The S100A4 Oncoprotein Promotes Prostate Tumorigenesis in a Transgenic Mouse Model

Abstract

S100A4, a calcium-binding protein, is known for its role in the metastatic spread of tumor cells, a late event of cancer disease. This is the first report showing that S100A4 is not merely a metastatic protein but also an oncoprotein that plays a critical role in the development of tumors. We earlier showed that S100A4 expression progressively increases in prostatic tissues with the advancement of prostate cancer (CaP) in TRAMP, an autochthonous mouse model. To study the functional significance of S100A4 in CaP, we generated a heterozygously deleted S100A4 (TRAMP/S100A4^+/−) genotype by crossing TRAMP with S100A4^−/− mice. TRAMP/S100A4^+/− did not show a lethal phenotype, and transgenes were functional. As compared to age-matched TRAMP littermates, TRAMP/S100A4^+/− mice exhibited 1) an increased tumor latency period (P < 0.001), 2) a 0% incidence of metastasis, and 3) reduced prostatic weights (P < 0.001). We generated S100A4-positive clones from S100A4-negative CaP cells and tested their potential. S100A4-positive tumors grew at a faster rate than S100A4-negative tumors in vitro and in a xenograft mouse model. The S100A4 protein exhibited growth factor–like properties in multimode (intracellular and extracellular) forms. We observed that 1) the growth-promoting effect of S100A4 is due to its activation of NFκB, 2) S100A4-deficient tumors exhibit reduced NFκB activity, 3) S100A4 regulates NFκB through the RAGE receptor, and 4) S100A4 and RAGE co-localize in prostatic tissues of mice. Keeping in view its growth-promoting role, we suggest that S100A4 qualifies as an excellent candidate to be exploited for therapeutic agents to treat CaP in humans.

Introduction

The S100A4 protein belonging to the S100 superfamily of calcium-binding proteins is reported to be highly expressed in embryonic macrophages and differentiating mesenchymal tissues during mouse development.¹ S100A4 overexpression has been reported to be positively correlated to rheumatoid arthritis, tissue fibrosis, and metastasis of tumors.² We recently provided a comprehensive report about the involvement of S100A4 with the prognosis of gastric, colon, pancreatic, thyroid, oral squamous cell, and non–small cell lung carcinoma.² S100A4 is reported to cross-talk with important cellular signaling pathways, which participate in the invasion of tumor cells to distant sites.^2-4 Since the majority of studies have related the S100A4 protein with the development of metastatic phenotypes in tumor cells, it is used as a metastatic marker.^2,5 Metastatic prostatic tumors are hard to treat, and we showed that S100A4 regulates the invasiveness of prostatic tumor cells by regulating the MMP-9 protein that gets tumor cells to loosen up from parent tissues and spread to distant sites.^6,7 Earlier, we showed that S100A4 levels are increased in prostatic tissues of patients with prostatic cancer (CaP) and in transgenic adenocarcinoma of the mouse prostate (TRAMP) mice.^3,4 Although numerous published studies are focused on the role of S100A4 during metastasis (of late-stage carcinomas), very little has been done to determine its status, role, and mechanism of action during the development of disease. The current study is the first to investigate 1) the role of S100A4 during prostate tumorigenesis in a relevant transgenic mouse model (that mimics human CaP development) and 2) the underlying mechanism. By employing S100A4-positive and S100A4-negative human CaP cells, xenograft models, and transgenic mouse models, we provide evidence that the S100A4 protein, both in its intracellular and extracellular states, plays a tumor-promoting role in the development of CaP by regulating the function of nuclear factor κ B (NFκB)/receptor for advanced glycation end products (RAGE) molecule. Taken together, we show that S100A4 plays a critical role during CaP development and could serve as a molecular target for novel therapies to treat this lethal disease.

Results

S100A4 deletion decreases prostate tumorigenesis in TRAMP/S100A4^+/− mice

To understand the role of S100A4 in the development and progression of CaP, we selected TRAMP and S100A4^−/− mice. TRAMP is an autochthonous mouse model that develops CaP in defined stages and mimics the development of human disease.⁸ In TRAMP mice, expression of the SV40 early genes (T and t antigen, Tag) is driven by the prostate-specific promoter probasin that leads to cell transformation within the prostate.⁸ It has been reported that 100% of male TRAMP mice develop CaP without any chemical or hormonal treatment.⁸ S100A4^−/− mice generated as described earlier⁹ were crossed with TRAMP mice for 5 generations, and mice with genotypes TRAMP/S100A4^+/+, TRAMP/S100A4^+/−, and TRAMP/S100A4^−/− were obtained. However, the generation of heterozygous TRAMP/S100A^+/− mice was significantly higher than the homozygous TRAMP/S100A4^−/− genotype. No lethality and no defect in prostate development were observed in TRAMP/S100A4^+/− mice. In TRAMP, TRAMP/S100A4^+/−, and TRAMP/S100A4^−/− mice, expression of SV40 viral antigens was intact. Knockdown of S100A4 expression did not influence the expression of SV40 viral antigens. Magnetic resonance imaging (MRI) was used to monitor the tumor development and measure the tumor volume at regular intervals of time in TRAMP and TRAMP/S100A4^+/− mice. As expected, 100% of TRAMP mice developed prostatic carcinoma at the age of 32 weeks (Fig. 1A). However, TRAMP/S100A4^+/− mice displayed a significant delay in tumor development (Fig. 1A). As compared to TRAMP, tumor-free survival was significantly (P < 0.001) higher in TRAMP/S100A4^+/− mice, and more than 50% of TRAMP/S100A4^+/− mice remained tumor free up to 35 weeks of age (Fig. 1A). Next, age-matched littermates were compared for tumor volume. This was performed both by employing MRI and sacrificing mice randomly from each group at 28 weeks of age. As is evident from Figure 1B, TRAMP/S100A4^+/− mice exhibited a highly reduced prostatic tumor size than TRAMP/S100A4^+/+ mice. Further, TRAMP/S100A4^+/− mice exhibited a significantly reduced gross weight of the prostate tissue alone and genitourinary (GU) apparatus as compared with TRAMP/S100A4^+/+ age-matched littermates (Fig. 1C). The average volume of tumors in TRAMP/S100A4^+/− mice was significantly (P < 0.001) lower than in TRAMP mice (Fig. 1D). In agreement with our previous findings, 18% to 22% of TRAMP mice exhibited liver metastasis at 28 weeks of age; however, no liver metastasis was observed in TRAMP/S100A4^+/− mice (Fig. 1E). Taken together, these data suggest that S100A4 plays an important role in the growth of prostatic tumors.

Figure 1.

Prostate tumorigenesis in TRAMP and TRAMP/S100A4^+/− mouse models. (A) Line graph represents the percentage of tumor-free TRAMP and TRAMP/S100A4^+/− mice as a function of age. (B) (i, ii) MRI analysis of prostate tumor growth in age-matched transgenic mice (28 weeks of age). (iii, iv) Photomicrographs of the pelvic region showing prostate tumors in age-matched mice. (v, vi) Photomicrographs of the urogenital system showing prostate tumors in age-matched transgenic mice. (C) Histogram showing GU and prostate weights. (D) Histogram showing average tumor volumes. (E) Histogram showing the percentage incidence of liver metastasis. (F) Histogram showing the average life span of TRAMP versus TRAMP/S100A4^+/− mice (FVB and C57B backgrounds). Each bar in the histogram represents the mean ± SE. *P < 0.05. (G) Photomicrographs of immunostaining of S100A4 of prostatic specimens. Arrows indicate staining for S100A4. Magnification, 40×.

Next, the implication of S100A4 deletion on the total life span of TRAMP mice was evaluated. We observed that the heterozygous deletion of S100A4 in TRAMP mice resulted in the prolongation of the life span of TRAMP mice in both types of backgrounds (i.e., FVB and C57B) (Fig. 1F). Importantly, TRAMP/S100A4^+/− mice exhibited a significant increase (65% higher) in life expectancy (P < 0.001) with a median survival of 65 weeks compared with the 42 weeks in TRAMP/S100A4^+/+ mice (Fig. 1F). Next, we determined the expression level of S100A4 by immunohistochemical analysis in prostatic tissues of mice. The immunoperoxidase staining of S100A4 was found to be stronger in TRAMP than TRAMP/S100A4^+/− mice (Fig. 1G). Immunostaining of prostate tissue showed that the S100A4 protein is present in epithelial cells as well as the stromal region of the dorsolateral prostate (Fig. 1G).

Comparative analysis of S100A4-positive and S100A4-null prostatic tumors

Since the heterozygous deletion of S100A4 significantly delayed the development of prostatic tumors in TRAMP mice, we questioned whether S100A4 is a growth-promoting gene. For this purpose, we determined and confirmed the intracellular expression of S100A4 in prostatic epithelial (RWPE1, PC3, Du145) and stromal (WPMY1) cells (Fig. 2A). It is noteworthy that LNCaP cells, which are slow growing in vitro and in vivo, do not express S100A4-negative tumors. We therefore selected LNCaP as a model that represents S100A4-negative tumors. Next, S100A4 was ectopically expressed in S100A4-negative LNCaP cells to generate S100A4-expressing LNCaP (Fig. 2B and 2C). Since both genotypes of cells (S100A4-positive and S100A4-negative LNCaP) are of common lineage, it provided us an opportunity to study the behavior of tumor cells (in terms of growth potential and tumorigenicity). As evident from the [3H]thymidine uptake analysis, S100A4-expressing LNCaP cells exhibited an increased rate of proliferation than S100A4-negative LNCaP cells (Fig. 2D). S100A4-expressing cells formed an increased number of colonies than S100A4-negative control LNCaP cells of the same lineage (Fig. 2E). Further, S100A4 was suppressed in PC3 cells by siRNA, and the proliferation rate was determined. As measured in terms of [3H]thymidine uptake, S100A4 suppression was observed to decrease the rate of proliferation (Fig. 2F), suggesting the regulatory role of S100A4 on cell proliferation. The current data corroborate our observation previously shown in PC3 cells.⁷ Next, stable clones of S100A4-expressing LNCaP and control (S100A4-negative and transfected with vector alone) cells were implanted as xenografts in male athymic mice, and tumor growth was measured weekly. S100A4-negative (LNCaP) tumors grew slowly in the beginning and did show a spike in growth after 4 weeks of implantation. The difference in tumor sizes of S100A4-expressing and S100A4-negative tumors was visually distinguishable and quantitatively significant (Fig. 3A and 3B). S100A4-expressing tumors grew at a rate of 28.57 mm³/d, and S100A4-negative tumors grew at a rate of 12.57 mm³/d (Fig. 3C and 3D). The growth-promoting potential of S100A4 can be ascertained from the observation that 100% of mice implanted with S100A4-expressing tumors reached a volume of 1,000 mm³ by 5 weeks after implantation (Fig. 3D). However, at this time point, S100A4-negative tumors exhibited an average volume of 380 mm³ (Fig. 3D). Next, we determined the rate of proliferation of cells within tumors by using an in vivo proliferation assay. By counting BrdU-positive cells as an indicator of proliferating cells, S100A4-expressing tumors exhibited an increased number of proliferating cells than S100A4-negative tumors (Fig. 3E). Taken together, these data further strengthen the hypothesis that S100A4 is important for the growth of CaP cells.

Figure 2.

Comparative analysis of growth of S100A4-positive and S100A4-negative human tumor cells in vitro. (A) Immunoblot image shows the intracellular S100A4 levels in human prostate stromal cells (WPMYI) and neoplastic epithelial cells (PC3, Du145, LNCaP). Pancreatic epithelial cells (PANC1) were used as a positive control. (B, C) Histograms and immunoblots represent the mRNA and protein expression of S100A4 as determined by real-time PCR and immunoblot analysis in S100A4-negative (LNCaP) and stable S100A4-positive (LNCaP) tumor cells. (A, C) Equal loading of proteins was confirmed by stripping the immunoblot and reprobing it for β-actin. (D, E) Histograms showing differences in the proliferative potential of S100A4-negative (LNCaP) and S100A4-positive (LNCaP) tumor cells as measured by [3H]thymidine uptake and soft-agar colony formation assays. (F) Histograms showing differences in the proliferative potential of PC3 and S100A4-suppressed PC3 cells as measured by a [3H]thymidine uptake assay. Each bar in the histogram (B, D-F) represents the mean ± SE. *P < 0.05. All experiments were repeated 3 times.

Figure 3.

Comparative analysis of growth of S100A4-positive and S100A4-negative human tumor cells in an athymic nude mouse model and quantification of extracellular S100A4 in culture media of CaP cells. (A, B) Representative photomicrographs show the visible differences between the S100A4-negative and S100A4-positive tumors implanted in athymic nude mice. (C) Graphical representation of data showing the growth of S100A4-negative and S100A4-positive tumors implanted in athymic mice. Data are represented as the mean ± SE. *P < 0.05 from the control group. (D) Line graph shows the number of mice reaching the preset end point of tumor volume of 1,000 mm³ in indicated weeks. (E) Histogram showing the proliferation of tumor cells in vivo (in terms of BrdU-positive cells/field). (F) Bands in the image represent the presence of the S100A4 protein in serum-free conditioned culture media of cells as assessed by a Slot-blot immunoreactivity assay. Recombinant S100A4 (dissolved in serum-free media) and PANC1 cells were used as a positive control. (G, H) Histogram showing levels of the S100A4 protein in the serum-free conditioned culture media of cells as measured by ELISA. Each bar in the histogram represents the mean ± SE. *P < 0.05. All experiments were repeated 3 times.

The S100A4 protein is secreted by prostatic cells in media

The presence of S100A4 in stromal as well as epithelial cell regions in prostatic tissues of TRAMP mice (Fig. 1G) prompted us to investigate whether S100A4 is a secretory protein. As evident from a Slot-blot immunoreactivity assay, serum-free media collected from cultures of WPMY1, PC3, and Du145 cells tested positive, while LNCaP cells tested negative for the S100A4 protein (Fig. 3F). We next quantified the levels of S100A4 secreted by prostatic cells by using S100A4-specific ELISA. When compared, stromal cells were observed to secrete more protein than epithelial cells (Fig. 3G). Since most of extracellular S100A4 was contributed by stromal cells, we further investigated whether S100A4 is indeed secreted by epithelial cells. For this purpose, S100A4 was 1) ectopically expressed in PC3 cells and 2) knocked down in another set of PC3 cells. Notably, induction of intracellular S100A4 expression was observed to increase the secretion of S100A4, as was evident from the extracellular S100A4 protein levels found present in the culture media (Fig. 3H). However, culture medium from S100A4-suppressed PC3 cells exhibited decreased S100A4 levels than control cells (Fig. 3H). These data suggest that the increase in intracellular S100A4 levels in CaP cells amounts to its subsequent release into the extracellular space and causes a spike in extracellular or secretory S100A4 protein levels. Preliminary studies from our laboratory show that S100A4 is detectable in the serum of humans, and currently, a clinical study in a cohort of CaP patients is underway in our department (data not shown).

Determining the functional relevance of extracellular S100A4

We next investigated the functional significance of the secretory S100A4 protein. PC3 and LNCaP cells were treated with the recombinant (rh) S100A4 protein, and the rate of proliferation was observed. As measured in terms of [3H]thymidine uptake, rhS100A4 treatment was observed to increase the rate of proliferation of CaP cells (Fig. 4A and 4B), suggesting that extracellular S100A4 possesses a growth-promoting property.

Figure 4.

Determining the functional significance of extracellular S100A4 in vitro and analyzing the role of RAGE for the function of S100A4 during prostate tumorigenesis. (A, B) Histogram showing the rate of [3H]thymidine uptake of PC3 cells and LNCaP cells when treated with the rhS100A4 protein. (C) Histogram showing the effect of extracellular S100A4 on the rate of proliferation (in terms of [3H]thymidine uptake) of cells in the presence of the universal receptor blocker BSA. (D) Immunoblot represents the effect of RAGE knockdown on RAGE expression in LNCaP cells. Knockdown or suppression of RAGE was achieved by transfecting CaP cells with RAGE siRNA. Control cells were transfected with scrambled siRNA. Immunoblots shown are representative images from 3 independent experiments. Equal loading was confirmed by stripping immunoblots and reprobing them for β-actin. (E) Histogram showing the significance of the RAGE receptor for progrowth activity of extracellular S100A4. [3H]thymidine uptake was measured in control and RAGE-suppressed cells treated with either a nonspecific peptide or the rhS100A4 protein. (F) Immunoblot showing levels of the S100A4/RAGE complex in prostatic tissues of TRAMP and TRAMP/S100A4^+/− transgenic mice as assessed by immunoprecipitation (IP). Cyclin E immunoprecipitate from tissues was used to confirm the specificity of the antibody-protein interaction (for IP experiments) and termed as a negative control. β-actin levels in 10% of input were used as a loading control. KO represents TRAMP/S100A4^+/− mice; WB represents Western blotting. (G) Representative photomicrographs showing the co-localization of S100A4 with the RAGE receptor in prostatic tissues of TRAMP and TRAMP/S100A4^+/− mice as assessed by an immunofluorescence assay. Green fluorescence shows S100A4 protein localization, and red fluorescence shows RAGE localization within prostatic tissues. The merged lane shows the co-localization of 2 proteins. DAPI was used as a nuclear staining control. Each bar in the histograms represents the mean ± SE. *P < 0.05. All experiments were repeated 3 times.

Extracellular S100A4 exerts proliferative effects via a membrane receptor in CaP cells

We examined whether growth-promoting effects of the secretory S100A4 protein are mediated through a receptor in tumor cells. LNCaP (S100A4-negative) cells were treated with rhS100A4 for 12 hours in the presence or absence of bovine serum albumin (BSA; widely used as a receptor-blocking agent). Notably, in the presence of BSA, the rhS100A4 protein did not induce the rate of proliferation of CaP cells, suggesting that extracellular S100A4 requires a membrane receptor on cells to induce its growth-promoting activity (Fig. 4C).

Extracellular S100A4 acts through the RAGE receptor

Recent studies showed that S100 family proteins have an affinity towards binding to a surface receptor known as RAGE in various cell types.^10,11 RAGE is an important component of the inflammation-associated pathway and highly expressed in CaP cells.¹² This prompted us to investigate if extracellular S100A4 requires RAGE for its growth-promoting function. RAGE was knocked down in LNCaP cells by siRNA, and cell proliferation was measured after the treatment of cells with rhS100A4 (Fig. 4D). As evident from [3H]thymidine uptake data, the rate of proliferation of RAGE-silenced cells did not change after rhS100A4 treatment, suggesting that extracellular S100A4 requires a surface receptor for its growth-promoting function (Fig. 4E).

S100A4 physically binds with the RAGE receptor in prostatic tissues of transgenic mice

Extracellular S100A4 was observed to require RAGE for its growth-promoting function (Fig. 4E). We next examined whether a physical interaction between S100A4 and RAGE exists in vivo and analyzed prostatic tissues of TRAMP and TRAMP/S100A4^+/− mice for S100A4/RAGE complex localization. Immunoprecipitation assays of prostatic tissues showed that S100A4 and RAGE proteins form a complex (Fig. 4F). Notably, prostatic tissues of TRAMP/S100A4^+/− mice exhibited reduced S100A4/RAGE complex levels (Fig. 4F). We next validated the physical co-localization of S100A4 and RAGE in prostatic tissues by employing an immunofluorescence-based co-localization assay. As shown in Figure 4G, S100A4 (green fluorescence) was found to co-localize with RAGE (red fluorescence) within prostatic tissues (Fig. 4G). The co-localization of 2 proteins was highly reduced in prostatic tissues of TRAMP/S100A4^+/− mice (Fig. 4G).

S100A4 induces the activation of NFκB in CaP cells

RAGE-induced signaling is reported to cross-talk with NFκB signaling in tumor cells.^13,14 CaP cells are reported to exhibit the constitutive activation of RAGE and NFκB.^15,16 S100A4, RAGE, and activated NFκB have been found to be increased at inflammation sites and share several downstream targets.^2,7,15-17 Based on our data and published reports, we questioned whether there is a possibility about the convergence of S100A4 and RAGE-mediated signaling to NFκB in CaP cells. To examine this possibility, we first determined whether S100A4 acts as a regulator of NFκB signaling in CaP cells. We determined NFκB activation and NFκB transcriptional factor activity in nuclear fractions of rhS100A4-treated LNCaP and S100A4-expressing LNCaP cells by employing immunoblotting and ELISA. Activation of NFκB is marked by the phosphorylation of its subunits and their nuclear translocation.¹⁸ rhS100A4 treatment was observed to increase the nuclear p65 subunit levels, suggesting the activation of NFκB signaling within tumor cells (Fig. 5Ai). An increased binding of NFκB to its target DNA is an indicator of its increased transcriptional activity. At 12 hours after rhS100A4 treatment, an increase in NFκB activity was observed in LNCaP (S100A4-negative) cells (Fig. 5Aii). S100A4-negative and S100A4-positive tumor cells were compared for NFκB activity. S100A4-expressing LNCaP cells were observed to exhibit increased NFκB activity than S100A4-negative cells (Fig. 5Bi and 5Bii). Next, we determined if knockdown of intracellular S100A4 has an effect on NFκB activity in S100A4-positive CaP cells. S100A4-suppressed PC3 cells exhibited decreased NFκB nuclear translocation and transcriptional activity (Fig. 5Ci and 5Cii). These data suggest that increased intracellular S100A4 causes the outflux of this protein, which in turn (in extracellular form) increases NFκB activity in CaP cells (Fig. 5B and 5C).

Figure 5.

Determining the role of S100A4 in regulating NFκB activity in prostatic tumor cells and prostatic tissues of TRAMP and TRAMP/S100A4^+/− transgenic mice. (A) (i) Immunoblots and (ii) histogram represent the effect of treatment of the S100A4 protein (rhS100A4) on the levels of nuclear NFκB (p65 subunit) in cells using immunoblots and transcriptional factor activity assays. (B) (i) Immunoblots and (ii) histogram represent the effect of ectopic expression of S100A4 in LNCaP cells on their NFκB activity measured in terms of nuclear levels of the p65-NFκB subunit (active NFκB). (C) (i) Immunoblots and (ii) histogram represent the effect of S100A4 knockdown on NFκB activity of PC3 cells measured in terms of nuclear levels of the p65-NFκB subunit. Knockdown or suppression of S100A4 was achieved by transfecting CaP cells with S100A4 siRNA. Control cells were transfected with scrambled siRNA. (Ai, Bi, Ci) Immunoblots shown are representative images from 3 independent experiments. Equal loading was confirmed by stripping immunoblots and reprobing them for lamin A/C. (Aii, Bii, Cii) Data are presented as the percentage of NFκB activity terms. (D) Photomicrographs show the phosphorylated p65-NFκB subunit levels in S100A4-negative LNCaP tumors and S100A4-expressing LNCaP tumors implanted in athymic mice and (Ei) prostatic tissues of TRAMP and TRAMP/S100A4^+/− mice (28 weeks old) as assessed by immunohistochemical analysis. Arrows point to the immunoreactive regions in tissues. Histograms represent the activity of NFκB in (Eii) prostatic tissues of TRAMP and TRAMP/S100A4^+/− mice (28 weeks old) and (F) RAGE-suppressed LNCaP cells following treatment with the rhS100A4 protein as assessed by a NFκB transcriptional factor activity assay. Data are presented as the percentage of NFκB activity terms. (G) Immunoblot represents NFκB activity in RAGE-suppressed LNCaP cells following treatment with the rhS100A4 protein, showing the significance of RAGE for the function of extracellular S100A4. (H) Pictorial representation of the model of the hypothesis (for the role of S100A4 during prostate tumorigenesis) generated from the current study. Each bar in the histograms represents the mean ± SE. *P < 0.05. All experiments were repeated 3 times.

Comparative status of NFκB in S100A4-positive and S100A4-negative tumors of the same lineage

Wild-type LNCaP cells form S100A4-negative tumor xenografts in mice. We generated stable S100A4-expressing LNCaP clones, which developed S100A4-positive tumors in mice (Fig. 2). This provided us an advantage of having 2 types of tumors of the same lineage. Immunohistochemical analysis of paraffin-embedded sections showed that S100A4-negative tumors exhibit significantly decreased phosphorylated p65-NFκB (activated NFκB) than S100A4-positive tumors (Fig. 5D).

TRAMP/S100A4^+/− transgenic mice exhibit decreased prostatic NFκB activity

We determined the NFκB level and activity in prostatic tissues of TRAMP and TRAMP/S100A4^+/− mice. By employing immunostaining, we observed that prostatic tissues of TRAMP mice (28 weeks of age) exhibit increased staining for the phosphorylated NFκB protein (activated NFκB) (Fig. 5Ei). However, their age-matched TRAMP/S100A4^+/− littermates exhibited decreased levels of activated NFκB (phosphorylated NFκB) in prostatic tissues (Fig. 5Ei). Next, we measured the NFκB transcriptional factor activity in prostatic tissues of TRAMP and their age-matched TRAMP/S100A4^+/− littermates by using an ELISA-based transcriptional activation assay. NFκB activity in prostatic tissues was significantly lower in TRAMP/S100A4^+/− mice than in age-matched TRAMP littermates (Fig. 5Eii). These findings show a positive association of S100A4 and NFκB activity during prostate tumorigenesis.

RAGE is required for S100A4-induced NFκB signaling in CaP cells

Since we observed that RAGE is required for the S100A4-induced proliferation of CaP cells, we next investigated whether RAGE is required for the S100A4-induced activation of NFκB in CaP cells. For this purpose, we measured NFκB activity in RAGE-suppressed LNCaP cells. RAGE-suppressed cells exhibited decreased NFκB activity (Fig. 5F). As expected, rhS100A4 protein treatment increased NFκB activity in LNCaP cells transfected with scrambled siRNA. It is noteworthy that rhS100A4 treatment did not cause any increase in NFκB activity in RAGE-silenced LNCaP cells (Fig. 5F). RAGE-suppressed LNCaP cells exhibited decreased nuclear p65-NFκB levels. Further, rhS100A4-treated cells devoid of RAGE did not exhibit any increase in nuclear NFκB levels, suggesting that RAGE is required for extracellular S100A4-induced NFκB activation in CaP cells (Fig. 5G).

Discussion

Emerging data have implicated S100 family proteins in the pathogenesis of various human disorders including inflammatory diseases, Alzheimer disease, and cancer.^2,19,20 Members of this family have unique individual features with the general property of calcium binding.²¹ Calcium-dependent conformational change is necessary for S100 proteins to interact with their protein targets to generate biological effects.²¹ S100A4 is known by several aliases such as metastatin (Mts1), fibroblast-specific protein (FSP1), p9Ka, 18A2, pEL98, 42A, CAPL, and calvasculin.² The human S100A4 gene is located at position 1q21 on chromosome 1, whereas it is placed at positions 3f3 and 2q34 in mouse and rat chromosomes, respectively.⁹ The S100A4 protein is reported to be capable of both calcium-dependent and calcium-independent interactions with the target molecule.²¹ S100A4 is highly expressed in cells representing the morphogenic transition from an epithelial to mesenchymal phenotype.²² Testing of biopsy samples and tumor specimens of human cancer patients has shown that malignant tissues express higher S100A4 protein levels than adjacent normal tissues.^2,22 S100A4 is reported to integrate pathways to generate a phenotypic response characteristic of cancer metastasis.²² We previously reported that S100A4 induces the invasiveness of CaP cells by regulating MMP-9 activity in tumor cells.⁷ The current study provides compelling evidence that, apart from being a metastatic protein, S100A4 also plays a key oncogenic role in the development of prostate tumors.

It is well established that inflammation has a tumor-promoting effect in tissues and that NFκB is an important player in the inflammation process.^13,14 Both inflammation and NFκB have been implicated for their tumor-promoting effects in CaP development.²³ Based on 1) reports that NFκB is involved in the pathogenesis of CaP diseases and 2) our current observation that S100A4 regulates NFκB in prostatic cells, we speculate that the sustained activation of NFκB within inflammation-laden prostatic tissues could be due to the activation of S100A4. Previously, Luo et al.²⁴ showed that downregulation of the NFκB pathway component (IKKα) slows down CaP growth in TRAMP mice, thus establishing the significance of NFκB in tumorigenesis in CaP development. Our findings carry significance as these showed the upstream regulator of NFκB during prostate tumorigenesis in TRAMP mice. This was corroborated by the observation that prostatic tissues of TRAMP/S100A4^+/− mice exhibited reduced NFκB activity and S100A4-negative human prostatic tumors exhibited reduced NFκB activity than S100A4-positive tumors (Fig. 5). Furthermore, we observed that the effect of S100A4 knockdown on MMP-9 activity (downstream target of NFκB) in CaP cells was similar to the treatment of the NFκB inhibitor (data not shown). The current study in human CaP cells in which NFκB activity was observed to be amenable to modulate change in S100A4 status (Fig. 5B and 5C) further strengthened our hypothesis that S100A4 acts as a regulator of NFκB activity. Based on our observations, we hypothesize that the heterozygous allelic loss of S100A4 resulted in the loss of NFκB activity, which in turn caused a reduction in the inflammatory environment within prostatic tissues and delayed tumor onset in TRAMP/S100A4^+/− mice.

RAGE, a member of the immunoglobulin superfamily, interacts with advanced glycation end products and several ligands.^25,26 The interaction of ligands with RAGE is reported to trigger the activation of MAPK, JAK/STAT, and NFκB.^14,26,27 RAGE, elevated in several cancers including CaP, is reported to play a role in the migration of tumor cells.^26,28 Knockdown of RAGE has been shown to inhibit the growth of CaP cells in vitro and in vivo.^15,26,29 S100A proteins are reported to exhibit an affinity of binding to RAGE.^10,11 The current study showed that the secretory S100A4 protein mediates its progrowth effects on CaP cells through RAGE. Supporting evidence in this direction was the observation of a physical interaction/co-localization of S100A4 and RAGE in prostatic tissues of the TRAMP model. This was evident from immunofluorescence and immunoprecipitation studies of RAGE and S100A4, showing reduced levels of the S100A4/RAGE complex in prostatic tissues of TRAMP/S100A4^+/− mice (Fig. 4E and 4F). The significance of this observation is that this would be the first report showing a S100A4/RAGE interaction in a prostatic mouse model.

Previous studies showed that RAGE induces NFκB.^14,27 Notably, this study provides evidence that the induction of NFκB by RAGE is regulated by S100A4. This is evident from the data that the exclusion of RAGE rendered extracellular S100A4 nonfunctional, with no effect on NFκB in prostatic tumor cells (Fig. 5F and 5G). Taken together, these data established the significance of the S100A4/RAGE/NFκB molecular circuitry in CaP cell growth. Based on our observations, we generated a global hypothesis stating that the heterozygous loss of S100A4 caused a reduced activation of the RAGE/NFκB interaction, which ultimately resulted in the inhibition of tumor development (Fig. 5H). To conclude, we suggest that S100A4 is an important oncoprotein that has a role in the development of CaP as well as its metastasis. This study is important as it identifies S100A4 as a target that could be used to develop new therapeutic agents that would help in inhibiting the progression of prostatic cancer as well as its metastasis, for which there are no therapies available. In this context, testing of small molecule inhibitors of S100A4 is underway in our laboratory. To develop the utility of S100A4 as a therapeutic target, further investigations in animal models are warranted.

Materials and Methods

Generation of S100A4^−/− mice

These were generated by Dr. Noona Ambartsumian (Danish Cancer Society, Copenhagen, Denmark) as described earlier.⁹

Generation of TRAMP/S100A4^+/− mice

Previously described S100A4^−/− and TRAMP mice were intercrossed for 5 generations to generate TRAMP/S100A4^+/− and TRAMP/S100A4^+/+ mice of nearly identical genetic backgrounds. Genotyping for TRAMP mice was performed as described earlier.⁸ Primers used for the S100A4 knockout allele were as follows: forward: 5′-GAG GTC CAT CTC TTA GAG AGT TGG C-3′; reverse: 5′-GCA CAT GTG CGA AGA AGC CAG AGT A-3′. Primers for the S100A4 wild-type allele were as follows: forward S100A4 exon 2: 5′-CTG CCC TTA GGT CTC AAC GGT TAC C-3′; reverse S100A4 exon 2: 5′-CCT CCT CCT GCA GAT GCA TCA CGT G-3′. The resulting bands were as follows: for the S100A4 knockout allele, a 218-bp product, and for the S100A4 wild-type allele, a 350-bp product.

Tumor development measurement

Tumor growth and volume in mice were measured by employing MRI once after every 8 weeks as described earlier.³⁰

Cell lines and plasmids

Human prostatic normal epithelial (RWPE1), CaP (LNCaP, PC3, Du145), and stromal myofibroblast (WPMY1) cells and the pancreatic carcinoma cell line PANC1 obtained from ATCC (Manassas, VA) were cultured in appropriate media under a 5% CO₂ condition. Preparation of the pcDNA3.1-S100A4 plasmid has been described previously.⁷

Chemicals and reagents

G418 was purchased from Invitrogen (Carlsbad, CA). Recombinant S100A4 was commercially purchased (Pro-Spec-Tany TechnoGene, Rehovot, Israel).

Detecting secretory S100A4 in culture medium

Cells (80% confluent level) were allowed to grow in serum-free culture media for 24 hours. At 24 hours, culture media were collected and analyzed for S100A4 levels by using a Slot-blot assay (Whatman, Florham Park, NJ) as per the vendor’s protocol.

Generation of stable cell lines

To generate S100A4-expressing LNCaP clones, cells were stably transfected with pCDNA3.1-S100A4 using Lipofectamine (Invitrogen) as described previously.³¹ Cells were than selected in the presence of G418 (400 µg/mL) starting at 48 hours after transfection. The selection of cells under antibiotics was continued for 4 weeks, and clones were tested for S100A4. The stable clones were maintained in RPMI containing 10% FBS and 300 µg/mL of G418.

RAGE receptor knockdown and its effect on the rate of proliferation of cells

RAGE knockdown was achieved by transfecting cells with RAGE siRNA (final concentration of 200 nM). Control cells were transfected with scrambled siRNA. Twelve hours after transfection, cells were exposed to rhS100A4. Control cells were treated with a nonspecific peptide control. Cells were allowed to grow for 36 hours, the last 16 hours of which in the presence of [3H]thymidine (0.5 µCi/mL). Incorporated [3H]thymidine was quantified as described earlier.^31,32

Measurement of NFκB activation

NFκB activity was measured in nuclear lysates by employing a NFκB transcription factor assay kit (Active Motif, Carlsbad, CA).

RAGE receptor knockdown and its effect on the NFκB transcriptional activity of cells treated with extracellular S100A4 (rhS100A4)

RAGE knockdown was achieved by transfecting cells with RAGE siRNA. Control cells were transfected with scrambled siRNA. At 24 hours after transfection, cells were treated with either a nonspecific peptide or rhS100A4 (2 µg/mL). After 24 hours of incubation with extracellular peptides, NFκB activity was measured in nuclear lysates of cells by using a NFκB transcriptional activity kit and immunoblotting. Equal loading was confirmed by stripping immunoblots and reprobing them for lamin A/C.

Treatment of cells with the rhS100A4 protein

Control cells were treated with a nonspecific peptide control. Cells at an 80% confluent level were incubated in serum-free media for 12 hours. At this time point, cells were added with fresh serum-free media containing rhS100A4. After the incubation of cells for 12 hours, nuclear lysates were prepared.

Investigating the relevance of S100A4 in the growth of tumors in xenograft mouse models

Athymic (nu/nu) male nude mice (6 weeks old) (HarlanTek, Madison, WI) were implanted with S100A4-negative (LNCaP stably transfected with an empty vector) and S100A4-positive (S100A4-expressing LNCaP) cells (3 × 10⁶) in 50 µL of RPMI and 50 µL of Matrigel (BD Biosciences, Bedford, MA) subcutaneously into the right flanks of each mouse. Tumor measurements and body weights were recorded weekly as described earlier.^31-33 Tumors were excised at the 35th day after implantation when 100% of S100A4-negative tumors reached a volume of 1,000 mm³. All procedures conducted were in accordance with Institutional Animal Care and Use Committee (IACUC) guidelines. Two hours before sacrifice, each animal received an intraperitoneal administration of BrdU (10mg/kg; Life Technologies, Grand Island, NY) to label proliferating cells with BrdU for measuring in vivo proliferation.³³

[3H]thymidine incorporation, colony formation, Luciferase activity, transfections, immunoprecipitation, immunohistochemistry, real-time PCR, and immunoblot assays and in vivo proliferation

These were performed as described earlier.^7,32-35

Immunofluorescence using confocal microscopy

This was performed by using the method of Robertson et al.³⁶ The antibodies used were rabbit anti-S100A4 (1:50) (Millipore, Bedford, MA) and mouse anti-RAGE (1:30) (Santa Cruz Biotechnology, Santa Cruz, CA). Secondary Alexa Fluor (Jackson Laboratories, West Grove, PA) and Cy3 (Invitrogen) antibodies were used in a dilution of 1:1,000. Nuclei were stained with DAPI (1:1,000) (Sigma, St. Louis, MO). Image stacks were captured by using a laser-scanning confocal microscope (C1^si Confocal Spectral Imaging System, Nikon Instruments, Melville, NY) and analyzed with EZ-C1 3.9 imaging software (Nikon Instruments).

Statistical analyses

The Student t test for independent analysis was applied to evaluate differences between the treated and untreated cells with respect to the expression of various proteins. A Kaplan-Meier survival analysis with a corresponding log-rank and linear regression analysis was used to measure the rate of mean tumor volume growth as a function of time. A P value of <0.05 was considered to be statistically significant.