Fatal autoimmunity in mice reconstituted with human hematopoietic stem cells encoding defective FOXP3

Abstract

Mice reconstituted with a human immune system provide a tractable in vivo model to assess human immune cell function. To date, reconstitution of murine strains with human hematopoietic stem cells (HSCs) from patients with monogenic immune disorders have not been reported. One obstacle precluding the development of immune-disease specific "humanized" mice is that optimal adaptive immune responses in current strains have required implantation of autologous human thymic tissue. To address this issue, we developed a mouse strain that lacks murine major histocompatibility complex class II (MHC II) and instead expresses human leukocyte antigen DR1 (HLA-DR1). These mice displayed improved adaptive immune responses when reconstituted with human HSCs including enhanced T-cell reconstitution, delayed-type hypersensitivity responses, and class-switch recombination. Following immune reconstitution of this novel strain with HSCs from a patient with immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome, associated with aberrant FOXP3 function, mice developed a lethal inflammatory disorder with multiorgan involvement and autoantibody production mimicking the pathology seen in affected humans. This humanized mouse model permits in vivo evaluation of immune responses associated with genetically altered HSCs, including primary immunodeficiencies, and should facilitate the study of human immune pathobiology and the development of targeted therapeutics.

Figure 1

Normal T-cell development and function in NSGAb°DR1 mice. (A) Human chimerism in blood, spleen, and bone marrow determined using human CD45 staining 20 weeks after reconstitution from a single DR1 allelically matched cord blood donor. n ≥ 6 per group pooled from 3 independent experiments. (B) Human immune cell populations as a percent of human leukocytes in spleens of reconstituted mice. cDCs, classical dendritic cells; pDCs, plasmacytoid dendritic cells; NK, natural killer. (C) Representative flow cytometry dot plots of developing T cells (left) quantified (right) showing double-positive (DP), CD4⁺, CD8⁺, double-negative (DN) and CD4⁺CD25⁺FOXP3⁺ Tregs. n ≥ 5 per group from 2 independent experiments. (D) The CDR3 region of the TCR-β locus was profiled using next-generation ultrahigh-throughput sequencing. Pie segments are displayed in different colors to highlight unique V-gene segments for NSG and NSGAb°DR1 mice. TCR-β repertoire Diversity was calculated using normalized Shannon’s entropy (P > .05). (E) Histogram of CDR3 nucleotide length in CD4⁺ T cells isolated from spleens of reconstituted NSG or NSGAb°DR1 mice. (F) The relative frequency or “evenness” of the CD4⁺ TCRβ repertoire was plotted for the top 100 clonotypes in 2 representative animals in each genotype, with each circle depicting an individual clonotype. (G) Delayed-type hypersensitivity response in OVA-immunized mice displayed as the difference in swelling between OVA-injected left footpad (lfp) and PBS-injected right footpad (rfp). NSGAb°DR1 (mismatched) refers to reconstitution using HSCs from a donor that was negative for the DRB1*01:01 allele. Data are pooled from 2 independent experiments. Bars are the mean ± standard error of the mean (SEM). Statistical analysis performed using 1-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (A), multiple Student t test corrected for multiple comparisons using Holm-Sidak method (C), nonlinear regression (E), unpaired Student t test (F-G). *P < .05, **P < .01, ***P < .001.

Figure 2

Increased B cell maturation and class switched immunoglobulins in reconstituted NSGAb°DR1 mice. (A) Representative flow cytometry dot plots of splenic B cells gated on CD45⁺CD19⁺ (left) and quantified (right) from 3 independent experiments with n ≥ 5 per group. (B) Serum immunoglobulin levels quantified by ELISA. Each data point represents an individual mouse from 3 independent experiments. Solid bars are the mean ± SEM. Statistical analysis was performed using multiple t test corrected for multiple comparisons using the Holm-Sidak method (A), Mann-Whitney test (B). *P < .05, **P < .01, ***P < .001.

Figure 3

Identification of a novel variant in FOXP3 associated with IPEX syndrome. (A) Sequence of the FOXP3 gene in the IPEX patient. (B) FOXP3 expression in Tregs (CD4⁺CD25^hi) and T_naive (CD4⁺CD25⁻) from the IPEX patient and healthy control shown by flow cytometry (left) and histogram the mean fluorescence intensity (right). (C) In vitro suppression assay using Tregs from a healthy control or IPEX patient and carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled allogeneic responder T cells.

Figure 4

NSGAb°DR1 mice reconstituted with IPEX hematopoietic stem cells causes mortality and multiorgan inflammation. (A) Longitudinal assessment of body weight change compared with weight at 10 weeks from 2 independent experiments. (B) Kaplan-Meier survival curve comparing reconstituted mice. The 18-week cutoff was selected based on 100% lethality in IPEX(DR1) mice by this time point in the first experimental cohort to compare variability across 2 independent experiments. Control(DR1) n = 10, IPEX(NSG) n = 10, IPEX(DR1) n = 19. (C) Human chimerism in spleen of reconstituted mice assessed by flow cytometry pooled from 2 experiments. Open circles are individual mice. Bars are the mean ± SEM. (D) Hematoxylin and eosin–stained sections from the lung, liver, small intestine (sm. int.), and colon of NSGAb°DR1 mice reconstituted with control bone marrow CD34⁺ HSCs or IPEX CD34⁺ HSCs, NSG mice reconstituted with IPEX CD34⁺ HSCs, and Foxp3^sf mice. Black arrowheads denote leukocyte infiltrate. Scale bars are 100 μm. Statistical analysis performed using 1-way ANOVA with Tukey’s multiple comparisons test (A,C), curve comparison using log-rank Mantel-Cox test (B). *P < .05, **P < .01.

Figure 5

Human T-cell inflammation in IPEX(DR1) mice. Immunohistochemical analysis detecting human immune cell infiltration using (A) anti-hCD3 staining on lung, liver, sm. int., and colon tissue sections (top) quantified by counting the number of CD3⁺ cells/20× field for 3 mice per group (bottom). (B) Representative flow cytometry dot plots of splenic T cells gated on TCR-β⁺ (left) with the frequency of CD4⁺/CD8⁺ T cells reported (right). Open circles are individual mice. (C) The absolute numbers of splenic T cells were quantified. Open circles are individual mice. All values are mean ± SEM. Statistical analysis performed using 1-way ANOVA with Tukey’s multiple comparisons test (A), 1-way ANOVA using Kruskal-Wallis multiple comparisons test (B-C). *P < .05, **P < .01, ***P < .001.

Figure 6

Loss of B-cell tolerance in IPEX mice. (A) Serum immunoglobulin levels quantified by ELISA from 2 independent experiments. (B) Autoantibody detection by indirect immunofluorescence on cryosections of frozen lung and liver tissue from a nonhumanized NSGAb°DR1 mouse. Frozen sections were incubated independently with sera from 4 individual reconstituted mice per group. Fluorescein isothiocyanate (FITC)-conjugated anti-human IgG was used as a secondary antibody combined with DAPI to label nuclei. (a) Denotes lung airway. Scale bars are 50 μm. (C) Heat map of antibody response to select murine antigens determined using an antigen microarray. Each column is the IgM and IgG serum reactivity from an individual mouse in each group with the averaged reactivity score for all antigens per group noted at the bottom. IgM P = .06, IgG P = .09 (D) The normalized cumulative autoreactive IgM and IgG antibody reactivity values for all 500 mouse antigens in arbitrary units (a.u.) were plotted against the % human B cells in each mouse (n = 5). Solid line marks the best-fit linear regression for each cohort with the R² value calculated using Pearson correlation coefficients. For each data point, the % of human chimerism is shown in parenthesis and for IPEX(DR1) mice, the age in weeks when mice became moribund and were killed. Statistical analysis was performed using Mann-Whitney test (A) and nonlinear regression correlation using Pearson correlation coefficients (D). ***P < .001.