Search for a command to run...
Abstract Introduction Durable responses remain rare in patients with Acute Myeloid Leukemia (AML) who are ineligible for intensive chemotherapy and stem cell transplantation. Although immunotherapies have transformed treatment outcomes in several cancer,s its impact in AML has been hampered by lack of efficacious immunogenic antigens. One type of emerging immunotherapies, therapeutic cancer vaccines, aims to direct the immune system against tumor-specific mutations. These vaccines have shown encouraging results in solid tumors, but their efficacy in AML is limited by low tumor mutational burden and availability of suitable targets. Recently endogenous retroviruses (ERVs) have been identified as highly and selectively expressed in AML blasts, thus introducing a novel group of antigen targets for therapeutic intervention (Alcazer et al, 2022). ERVs are remnants of ancient infectious viruses that integrated into the human genome. In healthy tissues, expression of ERVs is silenced or kept at low levels through strict epigenetic regulation. Hence, ERVs are annotated as part of the dark genome. However, due to the AML blast-preferential expression pattern, ERVs represent an underexplored opportunity for antigen-targeted immunotherapy in AML. Aim Here, we describe the discovery of ERVs expressed and shared across the AML patient population and our AI enabled identification of the optimal antigens presented on the Major Histocompatibility Complex (MHC). Moreover, we outline the design of an off-the shelf therapeutic cancer vaccine and demonstrate potent immune recognition in vitro and in vivo. Lastly, we outline the clinical development strategy for first-in-human testing in patients with AML. Methods & Results Genomic and transcriptomic data from AML blasts was sourced from dbGAP. To identify AML specific ERVs, so-called ERV-derived antigens, and to design an off-the-shelf vaccine, the curated data was served to our bioinformatic model, ObsERV™ (ref). ObsERV™ integrates ERV-derived antigen sequences with MHC population allele frequencies and patient tumor expression data and operates in three key steps: first it identifies ERVs expressed across the patient population; second it identifies ERV-derived subsequences with a high density of predicted MHC presented antigens referred to as “immunogenic hotspots”, and third, it filters all hotspots from ERVs expressed in healthy tissue to ensure tumor specificity. The final output of ObsERV is an optimized cancer vaccine comprising the predicted immunogenic hotspots that collectively maximize coverage of tumor-specific MHC-presented ERV antigens across patients. We identified a large number of ERVs expressed in AML and designed a novel AML vaccine, EVX-04, based on the curated data from 136 AML patients. EVX-04 was validated using both a hold-out data set and an independent validation data set. The vaccine contained at least eight MHC presented antigens present in all 15 AML patients in the hold-out dataset and in 49 out of 52 AML patients in the validation set. Preclinical evaluation The ability of the ERV-derived immunological hotspots in the EVX-04 vaccine to induce an antigen-specific T-cell response was tested by in vitro stimulation of human healthy donor peripheral blood mononuclear cells (PBMCs) with ERV antigen peptides from the hotspots and measurement of activated T cells using IFNg ELISpot analysis. The in vitro studies demonstrated that the selected ERV immunogenic hotspots induce significant antigen-specific T-cell responses in human PBMCs. Additionally, a murine surrogate EVX-04 cancer vaccine was evaluated in syngeneic mouse models of cancer. These experiments showed that prophylactic immunization with the surrogate EVX-04 vaccine potently inhibited tumor growth and elicited ERV antigen-specific immune recognition in vaccinated mice. Future Directions Despite improved response rates with the addition of venetoclax to hypomethylating agents, durable remissions and cure remain rare in newly diagnosed AML patients ineligible for stem cell transplantation. To address this unmet need - and building on a strong preclinical foundation and a favorable safety profile from prior cancer vaccine trials (Khattak et al, 2024) - we plan to initiate a first-in-human study of EVX-04 in combination with azacitidine and venetoclax. We believe these novel targets and therapeutic modality may allow for increased durability of clinical responses and improved quality of life for AML patients.