Unfortunately, the combination therapy did not extend OS for men with this disease, but there are lessons learned on how to develop immune checkpoint combination strategies for men with metastatic castration-resistant prostate cancer.

There are many possible reasons why the clinical trial failed.

• Prostate cancer has very few T cells, so targeting PD-1 or PD-L1 in a microenvironment where there are few T cells expressing these pathways may not lead to antitumor responses.
• There are multiple immunosuppressive pathways within the prostate tumor microenvironment, so the PD-1- or PD-L1-targeting agents may not target all these pathways.
• Prostate cancer has few mutations. As a result, the effector T cells may not be able to recognize an adequate number of antigens to induce an antitumor response.
• The impact of enzalutamide on the immune response in the tumor microenvironment is not yet known. It is unclear whether enzalutamide led to an increase of infiltrating T cells within the prostate cancer tumor environment.
• These patients had multiple prior therapies, which could have left them with poor immune responses when given this combination.

A hot tumor, which we see in melanoma, has lots of infiltrating T cells. So, blocking PD-1 or PD-L1 inhibitory pathways will lead to antitumor responses and clinical benefit. Cold tumors, such as those in prostate cancer, do not have many infiltrating T cells. So, treatment with anti-PD-1 or PD-L1 does not make sense. Combination therapy — where we drive T cells into the prostate tumor microenvironment and then block the inhibitory pathway, such as PD-1 or PD-L1 — may be the way to move forward.

Unfortunately, we don’t have data on how these drugs affect the microenvironment. Is it feasible to test agents for their impact on the microenvironment before starting phase 3 trials?

Access to tumor tissues is critical to help understand how the immune response is affected by different agents. Immune responses in the blood are not equal to immune responses in the tumor microenvironment. Changes in the tumor microenvironment could help guide combination strategies.

We need to know the impact of anti-PD-1 or anti-PD-L1 agents in the prostate tumor microenvironment. Is there an increase in T-cell repertoire tumor-infiltrating T cells? Do these T cells recognize prostate tumor antigens or mutations? What compensatory inhibitory pathways occur after anti-PD-1 or anti-PD-L1 monotherapy? These studies will be critical to move anti-PD-1 or anti-PD-L1 forward in the setting of prostate cancer.

Also, what is the rationale for combination therapy with anti-PD-1 or anti-PD-L1 plus enzalutamide? Although there have been small studies with enzalutamide, we do not know how the agent impacts the prostate tumor microenvironment.

We now know the biologic mechanisms of immune checkpoint therapy require tumor-infiltrating effector T cells for these therapies to work. We should be evaluating whether agents that are being tested can increase this. We also need to study longitudinal samples to figure out the rational combination that can target compensatory inhibitor pathways.

Biomarkers will be needed for patient selection. We need combinational biomarkers to reflect the tumor status and immune response.

Finally, successful combination strategies will require in-depth studies to understand the impact of each agent and its effect on the tumor microenvironment before we select them for phase 3 studies.

Advanced androgen-targeted therapies with agents such as enzalutamide have significantly prolonged survival and reduced skeletal morbidity in metastatic castration-resistant prostate cancer (mCRPC); however, durable remissions remain elusive. Although immune checkpoint inhibition has been successful in achieving durable remissions in a number of solid tumors, results in mCRPC have been disappointing. An experimental observation that PD-L1 was upregulated at the time of resistance to enzalutamide led to the hypothesis that a combination of these agents would improve outcomes. Unfortunately, in the randomized phase 3 IMbassador250 trial, the addition of atezolizumab failed to show any benefit over enzalutamide monotherapy. Sadly, the addition of immunotherapy came with an increased risk of toxicity, including three fatal immune-related adverse events. There were no unusual toxicities, and the rates of immune-related adverse events were not greater than what has been seen in other trials of checkpoint inhibitors in other solid tumors, but these risks should inspire caution when considering off-label use of immune checkpoint inhibitors.

Immunotherapy in mCRPC has a checkered past, with favorable results in phase 2 studies failing to persist in subsequent randomized studies. Whether with vaccines such as GVAX-PCa (Biosante) or immune checkpoint inhibitors such as ipilimumab (Yervoy, Bristol-Myers Squibb), it has been clear that some men with mCRPC benefit, but they represent a very small percentage. Until we can discover the defining tumor and/or host characteristics of those who respond, large studies in unselected populations are likely to fail. Furthermore, prostate cancer generally has a “cold” tumor microenvironment, and combination strategies likely will need to find a means of drawing the immune system into the tumor microenvironment. Combination studies with systemic radiation agents — such as radium Ra 223 dichloride (Xofigo, Bayer) and 177-Lu-PSMA-617 (Endocyte) — are underway; they are attempting to leverage immunogenic cell death from radiation as a signal to draw in the T cells. Given the unique success of the autologous activated cellular immunotherapy approach of sipuleucel-T (Provenge, Dendreon), additional cellular immunotherapy studies also are of considerable interest. Chimeric antigen receptor T-cell studies are underway at several institutions, as are studies of bispecific T-cell engaging antibody therapy. Immunological insights from these trials will emerge as studies of tissue and circulating immune responses are evaluated, which will hopefully guide future trial development in mCRPC.