Promoted Generation of T Helper 1-Like Regulatory T Cells After Transient Middle Cerebral Artery Occlusion in Type-2 Diabetic Mice

ABSTRACT Background Regulatory T cells (Tregs) play a remarkable role in modulating post-ischemic neuroinflammation. However, the characteristics of Tregs in diabetic ischemic stroke remain unknown. Methods Transient middle cerebral artery occlusion (MCAO) was conducted on leptin receptor-mutated db/db mice and db/+ mice. The number, cytokine production, and signaling features of Tregs in peripheral blood and ipsilateral hemispheres were evaluated by flow cytometry. Treg plasticity was assessed by the adoptive transfer of splenic Tregs into mice. The effect of ipsilateral macrophages/microglia on Treg plasticity was determined by in vitro co-culture analysis. Results db/db mice had more infiltrating Tregs in their ipsilateral hemispheres than db/+ mice. Infiltrating Tregs in db/db mice expressed higher transforming growth factor-β (TGF-β), interleukin-10 (IL-10), forkhead box P3 (Foxp3), interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), and T-box expressed in T cells (T-bet) in comparison to infiltrating Tregs in db/+ mice, suggesting promoted generation of T helper 1 (Th1)-like Tregs in the brains of db/db mice after stroke. The post-ischemic brain microenvironment of db/db mice significantly up-regulated IFN-γ, TNF-α, T-bet, IL-10, and TGF-β in infiltrating Tregs. Moreover, ipsilateral macrophages/microglia remarkably enhanced the expression of IFN-γ, TNF-α, and T-bet but not IL-10 and TGF-β in Tregs. db/db macrophages/microglia were more potent in up-regulating IFN-γ, TNF-α, and T-bet than db/+ macrophages/microglia. Interleukin-12 (IL-12) blockage partially abolished the modulatory effect of macrophages/microglia on Tregs. Conclusion The generation of Th1-like Tregs was promoted in the brains of type 2 diabetic mice after stroke. Our study reveals significant Treg plasticity in diabetic stroke. Abbreviations: Foxp3: forkhead box P3; IFN-γ: interferon-γ; IL-10: interleukin-10; IL-12: interleukin-12; MCAO: middle cerebral artery occlusion; PBS: phosphate-buffered saline; STAT1: Signal transducer and activator of transcription 1; STAT5: Signal transducer and activator of transcription 1; T-bet: T-box expressed in T cells; TGF-β: transforming growth factor-β; Th1: T helper 1; TNF-α: tumor necrosis factor-α; Tregs: regulatory T cells. Foxp3: forkhead box P3; IFN-γ: interferon-γ; IL-10: interleukin-10; IL-12: interleukin-12; MCAO: middle cerebral artery occlusion; PBS: phosphate-buffered saline; STAT1: Signal transducer and activator of transcription 1; STAT5: Signal transducer and activator of transcription 1; T-bet: T-box expressed in T cells; TGF-β: transforming growth factor-β; Th1: T helper 1; TNF-α: tumor necrosis factor-α; Tregs: regulatory T cells.


Introduction
Ischemic stroke is a significant cause of mortality and disability across the globe. Shortly after stroke, neuroinflammation is provoked and results in the activation of mast cells, astrocytes, microglia, and vascular endothelial cells. These cells produce abundant inflammatory mediators to facilitate leukocyte infiltration across the blood-brain barrier (Jayaraj et al. 2019). Acute neuroinflammation is detrimental because it triggers and accelerates secondary neurodegeneration in the penumbra area (Pluta et al. 2021). Infiltrating leukocytes including T cells would exacerbate post-ischemic neuron damage in an antigendependent or independent manner (Selvaraj and Stowe 2017;Zhang et al. 2021). Notably, forkhead box P3 (Foxp3)-expressing CD4 + T cells, also known as Tregs, can mitigate neuroinflammation, protect neurons, and promote functional recovery (Chen et al. 2013;Santamaria-Cadavid et al. 2020;Shi et al. 2021;Wang et al. 2021;Xia et al. 2016). The neuroprotective mechanisms of Tregs in ischemic stroke include: inhibiting the production of neutrophil-derived metalloproteinase-9 and maintaining the integrity of the blood-brain barrier; inhibiting the activation of T cells by producing anti-inflammatory factors such as interleukin-10 (IL-10) and transforming growth factor-β (TGF-β); preventing the inflammatory reaction of microglia/macrophages; inhibiting the activation of neurotoxic astrocytes; and producing osteopontin to support microglia-mediated brain repair (Shi et al. 2021;Wang et al. 2021).
Cumulative studies suggest that Tregs are heterogeneous and plastic in various disorders (Qiu et al. 2020). Plastic Tregs have the traits of T helper (Th) cells, such as the production of Th-associated cytokines and transcription factors. However, these Tregs still express Foxp3 and immunosuppressive mediators (Piconese et al. 2020). Among heterogeneous Treg populations, Th1-like Tregs co-express Foxp3, and T-box expressed in T cells (T-bet), and interferon-gamma (IFN-γ). They participate in the pathogenesis of autoimmune diseases including multiple sclerosis, type I diabetes, and colitis (Di Giovangiulio et al. 2019;Lee 2018). Th1-like Tregs are generated from Tregs under the education of IFN-γ and interleukin-12 (IL-12) (Kitz and Dominguez-Villar 2017). Although Th1-like Tregs are usually considered immunosuppressive and anti-inflammatory, they could become less immunosuppressive or even pro-inflammatory to initiate or exacerbate inflammatory disorders (Di Giovangiulio et al. 2019;Kitz and Dominguez-Villar 2017). To our knowledge, the presence of post-ischemic Th1-like Tregs has not been previously reported.
Hyperglycemia and diabetes are risk factors for not only ischemic stroke but also neuroinflammation. In diabetes mellitus type 1 and type 2 mouse models, hyperglycemia incurs neuroinflammation to compromise the blood-brain barrier (Rom et al. 2019). Furthermore, diabetes accelerates and enhances post-ischemic neuroinflammation (Shukla et al. 2017;Vannucci et al. 2001). In mouse diabetic stroke models, brain damage is exacerbated by impaired leptomeningeal collateral flow (Akamatsu et al. 2015) reduced heat-shock chaperone expression (Tureyen et al. 2011) and elevated pro-inflammatory tumor necrosis factor-α (TNF-α), interleukin-1 (IL-1), interleukin-6 (IL-6), and interferon-γ (IFN-γ) (Shukla et al. 2017;Zhang et al. 2019). Diabetes also alters the activation of macrophages, T cells, natural killer cells, and other immune cell populations after stroke (Shukla et al. 2017). Particularly, profound increases in infiltrating CD4 + T cells and CD8 + T cells occur in type 2 diabetic db/db mice after distal middle cerebral artery occlusion (MCAO) (Zhang et al. 2019). Enhanced post-ischemic neuroinflammation in the brains of diabetic mice would significantly affect the plasticity and function of infiltrating Tregs. However, the characteristics of infiltrating Tregs in diabetic stroke have not been reported.
In this research, we evaluated Treg quantity and function in db/+ mice and db/db mice after transient MCAO. We found that the number of Th1-like Tregs was higher in the ipsilateral hemispheres of db/db mice relative to db/+ mice. These Th1-like Tregs featured substantial expression of IFN-γ, TNF-α, and T-bet. Furthermore, ipsilateral macrophages/ microglia were involved in the generation of Th1-like Tregs. In summary, we discovered Th1-like Tregs in ipsilateral hemispheres and shed light on Treg plasticity after diabetic ischemic stroke.

Mouse strains
The animal research was approved by the China Three Gorges University Animal Care and Use Committee. The surgical procedures were conducted following the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. db/db and db/+ mice (12 weeks old) were purchased from Hunan SJA Laboratory Animal Co., Ltd. The mice were housed in our pathogen-free animal center with a 12-hour light/12-hour dark cycle and 40% humidity. The bedding was autoclaved and changed every week.

Transient MCAO
Transient MCAO was conducted based on the previous method (Shu et al. 2019). Anesthesia was induced by inhalation of 2.0% isoflurane and maintained by 1.0% isoflurane in 70% N 2 O and 30% O 2 on a V-10 Anesthesia system (VetEquip, Inc). A rectal probe was applied to measure core body temperature and a thermostat-controlled heating blanket was used to keep body temperature between 36.5-37.5°C. After shaving the fur and exposing the skin of the ventral neck, the surgical area was sterilized using 70% ethanol. A 1-cm long midline incision was cut on the neck, followed by careful dissection of the common carotid artery, external carotid artery, and internal carotid artery (ICA) from surrounding tissues and nerves. A silicon-coated 6-0 nylon monofilament was placed into the left ICA and moved forward until it occluded the middle cerebral artery (MCA). The occlusion lasted for 90 minutes. Then the monofilament was withdrawn to allow for reperfusion. Cerebral blood flow was monitored by a laser doppler flowmetry (ADInstruments) before and after occlusion and reperfusion. Mice having a residual cerebral blood flow<20% of initial levels during occlusion, cerebral blood flow>80% of initial levels within 10 minutes of reperfusion, and surviving for 3 to 5 days after MCAO were used in the following analysis.

Enrichment of blood leukocytes and ipsilateral immune cells
On post-MCAO day 3 and 5, mice were anesthetized by inhalation of 2.0% isoflurane. Peripheral blood was collected from the tail vein. Erythrocytes were removed by incubating cells in red blood cell lysis buffer (Beyotime) for 5 minutes at room temperature. Blood leukocytes were suspended in phosphate-buffered saline (PBS) and placed on ice before analysis. Each mouse was then trans-cardially perfused with 10 ml of ice-cold PBS. The ipsilateral hemisphere was taken, minced into small pieces, and digested at 37°C for 30 minutes in 1 ml of RPMI1640 medium supplemented with 0.5 mg/ml collagenase IV, 100 µg/ml DNase I, and 5 mM CaCl 2 . The digested hemisphere was pressed through a 70-µm cell strainer and mixed with 4 volumes of 30% percoll (Beyotime). The mixture was placed onto an equal volume of 70% percoll, followed by centrifugation at 500×g for 15 minutes at room temperature. Immune cells floating at the 30% percoll-70% percoll interface were collected and suspended in PBS. In some experiments, enriched blood leukocytes and ipsilateral immune cells were suspended at the density of 1 × 10 5 /ml in RPMI1640 medium supplemented with 10% fetal calf serum (FCS), followed by stimulation with 20 ng/ml phorbol myristate acetate (PMA) plus 1 μg/ml ionomycin and 10 μg/ml brefeldin A (All from Sigma-Aldrich) for 3 hours. After that, cells were subjected to flow cytometry analysis as described below.

Flow cytometry analysis
Fluorophore-conjugated antibodies are summarized in Table 1. To stain cell surface proteins, cells were suspended in PBS at the density of 5 × 10 5 /ml and stained with 2 µg/ml of each antibody on ice for 15 minutes. If there was no further intracellular staining, cells were washed with PBS once and subjected to analysis. For intracellular staining, cells were washed with PBS once and fixed with 200 µl of 3% paraformaldehyde for 10 minutes at room temperature. Then cells were permeabilized with 1 ml of 90% methanol-PBS for 30 minutes on ice. After that, cells were incubated with 2 µg/ml of each antibody for 30 minutes on ice. Cells were analyzed on a BD FACSCalibur™ cytometer or sorted on a BD FACSAria cell sorter (Both from BD Biosciences).

Real-time RT-PCR
RNAs were extracted using an Arcturus picopure RNA isolation kit (ThermoFisher Scientific) based on the vendor's protocol. cDNAs were prepared using a FastKing firststrand synthesis kit (Tiangen Biotech) following the manufacturer's instructions. cDNAs were mixed with a Fast SYBR green master mix (ThermoFisher Scientific). Real-time PCR was carried out on a 7300 Real-Time PCR System (Invitrogen). The relative expression of target transcripts was normalized to β-actin and computed by the 2 −ΔΔCt method. Primer information was summarized in Table 2.

Adoptive transfer
The spleens of normal db/+ and db/db mice were harvested and ground on 70-µm cell strainers to prepare single splenocyte suspensions. Erythrocytes were removed by incubating cells in red blood cell lysis buffer (Beyotime) for 5 minutes at room temperature. Treg-enriched CD4 + CD25 + T cells were isolated from splenocytes using the EasySep™ mouse CD4 + CD25 + regulatory T cell isolation kit II (Stemcell Technologies), following the manufacturer's manual. Treg-enriched CD4 + CD25 + T cells were then suspended at the density of 1 × 10 6 /ml in PBS and labeled with 2 µM carboxyfluorescein succinimidyl ester (CFSE, eBioscience™) at 37°C for 15 minutes in the dark. Labeled CD4 + +CD25 + T cells were washed with PBS once and suspended at the density of 1 × 10 7 /ml in PBS. Meanwhile, MCAO was performed on db/+ and db/db mice. A total of 1 × 10 6 labeled CD4 + CD25 + T cells (in 100 µl of PBS) were then injected into the peritoneal cavity of each db/+ or db/db mouse immediately after MCAO. Four groups were included in the assay: 1) db/+ Tregs transferred into ischemic db/+ recipients; 2) db/+ Tregs transferred into ischemic db/db recipients; 3) db/db Tregs transferred into ischemic db/+ recipients; 4) db/db Tregs transferred into ischemic db/db recipients.
TTGAACTGGCGTTGGAAGCACG CCACCTGTGAGTTCTTCAAAGGC T cells were co-cultured in each well of a 96-well V-shaped-bottom culture plate (Corning) for 24 hours at 37°C, in the presence or absence of 5 µg/ml anti-IL-12 neutralizing antibody (AF-419-SP, R&D Systems). After that, the whole cells were stained with the Pacific blueconjugated anti-CD3 antibody for 20 minutes on ice. CD3 + cells, i.e. Treg-enriched T cells were sorted by flow cytometry and subjected to real-time RT-PCR.

Statistics
Data were presented as means ± standard deviations. The unpaired Student's t-test or oneway ANOVA with post-hoc Tukey HSD test was used for statistical analysis. Each experiment was independently repeated 2 or 3 times. A P-value<.05 was considered statistically significant.

Identification of Tregs after ischemic stroke
db/+ and db/db mice were subjected to MCAO. On day 3 and day 5 after MCAO, the mice were euthanized and their blood and ipsilateral hemispheres were harvested. Infiltrating immune cells were enriched from ipsilateral hemispheres for flow cytometry analysis (Supplementary Figure S1). As illustrated in Figure 1, CD3 + CD4 + T cells were recognized in blood leukocytes and infiltrating immune cells, respectively. Subsequently, Foxp3 + CD25 + T cells, i.e. Tregs, were found in CD3 + CD4 + T cells. We quantified Treg frequencies in blood leukocytes and ipsilateral immune cells. As shown in Figure 2a, the frequency of blood Tregs was significantly higher in db/db mice compared with db/+ mice on either post-MCAO day 3 or day 5. This result was consistent with previous research reporting increased Tregs in mice with defective leptin signaling (Moraes-Vieira et al. 2014). In contrast, db/db mice and db/+ mice showed comparable frequencies of infiltrating Tregs on post-MCAO day 3. However, on post-MCAO day 5, db/db mice had a higher infiltrating Treg frequency than db/+ mice (Figure 2b). Interestingly, the number of total infiltrating immune cells was significantly higher in db/db mice compared to db/+ mice on either post-MCAO day 3 or day 5 (Figure 2c). As a result, the absolute number of infiltrating Tregs was also higher in db/db mice (Figure 2d). In addition, quantification of the mean fluorescence of Foxp3 revealed equivalent Foxp3 expression in blood db/db Tregs and blood db/+ Tregs (Figure 2e,f). However, in db/db mice, infiltrating Tregs expressed higher Foxp3 than their counterparts in db/+ mice on post-MCAO day 3 and day 5, respectively (Figure 2e,f).

T-bet expression and signal transducer and activator of transcription 1 (STAT1) phosphorylation in Tregs
T-bet is up-regulated when Tregs become Th1-like Tregs (Kitz et al. 2018). To confirm the identity of infiltrating Th1-like Tregs, we analyzed T-bet expression in infiltrating CD3 + CD4 + Foxp3 + Tregs. As illustrated in Figure 4a,b, T-bet expression was very low in either blood db/+ Tregs or blood db/db Tregs. However, T-bet was up-regulated in infiltrating db/+ Tregs and infiltrating db/db Tregs. Infiltrating db/ db Tregs expressed considerably higher T-bet than infiltrating db/+ Tregs. Because STAT1 phosphorylation is crucial for Th1 differentiation, we quantified phosphorylated STAT1 in Tregs. As shown in Figure 4c,d, phosphorylated STAT1 was minimal in either blood db/+ Tregs or blood db/db Tregs but increased in infiltrating db/+ Tregs and infiltrating db/db Tregs. However, the levels of phosphorylated STAT1 were not significantly different between infiltrating db/+ Tregs and infiltrating db/db

Induction of Th1-like Tregs in the ipsilateral microenvironment
We wonder whether the ipsilateral microenvironment of db/db mice was more capable of inducing Th1-like Tregs. To this end, we enriched splenic CD4 + CD25 + T cells (mostly Tregs) from normal db/+ mice and db/db mice, respectively. These cells were labeled with CFSE and intraperitoneally injected into db/+ mice or db/db mice immediately after MCAO was done to the recipients. Five days after MCAO, CFSE-labeled T cells, i.e. donorderived Tregs, were retrieved from the recipients' ipsilateral hemispheres by flow cytometry (Figure 5a), followed by quantification of transcripts of IFN-γ, TNF-α, and T-bet. As shown in Figure 5b and Table 3, db/+ donor-derived Tregs and db/db donor-derived Tregs expressed comparable levels of IFN-γ in db/+ recipients. In addition, in db/db recipients, both db/+ donor-derived Tregs and db/db donor-derived Tregs expressed higher IFN-γ than their counterparts in db/+ recipients. Notably, in db/db recipients, db/+ donorderived Tregs and db/db donor-derived Tregs expressed comparable IFN-γ. Similar changes in TNF-α and T-bet were observed (Figure 5b and Table 3). The transcripts of IL-10 and TGF-β also exhibited similar changes as IFN-γ (Figure 5c and Table 3). Tregs. db/+ to db/+: db/+ Tregs transferred into db/+ recipients. db/db to db/+: db/db Tregs transferred into db/+ recipients. db/+ to db/db: db/+ Tregs transferred into db/db recipients. db/db to db/db: db/db Tregs transferred into db/db recipients. N = 6 recipients per group. *: P < .05. **: P < .01. ***: P < .001. ns: not significant. One-Way ANOVA.  However, Foxp3 expression was comparable in all four groups (Figure 5c and Table 3). Therefore, compared with the ipsilateral microenvironment of db/+ recipients, the ipsilateral microenvironment of db/db recipients induced more Th1-like Tregs while enhancing the expression of IL-10 and TGF-β in Tregs.

Discussion
Tregs are reported to be neuroprotective in ischemic stroke by alleviating neuroinflammation (Liesz et al. 2013;Stubbe et al. 2013;Xie et al. 2014Xie et al. , 2015. However, some studies imply that Tregs are not neuroprotective or even detrimental in ischemic stroke (Kleinschnitz et al. 2013;Schuhmann et al. 2015). This controversy might be caused by Treg heterogeneity and plasticity. Under the influences of various cytokines, Tregs may differentiate into IFN- γ-expressing Th1-like Tregs or IL-17-expressing Th17-like Tregs (Ali et al. 2020;Qiu et al. 2020). Since IFN-γ and IL-17 are pro-inflammatory cytokines, Th1-like Tregs and Th17like Tregs would exacerbate pathological inflammatory disorders. In atherosclerosis, dysfunctional Th1-like Tregs that may permit further arterial inflammation and atherogenesis have been reported (Butcher et al. 2016). If Th1-like Tregs are present in stroked brains, post-ischemic neuroinflammation might be worsened. However, to our knowledge, no previous research has demonstrated the presence of Th1-like Tregs in ischemic stroke. In this study, we briefly characterized the phenotype and function of Th1-like Tregs in the stroked brains of diabetic and non-diabetic mice. Consistent with a previous report displaying markedly increased Treg number in db/db mice (Taleb et al. 2007) we also found a higher Treg frequency in the blood of db/db mice. This increase might be caused by the mutated leptin receptor in db/db mice, because leptin prevents Treg growth whereas defective leptin signaling elevates peripheral Tregs (De Rosa et al. 2007;Matarese et al. 2010). Besides, we also observed more immune cells in db/db mice after MCAO, consistent with previous studies demonstrating increased leukocyte infiltrates in db/db mice after MCAO (Tureyen et al. 2011;Zhang et al. 2019). We noticed that compared with blood db/+ Tregs, blood db/db Tregs expressed comparable Foxp3 but higher IL-10 and TGF-β, suggesting that db/db Tregs were more anti-inflammatory than db/+ Tregs. This finding is in agreement with former research showing improved Treg function in db/db mice (Taleb et al. 2007). Importantly, our data indicate that some infiltrating Tregs co-expressed IL-10, TGF-β, IFN-γ, and TNF-α in both db/db mice and db/+ mice, suggesting the presence of Th1-like Tregs in stroked brains of either non-diabetic or diabetic mice. To our knowledge, this is the first study showing Th1-like Tregs in ischemic stroke. Moreover, infiltrating db/db Tregs produced higher IL-10, TGF-β, IFN-γ, TNF-α, and T-bet than infiltrating db/+ Tregs. Therefore, both pro-inflammatory and anti-inflammatory cytokines were likely enhanced in infiltrating db/db Tregs. The adoptive transfer assay suggests that the microenvironment of stroked db/db brains robustly boosted the expression of IFN-γ, TNF-α, T-bet, IL-10, and TGF-β in infiltrating Tregs, no matter the sources of Tregs. The strong post-ischemic neuroinflammation in db/db mouse brains may trigger higher activation of the antiinflammatory activity of infiltrating Tregs but simultaneously induces Treg differentiation towards pro-inflammatory Th1-like Tregs. However, the exact factors responsible for these changes remain unclear.
This study also revealed that ipsilateral macrophages/microglia produced IL-12 to promote the expression of IFN-γ, TNF-α, and T-bet in Tregs. In the acute phase of stroke, reactive macrophages/microglia are polarized towards the pro-inflammatory M1 type (Boddaert et al. 2018;Mao et al. 2021). IL-12 is a hallmark of M1 macrophages and can trigger Th1-like Treg generation (Kitz and Dominguez-Villar 2017). Therefore, M1 macrophages/microglia contribute to the generation of Th1-like Tregs in stroked brains. Furthermore, we found that in comparison with db/+ macrophages/microglia, db/db macrophages/microglia expressed more IL-12 and thus induced more Th1-like Tregs. This result is again consistent with previous research indicating more severe neuroinflammation after cerebral Ischemia in diabetic mice (Zhang et al. 2019). Surprisingly, macrophages/microglia did not affect IL-10 expression and decreased TGF-β, suggesting that they were not responsible for the higher IL-10 and TGFβ in Figure 5. Besides macrophages/ microglia, multiple cell types including mast cells, conventional CD4 + T cells, CD8 + T cells, γδ T cells, dendritic cells, neutrophils, astrocytes, and vascular endothelial cells participate in post-ischemic neuroinflammation. One type or several types of these cells might robustly induce IL-10 and TGF-β expression in infiltrating Tregs, thus overwhelming the effect of macrophages/microglia on IL-10 and TGF-β. However, this hypothesis needs to be checked in future studies.
Although we showed Treg heterogeneity in stroked brains, the significance of Th1-like Tregs to neuroprotection is still unidentified. Th1-like Tregs might exacerbate inflammation due to the secretion of IFN-γ and TNF-α. However, anti-inflammatory IL-10 and TGFβ were also elevated in infiltrating Tregs. The outcome of stroke might be determined by the balance between Tregs and Th1-like Tregs, as well as the levels of cytokines produced by Th1-like Tregs. In the future, it is necessary to use Treg-specific T-bet-overexpressing or T-bet-deficient mice to identify the impact of Th1-like Tregs on neuroinflammation after stroke.

Conclusions
This study discovered Th1-like Tregs in the stroked brains of both control mice and db/db mice, shedding light on Treg plasticity after ischemic stroke. It also identified ipsilateral macrophages/microglia as the inducers of Th1-like Tregs, especially in db/db mice. Our findings deepen the understanding of Treg plasticity in diabetic stroke.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This study was supported by the Natural Science Foundation of Hubei Province (Grant# WJ2017×020).

Data availability statement
The data that support the findings of this study are available from the corresponding author, [L.S.], upon reasonable request.