Development of a mRNA Lipid Nanoparticle (mRNA-LNP) Cancer Vaccine to Prevent Leukemia Relapse after Stem Cell Transplant

Introduction The most common acute leukemia in adults is acute myeloid leukemia (AML). Approximately 20,050 new cases and 11,540 deaths are expected for 2022 in the US. Allogeneic stem cell transplant (alloSCT) remains the only cure for high-risk AML. AlloSCTs leverage the immune system to kill leukemic cells through a graft-versus-leukemia (GVL) response mediated by donor CD4+ and CD8+ T cells that recognize recipient leukemic minor histocompatibility antigens (mHAs) as foreign. Unfortunately, transplantation failure related to relapse is common. We postulate that proper targeting of mHAs could enhance leukemia-specific immune responses without off-target effects, improving relapse rates. Using computational modeling, we predicted a series of GVL mHAs, confirmed our first GVL mHA (UNC-GRK4-V) through targeted mass spectrometry, and generated a high-affinity T cell receptor (TCR) against it. Using UNC-GRK4-V as our target, we developed a novel mHA mRNA lipid nanoparticle (LNP) vaccine that we hypothesized could induce an antigen-specific immune response.

Methods A custom mRNA was ordered from TriLink that encoded the UNC-GRK4-V peptide linked to an MHC Class I trafficking domain (MITD) and was modified with a 5' cap and polyadenylated tail. We created mRNA LNPs by mixing an ethanol lipid solution made of ionizable lipid (SS-OP, SM-102, or MC3)/ DOPE/Cholesterol/DMG-PEG 2000 with an aqueous solution of mRNA (eGFP or UNC-GRK4-V mRNA) dissolved in a citrated buffer solution. We measured mRNA LNP sizes with a Malvern ZS90 and quantified encapsulation rates using a Quant-it RiboGreen RNA assay kit. Immature monocyte-derived dendritic cells (mo-DCs) were differentiated from monocytes collected from fresh blood and transfected with mRNA LNPs. As a control, we measured eGFP mRNA LNP transfection rates of mo-DCs using flow cytometry. CD8+ T cells were transduced with a retroviral supernatant containing a high-affinity UNC-GRK4-V T-cell receptor and stimulated. We co-cultured mRNA LNP transfected mo-DCs with transduced CD8+ T cells. IFN-g was measured from the co-culture supernatant using IFN-g ELISA assays. We used flow cytometry to determine the cytotoxic killing of transfected mo-DCs by transduced CD8+ T cells.

Results To identify the most efficient ionizable lipid platform for mRNA delivery and transfection, we tested three commercially available ionizable lipids: SS-OP, SM-102, and MC3. We found that all three ionizable lipids were easily reproducible with SS-OP 75.2nm ± SD 14.8, MC3 97.0nm ± SD 5.4, SM-102 102.1nm ± SD 0.651. The encapsulation rates of mRNA were 75.4% ± SD 6.3% for SS-OP, 70.5% ± SD 2.6% for MC3, and 73.7% ± SD 0.1% for SM-102. The transfection rate of mo-DCs using eGFP mRNA LNPs were 0.48%, 19.3%, and 62.5% for SS-OP, MC3, and SM-102, respectively. Based on poor transfection rates with SS-OP, we used MC3 and SM-102 LNP formulations to study cytokine expression and cytotoxic killing using a co-culture model of transfected mo-DC with transduced CD8+ T cells. We found a 16.8-fold increase in IFN-g release for the SM-102 based LNPs compared to control empty LNPs (p < 0.0001) and a 4.1-fold increase in IFN-g release for the MC3 based LNPs compared to control empty LNPs (p < 0.0001). We then evaluated SM-102 LNPs for cytotoxic killing as they demonstrated the highest IFN-g release. SM-102 demonstrated increased cell death of 13.5% versus 3.5% for control (empty SM-102) after being normalized for background mo-DC death.

Conclusions The SM-102 based UNC-GRK4-V mRNA LNP vaccine elicits robust, antigen-specific T-cell responses and could be a potent therapeutic for decreasing relapse rates post-alloSCT. Furthermore, advances in alloSCT have created regimens that eliminate traditional long-term calcineurin inhibitor use in post-alloSCT, allowing therapeutic cancer vaccines to be given much earlier after transplant.

No relevant conflicts of interest to declare.