Background
The tumor microenvironment plays a role in cancer cell proliferation, immune evasion, metastasis, and treatment resistance which is mediated by direct cancer cell contact and/or indirect cell signaling through cytokines, chemokines, and growth factors [
1,
2]. Successful targeting of T-cell immune checkpoint pathways has shown dramatic responses in this environment for multiple malignancies but has failed as a monotherapy in prostate cancer (PCa) [
3‐
9]. This may be due to the unique nature of the PCa tumor immune microenvironment [
10]. In PCa mouse models, the presence of an immunosuppressive B-cell subpopulation was associated with accelerated recurrence of castrate resistant PCa [
11] and suppressed the cytotoxic T-cell response normally associated with chemotherapy [
12]. In prostatectomy specimens, we have demonstrated previously that high B-cell density is associated with biochemical failure in high-risk patients [
11,
13]. While inhibition of B-cells may provide an alternative therapeutic approach, no clinical trials have attempted to modulate B-cells in the tumor microenvironment.
Rituximab is a well-tolerated monoclonal antibody against the CD20 antigen which is highly expressed on most B-cells. Rituximab was originally approved by the FDA for the treatment of Non-Hodgkin’s Lymphoma, with expanded indications to many non-malignant B-cell related diseases. In a PCa mouse model, anti-CD20 treatment decreased the number of tumor infiltrating B-cells and delayed the onset of castrate resistant disease [
11]. We sought to determine whether neoadjuvant treatment of high-risk PCa patients with the anti-CD20 immunotherapy, Rituximab, could reduce B-cell infiltration of PCa tumors.
Men with high-risk PCa were prospectively enrolled to undergo a single 1-month cycle of neoadjuvant rituximab prior to curative intent prostatectomy (NCT01804712). The primary objective was to determine the histological response of tumor infiltrating B-cells. Secondary objectives included assessing the impact on oncological outcomes, serum PSA response, immunohistochemical staining profiles of other immune cells, and observe patient safety and tolerability.
Discussion
Immunoregulation of the tumor microenvironment in human prostate cancer is not well understood [
1,
2,
15]. In the present study, we demonstrated that neoadjuvant rituximab significantly decreased B-cell density within tumors compared to concurrently assayed controls (p = 0.02) and appeared to reduce the density of tumor resident B-cells to levels comparable to adjacent non-neoplastic tissue, (p = 0.11 relative to controls). There was also a significant decrease in CD3+ T-cell density within the tumor regions demonstrating the inter-dependence between B and T-cells in prostate cancer. Neoadjuvant rituximab was well tolerated with no attributable serious adverse events. Therefore, these results provide evidence that rituximab is a safe modifier of the immune environment within prostate tumors.
Lymphocyte infiltration plays a central role in cancer cell apoptosis and has been linked to tumor stage and recurrence free survival [
2,
10,
14,
16]. Observational studies of B-cell density in PCa have had conflicting results, with some reporting lower B-cells counts [
17], while others described higher B-cell density only in high-grade tumors [
18]. The prognostic implications of B-cell density in PCa are also mixed. Petitprez and co-workers [
19] showed recently that a high density of CD8+ cells plus PD-L1+ cells was associated with higher risk of biochemical recurrence in lymph node-positive patients, but no association was found with CD20+ cells.
Our prior work has shown increased tumor infiltrating CD20+ B-cells were correlated with biochemical recurrence [
13], while others remark that B-cells have no association with biochemical recurrence [
20,
21]. All studies agree that B-cells are rare in PCa tumors, making the variations in results likely due to differences in analysis methods and lymphocyte quantification. To our knowledge this is the first trial to prospectively modulate B-cell density within PCa tumors. Pre-clinical work in PCa mouse models demonstrated a B-cell driven pathway that promoted castration resistance [
11,
12]. Furthermore, decreased or depleted B-cell numbers within the tumor were associated with inhibited tumor growth and delay of metastasis. The current study was a successful phase I translation of the mouse model, whereby rituximab was able to decrease B-cell density in the PCa tumor.
The majority of immunotherapies are based on augmenting CD8+ cytotoxic T-cells (CTL), which normally assist in eradicating tumor cells but can become quiescent over time [
22,
23]. Observational studies of PCa found that rare spontaneous regressions were associated with the lack of CD4+ T-regulatory cells while progression was associated with increased CD4+ T-regulatory cells, with no significant differences in CD4+ T-helper or CD8+ T-cytotoxic concentrations [
24,
25]. In animal models of advanced PCa, B-cell interactions were responsible for down-regulating T-cell dependent cytotoxicity, while B-cell depletion had no effect on the number of CD4+ Treg cells within the tumor [
12]. Following this result, we expected decreasing B-cells would have no effect on the number of T-cells in the treatment arm, and thus the decreased CD3+ density, a pan T-cell marker, came as a surprise. To verify our finding, all samples were re-analyzed by three blinded researchers, with the CD3+ density staying within 5% of the original findings. In different patient populations, rituximab has shown a range of T cell responses from no change in T-cell concentration [
24], to an increase in all T-cell subpopulations [
26], or dramatic decrease in CD4+ Treg cells [
27]. These studies highlight that enumerating lymphocytes is often not sufficiently informative about functional immune interactions. The clinical implications of our findings are as yet unknown, as PCa emergence may be related to unique inhibitory pathways [
28]. Future studies will concentrate on defining specific lymphocyte and monocyte subpopulations and associated signaling and to better understand B and T cell interactions in prostate tumor regions compared to benign regions.
Rituximab was chosen given its efficacy and safety profile. Initially approved in 1997 for the treatment of relapsed or refractory B-cell Non-Hodgkin’s Lymphoma (NHL), rituximab causes rapid depletion of circulating and tissue-based B-cells generally within the first three doses, with sustained depletion for up to 6–9 months after treatment [
29]. Rituximab has subsequently been approved for several additional indications including chronic lymphocytic leukemia (CLL), rheumatoid arthritis, Wegener’s granulomatosis and microscopic polyangiitis. Common adverse reactions observed in clinical trials of lymphoid malignancies were infusion reactions, fever, lymphopenia, neutropenia, chills, infection and asthenia (≥ 25%) [
30]. We noted four Grade 1 and 2 adverse events that were probably related to rituximab (infusion reaction and fatigue) occurring in three of the eight patients. Both infusion related reactions were able to be dose increased and complete the full cycle. This suggest a well-tolerated profile in the prostate cancer population.
The strengths of this study are the prospective design with rigorous, blinded assessment of B-cell density in tumor and adjacent tissue. Since the study started, two case reports have described treatment of metastatic PCa patients with rituximab. The first was a 66-year-old male who developed metastatic PCa; a lesion biopsy disclosed prominent CD20+ nests. He was treated with rituximab which resulted in a significant PSA decline [
31]. The second was a 79-year-old metastatic PCa patient with a concurrent B-cell lymphoma. On treating his lymphoma with rituximab, he had a dramatic decline in PSA and his skeletal lesions showed signs of resolving [
32]. To our knowledge, other than these reports, no other prospective studies of B-cell densities have been conducted.
This study adds to the rapidly evolving field of immunomodulating the tumor microenvironment of PCa. Other studies assessing are correlative retrospective tissue assessment [
24,
25] or are evaluations of circulating lymphoid cells [
33].
The current study is one of the few neoadjuvant immune trials to report results. Prior neoadjuvant studies either terminated or withdrew (NCT00577356/NCT01197209) or have not published results. Gao et al. [
28] demonstrated that neoadjuvant ipilimumab increased CD4+ and CD8+ cells within prostatectomy specimens, but did not mention effects on CD20+ cells.
Weaknesses include unknown impact on clinical outcomes, or impact of T-cell modulation. However, the primary outcome of the study was met, and the study design was not powered for secondary outcomes. Because of the small number of patients in this study it is hard to generalize our results, however this is the only prospective assessment of B-cell modulation in prostate cancer. Our method of lymphocyte annotation is not widely used, but we have published experience using this imaging quantification method for B cells [
13]. An important difference between the animal models that served as the background work for the trial design was that they used castrated PCa mouse models, whereas this patient cohort did not receive androgen deprivation therapy.
Recently, three studies showed that the presence of B cells in tumor regions in patients with metastatic melanoma or sarcoma who were treated with immune check-point blockade (ICB) immunotherapy were predictive of good prognosis [
34‐
36]. In particular, B cells that were in tertiary lymphoid structures (TLS) which contain aggregates of multiple immune cell types, were present in responding patients. B cells, therefore, have both positive and negative roles in tumor immunology, and thus, it is critical to elucidate the precise tumor infiltrating B cell and T cell states, or types, and their interactions within TLSs not only for prognosis but also to tailor immunotherapy regimens and improve the anti-tumor response in prostate cancer.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.