Neoadjuvant rituximab modulates the tumor immune environment in patients with high risk prostate cancer

Abstract

Background: Immunotherapeutic regulation of the tumor microenvironment in prostate cancer patients is not understood. Most antibody immunotherapies have not succeeded in prostate cancer. We showed previously that high-risk PCa patients have a higher density of tumor infiltrating B-cells in prostatectomy specimens. In mouse models, anti-CD20 antibody ablation of B-cells delayed PCa regrowth post-treatment. We sought to determine whether neoadjuvant anti-CD20 immunotherapy with rituximab could reduce CD20+ B cell infiltration of prostate tumors in patients.

Methods: An open label, single arm clinical trial enrolled eight high-risk PCa patients to receive one cycle of neoadjuvant rituximab prior to prostatectomy. Eleven clinical specimens with similar characteristics were selected as controls. Treated and control samples were concurrently stained for CD20 and digitally scanned in a blinded fashion. A new method of digital image quantification of lymphocytes was applied to prostatectomy sections of treated and control cases. CD20 density was quantified by a deconvolution algorithm in pathologist-marked tumor and adjacent regions. Statistical significance was assessed by one sided Welch's t-test, at 0.05 level using a gatekeeper strategy. Secondary outcomes included CD3+ T-cell and PD-L1 densities.

Results: Mean CD20 density in the tumor regions of the treated group was significantly lower than the control group (p = 0.02). Mean CD3 density in the tumors was significantly decreased in the treated group (p = 0.01). CD20, CD3 and PD-L1 staining primarily occurred in tertiary lymphoid structures (TLS). Neoadjuvant rituximab was well-tolerated and decreased B-cell and T-cell density within high-risk PCa tumors compared to controls.

Conclusions: This is the first study to treat patients prior to surgical prostate removal with an immunotherapy that targets B-cells. Rituximab treatment reduced tumor infiltrating B and T-cell density especially in TLSs, thus, demonstrating inter-dependence between B- and T-cells in prostate cancer and that Rituximab can modify the immune environment in prostate tumors. Future studies will determine who may benefit from using rituximab to improve their immune response against prostate cancer. Trial registration NCT01804712, March 5th, 2013 https://clinicaltrials.gov/ct2/show/NCT01804712?cond=NCT01804712&draw=2&rank=1.

Keywords: CD20; CD3; Immunotherapy; Neoadjuvant; PD-L1; Prostate cancer; Prostatectomy; Rituximab; Tumor infiltrating lymphocytes (TILs).

Fig. 1

Immunohistochemical staining of serial prostatectomy sections showed presence of CD20+ B-cells cells aggregated in immune cell foci known as tertiary lymphoid structures (TLSs). a Representative AperioScope scanned image of anti-CD20 stained prostatectomy section counter-stained with hematoxylin. Tumor regions outlined by pathologist markings in blue, T = tumor, NT = non-tumor, ×10 magnification. b Box inset enlarged at ×200 magnification shows CD20+ B-cells stained brown in bright-field. c Post-deconvolution image of CD20 staining. After Imagescope deconvolution algorithm is run the stained color intensity is represented as image pixels with high intensity (brown), intermediate (orange) and low (yellow) staining intensity

Fig. 2

Primary outcome analysis of B cell density in prostatectomy tissue showed decrease in B-cell density in tumor regions after neoadjuvant Rituximab treatment. a Boxplot and scatter plot for B cell density (Treated vs. Control), B-cell density in treated: 0.027 mm²/mm² (95% CI 0.021, 0.033) and 0.044 mm²/mm² (95% CI 0.028, 0.062) in the controls, p = 0.02 for Welch two sample t-test comparing B-cell density of Tumor in treated group and historic control group, i.e. The mean B cell density in tumor was significantly lower in treated subjects compared to historical control subjects. C = control samples, P = PROTUX Rituximab-treated, T = marked tumor region, N = adjacent normal tissue. b Within subject difference of B-cell density: tumor minus adjacent normal tissue (Treated vs. Controls)

Fig. 3

Secondary outcome analysis: serum PSA from Day 1 to Day 29 not significantly changed in Rituximab treated patients. Serum PSA did not appreciably change over the course of neoadjuvant rituximab (Day1: 8.1 ng/mL (± 4.3), Day 29: 8.36 ng/mL (± 4.88), p = 0.84)

Fig. 4

Secondary analysis showed decrease in CD3 + T-cell density. a CD3 plot Boxplot and scatter plot for CD3 density (Treated vs. Control), p = 0.995 for Welch two sample t-test comparing CD3+ T-cell density of Tumor in treated group and historic control group, i.e. the mean CD3+ T-cell density in tumor was significantly lower in treated subjects compared to historical control subjects. Mean CD3+ T-cell density in the tumor region of the Rituximab-treated group was 0.0223 mm²/mm² (95% CI 0.0116, 0.033) and 0.0419 mm²/mm² (95% CI 0.0305, 0.0533) in the control group. b Within subject difference of CD3 density: tumor minus adjacent normal tissue (Treated vs. Controls)

Fig. 5

Secondary analysis of PD-L1 immunohistochemical staining density. a Boxplot and scatter plot for PD-L1 density (Treated vs Control), mean PD-L1 density in the treated (0.0589) versus control samples (0.0833) was observed but was not significant: p = 0.245 for Welch two sample t-test comparing PD-L1+ cell density of tumor in treated group and historic control group, i.e. the mean PD-L1+ cell density in tumor was significantly lower in treated subjects compared to historical control subjects. b Within subject difference of PDL1 density: tumor minus adjacent normal tissue (Treated vs. Controls)