Expanded nonhuman primate tregs exhibit a unique gene expression signature and potently downregulate alloimmune responses

Abstract

We have established two complementary strategies for purifying naturally occurring regulatory T cells (Tregs) from rhesus macaques in quantities that would be sufficient for use as an in vivo cellular therapeutic. The first strategy identified Tregs based on their being CD4+/CD25(bright). The second incorporated CD127, and purified Tregs based on their expression of CD4 and CD25 and their low expression of CD127. Using these purification strategies, we were able to purify as many as 1x10(6) Tregs from 120 cc of peripheral blood. Cultures of these cells with anti-CD3, anti-CD28 and IL-2 over 21 days yielded as much as a 450-fold expansion, ultimately producing as many as 4.7x10(8) Tregs. Expanded Treg cultures potently inhibited alloimmune proliferation as measured by a carboxyfluorescein succinimidyl ester- mixed lymphocyte reaction (CFSE-MLR) assay even at a 1:100 ratio with responder T cells. Furthermore, both responder-specific and third-party Tregs downregulated alloproliferation similarly. Both freshly isolated and cultured Tregs had gene expression signatures distinguishable from concurrently isolated bulk CD4+ T-cell populations, as measured by singleplex reverse transcriptase-polymerase chain reaction (RT-PCR) and gene array. Moreover, an overlapping yet distinct gene expression signature seen in freshly isolated compared to expanded Tregs identifies a subset of Treg genes likely to be functionally significant.

Figure 1

Purification of Tregs from Rhesus macaques: A: Flow cytometric analysis of CD4 cells from rhesus macaques reveals a continuum of CD25 expression, such that ~1% of CD4+ cells can be identified as CD25^bright, just has been shown previously for human samples (2h) B: Representative flow cytometric plots of CD4+/CD25- cells (left panel) and CD4+/CD25^bright cells (right panel) after sorting using a FACSAria cell sorter. C: Representative analysis of FoxP3 expression in either purified bulk CD4 cells or CD4+/CD25bright putative Tregs. Top left panel: RT PCR analysis of control RNAs (18s and CD4) shows no difference in expression between bulk CD4 cells and putative Tregs. RT PCR analysis of FoxP3 (top right panel) shows significantly more FoxP3 message in putative Tregs compared to bulk CD4 cells. Bottom panels: FoxP3 expression as measured by intracellular flow cytometric analysis reveals minimal expression on Bulk CD4 cells but high expression on Tregs.

Figure 1

Purification of Tregs from Rhesus macaques: A: Flow cytometric analysis of CD4 cells from rhesus macaques reveals a continuum of CD25 expression, such that ~1% of CD4+ cells can be identified as CD25^bright, just has been shown previously for human samples (2h) B: Representative flow cytometric plots of CD4+/CD25- cells (left panel) and CD4+/CD25^bright cells (right panel) after sorting using a FACSAria cell sorter. C: Representative analysis of FoxP3 expression in either purified bulk CD4 cells or CD4+/CD25bright putative Tregs. Top left panel: RT PCR analysis of control RNAs (18s and CD4) shows no difference in expression between bulk CD4 cells and putative Tregs. RT PCR analysis of FoxP3 (top right panel) shows significantly more FoxP3 message in putative Tregs compared to bulk CD4 cells. Bottom panels: FoxP3 expression as measured by intracellular flow cytometric analysis reveals minimal expression on Bulk CD4 cells but high expression on Tregs.

Figure 1

Purification of Tregs from Rhesus macaques: A: Flow cytometric analysis of CD4 cells from rhesus macaques reveals a continuum of CD25 expression, such that ~1% of CD4+ cells can be identified as CD25^bright, just has been shown previously for human samples (2h) B: Representative flow cytometric plots of CD4+/CD25- cells (left panel) and CD4+/CD25^bright cells (right panel) after sorting using a FACSAria cell sorter. C: Representative analysis of FoxP3 expression in either purified bulk CD4 cells or CD4+/CD25bright putative Tregs. Top left panel: RT PCR analysis of control RNAs (18s and CD4) shows no difference in expression between bulk CD4 cells and putative Tregs. RT PCR analysis of FoxP3 (top right panel) shows significantly more FoxP3 message in putative Tregs compared to bulk CD4 cells. Bottom panels: FoxP3 expression as measured by intracellular flow cytometric analysis reveals minimal expression on Bulk CD4 cells but high expression on Tregs.

Figure 2

Both Treg and Bulk CD4 cells exhibit significant expansion when placed in culture conditions as described in Methods.

Figure 3

Sorting putative Tregs on the basis of their low expression CD127 also yields cells with high level expression of FoxP3.

Figure 4

Flow cytometric analysis of phenotype of either Bulk CD4 cells or Tregs. A: Representative flow plots showing CD25, FoxP3, CD27 and CD62L expression in either Treg cultures (blue), bulk CD4 cultures (red) or in Treg cultures treated with isotype control antibodies (green). Shown are histograms created after first gating on CD4+ cells. Expression patterns were measured at day 0, day 7, day 14 and day 21 after flow cytometric sorting. B: Representative dot-plot showing high level of expression of FoxP3 on Treg cultures and low level expression on Bulk CD4 cells. FoxP3 was measured after first gating on CD4+ cells. C: Representative data depicting the % of positive cells (red) and the mean fluorescence intensity (MFI, black) for CD25, FoxP3, CD27 and CD62L for both Treg and Bulk CD4 cultures measured at 21 days after culturing.

Figure 5

CFSE MLR analysis reveals in vitro suppressive activity of cultured Tregs. Representative CD4+ and CD8+ T-cell proliferation kinetics in MLR in the presence of varying ratios of Tregs to responder T cells are shown. Flowjo analysis was performed and “percent divided” (upper left corner of each panel) and “division index” (upper right corner of each panel) are indicated for each Treg:responder T cell ratio.

Figure 6

Representative CD4+ and CD8+ T-cell proliferation kinetics in MLR with and without recipient-specific or third-party Treg influence, showing “percent divided” (upper left corner of each panel) and “division index” (upper right corner of each panel).

Figure 7

Cluster diagram of GeneChip results from cultured CD4+/CD25^bright and CD4+/CD25- cells. Expression levels were measured in three replicate cultures of each cell type. Significant changes were evaluated following RMA normalization and analysis in GeneSpring. The 387 genes selected were clustered to reveal similarity patterns. Both gene tree and the condition tree were created by unsupervised analysis. Expression levels are shown relative to the average fluorescence intensity of all genes (shown as yellow). Colorbar estimates relative expression levels: Blue indicates lower than average expression, red indicates higher than average expression.

Figure 8

Cluster diagram of GeneChip results from freshly-sorted CD4+/CD25^bright and CD4+/CD25- cells. Expression levels were measured in four replicate preparations of each cell type. The 221 genes that met cutoff criteria (2.5-fold different levels between groups, and p<0.05) were used to create the cluster diagram. Both gene tree and condition tree were generated by unsupervised analysis. Colorbar estimates relative expression levels: Blue indicates lower than average expression, red indicates higher than average expression.