Increased Expression of Fibroblast Growth Factor 6 in Human Prostatic Intraepithelial Neoplasia and Prostate Cancer1

Home , FGF6

[CANCER RESEARCH 60, 4245–4250, August 1, 2000] Increased Expression of Fibroblast Growth Factor 6 in Human Prostatic Intraepithelial Neoplasia and Prostate Cancer1

Frederic Ropiquet, Dipak Giri, Bernard Kwabi-Addo, Alka Mansukhani, and Michael Ittmann2 Department of Pathology, Baylor College of Medicine and Department of Veterans Affairs Medical Center, Houston, Texas 77030 [F. R., D. G., B. K- A., M. I.], and Department of Microbiology, New York University School of Medicine, New York, New York 10016 [A. M.]

ABSTRACT have also been shown to increase the motility and invasiveness of a variety of cell types including prostatic epithelial cells (3, 5–7). Fibroblast growth factors (FGFs) are known to play an important role Finally, it has been shown that FGFs can inhibit apoptosis in the in the growth of normal prostatic epithelial cells. In addition to their appropriate context (8). Thus, FGFs have a broad range of biological effects on proliferation, FGFs can promote cell motility, increase tumor angiogenesis, and inhibit apoptosis, all of which play an important role in activities that can play an important role in tumorigenesis. tumor progression. To determine whether FGFs are overexpressed in We have shown previously that FGF2, FGF7, and FGF9 are present human prostate cancers, we analyzed 26 prostate cancer RNAs by reverse in high concentrations in normal human prostate (9–11). They are transcription-PCR for expression of FGF3, FGF4, and FGF6, which produced by prostatic stromal cells and can act as paracrine growth cannot be detected in normal prostate tissue by this technique. Fourteen factors for prostatic epithelial cells and, except for FGF7, as autocrine of 26 prostate cancers expressed FGF6 mRNA. No expression of FGF3 or growth factors for stromal cells. Thus, FGFs are important growth FGF4 was detected. An ELISA of tissue extracts of normal prostate, factors for maintenance of the normal prostate in vivo. high-grade prostatic intraepithelial neoplasia (PIN), and prostate cancer Evidence for the importance of FGFs in prostate cancer comes from for FGF6 showed that this growth factor was undetectable in normal studies of the Dunning rat model system. Yan et al. (12) have shown prostate but was present at elevated levels in 4 of 9 PIN lesions and in 15 that as these transplantable tumors progress from a mixed stromal- of 24 prostate cancers. Immunohistochemical analysis with anti-FGF6 antibody revealed weak staining of prostatic basal cells in normal prostate epithelial phenotype to a stromal independent phenotype, expression that was markedly elevated in PIN. In the prostate cancers, the majority of FGFs not originally present in the tumors (such as FGF3 and FGF5) of cases revealed expression of FGF6 by the prostate cancer cells them- occurs, and there are changes in the isoforms of FGFRs expressed, selves. In two cases, expression was present in prostatic stromal cells. consistent with autocrine stimulation of growth. We therefore sought Exogenous FGF6 was able to stimulate proliferation of primary prostatic to determine whether a similar autocrine production of FGFs occurs in epithelial and stromal cells, immortalized prostatic epithelial cells, and human prostate cancers. prostate cancer cell lines in tissue culture. FGF receptor 4, which is the We have found that approximately half of human prostate cancers most potent FGF receptor for FGF6, is expressed in the human prostate contain increased amounts of FGF6, and, in the majority of such in vivo and in all of the cultured cell lines. Thus, FGF6 is increased in PIN cases, FGF6 is expressed in the cancer cells. In PIN, the precursor and prostate cancer and can promote the proliferation of the transformed lesion for prostate cancer, the prostatic basal cells express a markedly prostatic epithelial cells via paracrine and autocrine mechanisms. increased level of FGF6. The biological importance of this overexpression is shown by our finding that FGF6 is a potent growth factor INTRODUCTION for normal and neoplastic prostatic epithelial cells in culture. Thus, The FGF3 gene family consists of at least 19 different genes FGF6 may play an important role as both a paracrine and an autocrine encoding related polypeptide mitogens. These growth factors interact growth factor in the initiation and progression of human prostate with a family of four distinct, high-affinity tyrosine kinase receptors cancer. designated FGFRs 1–4 (for review, see Ref. 1). These receptors consist of an extracellular portion containing three immunoglobulin- MATERIALS AND METHODS like domains and an intracellular tyrosine kinase domain and have Tissue Acquisition and Analysis. Prostate cancer tissues and samples of variable affinities for the different FGFs. In addition, FGFRs 1–3 all the uninvolved peripheral and transitional zone of the prostate were taken from undergo an alternative splicing event in which two alternative exons radical prostatectomies performed for treatment of clinical stage T1cN0 or (IIIb and IIIc) can be used to encode the COOH-terminal portion of T2N0 prostate cancer. Tissues were received fresh, and portions were snap- the third immunoglobulin-like loop, which results in receptor isoforms frozen in liquid nitrogen or used to establish primary cell cultures (see below). with dramatically altered binding specificity. Considerable progress Additional cancer tissues were obtained from three transurethral resections of has been made in defining the relative affinities of FGFs 1–9 for the prostate performed to treat obstruction in advanced prostate cancer and from a various receptors and their isoforms (2), although much less is known pelvic lymph node metastasis. The frozen tissues were then analyzed by frozen about FGFs 10–19. section to confirm the presence or absence of carcinoma or high-grade PIN, The FGFs are mitogenic for a many cell types, both epithelial and and, if present, the percentage of the tissue involved in these processes. The carcinoma tissues contained 30–80% cancer, and tissues with PIN contained mesenchymal. In addition, FGFs have a variety of other biological 30–50% PIN. To obtain nine tissues with such extensive high-grade PIN, it activities. Some FGFs, like FGF2, have potent angiogenic activity and was necessary to select tissues from over 800 frozen sections performed have been implicated as promoters of tumor angiogenesis (3, 4). FGFs previously on prostate tissues harvested from radical prostatectomies and, in some cases, to microdissect the tissue to enrich for PIN. All normal peripheral Received 11/5/99; accepted 6/2/00. and transition zone tissues were free of cancer and PIN. Additional frozen The costs of publication of this article were defrayed in part by the payment of page sections were prepared for immunohistochemistry in some cases, and the charges. This article must therefore be hereby marked advertisement in accordance with remaining tissue was used to prepare cell extracts or RNA. 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by Department of Veterans Affairs Merit Review funding. RNA Extraction, RT-PCR, and Northern Blotting. RNAs were ex- 2 To whom requests for reprints should be addressed, at Research Service, Houston tracted from tissues and cultured cells using the guanidine isothiocyanate Veterans Affairs Medical Center, 2002 Holcombe Boulevard, Houston, TX 77030. Phone: procedure, and reverse transcription was performed as described previously (713) 791-1414, ext. 4008; Fax: (713) 794-7938; E-mail: [email protected]. using 2 ␮g of total RNA (9). The cDNA product, corresponding to 0.1 ␮gof 3 The abbreviations used are: FGF, fibroblast growth factor; FGFR, FGF receptor; RT-PCR, reverse transcription-PCR; PIN, prostatic intraepithelial neoplasia; EGF, epi- total RNA, was used for PCR amplification as described previously. Primers dermal growth factor. for FGF6 were CACGAGGAGAACCCCTACA (forward primer, exon 1) and 4245

TCCCTTGGTACAAGTCTGA (reverse primer, exon 3). Positive control for added at a concentration of 400 ng/ml for2hatroom temperature. After FGF6 was skeletal muscle mRNA from Clontech Laboratories, Inc. (Palo Alto, washing as described above, detection was carried out by adding 100 ␮l/well CA). FGFR-4 primers were CTTGACCTCCAGCAACGATGA (forward of a 1:4000 dilution of streptavidin/horseradish peroxidase (Zymed, San Fran- primer) and GCACATCTAGCAGGTAGTTATAGCGG (reverse primer). For cisco, CA) and incubating for 20 min at room temperature. Wells were washed each set of PCR reactions, a control in which water was added rather than and incubated with a substrate consisting of a 1:1 solution of H2O2 and cDNA was included, and this reaction was uniformly negative. FGF3 and tetramethylbenzidine (Sigma) at a concentration of 0.1 mg/ml. Stop solution

FGF4 primers, Southern blotting of PCR products, Northern blotting, and (H2SO4) was added within 30 min, and absorbance at 450 nm was determined specific probes for FGF3, FGF4, and FGF6 were as described previously (9). using an ELISA plate reader. The sensitivity of this ELISA was found to be The FGFR-4 probe was a 600-bp BamHI fragment from the 3Ј end of the less than 1 pg/ml. No cross-reactivity with FGF2, FGF7, and FGF9 was human FGFR-4 plasmid SV2SE (13). This probe consists primarily of a 3Ј detected when these growth factors were added at 10 ng/well. untranslated sequence to eliminate cross-hybridization with other FGFR tran- scripts. RESULTS Cell Proliferation Assays. Primary epithelial cell cultures were established using prostatic tissue samples from areas in the peripheral zone of radical RT-PCR Analysis of FGF6 Expression in Prostate Cancers. prostatectomy specimens that were free from carcinoma. Primary epithelial Our initial goal was to screen prostate cancer RNAs for expression of ϫ 4 cells were plated on collagen-coated 35-mm dishes at 5 10 cells/dish in FGF RNAs that were not expressed by normal prostate. To maximize complete epithelial growth medium that includes bovine pituitary extract (a the sensitivity and specificity of this screening, we analyzed the source of FGFs), EGF, insulin, dexamethasone, cholera toxin, and BSA, as prostate cancer RNAs by RT-PCR using primers for FGFs 1–9 fol- described previously (9). We have previously shown such cultures to be of prostatic epithelial origin by immunohistochemistry (9). The next day, medium lowed by Southern blotting of the PCR products and hybridization was changed to epithelial growth medium without bovine pituitary extract or with a probe corresponding to the specific product of RT-PCR reac- EGF. Cells were kept in this incomplete medium as controls or were supple- tion. Using this extremely sensitive technique, FGFs 1, 2, 5, 7, 8, and mented with 10 ng/ml recombinant FGF6 (R&D Systems, Minneapolis, MN). 9 gave detectable bands in normal tissues (data not shown) and thus Cells were then trypsinized and counted using a Coulter counter at 2-day were not useful for this qualitative study to determine whether new intervals. FGFs were expressed in the prostate cancer tissues. However, no Prostate cancer cell lines (DU145, PC3, and LNCaP) were maintained in expression of FGFs 3, 4, or 6 was detected by this technique in RNA ϫ 4 RPMI 1640 with 10% FCS. Cells were plated at 5 10 cells/35-mm dish, and from the normal peripheral zone. To determine whether these FGFs the next day, cells were refed with 0.2% FCS with or without 10 ng/ml FGF6. are expressed in human prostate cancers, we analyzed a total of 26 Cells were counted at 2 or 4 days after growth factor addition. PNT1A, an clinically localized prostate cancer RNAs from tissues containing immortalized normal prostatic epithelial cell line (14), was maintained in RPMI 1640 containing 5% FCS. Cell were plated at 2.5 ϫ 104 cells/35-mm 30–80% carcinoma on frozen section by RT-PCR followed by South- dish and refed the next day with RPMI 1640 supplemented with 1% ITS ern blotting of the PCR products. Expression of FGF6 mRNA was (insulin, transferrin, and selenium; Sigma, St. Louis, MO) with or without 10 detectable in variable amounts in a subset of prostate cancers (Fig. 1), ng/ml FGF6. Cells were then counted on days 3 and 5. and no FGF6 was detected in the control normal prostate tissue. No Preparation of Cell Extracts. Prostatic tissue samples were weighed, expression of FGF3 or FGF4 was seen (data not shown). Overall, 14 pulverized in liquid nitrogen, and then homogenized by three strokes (each for of 26 prostate cancers had detectable FGF6 mRNA by this RT-PCR 10 s) on ice in a lysis buffer, as described previously (10), using 0.5 ml lysis analysis. buffer/200 mg tissue. The homogenate was then incubated for 30 min on ice, ELISA of FGF6 Concentration. To determine whether the ex- and insoluble material was removed by centrifugation for 1 min in a micro- pression detected by RT-PCR corresponded to increased expression of centrifuge at 4°C. biologically significant quantities of FGF6, we analyzed tissue ex- Immunohistochemistry. Frozen tissue sections were fixed in acetone for tracts from a total of 24 prostate cancers for FGF6 content by ELISA. 10 min, postfixed in methanol for an additional minute, and stored at Ϫ80°C. Immunohistochemical localization of FGF6 was carried out by the avidin- These samples included 20 clinically localized cancers, a pelvic biotin complex method as described previously (10). All sections were treated lymph node metastasis, and 3 locally advanced, androgen-indepen- with Autoblocker (R&D Systems) to inhibit endogenous peroxidase and avi- dent cancers. All tissue extracts contained 40–90% cancer by frozen din/biotin (Vector Laboratories, Burlingame, CA) to block endogenous biotin. section analysis. As controls, we also analyzed eight tissue extracts The sections were incubated with rabbit polyclonal anti-FGF6 antibody (200 from the normal peripheral zone and four extracts from the normal ng/ml; R&D Systems) at 4°C for 12 h. After liberal washing with PBS (pH transition zone. In addition, nine extracts from prostate tissue with 7.4), sections were incubated with the appropriate biotinylated secondary high-grade PIN, which is widely believed to be the precursor lesion antibody at a 1:200 dilution (Vector Laboratories). Sections were then washed for invasive prostate carcinoma, were analyzed. As shown in Fig. 2 with PBS containing 0.1% Tween 20 and incubated with avidin-biotin complex the normal control tissues did not express detectable quantities of (Vectastain Elite; Vector Laboratories) for 15 min. The antigen-antibody FGF6, whereas 15 of 24 prostate cancers contained between 0.5 and reaction was demonstrated using diaminobenzidine as a substrate, and the sections were then counterstained with hematoxylin. Staining of the cancer 2.6 ng FGF6/g tissue. No correlation between FGF6 content and cells was graded as 1ϩ (weak), 2ϩ (moderate), or 3ϩ (strong). The specificity differentiation, as determined by the Gleason score or clinical or of staining was confirmed by preincubation of a 1 ␮g/ml solution of anti-FGF6 pathological stage, was apparent, but the number of cases is small, and antibody with an equal volume of recombinant FGF6 at 10 ␮g/ml for1hat more extensive analysis will need to be performed to definitively room temperature before immunohistochemistry. This pretreatment completely abolished immunostaining for this antibody. ELISA of FGF6 Concentration. Each well of a 96-well plate was coated with 100 ␮l of a solution of anti-FGF6 monoclonal antibody (MAB 238; R&D Systems) at a concentration of 4 ␮g/ml overnight at room temperature in a sealed bag. The next day, wells were washed three times with PBS containing 0.05% Tween 20 and incubated overnight at 4°C with 300 ␮l of a blocking solution (Kirkegaard and Perry Laboratories, Gaithersburg, MD). The plate Fig. 1. Expression of FGF6 mRNA in human prostate cancer. RNAs from normal was then washed as described above, and standards and samples were added prostatic peripheral zone (Lane N) or clinically localized prostate cancers (Lanes 1–12) (100 ␮l/well). Samples were 100 ␮l of pure tissue extract. Wells were then were isolated and reverse transcribed, and cDNAs were analyzed by PCR with FGF6- specific primers, followed by transfer of PCR products to nylon membranes and hybrid- incubated for3hatroom temperature. After washing as described above, ization with a radiolabeled FGF6 probe. The washed blot was exposed to X-ray film for biotinylated polyclonal anti-FGF6 antibody (BAF 238; R&D Systems) was approximately 30 min. The FGF6 product was 207 bp. 4246

determine whether FGF6 expression is correlated with clinical or pathological parameters. We also detected between 0.4 and 0.9 ng FGF6/g tissue in four of nine high-grade PIN lesions. Thus, we found expression of FGF6 protein in approximately one-half of PIN lesions and invasive prostate carcinomas, which is similar to the percentage of prostate cancers expressing FGF6 mRNA by RT-PCR analysis. Immunohistochemical Analysis of FGF6 Expression. To determine the cellular localization of FGF6 in the prostatic tissues, we examined normal peripheral zone, PIN, and cancers by immunohistochemistry with affinity-purified rabbit polyclonal anti-FGF6 anti- bodies. As shown in Fig. 3, in the normal peripheral zone, weak staining of the basal cell layer was detectable (Fig. 3A). In one sample, we also found staining of skeletal muscle fibers (data not shown). It is well known that occasional skeletal muscle fibers can be found in normal prostate, and FGF6 has previously been shown to be expressed by skeletal muscle at other sites (15); therefore, the positive staining observed is consistent with prior observations. In the high-grade PIN lesions, we found a marked increase in the staining of the basal cells in the dysplastic acini (Fig. 3, B and C). This staining was distinctly stronger than that seen in adjacent, normal acini in the same section. Fig. 2. FGF6 content of normal and neoplastic prostate tissues. Tissue extracts were prepared as described in “Materials and Methods” from normal peripheral and transition Preincubation of the anti-FGF6 antibody with excess recombinant zone tissues (NL), prostate tissues with high-grade PIN, or prostate cancer tissues (CA), FGF6 completely abolished immunostaining (Fig. 3D). A total of 10 and the FGF6 content was determined by ELISA. For the cancer tissues, open circles are carcinomas with detectable FGF6 by ELISA were examined by im- the values for a metastasis to a pelvic lymph node (the highest value) and three transurethral resections in men with locally advanced, androgen-independent prostate cancer, and munohistochemistry. In eight of these carcinomas, staining of the closed circles are the values for clinically localized cancers. cytoplasm of the cancer cells was identified. This staining varied from

Fig. 3. Immunohistochemistry of prostate, PIN, and prostate cancer with anti-FGF6 antibody. Frozen sections of prostatic tissues were analyzed by immunohistochemistry using anti-FGF6 antibody as described in “Materials and Methods.” A, normal prostate showing weak staining of basal epithelial cells (ϫ400). B, low- power view of high-grade PIN showing intense staining of basal epithelial cells (ϫ40). C, high-power view of a PIN lesion (ϫ400). D, high-power view of a PIN lesion stained with antibody preincubated with excess recombinant FGF6 (negative control; ϫ400). E, prostate carcinoma showing strong staining of the cytoplasm of the cancer cells (ϫ400). F, prostate cancer with staining of the spindle-shaped fibroblastic cells within the cancer tissue (ϫ200).

4247

Fig. 4. Cell proliferation in response to FGF6. Cells were plated and treated with FGF6 (10 ng/ml) or kept in basal media containing insulin as the only growth factor (primary cultures and PNT1A) or 0.2% serum (prostate cancer cell lines) and counted at the indicated times. A, primary epithelial and stromal cell cultures. B, PNT1A, an immortalized prostatic epithelial cell line. C, prostate cancer cell lines DU145, PC3, and LNCaP. Error bars, SD of triplicate determinations.

light but diffuse (1ϩ) to intense and diffuse (3ϩ). The single case PC3 prostate cancer cell line had only a slight (but statistically with 1ϩ staining was derived from a metastatic cancer to a pelvic significant) proliferative response to FGF6. Thus, with the exception lymph node that was Ͼ90% cancer on frozen section and contained of PC3 cells, the normal and transformed prostatic epithelial cells 2.6 ng FGF6/g tissue by ELISA. In six of eight cases, there was 2ϩ tested showed a marked proliferative response to exogenous FGF6. staining for FGF6, characterized by variable weak to intense staining Expression of FGFR-4 in Primary Prostatic Cells and Prostate of the tumor cytoplasm. Finally, a single case of 3ϩ staining with Cancer Cell Lines. It has been shown by Ornitz et al. (2) that the uniform intense staining of the prostate cancer epithelial cells was FGFR with the most potent mitogenic response to FGF6 is FGFR-4. found (Fig. 3E). In two prostate cancers, there was no staining of the We therefore examined the expression of FGFR-4 in the normal cancer epithelial cells, but staining of stromal cells with fibroblastic peripheral zone of human prostate and in DU145, PC3, and LNCaP morphology was seen (Fig. 3F). prostate cancer cell lines by RT-PCR. The breast cancer cell line Mitogenic Activity of FGF6 on Prostatic Cells. To determine the MDA-MB231, which has been shown to express FGFR-4 (16, 17), role of FGF6 in the proliferation of normal and neoplastic prostatic was included as a positive control. As can be seen in Fig. 5A, all four epithelial cells, we assessed the ability of exogenous recombinant prostate samples and the prostate and breast cancer cell lines express FGF6 to induce proliferation in primary cultures of prostatic cells and FGFR-4. A control in which genomic DNA was added results in a in normal and neoplastic prostatic epithelial cells lines. As seen in Fig. larger 1.2-kb product, confirming that the primers span an intron and 4A, FGF6 was mitogenic for both primary cultures of prostatic epi- the product seen in the RT-PCR reactions is not due to genomic DNA thelial and stromal cells. The differences seen are highly statistically contamination. To determine the site of FGFR-4 expression in normal significant (P Ͻ 0.001, t test) at 4 and 6 days of treatment for both prostate, RNAs from primary cultures of prostatic epithelial and epithelial and stromal cells. Similar results were seen with PNT1A stromal cells were analyzed by RT-PCR. As seen in Fig. 5B, both cells, a cell line derived from normal prostatic epithelial cells immor- epithelial and stromal cells express FGFR-4. Northern blotting using talized by SV40 large T antigen (Fig. 4B). In addition, a marked 1 ␮g of polyadenylated RNA from the three prostate cancer cell lines proliferative response to exogenous FGF6 was also seen in the pros- and the immortalized normal prostatic epithelial cell line PNT1A tate cancer cell lines DU145 and LNCaP, which was again highly reveals a low but detectable expression FGFR-4 in all four cell lines, statistically significant by day 4 of treatment (P Ͻ 0.001, t test). The with PC3 cells having the lowest FGFR-4 expression (Fig. 6). 4248

FGF6 overexpression observed is shown by the ability of exogenous FGF6 to promote proliferation of normal, immortalized, and fully transformed prostatic epithelial cells. In addition, by promoting the growth of stromal cells, FGF6 may enhance the production of paracrine growth factors by tumor stroma (20), which can further promote tumor progression. FGF6 is normally expressed in the skeletal muscle and may play an important role as a growth factor in that tissue. Studies by Penault- Llorca et al. (21) have shown expression of FGF6 in 3 of 10 breast cancer cell lines and in 15% of primary breast cancer tissues. No FGF6 expression was detected in normal breast tissue. These same Fig. 5. Expression of FGFR-4 by RT-PCR. cDNAs corresponding to 0.1 ␮g of total authors detected overexpression of FGFR-4 in 32% of primary breast RNA were used for PCR with FGFR-4-specific primers, analyzed by gel electrophoresis cancers. Expression of FGFR-4 has been shown by others to be in 2% agarose gels, and stained with ethidium bromide. A, control, no cDNA (Lane C); prostate peripheral zone tissue (Prostate, Lanes 1–4); DU145 (Lane D), PC3 (Lane P), present in the majority of breast cancer cell lines (16, 17). Given that and LNCaP (Lane L) prostate cancer cell lines; MDA-MB231 breast cancer cells (Lane FGFR-4 is the most potent receptor for FGF6 (2), a potential autocrine M); and 1 ␮g of human genomic DNA (Lane G). B, control, no cDNA (Lane C), primary epithelial cultures (Lane E), and primary stromal cell cultures (Lane S). R4 indicates the loop involving FGF6 and FGFR-4 may be established in a subset of product of the predicted size for FGFR-4 amplified from cDNA (364 bp). The larger breast cancers. We have also found expression of FGFR-4 in all three (1200-bp) product in the genomic control lane is amplified across introns (leading to a prostate cancer cell lines and in primary prostate epithelial cultures. It larger product), thus confirming that the smaller product in the cDNA lanes is not a result of genomic DNA contamination of the cDNAs. should be noted that this expression is relatively low, and we did not detect it previously in the normal prostate (9); however, using im- proved PCR primers allowed us to detect this expression. Chandler et DISCUSSION al. (16) have also reported expression of FGFR-4 in DU145 cells using a RNase protection assay, which is consistent with our results, Members of the FGF family are known to play an important role as but they found barely detectable expression of FGFR-4 in PC3 cells, growth factors in the normal human prostate (9–11). We report here whereas we found FGFR-4 expression that was lower than but similar our finding of increased expression of FGF6 in PIN, the precursor of to that seen in DU145 cells. Whether this difference reflects the use of invasive prostate cancer, and in localized and metastatic prostate different probes and methods for RNA analysis or differences in the cancers. FGF6 is expressed at an extremely low level in the basal cells PC3 cells between different laboratories is not clear. It is of interest to of normal human prostate that is undetectable by RT-PCR or ELISA note that both prostate (22–24) and breast cancers (25) have been but can be detected by immunohistochemistry. Given that the basal reported to have increased expression of FGF8, a potent activator of cells constitute only a small fraction of the total cells in the prostate FGFR-4. Thus, in both breast and prostate cancer, there is evidence and that the expression of FGF6 was quite low on immunohistochem- that autocrine stimulation of FGFR-4 by FGF6 and FGF8 may occur. istry, it is certainly conceivable that FGF6 expression in normal It is known that the signaling properties of FGFR-4 differ from those prostate is below the limits of detection of RT-PCR, although this is of the other FGFRs (26–29) in that FGFR-4 only weakly stimulates a very sensitive technique. In a significant proportion of cases of mitogen-activated protein kinase activation (27) and is associated with

high-grade PIN, we observed an increased expression of FGF6 by the a unique Mr 85,000 serine kinase (27, 28) that may be important in basal cells that was obvious by immunohistochemistry, and we have FGFR-4 signaling. Johnston et al. (29) have shown that FGFR-4 is the confirmed this overexpression by the detection of FGF6 using ELISA. only FGFR that can promote membrane ruffling when tranfected into Because the basal cells in PIN are not neoplastic, it is likely that the COS-7 cells. Such membrane ruffling is associated with changes in dysplastic luminal cells directly or indirectly induce the expression of the actin cytoskeleton related to increased motility. Thus, FGFR-4 FGF6 by the basal cells. Because FGF6 contains a signal peptide and activation by autocrine stimulation may be more important in altering is actively secreted, the increased production of FGF6 by the basal motility or other properties when compared with similar stimulation cells may play a role in the increased proliferation of the dysplastic of other FGFRs and may therefore play a unique role in tumor epithelial cells seen in PIN (18). To our knowledge, this is the first progression. report of expression of a growth factor by prostatic basal cells in adult In addition to FGFR-4, other FGFR isoforms that can respond to prostate and of the increased expression of such a growth factor in PIN. In the prostate cancer tissues, we have shown increased expression of both FGF6 mRNA and protein. In most cases, FGF6 was localized in the cancer cells by immunohistochemistry, consistent with an autocrine growth stimulation. In two cases, we observed immunore- active FGF6 primarily in fibroblastic cells, consistent with the induc- tion of FGF6 production by these cells with potential paracrine effects on adjacent cancer cells. A similar mixture of overexpression by both epithelial and stromal cells has been described by Kornmann et al. (19) for the increased expression of FGF5 seen in pancreatic cancer; therefore, this observation is not without precedent. Thus, in prostate cancers, FGF6 that is normally secreted by the basal cells is no longer ␮ available as a paracrine growth factor, and in many cases, the cancers Fig. 6. Northern blot analysis of FGFR-4 expression. Northern blot using 1 gof polyadenylated mRNA from DU145, PC3, and LNCaP cell lines and immortalized normal appear to have developed mechanisms to compensate for this loss prostate epithelial cells (PNT1A) was probed with a FGFR-4 probe from the 3Ј-untrans- either by autocrine production of FGF6 or by inducing adjacent lated region of this gene, washed, and exposed to X-ray film for 2 days. The indicated FGFR-4 band is approximately 3.5 kb. The same filter was stripped and reprobed with stromal cells to produce it and, in fact, produce it in larger amounts glyceraldehyde 3-phosphate dehydrogenase probe (GAPDH), washed, and exposed to than in normal tissues. The potential biological importance of the X-ray film overnight. 4249

FGF6, such as FGFR-1 IIIc and FGFR-2 IIIc (2) are also expressed in 8. Fox, J., and Shanley, J. Antisense inhibition of basic fibroblast growth factor induces prostate cancer cells. We have shown that expression of FGFR-1 is apoptosis in vascular smooth muscle cells. J. Biol. Chem., 271: 12578–12584, 1996. 9. Ittmann, M., and Mansukhani, A. Expression of fibroblast growth factors (FGFs) and increased, particularly in poorly differentiated prostate cancers (20), FGF receptors in human prostate. J. Urol., 157: 351–356, 1997. and we have recently found that FGFR-1 is present only as the IIIc 10. Giri, D., Ropiquet, F., and Ittmann, M. FGF9 is an autocrine and paracrine prostatic 4 growth factor expressed by prostatic stromal cells. J. Cell. Physiol., 180: 53–60, isoform in prostate cancers. FGFR-1 IIIc could markedly enhance 1999. response to FGF6 in those cases where it is expressed. Similarly, there 11. Ropiquet, F., Giri, D., Lamb, D., and Ittmann, M. FGF7 and FGF2 are increased in is evidence that the DU145 prostate cancer cell line and some prostate benign prostatic hyperplasia and are associated with increased proliferation. J. Urol., 162: 595–599, 1999. cancer xenografts express the FGFR-2 IIIc isoform (30) that can bind 12. Yan, G., Fukobori, Y., McBride, G., Nikolaropolous, S., and McKeehan, W. Exon FGF6. However, FGFR-1 IIIc and FGFR-2 IIIc are not expressed in switching and activation of stromal and embryonic fibroblast growth factor (FGF)- cultured primary prostatic epithelial cells (9); therefore, in normal FGF receptor genes in prostate epithelial cells accompany stromal independence and malignancy. Mol. Cell. Biol., 8: 4513–4522, 1993. prostatic epithelial cells, most of the mitogenic effect of FGF6 must be 13. Partanen, J., Makela, T. P., Eerola, E., Korhonen, J., Hirvonen, H., Claesson-Welsh, through activation of FGFR-4. Additional studies are needed to define L., and Alitalo, K. FGFR-4, a novel acidic fibroblast growth factor receptor with a the expression of the various FGFR isoforms that bind FGF6 in distinct expression pattern. EMBO J., 10: 1347–1354, 1991. 14. Cussenot, O., Berthon, P., Berger, R., Mowszowicz, I., Faille, A., Hojman, F., Teillac, prostate cancer cells in vivo and the relative importance of each of P., Le Duc, A., and Calvo, F. Immortalization of human adult normal prostatic these receptors in the biological response of prostate cancer cells to epithelial cells by liposomes containing large T-SV40 gene. J. Urol., 146: 881–886, FGF6. 1991. 15. de Lapeyriere, O., Rosnet, O., Benharroch, D., Raybaud, F., Marchetto, S., Planche, In summary, we have found that FGF6 is expressed by normal J., Galland, F., Mattei, M., Copeland, N. G., Jenkins, N. A., Coulier, F., and prostatic basal cells in extremely small amounts, and expression in Birnbaum, D. Structure, chromosome mapping and expression of the murine Fgf-6 basal cells is markedly increased in PIN lesions. Approximately half gene. Oncogene, 5: 823–831, 1990. 16. Chandler, L. A., Sosnowski, B. A., Greenlees, L., Aukerman, S. L., Baird, A., and of prostate cancers express FGF6 in the cancer cell themselves or in Pierce, G. F. Prevalent expression of fibroblast growth factor (FGF) receptors and the tumor stroma. The acquisition of FGF6 expression by the prostate FGF2 in human tumor cell lines. Int. J. Cancer, 81: 451–458, 1999. 17. Ron, D., Reich, R., Chedid, M., Lengel, C., Cohen, O. E., Chan, A. M., Neufeld, G., cancers implies but does not prove that it may play a role in cancer Miki, T., and Tronick, S. R. Fibroblast growth factor receptor 4 is a high affinity cell proliferation or perhaps in other aspects of tumor progression. receptor for both acidic and basic fibroblast growth factor but not for keratinocyte Other FGFs, such as FGF2 (20) and FGF8 (22–24), have also been growth factor. J. Biol. Chem., 268: 5388–5394, 1993. 18. Harper, M. E., Glynne-Jones, E., Goddard, L., Mathews, P., and Nicholson, R. I. shown to be increased in prostate cancer, and a number of other Expression of androgen receptor and growth factors in premalignant lesions of the growth factors, including EGF and transforming growth factor ␣ (31), prostate. J. Pathol., 186: 169–177, 1998. have also been implicated in prostate cancer. The relative importance 19. Kornmann, M., Ishiwata, T., Beger, H., and Korc, M. Fibroblast growth factor-5 stimulates mitogenic signaling and is overexpressed in human pancreatic cancer: of each of these growth factors in prostate cancer progression has yet evidence for autocrine and paracrine actions. Oncogene, 15: 1417–1424, 1997. to be elucidated. Further work will be needed to define the role of the 20. Giri, D., Ropiquet, F., and Ittmann, M. Alterations in expression of FGF2 and its increased expression of FGF6 in human prostate cancer. receptor FGFR-1 in human prostate cancer. Clin. Cancer Res., 5: 1063–1071, 1999. 21. Penault-Llorca, F., Bertucci, F., Adelaide, J., Parc, P., Coulier, F., Jacquemier, J., Birnbaum, D., and DeLapeyiere, O. Expression of FGF and FGF receptor genes in human breast cancer. Int. J. Cancer, 61: 170–176, 1995. ACKNOWLEDGMENTS 22. Leung, H., Dickson, C., Robson, C., and Neal, D. Overexpression of fibroblast growth factor-8 in human prostate carcinoma. Oncogene, 12: 1833–1835, 1996. We thank Karen Schmidt for skilled technical assistance. 23. Tanaka, A., Furuya, A., Yamasaki, M., Hanai, N., Kuriki, K., Kamiakito, T., Koba- yashi, Y., Yoshida, H., Koike, M., and Fukayama, M. High frequency of fibroblast growth factor (FGF) 8 expression in clinical prostate cancers and breast tissues: immunohistochemically demonstrated by a newly established neutralizing mono- REFERENCES clonal antibody against FGF 8. Cancer Res., 58: 2053–2056, 1998. 24. Dorkin, T. J., Robinson, M. C., Marsh, C., Bjartell, A., Neal, D. E., and Leung, H. Y. 1. Johnson, D., and Williams, L. Structural and functional diversity in the FGF receptor FGF8 over-expression in prostate cancer is associated with decreased patient survival multigene family. Adv. Cancer Res., 60: 1–41, 1993. and persists in androgen independent disease. Oncogene, 18: 2755–2761, 1999. 2. Ornitz, D., Xu, J., Colvin, J., McEwen, D., MacArthur, C., Coulier, F., Gao, G., and Goldfarb, M. Receptor specificity of the fibroblast growth factor family. J. Biol. 25. Marsh, S. K., Bansal, G. S., Zammit, C., Barnard, R., Coope, R., Roberts-Clarke, D., Chem., 271: 15292–15297, 1996. Gomm, J. J., Coombes, R. C., and Johnston, C. L. Increased expression of fibroblast 3. Basilico, C., and Moscatelli, D. The FGF family of growth factors and oncogenes. growth factor 8 in human breast cancer. Oncogene, 18: 1053–1060, 1999. Adv. Cancer Res., 59: 115–165, 1992. 26. Wang, J. K., Gao, G., and Goldfarb, M. Fibroblast growth factor receptors have 4. Hector-Schulz, S., and Haghayegh, S. Basic fibroblast growth factor expression in different signaling and mitogenic potentials. Mol. Cell. Biol., 14: 181–188, 1994. human and murine squamous cell carcinomas and its relationship to regional endo- 27. Vainikka, S., Joukov, V., Wennstrom, S., Bergman, M., Pelicci, P. G., and Alitalo, K. thelial cell proliferation. Cancer Res., 53: 1444–1449, 1993. Signal transduction by fibroblast growth factor receptor-4 (FGFR-4): comparison 5. Miyake, H., Yoshimura, K., Hara, I., Eto, H., Arakawa, S., and Kamidono, S. Basic with FGFR-1. J. Biol. Chem., 269: 18320–18326, 1994. fibroblast growth factor regulates matrix metalloproteinases production and in vitro 28. Vainikka, S., Joukov, V., Klint, P., and Alitalo, K. Association of a 85-kDa serine invasiveness in human bladder cancer cell lines. J. Urol., 157: 2351–2355, 1997. kinase with activated fibroblast growth factor receptor-4. J. Biol. Chem., 271: 6. Galzie, Z., Fernig, D., Smith, J., Poston, G., and Kinsella, A. Invasion of human 1270–1273, 1996. colorectal carcinoma cells is promoted by endogenous basic fibroblast growth factor. 29. Johnston, C. L., Cox, H. C., Gomm, J. J., and Coombes, R. C. bFGF and aFGF induce Int. J. Cancer, 71: 390–395, 1997. membrane ruffling in breast cancer cells but not in normal breast epithelial cells: 7. Ropiquet, F., Huguenin, S., Villette, J. M., Ronfle, V., Le Brun, G., Maitland, N. J., FGFR-4 involvement. Biochem. J., 306: 609–616, 1995. Cussenot, O., Fiet, J., and Berthon, P. FGF7/KGF triggers cell transformation and 30. Carstens, R. P., Eaton, J. V., Krigman, H. R., Walther, P. J., and Garcia-Blanco, M. A. invasion on immortalized human prostatic epithelial PNT1A cells. Int. J. Cancer, 82: Alternative splicing of fibroblast growth factor receptor 2 (FGF-R2) in human 237–243, 1999. prostate cancer. Oncogene, 15: 3059–3065, 1997. 31. Sherwood, E. R., and Lee, C. Epidermal growth factor-related peptides and the epidermal growth factor receptor in normal and malignant prostate. World J. Urol., 4 Unpublished data. 13: 290–296, 1995.

4250

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2000 American Association for Cancer Research. Increased Expression of Fibroblast Growth Factor 6 in Human Prostatic Intraepithelial Neoplasia and Prostate Cancer

Frederic Ropiquet, Dipak Giri, Bernard Kwabi-Addo, et al.

Cancer Res 2000;60:4245-4250.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/60/15/4245

Cited articles This article cites 30 articles, 9 of which you can access for free at: http://cancerres.aacrjournals.org/content/60/15/4245.full#ref-list-1

Citing articles This article has been cited by 12 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/60/15/4245.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/60/15/4245. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.