Loss of Gimap5 promotes pathogenic CD4⁺ T cell development and allergic airway disease

Abstract

Background:

The GTPase of immunity-associated protein 5 is essential for lymphocyte homeostasis and survival. Recently, human GIMAP5 SNPs have been linked with an increased risk for asthma, while the loss of Gimap5 in mice has been associated with severe CD4⁺ T celldriven immune pathology.

Objective:

To identify the molecular and cellular mechanisms by which Gimap5-deficiency predisposes to allergic airway disease.

Methods:

CD4⁺ T cell polarization and the development of pathogenic CD4⁺ T cells were assessed in Gimap5-deficient mice and a human patient with a GIMAP5 loss-of-function (LOF) mutation. House dust mite (HDM)-induced airway inflammation was assessed using a complete Gimap5 LOF (Gimap5^sph/sph) and conditional Gimap5^fl/fl Cd4^Cre/ert2 mice.

Results:

GIMAP5 LOF mutations in both mice and humans are associated with spontaneous polarization towards pathogenic T_H17 and T_H2 cells in vivo. Mechanistic studies in vitro reveal that impairment of Gimap5-deficient T_H cell differentiation is associated with increased DNA damage, particularly during T_H1 polarizing conditions. The DNA damage in Gimap5-deficient CD4⁺ T cells could be controlled by TGFβ, thereby promoting T_H17 polarization. When challenged with HDM in vivo, Gimap5-deficient mice displayed an exacerbated asthma phenotype (inflammation and airway hyperresponsiveness) with increased development of T_H2, T_H17, and pathogenic T_H17/ T_H2 cells.

Conclusion:

Activation of Gimap5-deficient CD4⁺ T cells is associated with increased DNA damage and reduced survival that can be overcome by TGFβ. This leads to a selective survival of pathogenic T_H17 cells, but also T_H2 in humans and mice, ultimately promoting allergic airway disease.

Capsule Summary

GIMAP5 LOF mutations in both mice and humans are associated with polarization of CD4⁺ T cells towards T_H17 and T_H2 cells in vivo. Gimap5-deficient mice develop more severe allergic airway disease upon HDM sensitization.

Introduction

Asthma is a chronic airway inflammatory disorder that develops through a variety of mechanisms, each leading to different clinical manifestations and responsiveness to treatment. While the majority of asthma cases can be controlled with standard treatments, a subset of asthmatics develop “severe” or “refractory” asthma.¹ Recent work has shown that these cases are associated with the presence of T_H2/T_H17 CD4⁺ T cells and that these cells drive more severe disease than classical T_H2 CD4⁺ T cells.^2–4 Moreover, levels of IL-17A, or IL-17A producing CD4⁺ cells, in the lung, sputum, and blood of asthmatic patients correlate with disease severity.^2–11However, both genetic and environmental factors that contribute to T_H2 and T_H17-associated asthma development remain unclear.

The GTPase of immunity-associated protein 5 (Gimap5) is critical for lymphocyte survival, with Gimap5-deficient CD4⁺ T cells demonstrating impaired survival and elevated DNA damage upon antigenic stimulation.^12–14 Complete loss-of-function (LOF) mutations in GIMAP5, both in rodents and in humans, results in lymphopenia, loss of immunologic tolerance, and subsequent development of autoimmunity.^12–21 In the context of Gimap5-deficient (Gimap5^sph/sph) mice, this manifests as CD4⁺ T cells-driven colitis. Importantly, polymorphisms in GIMAP5 have been associated with increased risk for type 1 diabetes (T1D) and systemic lupus erythematosus (SLE), and a recent study revealed a SNP in the promoter region of GIMAP5 to be associated with asthma.^22–25 It remains unclear, however, how polymorphisms in GIMAP5 predispose individuals to these inflammatory diseases.

In the current study, we show that Gimap5-deficiency drives the development of pathogenic CD4⁺ T cells with an exaggerated TH17 cell response and a more severe asthma phenotype.

Methods

Study design

For the proposed experiments involving the studies of immune responses in vivo, both male and female mice were used. All mice used were generated on a C57BL/6J background and the background purity was confirmed by whole genome SNP analysis. To minimize confounding secondary factors arising from lymphopenia and other late-stage pathologies that develop in Gimap5^sph/sph mice, CD4⁺ T cells from 3–4 week-old mice were used unless otherwise noted. At this age, Gimap5^sph/sph mice have relatively normal numbers of CD4⁺ T cells with a normal frequency of naïve T cells that exhibit a quiescent phenotype comparable to WT mice. Since we were interested in T cell proliferation/survival in WT vs Gimap5-deficient cells, we expected to find strong differences in measurement (e.g. normal vs absence of T cell survival/proliferation). In cases where the differences in the observed phenomena were clear and distinct, a group size of 6 mice was suitable to give statistically significant (i.e. P<0.05) data. For these studies, we estimated that, with a sample size of 6, we have a 99% power to detect at least 25% reduction in the % of proliferating cells in stimulated Gimap5^sph/sph cells with a significant level (alpha) of 0.05 (two-tailed).

Mice and reagents

All experiments were performed according to US National Institutes of Health guidelines and were approved by the IACUC of The Cincinnati Children’s Hospital. C57BL/6J mice were obtained from Jackson. Deletion of Gimap5 in Gimap5^flox/floxCd4^cre-ert2 mice was induced by the administration of tamoxifen chow (40mg/kg body weight; Harlan Laboratories Teklad Diets). Gimap5^sph/sph mice were generated as previously described (6) and bred in-house in the vivarium of Cincinnati Children’s Hospital. All mice were maintained under specific pathogen-free conditions.

All antibodies used for flow cytometry were purchased from eBioscience or Biolegend unless otherwise noted. Purified α-mouse-CD3 (17A2) and α-mouse-CD28 (37.51) antibodies (Biolegend) were used for mouse T cell activation. A fixable viability stain was purchased from eBioscience. 7-AAD was purchased from BD. PMA, ionomycin, and Brefeldin A were obtained from Sigma. αIFNγ (XMG1.2) and mIL-23 were obtained from eBioscience. TGFβ1, mIL-4, and hIL-2 were purchased from Miltenyi Biotech. αIL-4, IL-6, and mIL-1β, mIL-12, and TGFβ3 were obtained from PeproTech.

Generation of Gimap5^flox/flox mice

CRISPR/Cas9 was used to insert LoxP sites flanking the second exon of Gimap5 (Fig. E3A). As the second exon contains the translation start site, deletion of this exon is expected to result into loss of Gimap5 protein (Fig. E5A). The accuracy of the insert was confirmed through sequencing (Fig. E3B). Presence of the LoxP sites was confirmed through PCR genotyping as shown in Fig. E3C. The PCR primers used are as follows: 5’ LoxP site PCR forward, 5’-GATC GTGCAACGACATGGG-3’, 5’ LoxP site PCR reverse, 5’-ACGACCTTGCCCGACTAGAG-3’, 3’ LoxP site PCR forward, 5’-CCACCCAAGGAGAGTTACCC-3’, 3’ LoxP site PCR reverse, 5’-GTCAAAGACAAGCTTCCACAGT-3’. Mice bearing the Gimap5^fl/fl allele were then crossed those carrying the Cd4^cre-ert2 and Foxp3^cre transgenes to produce Gimap5^fl/flCd4^cre-ert2 or Gimap5^fl/flFoxp3^cre offspring. Deletion of Gimap5 was induced in Gimap5^fl/flCd4^cre-ert2 CD4⁺ T cells through the administration of tamoxifen chow (40mg/kg body weight; Harlan Laboratories Teklad Diets).

Flow cytometry and T cell analyses

T cell proliferation was quantified by incubating Mojo-purified (Biolegend) CD4⁺ T cells in 5 μM CTV in 0.2% FBS for 20 min. Cells were cultured in supplemented IMDM media containing 10% FBS, 2% penicillin/streptomycin, 1% L-glutamine, and 50 μM BME. Cells were either left unstimulated or stimulated with αCD3 (1 μg/ml)/αCD28 (2 μg/ml) in the presence of TGFβ1 (2 ng/mL). After 3 days incubation, proliferation was evaluated by analyzing CTV dilution by flow cytometry. Viability was determined by staining with a fixable viability dye. DNA damage was evaluated by γH2AX staining in conjunction with intracellular 7-AAD staining for cell cycle analysis. To evaluate lymphocyte populations ex vivo, lymphocytes from spleen, mesenteric lymph node (mLN), mediastinal lymph node (mdLN), and lung were isolated and stained with fluorochrome-conjugated antibodies for mouse CD3, CD4, CD19, CD44, γH2AX, Tbet, RORγt, Gata3, and Foxp3. To evaluate CD4⁺ T cell cytokine production capacity, lymphocytes from the spleen, mesenteric lymph node, mediastinal lymph node, and lung were restimulated with PMA (50 ng/mL) and ionomycin (500 ng/mL) in the presence of Brefeldin A for 6h before staining with fluorochrome-conjugated antibodies for mouse IFNγ, IL-13, and IL-17A. Except where noted in in vitro experiments, all data was generated from live cells as determined by fixable viability dye staining.

In vitro CD4⁺ T cell polarization

To test CD4⁺ T cell polarization, Mojo-purified (Biolegend) naïve CD4⁺ T cells were isolated from the spleen and peripheral lymph nodes (cervical, brachial, axillary, and inguinal) of 3–4week-old mice and labeled with 5 μM CTV in 0.2% FBS for 20 min. Cells were cultured supplemented RPMI media containing 10% FBS, 2% penicillin/streptomycin, and 50 μM BME. Cells were stimulated with αCD3 (1 μg/ml)/αCD28 (2 μg/ml) alone or in the presence of polarizing cytokines (Table E1).

Mouse asthma model

Because Gimap5^sph/sph mice show progressive changes in differentiated CD4⁺ T cells with age,^12,14 IT injections started at 4 weeks of age, when a similar frequency of naïve CD4⁺ T cells can be observed compared to WT littermates.^12,14 WT C57BL/6 and Gimap5^sph/sph mice were administered nine, 25 μl intratracheal injections of 25 μg HDM or saline over 3 weeks. For older (6–8 wk) WT, Gimap5^fl/flCd4^cre-ert2, and Gimap5^fl/flFoxp3^cre mice, doses of 50 μg HDM in 50 μl saline were used. WT and Gimap5^fl/flCd4^cre-ert2 mice were placed on tamoxifen chow (40mg/kg body weight; Harlan Laboratories Teklad Diets) 7 days prior to the first I.T. injection in relevant experiments. In HDM experiments, mice were sacrificed 24h after the last exposure.

Airway hyperresponsiveness (AHR)

Mice were anesthetized with Ketamine, Xylazine, and Acepromazine (100, 20, and 10 mg/mL respectively mixed at a ratio of 4:1:1). Invasive measurements of airway responsiveness were made using the flexiVent apparatus (Scireq, Montreal, Canada). Mouse tracheas were cannulated with a 19-gauge blunt needle; mice were ventilated at 150 breaths/min, 3.0 cm water positive end-expiratory pressure. Two total lung capacity perturbations were then performed for airway recruitment before baseline measurement and subsequent methacholine challenges were performed. Dynamic resistance (R) was assessed following exposure to increasing concentrations of aerosolized methacholine (0, 12.5, 25, 50, and 100 mg/mL). The average of the 3 highest R values with a coefficient of determination of 0.9 or greater (as determined by flexiVent software) was used to determine the dose-response curve.

BALF cell collection and analysis

Bronchoalveolar lavage was performed by cannulation of the trachea. Lungs were lavaged with 1 mL PBS. After counting the number of cells with a hemocytometer, cells were spun onto slides and stained with the HEMA3 stain set (Fisher Scientific). After de-identifying samples, approximately 200 cells were counted and the number of macrophages, neutrophils, and eosinophils were calculated.

Isolation and characterization of lung-infiltrating cells

After lavage, lungs were removed and the upper right lobe was minced and digested in 2 ml RMPI + 0.5 mg/ml Liberase DL (Roche Diagnostics) and 0.5 mg/ml DNAse I (Sigma) at 37ºC for 30 min. Cells were then passed over a 70 μm cell strainer. Cells were washed with PBS and counted. ~1×10⁶ cells were stimulated with 50 ng/mL PMA + 500 ng/mL ionomycin for 6h. After 2 hours, Brefeldin A was added. CD4⁺ T cells were stained with a viability dye (eBioscience Live/Dead stain) and antibodies for CD3, CD4, CD44, and γδTCR; intracellular staining for IL-13, IL-17A, and IFNγ was performed after fixation and permeabilization according to manufacturer’s instructions (eBioscience). Foxp3 staining was performed on separate cells that had not been restimulated ex vivo.

Histology

Lung tissue was collected and immediately fixed in 10% buffered formalin solution, followed by routine paraffin embedding. Periodic acid-Schiff staining (PAS) and hematoxylin/eosin (H&E) staining was performed on 4 μm sections from the paraffin-embedded tissue blocks for conventional light microscopy analysis.

Analysis of human T cells

For all studies concerning human cells, informed consent was obtained and studies were approved by the CCHMC institutional IRB. PBMCs were isolated from whole blood of a GIMAP5-deficient patient, a healthy parent (heterozygous for GIMAP5 rs72650695), and unrelated healthy controls by Ficoll-Paque Plus density gradient centrifugation. Isolated PBMCs were cultured in RPMI, 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 2mM Lglutamine, and 10 mM HEPES and were stimulated with PMA (50 ng/mL) and ionomycin (500 ng/mL) in the presence of Brefeldin A and analyzed for IFNγ, IL-17A, and IL-13 production by flow cytometry.

Immunoblotting

Protein lysates from Mojo-isolated (BioLegend) CD4⁺ T cells were prepared according to standard methods from resting cells. Lysates were separated using 10% Bis-Tris Gels, transferred onto nitrocellulose, and immunoblotted with primary antibodies to Gimap5 (MAC421)¹² and βactin (CST).

Analysis and Graphics Software

Flow cytometry data were analyzed using FlowJo. All data were analyzed for statistical significance using GraphPad Prism4® software (GraphPad Software, San Diego, CA). Graphical abstract was created with BioRender. Figures components were generated in FlowJo or GraphPad Prism and assembled in Adobe Illustrator.

Statistical analysis

Data are expressed as mean ± SDs except for where otherwise noted. For studies comparing T cells from C57BL/6J, Gimap5^sph/sph; and/or Gimap5^fl/flCd4^cre-ert2 mice, Student’s two-tailed test or ANOVA followed by Sidak’s multiple comparisons test for three or more groups were used. Data were considered statistically significant if P values were <0.5.

Results

GIMAP5 LOF mutations in mice and human promote development of T_H17/T_H2 CD4⁺ T cells in vivo

Given that polymorphisms in GIMAP5 have been associated with increased risk for type 1 diabetes (T1D), systemic lupus erythematosus (SLE), and asthma.^22–25, and our previous studies revealed the development of CD4-dependent colitis in Gimap5-deficient mice,^14,15 we first investigated whether baseline differences in CD4⁺ T cell polarization exist in vivo. To assess this, we examined the CD4⁺ T cell compartment in the spleen and mesenteric lymph nodes (mLN) of 8-week old WT and Gimap5^sph/sph mice. As previously described, CD4⁺ T cells were reduced in the spleen of Gimap5^sph/sph mice (Fig. 1A). Importantly, however, an increased proportion of the remaining Gimap5^sph/sph CD4⁺ T cells were polarized to T_H2 cells (Gata3⁺CD4⁺) and T_H17 cells (RORγt⁺CD4⁺), but not T_H1 (Tbet⁺, CD4⁺) or Treg cells (Foxp3⁺, CD4⁺) compared to WT mice (Fig. 1B). Furthermore, Gimap5^sph/sph splenic CD4⁺ T cells more frequently produced IFNγ, IL-17A, and/or IL-13 compared to WT CD4⁺ T cells (Fig. 1C-E). Notably, Gimap5^sph/sph mice had expanded frequencies of CD4⁺ T cells that produced both IFNγ⁺IL-17A⁺ or IL-13⁺IL-17A⁺ (Fig. 1F,G). These cells have been associated with pathology in several autoimmune diseases and asthma, respectively.^2,3,26–31

FIG 1.

Enhanced In vivo polarization of Gimap5-deficent mouse and human CD4⁺ T cells. (A) Number of CD4⁺ T cells in the spleen of 8wk WT (black bars) and Gimap5^sph/sph (red bars) mice. (B) frequency of splenic CD4⁺ T cells expressing Foxp3, Tbet, Gata3, or RORγt (n=7). (C-G) Expression of IFNγ, IL-17A, and/or IL-13 in splenic CD4⁺ T cells after PMA/ionomycin stimulation (n=4). (H) Number of CD4⁺ T cells in the mesenteric lymph node of 8wk mice. (I) frequency of CD4⁺ T cells expressing Foxp3, Tbet, Gata3, or RORγt in the mesenteric lymph node (n=7). (J-N) Expression of IFNγ, IL-17A, and/or IL-13 in CD4⁺ T cells from the mesenteric lymph node after PMA/ionomycin stimulation (n=4). Bars represent mean values ± SD. Statistical significance is determined by Student’s two-tailed test. (O-S) Frequency of CD4⁺ T cells producing IFNγ, IL-17A, and/or IL-13 upon PMA/ionomycin stimulation. T cells isolated from the PBMCs of a GIMAP5 patient (GIMAP5^204-Pro/Pro), his heterozygous parent, and unrelated controls (n=2, 2, and 7, respectively). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Interestingly, the mLNs of Gimap5^sph/sph mice contained similar numbers of total CD4⁺ T cells compared to WT littermates, but a higher frequency expressed Tbet, Gata3, or RORγt compared to WT littermates (Fig. 1H,I). Similar to the spleen, an increased proportion of CD4⁺ T cells in the mLN were IFNγ⁺, IL-17A⁺, IL-13⁺, and the rarer IFNγ⁺IL-17A⁺ double producers (Fig. 1J-N). Together these observations suggest that loss of Gimap5 is associated with an increased polarization of effector CD4⁺ T cells, including the development of pathogenic CD4⁺ T cells that may drive immunopathology.

Previously, we described a patient with a LOF mutation in GIMAP5 (rs72650695; a missense variant leading to a Leu204→Pro amino acid change and a complete loss of protein expression)¹⁴ resulting in an immunodeficiency with features of autoimmunity. Functional characterization of CD4⁺ T cells from this patient revealed strikingly similar defects to Gimap5-deficient mouse CD4⁺ T cells.¹⁴ We therefore asked if the patient has similarly skewed CD4⁺ T cell compartment. Strikingly, a higher proportion of the patient’s circulating CD4⁺ T cells expressed IL-17A, IL-13, or IFNγ compared to the patient’s heterozygous mother or unrelated controls upon stimulation with PMA and ionomycin (Fig. 1O-Q). Furthermore, an increased frequency of CD4⁺ T cells co-expressed IL-17A⁺IL-13⁺ or IL-17A⁺IFNγ⁺ (Fig. 1R,S). While we were only able to study a single patient with GIMAP5-deficiency, the increased development of pathogenic peripheral T_H2 and T_H17 cells in the GIMAP5-defcient patient is consistent with our mouse data.

Notably, this predilection towards T_H17 polarization can be observed in in vitro culture systems. In previous studies, we observed that bulk Gimap5^sph/sph CD4⁺ T cells have an impaired proliferation and survival upon stimulation with αCD3/αCD28 (Fig. 2A-B) and only a small subset is able to survive and proliferate.^12,14 A more in-depth evaluation of surviving cells demonstrates expression of high levels of RORγt, indicating either a selective survival of T_H17polarized cells or a predisposition of surviving cells to adopt a T_H17 (Fig. 2C). Together, these data suggest that GIMAP5-deficiency in both humans and mice causes CD4⁺ T cells to preferentially adopt a T_H17 phenotype in vitro and in vivo.

FIG 2.

Preferential expression of RORγt by surviving Gimap5^sph/sph CD4⁺ T cells upon TCR stimulation. (A) Proliferation of WT and Gimap5^sph/sph CD4⁺ T cells after 3 days stimulation with αCD3 (1 μg/ml) + αCD28 (2 μg/ml). (B) Survival CD4⁺ T cells after 3d stimulation. (C) Expression of RORγt in live CD4⁺ T cells after 3d stimulation with αCD3/αCD28. Plots depict live cells (C), while bar graphs (B) represent mean values ± SD (n=4). All experiments were performed at least three times. Statistical significance is determined by Student’s two-tailed test. ****P<0.0001

Selective survival of Gimap5-deficient CD4⁺ T cells under different polarizing conditions

Both Gimap5-deficient rodents and humans develop lymphopenia^12–19 and it is thus unclear whether the increased prevalence of T_H17 cells in vivo is due to a skewing of polarization or selective survival of T_H17 cells. Similarly, the expression of RORγt by almost all surviving Gimap5-deficient CD4⁺ T cells could be due to the survival of cells, or merely as the consequence of surviving cells polarizing to T_H17 cells in vitro. To differentiate between these possibilities, naïve CD4⁺ T cells were isolated from 3–4-week-old wildtype and Gimap5^sph/sph mice and stimulated with αCD3/αCD28 in T_H0, T_H1, T_H2_, T_H17, or iTreg polarizing conditions for 3 days (detailed in Table E1) before evaluating survival and proliferation. As with total CD4⁺ T cells, naïve Gimap5^sph/sph CD4⁺ T cells exhibited reduced survival and proliferation when stimulated with αCD3/αCD28 (T_H0 conditions) compared to WT CD4⁺ T cells (Fig. 3A,B). A similar defect was observed in T_H1 polarizing conditions. This defect was primarily characterized by Gimap5^sph/sph cells unable to complete cell cycle, although a considerable proportion of cells died after cycling. Stimulation in T_H2 polarizing conditions resulted in a more bifurcated response, with a subset of cells exhibiting strong proliferation while a significant number of cells died without completing cell cycle. Strikingly, compared to WT, naïve Gimap5^sph/sph CD4⁺ T cell survival and proliferation was completely normal in T_H17 and iTreg polarizing conditions. Notably, both T_H17 and iTreg polarizing conditions include TGFβ1; we asked if TGFβ signaling promoted Gimap5^sph/sph CD4⁺ T cell survival and proliferation during Th17 polarization. Previously, Lee et al. identified different T_H17 polarizing conditions using combinations of TGFβ1, TGFβ3, IL-6, IL-23, and IL-1β (Table E1).³² To test the contribution of TGFβ and the T_H17 transcriptional program to Gimap5^sph/sph CD4⁺ T cell survival, we stimulated naïve CD4⁺ T cells in T_H17 polarizing conditions containing TGFβ1, TGFβ3, or anti-TGFβ1. While Gimap5^sph/sph CD4⁺ T cell survival and proliferation were improved in T_H17 polarizing conditions lacking TGFβ signaling over T_H0 conditions, survival was significantly reduced compared to T_H17 polarizing conditions containing TGFβ1 or TGFβ3 (Fig. 3C). These data suggest that TGFβ is a critical mediator of Gimap5-deficient CD4⁺ T cell survival ultimately promoting the development of pathogenic T_H17 cells.

FIG 3.

Preferential survival of Gimap5^sph/sph CD4⁺ T cells in T_H17 and iTreg polarizing conditions is associated with TGFβ. (A-B) Proliferation and survival of naïve CD4⁺ T cells from WT and Gimap5^sph/sph mice stimulated 3d in non-polarizing conditions (T_H0) or T_H1, T_H2, T_H17, or iTreg polarizing conditions. Blue histograms represent dead cells; red histograms show live cells. Hatched bars represent non-proliferated cells; solid bars represent proliferated cells. (C) Proliferation of live/dead CD4⁺ T cells after 3d stimulation in different T_H17 polarizing conditions. (D) γH2AX expression (DNA damage) in live CD4⁺ T cells in cycle after 3d stimulation in different polarizing conditions. Data represent mean values ± SD (n=4). All experiments were performed at least three times. Statistical significance is determined by ANOVA followed by Sidak’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Activation of Gimap5^sph/sph CD4⁺ T cells results in increased DNA damage that is mitigated by TGFβ1

We have previously reported that poor survival of Gimap5^sph/sph CD4⁺ T cells upon αCD3/αCD28 stimulation correlated with exacerbated DNA damage.¹⁴ We therefore asked whether the reduced survival in different polarizing conditions was associated with exacerbated DNA damage. Specifically, we assessed the level of phosphorylated histone H2AX (γH2AX) in live, dividing CD4⁺ T cells as a marker for double-stranded DNA breaks. Interestingly, high levels of γH2AX were observed even in WT CD4⁺ T cells when stimulated under T_H0, T_H1, or T_H2 conditions, while T_H17 and iTreg conditions were associated with comparatively modest DNA damage after 3 days of polarizing conditions (Fig. 3D). Importantly, Gimap5^sph/sph CD4⁺ T cells consistently showed elevated DNA damage in T_H0, T_H1 or T_H2 conditions compared to WT (Fig. 3D), while γH2AX levels in both WT and Gimap5^sph/sph CD4⁺ T cells were reduced in polarizing conditions containing either TGFβ1 or TGFβ3. Notably, elevated DNA damage was also observed in Gimap5^sph/sph T_H1 (Tbet⁺) cells directly ex vivo, indicating that this phenotype remains relevant in vivo (Fig E1). Together, these data indicate that Gimap5^sph/sph CD4⁺ T cells are unable to control DNA damage that occurs in T_H0, T_H1 and T_H2 polarizing conditions.

Given the preferential survival of Gimap5-deficient CD4⁺ T cells stimulated in TGFβ-containing- T_H17 and iTreg polarizing conditions, we next asked if TGFβ1 signaling was sufficient to rescue CD4⁺ T cells cultured in polarizing conditions associated with poor survival. Strikingly, proliferation and survival of Gimap5^sph/sph CD4⁺ T cells were almost completely rescued with the addition of TGFβ1 in these polarizing conditions (Fig. 4A,B).

FIG 4.

TGFβ1 rescues survival and proliferation of naïve Gimap5^sph/sph CD4⁺ T cells. Proliferation and survival of naïve WT and Gimap5^sph/sph CD4⁺ T cells after 3d stimulation in (A) T_H0 or (B) T_H2 polarizing conditions ± 2 ng/mL TGFβ1. Blue histograms represent dead cells; red histograms show live cells. Survival, proliferation, and γH2AX expression of naïve CD4⁺ T cells stimulated in (C-D) T_H0, (E-F) T_H1, (G-H) or T_H2 polarizing conditions. At 1 or 2 days after initial stimulation, 2 ng/ml TGFβ1 was added to the indicated culture. Data represent mean values ± SD (n=4). Statistical significance is determined by ANOVA followed by Sidak’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Notably, TGFβ is cleaved to its active form in the lung (and other tissues) after tissue injury,^33–36 resulting in a TGFβ-rich environment. As naïve T cells are first activated in lymph nodes before migrating to the affected tissue, we asked if exposure to TGFβ1 during the cycling phase when DNA damage is most pronounced was sufficient to promote survival of Gimap5^sph/sph CD4⁺ T cells. Indeed, the addition of TGFβ1 after 1 or 2 days of stimulation/polarization robustly improved Gimap5^sph/sph CD4⁺ T cell survival, proliferation, and control of DNA damage in T_H0, T_H1, and T_H2 polarizing conditions, although it was insufficient to completely abrogate the effects of Gimap5-deficiency (Fig. 4C-H). These data demonstrate that, independent of T_H17 or iTreg polarization, TGFβ1 is sufficient to rescue the survival of Gimap5^sph/sph CD4⁺ T cells. Moreover, TGFβ1 signaling is not required at the time of initial αCD3/αCD28 stimulation. These studies thus identify TGFβ as an important mediator during CD4⁺ T cell differentiation limiting DNA damage and improving survival and proliferation of activated Gimap5^sph/sph CD4⁺ T cells and promoting the selective development of pathogenic T_H17 cells.

Gimap5-deficiency exacerbates allergic airway pathology

A recent study identifies GIMAP5 as a modifier gene in asthma.²⁵ Specifically, a 5’ SNP (rs6965571) in the promoter region of GIMAP5 is associated with ~3-fold increased risk of asthma in a prospective Swedish birth cohort²⁵, although the mechanisms responsible for increased asthma risk are unclear. Given our findings that both GIMAP5-deficient mice and human display a high frequency of T_H2 and T_H2/T_H17 pathogenic CD4⁺ T cells in vivo, and that T_H2/T_H17 allergic responses exacerbate the asthma phenotype, we assessed the role of Gimap5 in allergic airway disease. Based on the changes in T cell polarization described above, we predicted that loss of Gimap5 would increase sensitivity to allergic airway disease and we therefore used a mild model of experimentally-induced allergic airway disease in C57BL/6J mice.^37,38 Specifically, we sensitized Gimap5^sph/sph and control C57BL/6 wildtype (WT) mice with nine intratracheal (IT) injections of 25 μg house dust mite (HDM) during a 3-week period as previously described,³⁹ after which immune pathology was assessed. One day following the last exposure, mice were evaluated for airway hyperresponsiveness (AHR) and BAL inflammatory profile. Strikingly, HDM-sensitized Gimap5^sph/sph mice exhibited significantly higher AHR compared to WT mice and saline-treated mice (Fig. 5A). The observed increase in AHR corresponded with a significantly increased infiltration of eosinophils and macrophages in the bronchoalveolar lavage fluid (BALF) in HDM-sensitized Gimap5^sph/sph mice (Fig. 5B-D).

FIG 5.

Gimap5^sph/sph mice exhibit increased AHR, CD4⁺ T cell polarization, and airway pathology upon HDM sensitization. C57BL/6 WT (black bars) and Gimap5^sph/sph (red bars) mice were administered 25 μg HDM 9 times IT over three weeks. (A) Airway hyper-responsiveness (AHR) in response to methacholine challenge 1 day after the last HDM challenge (saline: n=3; HDM: n=5–6). (B-D) Number of macrophages, neutrophils, and eosinophils in the BAL fluid (saline: n=3; HDM: n=4–7). (E-F). Number of (E) total lung-infiltrating cells and (F) CD4⁺ T cells (in 2 lobes). (G-J) Percentage of lung-infiltrating CD4⁺CD3⁺ T cells expressing IL-13, IL-17A, IL-13 and IL-17A, or IFNγ upon restimulation with PMA/ionomycin for 6h (saline: n=3; HDM: n=6–7). (K-L) Representative (K) H&E and (L) PAS images of lung tissue from WT and Gimap5^sph/sph mice given saline or HDM. Data represents mean values ± SD. Statistical significance is determined by ANOVA followed by Sidak’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

To determine whether exacerbated AHR and eosinophil/macrophage infiltration correlated with increased development of pathogenic CD4⁺ T cells, total lung infiltrating cells from the upper right lobe for each mouse were quantified, restimulated with PMA and ionomycin, and evaluated for cytokine production. The lungs from HDM-sensitized Gimap5^sph/sph mice had significantly more inflammatory infiltrate (Fig. 5E) and, importantly, exhibited a four-fold increase in infiltrating CD4⁺ T cells compared to WT mice (Fig. 5F). The latter involved a marked increase in CD4⁺ T cells producing IL-13 or IL-17 (Fig. 5G,H) and importantly also included a significant increase in pathogenic IL-13⁺IL-17A⁺ CD4⁺ T cells (Fig. 5I) that have been associated with severe immune pathology/disease. Similarly, lung-draining mediastinal lymph nodes of Gimap5^sph/sph mice displayed an increase in IL-13⁺, IL-17A⁺, and IL-13⁺IL-17A⁺ CD4⁺ T cells (Fig. E2). Furthermore, HDM-sensitization resulted in distortion of normal lung architecture, increased airway mucus production, marked edema, and lymphocytic infiltration in the interstitial space, representative of perivascular inflammation (Fig. 5K-L). Together, these data exhibit that loss-of-function mutations in Gimap5 exacerbate the asthma phenotype which is associated with a significant expansion of pathogenic T_H2/T_H17 cells.

Polarization defects in Gimap5-deficient CD4⁺ T cells are cell-intrinsic

Besides CD4⁺ T cells, Gimap5 is expressed in innate lymphoid cells (ILCs) and our previous studies indicate NK cell development/survival is markedly affected in Gimap5-deficient mice.^12,19 Given the critical role for ILC2 and ILC3s in CD4⁺ T cell differentiation and asthma,^40–45 we next asked if the observed CD4⁺ T cells defects were cell intrinsic or whether other cell types might contribute to the impaired polarization of CD4⁺ T cells. To do so, we generated the Gimap5^fl/fl mouse using CRISPR/Cas9 on a C57BL/6 background. Loxp sites were placed flanking Exon 2 of Gimap5 (Fig. E3) and subsequent Gimap5^fl/fl mice were crossed to Cd4^cre-ert2 mice, allowing the tamoxifen-inducible deletion of Gimap5 in CD4⁺ T cells (cKO). Effective ablation of Gimap5 expression in CD4⁺ T cells was achieved with the administration of tamoxifen in chow (Fig. 6A).

FIG 6.

Impaired survival and increased DNA damage associated with Gimap5-deficiency is CD4⁺ T cell intrinsic. (A) Expression of Gimap5 in CD4⁺ T cells isolated from WT and Gimap5^fl/flCd4^cre-ert2 mice after tamoxifen-induced deletion of Gimap5. Immunoblot shows three independent cKO mice. (B-C) Proliferation and survival of naïve CD4⁺ T cells from WT and Gimap5^fl/flCd4^cre-ert2 mice stimulated 3d in nonpolarizing conditions (T_H0) or T_H1, T_H2, T_H17, or iTreg polarizing conditions. Blue histograms represent dead cells; red histograms show live cells. (D) γH2AX expression in live, cycling CD4⁺ T cells after 3d stimulation in T_H0, T_H1, T_H2, T_H17, or iTreg polarizing conditions. Data represent mean values ± SD (n=4–5). Statistical significance is determined by ANOVA followed by Sidak’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Similar to CD4⁺ T cells isolated from the Gimap5^sph/sph mouse, Gimap5^fl/flCd4^cre-ert2 naïve CD4⁺ T cells demonstrated impaired survival and proliferation in T_H0_, T_H1, and T_H2 polarizing conditions compared to WT controls. Furthermore, no significant difference in survival and proliferation between WT and cKO CD4⁺ T cells was observed in T_H17 polarizing conditions (Fig. 6B-C). Together, these results correlated with DNA damage observed in cycling CD4⁺ T cells with cKO CD4⁺ T cells displaying elevated γH2AX compared to WT CD4⁺ T cells stimulated in T_H0, T_H1, T_H2, and iTreg polarizing conditions. Similar to Gimap5^sph/sph CD4⁺ T cells, no difference in survival/proliferation was observed in T_H17 polarizing conditions. These data confirm that the impaired survival and increased DNA damage in stimulated Gimap5-deficient CD4⁺ T cells are cell intrinsic.

Loss of Gimap5 in CD4⁺ T cells is sufficient to drive exacerbated AHR and pathology

As pathogenic CD4⁺ T cells are one of the primary drivers of HDM-induced lung pathology, we asked if Gimap5-deficiency in CD4⁺ T cells alone was sufficient to drive more severe disease in the HDM model of allergic asthma. To do so, one week after starting the tamoxifen chow, allergic airway disease was induced WT and Gimap5^fl/flCd4^cre-ert2 mice as above. Comparable to Gimap5^sph/sph mice, conditional deletion of Gimap5 in CD4⁺ T cells was sufficient to drive increased AHR in HDM-sensitized mice (Fig 7A). While exacerbated AHR was not specifically associated with changes in the numbers of macrophages and neutrophils in the BALF, the number of infiltrating eosinophils trended upwards in HDM-sensitized Gimap5^fl/flCd4^cre-ert2 mice compared to WT mice (Fig. 7B-D). Interestingly, while the number of lung-infiltrating CD4⁺ T cells in HDM-sensitized Gimap5^fl/flCd4^cre-ert2 mice was similar, or even slightly reduced, compared to WT mice (Fig. 7E-F), a higher proportion expressed an activated phenotype (CD44^hi) (Fig. 7G). Furthermore, they exhibited an increased frequency of CD4⁺ T cells producing IL-13, IL-17A, or both IL-17A and IL-13 together (Fig. 7H-J), while the frequency of IFNγ⁺ CD4+ T cells was reduced compared to WT mice (Fig.7 K). Notably, similar levels of regulatory CD4⁺ T cells were observed in the lungs WT and Gimap5^fl/flCd4^cre-ert2 mice, indicating that the development of iTreg cells and/or recruitment is not affected in this model (Fig. E4A). B cells were likewise unaffected (Fig. E4B). Similar to the Gimap5^sph/sph mice, HDM sensitization resulted in edema with perivascular inflammation and increased mucus production in the lungs of cKO mice (Fig. 7L-M). These data reveal that the conditional loss of Gimap5 in CD4⁺ T cells is sufficient to drive the development of pathogenic CD4⁺ T cells and exacerbate allergy-like lung disease upon HDM-sensitization.

FIG 7.

Gimap5-deficiency in CD4⁺ T cells is sufficient to drive increased AHR, CD4⁺ T cell polarization, and airway pathology upon HDM sensitization. C57BL/6 WT (black data points) and Gimap5^fl/flCd4^cre-ert2 (red data points) mice were administered 50 μg HDM 9 times IT over three weeks. (A) Airway hyper-responsiveness (AHR) in response to methacholine challenge 1 day after the last HDM challenge (saline: n=4–5; HDM: n=7–8). (B-D) Number of macrophages, neutrophils, and eosinophils in the BAL fluid (saline: n=4–5; HDM: n=8). (E-F). Number of (E) total lung-infiltrating cells (in 2 lobes) and (F) frequency of CD4⁺CD3⁺ T cells therein. (G) Frequency of CD4⁺CD3⁺ T cells expressing an activated (CD44^hi) phenotype. (H-K) Percentage of lung-infiltrating CD4⁺CD3⁺ T cells expressing IL-13, IL-17A, IL-13 and IL-17A, or IFNγ upon restimulation with PMA/ionomycin for 6h (saline: n=4–5; HDM: n=8). (L-M) Representative (L) H&E and (M) PAS images of lung tissue from WT and Gimap5^fl/flCd4^cre-ert2 mice given saline or HDM. Data represents mean values ± SD. Statistical significance is determined by ANOVA followed by Sidak’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

As we have previously shown that Gimap5-deficient Tregs have impaired suppressive capacity, we next asked defective Tregs were responsible for the development of more severe airway disease. Interestingly, selective ablation of Gimap5 in Tregs was insufficient to drive exacerbated airway disease and development of pathogenic CD4⁺ T cells, indicating that loss of Gimap5 intrinsically promotes the polarization of pathogenic CD4⁺ T cells (Fig E5).

Discussion

Herein we describe a key role for Gimap5 in the prevention of pathogenic CD4⁺ T cell development and show that loss of Gimap5 is associated with an increased risk for allergic lung disease. Specifically, HDM sensitization of Gimap5-deficient mice results in severe airway hyperresponsiveness, eosinophil recruitment to the lung, and an exacerbated T_H2/T_H17 response. The increased airway inflammation was also observed in conditional loss of Gimap5 in CD4⁺ T cells demonstrating that these cells are the primary driver of HDM-induced lung pathology. Moreover, even in the absence of HDM sensitization, Gimap5^sph/sph CD4⁺ T cells progressively adopt an inflammatory phenotype, with T_H17 cells being particularly prominent. Importantly, the abnormal CD4⁺ T cell polarization profile is also observed in a GIMAP5-deficient patient that exhibited an increased development of pathogenic CD4⁺ T cells in the periphery. Thus, despite the observed lymphopenia in both Gimap5-deficient mice and humans, the selective survival of pathogenic T_H17 cells may be a key driver of immune pathology in allergic airway disease.

Our current findings that loss of GIMAP5 leads to pathogenic CD4⁺ T cell development may help to understand how GIMAP5 polymorphism rs6965571 in the promoter region of GIMAP5 predisposes to asthma development.²⁵ The rs6965571 SNP is adjacent to a Foxo1 binding site (upstream) and an NFATc1 binding motif (downstream) and could affect the binding of these transcription factors (TFs) to regulate GIMAP5 expression. Interestingly, our previous studies identified an impaired Foxo1 expression¹⁵ and NFATc1 activity upon stimulation of Gimap5-deficient CD4⁺ T cells.¹⁴ Thus, a potential feed-forward loop exists in which Gimap5 expression is required for the expression and activity of TFs that in turn regulate the transcriptional expression of GIMAP5. Thus the SNP may cause a progressive loss of GIMAP5 expression following T cell activation leading to an increased development of T_H2/T_H17 cells that promote immune pathology in the lung. Previously, the presence of T_H2/ T_H17 cells has been associated with more severe disease, and adoptive transfer experiments in mice have demonstrated that these cells are more pathogenic than classic T_H2 cells.^2–4 Furthermore, elevated IL-17A, either from these double producers or classic T_H17 cells, has been linked to more severe disease and steroid resistance in human patients.^5–11 This elevated IL-17A can drive increased pathology directly through recruitment of neutrophils or, as shown by Hall et al., or indirectly by modulating T_H2 pathogenicity by enhancing IL-13/STAT6 signaling.^46,47 Thus a critical future question to address is to identify the effect of the rs6965571 on GIMAP5 expression upon T cell activation and to establish whether patients with this SNP have an increased number of pathogenic (T_H17) CD4⁺ T cells.

While DNA damage occurs normally upon mitogenic stimulation in CD4⁺ T cells, a key finding is the exacerbated DNA damage in cycling Gimap5-deficient CD4⁺ T cells that is associated with reduced T cell survival that is independent of Fas-signaling or intrinsic apoptotic cascades.¹⁴ Importantly, our previous studies identified a key role for Gimap5 in regulating GSK3 activity and subsequent control of the DNA damage response during CD4⁺ T cell proliferation.¹⁴ Here we show that DNA damage was particularly associated with T_H1 and T_H2 polarizing conditions. Interestingly, the presence of TGFβ could limit the extent of DNA damage even in “T_H1” and “T_H2” polarizing conditions. Previous studies have shown that TGFβ signaling can promote DNA damage repair and cell viability by inducing activation of ataxia-telangiectasia-mutated (ATM),^48,49 suggesting that TGFβ signaling can overcome the exacerbated DNA damage observed in activated Gimap5-deficient CD4⁺ T cells. Moreover, we show that initiation of TGFβ signaling even 2 days after initial TCR stimulation is sufficient to reduce cell death and DNA damage in vitro in Gimap5-deficient CD4⁺ T cells. In vivo, this would translate towards improved survival of newly activated CD4⁺ T cells encountering TGFβ in the lung or other TGFβ-rich tissues. We posit that the ability of TGFβ to overcome the survival defects associated with Gimap5-deficiency could affect the CD4⁺ T cell compartment by promoting the selective survival of T_H17 cells and other pathogenic CD4⁺ T cells infiltrating the TGFβ-rich asthmatic lung, where they drive further immune pathology.

Somewhat paradoxically, Gimap5-deficient mice fail to induce iTregs in vivo and retained Tregs have reduced suppressive capacity.^14,15 It remains unclear why TGFβ rescues T_H17 CD4⁺ T cell development, but not iTreg polarization and function, although our previous data suggest that the uncontrolled activity of GSK3β in the absence of Gimap5 may be critical to this process. Several independent studies demonstrate that GSK3β activity promotes T_H17 polarization while impairing iTreg activity, providing one possible mechanism by which T_H17 polarization is favored in the absence of Gimap5.^50–56 Furthermore, Foxo family members, which are poorly expressed in Gimap5-deficient CD4⁺ T cells,¹⁵ are key TFs in Treg development and function.^57,58 Impaired expression of these TFs is predicted to be a critical factor in the inability of Gimap5^sph/sph mice to develop iTregs. This dysbalance between pathogenic CD4⁺ T cells and functional regulatory T cells is one of the primary driving factors behind the immune pathologies associated with Gimap5 deficiency. Unfortunately, the inability of TGFβ to rescue iTreg function, its promotion of the development of pathogenic T_H17 CD4⁺ T cells, and its critical role in the resolution of inflammation makes it an impractical target in clinical applications. However, therapeutic targeting of other links in the GIMAP5-pathway, such as GSK3β, may offer more promise in clinical options for patients with GIMAP5-associated asthma.

Key Messages.

• GIMAP5 LOF mutations in mice and humans result in increased T_H2/T_H17 polarization in vivo.

• Gimap5-deficienct CD4⁺ T cells exhibit increased DNA damage and reduced survival upon activation that can be overcome by TGFβ.

• Gimap5-deficient mice develop more severe allergic airway disease upon HDM sensitization.

Abbreviations used

AHR

Airway hyper responsiveness

BALF

Bronchoalveolar lavage fluid

cKO

conditional knockout

GIMAP5

GTPase of Immunity-associated protein 5

HDM

House dust mite

IT

Intratracheal

iTreg

induced regulatory T cell

LOF

Loss-of-function

γH2AX

phosphorylated histone H2AX

mLN

mesenteric lymph node

mdLN

mediastinal lymph node

PMA

Phorbol 12-myristate-13-acetate

SLE

Systemic lupus erythematosus

SNPs

Single nucleotide polymorphisms

T1D

Type 1 diabetes

WT

Wildtype