Technical Feasibility of Tissue Microarray (TMA) Analysis of Tumor-Associated Immune Response in Prostate Cancer.

Introduction: The androgen receptor (AR) regulates immune-related epithelial-to-mesenchymal transition (EMT), and prostate cancer (PCa) metastasis. Primary tumor-infiltrating lymphocytes (TILs) [CD3+, CD4+, and CD8+ TILs] are potential prognostic indicators in PCa, and variations may contribute to racial disparities in tumor biology and PCa outcomes. Aim: To assess the technical feasibility of tumor microarray (TMA)-based methods to perform multi-marker TIL profiling in primary resected PCa.
Methods: Paraffin-embedded tissue cores of histopathologically-confirmed primary PCa (n = 40; 1 TMA tissue specimen loss) were arrayed in triplicate on TMAs. Expression profiles of AR, CD3+, CD4+, and CD8+ TILs in normal prostate, and the center and periphery of both the tumor-dominant nodule and highest Gleason grade were detected by IHC and associated with clinical and pathological data using standard statistical methodology. An independent pathologist, blinded to the clinical data, scored all samples (percent and intensity of positive cells).
Results: TMAs were constructed from 21 (53.8%) Black and 18 (46.2%) White males with completely-resected, primarily pT2 stage PCa [pT2a (n = 3; 7.7%); pT2b (n = 2; 5.1%); pT2c (n = 27; 69.2%); pT3a (n = 5; 12.8%); mean pre-op PSA = 8.17 ng/ml]. The CD3, CD4, CD8, and CD8/CD3 cellular protein expression differed from normal in the periphery of the dominant nodule, the center of the highest Gleason grade, and the periphery of the highest Gleason grade (P < 0.05). Correlations between TIL expression in the center and periphery of the dominant nodule, with corresponding center and periphery of the highest Gleason grade, respectively, were robust, and the magnitude of these correlations differed markedly by race (P < 0.05). Conclusions: Multi-marker (AR, CD3, CD4, CD8) profiling with IHC analysis of TMAs consisting of primary, non-metastatic resected prostate cancer is technically feasible in this pilot study. Future studies will evaluate primary tumor immunoscore using semi-quantitative, IHC-based methodology to assess differences in the spectrum, quantity, and/or localization of TILs, and to gain insights into racial disparities in PCa tumor biology and clinical outcomes.

Introduction

Prostate cancer (PCa) is the second-most frequently diagnosed and fifth-most lethal malignancy amongst men 1. It is particularly troubling that PCa has the highest disparity in incidence and mortality between African American (Black) and Caucasian American (White) men 2, 3, more so than any other malignancy 4-6. Black men are nearly three to four times as likely to die of PCa compared to White men (50.9 versus 17.8 cases per 100,000 men, respectively) 5, 7-10.

There are also racially-based differences in PCa molecular biology. Black men tend to present with more advanced stage, more aggressive disease, and with adverse pathological indicators of disease than men in other ethnic groups 4, 5, 11-13. In addition to having variations in genes linked to tumorigenesis (e.g., androgen receptor 14-17, growth factors 18-20, p53 21, and BCL-2 22), Black men also have genetic variations in their genome-wide methylation 23, 24 and tumor-associated immune response. For instance, work from our group and others using genome-wide analyses of PCa tumor specimens from Black men revealed enrichment in the expression of gene sets related to lymphocyte and mononuclear cell proliferation, T-cell activation and proliferation, lymphocyte and leukocyte activation, and integrin-mediated signaling 25-30. Racial disparities also give rise to differences in how men with PCa are treated, wherein, even after adjusting for socioeconomic and clinical factors, facilities tended to favor more definitive therapy (i.e., surgery and radiation) for White men than Black men 31, 32.

The focus of this study is on immune-related parameters and PCa outcome. While genetic data suggest race affects immune-related gene expression, these differences have yet to be validated using tumor specimens and knowledge of long-term patient outcomes. The role of increased levels of tumor-infiltrating lymphocytes (TILs) in cancer appears to be context- and site-dependent 12, 33, and seems to vary with the specific T-cell population in question. In the prostate, CD8+/Foxp3+ and CD4+/CD25+ regulatory T (T-reg) cells appear to be the predominant TIL populations 34-36. Both CD8+ and CD4+ T-reg cells suppress the immune response in the prostate, in a TLR8 ligand-dependent manner 37. We postulate that factors in the prostate tumor microenvironment transform these normally beneficial cells into promoters of cancer invasiveness 38.

Results describing the clinical significance of CD4+ and CD8+ populations in PCa are limited, but supportive. In a study of 188 patients treated for localized PCa, the strong expression of TILs was correlated with reduced PSA recurrence-free survival 39. Similarly, a small study (n = 82) found correlations between CD8+ infiltrates and biochemical relapse in PCa patients following radical prostatectomy 40. A small case-controlled study revealed that, when stratified by CD4+ T-reg infiltration, men in the highest quartile of expression experienced a nearly two-fold increase in lethal PCa, compared to the lowest quartile of expression 36. Recently, a study of 22 patients with locally advanced prostate cancer patients treated with salvage radiotherapy who experienced biochemical relapse after radical prostatectomy found significant correlations between immunological tumor microenvironment and patient outcome 41. Higher CD8 and CCR7 TIL scores and lower CD45 and FoxP3 TIL scores were associated with prolonged progression-free and overall survival 41.

In the prostate gland, immune mediators in tumor-adjacent stroma alter the regulation of genes responsible for epithelial-mesenchymal transition (EMT), which is known to perpetuate tumor invasiveness and metastasis, and is regulated by androgen signaling (42, 43, and reviewed in 44). This is of particular interest because PCa in Black men tends to have higher AR activity, due to selective alterations in the CAG repeats within the AR gene 17, 45-47. Black men with PCa are also 27% more likely to stain positive for AR and have 81% greater nuclear expression of AR 14, 47. The “EMT signature” also appears to correlate with resistance to radiation therapy and anti-androgens in a subset of PCa patients 48. With this increased androgen responsiveness, one could posit that the tumorigenic epithelial cells within Black men would more likely undergo EMT and, thereby distant disease spread as a direct result of increased concentrations of TILs in the tumor microenvironment. Studies are needed to test the validity and clinical significance of this hypothesis.

Given the strong correlation between the type, density, and spatial distribution of infiltrating lymphocytes and survival, there are therapeutic implications and support for immune therapies to enhance anti-tumor immune responses 49, 50. Immunotherapy, or the use of agents to recruit the host's immune system to target and destroy malignant cells, is particularly attractive for PCa. This is because PCa has a relatively indolent rate of progression, which gives the immune system time to mount an effective, durable response, and the expression of tumor-specific antigens (i.e., prostate-specific antigen, prostate acid phosphatase, and prostate-specific membrane antigen) that could be targeted by immune modulators 49, 51, 52. This therapeutic approach, termed “antigen-specific vaccination,” is promising, as evidenced by the success with Sipuleucel-T. Sipuleucel-T is an autologous vaccine derived from harvested mononuclear cells from the patient, incubated with recombinant prostatic acid phosphatase (PAP) antigen, and reinfused into the patient, leading to a PAP-specific CD4+ and CD8+ T-cell response 51, 53, 54. This was one of the first vaccines approved by the Food and Drug Association for use as a cancer treatment in men with castrate-refractory metastatic PCa 55.

In the double-blind, placebo-controlled, multi-center Phase 3 trial, Sipuleucel-T Immunotherapy for Castration-Resistant Prostate Cancer, there were notable improvements in overall survival (increase of 4.1 months) and a decreased risk of death (22% reduction) in the Sipuleucel-T cohort, but no effect on time to disease progression 53. Responsiveness to therapy varied significantly based on age, variations in antibody titers, and clinicopathological features 53, 54, 56. Various other types of vaccination approaches (e.g., vector-based vaccines, personalized peptide vaccination) and immunomodular therapies have been explored, although they remain in the early-to-mid stages of clinical investigation, and with mixed clinical results 49, 57. We and others in the field believe that development and validation of reliable methods to screen and select PCa patients for immunotherapy-based regimens, based on density, location, and quantity of clinically-relevant immune infiltrates in the prostate microenvironment, will support the use and effectiveness of this promising type of anti-cancer therapy 49, 51, 52.

Tissue microarray (TMA)-based immunoscoring of PCa specimens has the potential to: (1) shed new light on the immune-related causes of PCa disparities, particularly in the context of androgen responsiveness; (2) enable reliable screening of patients in PCa immunotherapy trials; and, (3) provide a way to serially track changes in the tumor microenvironment in response to immunotherapy. However, the technical methods enabling such an assay have yet to be established. The current study assesses the technical feasibility of TMA expression profiling of TILs in primary resected PCa, and further demonstrates its sensitivity amongst tumor specimens from a racially diverse cohort.

Methods

Patients

We utilized a prostate tissue biorepository collected between 2003 and 2013 by Virginia Urology and Urology of Virginia. The tissue banks contain over 6,500 archival formalin-fixed paraffin-embedded (FFPE) post-prostatectomy blocks, 25% (1,630) of which are from Black men. De-identified information limited only to relevant clinical descriptors (original pathological assessment, PSA values, age, race, etc.) was received along with relevant long-term follow-up data, including events related to PSA recurrence, disease progression, and disease-related mortality.

Tissue Microarray Construction

Pathologists with expertise in genitourinary (GU) pathology confirmed the diagnosis of PCa from patient tissue stored in the prostate tissue biorepository. Confirmation of PCa diagnosis was through hematoxylin and eosin (H&E)-stained tissue slide review. H&E sections were labeled for FFPE tissue block core specimen targeting to provide quality assurance (QA) and quality control (QC) when constructing the TMA.

After archived slide review, diagnostic confirmation, and FFPE tissue block targeting and corresponding slide labeling, TMAs were constructed from targeted areas on the paraffin block with a tissue-array device. Triplicate cores (0.6-mm diameter; 0.28-mm2 area) from the primary tumor from each patient in the study were placed in a recipient TMA paraffin block (Figure ). The multi-tissue block was pre-heated for two hours at 55˚C to adhere the tissue cores to the recipient block. The block was then serially sectioned to obtain 4-µm slices of the TMA, which includes per patient cores of: (i) the largest, or dominant, nodule; and, (ii) the area with the highest Gleason-graded PCa. At each location, and in triplicate, we stained and quantified AR+ cells, and TILs (CD3+, CD4+, and CD8+ T cells) both at the margin (periphery) of the region and at the region's center (Figure ). For each of the two locations (dominant nodule and highest Gleason grade) per patient, we generated three sets of CD3+/CD4+/CD8+ center and margin (periphery) data, as well as AR+/AR intensity/AR H-score center and margin (periphery) data.

Immunohistochemistry and Quantitation Methodology

Pathologist-selected sections were stained using standard H&E protocol for pathological assessment, as specified above, and then were immunostained for CD8 (C8/144; Dako, Carpinteria, CA), CD3 (2GV6; Ventana, Tucson, AZ), and CD4 (SP35; Thermo Scientific, CA) reactivity using automated slide stainers (Autostainer; DAKO, Denmark), precisely according to Worldwide Immunoscoring Task Force (WITF) standards for CD8 and CD3 58. For CD4, the staining protocol was according to manufacturer recommendations and aligned with the WITF protocol. Each marker was stained and quantitated in triplicate.

After the FFPE tissue in the TMA was cut into 4-μm sections and placed on the positively-charged poly-lysine-coated slides, the sections were dried at 70 ˚C, deparaffinized using a series of xylenes, and rehydrated in graded alcohols and distilled water. Following deparaffinization, the slides were treated with methanol peroxide to block endogenous peroxidases. Antibody validation studies were conducted to specifically optimize and determine each condition for the IHC protocol, including the appropriate primary antibody dilution, blocking reagent, incubation times, and antigen recovery. Immunostaining was performed using an Ab-specific auto-stainer. Following antigen retrieval, slides were blocked and labeled with primary antibody. Primary antibody binding sites were visualized using a non-avidin biotin polymer peroxidase detection system. Diaminobenzidine (DAB) / 1% hydrogen peroxide was used as the chromogen. The DAB reaction was terminated once optimal color visualization was determined. Following the DAB reaction, the slides were counterstained with hematoxylin, treated with xylene and graded alcohols, and mounted. Controls in each antibody run consisted of isotype-specific antibodies that matched the primary antibody of interest.

Interpretation of stained slides was performed by microscopic examination. In general, a morphologic review of the tissue on the slide was performed to determine whether an adequate amount of tissue was present for evaluation, and whether the designated tissue of interest was appropriately represented. One sample failed to meet the above standards and was rejected from the analysis.

The scoring system included an analysis of staining intensity. The staining intensity of the test article was judged relative to the intensity of a control slide containing an adjacent section stained with an irrelevant, negative-control antibody that was species- and isotype-matched to the test article. Staining of the section labeled with the negative-reagent control was considered “background.” A “0” indicated no staining relative to background; “1+” indicated weak staining; “2+” indicated moderate staining; and “3+” indicated strong staining. A total number of manually-counted staining cells was determined by the scoring genitourinary pathologist. Subcellular localization of the antibodies (nuclear, cytoplasmic, or membranous staining) was also assessed.

A semi-quantitative scheme was used for AR scoring expression. One component of this scheme was a value proportional to the amount of epitope present on individual cells as assessed by the intensity of product present (AR+ %). In addition, AR intensity was determined using a semi-quantitative scale: “0” = no staining; “1+” = faint staining; “2+” = moderate staining; and “3+” = prominent staining. Considering the nature of heterogeneous staining of PCa, we also used histological score (HSCORE), which is a measure of both the intensity and distribution of staining, to measure the immunohistochemical staining of AR. The HSCORE was calculated using the equation: HSCORE ∑Pi(i + 1), where Pi is the percentage of stained epithelial cells for each intensity and i is the intensity of staining.

Statistical Analysis

Summary statistics were obtained using established methods. Means ± standard deviation, or counts and perentages were calculated, as appropriate. Associations between ordered categorical factors and cellular protein expression were analyzed using Spearman's correlation coefficient. Analyses of continuous data were conducted using generalized estimating equations and an exchangable working correlation matrix. A gamma distribution with a log link was used to model the data.

We also explored the strength of the relationship between cellular expression of CD3, CD4, and CD8 in the center and margin (periphery) of the dominant nodule, along with expression in the center and margin (periphery) of the highest Gleason-graded region. We determined the degree to which this relationship differed by race of the donor using the Fisher r-to-z transformation for Spearman correlation coefficients 59.

Statistical analysis was performed using JMP v8.0 statistical software (JMP® and SAS, Cary, NC). A P-value < 0.05 (two-tailed) was considered statistically significant. Results are presented without adjustment for multiple comparisons, due to the exploratory nature of this investigation.

Results

The study population is summarized in Table according to patient, tumor, and disease stage characteristics. TMAs were constructed from 21 (53.8%) African American, and 18 (46.2%) Caucasian males, with mean pre-op serum PSA levels of 8.17 ± 11.20 ng/mL, and completely resected prostate cancer of mostly pathological T2 stage [pT2a (n = 3; 7.7%); pT2b (n = 2; 5.1%); pT2c (n = 27; 69.2%); pT3a (n = 5; 12.8%)]. Most of the resected tumors demonstrated peri-neural invasion (n = 33; 84.6%), and most of the specimens had surgical Gleason grade of either 3+3 (n = 7; 17.9%), 3+4 (n = 16; 41.0%), or 4+3 (n = 13; 33.3%). One sample failed to meet standards for TMA sampling, processing, and analysis, and was therefore excluded from the study.

The CD3, CD4, sand CD8/CD3 cellular protein expression differed from normal in the dominant nodule periphery, highest Gleason grade center, and highest Gleason grade periphery regions (all P < 0.05; Table ). The CD4 and CD8/CD3 cellular protein expression also differed from benign in the dominant nodule center (P < 0.05).

With respect to AR staining, the percent of AR+ cells differed from normal in the dominant nodule periphery, as well as the highest Gleason grade periphery (P < 0.01). Pairwise comparisons for the intensity of AR staining revealed differences between: the center and periphery of the dominant nodule center; the highest Gleason grade center and periphery; the center of the dominant nodule vs. the center of the highest Gleason grade; the periphery of the dominant nodule vs. the periphery of the highest Gleason grade; and the dominant nodule center vs. highest Gleason grade periphery (all P < 0.05). The AR H-score differed significantly between: normal vs. periphery of both the dominant nodule and the highest Gleason grade; the center and periphery of the highest Gleason grade; and the center of the dominant nodule vs. the periphery of the highest Gleason grade (P < 0.05; Table ).

Trends in mean CD3+, CD4+, and CD8+ IHC quantitative staining were identified within the dominant nodule periphery according to peri-neural invasion (Table ). Analysis of the periphery of the dominant nodule and the center of the highest Gleason grade region identified a trend of increasing AR intensity and AR H-score with Gleason score (Tables and 5). We also identified an inverse trend between mean CD3 positivity in the center of the region with the highest Gleason grade when stratified according to the highest surgical Gleason grade (Table ).

Specifically referring to the periphery of the highest Gleason grade region, we found noteworthy differential biomarker (AR and CD3+, CD4+, CD8+ TILs and the CD8/CD3 ratio) expression (Table ). When comparing AR intensity and AR H-score according to race, AR expression was higher amongst samples from White men compared to samples from Black men, and its expression increased with Gleason grade (Table ). Similarly, CD8 positivity correlated with increasing pre-treatment serum PSA. Interestingly, significant differential expression was identified across all AR categories according to surgical Gleason grade (P < 0.05; Table ). Differential expression of CD3+, CD4+ and CD8+ tumor cells within this region was also noted when comparing the pathological stage of resected primary PCa (Table ). Relatively higher CD3, CD4, and CD8 positivity was also seen, indicating peri-neural invasion (Table ).

As seen in Table , the correlations between TIL positivity in the center and periphery of the dominant nodule, with corresponding center and periphery of the highest Gleason-graded region, respectively, were robust—most exceeding 0.70. The magnitude of these correlations differed markedly by race, significant in every instance except for CD4+ and AR H-score collected at the centers of the dominant nodule and highest Gleason-graded regions.

Discussion

The primary goal of this study was to establish methods for high-throughput tumor microarray (TMA)-based immunoprofiling of TILs and AR in primary resected PCa specimens. Our secondary objective was to use this platform to begin to test the hypothesis that differences in TIL type and quantity within resected PCa specimens of Black and White men correlate with known disparities in disease staging, severity, and possibly outcome 4-6, 10.

To improve our ability to both quickly and quantitatively characterize TIL populations in resected primary PCa specimens, to conclusively assess the role of TIL invasion in PCa outcome (particular with respect to health disparities), and to inform discussion of provider recommendations for care and treatment, methods were needed for a high-throughput IHC-based platform that could perform TIL profiling on FFPE prostate tissue. Guided by successful parallel efforts to profile TILs in other tumor types (e.g., breast 60, colorectal 61-63, and ovarian 64), and using a preliminary cohort collected from a unique prostate tissue biorepository (see Methods), we report on the successful use of TMAs and a scoring system (Figure ) to quantify the expression of CD8, CD3, CD4, and AR in multiple clinically-relevant regions of the primary resected PCa sample (e.g., the largest, or dominant, nodule, and the area with the highest Gleason-graded cancer; Figure ).

Despite the small sample size of this preliminary cohort, we found statistically relevant differences in the expression of TIL markers (CD3, CD4, CD8 and the CD8/CD3 ratio) between normal/benign tissues and within/between the different regions of the tumor, and also in relation to some measures of disease severity [e.g., surgical Gleason grade (Tables )], as expected, given the known involvement of TILs in PCa tumorigenesis 36, 37, 39, 41, 65. These data support the technical validity of this platform. There were also trends (though not yet statistically significant) suggesting correlations between TIL markers and peri-neural invasion, surgical Gleason grade, and pre-treatment serum PSA. We also identified expected variations in the expression of AR, in both intensity and H-score, like previous studies 14, 47.

A second objective of this study was to begin to explore differences in TIL invasion and AR expression in racially diverse PCa specimens. This premise was based on studies that found racially-based differences in the expression of genes related to TIL proliferation, activation, and signaling 25-30, as well as androgen signaling 14-17. Despite a small sample size in this study, we were already able to find statistically relevant racial differences in the intensity of AR expression at the periphery of the highest surgical Gleason graded tissue (Table ). Most strikingly, we also found strong correlations between the TIL expression profiles at clinically-relevant sites in the PCa specimens of Black men, but not necessarily White men (Table ). For example, if a PCa specimen had high expression of CD3 at the center of the dominant nodule and was from a Black man, in our study, that specimen always also had high expression of CD3 at the center of the highest Gleason-graded region (Spearman correlation of 1.00; Table ). Correlation was similarly high for the expression of all TIL markers and AR intensity and scoring for PCa specimens from Black men. This suggests that PCa tumors from Black men may have more homogenous TIL invasion and AR expression across the tumor, and that profiling these markers at any clinically-relevant site (e.g., dominant nodule or area of highest Gleason grade) would provide as much information as profiling at all sites. Conversely, this correlation was not as consistent in PCa specimens from White men, suggesting that TIL infiltration in White men may be more heterogeneous and that certain combinations of marker/location may correlate more significantly with disease outcome than others. Future studies testing this hypothesis are warranted.

It is important to note the limitations of this study. First and foremost, as this was a preliminary study aimed primarily at developing methodology, the sample size was small and is likely to have contributed to the high variability in scoring data. This variability is expected to become smaller as additional samples are profiled and will ultimately serve to refine our ability to detect statistically relevant differences in TIL profiling between subgroups of PCa patients. Second, to best inform larger-scale, future studies utilizing this platform, we explored and report upon a wide variety of factors and parameters, including but not limited to: four markers of TILs, three measures of AR status, four clinically-relevant geographical locations within the tumor, and a host of clinicopathological variables (i.e., pre-treatment PSA, Gleason grade at biopsy and surgery, clinical and pathological stage, and presence of peri-neural invasion). While the study was not sufficiently powered to assess differences within and between all these variables, we think it is important to report on these findings since there were notable trends and patterns in the data that are provocative and will generate testable hypotheses for future studies.

Colorectal cancer has been the model for proof-of-principle research, development, and validation of the Immunoscore. Noteworthy progress has been made in the immunophenotyping of tumors of the colon and rectum to enhance the: characterization of the tumor microenvironment or immune landscape, staging of disease, prognostic scoring, predicting disease outcome, and providing clinical decision support (66-69). Galon et al. as part of a global task force have sought to gain prognostic, predictive and mechanistic insights through a standardized, immune classification and stratification system for colorectal cancer, the Immunoscore. This worldwide collaborative group has demonstrated the predictive value of the Immunscore and its researchers have recently found the Immunoscore to be superior to microsatellite instability in predicting recurrence-free and disease-specific survival in colorectal cancer (70).

The current Immunoscore classification for colorectal cancer, is based on positivity to CD3 and CD8—both along the margin and center of the tumor (67). CD45RO was of interest initially but was not chosen by the Task Force due to high background staining and technical limitations. A low density of cell types in both tumor regions results in an Immunoscore of “0”; when high densities are found, the sample is given an Immunoscore of “4”.

Our intent is to develop a modified Immunoscore, incorporating quantitation of CD4+ cell populations, given their supposed involvement in prostate tumorigenesis. In this study we showed the technical feasibility of TMA expression multi-marker profiling of TILs in primary resected PCa in a racially diverse cohort. Specifically, CD3, CD4, CD8, and CD8/CD3 cellular protein expression differed from normal in the periphery of the dominant nodule, the center of the highest Gleason grade, and the periphery of the highest Gleason grade (P < 0.05). Correlations were found to be robust between TIL expression in the center and periphery of the dominant nodule, with corresponding center and periphery of the highest Gleason grade, respectively, and the magnitude of these correlations differed significantly according to race (P < 0.05).

Significant correlations have been found between the type, density, and spatial distribution of infiltrating lymphocytes and outcome in patients with PCa. Immunotherapy-based approaches to enhance host anti-tumor immune responses are exciting and transformative areas of current research in difficult-to-treat cancer such as castrate-refractory metastatic PCa. The development and validation of methods such as biomarker-based clinical decision support tools that reliably quantify and localize clinically-relevant immune infiltrates in the prostate microenvironment [harnessing high-throughput tumor microarray (TMA)-based immunoprofiling] will be important to screening and selecting PCa patients for immunotherapy-based anti-PCa therapy. We believe that this will enable the delivery of personalized prostate cancer care with precision.

Conclusion

In conclusion, these findings support the continued use of a TMA-based platform to characterize TIL infiltration in PCa in a larger cohort, particularly with respect to short- and long-term survival outcomes and responses to various modalities of treatment and in relation to race.

Table 1

Distribution of Patient, Tumor, and Disease-Stage Characteristics

Category	Characteristic	n (%)
Race	Black	21 (53.8)
	White	18 (46.2)
Pre-treatment PSA	<10	35 (87.2)
(ng/mL)	10-20	2 (5.1)
	>20	3 (7.7)
Clinical Stage	T1c	27 (69.2)
	T2a	5 (12.8)
	T2b	2 (5.1)
	T2c	3 (7.7)
	Unknown	2 (5.1)
Biopsy Gleason Score	3+3	15 (38.5)
	3+4	12 (30.8)
	4+3	5 (12.8)
	4+4	5 (12.8)
	4+5, 5+4	2 (5.1)
Pathological Stage	T2a	3 (7.7)
	T2b	2 (5.1)
	T2c	27 (69.2)
	T3a	5 (12.8)
	T3b	1 (2.5)
	T4	1 (2.5)
Perineural Invasion	Absent	6 (15.4)
	Present	33 (84.6)

Table 2

Immunohistochemical Scoring in Normal and Primary Prostate Cancer Lesions

		IHC Markers Mean positivity ± Standard Deviation
Region	Location	CD3+	CD4+	CD8+	CD8+/CD3+	AR+ %	AR Intensity	AR H-Score
Normal	-	96.92±40.44	25.49±21.27	64.54±27.60	0.69±0.20	61.49±25.74	1.65±0.68	1.75±0.95
Dominant Nodule	Center	118.03±79.77	50.50±27.51*	65.71±56.81	0.54±0.19*	66.26±35.44	1.47±0.80	1.80±1.12
	Periphery	132.77±92.46*	52.00±52.44*	72.92±47.97	0.59±0.18*	72.82±26.45*	1.74±0.75+	2.11±0.97*
Highest Gleason Grade	Center	137.42±97.63*	64.13±86.85*	71.21±57.26	0.53±0.18*	65.21±34.22	1.61±0.79+	1.85±1.12
	Periphery	133.23±97.61*	56.66±56.04*	71.97±48.72	0.57±0.18*	74.62±24.69*	1.90±0.75+#0	2.27±0.96*+0
P-value	-	0.031	<0.001	0.304	0.001	0.002	0.006	0.015

Statistically significant difference (at least P < 0.05) between the indicated region and: *Normal; +Center of the dominant nodule; #Periphery of the dominant nodule; or 0Center of the highest Gleason-graded region.

Table 3

Immunohistochemical Scoring at the Periphery of the Dominant Nodule in Relation to Clinical and Pathological Characteristics

	IHC MarkersMean positivity ± Standard Deviation
Characteristic	CD3+	CD4+	CD8+	CD8+/CD3+	AR+ %	AR Intensity	AR H-Score
Race
Black (n = 21)	145.57±112.35	58.52±66.03	77.90±59.01	0.56±0.16	67.14±32.12	1.57±0.81	1.90±1.10
White (n = 18)	117.83±61.80	43.94±28.01	67.11±31.30	0.62±0.21	79.44±16.26	1.94±0.64	2.35±0.74
P-value	NS	NS	NS	NS	NS	NS	NS
Pre-treatment PSA (ng/mL)
<10 (n = 34)	124.68±91.89	47.12±44.70	65.47±36.30	0.58±0.17	73.24±25.67	1.71±0.76	2.09±0.96
10-20 (n = 2)	227.50±123.74	114.50±153.44	188.50±108.19	0.82±0.03	50.00±56.57	1.50±0.71	1.45±1.77
>20 (n = 3)	161.33±64.36	72.50±53.03	80.33±41.74	0.51±0.27	83.33±11.55	2.33±0.58	2.73±0.06
P-value	0.079	NS	NS	NS	NS	NS	NS
Clinical Stage
T1c (n = 27)	130.96±98.62	51.15±50.30	68.67±39.33	0.58±0.19	73.00±26.86	1.67±0.78	2.07±0.97
T2a (n = 5)	100.50±52.67	33.00±21.18	58.80±30.41	0.60±0.10	82.00±13.04	2.00±0.71	2.52±0.86
T2b (n = 2)	113.00±38.18	16.00±14.14	79.00±46.67	0.67±0.19	50.00±56.57	1.50±0.71	1.45±1.77
T2c (n = 3)	171.67±125.16	139.50±118.09	136.00±113.58	0.75±0.12	70.00±34.68	2.00±0.00	2.10±1.04
P-value	NS	NS	NS	NS	NS	NS	NS
Biopsy Gleason Score
3+3 (n = 15)	113.87±86.57	38.14±32.35	60.80±35.00	0.58±0.15	76.00±24.44	1.67±0.72	2.15±0.88
3+4 (n = 12)	168.67±102.61	68.25±64.57	90.83±37.56	0.58±0.19	65.83±29.06	1.50±0.80	1.76±0.99
4+3 (n = 5)	128.40±87.38	26.80±19.32	62.60±38.30	0.56±0.23	68.00±33.47	2.00±1.00	2.18±1.32
4+4 (n = 5)	119.00±110.78	72.20±85.79	88.00±99.33	0.70±0.23	78.00±26.83	2.00±0.00	2.34±0.80
4+5,5+4 (n = 2)	104.50±31.82	59.50±24.75	44.50±10.61	0.43±0.03	90.00±00.00	2.50±0.71	3.15±0.64
P-value	NS	NS	NS	NS	NS	NS	NS
Surgical Gleason Score
3+3 (n = 7)	99.71±43.28	37.00±24.17	67.29±39.59	0.65±0.16	61.43±28.54	1.29±0.76	1.53±0.78
3+4 (n = 16)	147.00±117.15	60.40±61.61	71.63±39.75	0.55±0.17	76.25±27.54	1.82±0.75	2.26±0.98
4+3 (n = 13)	129.62±66.71	40.92±29.01	67.39±35.27	0.54±0.16	70.77±26.29	1.85±0.80	2.09±1.06
4+4 (n = 1)	46.00±0.00	14.00±0.00	49.00±0.00	1.07±0.00	90.00±0.00	2.00±0.00	2.70±0.00
4+5,5+4(n = 2)	198.50±164.76	132.50±127.99	151.00±161.22	0.65±0.28	90.00±0.00	2.00±0.00	2.70±0.00
P-value	NS	NS	NS	NS	NS	NS	NS
Pathological Stage
T2a (n = 3)	94.00±38.04	37.33±16.17	73.67±49.22	0.74±0.21	80.00±10.00	1.67±0.58	2.13±0.55
T2b (n = 2)	69.50±12.02	18.00±18.39	35.50±9.19	0.53±0.22	75.00±21.21	2.00±1.41	2.40±1.70
T2c (n = 27)	132.37±98.63	53.62±51.14	69.04±37.31	0.58±0.17	71.85±28.29	1.70±0.78	2.07±1.00
T3a (n = 5)	147.00±74.40	35.20±9.72	77.80±33.86	0.55±0.17	70.00±33.91	1.60±0.55	1.92±1.06
T3b/T4 (n = 2)	224.00±128.69	129.00±132.94	149.50±163.34	0.55±0.41	80.00±1.14	2.50±0.71	2.75±0.07
P-value	0.043	NS	NS	NS	NS	NS	NS
Perineural Invasion
Absent (n = 6)	64.67±30.74	17.20±14.62	40.33±27.03	0.60±0.14	76.67±17.51	1.83±0.75	2.27±0.99
Present (n =33)	145.15±94.74	57.27±54.18	78.85±48.81	0.58±0.19	72.12±27.92	1.73±0.76	2.08±0.97
P-value	0.005	0.024	0.011	NS	NS	NS	NS

Table 4

Immunohistochemical Scoring at the Center of the Dominant Nodule in Relation to Clinical and Pathological Characteristics

	IHC MarkersMean positivity ± Standard Deviation
Characteristic	CD3+	CD4+	CD8+	CD8+/CD3+	AR+ %	AR Intensity	AR H-Score
Race
Black (n = 21)	138.00±82.72	52.85±87.27	76.55±62.19	0.52±0.20	63.30±37.48	1.45±0.83	1.75±1.20
White (n = 18)	95.83±72.19	47.89±67.47	53.67±49.10	0.55±0.18	69.56±33.80	1.50±0.79	1.86±1.07
P-value	NS	NS	NS	NS	NS	NS	NS
Pre-treatment PSA (ng/mL)
<10 (n = 34)	121.62±81.81	54.86±80.87	67.77±58.49	0.54±0.19	68.32±33.45	1.53±0.79	1.90±1.10
10-20 (n = 2)	139.00±99.81	6.00±1.78	30.00±24.33	0.40±0.09	53.67±46.69	1.33±0.58	1.30±1.14
>20 (n = 3)	70.33±0.00	16.00±0.00	30.00±0.00	0.74±0.00	0.00±0.00	0.00±0.00	0.00±0.00
P-value	0.079	NS	NS	NS	NS	NS	NS
Clinical Stage
T1c (n = 27)	135.78±82.97	61.44±88.67	74.89±61.93	0.53±0.17	68.41±33.90	1.41±0.69	1.76±1.05
T2a (n = 5)	79.60±50.13	29.00±23.05	44.80±27.28	0.62±0.27	68.00±38.99	1.40±0.89	1.90±1.20
T2b (n = 2)	76.50±88.39	4.50±2.12	53.00±27.28	0.48±0.37	45.00±63.64	1.50±2.12	1.80±2.55
T2c (n = 3)	108.50±34.65	48.50±41.72	61.50±27.58	0.55±0.08	35.50±48.79	2.00±1.41	1.40±1.98
P-value	NS	NS	NS	NS	NS	NS	NS
Biopsy Gleason Score
3+3 (n = 15)	109.60±70.83	52.73±20.91	68.73±68.94	0.60±0.22	62.47±38.54	1.47±0.74	1.71±1.18
3+4 (n = 12)	162.33±80.97	61.42±80.25	86.50±48.73	0.54±0.15	59.25±37.39	1.00±0.43	1.26±0.86
4+3 (n = 5)	72.20±68.53	24.80±41.55	42.00±44.86	0.50±0.15	70.00±40.62	1.40±0.89	1.92±1.17
4+4 (n = 5)	96.25±100.61	53.00±59.21	40.25±40.48	0.41±0.18	85.00±10.00	2.50±0.58	2.95±0.44
4+5,5+4(n = 2)	73.50±54.45	27.50±7.78	28.50±17.68	0.41±0.06	90.00±00.00	2.50±0.71	3.15±0.64
P-value	NS	NS	NS	NS	NS	NS	NS
Surgical Gleason Score
3+3 (n = 7)	175.29±107.94	89.71±97.40	95.00±60.89	0.61±0.24	50.14±38.95	1.00±0.58	1.13±0.98
3+4 (n = 16)	121.25±65.35	54.81±92.94	74.00±65.33	0.58±0.16	66.69±37.66	1.44±0.63	1.72±1.12
4+3 (n = 13)	91.54±70.01	28.85±36.97	46.08±37.75	0.46±0.18	70.77±32.78	1.69±1.03	2.12±1.15
4+4 (n = 1)	16.00±0.00	6.00±0.00	6.00±0.00	0.50±0.00	90.00±0.00	2.00±0.00	2.70±0.00
4+5,5+4(n = 2)	112.00±0.00	33.00±0.00	33.00±0.00	0.37±0.00	90.00±0.00	2.00±0.00	2.70±0.00
P-value	NS	NS	NS	NS	NS	NS	NS
Pathological Stage
T2a (n = 3)	196.33±82.04	148.67±125.13	119.00±86.89	0.56±0.17	60.33±51.38	1.00±0.00	1.20±1.04
T2b (n = 2)	67.50±51.62	14.50±19.09	33.00±0.00	0.69±0.53	45.00±26.15	1.50±2.12	1.80±2.55
T2c (n = 27)	123.33±79.88	49.89±76.04	68.33±55.94	0.54±0.16	66.93±33.34	1.44±0.70	1.74±1.03
T3a (n = 5)	85.00±52.88	18.80±19.49	45.50±38.71	0.46±0.19	74.00±41.59	1.80±1.10	2.40±1.39
T3b/T4 (n = 2)	6.00±0.00	3.00±0.00	2.00±0.00	0.33±0.00	74.00±41.59	2.00±0.00	2.10±0.00
P-value	0.068	0.034	NS	NS	NS	0.095	NS
Perineural Invasion
Absent (n = 6)	89.67±53.75	22.83±16.13	43.33±30.21	0.51±0.30	56.83±44.32	1.67±1.03	1.85±1.51
Present (n=33)	123.34±83.33	55.69±83.39	69.91±59.92	0.54±0.16	68.03±34.09	1.43±0.76	1.79±1.07
P-value	NS	NS	NS	NS	NS	NS	NS

Table 5

Immunohistochemical Scoring at the Center of the Highest Gleason-Graded Region in Relation to Clinical and Pathological Characteristics

	IHC MarkersMean positivity ± Standard Deviation
Characteristic	CD3+	CD4+	CD8+	CD8+/CD3+	AR+ %	AR Intensity	AR H-Score
Race
Black (n = 21)	137.20±83.37	55.20±86.45	76.45±61.57	0.54±0.18	63.30±35.75	1.50±0.76	1.73±1.16
White (n = 18)	137.67±113.91	74.06±88.68	65.39±53.21	0.51±0.19	67.33±33.34	1.72±0.83	1.97±1.08
P-value	NS	NS	NS	NS	NS	NS	NS
Pre-treatment PSA (ng/mL)
<10 (n = 34)	137.35±100.61	62.74±87.99	71.65±58.99	0.53±0.18	68.44±33.74	1.59±0.78	1.92±1.11
10-20 (n = 2)	139.67±96.42	85.33±104.67	59.00±52.60	0.39±0.16	40.33±34.06	2.00±1.00	1.40±1.23
>20 (n = 3)	133.00±0.00	48.00±0.00	93.00±0.00	0.70±0.00	30.00±0.00	1.00±0.00	0.60±0.00
P-value	NS	NS	NS	NS	NS	NS	NS
Clinical Stage
T1c (n = 27)	135.78±82.97	61.44±88.67	74.89±61.93	0.53±0.17	68.41±33.90	1.41±0.69	1.76±1.05
T2a (n = 5)	79.60±50.13	29.00±23.05	44.80±27.28	0.62±0.27	68.00±38.99	1.40±0.89	1.90±1.20
T2b (n = 2)	76.50±88.39	4.50±2.12	53.00±27.28	0.48±0.37	45.00±63.64	1.50±2.12	1.80±2.55
T2c (n = 3)	108.50±34.65	48.50±41.72	61.50±27.58	0.55±0.08	35.50±48.79	2.00±1.41	1.40±1.98
P-value	NS	NS	NS	NS	NS	NS	NS
Biopsy Gleason Score
3+3 (n = 15)	110.87±70.26	53.87±95.12	68.67±68.66	0.59±0.20	60.47±38.90	1.53±0.74	1.71±1.21
3+4 (n = 12)	152.25±84.88	58.67±80.89	80.33±53.25	0.52±0.18	57.58±36.82	1.17±0.58	1.35±0.95
4+3 (n = 5)	187.80±137.17	84.80±74.85	83.40±45.53	0.49±0.14	72.00±26.83	1.80±0.84	2.04±0.88
4+4 (n = 5)	96.25±100.61	53.00±59.21	40.25±40.48	0.41±0.18	85.00±10.00	2.50±0.58	2.95±0.44
4+5,5+4(n = 2)	204.00±239.00	144.50±173.24	67.00±72.12	0.39±0.10	90.00±0.00	2.50±0.71	3.15±0.64
P-value	NS	NS	NS	0.047	NS	NS	NS
Surgical Gleason Score
3+3 (n = 7)	175.29±107.94	89.71±97.40	95.00±60.89	0.61±0.24	50.14±38.95	1.00±0.58	1.13±0.98
3+4 (n = 16)	123.06±63.64	55.56±92.56	73.44±65.57	0.55±0.16	66.69±37.66	1.50±0.63	1.78±1.14
4+3 (n = 13)	125.92±106.74	49.77±58.22	56.92±42.53	0.46±0.17	67.69±28.91	2.00±0.91	2.18±1.10
4+4 (n = 1)	16.00±0.00	6.00±0.00	8.00±0.00	0.50±0.00	90.00±0.00	2.00±0.00	2.70±0.00
4+5,5+4(n = 2)	112.00±0.00	267.00±0.00	118.00±0.00	0.32±0.00	90.00±0.00	2.00±0.00	2.70±0.00
P-value	NS	NS	NS	NS	NS	0.002	0.017
Pathological Stage
T2a (n = 3)	196.33±54.02	148.67±47.62	119.00±86.89	0.56±0.17	60.33±51.38	1.00±0.00	1.21±1.03
T2b (n = 2)	67.50±66.16	14.50±58.32	33.00±0.00	0.69±0.53	45.00±63.64	1.50±2.12	1.80±2.55
T2c (n = 27)	119.56±18.01	49.30±15.87	65.56±56.55	0.53±0.16	65.07±33.62	1.56±0.70	1.79±1.06
T3a (n = 5)	203.80±41.85	85.00±36.88	79.00±49.04	0.41±0.20	78.00±26.83	2.00±0.71	2.46±1.11
T3b/T4 (n = 2)	251.00±93.57	206.00±82.47	118.00±0.00	0.47±0.00	60.00±0.00	3.00±0.00	2.40±0.00
P-value	NS	NS	NS	NS	NS	0.026	NS
Perineural Invasion
Absent (n = 6)	88.0±55.68	23.67±15.38	44.67±28.47	0.57±0.26	51.83±44.23	1.67±1.03	1.70±1.52
Present (n=33)	146.69±101.56	71.72±92.67	76.19±60.17	0.52±0.17	67.72±32.27	1.59±0.76	1.87±1.05
P-value	NS	NS	NS	NS	NS	NS	NS

Table 6

Immunohistochemical Scoring at the Periphery of the Highest Gleason Graded Region in Relation to Clinical and Pathological Characteristics

	IHC MarkersMean positivity ± Standard Deviation
Characteristic	CD3+	CD4+	CD8+	CD8+/CD3+	AR+ %	AR Intensity	AR H-Score
Race
Black (n = 21)	144.24±112.56	60.10±64.91	76.19±58.78	0.55±0.16	70.48±29.41	1.67±0.80	2.02±1.01
White (n = 18)	120.50±77.93	52.41±44.29	67.06±34.54	0.60±0.19	79.44±17.31	2.17±0.62	2.54±0.83
P-value	NS	NS	NS	NS	NS	0.045	NS
Pre-treatment PSA (ng/mL)
<10 (n = 34)	126.12±98.65	51.27±50.92	64.47±38.90	0.56±0.17	73.82±26.06	1.85±0.78	2.23±1.02
10-20 (n = 2)	161.33±40.77	79.50±12.02	89.33±6.43	0.58±0.18	76.67±15.28	2.33±0.58	2.50±0.21
>20 (n = 3)	213.00±0.00	125.50±137.89	173.50±129.40	0.79±0.07	85.00±7.07	2.00±0.00	2.55±0.17
P-value	NS	NS	NS	NS	NS	NS	NS
Clinical Stage
T1c (n = 27)	126.67±97.14	51.22±48.63	64.85±39.67	0.56±0.19	72.59±26.69	1.78±0.80	2.15±0.99
T2a (n = 5)	158.40±129.58	63.20±75.42	78.20±41.82	0.58±0.13	82.00±13.04	2.00±0.71	2.52±0.86
T2b (n = 2)	98.50±17.68	27.00±1.41	64.00±25.46	0.64±0.14	85.00±7.07	2.00±0.00	2.55±0.21
T2c (n = 3)	171.67±125.16	139.50±118.09	136.00±113.58	0.74±0.12	70.00±34.64	2.00±0.00	2.10±1.04
P-value	NS	NS	NS	NS	NS	NS	NS
Biopsy Gleason Score
3+3 (n = 15)	114.93±86.64	40.21±30.55	58.00±36.39	0.54±0.15	75.33±24.16	1.87±0.74	2.28±0.91
3+4 (n = 12)	157.67±103.41	63.75±64.16	85.75±39.11	0.58±0.18	65.83±29.06	1.50±0.80	1.75±0.99
4+3 (n = 5)	97.80±59.63	38.80±28.12	56.40±29.26	0.61±0.13	84.00±15.17	2.60±0.55	3.02±0.74
4+4 (n = 5)	119.00±110.78	72.20±85.79	88.00±99.33	0.70±0.25	78.00±26.83	2.00±0.00	2.34±0.80
4+5,5+4(n = 2)	249.00±172.53	135.00±82.02	93.00±57.98	0.39±0.03	90.00±0.00	2.50±0.71	3.15±0.64
P-value	NS	NS	NS	NS	NS	NS	NS
Surgical Gleason Score
3+3 (n = 7)	99.71±43.28	37.00±24.17	67.29±39.59	0.65±0.16	61.43±28.54	1.29±0.76	1.53±0.78
3+4 (n = 16)	147.94±116.95	61.60±60.66	69.38±41.21	0.52±0.16	75.63±27.32	1.94±0.77	2.36±1.03
4+3 (n = 13)	107.77±48.93	41.54±24.28	59.85±30.98	0.55±0.13	76.92±20.57	2.15±0.69	2.45±0.94
4+4 (n = 1)	46.00±0.00	14.00±0.00	49.00±0.00	1.07±0.00	90.00±0.00	2.00±0.00	2.70±0.00
4+5,5+4(n = 2)	343.00±39.60	208.00±21.21	199.50±92.63	0.60±0.34	90.00±0.00	2.00±0.00	2.70±0.00
P-value	NS	NS	NS	NS	0.032	0.041	0.032
Pathological Stage
T2a (n = 3)	94.00±38.04	37.33±16.17	73.67±49.22	0.74±0.21	80.00±10.00	1.67±0.58	2.13±0.55
T2b (n = 2)	69.50±12.02	18.00±18.38	35.50±9.19	0.53±0.22	75.00±21.21	2.00±1.41	2.40±1.70
T2c (n = 27)	128.07±97.21	52.65±49.45	65.22±37.70	0.56±0.17	71.48±28.11	1.81±0.79	2.14±1.02
T3a (n = 5)	161.80±121.40	65.00±71.93	79.00±40.06	0.54±0.15	88.00±8.37	2.20±0.45	2.84±0.67
T3b/T4 (n = 2)	255.00±84.85	155.50±95.46	179.50±120.92	0.66±0.25	75.00±21.21	2.50±0.71	2.55±0.21
P-value	0.036	NS	NS	NS	NS	NS	NS
Perineural Invasion
Absent (n = 6)	64.83±30.63	19.40±13.09	39.33±27.77	0.58±0.16	76.67±17.51	2.00±0.63	2.35±0.88
Present (n=33)	145.73±100.65	62.30±57.97	77.91±49.62	0.57±0.18	74.24±25.98	1.88±0.78	2.25±0.98
P-value	0.009	0.029	0.018	NS	NS	NS	NS

Table 7

Spearman Correlations Between Cellular Protein Expression in the Center of the Dominant Nodule vs. the Center of the Highest Gleason Graded Region (left) and the Periphery of the Dominant Nodule vs. the Periphery of the Highest Gleason Graded Region (right) for African American and Caucasian Patients, with Tests of Differences

	Correlation between centers of DN and HGG						Correlation between peripheries of DN and HGG
Marker	Black (n = 20)	White (n = 18)	Z*	P-value		Marker	Black (n = 20)	White (n = 18)	Z*	P-value
CD3+	1.00	0.313	9.81	<0.001		CD3+	0.999	0.595	8.79	<0.001
CD4+	0.853	0.597	1.63	0.103		CD4+	0.984	0.714	4.28	<0.001
CD8+	0.998	0.399	8.56	<0.001		CD8+	0.990	0.254	6.74	<0.001
CD8/CD3	0.953	0.793	2.21	0.027		CD8/CD3	0.991	0.719	5.06	<0.001
AR+%	0.986	0.862	3.32	<0.001		AR+%	0.970	0.699	3.46	<0.001
AR Intensity	0.976	0.853	2.65	0.008		AR Intensity	0.900	0.599	2.20	0.03
AR H Score	0.968	0.893	1.76	0.08		AR H Score	0.945	0.609	3.04	0.002

*Fisher r-to-z transformation.

DN, dominant nodule; HGG, highest Gleason grade.