recombinase (ARR2Probasin-Cre transgenic line, PB-Cre4), (NCI), ROSA26 lox-stop-lox LacZ (15) (Jackson Laboratories) or floxed FAK (16) were bred onto a C57Bl/6 background. TRAMP and Pb-Cre alleles were maintained on the same individual as heterozygotes; LacZ and floxed FAK alleles were maintained as homozygotes. TRAMP/Pb-Cre⁺ mice were bred with LacZ/FAK^fl/fl to generate TRAMP^+/−, Pb-Cre^+/−, FAK^fl/fl, LacZ^+/+ mice. Pb-Cre was maintained only in males; all other transgenes were carried on either sex.

Genotyping

Mice were genotyped using tail DNA isolated from pups 10 to 14 days old and the Purelink genomic DNA kit (Invitrogen). Transgene identification was performed by PCR using the following primer sets:

• TRAMP: - 5’- CAG AGC AGA ATT GTG GAG TGG-3’, 5’-GGA CAA ACC ACA ACT AGA ATG CAG TG −3’;

• Pb-Cre : 5’-TTC CCG CAG AAC CTG AAG ATG-3’, 5’-CGC CGC ATA ACC AGT GAA AC-3’;

• FAK^fl/fl: 5’-GAG AAT CCA GCT TTG GCT GTT G-3’, 5’- GAA TGC TAC AGG AAC CAA ATA AC-3’;

• Rosa26-LacZ: #1 5’-AAA GTC GCT CTG AGT TGT TAT-3’, rosa26 #2 5’-GCG AAG AGT TTG TCC TCA ACC-3’, rosa26 #3 5’-GGA GCG GGA GAA ATG GAT ATG-3’.

TRAMP sequences were amplified using 1 cycle 95°C 5 min; 10 cycles 94°C 10 sec, 53°C 30 sec, 68°C 3 min; 20 cycles 94°C 10 sec, 53°C 30 sec, 68°C 3 min 20 sec; and 1 cycle 68°C 7 min to generate a 474 bp product. The remaining transgene sequences were amplified using 1 cycle 95°C 5 min; 10 cycles 94°C 10 sec, 57°C 30 sec, 68°C 3 min; 20 cycles 94°C 10 sec, 57°C 30 sec, 68°C 3 min 20 sec; and 1 cycle 68°C 7 min to produce 166 bp Pb-Cre, 400 bp FAK ^fl/fl (290 bp wildtype) and 300 bp LacZ (650 bp wildtype) fragments.

To detect sequence alterations in the FAK gene as a result of Cre-induction, DNA from prostate or tumor tissue was extracted and DNA was amplified using 1 cycle 95°C 5 min; 35 cycles 94°C 30 sec, 62°C 30 sec, 72°C 2 min; one cycle 72°C 6 min; using the following primers: 5’-GAA TGC TAC AGG AAC CAA ATA AC-3’ and 5’-GAC CTT CAA CTT CTC ATT TCT CC-3’. The floxed FAK allele is 1800 bp, wild type allele 1690 bp and recombined FAK 300 bp. All PCR products were resolved on 2% agarose gels containing ethidium bromide.

Castration and drug treatment

Radical orchiectomy was performed on mice 15–16 weeks of age anesthetized with IP injection of 0.018 ml/g 2.5% tribromoethanol. A transverse incision was made through the shaved abdominal skin and wall just superior to the preputial glands. Testicles were everted one at a time and removed by cauterization. The abdominal wall was sutured and the skin glued, stapled or sutured. Analgesia (bupivicane) was administered at the incision. Following recovery from surgery (4–5 days), mice were treated for 4 weeks PO, BID with 33 mg/kg PF-562,271 suspended in paraffin oil vehicle or vehicle alone.

Immunohistochemistry and β-galactose staining

Following euthanasia, the total utogenital tract was removed, weighed, the prostate and/or tumor isolated, weighed and prepared for IHC or β-galactose (β-gal) staining. Additional organs including lymph node, kidney, liver and lung were examined grossly for metastasis, removed and prepared for histology. Prostates and tumors were washed twice in PBS, fixed for 30 minutes at room temperature in 0.1 M sodium phosphate (pH 7.3), 20mM Tris (pH 7.3), 5 mM EGTA, 2 mM MgCl₂, 0.25% glutaraldehyde, and 1% formaldehyde. Fixed tissues were washed twice in PBS and stained for β-gal activity in 0.1M sodium phosphate (pH 7.3), 20 mM Tris (pH 7.3), 2 mM MgCl₂, 5 mM potassium ferrocyanate, 5 mM potassium ferricyanate, 0.1% deoxycholate, 0.2% NP40, and 1 mg/ml X-gal (in DMF) 16 hours at room temperature in the dark. Tissues were de-stained following several PBS washes and either cleared for whole mount by incubating 16 hours in 2:1 benzylbenzoate:benzyl alcohol followed by methanol dehydration for 48 hours or fixed for histology. Selected tissues were fixed in zinc formalin, embedded in paraffin, sectioned and stained with hematoxylin and eosin (H&E). IHC was performed on additional sections for FAK (1:200, Cat # sc-577, Santa Cruz), T antigen (1:200, Cat # 554149, BD Pharminogen), Synaptophysin (1:50, Cat # 611880, BD Pharminogen), and E-cadherin (1:25, Cat # M3612, DAKO). Prior to staining, de-paraffinized, hydrated sections were placed in citrate buffer (10mM sodium citrate, pH 6.0) for microwave antigen retrieval followed by incubation in 3% H₂O₂ for 5 min at room temperature. FAK IHC was performed using ABC Vectastain (Vector Laboratories), and slides were stained with DAB and then counterstained with hematoxylin. For the remaining antibodies, tissues were blocked using the M.O.M. kit (Vector Laboratories), stained with DAB, and counterstained with hematoxylin.

Western Analysis

Prostate or tumor tissue was homogenized and lysed in supplemented RIPA buffer (50 mM Hepes, 0.15 M NaCl, 2 mM EDTA, 0.1% Nonidet P-40, and 0.05% sodium deoxycholate, pH 7.2) containing EDTA-free protease inhibitor cocktail (Roche), 1 mM Na₃VO₄, 40 mM NaF, and 10 mM Na₄P₂O₇. Proteins from 25–50 µg whole cell lysates were resolved on 8% SDS-PAGE, transferred to nitrocellulose and blotted for FAK (clone 4.47, Millipore), β-actin (Millipore), or β-galactosidase (Invitrogen).

Results

Disruption of FAK expression in the prostate of normal and TRAMP mice

To assess the role of FAK in prostate cancer progression in TRAMP mice, normal and TRAMP mice were interbred with conditional FAK knockout mice (FAK^fl/fl) and further interbred with the ARR2Probasin-Cre transgenic line, PB-Cre4, in which the Cre recombinase is under the control of a modified rat prostate-specific probasin promoter (Pb-Cre). To have a readily detectable marker for Cre activity, we introduced the GTRosa26 Cre-inducible β-galactosidase reporter (15) into the normal and TRAMP genetic backgrounds. The resulting mice carried the TRAMP allele, FAK^fl/fl, Rosa26-LacZ and either expressed Pb-Cre (Pb-Cre⁺) or lacked Pb-Cre expression (Pb-Cre⁻). All mice were maintained on a C57Bl/6 background.

PCR analysis revealed that recombination of FAK^fl/fl was readily detectable in the prostate tissue of normal and TRAMP mice expressing Pb-Cre whereas no recombination was observed in other tissues (e.g., kidney) (Fig.1A). Normal mice expressing Pb-Cre and FAK^fl/fl were fertile and phenotypicially normal up to 40 weeks of age; no detectable alteration in prostate architecture was observed and robust β-galactosidase activity was present in 60–80% of the individual glands (Fig.1B, C, D) (17).

Figure 1. Prostate development in Pb-Cre⁺, FAK^fl/fl mice.

(A) Genomic DNA isolated from prostates (P) and kidneys (K) was subjected to PCR amplification of the FAK locus to identify floxed (1800 bp) and recombined (300 bp) FAK alleles. (B) Urogenital track (top panel) and individual prostate glands (bottom panels) were isolated and stained for β-galactosidase activity (blue). Arrows indicate prostate glands. (C) Individual prostate glands from Pb-Cre⁺ and Pb-Cre⁻ mice were dissected to reveal the ductal tree. (D) Sections of prostate from Pb-Cre⁺, FAK^fl/fl mice stained for β-galactosidase activity. Secretory epithelial cells from 60–70% of the glands stained positively (blue). Magnification: 400x

Appearance of lesions in FAK^fl/fl mice

As expected, TRAMP/FAK^fl/fl mice that did not express Pb-Cre developed adenocarcinoma as early as 7 weeks after birth. These lesions, as described by others (12, 13), were characterized by multifocal atypical hyperplastic lesions within the lining epithelial cells. As illustrated in Fig. 2 and tabulated in Table 1, by 20 weeks of age, 49% of the TRAMP animals exhibited adenocarcinoma (AD). Between 11 and 20 weeks about 16% of tumor-bearing animals exhibited neuroendocrine (NE) carcinomas or a mix of NE and adenocarcinoma (Table 1, (12, 13)). By 21–30 weeks, all animals except one exhibited either AD alone, NE tumors alone, or both, approximately 21%, 7% and 5% respectively (Table 1). NE tumors were typically large and grew rapidly; they were locally invasive and exhibited metastasis to lymph nodes, kidney and lung (data not shown). Immunohistochemical analysis revealed that virtually all adenocarcinomas were Tag positive, E-cadherin positive and synaptophysin negative. NE tumors characteristically were Tag positive, E-cadherin negative, and synaptophysin positive at this stage of development (Fig.3). FAK staining was weak in normal prostate epithelial cells but was clearly enhanced in both AD and NE tumors (Fig.3). TRAMP/FAK^fl/fl mice that expressed Pb-Cre showed no significant differences in time to appearance of AD or NE tumors compared to Pb-Cre⁻ mice (Table 1). The relative number of animals with AD was not statistically different between Pb-Cre⁺ and Pb-Cre⁻ mice. In addition there was no difference in the percent of mice exhibiting NE or mixed AD/NE tumors in the Pb-Cre⁺ cohort. No statistically significant differences in urogenital weights were observed at early, intermediate, or late times after transgene expression (Fig. 2).

Figure 2. Tumor distribution in FAK^fl/fl mice.

Schematic representation of the onset and duration of well differentiated adenocarcinoma (AD) and neuroendocrine (NE) tumors in the TRAMP model (top panel). (A) Percentage of mice with AD (dark gray), NE (light gray) or no tumor (black) in Pb-Cre⁺, FAK^fl/fl (Pb+) or Pb-Cre⁻, FAK^fl/fl (Pb-) mice. (B) Weight of the urogenital track including seminal vesicles, prostate, and any tumor from TRAMP⁻ (Tr-), TRAMP⁺, Pb-Cre⁺, FAK^fl/fl (Tr+, Pb+) or TRAMP⁺, Pb-Cre⁻, FAK^fl/fl (Tr+, Pb-) mice. Statistical significance was determined using the Kruskal-Wallis test followed by the Dunn’s Multiple Comparison post-test. The numbers above each graph represent the number of mice analyzed in each group.

Table 1.

Tumor incidence by age and genotype.

Age (weeks)		Normal		AD*		AD/NE*		NE*		Total
Pb-Cre status		−	+	−	+	−	+	−	+	−	+
<10	#	1	2	7	8					8	10
	(%)	(2)	(4)	(13)	(15)					(14)	(19)
11–20	#		1	20	14	6	4	3		29	19
	(%)		(2)	(36)	(26)	(11)	(8)	(5)		(52)	(36)
21−30	#		1	12	12	4	2	3	9	19	24
	(%)		(2)	(21)	(23)	(7)	(4)	(5)	(17)	(34)	(45)
Total		1	4	39	34	10	6	6	9	56	53

^*AD, denotes animals in which only adenocarcinoma was detected, NE, denotes animals with only neuroendocrine tumors, AD/NE denotes animals have both adenocarcinoma and neuroendocrine tumors.

Figure 3. Immunohistochemical analysis of mouse prostates.

Representative sections of TRAMP⁻ prostates (Normal) or TRAMP⁺ AD and NE tumors were stained with H&E to reveal tissue architecture. Sections were stained for T antigen, FAK, E-cadherin and synaptophysin as indicated.

FAK recombination and expression in TRAMP/ FAK^fl/fl mice

PCR genotyping of lesions revealed significant recombination of the FAK^fl/fl allele in prostates with adenocarcinoma whereas no recombination was observed in normal kidney tissue (Fig. 4A, left panel). In contrast, NE tumors showed little evidence of recombination of the FAK^fl/fl allele (Fig. 4A, right panel). These data were consistent with observations based on Western blotting of multiple AD and NE lesions (Fig. 4B), in which there was readily detectable expression of FAK irrespective of the age of the animal or the type of lesion. The expression of β-galactosidase was detectable in all normal prostate samples as well as in prostates with adenocarcinomas (Fig. 4B, left panel). Loss of β-galactosidase expression was typically observed in late stage AD, e.g., animals greater than 20 weeks of age. Interestingly NE tumors were always negative for β-galactosidase expression, irrespective of age. These data support the observations from PCR genotyping and indicate that NE tumors develop from cells that fail to undergo Cre-mediated recombination at the floxed FAK or the floxed Rosa26 locus. In accordance with these observations, we also failed to observe β-galactosidase staining in tissue sections from NE carcinomas (data now shown). In contrast, typically 60–70% of glands exhibiting adenocarcinoma were positive for β-galactosidase (Fig. 3). These data indicate that deletion of FAK and loss of FAK expression does not prevent the outgrowth of AD. However, the fact that virtually all NE tumors had FAK expression and were derived from cells that appeared to fail to undergo Cre-mediated recombination indicates that sustained FAK expression may be required for the efficient outgrowth of NE carcinomas.

Figure 4. Incidence of Cre-mediated recombination in mouse prostate tumors.

(A) Genomic DNA isolated from prostate tumors (P; AD, left panel, or NE, right panel) or kidney (K) was subjected to PCR amplification of the FAK locus to identify floxed (1800 bp) and recombined FAK (300 bp). (B) Western analysis was performed on protein isolated from prostates of Pb-Cre⁺ mice. Lysates were blotted for β-galactosidase (β-gal), total FAK or actin (loading control). TRAMP and FAK^fl/fl status is indicated. N- normal tissue, AD – adenocarcinoma, NE – neuroendocrine.

Appearance of NE tumors in TRAMP mice treated with the FAK inhibitor PF-562,271

Castration of TRAMP mice at 12–14 weeks of age leads to regression of adenocarcinomas and the efficient outgrowth of NE tumors (14, 18). To test whether FAK was required for the outgrowth of NE carcinomas, mice were castrated at 15–16 weeks of age, a time in which AD and/or NE lesions were observed in greater that 95% of the TRAMP mice. Following recovery from surgery, castrated mice were assigned to either a treatment group that received the FAK inhibitor, PF-562,271 at a dose of 33 mg/kg, BID, PO for 4 weeks or a group treated in a parallel fashion with drug carrier only. Control groups included non-TRAMP mice treated with PF-562,271 as indicated above. This treatment regimen has been demonstrated to efficiently inhibit the growth of several different human tumor cells lines implanted subcutaneously in immunocompromised mice as well as reduce the level of FAK phosphorylation on tyrosine397 (9). Control non-TRAMP mice, as expected, did not get lesions and were healthy and viable following 4 weeks of PF-562,271 treatment (data not shown). Castrated animals that did not receive PF-562,271 developed large, highly invasive NE carcinomas; 6 of 8 mice developing NE tumors exhibited lymph node metastases, 3 of 6 mice with lymph node metastasis also had metastases within the kidney or lung with 1 of the 3 having metastases in both the kidney and the lung. Castrated mice treated with PF-562,271 developed fewer NE carcinomas and those that did arise were considerably smaller (Figure 5B, C). Lesions developing in the castrated mice that were not treated with PF-562,271 exhibited the typical NE phenotype, as they were T-antigen positive and synaptophysin positive (data not shown). Non-neoplastic glands present in castrated, PF-562,271 treated animals showed evidence of atrophy with dilated glands that were lined by flattened epithelial cells with minimal secretory cytoplasm (data not shown) The above observations confirm a role for FAK in the outgrowth of NE carcinomas in the setting of androgen independence.

Figure 5. Pharmacologic inhibition of FAK inhibits growth of castrate resistant tumors.

(A) Schematic representation of prostate cancer onset in the TRAMP model and the treatment strategy. Mice were castrated at 16 weeks of age and treated with PF-562,271 for 4 weeks starting 5 days post-castration. (B) Percentage of mice with prostate tumors after castration and treatment with PF-562,271 (PF-271, 10 mice) or vehicle (9 mice). TRAMP⁻ (Normal) mice were castrated and treated with PF-562,271 (10 mice) or vehicle. (C) The weight of prostates, including tumor if present, was determined from TRAMP⁻ (Normal) mice, those castrated and treated with PF-271, or mice castrated and treated with vehicle. Statistical significance was determined using the Kruskal-Wallis test followed by the Dunn’s Multiple Comparison post-test. *p<0.05 compared to Normal; **p<0.05 compared to vehicle.

Discussion

Our studies point to several important roles for FAK in the progression of cancer in the TRAMP model. First, lost of FAK expression by targeted deletion in prostate epithelial cells did not alter the time to appearance or the frequency of well differentiated adenocarcinoma indicating that FAK expression is not necessary for the development of these of early intraepithelial lesions. In contrast, all neuroendocrine tumors expressed FAK and appeared to have escaped Cre-mediated recombination at both the FAK and the Rosa26-LacZ loci. These observations suggest a role for FAK function in progression to the more aggressive neuroendocrine tumors, and a less important role in development of early in-situ lesions. It is possible that the early progenitor cells that give rise to neuroendocrine tumors become transformed by expression of T-antigen but fail to express Cre, thus accounting for the observed expression of FAK and the lack of expression of β-galactosidase. To test this possibility we used the pharmacological inhibitor of FAK, PF-562,271 and found that this inhibitor significantly limited the outgrowth of neuroendocrine tumors following castration of mice at 15 weeks of age. These data suggest that integrin signaling through FAK is an important component of cancer progression in the TRAMP model and that treatment modalities targeting FAK should be considered as a therapeutic strategy for patients with castrate-resistant cancer.

The prostate epithelium in the mouse is comprised of luminal, basal and neuroendocrine cells (12). Recent studies suggest that in the TRAMP model, neuroendocrine carcinomas arise independently from atypical hyperplasias/adenocarcinomas or other epithelial lesions in a population of bipotential progenitor cells that express both markers of epithelial (E-cadherin) and neuroendocrine (synaptophysin) lineages (13). In our studies we observed that neuroendocrine tumors expressed T antigen and synaptophysin, while deficient in the expression of E-cadherin. Neuroendocrine tumors from Pb-Cre⁺ mice expressed FAK and lacked β-galactosidase expression, indicating that Cre-mediated recombination had failed to take place in these tumor cells. These observations would suggest that oncogenic transformation and/or progression of tumors with the neuroendocrine phenotype occurs only in mice that have escaped FAK deletion and hence have FAK activity. In contrast, the oncogenic transformation and outgrowth of epithelial cells that give rise to the atypical hyperplastic phenotype characteristic of early adenocarcinomas do not appear to be influenced by the loss of FAK expression. This is consistent with our observations that loss of FAK expression in the normal prostate does not appear to influence normal glandular structure or function. Thus, maintenance of the epithelial character of early Tag induced hyperplasia appears to bypass the requirement for FAK signaling. It is possible that in these glandular structures, PYK2 may compensate for the loss of FAK expression, although we observe no clear increase in PYK2 expression in prostates of FAK deficient mice (data not shown).

Androgen appears to play an important role in regulating the progression of cancer in the TRAMP models, mimicking to some degree its role in human cancers. Maintenance of Tag induced adenocarcinoma requires androgen, since castration results in the loss of these lesions (14). The relative synchronous outgrowth of neuroendocrine tumors following castration indicates their androgen-independent progression. At this time it is not clear whether FAK is required for the growth of neuroendocrine tumors or is important for some aspect of transdifferentiation of epithelial cell progenitors to the neuroendocrine tumor phenotype. The treatment of TRAMP mice with PF-562,271 from 10 to 14 weeks did not block the appearance of adenocarcinomas, consistent with the genetic ablation studies.

Genetic approaches using Cre-mediated recombination to delete FAK have been used to assess the role of FAK in several cancers (19–22). Targeted disruption of FAK in the mammary epithelium impairs mammary tumor development in the MMTV-Polyoma virus middle T-antigen mouse tumor model. In these studies, FAK expression was required for the transition of premalignant hyperplasias to carcinomas and their subsequent metastases. Mice that lacked functional FAK expression did develop mammary tumors, yet there was a decrease in the number of hyperplasic lesions. Interestingly, the late stage tumors and lung metastases arising in homozygously floxed FAK mice did not exhibit Cre-mediated recombination, consistent with a role for FAK function in mammary tumor progression in this model. Similar observations have been reported by Provenazano et al (20). In these studies, FAK was not required for tumor initiation but was required for tumor progression. Interestingly, late-stage tumors that lacked FAK did not show evidence of invasion suggesting that FAK is required for progression to the infiltrative phenotype. A role for FAK has been confirmed in breast cancer models driven with other oncogenes. Pylayeva et al (21) showed that upon silencing of Fak in mouse mammary tumor cells transformed by activated Ras, cells became senescent and lost their invasive ability. Loss of FAK expression in Neu-transformed cells induced growth arrest and apoptosis, albeit more efficiently if integrin β4–dependent signaling was also inactivated. These observations in mouse breast cancer models parallel the observations in previous reports showing that FAK deletion in papillomas blocks conversion to squamous cell carcinomas (22) and are consistent with finding described above.

The preclinical data indicating a role for FAK in breast and prostate cancer, as well as preclinical data with PF-562,271 suggest that FAK inhibitors may prove efficacious in the clinical setting. Preliminary reports of data from a Phase I clinical trial provide evidence that PF-562,271 is well tolerated in cancer patients upon extended oral administration with some patients showing durable stable disease [L. L. Siu, H. A. Burris, L. Mileshkin, D. R. Camidge, D. Rischin, E. X. Chen, S. Jones, D. Yin, H. Fingert, Journal of Clinical Oncology, 2007 ASCO Annual Meeting Proceedings Part I. Vol 25, No. 18S (June 20 Supplement), 2007: 3527]. Future trials in breast and prostate patients will provide new information as to the effectiveness of FAK inhibitors for treating late stage cancers.

Abbreviations

FAK

focal adhesion kinase

TRAMP

transgenic adenocarcinoma of mouse prostate

SV40

simian virus 40

Tag

SV40 T antigen

Pb

probasin

β-gal

β-galactosidase

IHC

immunohistochemistry

PBS

phosphate buffered saline

BID

twice daily

PO

by mouth (oral gavage)

NE

neuroendocrine

AD

adenocarcinoma