Page 247 - 2020_01-Haematologica-web

Targeting CARVs using ex-vivo-expanded VSTs
AB
CD
Figure 1. Generation of polyclonal multi-respiratory virus-targeted cells (multi-R-VST) from healthy donors. (A) A schematic representation of the multi-R-VST gener- ation protocol. (B) Fold expansion achieved over a 10-13 day period based on cell counting using trypan blue exclusion (n=12). (C and D) Phenotype of the expanded cells (mean±Standard Error of Mean, n=12) (SEM).
Results
Generation of polyclonal multi-respiratory virus-targeted cells from healthy donors
To investigate the feasibility of generating VST-specific T-cell lines containing sub-populations of cells reactive against Influenza, RSV, hMPV, and PIV-3 we utilized a pool of overlapping peptide libraries spanning immuno- genic antigens from each of the target viruses (Influenza – NP1andMP1;RSV–NandF;hMPV–F,N,M2-1andM; PIV-3 – M, HN, N and F) to stimulate PBMC before cul- ture in a G-Rex10 in cytokine-supplemented VST medium (Figure 1A). Over 10-13 days we achieved an average 8.5- fold increase in cells (Figure 1B) [increase from 0.25x107 PBMC/cm2 to mean 1.9±0.2x107 cells/cm2 (median: 2.05x107, range: 0.6-2.82x107 cells/cm2; n=12)], which were comprised almost exclusively of CD3+ T cells (96.2±0.6%; mean±SEM), with a mixture of cytotoxic (CD8+; 18.1±1.3%) and helper (CD4+ ; 74.4±1.7%) T cells (Figure 1C), with no evidence of regulatory T-cell out- growth, as assessed by CD4/CD25/FoxP3+ staining (Online Supplementary Figure S1). Furthermore, the expand- ed cells displayed a phenotype consistent with effector function and long-term memory as evidenced by upregu- lation of the activation markers CD25 (50.2±3.8%), CD69 (52.8±6.3%), CD28 (85.8±2%) as well as expression of central (CD45RO+/CD62L+: 61.4±3%) and effector mem- ory markers (CD45RO+/CD62L−: 20.3±2.3%), with mini- mal PD1 (6.9±1.4%) or Tim3 (13.5±2.3%) surface expres- sion (Figure 1C and D).
Anti-viral specificity of multi-respiratory virus-targeted cells
To next determine whether the expanded populations were antigen-specific, we performed an IFNγ ELIspot assay using each of the individual stimulating antigens as an immunogen. All 12 lines generated proved to be reactive against all of the target viruses (Table 1 and Online Supplementary Figure S2). Figure 2A summarizes the magni- tude of activity against each of the stimulating antigens, while Online Supplementary Figure S3 shows the response of our expanded VST to titrated concentrations of viral anti- gen. Of note, over the 10-13 days in culture we achieved an enrichment in VST of between 14.6±4.3-fold (PIV-3-HN) and 50.4±9.9-fold (RSV-N) (Figure 2B). The precursor fre- quencies of CARV-reactive T cells within donor PBMC are summarized in Online Supplementary Figures S4 and S5. Taken together these data suggest that respiratory VST reside in the memory pool and can be readily amplified ex vivo using GMP-compliant manufacturing methodologies.
To next evaluate whether viral specificity was contained within the CD4+ or CD8+ or both T-cell subsets we per- formed ICS, gating on CD4+ and CD8+ IFNγ-producing cells. Figure 2C shows representative results from one donor with activity against all four viruses detected in both T-cell compartments [(CD4+: Influenza – 5.28%; RSV – 11%; hMPV – 6.57%; PIV-3 – 3.37%), (CD8+: Influenza – 2.26%; RSV – 4.36%; hMPV – 2.69%; PIV-3 – 2.16%)] while Figure 2D shows a summary of results for nine donors screened, confirming that our multi-R-VST are polyclonal and poly-specific.
haematologica | 2020; 105(1)
237