Augmented Expansion of Treg Cells From Healthy and Autoimmune Subjects via Adult Progenitor Cell Co-Culture

Abstract

Recent clinical experience has demonstrated that adoptive regulatory T (Treg) cell therapy is a safe and feasible strategy to suppress immunopathology via induction of host tolerance to allo- and autoantigens. However, clinical trials continue to be compromised due to an inability to manufacture a sufficient Treg cell dose. Multipotent adult progenitor cells (MAPC^Ⓡ) promote Treg cell differentiation in vitro, suggesting they may be repurposed to enhance ex vivo expansion of Tregs for adoptive cellular therapy. Here, we use a Good Manufacturing Practice (GMP) compatible Treg expansion platform to demonstrate that MAPC cell-co-cultured Tregs (MulTreg) exhibit a log-fold increase in yield across two independent cohorts, reducing time to target dose by an average of 30%. Enhanced expansion is coupled to a distinct Treg cell-intrinsic transcriptional program characterized by elevated expression of replication-related genes (CDK1, PLK1, CDC20), downregulation of progenitor and lymph node-homing molecules (LEF1 CCR7, SELL) and induction of intestinal and inflammatory tissue migratory markers (ITGA4, CXCR1) consistent with expression of a gut homing (CCR7lo β₇hi) phenotype. Importantly, we find that MulTreg are more readily expanded from patients with autoimmune disease compared to matched Treg lines, suggesting clinical utility in gut and/or T helper type1 (Th1)-driven pathology associated with autoimmunity or transplantation. Relative to expanded Tregs, MulTreg retain equivalent and robust purity, FoxP3 Treg-Specific Demethylated Region (TSDR) demethylation, nominal effector cytokine production and potent suppression of Th1-driven antigen specific and polyclonal responses in vitro and xeno Graft vs Host Disease (xGvHD) in vivo. These data support the use of MAPC cell co-culture in adoptive Treg therapy platforms as a means to rescue expansion failure and reduce the time required to manufacture a stable, potently suppressive product.

Keywords: adult progenitor cells; autoimmune diseases; co-culture; ex vivo expansion; regulatory T (Treg) cells.

Figure 1

Treg and MulTreg isolation and expansion. (A) Schematic of the expansion protocol for Treg and MulTreg. (B) The fold expansion of Treg and MulTreg lines vs ex vivo (fresh Tregs plated post sort) was calculated for each donor at each time point. Datapoints represent the mean +/- SEM for 10 donors. (C) Fold expansion (from ex vivo) at d30 for the UK (KCL) cohort (Treg and MulTreg lines grown from 10 individuals). (D) Average fold expansion (from ex vivo) at d30 for the validation (Belgian) cohort (Treg and MulTreg lines grown from 5 individuals). (E) Fold expansion in MulTreg versus Treg expansion was calculated for each donor at each time point in the combined cohort: Treg and MulTreg lines grown from 15 individuals. All p and q values from single or multiple (corrected) matched-pairs Wilcoxon-signed rank test. **q < 0.01, ***q < 0.0005.

Figure 2

MulTreg are stably committed FoxP3hi CD4+ Treg cells exhibiting a distinct tissue homing phenotype. (A) Example of FoxP3 staining in activated CD4 T cells and Treg or MulTreg lines. (B) Frequency of CD4+ T cells in live gated events, n=8 pairs. (C) Frequency of CD4+ T cells expressing FoxP3 in activated bulk CD4 T cells (n = 2) Treg and MulTreg cell lines, n = 6 pairs. (D) Median fluorescence intensity of FoxP3 staining in activated CD4 T cells (n = 2), Treg and MulTreg cell lines, n = 6. (E) Mean percentage of methylation on 9 CpG islands of the FOXP3 Treg-specific demethylated region (TSDR), n = 2. (F) Representative flow cytometry plots of expanded lines following restimulation for 5h with PMA/Io and intracellular staining with the cytokines indicated, gated on live CD4 T cells. (G, H) Representative flow cytometry plots (G) or histograms (H) of expanded lines showing expression of the markers indicated. (I, J) Relative MFI of markers indicated amongst Treg and MulTreg lines (expressed as log2FC in favor of MulTreg) from (I) n = 5 donors in the KCL (Discovery) cohort and (J) n = 5 donors in the Belgian (Validation) cohort. (K–M) Combined analysis of KCL and Belgian cohorts (n = 10) showing frequency (K) or relative MFI (L) of marker expression indicated and the ratio of β₇ integrin to CCR7 MFI, defined as the tissue homing index (M). Data points represent individual donors, Error bars = SEM of 5 or 10 (L) donors. Stats derived from single or multiple Wilcoxon matched-pairs signed rank test. NS, not significant.

Figure 3

MAPC expanded Tregs exhibit superior suppression of antigen specific responses. (A–C) Suppression assay using 3rd party PBMC stimulated with CD3/CD28 microbeads for 6 days in the presence or absence of Treg or MulTreg lines. (A). Flow cytometry plots showing proliferation in responder T cells from PBMC stimulated with CD3/CD28 microbeads (1:10 bead to cell) in the presence or absence of autologous, expanded Treg or MulTreg lines at a ratio of 1:2 Treg : PBMC. (B) Bar graph displaying % suppression of responder T cell proliferation at different ratios of Treg or MulTreg to PBMC. (C) Bar graph displaying cytokine levels in tissue culture supernatant from suppression assays in panel (B) (ratio 1:2 Treg : PBMC). n = 5. (D–F) Suppression assay using autologous PBMC stimulated with Flu-HA for 6 days in the presence or absence of Treg or MulTreg lines. (D) Flow cytometry plots showing proliferation in responder CD4 T cells from PBMC stimulated with Flu-HA in the presence or absence of autologous, expanded Treg or MulTreg lines at a ratio of 1:5 Treg : PBMC. (E) Bar graph displaying % suppression of responder CD4+ T cell proliferation at different ratios of Treg : PBMC, CD4 (left) and CD8 (right). (F) Bar graph displaying cytokine levels in tissue culture supernatant from suppression assays in panel (E) (ratio of 1:5). Error bars represent the SEM of 5 donors. Stats from Wilcoxon matched-pairs analysis (B, E) or Friedman test (C, F). *p < 0.05, **p < 0.01, ns, not signficant.

Figure 4

MulTreg effectively control the immune response in a humanized mouse model of xeno-GvHD. NSG mice were inoculated with 10⁷ PBMCs +/- 10⁷ Tregs or MulTregs or PBS as a control (PBMC alone n = 7, PBMC and Treg n = 10, PBMC and MulTreg n = 12, PBS alone n = 5). Survival of mice administered with MulTreg and Treg in addition to PBMC was significantly longer than those given PBMC alone. Log-rank Mantel-Cox Test (*p < 0.05, ***p < 0.001).

Figure 5

Characteristics of MulTregs from patients with Crohn’s Disease. (A) Average fold expansion (from ex vivo) for Treg and MulTreg lines grown from 4 individuals with Crohn’s disease and 2 individuals with type 1 diabetes (T1D). Wilcoxon matched-pairs signed rank test (n = 6). (B) Paired flow cytometric analysis of marker frequencies in Treg and MulTreg cells (n=3). (C) Average percentage of methylation on 9 CpG islands of the FOXP3 Treg-specific demethylated region (TSDR) (n = 3). Friedman test (ns). (D, E) Suppression assay using 3rd party PBMC stimulated with CD3/CD28 microbeads (1:10 bead to cell) for 6 days in the presence or absence of Treg or MulTreg lines before analysis at d6. (D) Bar graph displaying % suppression of responder CD3+ T cell proliferation at different ratios of Treg : PBMC (n = 4). (E) Bar graph displaying cytokine levels in tissue culture supernatant from suppression assays in panel (D) (ratio 1:2) (n = 3), Friedman test. Data in (A–D) analysed via Wilcoxon matched-pairs analysis non-significant unless stated. **q < 0.01. ns, non-significant.

Figure 6

RNAseq analysis of Treg and MulTreg. Pairs of Treg and MulTreg expansions from n = 4 healthy donors and n = 2 patients with T1D were analysed by bulk RNAseq. (A) Volcano plot of all genes analysed highlighting differentially expressed genes (DEGs) (q < 0.05) in color, with genes of interest labelled. X axis represents log2 fold change in favor of Multreg. (B) Unsupervised principal component analysis showing global gene expression in all 14k transcripts. Plot shows the first two components (PC1 and 2) that account for 14.2% and 12.6% of the variance respectively. z PCs 3 and 4 accounted for 10.3 and 10%, respectively and PCs 5-11 contributed the remaining variance (range 9.2-6.5%). (C) IMPACT pathway analysis displaying pathways in which DEGs are significantly accumulated (pAcc) or over-represented (ORA) following Bonferroni correction. (D–H) Bar plots of genes in pathways highlighted by IMPACT analysis (D–F) or related to exhausted or progenitor T cell states (G, H). (I) Gene set enrichment analysis (GSEA) Multiplot showing enrichment of the gene sets indicated from references in the main text. * < 0.05 Wilcoxon-paired test. Data points in (D–H) represent individual donors. ns, not signficant.