The cells' ability to proliferate in response to growth factor stimulation is significantly altered during cancer progression. To investigate the mechanisms underlying these alterations in prostate cancer, the role and expression of β1A integrin and type 1 insulin-like growth factor receptor (IGF-IR), known to contribute to cell proliferation and transformation, were analyzed. Using small interfering RNA oligonucleotides to down-regulate β1A, we show that β1A expression is required for IGF-IR–mediated prostate cancer cell proliferation and anchorage-independent growth. In vivo, using age-matched transgenic adenocarcinoma of mouse prostate (TRAMP) mice at different stages of prostate cancer [prostatic intraepithelial neoplasia, PIN; well-differentiated adenocarcinoma, WD; and poorly differentiated adenocarcinoma, PD], the expression of β1A and of IGF-IR was studied. β1A and IGF-IR expression levels were concurrently up-regulated in high PIN and WD, whereas their expression did not correlate in late-stage PD. In contrast to the up-regulated expression of β1A, the levels of β1C, a β1 cytoplasmic variant that inhibits cell proliferation, were down-regulated in all stages of prostate cancer. A similar expression pattern was observed for a β1C downstream effector, Grb2-associated binder-1 (Gab1) which is known to inhibit IGF-IR phosphorylation. To analyze in vitro the mechanistic implications of β1A, β1C, and Gab1 deregulation in prostate cancer, we investigated whether expression of either β1 variant in β1-null cells affected IGF-IR localization. We found that IGF-IR and β1A were colocalized in highly specialized integrin signaling compartments, designated focal contacts. However, in the presence of β1C, IGF-IR remained diffuse on the cell surface and did not localize to focal contacts. The findings that β1 integrins and IGF-IR are concurrently deregulated and that expression of β1 integrins is necessary to achieve appropriate IGF-IR intracellular distribution point to the important role that the cross-talk between these receptors may have during prostate cancer progression and will be helpful in formulating new therapeutic strategies.
- Prostate Cancer
- Focal Contacts
Prostate cancer remains a significant health issue for aging men and is the second leading cause of cancer death among American men ( 1). Whereas substantial advances have been made in the diagnosis and treatment of prostate cancer, the underlying causes of this disease remain to be fully investigated. The development of prostate cancer proceeds through a series of defined stages, including preinvasive disease (prostatic intraepithelial neoplasia, PIN), invasive cancer, and androgen-dependent or androgen-independent metastases ( 2, 3).
The ability of neoplastic prostate cells to grow and invade is mediated by interactions with the surrounding cells and extracellular matrix–mediated by integrins ( 4). Integrins are crucial regulators of differentiation, growth, survival, migration, and invasion ( 5, 6). In prostate cancer, tumor cells have a markedly different surrounding matrix than normal cells; thus, changes in the integrin profile may be functionally relevant and contribute to cancer progression ( 7, 8). A number of studies have reported changes in integrin expression as prostate cancer progresses to an advanced stage ( 7, 9, 10).
Integrins are heterodimers consisting of an α and β subunit ( 6). The cytoplasmic domain of the β1 subunit, in its canonical form (β1A), is highly conserved and is a positive regulator of cell proliferation, development, migration, and integrin localization ( 11, 12). The β1C integrin is an alternatively spliced variant of the β1 subfamily that contains a unique sequence in its cytoplasmic domain ( 12). Previous studies have shown that either full-length β1C or its cytoplasmic domain inhibits prostate cancer cell proliferation ( 12). β1C is expressed in the nonproliferative, differentiated epithelium and is selectively down-regulated in prostatic adenocarcinoma ( 13– 16).
Several reports in the literature suggest that there are physical and functional interactions between integrins and components of growth factor signaling pathways, including insulin-like growth factor (IGF) or its downstream signaling proteins ( 17). However, the role of these interactions between type 1 IGF receptor (IGF-IR) and integrins in regulating the progression of prostate cancer is still not clear. The IGF axis is an important modulator of growth and development, and changes in this axis may have important implications in malignant growth ( 18, 19). The IGF and its receptor (IGF-IR) are found in human prostate stroma and epithelial cells ( 20). There is, however, no consensus regarding the relative levels of IGF-IR expression in benign and malignant prostate epithelium and the role of IGFs in metastases. In 1996, Tennant et al. reported that IGF-IR expression is significantly lowered at the protein and RNA level in malignant versus benign prostate epithelium ( 21). Similarly, Chott et al. reported that IGF-IR is expressed in luminal and basal epithelium in nonneoplastic prostate but is lost in metastatic androgen-independent prostate cancer ( 22). However, Kaplan et al. did fail to show a significant increase in IGF-IR mRNA during primary prostate tumor progression in transgenic adenocarcinoma of mouse prostate (TRAMP) mice, indicating the possibility of posttranscriptional regulation of IGF-IR but showed reduced levels of IGF-IR mRNA in metastatic lesions and in androgen-independent disease ( 23). Subsequently, Nickerson et al. reported that progression to androgen independence is associated with increased expression of IGF-I and IGF-IR as well as increased tyrosine phosphorylation of IGF-IR ( 24). Finally, in 2002, another study has shown that IGF-IR was significantly up-regulated at the protein and mRNA level in primary prostate cancer compared with benign prostatic epithelium ( 25).
Most signaling pathways stimulated by receptor tyrosine kinases use scaffolding adaptors like insulin receptor substrate (IRS) or Gab known to be involved in IGF signaling axis ( 26, 27). Upon the addition of growth factors, Grb2-associated binder (Gab1) becomes tyrosine phosphorylated and recruits Src homology-2 domain-containing tyrosine phosphatase 2 (Shp2; ref. 28). In our previous study, we have shown that the Gab1/Shp2 complex is recruited to the cell surface by integrins. Upon recruitment to cell surface, Gab1/Shp2 complex causes IGF-IR dephosphorylation and causes change in cell adhesion to laminin in response to IGF-I ( 29). Previous studies have shown that overexpression of Gab1 increases growth factor responsiveness and tumorigenicity in NIH3T3 cells ( 26). A recent study has shown that overexpression of Gab1 lacking the pleckstrin homology domain enhances (by 3-fold), whereas wild-type Gab1 inhibits, epidermal growth factor–induced soft agar colony formation of preneoplastic Syrian hamster cells ( 30). However, studies carried out to investigate whether changes in expression levels of Gab1 occur during cancer progression have not been done.
In the present study, we show that IGF-IR–mediated prostate cancer cell proliferation and anchorage-independent growth is inhibited by β1A down-regulation. Our study highlights a role for the β1A integrin as modulator of IGF-IR–mediated functions in prostate cancer.
Materials and Methods
Antibodies. The following mouse monoclonal antibodies (mAb) were used: to all human or mouse β1 integrins, Clone-18 (BD Bioscience, San Jose, CA); to human β1 integrin (K-20; Immunotech, Marseille, France); to phosphatidylinositol 3-kinase (PI3-K; p85; BD Bioscience); mouse IgM specific to β1C integrin, 7D5/BF10 ( 31); to vinculin (Chemicon, Temecula, CA). The following rabbit polyclonal antibodies were used: to all β1 integrins (M-106; Santa Cruz Biotechnology, Santa Cruz, CA); to Akt (Cell Signaling, Beverly, MA); to extracellular signal-regulated kinase-2 (ERK-2, Santa Cruz Biotechnology); to cytoplasmic domain of β1A integrin (kindly provided by Dr. E. Ruoslahti, Cancer Research Center, The Burnham Institute, La Jolla, CA); antibody to IGF-IR− (Santa Cruz Biotechnology); to β5 integrin (Chemicon); to β1C integrin ( 13); to Gab1 (Upstate Biotechnology, Lake Placid, NY). Normal mouse IgM (ni-Ab) and normal rabbit IgG (rIgG) were from Calbiochem (La Jolla, CA).
Transgenic mice. TRAMP mice on a pure C57BL/6 background were bred in the colony at the University of Alabama at Birmingham School of Medicine. Transgenic females were bred with nontransgenic littermates (NT) males because transgenic males tend to develop prostate tumors. All mice were maintained in a climate-controlled environment with a 12-hour light/12-hour dark cycle and diet and water supplied ad libitum. Breeders were fed standard pellet mouse feed (Harlan Teklad 7012, Madison, WI). Following weaning at 3 to 4 weeks of age, the gender of offspring was determined, males were separated from females, and a tail biopsy was collected from each mouse. Tail DNA, isolated by standard procedures, was used for determination of transgene incorporation by PCR as described previously ( 32– 34). All TRAMP mice develop premalignant changes resembling human PIN, which ultimately progress to poorly differentiated (PD) prostatic adenocarcinomas that metastasize to distant sites, primarily the lymph nodes, bones, and lungs ( 33). The present study focused on PD (grade 6) as an example of advanced prostate cancer, WD (grade 4) as an example of an early-stage of adenocarcinoma, prostates grade 3 as an example of PIN, and normal prostate (grade 1) from age-matched NT littermates.
Animal care and treatments were conducted in accordance with established guidelines and protocols approved by the University of Alabama Animal Care Committee.
Preparation and analysis of tissues. At necropsy, the urogenital tract, consisting of the prostate, the seminal vesicles, and the emptied bladder, was weighed. A fragment of the DLP, as well as periaortic lymph nodes, were rapidly frozen in liquid nitrogen and stored at −80°C, until further processed. The rest of the DLP was fixed in an acid alcohol solution containing 96% ethanol, 1% glacial acetic acid, and 3% distilled water, as described ( 35). Lymph nodes were fixed in 10% neutral buffered formalin phosphate. The fixed tissues were then embedded in paraffin and 4- to 5-μm sections were mounted on Colorfrost/Plus microscope slides (Fisher Scientific, Pittsburgh, PA). Sections were stained with H&E (lymph nodes) or with Gomori trichrome staining (DLP). DLP sections were evaluated blindly and scored using established criteria ( 32, 34) as normal prostate (score 1), low PIN (score 2), high PIN (score 3), well-differentiated (WD, score 4), moderately differentiated (MD, score 5), or poorly differentiated (PD, score 6). Prostate sections from each mouse were evaluated blindly by Dr. Isam A. Eltoum, a certified Pathologist (Department of Pathology, University of Alabama at Birmingham, Birmingham, Al; refs. 32, 34). Tumors showing grade 2 to 5 were heterogeneous and the pathologist graded them depending on the type of tissue that is most prevalent. The MMHCC Pathology Committee has published very elaborate guidelines in Cancer Research ( 36). Dr. Isam A. Eltoum, the pathologist who has examined these sections has followed those guidelines during his assessments of the tumor grade.
Data analysis. Analysis of the expression levels of IGF-IR and β1 integrin was done using SigmaStat, version 3.0 (SysStat, Inc., Evanston, IL). When the effect of one variable was examined, one-way ANOVA was done if data in multiple groups were normally distributed with equal variance. The Kruskall-Wallis one-way ANOVA on ranks was used if three or more groups were compared and the data were not normally distributed with equal variance. If the latter two tests indicated that the groups were significantly different, a Tukey or Dunn test was carried out, respectively, to determine which of the groups were significantly different from the others. P values for the appropriate tests are given with the results.
Prostate tissue lysis. TRAMP mice were maintained as described above until they were 28 weeks of age. DLPs were collected at necropsy. In this study, we used frozen prostate samples from TRAMP graded 3, 4, and 6 as well as prostate from age-matched NT. Tumor specimens were extracted as follows: specimens were homogenized on ice in a Dounce homogenizer in 100 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 0.1% Triton X-100, 1 mmol/L benzamidine, 10 μg/mL soybean trypsin inhibitor, 10 μg/mL leupeptin, 1 mmo/L phenylmethylsulfonyl fluoride, 1 μg/mL pepstatin A, 1 μmol/L calpain inhibitor, 1 mmol/L Na3VO4, and 1 mmol/L Na4O7P2. SDS was then added to a final concentration of 5%, after which the extracts were boiled for 5 minutes. After boiling, samples were centrifuged at 14,000 × g for 20 minutes, and supernatant was collected for protein determination using bicinchoninic acid protein assay reagent (Pierce, Rockford, IL).
Immunoblot analysis. Proteins from TRAMP prostate tissue lysate (PIN, WD, and PD) or from NT were electrophoresed using 7.5% SDS-PAGE under reducing condition and transferred onto polyvinylidene difluoride membranes, then probed sequentially with mAb to β1 integrins (Clone-18) or to PI3-K or to IGF-IR-β, mouse IgM to β1C integrin (7D5/BF10; ref. 31), normal mouse IgM, rabbit polyclonal antibody to β1C integrin or Gab1, or to β5 integrin. Proteins were detected using enhanced chemiluminescence.
Immunofluorescence microscopy. GD25 murine cells transiently transfected with human β1A or β1C integrin or GD25 cells stably transfected with human β1A integrin using LipofectAMINE 2000 (Invitrogen, San Diego, CA) were cultured as described previously ( 37). Cells were seeded onto 10 μg/mL fibronectin-coated coverslips and processed for immunofluorescence microscopy as described previously ( 38). The primary antibodies used were K-20 to human β1 integrin or a mAb to vinculin or a polyclonal antibody to IGF-IR-β. The secondary antibodies used were goat anti-rabbit IgG coupled to FITC or goat anti-mouse IgG coupled to rhodamine (The Jackson Laboratory, West Grove, PA). One hundred cells per experiment to study the colocalization of IGF-IR and β1 integrins were examined and experiments were done >10 times.
Small interfering RNA–mediated down-regulation of β1A integrin. This study used small interfering RNA (siRNA) to target a region specific to the β1A transcript. The siRNA was designed using conventional criteria ( 39) and corresponded to nucleotides 2425 to 2443 of the coding region in human β1A integrin. The sequence of siRNA was sense strand 5′-AUGGGACACGGGUGAAAAUTT-3′ and antisense strand 5′-AUUUUCACCCGUGUCCCAUTT-3′. These sequences were submitted to BLAST search to ensure that only the β1A integrin was targeted by the β1A integrin siRNA and that control sequences were not homologous to any known genes. siRNA specific to β1A integrin or control were synthesized by Dharmacon (Lafayette, CO). Mouse β1A integrin also have the same last 18 nucleotides in its coding regions as the human β1A integrin and correspond to nucleotides 2502 to 2519 of mouse β1A integrin coding region.
PC3, R−, or R+ cells were cultured as described before ( 29). TRAMP-C2 (American Type Culture Collection, Manassas, VA) cells were grown in DMEM, 2 mmol/L l-glutamine, 5 mg/mL bovine insulin, 10 nmol/L dihydrotestosterone, 5% fetal bovine serum (FBS), and 5% Nu-Serum (BD Biosciences, Bedford, MA). TRAMP-C2, R+, R−, or PC3 cells were transfected with siRNA (100 nmol/L) using oligofectamine (Invitrogen) in serum-free medium. After a 4-hour incubation, cells were fed with complete growth medium described above and incubated for additional 48 hours. After incubation, cells were lysed and proteins were separated on SDS-PAGE under reducing conditions. Proteins were immunoblotted with a mAb to β1 integrins (Clone-18) or to Akt or to ERK.
Sulforhodamine B assay. Proliferation was measured using sulforhodamine B (SRB) assay. PC3, R+, or TRAMP-C2 cells were transfected with siRNA to β1A integrin or control siRNA or oligofectamine alone for 48 hours. Cells were detached and seeded on 96-well cell culture plate (7.5 × 103 cells per well) and allowed to attach for 5 hours in the presence of serum. After attachment, cells were washed and incubated in the presence or in the absence of IGF-I (100 ng/mL) in serum-free medium for 48 hours. After incubation, cells were fixed and stained with SRB as described before ( 29). As a plating control, one set of plates after 5 hours of cell attachments was washed, fixed, and stained with SRB.
Anchorage-independent growth assay. Cell growth in soft agar was assayed by scoring the number of colonies formed in medium (DMEM supplemented with 10% FBS) containing 0.3% agarose, with a 0.5% agarose medium underlay as described ( 40). R− or R+ (2 × 103) or DU145 (5 × 103) cells were seeded on 60-mm diameter plates in triplicate in the presence of IGF-I (100 ng/mL). Cells were fed with 1.5 mL of suspension medium (DMEM supplemented with 10% FBS) every 3 days. The number of colonies larger than 100 μm were counted after 14 days (R+ cells) or after 21 days (DU145 cells). Cell viability was measured 48 hours after siRNA transfection using trypan blue exclusion and plotted as percentage of total viable cells.
Immunohistochemistry. Paraffin-embedded prostate tissue sections from eight mice were baked for 2 hours at 60°C and washed thrice with xylene to deparaffinize the slides. The sections were rehydrated using ethanol and then distilled water. The sections were boiled in sodium citrate using a steamer at 100°C to enhance antigen retrieval. Endogenous peroxidase was quenched with 3% H2O2 for 5 minutes. Blocking was achieved with 50% normal goat serum in TBS containing 0.2% bovine serum albumin for 20 minutes at room temperature. The sections were then incubated with primary antibodies (rabbit polyclonal antibody to β1 integrin, M-106, 0.8 μg/mL or ni-rIgG, 0.8 μg/mL) overnight at 4°C, followed by biotinylated goat anti-rabbit IgG (3 μg/mL, Vector Laboratories, Burlingame, CA) for 30 minutes at room temperature. The sections were incubated with streptavidin-peroxidase complex (5 μg/mL, Vector Laboratories) for 30 minutes at room temperature. Each incubation step was followed by three washes with TBS. Finally, 3,3′-diaminobenzidine tetrahydrochloride dehydrate was used as chromogen (Sigma, St. Louis, MO). Sections were then stained using hematoxylin for 1 minute. Tissue sections were examined on an Olympus BX41 microscope and photographed using Olympus DP12 camera. The immunostaining of β1A integrin was independently evaluated by four investigators, including Dr. David S. Garlick (DVM, Pathologist, Department of Cancer Biology, University of Massachusetts Medical School), H.L.G., M.B., and L.R.L. The expression of β1 integrin in prostate tissue sections of NT or TRAMP was evaluated and recorded as a separate observation.
β1A expression is essential for insulin-like growth factor-I–mediated prostate cancer cell proliferation. To investigate the role of β1A in IGF-I stimulated prostate cancer cell proliferation, we down-regulated the expression of β1A in PC3 or TRAMP-C2 cells using siRNA specific to β1A integrin. As shown in Fig. 1A and B , transfection of β1A siRNA significantly down-regulated β1A expression compared with control siRNA, without affecting the expression of other molecules such as Akt ( Fig. 1A), ERK ( Fig. 1B), or cell viability (data not shown). To investigate the role of β1A in IGF-I stimulated cell proliferation, we transfected PC3 ( Fig. 1C) or TRAMP-C2 ( Fig. 1D) cells with β1A siRNA or control siRNA and measured their proliferative response in the presence or in the absence of IGF-I. Cells transfected with β1A siRNA did not respond to IGF-I ( Fig. 1C and D, left), whereas IGF-I increased cell proliferation when the cells were transfected with control siRNA ( Fig. 1C and D, left). Cells transfected with control siRNA or β1A siRNA did attach to tissue culture plates in a comparable manner; therefore, the effect observed was not a consequence of a differential ability of these cells to adhere to the plate ( Fig. 1C and D, right). Based on these results, we conclude that β1A integrin expression is essential for IGF-I-mediated prostate cancer cell proliferation.
β1A is required for type 1 insulin-like growth factor receptor–mediated cell anchorage-independent growth. IGF-I binds to its receptor, IGF-IR, and stimulates tumor cell growth and cell survival ( 18). Cells devoid of IGF-IR are unable to undergo transformation, suggesting a crucial role for IGF-IR in the establishment and maintenance of the transformed phenotype ( 41). To investigate the role of β1A in IGF-IR-mediated prostate cancer cell transformation, we used R− or R+ cells. R− cells are 3T3-like fibroblasts established from mouse embryos with a targeted disruption of igf-ir gene and cannot be transformed by SV40 T antigen, or an activated Ha-ras, or a combination of both ( 42– 44). R+ cells were generated by transfecting wild-type IGF-IR in R− cells ( 42– 44). We down-regulated β1A in R+ cells using siRNA specific to β1A integrin and measured the effect of IGF-I on cell proliferation. We found significant down-regulation of β1A integrin upon siRNA transfection in R+ cells ( Fig. 2A ). As shown in Fig. 2B, β1A down-regulation prevented IGF-I stimulated R+ cell proliferation ( Fig. 2B, left), but it did not affect cell attachment ( Fig. 2B, right). In R− cells, we observed significant cell death (90%) after transfection with β1A siRNA as compared with R+ cells, indicating a crucial role for β1A integrin in cell survival in the absence of IGF-IR ( Fig. 2C). An analysis of the effect of β1A integrin down-regulation on the cells' ability to grow in anchorage-independent manner shows that R+ cells formed colonies in soft agar when transfected with control siRNA, but their ability to grow in soft agar was completely inhibited by β1A down-regulation ( Fig. 2D). R− cells did not form colonies in soft agar ( Fig. 2D) confirming a previous published observation that IGF-IR is required for anchorage-independent growth ( 40). Similar results were obtained in human prostate cancer cells, DU145, whose viability was not affected by β1A integrin down-regulation, although a complete inhibition of anchorage-independent growth was achieved (data not shown). These results show that β1A integrin expression plays a crucial role in IGF-IR-mediated cell transformation.
β1A and type 1 insulin-like growth factor receptor expression is up-regulated in prostatic intraepithelial neoplasia and well-differentiated prostate carcinoma. Because we have shown that the β1A integrin regulates IGF-IR-mediated transformation ( 29), we investigated whether the expression levels of β1A and IGF-IR correlate during prostate cancer progression in vivo. We studied the expression profile of the β1A integrin and IGF-IR in prostate tissues from TRAMP mice or from NT. We have analyzed five prostates showing tumor grade 3, five prostates showing tumor grade 4, and four prostates showing tumor grade 6. As shown in Fig. 3A-B , we found a significant increased expression of IGF-IR and β1A integrin in all prostate tissues showing grades 3 and 4 compared with prostates from NT mice (P < 0.011). The concurrent up-regulation of β1A and IGF-IR in the early stages of neoplastic transformation in TRAMP mice is consistent with our model that β1A and IGF-IR form a functional complex. In PD tumors, however, the expression of β1A was significantly decreased relative to WD tissues, whereas levels of IGF-IR were comparable with WD tissues, suggesting that enhanced expression of β1A is not required to maintain the proliferative capacity of late-stage tumor cells ( Fig. 3C). Immunohistochemical analysis of TRAMP prostate tissue sections was done in a qualitative manner to determine the localization of β1 integrin. As shown in Fig. 4 , we consistently found that the β1 integrin was expressed in epithelial and stroma cells in WD but was down-regulated in PD tissue specimens. These results indicate that a concurrent expression of IGF-IR and β1 occurs in PIN and WD but not in PD prostate cancer in TRAMP mice.
Down-regulation of β1C integrin and Gab1 expression in prostate carcinoma. The β1C integrin is expressed in human nonproliferative, differentiated epithelium and is selectively down-regulated in prostatic adenocarcinoma ( 13– 16). Similarly, we show here using the TRAMP model of progressive prostate carcinoma that β1C is expressed in normal mouse prostate ( Fig. 5A ), but its expression is down-regulated in PIN, WD, and PD adenocarcinomas ( Fig. 5B). Detergent lysates of PNT1A cells were used as a positive control ( Fig. 5B, lane 1). In contrast, β5 integrin expression was found in normal (NT) and TRAMP prostates at similar levels ( Fig. 5B). These data indicate that selective down-regulation of β1C is an early event during prostate cancer progression.
In our previous study, we have shown that β1C recruits Gab1 to the cell surface and that Gab1 mediates β1C-negative regulation of IGF-IR signaling ( 29). A potential correlation of the expression of β1C and Gab1 was investigated. We analyzed Gab1 expression in tissue lysates from NT, and from PIN, WD and PD TRAMP prostates using an antibody to Gab1. Consistent with the expression profile of β1C integrin, Gab1 seems down-regulated in all TRAMP prostate specimens of grades 3, 4, and 6 compared with normal (NT) prostate tissue specimens ( Fig. 5C). These results indicate that expression of β1C integrin and Gab1 correlates in both normal and neoplastic prostate tissues.
β1A integrin recruits type 1 insulin-like growth factor receptor to focal contacts. In our previous study, we have shown that β1A but not β1C associates with IGF-IR and that β1C expression down-regulates IGF-IR-mediated signaling ( 29). To investigate the mechanism through which β1 integrin regulates IGF-IR signaling, we hypothesized that β1 might regulate the localization of IGF-IR in specific cellular signaling compartments. The experiments were done using GD25 cells that do not express β1 integrin, stably or transiently transfected with β1A or β1C. In β1A stable transfectants, we found that IGF-IR was localized in focal contacts, where vinculin ( Fig. 6A-C ), β1A and several signaling molecules, such as Src and focal adhesion kinase ( 45), are known to be localized in spread cells. In transiently transfected cells, β1A and IGF-IR colocalized in focal contacts ( Fig. 6D-F), whereas upon β1C expression, IGF-IR was not found to be localized in focal contacts ( Fig. 6G-I, arrow). Moreover, in nontransfected GD25 cells, the IGF-IR was found to be diffuse and not localized to focal contacts, suggesting that β1A integrin is essential for the localization of IGF-IR to focal contacts ( Fig. 6G-I, arrowhead). In conclusion, these data indicate that IGF-IR localization is controlled by specific β1 integrins.
The novel observation of this study is that down-regulation of the β1A integrin via siRNA oligonucleotides results in inhibition of IGF-IR-mediated cell proliferation and anchorage-independent growth. It is also shown that whereas β1A and IGF-IR are up-regulated, the β1C integrin cytoplasmic variant and Gab1, a β1C downstream effector, are down-regulated during prostate cancer progression. Finally, β1A's ability to modulate IGF-IR localization in focal contact signaling compartments, is shown to be the mechanism through which β1A controls IGF-IR signaling.
This is the first study that uses siRNA oligonucleotides to integrins to prove a role for the integrin/growth factor receptor cross-talk in prostate cancer and suggests the use of siRNA to down-regulate integrins as a novel therapeutic approach for this disease. The results show that IGF-mitogenic and transforming activities but neither basal proliferation nor cell viability are prevented by siRNA to β1A integrin. Based on this observation, we conclude that β1A expression is essential to activate IGF-IR.
It is well established that activated IGF-IR sends signals required for growth and transformation ( 41). Our results indicate that the synergistic action of IGF-IR and β1A in stimulating mitogenic signals during initiation of carcinogenesis is required. In conclusion, a model can be formulated in which differential expression and function of β1 integrins in prostate cancer modulates cell responses to IGF via IGF-IR. We suggest a paradigm shift that changes the point of view that (a) β1 integrins per se contribute to tumor progression ( 46, 47) and that (b) the IGF-IR per se contributes to tumor progression ( 18). By elucidating the players that contribute to prostate cancer progression, future studies will provide new insights into the current therapeutic approaches that target either integrins or IGF-IR.
The molecular basis and the cross-talk between the β1 integrins and IGF-IR in prostate cancer progression have not been investigated. Our previous study has unraveled a novel mechanism mediated by β1 integrin cytoplasmic variants that differentially regulates IGF-mediated prostate cell adhesion to a basement membrane protein, laminin ( 29). The relevance of this observation to prostate cancer, where up-regulation of the IGF-IR has been described ( 24), is that β1A by limiting cancer cell adhesion to basement membrane proteins and promoting cell proliferation in response to IGF presumably allows the tumor mass to expand and invade. In contrast, β1C, a cytoplasmic variant known to be down-regulated in prostate cancer, has the opposite effect; it completely prevents IGF-mediated cell proliferation and tumor growth, and promotes cell adhesion to laminin. Interestingly, the loss of IGF-IR in the adult prostate seems to cause the cells to become greatly disorganized and loose attachment to basement membrane while remaining viable. 4 Consistent with the hypothesis that β1A integrin and IGF-IR act as a complex that promotes cell proliferation but inhibits cell adhesion to basement membrane, in this study, β1A and IGF-IR are found to be concurrently up-regulated in TRAMP prostate tissues in the early stages of neoplastic transformation, when disruption of the basement membrane occurs.
We describe a concurrent increase in expression of both β1A integrin and IGF-IR in PIN and WD tumors in a defined model of prostate cancer, designated TRAMP; in contrast, PD tumors, where β1A is down-regulated, no longer show this correlation. As previously described for human prostate cancer ( 7), the present study shows that the expression of β1 (β1A and β1C) is altered during cancer progression in TRAMP mice. The findings show, specifically, that β1A integrin expression is up-regulated, whereas β1C is down-regulated during prostate tumor progression. This suggests that up-regulated expression of β1A and IGF-IR might contribute to uncontrolled cell division during initiation of prostate cancer; however, increased levels of both molecules are not required in the late stages of tumor progression. It is, thus, expected that lack of expression of β1A will adversely affect initiation of prostate cancer growth. This is likely to be the result because, similarly, targeted disruption of β1 has been recently shown to prevent initiation of mammary tumorigenesis in vivo in a transgenic mouse model ( 47).
The mechanism by which the β1 integrin cytodomain differentially modulates prostate cancer cell functions in response to IGF and allows IGF-IR to dissect downstream signaling pathways occurs via differential recruitment of IGF-IR downstream effectors to the plasma membrane. A specific association between integrins and IGF-IR effector molecules, specifically β1A and IRS-1, β3 integrins and IRS-1, β3 integrins and DOK1, and β1C and Gab1/Shp2-phosphatase complex, is likely to explain many activities attributed to IGF-IR ( 29, 48– 50). In this study, we show that both β1 integrins and IGF-IR are found to be colocalized in focal contacts where a significant amount of integrin downstream signaling is coordinated. In contrast, expression of β1C that is known to remain diffuse on the cell surface, prevents, or is not able to support IGF-IR localization in focal contacts. The relevance of these findings to prostate cancer is high because both β1C and Gab1 are down-regulated in prostate cancer when β1A is up-regulated. Presumably, in normal tissues, by preventing IGF-IR localization to focal contacts, β1C potentiates Gab1/Shp2 signaling that in turn keeps IGF-IR in a nonphosphorylated state. In contrast, in neoplastic tissues, an appropriate localization of IGF-IR in focal contacts mediated by β1A association with IGF-IR maintains IRS-1 signaling active and supports tumor progression.
In conclusion, deregulated expression of integrins and of IGF-IR and deregulation of their cross-talk is likely to be responsible for several alterations of cell functions and intracellular signaling in prostate cancer. Therefore, these alterations should be regarded as highly significant events in the search for novel therapeutic targets in prostate cancer.
Grant support: NIH grants RO1 CA-89720 and CA109874 (L.R. Languino) and RO1 CA-82807 (N.M. Greenberg), Army grants PCRP DAMD17-98-1-8506 and PCRP DAMD PC040221 (L.R. Languino), and American Institute for Cancer Research grant AICR 00A023 (A. Elgavish).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Natalia Teider for technical assistance with the immunofluorescence studies, Drs. David S. Garlick and Isam A. Eltoum for analyses of immunohistochemical staining, and Marjory Thomas for helping with the article preparation.
↵4 N.M. Greenberg, unpublished results.
- Received December 3, 2004.
- Revision received March 24, 2005.
- Accepted May 13, 2005.
- ©2005 American Association for Cancer Research.