Immune Infiltration and Prostate Cancer

1Prostate cancer (PCa) is the most common non-cutaneous cancer and second leading cancer-related cause of death of men in the United States.. Localized PCa can be controlled with surgery or radiation over 80% of the time. If prostate tumors are metastatic at presentation, neither surgery nor radiation will confer significant clinical benefit and androgen deprivation therapy (ADT) represents the sole therapeutic option for such patients. Unfortunately, resistance to hormonal targeting develops in most patients within 2 years, a state that is termed metastatic castration-resistant PCa (mCRPC) and is virtually untreatable. Until recently, Docetaxel was the lone accepted remedy for mCRPC and was associated with a 2- to 3-month enhancement of survival. The advent of newer chemotherapeutic agents such as cabazitaxel and approval of next generation drugs abiraterone and enzalutamide have improved upon this effect, yet resistance ultimately ensues, possibly via mutations in the androgen receptor (AR) gene.

2Relative to other forms of cancer, the discovery of valuable predictive markers of disease progression has been particularly challenging for PCa. Genetic markers capable of identifying clinically aggressive disease remain elusive. In addition, PCa is unique among solid tumors in that imaging techniques are insufficient for definitive identification and diagnosis of malignancy. Instead, the pre-biopsy serum prostate specific antigen (PSA) value is used as an indicator of potential underlying disease, but the utility of this screening method has been intensely scrutinized. In fact, the United States Preventative Services Task Force (USPSTF) recently recommended against routine PSA screening. Importantly, the ramifications of this ruling may be an increase in the proportion of men diagnosed at later stages of PCa. Thus, the identification of novel therapeutic targets for and biomarkers of mCRPC is severely needed.

3A multitude of studies investigating alternative PCa treatment interventions or predictive markers have focused on components of the inflammatory response, with promising results. A relationship between inflammation and cancer was first proposed by Virchow in 1863 and chronic inflammation is now well accepted as a driving force behind many malignancies, such as gastric, colon, esophageal and lung cancers, and hepatocellular carcinoma. While the role of the immune system in PCa is less clear, it has been implicated both as a possible etiologic agent and a tumor promoter. A recent report from the REDUCE study suggesting that aspirin and/or NSAID use decreased risk of PCa in men with a negative biopsy supports this notion. This review will explore the evidence regarding immune infiltration as a potential mediator and indicator of aggressive PCa, as well as a target for PCa therapeutics. Because this topic has been comprehensively evaluated elsewhere, only recent literature will be addressed here.

4As with many other cancers, PCa arises from accumulation of acquired genetic and epigenetic alterations. Currently, age, ethnicity, and family history are the only known risk factors for development of prostate neoplasia. These genetic traits may partially explain the geographical variance in PCa incidence and mortality, with rates in Southeast and East Asia being much lower than those in the United States and Western Europe. By contrast, the elevated rate of PCa observed in Chinese and Japanese men within one generation of relocation to the West indicates that environmental factors, particularly lifestyle factors such as diet, also contribute to the emergence of PCa. One likely link between environmental stimuli and PCa occurrence is chronic inflammation. Support for this hypothesis comes from reports of focal atrophic lesions in the prostate containing a relatively high percentage of proliferating epithelial cells and acute or chronic inflammatory infiltrate. These areas, designated proliferative inflammatory atrophy (PIA), comprise an extensive portion of the peripheral zone of the prostate, the primary site where PCa originates, and have been suggested to be precursors of prostatic intraepithelial neoplastic (PIN) lesions, although one recent retrospective epidemiological study refuted this concept. Additional evidence comes from a recent meta-analysis assessing 20 case–control studies conducted between 1990 and 2012 that revealed a significant positive relationship between self-reported prostatitis and PCa. However, the authors were unable to demonstrate a causal link between the two due to the absence of cohort studies in this analysis. Numerous other epidemiological investigations have examined the connection between prostatitis and subsequent PCa risk and while early studies described an elevated risk of malignancy or recurrence due to presence of inflammation, subsequent studies have produced conflicting results. In addition to differences in study populations and potential selection and other forms of bias, plausible explanations for these inconsistencies include tissue sample size and composition, definition and measurement of inflammation, tissue regions assessed, and categories of prostatitis examined. In order to obtain a more accurate reflection of the relationship between prostatitis and PCa, improved standardization of these measurements, incorporation of asymptomatic inflammatory prostatitis, and a prospective design in future epidemiological studies are needed. Importantly, many investigations defined inflammation with discrete variables, such as present or absent or severe, moderate or mild rather than with a detailed analysis of the cellular composition and distribution of the infiltrate. Although this classification system may be obligatory for certain statistical analyses, it is a major limitation of these investigations and may be masking true associations that exist. Furthermore, these studies may benefit from experimental research designed to provide insight regarding the kinetics, distribution, type, and components of inflammation that are important for prostate tumorigenesis.

5Despite the abundance of data evaluating the connection between prostatitis and PCa, the sources of prostatic inflammation have yet to be fully elucidated. Proposed causes can be divided into three broad categories: (1) microorganism induced, including E. coli, Propionibacterium acnes, and those associated with urine reflux and sexually transmitted diseases, (2) non-microorganism induced, such as uric acid crystals formed during urine reflux, estrogen, dietary factors, especially heterocyclic amines (HCA) 2-amino-1-methyl-6-phenylimidazo[4,5-b] pyridine (PhIP) generated from cooking meats at high temperatures, obesity, and physical trauma such as corpora amylacea, and (3) treatment- or tumor-elicited inflammation.

6Some of these hypotheses were tested using mouse models. For example, Shinohara et al. reported that intraurethral inoculation of mice with a prostatectomy-derived strain of P. acnes elicited prolonged and lobe-specific severe acute and chronic inflammation that was associated with increased proliferation and reduced Nkx3.1 and AR expression. Inflammatory infiltrate consisted of luminal neutrophils and stromal mononuclear cells. Similar observations were noted in experiments using E. coli. Introduction of this bacterium into mouse urogenital tracts provoked a dramatic decline in Nkx3.1 and AR expression and a corresponding enhancement of proliferation as well as the basal cell marker p63. These changes were pursuant to an early influx of neutrophils into prostatic lumen and stroma followed by a surge in stromal mononuclear cells and later a marked increase in plasma cells. Upregulation of pro-inflammatory genes, including IL-6 and COX-2, was also detected using a comparable system. The effects of these events on PCa pathogenesis were examined by combining bacteria CP9-induced prostatitis and a mouse PCa model and revealed that tissue damage resulting from acute prostatitis facilitates accumulation of stromal α-smooth muscle actin-positive myofibroblasts and the conversion of basal cells into luminal cells, which accelerates PCa progression.

7Non-bacterial inflammatory stimuli were also evaluated with respect to PCa development. A novel model of prostatitis involving ubiquitous overexpression of aromatase (AROM+) in mice exhibited early and persistent prostatic mast cell infiltration with subsequent accumulation of neutrophils, T lymphocytes, and macrophages, as well as increased expression of numerous chemokines. This chronic inflammation preceded formation of PIN lesions containing and surrounded by immune cells. Finally, Birbach et al. expressed a constitutively active form of the inflammatory mediator IκB kinase 2 in prostate epithelium of mice harboring a heterozygous Pten deletion and detected a robust inflammatory response, primarily encompassing neutrophils and macrophages in the prostate, that was accompanied by increased tumor size due to enhanced epithelial and stromal proliferation, generation of cribiform structures and stromal remodeling, but no invasion or metastasis.

8Together, these data corroborate the ability of inflammatory stimuli to generate an immune response that modulates the local environment into one that is conducive to malignant transformation and progression and implicate chronic inflammation as a causative factor in PCa. More studies combining prostatitis and murine PCa models are needed to confirm these effects and determine the mechanisms by which these agents promote disease. Of course, these results also must be confirmed in humans.

9As previously described, epidemiological studies exploring the relationship between prostatitis and PCa development may be limited by the absence of characterization of cellular phenotypes displayed by tumor-infiltrating leukocytes (TIL). However, many studies assessing the association between specific immune cells and PCa are now available and those conducted between 2008 and 2014 are summarized according to cell type in Table 1. Most of these investigated CD3+, CD4+, or CD8+ cells in prostatectomy or biopsy tissue of PCa patients and found such cells to be pro-tumorigenic. For example, a cross-sectional study of lymphocyte infiltration in tumor tissue microarrays from patients with PSA recurrence reported that extremely high or low CD3+ cell counts were correlated with reduced PSA recurrence-free survival (RFS). Importantly, CD3 was the only T-cell marker included in this study and no evaluations of T-cell subsets were performed. Thus, the variation in clinical outcomes across CD3+ cell quartiles may at least partially reflect differences in T-cell populations. In fact, the authors mentioned unpublished data indicating that higher numbers of regulatory T-cells (Treg) were associated with more advanced tumor stage and PSA-RFS. A likely assumption is that the CD3+ cells whose low numbers conferred poor prognoses were protective in nature. Thus, it is critical to determine not only the cell type but also the subtype associated with different outcomes in order to better direct therapeutic efforts. By contrast, the spontaneous TRAMP mouse PCa model was used to illustrate that the trafficking of a naturally arising clonally expanded PCa histone H4-specific population of CD44+CD8+ T-cells to prostate tissue lessened tumor burden but failed to improve survival. Several possible explanations for the lack of survival impact exist. First, it may be that the anti-tumorigenic CTL were initially successful in eliminating tumor cells but were eventually rendered inactive by a suppressive tumor microenvironment. Alternatively, it is conceivable that the antigen-specific destruction of H4-expressing tumor cells selected for less immunogenic and possibly more adept H4-negative tumor cells, or instead, that this response exerted pressure on cancer cells to downregulate MHC or undergo immunoediting of tumor antigens and thereby escape immune recognition and destruction. In line with the former hypothesis, several investigations demonstrated that tumor-infiltrating CTL expressed PD-1, one of several indicators of T-cell exhaustion, in PCa patients. This is an essential consideration for immunotherapeutic efforts and emphasizes the importance of evaluating both the cellular subtype and functional status when assessing TIL.

10Another cell type that has received much attention and shown relatively consistent results in the context of PCa progression is the tumor-associated macrophage (TAM). All studies analyzed demonstrated a pro-tumorigenic effect of TAM, although one report specified that this was true only for M2 type macrophages. However, only a fraction of the literature displayed a link between TAM and recurrence. In particular, Nonomura et al. determined TAM count to be a significant prognostic factor of clinical outcome in PCa patients. A second study that found TAM to be related to recurrence indicated this to be true only in the control but not ADT treated group. Thus, TAM count may be a good biomarker for disease progression, but further research is needed to identify the specific subset of macrophages and PCa patients for which this is applicable. TAM may also represent a viable therapeutic target for PCa, as demonstrated by Escamilla et al., who showed a reduction in castration-induced recruitment of pro-tumorigenic TAM and delayed emergence of CRPC in a mouse Myc-CaP PCa model following CSF1R inhibition.

11As expected, presence of immunosuppressive immune cells, such as Treg, myeloid derived suppressor cells (MDSC), and according to one report, TH17 cells, in prostate tissue was suggestive of more aggressive disease, but surprisingly few studies were able to equate this with clinical prognosis. Neutrophils and mast cells have also been fairly well studied, but investigations involving both cell types have generated conflicting results concerning their influence on PCa. One source of contention may concern the area of tissue evaluated. This point is nicely illustrated by Johannson et al., who implicated peritumoral but not intratumoral mast cells as independent prognostic variables of PCa progression. In addition, work from Pittoni et al. emphasized that care must be taken to evaluate potential negative outcomes when manipulating immune cell populations with therapies. In the TRAMP model of murine PCa, targeting mast cells prevented growth of well-differentiated tumors but simultaneously stimulated progression of more aggressive, neuroendocrine type tumors. Similarly, Tang et al. observed a late IL-2-induced expansion of Treg following castration and immunization that negated the initial anti-tumorigenic CD8+ T-cell response.

12Relative to other immune cells, B cells have been generally overlooked in terms of PCa. However, the importance of these cells in PCa progression was established in a landmark and subsequent papers using subcutaneous Myc-CaP and spontaneous TRAMP mouse PCa models to show that castration-induced cell death and resulting hypoxia stimulated CXCL13-mediated infiltration of lymphotoxin β (LTb)-expressing B lymphocytes into regressing tumors. These cells subsequently activated IKKα, STAT3 and BMI1 in cancer cells to promote CRPC progression. Notably, interference with any aspect of this pathway delayed emergence of hormone-resistant tumors. These initial findings were extended to reveal that the identical mechanism operates in androgen-dependent regeneration of normal prostate progenitor cells. Interestingly, an independent investigation reported significant reductions in human and murine PCa tumor growth following combination therapy with HIF-1α and phospho-STAT3 inhibitors. Together, these data suggest that the B cell response to tissue injury or death following androgen depletion drives restorative or aberrant proliferation in the prostate. In addition, a recent report implicates immunosuppressive IgA+PD-L1+IL-10+ plasma cells in impeding response to low-dose chemotherapy in a T-cell-dependent manner. The presence of CD20+ B cells has been documented in human PCa as well, and the density of these cells was shown to be higher in malignant than benign tissue. Furthermore, a case study of an advanced PCa patient treated with the B cell depleting antibody Rituximab produced fairly encouraging results. A Phase 0 clinical trial evaluating Rituximab as neoadjuvant therapy in patients prior to radical prostatectomy is currently ongoing. Thus, although research is still in preliminary stages, B cells may represent an emerging target for PCa immunotherapeutics. However, it is likely that a damage response and/or an immunogenic stimulus is needed to make these cells relevant in prostate carcinogenesis or that only particular subsets, such as those expressing LTβ, or IgA are crucial.

13The emergence of immune cells as important mediators of PCa progression has spawned a multitude of clinical trials targeting these cells via different approaches, including therapeutic vaccines, checkpoint inhibitors/immune modulators, adoptive T-cell therapy, and monoclonal antibodies. In theory, PCa should be particularly conducive to this type of treatment due to its long latency period. However, because PCa is commonly diagnosed late in life, therapeutic vaccines may not be as efficacious as desired due to the age-related decline in immune responsiveness. Despite this fact, Sipuleucel-T was the first immunotherapy approved by the FDA for PCa and has shown an overall survival advantage in mCRPC patients, although no effect on progression-free survival was noted. The lack of short-term response observed in this study may at least partially be partially attributed to the fact that immunotherapeutic agents are not directly cytotoxic, as are more standard treatments. Due to space limitations, individual studies will not be discussed in detail here but are summarized in Table 2. One important message from the combined studies is that, while extremely promising, these immunotherapeutic methodologies still need to be optimized as no individual treatment has yet elicited a complete response. Complete and durable responses will likely require combinations of current immunotherapy and/or more standard approaches, such as chemotherapy, radiation, or ADT, which induce an immunogenic stimulus or enhance T-cell responses. In addition, ideal target patient populations and treatment protocols still need to be identified. For example, a study combining PROSTVAC and ipilimumab discovered that patients with lower PD-1+Tim-3NEGCD4EM, higher PD-1NEGTim-3+CD8+, and more CTLA-4NEG Treg at baseline survived longer. Furthermore, an increase in Tim-3+ natural killer cells following vaccination conferred survival benefits. Similarly, Klyushnenkova et al. illustrated the importance of breaking immunological tolerance to “self” antigen by targeting CD25+ Treg during CTLA-4 blockade in mice. Although this research has several shortcomings, it highlights the potential value of such combination therapies.

14The presence of inflammatory cells in the prostate is well documented. Evidence suggests that these cells may drive PCa development. Additionally, growing tumors can induce recruitment of immune cells into the prostate microenvironment and initiate a reciprocal interaction that promotes disease progression. Surprisingly, information regarding precise profiles of prostate tumor-infiltrating leukocytes is extremely limited. Individual cells and the overall immune response have the potential to be anti- or pro-tumorigenic, depending on cellular phenotypes, combinations, localization and the tumor microenvironment. Thus, the identities, traits, functional status, distribution, kinetics, and interactions of inflammatory cells in the prostate and PCa must be fully characterized in order to elevate and combine the already promising current immunotherapies. Furthermore, efforts must be directed at elucidating better indicators of potential response to facilitate identification of ideal target populations, as well as biomarkers to assess therapeutic efficacy. Progress in these areas will strengthen the existing optimism that immunotherapy will become part of a successful standard care regimen for future PCa patients.

15The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.