Severe sepsis, defined as infection complicated by organ dysfunction, is a major public health problem. The organ dysfunction occurs due to a dysregulated immune response. For several years, it was assumed that the dysregulation was predominantly hyper inflammatory; but a large number of trials that blocked the hyper inflammatory response failed to improve outcomes. Prior studies suggest that many septic patients have immunosuppression of both innate and adaptive immune response, which is associated with increased risk of organ dysfunction, secondary infection, and mortality. The mechanisms of this immunosuppression bear similarities to the immune compromise seen in cancer. In particular, over activation of immune checkpoint proteins, particularly up regulation of programmed death 1 (PD-1) protein, and one of its ligands, PD-L1, may play important roles in both diseases. In preliminary work using animal models of sepsis, we and others showed that anti-PD-L1 therapy reduced organ dysfunction, restored protective immune response, and improved survival. Meanwhile, a randomized trial of BMS-936559, an anti-PD-L1 monoclonal antibody, demonstrated good drug safety and tolerability and improved outcomes in patients with cancer. Our goal, therefore, is to test anti-PD-L1 therapy in humans with severe sepsis. We have established a collaboration consisting of academic experts in sepsis and in adaptive clinical trial design and investigators from Bristol Myers Squibb (BMS) to conduct a phase 2, multicenter, randomized clinical trial of BMS-936559 in severe sepsis patients. We will understand the mechanisms through which anti-PD-L1 will improve outcomes, personalize therapy, and test the safety and identify the optimal dose. Since it is unclear which patients will respond to anti-PD-L1 therapy, we propose to use an adaptive enrichment trial design. We will start enrollment in a broad population of severe sepsis patients using two doses of anti-PD-L1 or placebo and measure biomarkers of immunosuppression prior to administering anti-PD-L1. During the trial, we will change the enrollment criteria according to pre-specified decision rules based on response to anti-PD-L1 in the biomarker-identified subgroups, enriching for subgroups that benefit and dropping those that are show no benefit or are harmed. Additionally, we will use response-adaptive randomization to alter the randomization probabilities such that patients will be preferentially assigned to the better-performing dosage arm. We will also measure biomarkers during the course of therapy to provide insight into the mechanisms. Adaptive trials require extensive simulations to finalize the decision rules for enrichment and response adaptive randomization. This proposal is a planning grant to execute detailed simulations to finalize the trial design, the study logistics, and the oversight procedures. In summary, the phase 2 trial will have important implications to understand the mechanisms through which immune checkpoint proteins improve outcomes and to develop novel trial designs for sepsis.
Severe sepsis is a major health problem and few therapies have improved outcomes of these patients. We will use a novel approach to test a new treatment that has improved outcomes in patients with cancer. If successful, this treatment will hasten recovery and save thousands of lives each year.
