Elevation of Stromal-Derived Mediators of Inflammation Promote Prostate Cancer Progression in African-American Men

Despite advances in the last 20 years in the diagnosis and treatment of prostate cancer, the incidence and death rate is still significantly higher for African Americans (AA) compared with European-American (EA) men (1). Cancer health disparity (CHD) represents a major public health concern in the United States. Although socioeconomic factors may be responsible to a certain extent, it is now appreciated that intrinsic differences in genetics and tumor biology make AA men more prone to aggressive prostate cancer. Therefore, understanding the impact of biological variability in prostate cancer is vital to reduce the observed cancer outcome gaps between AA and EA.

Tumor evolution (from local carcinogenesis to distant metastasis) is strongly influenced by the microenvironmental conditions encountered by cancer cells. The microenvironment is viewed as a part of the extended phenotype of tumors that influences the proliferation and migration of cancer cells. The tumor microenvironment (TME) is rich in cellular (inflammatory and immune cells, fibroblasts, endothelial cells, nerves, and others) and noncellular elements (enzymes, growth factors, matrix elements) providing a framework that can modulate cancer cell phenotype. Fibroblasts are the most abundant cell type in connective tissues and in cancer comprise a heterogeneous population of cell types collectively called carcinoma associated fibroblasts or CAF (2–4). Myofibroblasts, or activated fibroblasts expressing α-smooth muscle actin (αSMA) together with nonmyofibroblasts cells present in the TME make the core of the CAF population (5). CAF are known drivers of tumor progression, exerting their effects via paracrine and juxtacrine mechanisms (6).

Microarray studies revealed that immune-inflammatory response (IIR), apoptosis, and focal adhesion were differentially increased in AA versus EA prostate tumors (7). Alterations in the immune-inflammatory cell profile in the TME may increase the aggressive nature of prostate cancer in AA (8). It has been recognized that CAFs contribute significantly to the modulation of the IIR in breast cancer (9). We have shown that chronic activation of proinflammatory cytokines and chemokines in prostate CAF can have profound effects on tumor progression (3). However, the biological effects of fibroblasts in the TME of AA patients has not been studied due to the lack of suitable models. Here, we isolated fibroblasts from human prostate tissues from AA and EA patients with prostate cancer and compared the biological response of several prostate cancer cell lines. Our data suggest that fibroblasts from AA (PrF^−AA) can elicit a significant increase in the proliferation and motility of prostate cancer cells compared with EA (PrF^−EA) in vitro and in vivo. We show that among the proinflammatory factors secreted by PrF^−AA, paracrine activation of TrkB signaling by stromal-derived brain-derived neutrophic factor (BDNF) can increase the tumorigenicity of prostate cancer cells and potentially the progression of tumors in AA patients with prostate cancer.

BPH1 is a nontumorigenic human prostate epithelial cell line isolated from a patient undergoing transurethral resection of the prostate. This cell line was obtained from our stocks (ATCC cell authentication STRA3418) and maintained in 10% serum-based RPMI1640 (10). Lymph node metastatic LNCaP cells and derivative C4-2B were purchased from the Characterized Cell Line Core Facility (MD Anderson Cancer Center, University of Texas, Houston, TX) and cultured in T-Media with 10% FBS and Supplements. Lumbar bone metastatic PC3 cells were purchased from the ATCC (CRL-1435) and cultured in DMEM+10% FBS.

Nontumorigenic E006AA from a 50-year-old black male with localized prostate cancer obtained from EMD Millipore (Billerica; SCC102) and its tumorigenic derivative E006AA-ht (ATCC CRL-3277) were cultured in 10% DMEM high glucose (EMD Millipore SLM-120B). Bone metastatic MDA-PCa2b from a 63-year-old black male (ATCC CRL-2422) was cultured in 20% FBS F12-K medium.

Upon written informed patient consent and local ethical committee approval (Institutional Review Board–approved collection protocol) deidentified human prostatic tissue samples were obtained from patients with prostate cancer undergoing Robotic-assisted laparoscopic prostatectomy (RALP) through the University of Chicago Human Tissue Resource Core facility. Tissue was cut into small pieces (1 mm³) and digested overnight at 37°C in collagenase solution [0.28% collagenase 1 (Sigma-Aldrich), 1%DNase I (Sigma), 10% FCS (Life Technologies- Invitrogen), in RPMI1640]. Fibroblast (PrF) cultures were established by plating in RPMI-1640 plus 10% FCS and penicillin/streptomycin (11, 12). To determine the absence of contaminating epithelial cells, samples were screened for the lack of wide-spectrum cytokeratin (WSCK) and expression of vimentin (Sigma) and smooth muscle actin (Sigma) to confirm their fibroblastic nature. To ensure proper biological function, PrFs were used at low passage (<10; ref. 13).

Cell lines were validated by vendors and Mycoplasma screening was routinely performed using the MycoAlert Kit (Lonza Inc).

PrF were cultured to 90% confluence in complete medium and then replaced with 0.1%BSA-RPMI medium and conditioned (PrF-CM) for 48 hours before applied to the HXL Human Cytokine Antibody Array Kit (ARY022B, R&D Systems Inc.) following the manufacturer's protocol. Images were taken using the Bio-Rad MP Analyzer. Densitometric analysis was performed using the Protein Array Analyzer plug in from ImageJ (NIH, Bethesda, MD). Up- (≥1.5-fold) or downregulation (≤0.5-fold) in cytokine secretion were considered significant (P < 0.01) in proteins showing a signal density value >200 pixel.

A total of 250k PrF were mixed with 100k E006AA-ht cells in collagen. After overnight incubation at 37°C, recombinants were grafted under the renal capsule of 10-weeks-old intact male CB17Icr/Hsd-SCID mice (Harlan) supplemented with 5-mg testosterone pellets placed in the subcutaneous compartment. Mice were sacrificed after 8 weeks, kidneys removed, and grafts were imaged and cut into halves before processing for histology. Area occupied by the xenografts was measured using ImageJ (3). Animal studies were approved by the NorthShore Institutional Animal Care and Use Committee.

Gene expression studies have suggested differences in the TME of AA compared with EA (7). To identify cellular drivers of tumorigenesis in the TME, we performed IHC on cohorts of AA (n = 10) and EA (n = 10) patients with prostate cancer (Supplementary Table S1) obtained from the NorthShore Urobank Tissue Repository and focused our attention on three major cellular components:

Histologic examination of tissue samples revealed intense desmoplastic changes in the TME of both AA and EA patients. Assessment of collagen deposition by Masson Trichrome staining and picosirius red staining showed a significant increase in collagen deposition in AA patients (Fig. 1A–B; Supplementary Fig. S1A and S1B). In addition, higher expression of tenascin C (a potential stromal marker of prostate cancer progression; ref. 14) and coexpression of αSMA and vimentin (marker of myofibroblast conversion; ref. 5), were observed in AA compared with EA (Fig. 1C–F; Supplementary Fig. S1B).

Hematoxylin and eosin analysis showed larger areas of inflammatory infiltrates in AA compared with EA (Fig. 1G–H; Supplementary Fig. S1B). IHC analysis of leukocyte distribution in the tumor microenvironment is shown in Fig. 2A. Fujii and colleagues (15) showed increased infiltration of lymphocytes and macrophages in prostate cancer regions compared with benign areas. Therefore, we focused our attention on the CD3⁺ T cells, CD68⁺ (M1), and CD163⁺ (M2) macrophages infiltrates. The infiltration of CD3⁺ T cells was 3.5-fold higher in AA. In addition, both CD68 and CD163 macrophages showed a 2.0- and 2.6-fold increase in AA compared with EA (Fig. 2B, left). Interestingly the CD68/CD163 ratio in AA patients was slightly, but not significantly lower (1.5) compared with the EA cohort (1.9; Fig. 2B, right).

The most common method for semiquantitative evaluation of neoangiogenesis is microvessel density (MVD) using endothelial markers such as CD31, CD34, CD105, and von Willebrand factor (vWF; refs. 16, 17). Using a CD31 antibody and found an increased MVD in AA compared with EA patients (Supplementary Fig. S1A, c–d and S1B). Altogether, these observations suggest that alterations in the TME seen in prostate cancer are more pronounced in AA compared with EA. Amplified stromal remodeling in the TME could positively influence the progression of prostate cancer in AA patients.

To better understand the role of TME stromal cells on prostate cancer-HD biology, we isolated prostate fibroblasts (PrF) from prostate cancer specimens of 5 AA (PrF^−AA) and 5 EA (PrF^−EA) patients with similar clinicopathologic characteristics (Supplementary Table S2; ref. 4).

Myofibroblasts exhibit properties intermediate between fibroblasts and smooth muscle cells and coexpress a set of markers, including the intermediate filament vimentin, αSMA, and fibroblast activated protein 1 (FAP1; refs. 2, 6). To determine whether fibroblasts from AA display a different profile of myofibroblast markers compared with EA-derived cells we quantified the expression of several proteins associated with the activation of stromal cells using Western blot analysis. Expression of αSMA, vimentin, and FAP1 were significantly elevated in PrF^−AA compared with PrF^−EA (Fig. 3A). Caveolin1 (CAV1) expression was significantly lower in PrF^−AA. Although CAV1 expression in cancer cells increases with progression, its loss in prostate cancer stroma correlates with reduced relapse-free survival, suggesting a role in tumor progression (18). Expression of the mesenchymal stem cell marker CD90 was significantly higher in PrF^−AA compared with PrF^−EA.

The tumor stroma is composed of a myriad of cell populations, including the accumulation of nonproliferative myofibroblast and proliferative CD34/Vimentin fibroblasts and other cells with mesenchymal stem cells properties (4, 19, 20). PrF^−AA showed a slight (not statistically significant) increase in basal proliferation rate compared with PrF^−EA (Supplementary Fig. S2A). Because androgens exert mitogenic effects on prostate fibroblasts in vitro (21), PrF were cultured under hormone-deprived conditions (CSS), and exposed to 10 nmol/L testosterone (T) and their proliferation measured after 5 days. Compared with the CSS-control group, T-proliferative response was significantly higher in PrF^−AA compared with PrF^−EA (Fig. 3B). Higher AR expression in PrF^−AA (Fig. 3A) may account for this response (Fig. 3B).

Several mitogens have been shown to have a positive effect on prostate fibroblasts proliferation, these include FGF2, TGFß and PDGF (22). After 6 days of exposure, FGF2 significantly increased the proliferation of PrF^−AA compared with PrF^−EA. PDGF did not significantly affect the proliferation of PrF. In contrast, TGFß1 addition had a slight negative effect on proliferation in PrF cell lines (Fig. 3C).

These data showed that prostate fibroblasts from AA patients have an increased myofibroblastic component and suggest the presence of a population of mesenchymal-like cells (CD90⁺) that may be more responsive to stimulation by several growth factors compared with EA-derived cells.

After activation, myofibroblasts secrete a plethora of soluble factors that promote the proliferation of cancer cells (23). To determine whether the increased myofibroblast activation seen in PrF^−AA has a differential effect on prostate cancer cell proliferation compared with PrF^−EA, conditioned media (CM) from PrF was collected. Prostate cancer cell lines representing both EA and AA groups were exposed to CM (Fig. 4A). Four prostate cancer cell lines from EA (LNCaP, C4-2B, PC3, and 22RV1) and three AA prostate cancer cell lines (E006AA-Par, E006AA-ht, and MDA-PCa2b) with progressive tumorigenic potential and one nontumorigenic but genetically initiated line (BPH1) were used. Compared with control medium (0.5% serum) all prostate cell lines showed increased proliferation in the presence of PrF^−CM (Supplementary Fig. S2B). BPH1 response to either AA or EA PrF-CM was not significantly different. Growth response to PrF^−AA-CM was significantly higher compared with PrF^−EA-CM in all prostate cancer cells. Both LNCaP and C4-2B cell lines showed the highest differential proliferative response when comparing PrF^−AA and PrF^−EA effects. These two cell lines are known for their particular increased tumorigenic response in the presence of stromal cells in vitro and in vivo (24, 25).

Reciprocal interactions between cancer cells and myofibroblasts through direct juxtacrine and/or paracrine signals regulate invasion and metastasis and are reflected in poor clinical prognosis (26–28). To determine whether PrF^−AA and PrF^−EA differ in their promotility effects, prostate cancer cells were exposed to CM from PrF in a wound-healing assay for 24 hours. Prostate cancer cells consistently close the wound faster in the presence of PrF^−AA-CM compared with PrF^−EA-CM (Fig. 4B). Although the BPH1 proliferative response was not different (Fig. 4A), PrF^−AA-CM significantly enhanced its motility compared with PrF^−EA-CM. Overall, these data suggest that PrF in the TME of AA produce a factor(s) acting in a paracrine fashion to differentially enhance the tumorigenicity (proliferation and motility) of prostate cancer cells compared with EA fibroblasts.

Prostate cancer cells show enhanced in vivo tumorigenicity in the presence of stromal cells (24, 25). To determine whether the effects observed in vitro are translated to protumorigenic effects in vivo, we recombined E006AA-ht cells with PrF^−AA and PrF^−EA and grafted under the kidney capsule of immunocompromised mice. The kidney capsule model offers two major advantages for in vivo studies; (i) high take rate (in our study 100% for all groups) and (ii) the ability to quantify local invasion into the kidney parenchyma. Four different groups were compared. E006AA-ht cells alone formed invasive tumors (Fig. 5A, a). E006AA-ht cells are mildly tumorigenic in the subcutaneous xenograft model (29). Xenografts composed of E006AA-ht + PrF^−EA or E006AA-ht + PrF^−AA formed invasive tumors significantly larger than E006AA-ht cells alone (Fig. 5A, b). Tumor size was not only significantly larger in E006AA-ht + PrF^−AA compared with the other groups, but almost completely destroyed the renal parenchyma (Fig. 5A, c). Histologically, these tumors were composed of poorly differentiated and highly invasive cancer cells,(Fig. 5B; ref. 29). PrF grafted in isolation did not grow (Fig. 5B, d). However, addition of PrF enhanced the invasiveness of E006AA cells. Compared with PrF^−EA (Fig. 5B, b), PrF^−AA (Fig. 5B, c) revealed sheets and nests of large, pleomorphic tumor cells invading deeply into the renal parenchyma with foci of necrosis. Mitotic figures, including atypical mitoses were abundant. The remodeling of the surrounding tumor matrix showed a high density of capillary-lined vessels (increased angiogenesis) and leakage of red blood into the adjacent fibroblastic-rich stroma. Large pockets of inflammation were observed within these tumors. All of these changes mimic some of the phenotypic changes observed in the stroma of our AA cohort of patients with prostate cancer (Fig. 1), suggesting that this model could be useful to study the effects of PrF in CHD.

Transcriptome analysis by RNASeq (see Supplementary Materials) of PrF fibroblasts from AA and EA patients using RNAseq revealed significant differences. The expression patterns of PrF^−AA and PrF^−EA were clearly separated on the basis of principal component analysis, with 89 genes differentially expressed (fold-change ≥1.5 and P ≤ 0.05) between the 2 groups (Fig. 6A and B). Functional enrichment annotation tools showed statistical association of differentially expressed genes with the immune system regulation and cytokine signaling GO biological processes, among others (Fig. 6C and D; Supplementary Table S3). Gene expression of a small panel of genes was validated by qPCR (Fig. 6E). Upregulation of cytokines in PrF suggest the presence of secreted factors acting on a paracrine manner that could be responsible of the proliferative/invasive properties. Therefore, we assessed the levels of secreted proinflammatory cytokines from eight PrF^−AA (n = 4) and PrF^−EA (n = 4) using a cytokine array (Fig. 6). Out of 102 cytokines quantified, a total of 7 cytokines in PrF^−AA were at least 1.5-fold (significantly) higher compared with PrF^−EA (Fig. 6F). Only four cytokines (CXCL1, HGF, and Pentraxin-3) were reduced by 1.2-fold in PrF^−AA compared with PrF^−EA (Supplementary Fig. S3). Because the putative tumor stromal marker FAP (a paralog of DPPIV) showed high differential secretion in PrF^−AA (Fig. 3E and F), we decided to stain prostate cancer tissues from AA and EA to determine the expression pattern. Similar to the Western blot results, FAP staining was more abundant in the stroma of AA compared with EA patients with prostate cancer (Supplementary Fig. S4A and S4B). These results suggest that a profile of secreted cytokines by PrF^−AA compared with PrF^−EA may be responsible for the paracrine effects observed in vitro and in vivo.

Increased expression of the neurotropin BDNF and its receptor TrkB have been observed in prostate cancer tissues (30, 31). Assessment of TrkB receptor expression showed significant higher levels in AA (E006AA and E006AA-ht) compared with EA (LNCaP and C4-2B) cells (Fig. 7A). Similar results were observed when using a pan Trk and TrkA-specific antibodies. Next, to determine whether higher expression of the TrkB can influence BDNF response, we evaluated the proliferation and motility of prostate cancer cell lines in the presence of the ligand. We cultured prostate cancer cells for 6 days in the presence of 50 ng/mL and assessed the proliferation by crystal violet staining. BDNF enhanced the proliferation of all cell lines with the highest response observed in E006AA and E006AA-ht cells (Fig. 7B). Next, to evaluate whether targeting the BDNF/TrkB axis can affect the proliferative paracrine effects exerted by PrF^−AA, we treated prostate cancer cells with the TrkB-specific antagonist ANA-12 for 6 days in the presence of PrF-CM. Addition of ANA-12 significantly inhibited the proliferation of prostate cancer cells. Compared with PrF^−EA-CM, PrF^−AA-CM-induced proliferation was significantly more inhibited in E006AA cell lines than LNCaP and C4-2B cells (Fig. 7C).

Neurotropins have shown to modulate the migratory and invasive capabilities of prostate cancer cells (31). To determine whether BDNF have disparate effects on prostate cancer cells from EA and AA, we exposed cells to 50 ng/mL BDNF and their motility evaluated by WHA assay. Compared with control (0.5% serum), BDNF was able to increase the motility of all cell lines. Next, we evaluated whether addition of ANA-12 can impact the paracrine effects induced by PrF-CM (Fig. 4B). As shown in Fig. 7D, ANA-12 significantly inhibited PrF-induced motility of prostate cancer cells. Compared with PrF^−EA, inhibition of PrF^−AA-induced motility was significantly higher. The response of AA representative cell lines was more pronounced than EA cells. These results suggest that activation of the BDNF/TrkB axis has a significant contribution in the PrF paracrine effects on prostate cancer proliferation and motility. These effects are more prominent on cell lines with higher TrkB receptor expression.

TrkB activation by BDNF could mediate tumor progression by activating PI3K/AKT pathway (32). We assessed activation of the Akt pathway by BDNF in the EA and AA cell lines. As shown in Fig. 7E, BDNF increased Akt phosphorylation in all cell lines compared with basal levels. Exposure to PrF^−EA-CM showed disparate results with some cells (BPH1, C4-2B, and E006AA) displaying a higher response. However, PrF^−AA-CM effects on Akt phosphorylation was significantly higher than PrF^−EA-CM in AA prostate cancer cell lines. P-Akt levels in EA prostate cancer cells were similar under both PrF^−EA and PrF^−AA CM. Overall, these results suggest that activation of the BDNF/TrkB axis by tumor stromal cells may play a role in the progression of AA patients with prostate cancer by increasing the tumorigenicity of cancer cells through activating the PI3/AKT pathway.

Many studies are now focused on the impact and significance of the TME on tumor biology. However, few studies have assessed how TME might differ between men of different race. Although direct evidence of causality is lacking, differences from pathway-level and disease association analyses of tumor gene expression data show significant elevation of processes related to immune-inflammatory response and cytokine signaling in AA patients (7). Fibroblasts present in TME actively communicate with cancer and immune cells and participate in the growth and aggressiveness of the tumor by secreting chemokines, cytokines, growth factors, and other inflammatory mediators (6, 33). This multifaceted communication is crucial to provide the appropriate microenvironment for tumorigenesis, angiogenesis, and metastasis. However, direct evidence of racial disparities in the biological function of fibroblasts has not been previously determined. In this study, we isolated prostate fibroblasts from AA and EA patients with prostate cancer and characterized their behavior in vitro and in vivo. We found significant differences in the prostate tumor stroma of AA versus EA patients. Increased collagen deposition and myofibroblasts, evidence of the presence of reactive stroma were observed in TME of AA. The prognostic significance of reactive stroma has been well documented in organ-confined prostate cancer (34), but, to our knowledge, was not previously assessed in CHD. A proper histologic assessment of the reactive stromal grading system have been recently proposed (35). Although we included a small number of patients, our data strongly suggest that a comprehensive study with a larger cohort of patients is necessary to determine the role of reactive stroma in the prognosis of prostate cancer in AA patients. In addition, we also noticed increased IHC expression of the extracellular matrix (ECM) protein; tenascin-C in AA patients. High tenascin-C found in CAF have been reported to be associated with poor prognosis in prostate cancer (14). Interestingly, increased tenascin-C is found in keloid disease that affects dark-skinned patients and has been described for its inability to restrain the wound-healing process similar to reactive stroma in the TME (36). Increased matrix deposition is known to favor tumor growth and invasion; therefore, better understating of the role of ECM proteins such as tenascin-C in CHD is needed.

In addition to our observed changes in the ECM, very few studies have been performed to investigate the significance of inflammation, with some strong evidence suggesting that there is disparate inflammatory TME in prostate cancer of AA versus EA (37, 38). Our analysis of tumor samples from localized prostate cancer revealed increased chronic inflammatory infiltrates, mainly CD3⁺ T cells and CD68/CD163 macrophages. Our data not only confirm previous observations but also highlights the relevance to study the mechanisms leading to the recruitment as well as the role of these chronic inflammatory infiltrates in prostate cancer of different ethnicities.

The proliferation of prostate fibroblasts isolated from AA patients did not differ significantly to those from EA, however addition of testosterone and growth factors FGF2 and PDGF significantly increased the growth of PrF^−AA compared with PrF^−EA. Testosterone has been shown to induce mitogenic effects in prostate fibroblasts (21). This androgenic stromal response by fibroblast from AA has not been previously reported. It is well known that within tumor fibroblasts there is a population of mesenchymal cells that have a positive proliferative response to FGF2 and PDGF (39). We observed higher proliferation response to these two growth factors in fibroblast from AA patients, suggesting the presence of a subpopulation of undifferentiated mesenchymal-like fibroblasts with high proliferative response. Myofibroblasts are a key hallmark of cancer stroma and the cell population responsible for the CAF protumorigenic effects. FAP is a well-defined CAF marker and because of its multiple roles in neoangiogenesis, invasion, and metastasis, is currently evaluated for its potential use for cancer immunotherapy (40). We found increased expression of FAP in PrF^−AA compared with PrF^−EA, suggesting a potential therapeutic utility for AA patients.

CAF promote tumor cell malignancy through stromal–epithelial interaction by secretion of a wide range of soluble growth factors that function via paracrine or juxtacrine modulation of receptors on adjacent tumor cells (3, 6, 13). We observed a significant increased proliferation and migration of prostate cancer cell lines exposed to media conditioned by PrF^−AA compared with PrF^−EA. In addition to the in vitro differences, we developed an in vivo system to study the effects of PrF^−AA on prostate epithelial cells using a mildly tumorigenic prostate cancer cell line (E006AA-ht) from an AA patient. Identification of CAF with protumorigenic properties is performed by tissue recombination and renal xenografting model using a nontumorigenic cell line (BPH1) from an EA patient (41). Compared with PrF^−EA, fibroblast from AA patients significantly increased the tumorigenicity of E006AA-ht cells. Tumors showed increased angiogenesis, inflammatory infiltrates, and proliferation compared with PrF^−EA, suggesting that the nature of the stroma is critical for the fate of cancer cells. Because of their increased tumorigenic response in vitro by PrF^−AA, the E006AA and the MDA-PCa2b could potentially be used in the tissue recombination model. We are currently assessing the feasibility of these cell lines in combination with different types of stromal cell lines. This model can be used to test potential targets in stromal cells.

We have shown previously that blocking the paracrine effects of cytokines and chemokines can have profound effects on tumor growth (3, 13). To identify potential mediators, we screened a panel of secreted factors associated with proinflammatory and protumorigenic properties. As shown in Fig. 6A, PrF^−AA display a particular profile of cytokines compared with PrF^−EA. Notably growth factors VEGF, FGF7, and BDNF have been associated in prostate cancer progression (42–44).

Racial differences in the activation of tyrosine kinase has not been previously addressed. In vitro, we observed increased proliferation and motility in AA prostate cancer cell lines in the presence of BDNF ligand. We also show that paracrine effects exerted by PrF could be significantly impaired in the presence of a TrkB-specific antagonist ANA-12. Downstream effects on Akt activation could be implicated in these biological effects. Further in vivo studies are required to determine the contribution of TrkB signaling activation under BDNF and utility of targeting this pathway to prevent tumor progression in AA patients with prostate cancer. Overall our preliminary observations suggest that the BDNF/TrkB axis could potentially be involved in the speed of progression in prostate cancer tumors in AA patients.

Apart from their direct effects on prostate cancer cells these factors may influence the recruitment of inflammatory infiltrates increasing the progression of prostate cancer tumors in AA patients.

In summary, our study indicates ethnical differences in the tumor stroma of patients with prostate cancer. Isolation of human fibroblasts from AA and EA patients revealed molecular and biological differences. Our data show that PrF from AA contains a particular profile of factors different from the ones present in EA patients. Further studies are needed to determine the role of each of these factors on prostate cancer tumor biology in AA. Better understanding of the underlying mechanisms responsible for the induction of these factors as well as the biological consequences on prostate cancer cells are needed to identify targets with therapeutic potential benefit.

No potential conflicts of interest were disclosed.

Conception and design: S.L. Zheng, J. Xu, O.E. Franco

Development of methodology: O.E. Franco

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Gillard, R. Javier, B.L. Pierce, O.E. Franco

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Gillard, Y. Ji, S.E. Crawford, D.J.V. Griend, O.E. Franco

Writing, review, and/or revision of the manuscript: M. Gillard, J. Xu, S.E. Crawford, O.E. Franco

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Ji, S.L. Zheng, D.J.V. Griend, O.E. Franco

Study supervision: O.E. Franco

Other (manuscript preparation and discussions): C.B. Brendler

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.