Autoantibody against β1-adrenoceptor promotes the differentiation of natural regulatory T cells from activated CD4+ T cells by up-regulating AMPK-mediated fatty acid oxidation

Therapeutic adoptive transfer of natural regulatory T cells (nTreg, CD4+ CD25+ Foxp3+ T cells) or in vivo selective expansion of nTreg cells has been demonstrated to improve the cardiac function in various cardiovascular disease models. The differentiation of nTreg cells is mediated by catecholamines via β1-adrenergic receptor (β1-AR) activation. Autoantibody against β1-adrenoceptor (β1-AA) as a β1-AR agonist is closely associated with the occurrence and deterioration of cardiac dysfunction. However, whether β1-AA has any impact on nTreg cells has not been reported. The aim of the present study was intended to assess the potential impact of β1-AA on nTreg cell differentiation and explore the underlying mechanism. It was found that the expression of multiple proteins involved in nTreg cell differentiation, immunosuppressive function, and migration was up-regulated in mice after β1-AA administration, suggesting that β1-AA may promote nTreg cell activation. In vitro, β1-AA promoted nTreg cell differentiation by up-regulating mitochondrial fatty acid oxidation (FAO) in activated CD4+ T cells via AMP-activated protein kinase (AMPK) activation and mitochondrial membrane potential reduction. In addition, the AMPK agonist facilitated β1-AA-mediated FAO and nTreg cell differentiation. To further confirm the role of AMPK in β1-AA-mediated nTreg cell differentiation, β1-AA was acted on the CD4+ T cells isolated from AMPK-deficient (AMPK−/−) mice. The result showed that the effect of β1-AA on nTreg cell differentiation was attenuated markedly after AMPK knockout. In conclusion, AMPK-mediated metabolic regulation targeting for nTreg cell restoration may be a promising therapeutic target for β1-AA-positive patients with cardiac dysfunction.


Introduction
CD4 + T cells are known as the most important participant in adaptive immunity of the organism. Over-activation of CD4 + T cells and disproportion of their subpopulations play an important role in the pathogenesis of various cardiovascular diseases. Functionally, CD4 + T cells are classified as two major categories: effector T cells and regulatory T (Treg) cells 1 , among which natural Treg (nTreg, CD4 + CD25 + Foxp3 + T) cells play a critical role in inhibiting the immune response of effector T cells and maintaining immune tolerance 2,3 . Therapeutic adoptive transfer of nTreg cells or in vivo selective nTreg cell expansion has been demonstrated to attenuate post-infraction left ventricular remodeling, relief myocardial injury, and eventually improve the cardiac function in diverse cardiovascular disease models 4,5 . Studies have confirmed that the development and function of nTreg cells are regulated by catecholamines via the expression of α-, β 1 -, and β 2 -adrenergic receptors (β 1/2 -ARs) [6][7][8] . Compared with effector T cells, β 1 -AR expression in nTreg cells is more advantageous than β 2 -AR expression 8 , but the effect of β 1 -AR activation on nTreg cells remains unclear.
Autoantibody targeting the second extracellular loop of β 1 -adrenoceptor (β 1 -AA) is commonly detected in circulating blood of the patients with cardiac dysfunction caused by etiologies like dilated cardiomyopathy, ischemic heart disease, and arrhythmia [9][10][11] . β 1 -AA was found to exhibit the agonist-like effects on β 1 -AR, such as increasing the intracellular calcium level promoting the beating frequency of neonatal rat cardiomyocytes and inducing cAMP production [12][13][14] . The positive rate of β 1 -AA was reported to be as high as 80% in different cardiac dysfunction models 15 . Moreover, LVEF of the cardiac dysfunction patients improved obviously after removing β 1 -AA by immunoadsorption (IA) treatment 16 . However, it is not elucidated about the underlying mechanism related to β 1 -AA-induced cardiac dysfunction. Our previous and other studies found that in β 1 -AA-positive murine, not only the cardiac function was decreased but accompanied by an increase in the peripheral CD4 + / CD8 + T cell ratio; in addition, part of the myocardium was infiltrated by large number of T cells 17 . In vitro, β 1 -AA isolated from the sera of cardiac dysfunction patients promoted proliferation of CD4 + T cells through the β 1 -AR/cAMP pathway 14 . Furthermore, accompanied by cardiac function improvement of the β 1 -AA-positive cardiac dysfunction after IA treatment, the number of circulating nTreg cells increased significantly 18,19 . It was shown that nTreg cell proportion in rat peripheral blood was inhibited by β 1 -AR blocker propranolol 20 . However, whether β 1 -AA as a agonist-like substance of β 1 -AR can exert a direct effect on nTreg cells has not been reported. Therefore, the present study was intended to assess the potential impact of β 1 -AA on nTreg cell activation and differentiation, and the underlying mechanism was explored in an attempt to etiologically find a potential therapeutic target for β 1 -AA-positive cardiac dysfunction patients.
A metabolic shift toward fatty acid metabolism was associated with the increased nTreg cell differentiation induced by β 1 -AA The mechanism underlying β 1 -AA-mediated nTreg differentiation was explored in the further experiments. Knowing that mitochondrial fatty acid oxidation (FAO) is a decisive factor for CD4 + T cell differentiation, which promotes CD4 + T cell differentiation towards Treg cells as opposed to an effector phenotype 26,27 . To assess whether FAO was affected by β 1 -AA in CD4 + T cells, the uptake of palmitate was measured in the course of stimulation. Flow cytometry demonstrated that the absorption of palmitate was increased in activated CD4 + T cells with anti-CD3/ CD28 mAbs after β 1 -AA stimulation (Fig. 3a). Therefore, we postulated that the metabolic alteration in activated CD4 + T cells induced by β 1 -AA participated in enhanced nTreg cell differentiation. Etomoxir, a selective inhibitor of carnitine palmitoyltransferase I 28 , was used to confirm our hypothesis. Indeed, it was found that etomoxir was able to reverse the effect of β 1 -AA in nTreg cell differentiation (Fig. 3b). Thus, the enhanced nTreg cell differentiation elicited by β 1 -AA is accompanied by a metabolic shift toward FAO in activated CD4 + T cells and is reversible by FAO inhibitor. AMP-activated protein kinase (AMPK) activation is known to enhance mitochondrial FAO in response to decreased ATP level 29 , which is a crucial pathway attributable to nTreg cell differentiation 27,30 . AMPK is mainly activated by increased AMP level and phosphorylation of a threonine residue (Thr-172) 31 . To investigate the role of AMPK in β 1 -AA-induced nTreg cell differentiation, AMPK phosphorylation and ATP levels were estimated in the primary CD4 + T cells isolated from the splenic tissue of β 1 -AA-positive mice. Significant decrease in ATP levels has been observed in the primary CD4 + T cells since the fourth week of β 1 -AR mAb administration until the 12th week (Fig. 4a), accompanied by constantly increased Thr (172)-AMPKα phosphorylation (Fig. 4b). Unlike the expression changes in the inhibitory phosphorylation site serine 491 32 , which increased at the fourth week, and then decreased at the eighth week (Supplemental Fig. 2), Thr (172)-AMPKα phosphorylation have been constantly increased since the fourth week of β 1 -AR mAb administration until the 12th week. What is more, the direct AMPK activator, 5-aminoimidazole-4-carboxamide riboside (AICAR) facilitated β 1 -AA-mediated nTreg cell differentiation in vivo (Fig. 4c) and promoted palmitate absorption in activated CD4 + T cells with anti-CD3/ CD28 mAbs (Fig. 4d). Metformin, known as a indirect activator of AMPK by lowering the energy supply 30 , exhibited similar effects on β 1 -AA-mediated nTreg cell differentiation and palmitate absorption ( Fig. 4c-d).
To further explore the role of AMPK in β 1 -AAinduced nTreg cell differentiation, β 1 -AA was utilized Fig. 2 The proportion of nTreg cells among CD4 + T cells after β 1 -AA stimulation. a Percentage of nTreg cells in the preactivated CD4 + T cells at the 24th hour after β 1 -AA stimulation with or without metoprolol. b Percentage of nTreg cells in quiescent CD4 + T cells at the 24th hour after β 1 -AA stimulation. Data are presented as means ± SD (n = 5 per group). **P < 0.01 vs. vehicle group; ## P < 0.01 vs. 10 −8 mol/L β 1 -AA group on the CD4 + T cells isolated from AMPK-deficient (AMPKα2 −/− ) mice ( Fig. 5a, b). Percentage of the circulating CD4 + CD25 + Treg cells in AMPKα2 −/− mice was lower than that in wild=type mice, and it was further reduced by 4-week β 1 -AA administration (Fig. 5c). In vitro, the result showed that nTreg cell proportion in the preactivated AMPK −/− CD4 + T cells was lower than that in CD4 + T cells isolated from wild-type mice after β 1 -AA stimulation (10 −7 mol/L) for 24 h (Fig. 5d). The evidence suggesting that knockout of the AMPKα2 gene decreased the effect of β 1 -AA in promoting nTreg cell differentiation. To sum up, these data demonstrate that AMKP activation plays a moderate positive role in nTreg cell differentiation mediated by β 1 -AA.

Enhancement of fatty acid metabolism mediated by MMP reduction promoted β 1 -AA-induced nTreg cell differentiation
The mitochondrial FAO is closely associated with mitochondrial membrane potential (MMP) level. When MMP is reduced, intracellular absorption and utilization of palmitate are enhanced, and FAO is up-regulated 33,34 .
To determine whether β 1 -AA-induced nTreg cell differentiation resulted from MMP alteration in CD4 + T cells, cyclosporin A, an MMP stabilizer, was used in the course of β 1 -AA stimulation. It was found that the enhanced palmitate absorption elicited by β 1 -AA in activated CD4 + T cells with anti-CD3/CD28 mAbs was reversed drastically by cyclosporin A and metoprolol (Fig. 6a). As shown by JC-1 staining, a concurrent reduction in the MMP of preactivated CD4 + T cells was found in the β 1 -AA group (Fig. 6b, c). Subsequently, cyclosporin A inhibited the elevation of nTreg cell differentiation mediated by β 1 -AA (Fig. 6d). Therefore, enhanced fatty acid metabolism mediated by MMP reduction in preactivated CD4 + T cells is one of the mechanisms underlying nTreg cell differentiation induced by β 1 -AA.
Discussion β 1 -AA was first identified in the sera of patients with dilated cardiomyopathy by Wallukat et al. in 1987 35 . Subsequently, ample evidence has confirmed the pathogenic effect of β 1 -AA in cardiac dysfunction 9,11,15 . However, both our study and others found that β 1 -AA was positive in about 10% of healthy individuals of different age groups [36][37][38] , suggesting that β 1 -AA may participate in maintaining physiological homeostasis, although the correlative mechanism is unclear. nTreg cells play a very important role in maintaining the balance of the immune system by inhibiting effector T cells 2,3 , and their differentiation and function are regulated by the sympathetic nervous system 7,8 . Nevertheless, whether β 1 -AA as a agonist-like substance of β 1 -AR, could exert a direct effect on nTreg cells has not been reported. Therefore, the present study sought to assess the potential impact of β 1 -AA on nTreg cell differentiation and explore the underlying mechanism. It was found that β 1 -AA promoted nTreg cell differentiation by up-regulating fatty acid metabolism in activated CD4 + T cells via the AMPK pathway and MMP reduction.
To study the effect of β 1 -AA on Treg cells, a passive immunization mouse model was established successfully with the highly active and purified β 1 -AR mAb (Supplemental Fig. 3). Naïve CD4 + T cells differentiate into different subsets (Th1, Th2, Th17, or Treg cells) to establish immune tolerance and defense against pathogens. To quantify CD4 + T cell-related cytokine levels, the levels of Th1 cytokine (IFN-γ), Th2 cytokine (IL-4), Th17 cytokine (IL-17), and Treg cytokine (IL-10) in the  Fig. 4D) increased in mice 8 weeks after β 1 -AA administration, indicating that β 1 -AA promoted a systemic activation of CD4 + T cell in vivo. What is more, the expression of multiple proteins related to nTreg cell differentiation, Fig. 4 The role of AMPK in β 1 -AA-induced nTreg cell differentiation. ATP levels (a) and Thr(172)-AMPKα phosphorylation (b) were estimated in the primary CD4 + T cells isolated from the splenic tissue of β 1 -AA-positive mice at different time points during β 1 -AR mAb administration (n = 4 per group). c Percentage of nTreg cells in activated CD4 + T cells after β 1 -AA stimulation with or without AICAR (n = 5 per group). d The absorption level of palmitate in activated CD4 + T cells after β 1 -AA stimulation with or without AICAR/metformin (n = 5 per group). Data are presented as means ± SD. a, b: **P < 0.01 vs. 0 week since β 1 -AR mAb administration; ## P < 0.01 vs. the eighth week. c, -d **P < 0.01 vs. vehicle group; ## P < 0.01 vs. β 1 -AA group immunosuppressive function, and migration increased in mice peripheral blood, suggesting that β 1 -AA was able to promote nTreg cell activation. However, ultrasound analysis showed that β 1 -AA-induced cardiac dysfunction in mice, as illustrated by decreases in LVEF (Supplemental Fig. 5A), fractional shortening (Supplemental Fig. 5B), and cardiac output (Supplemental Fig. 5C), accompanied with a decreased proportion of circulating CD4 + CD25 + Treg cells (Supplemental Fig. 6). These results are consistent with the finding of many other studies that Treg cell frequency in cardiac dysfunction patients was decreased significantly [39][40][41] , and the number of Treg cells was positively correlated with LVEF, and negatively correlated with the NT-proBNP level 42 . Nevertheless, increased Treg cell infiltration was observed in the myocardium of mice with cardiac dysfunction 43 . For this reason, we explored whether β 1 -AA had a direct effect on nTreg cell differentiation in our subsequent experiment in vitro.
It was found that β 1 -AA induced a metabolic shift towards FAO in activated CD4 + T cells, thus promoting nTreg cell differentiation. In addition, the effect of β 1 -AA in promoting nTreg cell differentiation could be reversed drastically by the β 1 -AR-specific blocker metoprolol. Other studies also demonstrated that atecholamines such as epinephrine and norepinephrine increased the  26,27,45 . In other words, Treg cell differentiation depends on FAO, and we found Fig. 6 Effect of mitochondrial membrane potential (MMP) on β 1 -AA-induced nTreg cell differentiation. a The absorption level of palmitate in activated CD4 + T cells after β 1 -AA stimulation with or without cyclosporin A. b, c Percentages of JC-1 aggregate and JC-1 monomer in activated CD4 + T cells after β 1 -AA stimulation. d Percentage of nTreg cells in activated CD4 + T cells after β 1 -AA stimulation with or without cyclosporin A. Data are presented as means ± SD (n = 5 per group). **P < 0.01 vs. vehicle group; ## P < 0.01 vs. β 1 -AA group that the FAO inhibitor etomoxir reversed β 1 -AA-mediated nTreg cell differentiation. MMP which effects fatty acid uptake 33,34 and AMPK activation 29,30,46 are two pivotal regulator for FAO. Furthermore, the underlying mechanism involved in β 1 -AA-mediated nTreg cell differentiation was explored.
Indeed, β 1 -AA reduced MMP of activated CD4 + T cells, and the MMP stabilizer cyclosporin A drastically reversed β 1 -AA-induced fatty acid absorption enhancement and nTreg cell differentiation. Our previous study 29,30,47 demonstrated that β 1 -AA-induced cardiomyocyte apoptosis by reducing MMP. Similarly, the present study demonstrated that the level of CD4 + T cell apoptosis was increased significantly after β 1 -AA stimulation shown by Annexin V-FITC and propidium iodide double staining (Supplemental Figure 7). These findings suggest that β 1 -AA promoted nTreg cell differentiation through upregulating FAO and reducing MMP, and this effect is closely associated with β 1 -AA-induced CD4 + T cell apoptosis. In addition, MMP reflects the integrity of mitochondrial function and is a key indicator of mitochondrial function 48 . The mitochondrial function alteration may participate in nTreg cell differentiation elicited by β 1 -AA.
AMPK is a key regulatory molecule in response to energy deprivation of the organism 29 . It provides energy quickly by promoting FAO and inhibiting the activity of acetyl coenzyme A carboxylase 49,50 . AMPK-dependent metabolic regulation plays an important role in Treg cell differentiation 30,46 . We found that the ATP level in the primary CD4 + T cells isolated from the splenic tissue of β 1 -AA positive mice was significantly lower than that in the vehicle group, which was accompanied by enhanced Thr(172)-AMPKα phosphorylation. It was reported that AICAR, a pharmacological analog of AMPK 30,46 , promoted Treg cell differentiation without affecting effector T cells 51 . Besides, norepinephrine induced AMPK activation via the cAMP/β-AApathway 52 . The present study showed that both the direct AMPK activator AICAR and the indirect AMPK activator metformin facilitated β 1 -AAmediated FAO enhancement and nTreg cell differentiation in vitro. Moreover, AICAR promoted IL-2 level in the supernatant of activated CD4 + T cells after β 1 -AA stimulation, which is crucial for nTreg cell differentiation (Supplemental Figure 8). To further confirm the role of AMPK in β 1 -AA-mediated nTreg cell differentiation, β 1 -AA was acted on the CD4 + T cells isolated from the AMPKα2 −/− mice. It was found that knockout of the AMPKα2 gene reduced the effect of β 1 -AA in promoting nTreg cell differentiation markedly, confirming that AMPK-induced FAO is a key mechanism underlying β 1 -AA-mediated nTreg cell differentiation.
According to the published paper 52,53 , the β 2 -AR/ cAMP/PKA pathway played a positive moderate role in the immunosuppressive activity of Treg cells. Indeed, cAMP levels increased in the supernatants of nTreg cells after 30-min β 1 -AA stimulation, which is also the downstream signal molecule of β 1 -AR (Supplemental Figure 9A). However, immunofluorescence staining showed that the fluorescein-labeled β 1 -AA was incapable of binding to β 2 -ARs on Treg cells compared to the anti-β 2 -AR mAb (Supplemental Figure 9B). Moreover, by contrast to the activation effect of β 2 -AR pathway, proliferation assay of the CD4 + CD25 − T effector cells revealed that compromised suppressive activity of nTreg cells were resulted from 48h β 1 -AA administration (Supplemental Figure 10). The inhibitory effect of nTreg cells on Teff cells is mainly mediated by IL-10 production 54 . In view of the fact that nTreg cell dysfunction may induced by β 1 -AA, IL-10 level was measured in the supernatant of nTreg cells after β 1 -AA stimulation. It was found that IL-10 secretion from nTreg cells was suppressed by β 1 -AA at concentrations of 10 −6 and 10 −7 mol/L compared with vehicle groups (Supplemental Figure 11). However, 10 −8 mol/L β 1 -AA had a remarkable opposite effect, and 10 −9 mol/L β 1 -AA did not appear to be a factor (Supplemental Figure 11). The evidences above indicated that β 1 -AA had bidirectional impact on the immunosuppressive function of nTreg cells. Yet, the decline in cardiac function induced by β 1 -AA (Supplemental Fig. 5) seems to outweigh the potentially beneficial effects on nTreg cell restoration. Nevertheless, like many other physiological processes, the influence of β 1 -AA on organism is a double-edged sword with therapeutic potential that is associated with the concentration of β 1 -AA.

Limitation and clinical perspective
Till now, the influence of β 1 -AR gene knockout on Treg maturation and function has not been reported. In order to investigate the role of β 1 -AR in β 1 -AA-induced nTreg differentiation, our lab had already acquired three pairs of homozygous β 1 -AR gene knockout mice (C57BL/6J background) from Nanjing BioMedical Research Institute of Nanjing University recently. However, yet the number of available transgenic mice is not sufficient to build our model.
Cardiac dysfunction associated with myocardial injury triggers β 1 -AA generation in different cardiac dysfunction models, and the positive rate of β 1 -AA is nearly 80% 14 . However, there is no specific and effective therapeutic strategy for β 1 -AA-positive patients. β 1 -AR blockers cannot entirely reverse the injurious effect of β 1 -AA on cardiomyocytes 55,56 . The present study showed that the impacts of β 1 -AA on nTreg cell differentiation cannot be fully counteracted by β 1 -AR-specific blocker metoprolol (Fig. 2a, Fig. 3a, and Fig. 6a-c), indicating that there are other mechanisms involved except for a receptor pathway.
It is therefore an urgent task to find a more effective therapeutic target specific for β 1 -AA-positive cardiac dysfunction patients. The present study demonstrated that AMPK-mediated metabolic regulation targeting for nTreg cell restoration might be a promising therapeutic target for β 1 -AA-positive cardiac dysfunction patients.

Synthesis and identification of β 1 -AR mAb
The sequence (amino-acid residues 197-222) of the second extracellular loop of the β 1 -AR: H-W-W-R-A-E-S-D-E-A-R-R-C-Y-N-D-P-K-C-C-D-F-VT-N-R-C was synthesized by solid-phase method using an automated peptide synthesizer. Subsequently, 0.5 mg synthetic polypeptide was coupled with the carrier protein keyhole limpet hemocyanin and bovine serum albumin (BSA) to acquire immunogenicity. The coupled polypeptide was applied to BALB/c mice to create active immunization and induce the production of β 1 -AR-ECII-specific antibodies. Finally, these specific antibodies were fused with the hybridoma cell line to synthesize mAbs specific to β 1 -AR-ECII. The synthesis of β 1 -AR-ECII peptide was conducted by Qiang Yao Bio Scientific Commercial Development Co., Ltd (Shanghai, China), and the hybridoma cells secreting β 1 -AR mAb were constructed by AbMax Biotechnology Co., Ltd (Beijing, China).
To induce the generation of ascites containing β 1 -AR mAb, log-phase hybridoma cells were injected intraperitoneally to female BALB/c mice aged 10 weeks at a dose of 10 6 cells per mL, 0.5 mL per mouse biweekly. The ascites was collected and then purified by using Protein G Affinity Chromatography Column (GE Healthcare Life Sciences, USA). The specificity and activity of the purified β 1 -AR mAbs were determined by enzyme-linked immunosorbent assay (ELISA) and the neonatal mouse cardiomyocyte beating experiment, respectively.

ELISA
The specificity of the purified β 1 -AR mAb and the OD value of β 1 -AA in mice serum were detected by ELISA. Briefly, the β 1 -AR-ECII peptide was dissolved in 100 mM 10 µg/mL Na 2 CO 3 solution (pH = 11.0) at 4°C overnight. The embedded 96-well plate was incubated with 1% BSA at 37°C for 1 h, and then cultured with the primary antibody. Biotin-labeled anti-mouse immunoglobulin G (IgG) was diluted with the sealing solution at a ratio of 1:3000 and cultured at 37°C for 1 h. Horseradish enzyme-labeled streptavidin was diluted at a ratio of 1:2000 and cultured at 37°C for 1 h. The substrate ABTS (2,2′-azino-di-(ethyl-benzthiazoline) sulfonic acid) was dissolved in the substrate buffer with a final concentration of 1.1 mmol/L and cultured at 37°C for 30 min. The optical density (OD) value of each well was measured at 405 nm. The titer of β 1 -AR mAb was determined by the positive/negative (P/N) ratio using the following equation: P/N = (sample OD-blank control OD)/(positive control OD-blank control OD). The positivity or negativity of β 1 -AA was determined by P/N ≥ 2.1 or P/N ≤ 1.5, respectively. Twenty-four C57BL/6 mice which were β 1 -AA negative confirmed by ELISA were equally randomized to three groups: vehicle group, β 1 -AR mAb group, and negative IgG group. Mice in β 1 -AR mAb group received intraperitoneal injection of β 1 -AR mAb at a dose of 5 μg/g biweekly. Mice in the vehicle group received the same dose of normal saline, and the mice in the negative IgG group received the same dose of negative IgG.

Protein microarray chip analysis
The expression of Treg cell-related proteins and cytokines in the β 1 -AA-positive mice was detected using a biotin-labeled mouse protein chip reagent kit. Briefly, each chip well was added with 100 µL sealing solution, cultured on the rocking bed for 30 min at room temperature. After sucking out the sealing solution, each well was added with 100 µL serum, cultured by oscillation at 4°C overnight, and centrifuged at 13,000 rpm for 8 min. Each well was added with 70 µL biotin-labeled antibody and cultured at room temperature for 1 h after two washes of the plate. Then, 70 µL fluorant Cy3-streptomycin avidin was added to each well and cultured by oscillation at room temperature for 2 h. Finally, the fluorescent signal was detected by the Cy3 or green channel (532 nm).

Flow cytometric sorting for CD4 + T cells
Specific fluorescent antibody-labeled CD4 + T cells were separated from mouse spleen mononuclear cells by flow cytometry. C57BL/6 mice aged 10 weeks were euthanized by cervical dislocation to isolate splenocytes. Then, a single-cell suspension was prepared by using mechanical trituration method where the tissue was ground through a 300-mesh sieve. Subsequently, mouse spleen mononuclear cells were re-suspended using 50% and 70% Percoll separating media (GE Healthcare Life Sciences, USA). Density-gradient centrifugation was undertaken at 2500 rpm for 25 min following red blood cell lysis. The steps mentioned above were performed at a fast pace, at 4°C or on ice. For cell surface staining, the antibodies (FITCanti-CD4) were incubated with the single-cell suspension for at least 30 min at 4°C. FITC-anti-CD4 antibody was purchased from BD Bioscience (USA). Cells were sorted with the flow cytometer FACS Aria II (Becton, Dickinson and Company).
Effect of β 1 -AA on CD4 + T cell absorption of palmitic acid and subsequent FAO BODIP is a fat-soluble fluorescent probe. Coupling of BODIP and palmitate can be used to observe the cellular FAO level (Life Technologies, USA). A 5 mmol stock solution was prepared by dissolving the BODIPY-palmitate into dimethyl sulfoxide, and then diluted in PBS buffer to a working concentration of 0.5 µmol. All intervention factors and BODIPY-palmitate were added to the activated CD4 + T cells and cultured at 37°C for 48 h. Finally, the fluorescent intensity of each tube was detected with the FACS Aria II flow cytometer (Becton, Dickinson and Company).

Measurement of the ATP content
The ATP content in CD4 + T cells of β 1 -AA-positive mice was determined using a kit purchased from Beyotime Institute of Biotechnology (China) according to the manufacturer's protocol.

Western blotting
The expression of AMPK in CD4 + T cells was determined by Western blot analysis. CD4 + T cells were separated from the splenic tissues of the β 1 -AA-positive mice and immediately lysed. The supernatant protein was extracted by centrifugation. The supernatant protein was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis at a 50 μg sample volume. After electrophoresis, the PVDF membranes were transferred and blocked with 5% non-fat milk powder in TBST buffer for 1 h, and then incubated with anti-AMPKα2 mAb (1:1000; Abcam, Cambridge, UK), anti-phospho-AMPKα2 (Thr-172) mAb (1:1000; Abcam, Cambridge, UK) or antiphospho-AMPKα2 (S491) mAb (1:1000; Abcam, Cambridge, UK) or anti-GAPDH mAb (Cell Signaling Tech., Danvers, MA, USA) at 4°C overnight. The membranes were incubated with the corresponding secondary antibodies. Finally, the grayscale values of the straps were analyzed by Image J software after development.

JC-1 staining
When the cellular MMP remained stable, JC-1 aggregated in the mitochondrial matrix, producing red fluorescence; when MMP was reduced, JC-1 was present as a monomer in the cellular matrix, producing green fluorescence. Therefore, changes in MMP can be detected by observing the percentage of the red and green fluorescence. Briefly, 48 h after stimulation of activated CD4 + T cells (10 6 cells per well) with β 1 -AA, cells were resuspended in 0.5 mL RPMI medium 1640 (Hyclone, USA). After the addition of 0.5 mL JC-1 dye, cells were cultured in a 37°C incubator for 20 min. Then, cells were re-suspended by the addition of 300 μL JC-1 staining buffer after being washed with JC-1 staining buffer in a centrifuge at 600 × g and 4°C for 3 min twice. Finally, the red and green fluorescent intensity of each tube were detected and analyzed using the FACS Aria II flow cytometer.

Statistical analysis
Data are presented as mean ± SD. Statistical analysis was performed with the SPSS Statistics software (version 16.0, SPSS Inc., Chicago, IL, USA). The differences between groups were analyzed using independent sample t tests, one-way or two-way analysis of variance (between different mice strains). Histograms were produced by GraphPad Prism 6 (GraphPad Software Inc., USA). A P value <0.05 was considered statistically significant. Differences in the heat map of cluster analysis were statistically significant when the fold change between two groups was >1.5.