Search for targets of anti-cancer immunity in patients receiving an off-the-shelf cancer's dark matter vaccine, DPV-001, combined with anti-PD-1 +/- anti-GITR in head and neck squamous cell cancer (HNSCC)

Background: Despite advances in cancer genomics and immunotherapy, most cancer vaccine strategies target a limited set of canonical or mutated cancer antigens, capturing only a fraction of tumor antigenicity. Malignant cells generate a vast, largely unexplored repertoire of non-canonical “dark matter” antigens arising from the dark genome. Many of these antigens are frequently absent from normal tissues and enriched by and possibly responsible for oncogenic dysregulation. We have developed a vaccine strategy, DPV-001, that includes canonical and more than 400 dark matter cancer antigens and have evaluated DPV-001 in a combination immunotherapy trial for patients with HNSCC. This phase 1b trial tripled response rates over historical experience with anti-PD-1 and encouraged identification of the antigens targeted by the therapeutic anti-cancer immune response. Here we report the strategy we are employing and preliminary results on identifying targeted cancer antigens from patients on this trial. Methods: Eligible patients were randomized 1:1 to receive DPV-001 with or without a GITR agonist antibody (INCAGN-1876), with all patients receiving sequential PD-1 blockade (retifanlimab) starting day 15. Patient sera were analyzed by HuScan™ phage display immunoprecipitation (PhIP), covering 29,371 human proteins and splice isoforms. Antibody development was used as a surrogate for antigen-specific T cell responses. The canonical and dark matter cancer antigens in DPV-001, and an HNSCC HLA peptidome database, are being used to focus investigations of targets recognized by patients that received DPV-001. Results: PhIP detected Ab was assessed in 11 patients, including 7 patients from Arm 2 (received GITR). One-week post-infusion GITR serum levels were characterized as low (Avg 163 reads/million) in 3 patients and high (avg 585 reads/million) in 4 patients (p < 0.002). There were no significant differences between GITR Low and GITR High groups when comparing their smoking history, CPS score, or body weight. At baseline, patients in the GITR Low group had significantly fewer strong Ab signals (Avg 148) to the NCBI 35.1 proteome targets (>10x beads only maximum signal) compared to the GITR High group (avg 596, p=0.002) or Arm 1 (No-GITR, Avg 406, p=0.042). Characterization of B and T cell responses to HNSCC antigens is ongoing. Conclusions: This 18 pt trial of DPV-001 and sequenced PD-1 +/- GITR shows a promising RR. Identification of the antigens targeted in patients with objective clinical responses may enable development of a new generation of cancer vaccines.