Racial disparity in prostate cancer in the African American population with actionable ideas and novel immunotherapies

Abstract Background African Americans (AAs) in the United States are known to have a higher incidence and mortality for Prostate Cancer (PCa). The drivers of this epidemiological disparity are multifactorial, including socioeconomic factors leading to lifestyle and dietary issues, healthcare access problems, and potentially tumor biology. Recent findings Although recent evidence suggests once access is equal, AA men have equal outcomes to Caucasian American (CA) men, differences in PCa incidence remain, and there is much to do to reverse disparities in mortality across the USA. A deeper understanding of these issues, both at the clinical and molecular level, can facilitate improved outcomes in the AA population. This review first discusses PCa oncogenesis in the context of its diverse hallmarks before benchmarking key molecular and genomic differences for PCa in AA men that have emerged in the recent literature. Studies have emphasized the importance of tumor microenvironment that contributes to both the unequal cancer burden and differences in clinical outcome between the races. Management of comorbidities like obesity, hypertension, and diabetes will provide an essential means of reducing prostate cancer incidence in AA men. Although requiring further AA specific research, several new treatment strategies such as immune checkpoint inhibitors used in combination PARP inhibitors and other emerging vaccines, including Sipuleucel‐T, have demonstrated some proven efficacy. Conclusion Genomic profiling to integrate clinical and genomic data for diagnosis, prognosis, and treatment will allow physicians to plan a “Precision Medicine” approach to AA men. There is a pressing need for further research for risk stratification, which may allow early identification of AA men with higher risk disease based on their unique clinical, genomic, and immunological profiles, which can then be mapped to appropriate clinical trials. Treatment options are outlined, with a concise description of recent work in AA specific populations, detailing several targeted therapies, including immunotherapy. Also, a summary of current clinical trials involving AA men is presented, and it is important that policies are adopted to ensure that AA men are actively recruited. Although it is encouraging that many of these explore the lifestyle and educational initiatives and therapeutic interventions, there is much still work to be done to reduce incidence and mortality in AA men and equalize current racial disparities.


| INTRODUCTION
African Americans (AAs) in the United States have higher incidences and mortality for a number of cancers. 1 Their lifetime risk of prostate cancer (PCa) is 1 in 6, with a lifetime risk of PCa specific mortality of 1 in 23. 1 For non-Hispanic Whites their risk of PCa is lower with incidence of 1 in 8, and mortality of 1 in 42. 1 The drivers of this epidemiological disparity in incidence and mortality are not clearly defined and complex, ranging from social networks also causing lifestyle and dietary issues, healthcare access problems, comorbidities, and their influences on pelvic inflammation and tumor microenvironment, as well as tumor biology with associated differences in genomic, molecular, and immunological pathways. [1][2][3] Which of these drivers is the most significant has been debated in the literature for some time, but, although there are clearly differences in tumor biology, more recent reports suggest once socioeconomic differences are removed, prostate cancer specific mortality rates for African and Caucasian men are similar. [4][5][6] Examining retrospective data in PCa patient cohorts from the Veterans Affairs health care system, as well as from Surveillance, Epidemiology, and End Results (SEER) database and four Radiation Therapy Oncology Group trials with long-term follow-up, showed once AA men have equal access to healthcare and standardized treatment, prostate cancer specific mortality (PCSM) is similar for AA and CA men. 4,5 Similarly, Tewari et al 6 in a PCa patient cohort from the Henry Ford Health Care System found differences in PCSM no longer existed after multivariate analysis adjustment for socioeconomic status. Nevertheless, the higher incidence of PCa in AA men remains, and given their reduced access, one would expect the reverse. Clearly, any discussion on racial disparity with a view to producing equity in PCa incidence and management outcomes needs to be multifactorial and is complicated by the fact that "race" is a social construct rather than a biological definition. AA men have diverse and genetically heterogeneous ancestry, not only within Africa but also Europe and the Americas, and so their molecular and immunological response to cancer as well as their tumor biology is likely to be equally diverse. 7 In addition to this recent epidemiological evidence, there have been reviews of the molecular and genomic aberrations in African American (AA) men, and detailed analyses of the molecular and genomic pathways of PCa for all ethnicities at times highlighting potential therapeutic targets. [8][9][10][11] Bhardwaj et al 8 13 Interesting studies on noncoding and micro-RNAs have shown differential expression in AA men as compared to CA men, 14 and with specific reference to the tumor microenvironment (TME), Gillard et al 15 have shown a more reactive stroma with enhanced chronic inflammatory infiltrate as well as fibroblast function in AA men.
As yet, to our knowledge, there have been no reports leveraging our understanding of molecular differences in PCa in AA men into clinical practice. This review will first frame PCa oncogenesis across all ethnicities in Hannahan and Weinberg's "Hallmarks of Cancer", 16 which is presented concisely in Table 1. This table aims to be used as a reference to benchmark molecular and genomic differences in AA men that have emerged in the recent literature, including differences in the tumor microenvironment and noncoding RNA. Once highlighted, and bearing the recent epidemiological evidence in mind, these key insights will be used as a stepping stone to devising actionable ideas and novel immunotherapies with the ultimate aim of providing at least some guidelines for reducing the incidence and mortality of PCa in the United States in African American men.

| GENOMIC AND MOLECULAR FUNDAMENTALS IN THE ONCOGENESIS AND PROGRESSION OF PROSTATE CANCER
In their original paper, Hanahan and Weinberg 86 outline 6 "Hallmarks of Cancer" as an organizing principle to explain oncogenesis and its T A B L E 1 Summary of prostate cancer oncogenic pathways  In keeping with the process of wound healing, the TME and process of EMT promotes increase protease activity, angiogenesis and inflammatory cells, but rather than laying down granulation tissue they cause stromal reactivity and tumor cell proliferation.
A recent interesting study examined EMT in a mouse model focused on bone marrow dissemination . They confirmed ex vivo tumor cells showed increased angiogenic, proliferative and migratory features with altered mesenchymal markers including E-cadherin, Snail, ZO-1, and vimentin. 15,[82][83][84][85] progression. This included constitutive proliferation, uncontrolled growth, loss of cell cycle control, overriding cell death, tumor-induced angiogenesis, and tumor Invasion and metastatic cascade." Later, they acknowledged the importance of "genetic instability, inflammation, metabolic reprogramming, escaping antitumor immune response, and tumor microenvironment" (see Figure 1 and Table 1). The genomic and molecular pathways to the development and progression of PCa have also been well described, 10,11,25 and this knowledge will be reframed across all ethnicities under Hanahan and Weinberg's Hallmarks. 16 Each of the Hallmarks' genetic or pathway abnormalities have been outlined concisely, with specific reference to prostate cancer in Table 1. This More study is required in this area.
ETS NEGATIVE TUMORS: this accounts for the other half of prostate cancers and has been putatively subclassified into three groups according to three dominant different genomic drivers. These are SPINK1 overexpression, SPOP mutation, and CHD1 deletion. 10,11 The details of these molecular pathways are summarized concisely in Table 1.

| SPECIFIC FINDINGS IN AFRICAN AMERICAN MEN
Having outlined the fundamentals of oncogenesis for prostate cancer according to the "Hallmarks of Cancer" 16 in Table 1, studies highlighting specific differences for AA versus CA men will be discussed under the same headings. Given that some of the studies broadly investigate differences using, for example, GWAS, there is inevitable crossover of data, and some of these Hallmarks have been discussed in combina-

| Loss of cell cycle control
Telomere shortening occurs early in prostate cancer oncogenesis, possibly due to oxidative stress and local inflammation, and it can also increase genomic mutational burden. 41 41,149,150 and the MYC transcription factor is known to bind to the TERT promoter. 151 In AA men, both c-MYC oncogene expression and telomerase activity are elevated and at a level that is significantly higher than for CA men, 9 so maintaining the cancer cells' replicative potential.

| Tumor Invasion and metastatic cascade
Ali et al 152  found 17% of patients had a CDC27 to OAT gene fusion, the latter of which they note is influenced by AR signaling. [156][157][158] Epigenetic differences affecting WNT signaling and stemness: Differing epigenetic changes between AA and CA men have been shown in a number of studies and have been concisely reviewed by Karakas et al. 96 The genes RARβ2, SPARC, TIMP3, and NKX2-5 are more highly methylated in AA compared to CA men, and NKX2-5 and TIMP3 are hypermethylated, even in benign prostatic tissue of AA men. 159 Devaney et al 160

| Genetic variations and chromosomal instability
Differences in tumor biology resulting from pathways maintaining genomic stability as well as mutational burden between AA and CA men have been shown. 8

| Genetic polymorphisms
Genetic polymorphisms can be used as markers of genetic susceptibility as well as for disease prognosis. A study by Freedman et al, 171 using whole-genome admixture mapping analysis, identified a 3.8 mb region at 8q24 associated with PCa risk in AA men. It contained 9 genes, and a follow-up study demonstrated seven SNPs conferring a higher risk of PCa to AA men over CA men. 96,172 One of these nine genes was the C-MYC oncogene, discussed above, with the others listed in Table 2 182 There is also an association between AR transcriptional activity and genomic instability, 180 and AR signaling has been shown to regulate DDR. 183 One of the most prevalent somatic mutations in the AR is Thr877Ala, which has discussed in the AR hypermutation section above, but how AR signaling influences genomic instability is not fully understood.   189 and other mitochondrial DNA mutations have been associated with increased tumorigenicity in mice experiments. 190 Another study had found reduced mitochondrial DNA content correlates with adverse outcomes in AA patients. 191 Finally, Chaudhary et al 192 found reduced apoptosis related to mitochondrial dysfunction and defective heat shock proteins in AA men with prostate cancer, more specifically related to differences in mitochondrial unfolded protein stress response. 192 More studies are required to further investigate these differences and potentially develop actionable ideas of clinical benefit related to them.

| Tumor microenvironment, Pro-tumor inflammation, Escaping antitumor immune response, and Tumor-induced angiogenesis
In comparison to Asian and Caucasian populations, those of African descent have rarer allele frequencies and more diversity in nucleotide composition. This results in differences in immune landscape that were adapted to their ancestral origins 193 that have been shown to be specific to tissue types, but not specific cancers. 194 The influence of these differences on immune surveillance, and the innate and adaptive immune responses, has been suggested to affect not only oncogenesis but also response to treatment and cancer outcomes.
Tumor microenvironment, inflammatory axis, and cytokines: The importance of the tumor microenvironment (TME) in PCa oncogenesis and disease progression has been established but few studies have examined its association with racial disparity. 15  Studies have also found that, in prostate cancer, about 50% of disparities are seen at the 8q24 locus. 174 The long noncoding RNA, plasmacytoma variant translocation 1 (PVT1) at 8q24.21 plays an oncogenic role in prostate cancer, 219 and has been implicated in prostate cancer invasion and metastasis. 220 Tonon et al in a recent study reported higher expression of PVT1 in tumors from African Caribbean men compared to French Caucasians. 182 These studies demonstrate the relevance of exploring the expression and function of noncomplex molecular regulatory networks with coding genes, and may help develop better prognostic and therapeutic markers for prostate cancer patients of different ethnicity.

| ANATOMICAL AND MOLECULAR CORRELATIONS
Another recent study explored potential anatomical differences and molecular correlations in PCa between AA and CA men. 221 In a retrospective cohort of close to 300 men, they confirmed previous findings that ETS positive tumors (with either ERG or other ETS transcription factor fusions) were commoner in CA men, whereas SPINK overexpression was more common in AA men. In keeping with the studies of Sundi et al, 222 they found in cases defined as low risk preoperatively, anterior location was more common in AA men than CA men (50% vs 20% respectively). 221 However, racial differences in molecular subtype did not persist when tumors were analyzed by location, and anterior tumors had higher volume, lower PSA density, and higher risk-genomic classifier scores, suggesting a possible propensity to increased disease progression in the future. 221

| DIFFERENCES FOR AA MEN IN PUTATIVE DISEASE SUBCLASSES
Khani et al, 223 examining radical prostatectomy specimens in over 100 AA men and CA men, allowed a direct comparison of some of the Decipher). Given the differences in tumor biology that have been described, the hope is that research will develop biomarkers and gene panels to provide prognostic information and aid in treatment decisions specifically for the AA population.

| TREATMENT
Prostate cancer treatment has seen a number of advances in recent years and, in response to a deepening understanding of molecular pathways underlying the evolution of the disease, the concept of "Precision Medicine" has emerged. 254 This aims to provide individualized treatment, potentially in combination, once a patient's prostate cancer has undergone genomic profiling, over and above other treatments that have gained widespread acceptance (eg, cytotoxic therapy with radium-223 or cabazitaxel, anti-AR signaling with enzalutamide or abiraterone, and antitumor immunotherapy with sipuleucel. [255][256][257][258][259][260] Performing genomics on patients' solid tumor tissue, after biopsy or surgical treatment, to identify abnormalities in AR signaling, DNA repair, PI3K, WNT, and cell cycle pathways, as well as examining for immune/tumor microenvironment response signatures, allows the integration of clinical and genomic data to choose appropriate therapeutic trials for individual patients. 254 Furthermore, potential biomarkers described above from research on RNA sequencing, proteomic analysis, and immune response gene expression in AA men may also reveal targets for exciting novel immunotherapies, not only highlighting an individual patient's genomic risk profile, but also those who may develop therapeutic resistance, so earmarking them for specific combination therapies. Currently, some of the commonest abnormalities found on genomic profiling relate to deficiencies in homologous recombination DNA repair (genes such as ATM, PALB2, BRCA1, BRCA2, CHEK2, and the FANC (Fanconi's anemia) genes) or in microsatellite instability resulting from deficiencies in mismatch repair genes (eg MLH1, PMS2, MSH2 and MSH6). 260 These will be considered individually, although it is worth noting that, at the time of writing, only DDR and MSI-H treatments are currently actionable.
8.1 | Treatment for homologous recombination DNA (HRD) repair deficiencies PARP inhibition: PARP functions to repair single-strand DNA breaks.
If there is no PARP function, cells convert single strand to double strand DNA breaks, which are dealt with by HRD repair. When PARP is inhibited, and HDR repair pathways are deficient, the chromosome becomes unstable, and cell death occurs due to overwhelming genomic damage. 260 In the phase 3 PROFound trial assessing olaparib vs enzalutamide or abiraterone (having had standard ADT) for mCRPC, the PARP inhibitor gave a nearly 4 month progression-free survival advantage. 261 The patient cohort receiving olaparib had at least one deficiency in ATM, BRCA1, or BRCA2, and so in this study, the effect is limited to patients with DDR deficiency. Similarly, a prior recent phase 2 trial of the PARP inhibitor, Olaparib, not specific for AA men, but in 50 patients with castration resistant metastatic disease, reported encouraging results. All patients had received multiple treatments prior to the trail, were resistant to taxanes, and 22% had more than 50% reduction in PSA values. Of the 32% of patients' biomarker positive for HDR deficiencies (as in deficient of the above genes on profiling), 88% responded and clinical outcomes were significantly better than for biomarker-negative patients. 262 As far as we are aware, differential responses between AA biomarker positive and CA biomarker positive men have not been evaluated and require further study.
Checkpoint immunotherapy: Following the success of sipileucel-T, 259 there has been a focus on immunotherapy for prostate cancer, 260 and specifically checkpoint inhibitors linked to PD1 and CTL4 pathways. However, a phase 1 trial for Nivolumab (antiPD1) 263 and two phase 3 trials for Ipilimumab (anti-CTL4) 264,265 failed to show any convincing treatment responses. 260 Other trails with anti-PD1s (eg Pembrolizumab) have shown some PSA and radiographic response, 260,266,267 raising the possibility of a subset of patients who may respond. On the basis that some BRCA deficient tumors have responded to PD1 inhibition, trials are exploring its use in tumors biomarker positive for HDR and MMR abnormalities. 260,268 How this may direct treatment for populations of AA men will also require further investigation (see Table 1.).

| Treatment for microsatellite instability-high tumors (MSI-high) and mismatch repair (MMR) gene deficiencies
Only approximately 5% of advanced prostate cancers have MSI-high and MMR abnormalities, 260 and trials have been targeting this population with PD1 inhibition (eg, Pembrolizumab) in lung, colorectal tumors, and malignant melanoma, where the mutational load is high. [269][270][271] However, a recent trial of castration resistant metastatic PCa with MSI, using combined enzalutamide with pembrolizumab, showed complete PSA response in 3 of 10 patients, one of whom had combined MSI and PDL-1 tumor expression. 260,266 Similarly, biallelic somatic CDK12 mutations have been found in nearly 7% of mCRPC, which were not only associated with immune cell infiltration and enhanced checkpoint protein expression, but also increased efficacy of pembrolizumab PD-1 inhibitor treatment. 272 Once again, more study is required to assess how these responses may differ in a population of AA men.

| Other immunotherapeutic initiatives
The use of Sipuleucel-T for patients with mCRPC has been well described but a recent subanalysis of data from the original PROCEED trial demonstrated an 0S advantage of 9.5 months for AA men over CA men for all patients, as well as nearly 21 months when comparing AA to CA men with PSAs under 30. 273 More study is required to understand the immunogenetics underlying this difference, and potentially develop focused AA therapies to exploit it further.
As mentioned earlier, our group has been applying the integrative use of clinical and genomic data, to uncover individual patients who may have high risk disease (with high risk of biochemical relapse), in combination with an immune/TME response signature. Such patients have more CD8/3 T-cells, lower AR-receptor activity, more inflammation, and lower levels of immune suppressive inflammatory cells (T-reg cells and MDSC cells). By identifying specific cohorts of AA and CA men in this way and offering them "Precision Medicine" with entry in targeted therapeutic trials, the differential response of AA men can be studied. The results of these investigations will be available in the future.

| CONCLUSION
This review has outlined the molecular and genomic fundamentals of prostate cancer for all ethnicities, framed in Hanahan and Weinberg's updated "Hallmarks" of cancer.
Using this framework as a benchmark, the molecular and genomic differences for AA men have been described, which may in part explain the racial disparity in disease incidence, even though recent epidemiological studies have shown equal access for AA men results in equal outcomes. 4,5 Emphasizing genetic abnormalities in prostate cancer, recommendations have been made for genomic profiling to integrate clinical and genomic data for the purposes of diagnosis, prognosis, and treatment planning for "Precision Medicine". Treatment options have also been discussed, with a concise description of recent work in AA specific populations, showing an immune response signature, and detailing a number of targeted therapies. A summary of the current clinical trials that are active or recruiting can be seen in Table 3 (see below). It is encouraging that many of these are exploring lifestyle and educational initiatives, as well as therapeutic interventions, but clearly there is much work to be done to reduce incidence and mortality in AA men, and equalize current racial disparities.