A combination of cyclophosphamide and interleukin-2 allows CD4+ T cells converted to Tregs to control scurfy syndrome

Abstract

Immunodysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome is caused by mutations in forkhead box P3 (FOXP3), which lead to the loss of function of regulatory T cells (Tregs) and the development of autoimmune manifestations early in life. The selective induction of a Treg program in autologous CD4+ T cells by FOXP3 gene transfer is a promising approach for curing IPEX. We have established a novel in vivo assay of Treg functionality, based on adoptive transfer of these cells into scurfy mice (an animal model of IPEX) and a combination of cyclophosphamide (Cy) conditioning and interleukin-2 (IL-2) treatment. This model highlighted the possibility of rescuing scurfy disease after the latter's onset. By using this in vivo model and an optimized lentiviral vector expressing human Foxp3 and, as a reporter, a truncated form of the low-affinity nerve growth factor receptor (ΔLNGFR), we demonstrated that the adoptive transfer of FOXP3-transduced scurfy CD4+ T cells enabled the long-term rescue of scurfy autoimmune disease. The efficiency was similar to that seen with wild-type Tregs. After in vivo expansion, the converted CD4FOXP3 cells recapitulated the transcriptomic core signature for Tregs. These findings demonstrate that FOXP3 expression converts CD4+ T cells into functional Tregs capable of controlling severe autoimmune disease.

Conflict of interest statement

Conflict-of-interest disclosure: The authors declare no competing financial interests.

© 2021 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

The efficiency of FOXP3 expression by lentiviral vectors. (A) Vector maps showing the design of the 4 vectors and their mock counterparts within the pCCL backbone. Two bidirectional vectors: LNGFRp-eFOXP3 expressing FOXP3 under the control of EF1α promoter and expressing the reporter protein ΔLNGFR under the control of PGK promoter, and its mock counterpart LNGFRp-e, and LNGFRe-pFOXP3 expressing FOXP3 under the control of PGK promoter and expressing the reporter protein ΔLNGFR under the control of the EFS promoter, and its mock counterpart LNGFRe-p. Two bicistronic vectors under the control of EF1 promoter, both based on a self-cleaving T2A sequence: eLNGFR.t2a.FOXP3 and its mock counterpart eLNGFR.t2a and eFOXP3.t2a.LNGFR, and its mock counterpart e.t2a.LNGFR. (B) Quantification of the titers of vector used for transduction. (C) Representative flow cytometry dot plots showing the expression of hFOXP3 and ΔLNGFR on WT murine CD4+ T cells 5 days after transduction with all the constructs expressing FOXP3. The correlation between FOXP3 expression and ΔLNGFR expression was quantified by calculating the Spearman correlation coefficient (r2 = 0.51, 0.54, 0.66, and 0.61 for the LNGFRp-eFOXP3, LNGFRe-pFOXP3, eLNGFR.t2a.FOXP3, and eFOXP3.t2a.LNGFR vectors, respectively). (D) Transduction efficacy quantified as the percentage of ΔLNGFR on day 5 after transduction in WT and scurfy CD4+ T cells. Transduction efficiency was significantly higher with the LNGFRp-eFOXP3 vector than with the LNGFR.t2a.FOXP3 vector for both WT CD4+ T cells and scurfy CD4+ T cells (P = .002 and .007, respectively, in a Mann-Whitney test). In contrast, transduction efficacy with the mock vectors was higher with the T2A construct (P = .02 and .04 in WT and scurfy CD4+ T cells, respectively). n = 3 independent experiments for WT CD4+ T cells and n = 2 independent experiments for scurfy CD4+ T cells. (E) The geometric MFI for FOXP3 expression was quantified on day 5 posttransduction (gated on CD4+ ΔLNGFR+) in WT cells or scurfy cells. The hFOXP3 MFI in WT CD4 T cells was similar with LNGFRp-eFOXP3 and LNGFR.t2a.FOXP3 vectors. The MFI was significantly higher in scurfy CD4 T cells transduced with the LNGFRp-eFOXP3 vector than in the same cell type transduced with LNGFR.t2a.FOXP3 (P = .04 in a Mann-Whitney test). n = 3 independent experiments for WT CD4+ T cells and n = 2 independent experiments for scurfy CD4+ cells. ψ, psi packaging element; *P < .05; **P < .01; APC, allophycocyanine; CMV, cytomegalovirus; FITC, fluorescein isothiocyanate; LTR, long terminal repeat; ns, not significant; PolyA, polyadenylation sequence; SIN, self-inactivating; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element.

Figure 2.

A specific combination of Cy conditioning, IL-2 treatment, and Treg transfer rescues scurfy syndrome. (A) Rescue of scurfy mice. Scurfy male mice (XSf/Y.Rag1+/− on a CD45.2 background) were conditioned by an intraperitoneal injection of 50 mg/kg Cy on day 10 and then received 5 × 105 congenic CD45.1 WT Tregs on day 14. Next, 1000 IU/g IL-2 was injected intraperitoneally once a day for 5 days and then once a week. In an initial experiment, all mice were euthanized on day 50 for flow cytometry analysis. Survival was analyzed in a second experiment (n = 3 mice per group). (B) The scurfy disease score was rated (as described in “Materials and methods”) in Treg-treated mice and vehicle (PBS)-treated mice. Differences were apparent after day 31 (P = .003 in a Mann-Whitney test) and after day 38 (P < .001). (C) Flow cytometry analysis of CD45.1 chimerism, gated on CD4+ T cells in the lymph nodes, spleen, blood, liver, and lung. IL (n = 3 mice per group). (D) Survival of scurfy mice untreated (PBS), treated with Cy only (Cy), Cy and IL-2 (Cy+IL-2+PBS or Cy, IL-2, and 5 × 105 Tregs (Cy+IL-2+Treg). There was a significant difference (P = .0004 in a log-rank test) between Cy+IL-2+PBS and Cy+IL-2+Treg groups (n = 6 mice per group). (E) A representative flow cytometry histogram of CD62L expression, gated on CD4+ T cells in mice that had received Treg or vehicle (n = 3 mice per group). ***P < .001; ****P < .0001.

Figure 3.

Scurfy CD4 T cells engineered with an LNGFR.FOXP3 vector rescue scurfy mice after disease onset. Male scurfy mice (XSf/Y.Rag1+/− on a CD45.2 background) were conditioned by an intraperitoneal injection of 50 mg/kg Cy on day 10 and then received either vehicle, 5 × 105 congenic CD45.1 WT Tregs, CD4LNGFR transduced scurfy cells, or 0.75 × 106 CD4LNGFR.FOXP3 transduced scurfy cells on day 14. Next, 1000 IU/g IL-2 were injected intraperitoneally once a day for 5 days and then once a week. Data from at least 2 independent experiments are shown. (A) The mean plus or minus SD scurfy disease score in mice treated with Tregs, CD4LNGFR, and CD4LNGFR.FOXP3, vs vehicle-treated mice (P = .01, .26, and .02 in a Mann-Whitney test, respectively) on day 42 (n ≥ 3 per group). (B) All of the mice were euthanized on day 50 for flow cytometry analysis. CD45.1 chimerism was analyzed on day 50 (gated on CD4+ T cells) for the lymph nodes, spleen, blood, liver, and lung in mice receiving WT Tregs or PBS. (C) ΔLNGFR chimerism was analyzed on day 50 (gated on CD4+ T cells) in the lymph nodes, spleen, blood, liver, and lung in mice receiving CD4LNGFR.FOXP3 cells, CD4LNGFR cells, or PBS. The level of chimerism was higher in CD4LNGFR.FOXP3-treated mice than in CD4LNGFR-treated mice (P = .001). (D) Survival of scurfy mice from 2 independent experiments. Treatment with IL-2 did not increase survival relative to treatment with Cy or CD4LNGFR cells. Mice treated with Tregs and CD4LNGFR.FOXP3 cells survived significantly longer than Cy-treated mice (P < .0001 and .0195, respectively; n ≥ 5 per group). Follow-up was continued up to 110 days. *P < .05; **P < .01; ***P < .001; ****P < .0001.

Figure 4.

CD4LNGFR.FOXP3 cells partly maintain the Treg signature after adoptive transfer on day 50 of life. All of the transcriptomic data came from a single experiment in which CD4LNGFR.FOXP3 cells, CD4LNGFR cells, and WT Tregs were isolated from the corresponding treated mice (respectively; n = 4, n = 2, and n = 2) euthanized at 50 days old. (A) A volcano plot (the fold change [FC] vs the P value) of the transcriptomes of CD4LNGFR.FOXP3 vs CD4LNGFR cells on day 50. The murine Treg-upregulated signature (in red), downregulated signature (in blue), and murine core Treg gene annotations are highlighted. The values in the top half represent the number of corresponding Treg signature genes induced (right) or repressed (left), with the number of upregulated signature genes in red and the number of downregulated signature genes in blue. P values for the Treg signature enrichment were obtained in a χ2 test. (B) A heat map showing expression of the core Treg genes. The values correspond to the FC for each gene in each sample, normalized against the mean value for CD4LNGFR cells. (C) The Treg signature score for the murine core Treg genes in CD4LNGFR.FOXP3 cells, where 0 and 1 correspond to expression in Tconvs (CD4.LNGFR) and Tregs, respectively. The mean plus or minus SD signature score for CD4LNGFR.FOXP3 cells is represented. (D) A heat map showing all of the significant differentially expressed genes (absolute FC > 2; P < .05) when comparing CD4LNGFR.FOXP3 cells with CD4LNGFR cells (n = 677). The log10 FC for the 3 groups of mice is shown. The Treg-downregulated signature belongs to the most downregulated genes (green). Most of the upregulated genes are involved in the cell cycle (GO:007049; false discovery rate < 0.001) whereas the most upregulated genes in both Tregs and CD4LNGFR.FOXP3 came from the Treg-upregulated signature (red). The gene names are given beside each group.