Prostate cancer is the most common non-cutaneous malignancy and the second-leading cause of cancer-related mortality among men in the USA, with distant disease having a 5-year survival rate of 29.8% [1]. Though metastatic disease is initially responsive to androgen blockade, over time this treatment selects for a castration-resistant prostate cancer (CRPC) population with modern systemic treatments providing a median survival of 2.8 years, albeit with significant quality of life detriment due to treatment-related effects and disease progression [2].

The combination of immunotherapy and radiotherapy is an emerging clinical treatment paradigm, a growing research sector, and a critical research domain supported by the Radiation Biology Task Force [3]. X-ray radiation treatment (XRT) can activate both the adaptive and innate immune systems through directly killing tumor cells, causing mutations in tumor-derived peptides, and causing localized inflammation that increases immune cell trafficking to tumors [4, 5]. Most importantly, the activated immune system may cause tumor-directed treatment responses away from the site of irradiation, i.e., an abscopal treatment effect, which has the potential to treat disease throughout the body.

However, prostate tumors are considered poorly responsive to immunotherapy due to their low genetic mutational load, their lack of activated tumor-infiltrating lymphocytes, and specific genetic alterations that influence the immune landscape [6, 7]. Studies have shown over 50% of aggressive prostate cancers express high levels of PD-L1, a key factor in suppressing the local immune response [8]. A negative regulator of the immune response, Tregs have also been found to be enriched in both the tumor and peripheral blood of patients with prostate cancer [9, 10]. Altogether, prostate cancer has mechanisms to evade and inhibit anti-tumor immunity.

Clinical trials have studied immune checkpoint inhibition for CRPC. One phase III randomized trial of 799 patients tested 8 Gy XRT to a CRPC bone metastasis followed by either placebo or 4 cycles of ipilimumab (anti-CTLA4) and found the combination provided a statistically significant 7 mo median survival increase in a post-hoc analysis of a predefined subgroup of patients with low tumor burden (22.7 mo vs 15.8 mo, p = 0.0038) [11]. However, only a trend for improved overall survival was seen for the whole cohort (11.2 mo vs 10.0 mo, p = 0.053) and therefore the trial did not provide enough evidence to meet its primary endpoint and influence clinical practice [11]. A subsequent phase III randomized clinical trial for 600 patients with low tumor burden tested ipilimumab versus placebo, without XRT [12]. However, this ipilimumab-only approach failed to show any overall survival benefit and only a marginal progression-free survival benefit and prostate-specific antigen (PSA) response was seen, suggesting that the combination with XRT produces a superior treatment response in patients with low disease burden [12]. These large randomized clinical trials indicate there is a potentially powerful treatment approach when combining radiotherapy with immunotherapy for CPRC, but the optimal treatment combination has not yet been found for most patients to derive benefit.

This project builds upon the findings of these clinical trials to develop preclinical models that can be used to optimize the treatment approach. Anti-PD-1 and anti-PD-L1 antibodies are immune checkpoint inhibitors that target tumor-immune cell interactions and clinically have a reasonably favorable side-effect profile in patients. This suggests PD-inhibitors may be superior to anti-CLTA4 agents, which primarily block the interaction between immune cells without directly involving the tumor. However, PD-1 agents alone show little response in treating CRPC in early phase clinical trials [13]. Nevertheless, logically following the clinical trials described above, we hypothesized that combination PD-based immunotherapy-radiotherapy (iRT) approach would trigger a robust treatment response against CRPC that is mediated through the immune system, causing both local and distant (abscopal) effects, while likely being better tolerated in patients than an anti-CTLA4 approach.

There is evidence to suggest that the tumor-dependence on PD-1/PD-L1 immunosuppression is enhanced in lesions that respond to radiation [14]. Therefore, we examine a combination of immune checkpoint inhibition and radiotherapy for CRPC that causes local and abscopal treatment effects mediated by activated immune cells.

We developed a PD-based iRT approach for CRPC in an immunocompetent castrated syngeneic FVB mouse model using subcutaneous Myc-CaP tumor grafts [16, 18]. Expression of PD-L1 in Myc-CaP cells increases after irradiation (Additional file 1: Figure S1). Compared to mice treated with antibody alone, XRT (20 Gy in 2 fractions) to the leg tumor graft causes a local response in the irradiated tumor and a robust abscopal effect with regression of an unirradiated distant tumor graft (Fig. 1a and b). At 21 days, the mean graft volume for anti-PD-1 alone was 2094 mm³ (N = 18 grafts) compared to iRT irradiated grafts 726 mm³ (N = 9 grafts) (p = 0.04) and unirradiated grafts 343 mm³ (N = 9 grafts) (p = 0.0066). At 17 days, the mean graft volume for anti-PD-L1 alone was 1754 mm³ (N = 16 grafts) compared to iRT irradiated grafts 284 mm³ (N = 8 grafts) (p = 0.04) and unirradiated grafts 556 mm³ (N = 8 grafts) (p = 0.21). No significant differences were observed between the leg and flank graft volumes within each treatment group, so both grafts were included in the antibody alone data. Additional tumor graft volume data is in Additional file 1: Figure S2.Fig1Fig. 1

Castration-resistant prostate cancer is successfully treated by immune checkpoint inhibitor combined with radiotherapy, with effects on the irradiated and unirradiated tumors, and increased survival. a-d. Myc-CaP tumor graft volumes (a and b) and survival (c and d) for mice treated with immune checkpoint inhibitor monotherapy and given in combination with XRT to the leg graft. Significantly decreased tumor graft volume and significantly increased median survival was observed. Error bars represent ± SEM

Although clinical data suggests limited effects of immune checkpoint inhibitor treatment for CRPC, this preclinical model indicates robust responses are achievable using when combining anti-PD-1 or anti-PD-L1 treatment with XRT. A syngeneic mouse model was selected to allow treatment effects to be studied in the presence of an intact immune system. The highly-aggressive Myc-CaP model in the castration-resistant setting was selected to investigate treatment efficacy.

Tumor graft growth was significantly diminished by the combination treatment of immune checkpoint inhibitor and XRT compared to drug alone. Remarkably, unirradiated distant tumor grafts also responded to combination treatment, suggesting an abscopal treatment effect. Most importantly, significant increases in median survival were observed compared to antibody treatment alone: 70% longer for anti-PD-1 and 130% for anti-PD-L1. Importantly, no increased toxicity was observed for combination immuno-radiotherapy treatment compared to monotherapy. However, a notable limitation of this preclinical model is that the combination treatment was not found to be durable after a single treatment cycle (8 days), with no mice completely clearing their tumor grafts. It is possible that repeat dosing by immune checkpoint inhibitor would extend the treatment effect, as found in clinical studies using immune checkpoint inhibitors, but this was not investigated in this preclinical model. Furthermore, additional treatment combinations are currently being tested to determine the best approach, including varying the timing/sequencing of therapies and the radiation dose/fractionation.

To further understand the mechanism for decrease in tumor growth resulting from combination XRT and anti-PD-L1 antibody treatment, flow cytometry was used to characterize the tumor immune microenvironment. When analyzing only live cells, there was a higher percentage of CD8+ cytotoxic T cells in the tumors of control mice compared to flank tumors from mice that received systemic anti-PD-L1 antibody treatment and radiation treatment to the leg tumors. However, flow cytometry showed strong differences in activation between the tumor infiltrating lymphocytes in the control group compared to treated mice. Both flank and leg tumors from treated mice had significantly more CD8+ cytotoxic tumor infiltrating T cells expressing PD-1. Additionally, the XRT-treated leg tumors showed a significantly higher percentage of CD8+ cytotoxic T cells expressing CD44, a marker of T cells that are active after antigen presentation. Although the decreased T cell infiltration in treated tumors does not indicate a mechanism for decreased tumor growth in mice treated with radiation and anti-PD-L1 antibodies, the differences in activation can potentially account for these differences. The increased expression of both PD-1 and CD44 suggests that the tumors from mice treated with radiation and anti-PD-L1 are experiencing increased rates of tumor antigen presentation, which could be one mechanism for decreased tumor growth in the treated mice. It is also possible that CD335+ NK cells play a role in the tumor microenvironment, as supported by the flow cytometry data indicating an increase in the mice treated with anti-PD-L1 and XRT. Lastly, the survival advantage is lost when blocking CD8 in the mice, suggesting a key mechanistic role for CD8+ cells in the immune response. Additional mechanistic roles of the immune cells are being investigated, since the immunity triggered by combination immune checkpoint and radiotherapy is complex. [20]

Emerging clinical data indicates about 3% of patients with prostate cancer have a high tumor mutation burden (microsatellite instability-high or mismatch repair deficit) and they are responsive to anti-PD-1/PD-L1 agents, with 45% (5 of 11 patients) experiencing durable clinical benefit [21]. National Comprehensive Cancer Network guidelines for metastatic CRPC include consideration of testing tumor mutation burden and second-line treatment by pembrolizumab. As clinical trials develop to test PD-agents for prostate cancer treatment, it is important to recognize that an immune checkpoint treatment combined with radiotherapy may provide an even greater response rate than monotherapy. The preclinical model presented herein provides a framework for further investigating the optimal approach for combining radiotherapy and PD-agent that can be carried into future clinical trials.

Using an immune-intact mouse model for the important clinical entity CRPC, survival is dramatically improved by 70–130% when radiotherapy is combined with anti-PD-1 or anti-PD-L1 immune checkpoint inhibitor, respectively, compared to monotherapy. The immuno-radiotherapy treatment response mechanism involves CD8+ cells, suggesting activation of the immune system that is not observed with monotherapy. An abscopal treatment effect was observed for an unirradiated tumor distant from an irradiated one in the same animal, suggesting the potential for the immune system treating widespread metastatic disease. These data provide strong preclinical evidence for a combination treatment approach for CRPC using radiotherapy and immune checkpoint inhibitor, which can inform the design of future clinical trials.