Targeting neoantigens to APC-surface molecules improves the immunogenicity and anti-tumor efficacy of a DNA cancer vaccine

Abstract

Introduction: Tumor-specific mutations generate neoepitopes unique to the cancer that can be recognized by the immune system, making them appealing targets for therapeutic cancer vaccines. Since the vast majority of tumor mutations are patient-specific, it is crucial for cancer vaccine designs to be compatible with individualized treatment strategies. Plasmid DNA vaccines have substantiated the immunogenicity and tumor eradication capacity of cancer neoepitopes in preclinical models. Moreover, early clinical trials evaluating personalized neoepitope vaccines have indicated favorable safety profiles and demonstrated their ability to elicit specific immune responses toward the vaccine neoepitopes.

Methods: By fusing in silico predicted neoepitopes to molecules with affinity for receptors on the surface of APCs, such as chemokine (C-C motif) ligand 19 (CCL19), we designed an APC-targeting cancer vaccine and evaluated their ability to induce T-cell responses and anti-tumor efficacy in the BALB/c syngeneic preclinical tumor model.

Results: In this study, we demonstrate how the addition of an antigen-presenting cell (APC) binding molecule to DNA-encoded cancer neoepitopes improves neoepitope-specific T-cell responses and the anti-tumor efficacy of plasmid DNA vaccines. Dose-response evaluation and longitudinal analysis of neoepitope-specific T-cell responses indicate that combining APC-binding molecules with the delivery of personalized tumor antigens holds the potential to improve the clinical efficacy of therapeutic DNA cancer vaccines.

Discussion: Our findings indicate the potential of the APC-targeting strategy to enhance personalized DNA cancer vaccines while acknowledging the need for further research to investigate its molecular mechanism of action and to translate the preclinical results into effective treatments for cancer patients.

Keywords: APC-targeting; CCL19; DNA vaccine; cancer immunotherapy; neoantigens.

Figure 1

Design and in vitro characterization of APC-targeting pDNA constructs. (A) Schematic design of the pDNA construct and the encoded homodimeric fusion protein. Plasmid inserts containing the sequence of different APC-binding molecules, the dimerization module h1h4CH3, and 5 CT26-derived neoepitopes. GL linkers connect the three modules. The non-targeted control (NT_Neo5) and constructs harboring Fv αDEC205, Fv αClec9a, and Clec9a ligands had inserted secretion signals upstream the first coding element. Created with BioRender.com. (B) In vitro expression and molecular weight characterization of the APC-targeting fusion proteins by immunoblotting against the CH3 element in the supernatant of transfected HEK293T cells under reducing conditions. Representative of 3 independent experiments.

Figure 2

Tumor rejection and T-cell responses elicited by APC-targeting pDNA vaccines harboring different APC-binding molecules. BALB/c mice received five weekly immunizations of 5 μg pDNA in a prophylactic setup. Two weeks after the first immunization, mice were inoculated with 5 x 10⁵ CT26 tumor cells s.c. in the right flank (n =13 mice per group). The studies were terminated 20 days after tumor inoculation, and the spleen compartments were analyzed for TNFα and IFNγ-producing CD8+ and CD4+ T cells by ICS (n = 3-5 mice per group). The complete gating strategy used to identify specific the T-cell subsets is exemplified in Supplementary Figure 7B . (A, E) Mean tumor volume (mm³) ± SEM over time for each treatment group with LOCF. (B, F) Area under the curve (AUC) of individual tumors split by group. Mean ± SD. (C, G) Frequency of reactive splenic CD8+ T-cell and (D, H) CD4+ T-cells. Mean ± SD. Statistics: (B–D) APC_Neo5 groups were compared to the Mock DNA and NT_Neo5 groups by Kruskal-Wallis test and Dunn´s multiple comparison test. (F–H) APC_Neo5 groups were compared to Mock DNA group groups by Kruskal-Wallis test and Dunn´s multiple comparison test. Only comparisons where p-val< 0.1 are display in the figures. *p< 0.05, **p< 0.01, ***p< 0.001.

Figure 3

Contribution of the secretion signal, the covalent link between CCL19 and the neoepitopes, and dimerization module to the efficacy of APC-targeting pDNA vaccines. BALB/c mice received five weekly pDNA immunizations in a prophylactic setup. Seven or nine days after the first immunization, tail-vein blood was collected and stained with a neoepitope-specific MHC-I multimer to analyze the frequency of C1-specific CD8+ T cells in circulation (n = 4-7 mice per group). The complete gating strategy used to identify the CD8+ T-cell subset is exemplified in Supplementary Figure 7A . Two weeks after the first immunization, mice were inoculated with 5 x 10⁵ CT26 tumor cells s.c. in the right flank (n =13 mice per group). The studies were terminated 18 days after tumor inoculation. In the first experiment (C–E) mice received: i) 5 μg CCL19_Neo5, ii) 5 μg NT_Neo5 in combination with 5 μg pCCL19, iii) 5 μg Neo5 in combination with 5 μg pCCL19, iv) 5 μg CCL19 or v) 10 μg of Mock pDNA per immunization. In the second experiment (F–H) mice received 5 μg pDNA per immunization. (A, B) Schematic design of pDNA constructs. Created with BioRender.com. (C, F) Mean tumor volume (mm³) ± SEM over time for each treatment group with LOCF. (D, G) Area under the curve (AUC) of individual tumor volume split by group. Mean ± SD (E, H) % of C1-specific CD8+ T cells in circulation 9 and 7 days after the first immunization, respectively. Mean ± SD. Statistics: (D, G) Kruskal-Wallis test and Dunn´s multiple comparison test. (E, H) One-way ANOVA and Šidák´s multiple comparison test. All the comparisons performed are displayed in the figures. ns: p ≥ 0.05, *p< 0.05, ***p< 0.001, ****p< 0.0001.

Figure 4

APC-targeting of pDNA-encoded neoepitopes elicits tumor prevention at low DNA doses. BALB/c mice received five weekly pDNA immunizations in a prophylactic setup. Sixteen days after the first immunization, tail-vein blood was collected and stained with a neoepitope-specific MHC-I multimer to analyze the frequency of C1-specific CD8+ T cells in circulation (n = 3-5 mice per group). The complete gating strategy used to identify the CD8+ T-cell subset is exemplified in Supplementary Figure 7A . Two weeks after the first immunization, mice were inoculated with 2,5 x 10⁵ CT26 tumor cells s.c. in the right flank (n =13 mice per group). The study was terminated 22 days after tumor inoculation. (A) Mean tumor volume (mm³) ± SEM over time for each treatment group with LOCF. (B) Tumor growth of individual mice over time. (C) The area under the curve (AUC) of individual tumors split by group. Mean ± SD (D) Frequency of C1-specific CD8+ T cells in circulation. Mean ± SD. Statistics: (C) Kruskal-Wallis test and Dunn´s multiple comparison test. (D) One-way ANOVA and Šidák´s multiple comparison test. All the comparisons performed are displayed in the figures. ns: p ≥ 0.05, *p< 0.05, **p< 0.01, ****p< 0.0001.

Figure 5

Neoepitope-encoding APC-targeting pDNA immunotherapy induces therapeutic tumor control. BALB/C mice were inoculated with 2 x 10⁵ CT26 tumor cells s.c. in the right flank (n =13 mice per group). One day after, the mice received five EP-assisted immunizations of 50 μg pDNA spaced over three to four days. Ten days after tumor inoculation, tail-vein blood was collected and stained with a neoepitope-specific MHC-I multimer to analyze the frequency of C1-specific CD8+ T cells in circulation (n = 5-6 mice per group). The study was terminated 21 days after tumor inoculation, and the spleen compartment was analyzed for TNFα and IFNγ-secreting CD8+ and CD4+ T cells by ICS (n = 4-7 mice per group). The complete gating strategies used to identify specific T-cell subsets is exemplified in Supplementary Figure 7 . (A) Mean of group tumor volume (in mm³) ± SEM over time with LOCF. (B) Tumor growth of individual mice over time. (C) The area under the curve (AUC) of individual tumors split by group. Mean ± SD (D) Kaplan-Meier curve depicting % of tumor-free mice in each group over time. (E) Frequency of C1-specific CD8+ T in circulation. Mean ± SD. (F) Frequency of reactive CD8+ T cells and (G) CD4+ T cells. Mean ± SD. (H) IFNγ SFU/5x10⁵ splenocytes by ELISpot. Statistics: (C, E–G) Kruskal-Wallis and Dunn´s multiple comparison test. All the comparisons performed are displayed in the figures. ns: p ≥ 0.05, *p< 0.05, **p< 0.01 (D) Mantel-Coxt test with Bonferroni correction for multiple comparisons, *p< 0.0167.