6-8-weeks-old wild-type C57BL/6J mice were purchased from Jiangsu JicuiYaokang Biotechnology Co., Ltd (Jiangsu, China). The animals were kept in a pathogen-free environment. All experimental procedures were approved by the animal research committee of Fudan University. Both male and female mice are used for bone marrow isolation. Female mice are used for chitin administration experiments.

L-929 cells was obtained from the cell bank of the Chinese Academy of Science (Shanghai, China). Cells were cultured in RPMI 1640 medium (Gibco, USA) supplemented with 10% FBS (VivaCell, China) and 1% penicillin/streptomycin (Invitrogen, USA). Cells were maintained at 37°C in a humidified 5% CO₂ atmosphere.

Bone marrow-derived macrophages (BMDMs) were obtained as previously described (18). Briefly, bone marrow cells were obtained from wild-type C57BL/6 mice aged 6 to 8 weeks and cultured in RPMI 1640 supplemented with 10% FBS, 1% penicillin/streptomycin and 30% supernatants of L-929 conditioned medium. On day 7, attached macrophages were washed and harvested for stimulation.

Peritoneal exudate cells (PECs) were isolated from peritoneum by lavage using 4 mL PBS for 3 times, and were collected for analysis.

Differentiated BMDMs were incubated for 24 hr in culture medium containing IL-4 (PeproTech, USA) at a concentration of 20 ng/mL to induce M2 macrophage polarization. BMDMs were treated by 100 ng/mL LPS and 50 ng/mL IFNγ to induce M1 macrophage polarization. In some experiments, BMDMs were cultured in BCAA free RPMI 1640 medium (USbiological, USA) with or without 0.2 mM, 0.4 mM, 0.8 mM Val, Leu, Ile (Sigma, USA) or their combination.

BMDMs were transfected with 20 μM stock solution of small interfering RNAs (siRNA) for 24 hr by using Lipofectamine™ RNAiMAX Transfection Reagent (Invitrogen, USA) according to manufacturer’s instructions. Serum-free Opti-MEM (Gibco, USA) was used to dilute the siRNA reagents. si-SLC25A44, si-BCAT2, si-BCKDHA and si-NC oligo were purchased from GenePharma (Shanghai, USA). The mouse SLC25A44 siRNA sequence was 5′-CUCUCGGCAAGAAUCAUCUTT-3′, The mouse BCAT2 siRNA sequence was 5′-GGGAGAACCUUGGCUUCUUCU-3′, The mouse BCDKHA siRNA sequence was 5′-GCUΜGAGUUCAUCCAGCCCAA-3′. After 24 hr, the cells were cultured with IL-4 (20 ng/mL) to induce their polarization into M2.

Total RNA was extracted from BMDMs using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s instructions. cDNA was synthesized with PrimeScript RT Master Mix (TaKaRa, Japan). Quantitative RT-PCR was performed using SYBR Green Master Mix (Yeasen, China). The relative mRNA expression level was calculated by the 2^-ΔΔCt method (fold expression), β-actin was used as an internal reference. The sequences of primers are listed in Supplementary Table 1 .

For immunofluorescence, BMDMs were seeded in coverslips. After adhesion, cells were polarized and treated with BCAAs. Cells were then fixed using 4% PFA for 20 min, rinsed three times with PBS, and permeabilized and blocked with PBS containing 1%BSA and 0.3%Triton X-100 for 1 hr at room temperature. The cells were then stained using a rabbit polyclonal anti-RELMα antibody (Peprotech, USA) at 1:200 dilution in a humid chamber at 4°C overnight.

After primary antibody incubation, macrophages were stained with rabbit anti-IgG (Alexa Fluor^® 488, Invitrogen, USA) at 1:200 dilution at room temperature in the dark for 1 hr. The nuclear staining was done using Mounting Medium containing DAPI (Yeasen, China). Images were assessed by fluorescence microscopy with a ×40 objective lens (EVOS™ M7000, Thermo, USA). Quantification of the images was performed using Image J software.

Antibodies for flow cytometry were purchased from Thermo Fisher Scientific. The antibodies used were anti-CD45 (PE), anti-F4/80 (FITC), anti-Siglec-F (PE-Cyanine7), anti-CD11b (APC- Cyanine7), anti-CD206 (APC). Cells were washed in FACS buffer (3% (vol/vol) FBS and 1 mM EDTA in PBS, pH 7.4) once, then incubated with each antibody in 50 μL FACS buffer for 30 min and washed once with 1 mL FACS buffer. Cells were analyzed on the BD LSRII flow cytometer, and data analyzed using Flowjo software.

10 mg chitin (Solarbio, China) was suspended in 3 mL PBS, sonicated with a JY92-IIDN device (SCIENTZ, China) at 150 W, work 7 s and pause 6 s two times for 10 min. After filtration with 100 μm cell strainer, chitin was diluted in 10 mL PBS, 100 μL of chitin was injected once intraperitoneally and PECs were collected 2 d after administration (19, 20).

BCAA (valine, leucine, isoleucine) solution was prepared in PBS (pH 7.4) as isomolar mixtures and sterilized by filtration through a 0.22 μm filter. Mice were intraperitoneally injected with 500 μL of 30 μmol, 60 μmol or 90 μmol BCAA solution (21) after chitin administration.

IL-4 treated BMDMs or non-treated BMDMs were washed once with PBS. Pre-chilled 80% (vol/vol) methanol was added and then incubated at -80°C overnight. The metabolite-containing supernatants were centrifuged for 10 min at 14000 g and 4°C, and supernatants were dried under nitrogen flow (22). The metabolome extracts were reconstituted in acetonitrile and H₂O (1:1, vol/vol), vortex for 1 min, and centrifuged for 15 min at 14000 rpm and 4°C. Targeted LC-MS were performed by QTRAP 6500+ Liquid chromatography-mass spectrometer (AB SCIEX, USA) interfaced with ExionLC™ AD HPLC system (AB SCIEX, USA). Mobile phases A (95% H2O; 5% acetonitrile; 20 mM ammonium acetate; 10 mM ammonia; 50 µL 0.05 M medronic acid) and mobile phases B (100% acetonitrile) are used. Chromatographic separation was performed on a HILIC column (iHILIC-(P) Classic column, 5 μm, 150 × 2.1 mm) with the following gradient: 0-2 min 15% A; 2-7 min 15-40% A; 7-12 min 40-65% A; 12-12.1 min 65-80% A; 12.1-15.9 min 80% A; 15.9-16 min 80-15% A; 16-23 min 15% A. The flow rate was set as 0.2 mL/min, and the injection volume was 5 µL. The column temperature was 30°C and autosampler temperature was 6°C. The missing value was addressed according to the 80% rule (23): a metabolite was considered detectable if it appeared in at least 4 out of 5 samples in a group. In order to remove potential variations in total metabolite abundance across samples, the total intensity of metabolite peaks within each sample was adjusted to the mean total intensity across all samples. Each metabolite peak intensity is then divided by the normalization factor, standardizing the total metabolite abundance within each sample (24). The peak area list with compound names was analyzed by MetaboAnalyst 6.0. Kyoto Encyclopedia of Genes and Genomes (KEGG) was used for the enrichment analysis of the significantly upregulated or downregulated metabolites.

The oxygen consumption rate (OCR) was measured by the Seahorse XFe96 Extracellular Flux Analyzer (Agilent, USA). BMDM cells were plated at 8 × 10⁴ cells/well in a Seahorse 96-well microplate (Agilent, USA) and treated as indicated. The cells were incubated in Seahorse XF DMEM medium (Agilent, USA) for 1 hr at 37°C in measuring chamber before the assay. Oxygen consumption rate was measured under basal conditions and in response to sequential injections of 1 μM oligomycin, 1.2 μM FCCP (Carbonyl cyanide 4-trifluoromethoxy phenylhydrazone), and 1 μM rotenone.

Statistical analysis was performed with GraphPad Prism 9 software. Data are shown as mean ± SEM. Groups were compared by a Student’s two-tailed t test and two way ANOVA (followed by Tukey’s multiple comparison tests). P values of < 0.05 were considered to be statistically significant. The number of experiments and significance levels are presented in the legend of the figures.

To systematically identify the metabolic pathways that are rewired in macrophages upon alternative activation, we performed metabolomic analysis comparing murine bone marrow-derived macrophages (BMDMs) stimulated with IL-4 for 24 hr (M2) to unstimulated BMDMs (M0) ( Figure 1A ). We totally detected 186 metabolites in the samples, and we identified upregulated or downregulated metabolites in M2 relative to M0 cells (>1.5-fold, P < 0.05). Principal component analysis (PCA) and heatmap of the metabolites abundance revealed a significant separation between M0 and M2 macrophages ( Figures 1B, C ). In addition to several pathways including citrate cycle and arginine metabolic pathway, which are known to be altered in M2 macrophages, we discovered that BCAA metabolism was one of the most profoundly impacted pathways ( Figure 1D ). The abundance of Val and Leu/Ile were significantly decreased in M2 macrophages compared to M0 macrophages ( Figure 1E ). Additionally, one of the M0 samples did not cluster with the others in the PCA plot ( Figure 1B ). To determine if this outlier affected BCAA levels, we analyzed the relative intensities of Val and Leu/Ile both with and without this M0 outlier. The results showed a significant decrease in M2 macrophages, regardless of whether the outlier was included ( Supplementary Figure 1 ). These results suggested that BCAA levels were decreased in M2 macrophages.

We next determined whether BCAAs affect M2 macrophage polarization. Differentiated BMDMs were treated with 20 ng/mL of IL-4 for 24 hr with or without BCAAs. The mRNA expression level of the key M2 macrophage genes were determined by RT-qPCR. Treatment with BCAAs increased the mRNA expression levels of Arg1, Retnla, Mrc1, Mgl1 and Mgl2 in M2 macrophages, but did not increase the expression level of Ym1. BCAAs did not change the expression of M2 macrophage-related genes under M0 conditions ( Figures 2A–F ). In addition, BCAAs treatment increased RELMα protein level as assessed by immunofluorescence ( Figures 2G, H ), as well as the frequency of CD206⁺CD11b⁺F4/80⁺ cells ( Figure 2I ). The gating strategy was shown in Supplementary Figure 2 .

To understand the dose-dependent effect of BCAAs on M2 macrophage polarization, we treated BMDMs with increasing doses of BCAAs for 12, 24 and 48 hours during M2 macrophage polarization. The result showed that BCAAs increased Retnla expression in a dose-dependent manner at both 24 and 48 hours and increased Arg1 expression at 24 hours. In contrast, BCAA treatment suppressed Ym1 expression at 24 and 48 hours. At the 12 hour-time point, BCAA treatment had no significant effect on the expression of Arg1, Retnla or Ym1 ( Supplementary Figure 3 ). In addition, to understand whether BCAA supplementation specifically modulate M2 macrophage polarization, we also determined the effect of BCAA supplementation on M1 macrophage polarization. We found that BCAA supplementation didn’t increase the expression level of M1 markers in polarized M1 macrophages, including Il1β, Il6, Nos2, Tnfa ( Supplementary Figure 4 ). Next, to investigate the functional role of BCAAs in alternatively activated macrophages polarization in vivo, we subjected wildtype (WT) mice to an in vivo model of chitin-induced M2 polarization. Chitin, a major structural component of helminths and fungi, was suggested to induce M2 macrophage polarization and subsequent eosinophil recruitment when injected peritoneally (19, 20). We administered mice with chitin along with increasing doses of BCAAs ( Figure 3A ). The results indicate that BCAAs start to promote M2 macrophage polarization in peritoneal macrophages at a dose of 90 µmol, corresponding to 4.5 mM in circulation, given that the average mouse body weight is approximately 20 g. We found that chitin-induced recruitment of eosinophils was significantly increased in BCAAs treated animals ( Figures 3B, C ). BCAA supplementation increased the percentage of CD11b⁺F4/80⁺CD206⁺ cells in PECs ( Figures 3D, E ). In addition, the expression of M2 macrophage-associated genes Arg1, Ym1, Mgl1 and Mgl2 in chitin-elicited PECs were also increased in mice administrated with BCAAs ( Figures 3F–I ). Additionally, we measured Ym1 expression in peritoneal cells at 24 and 48 hours after chitin administration in mice treated with 90 µmol BCAAs. The results showed that Ym1 levels decreased at 24 hours in mice treated with both chitin and BCAAs, but increased at 48 hours compared to mice treated with chitin alone. This result, at least in part, suggests that the varying effects of BCAA supplementation on Ym1 expression may be influenced by the timing of detection ( Supplementary Figure 5 ). Collectively, these data indicated that BCAAs promoted M2 macrophage in vitro and in vivo.

To explore the impact of individual BCAAs on M2 macrophage polarization, we supplemented BMDMs with leu, Ile and Val, either individually or in combination, under M0 or M2 conditions. We found that while Ile and Val exhibited some capacity to enhance M2 macrophage polarization, their effects were not as pronounced as those of Leu alone or the combination of all three BCAAs. The results suggested that Leu exerts the most pronounced effect on alternative macrophage polarization ( Figure 4 ).

The metabolites of BCAA catabolism, such as acetyl-CoA and succinyl-CoA, enter the TCA cycle and fuel oxidative phosphorylation (OXPHOS), which is established to be critical for M2 polarization (25). Notably, the relative level of acetyl-CoA was upregulated in M2 macrophages, indicating enhanced OXPHOS in M2 macrophages ( Supplementary Figure 6 ). We speculated BCAA catabolism serves to fuel the OXPHOS and supports M2 polarization. Thus, we tested the oxygen consumption rate (OCR) by Seahorse XFe96 Analyzer. OCR of M2 BMDMs increased strongly, and BCAAs enhanced the OCR of M2 BMDMs ( Figures 5A–D ), basal respiration, maximal respiration and ATP production all upregulated. Hence, BCAA supplementation promoted oxidative phosphorylation in M2 macrophages.

SLC25A44 transports BCAAs into mitochondria (10). BCAAs are transaminated by BCATs to form BCKAs, which is reversible. BCAT2 plays the role in mitochondria, while BCAT1 works in cytoplasm. BCKDH catalyzes the irreversible oxidation of BCKAs. The metabolite of BCAA catabolism enter TCA cycle as acetyl-CoA and succinyl-CoA. As BCAAs were reduced in M2 BMDMs ( Figure 1 ) and supplementation of BCAAs promoted M2 polarization ( Figure 2 ), we further examined the expression level of key genes in BCAA catabolism, including Bcat1, Bcat2, Bckdha and Slc25a44 at 0, 4, 8, and 24 hours after IL-4 treatment. The result showed that Bcat1 was nearly undetectable in BMDMs, while Bcat2, Bckdha and Slc25a44 were all expressed ( Supplementary Figure 7A ). The mRNA levels of Bcat1 and Slc25a44 increased after 4 hours of IL-4 stimulation and then decreased over time. In contrast, Bcat2 mRNA levels decreased following IL-4 treatment, while Bckdha levels remained unchanged until 24 hours, when they also decreased ( Supplementary Figures 7B–E ). We have assessed the protein level of BCKDHA in M0 and IL-4 induced M2 macrophages, and the protein level of BCKDHA was upregulated from 2 to 8 hours after IL-4 treatment and returned to baseline levels 24 hours post IL-4 treatment ( Supplementary Figure 8 ). These results suggest that the expression level of key enzymes and mitochondrial transporter involved in BCAA catabolism were unchanged or moderately downregulated.

To explore the effect of BCAA transport into mitochondria on M2 macrophage polarization, we knocked down Slc25a44 with siRNA in BMDMs, and then polarized BMDMs toward M2 condition with 20 ng/mL IL-4 for 24 hr post transfection. With approximately 60% RNA silence efficiency, Slc25a44 inhibition reduced the mRNA expression level of Arg1, Retnla, Mrc1, Ym1, Mgl1 and Mgl2 ( Figures 6A–G ). Additionally, knockdown of Slc25a44 decreased the protein level of RELMα in M2 macrophages, as assessed by immunostaining ( Figures 6H, I ). The results suggested that inhibiting BCAA transport into mitochondria via SLC25A44 impairs BMDM M2 polarization.

In mitochondria, BCAAs were transaminated by BCAT2 to generate BCKAs. To investigate the effect of BCAA transamination on M2 macrophage polarization, we knocked down Bcat2 with siRNA and subsequently induced M2 polarization in BMDMs. We found that siRNA-mediated inhibition of BCAT2 resulted in a significant decrease in the expression levels of M2 marker genes, including Retnla, Ym1, Mgl1 and Mgl2, but there were no obvious decrease in Arg1 and Mrc1, ( Figures 7A–G ). Meanwhile, the protein levels of RELMα were reduced in M2 macrophages following Bcat2 knockdown ( Figures 7H, I ). These results indicated that restraining BCAA transamination via BCAT2 reduced M2 macrophage polarization.

BCKAs undergo oxidative decarboxylation mediated by the BCKDH complex, comprising catalytic subunits E1α, E1β, E2 and E3 encoded by the BCKDHA, BCKDHB, BCKDE2 and BCKDE3 genes, respectively. We then explored the role of BCKA decarboxylation in M2 macrophage polarization. Using siRNA, we knocked down Bckdha in BMDMs and induced M2 macrophage polarization. The results revealed that Bckdha knockdown reduced the expression levels of M2 marker genes, including Arg1, Retnla, Ym1, Mgl1 and Mgl2, while the expression level of Mrc1 remained unchanged ( Figures 8A–G ). Additionally, the protein level of RELMα was decreased in M2 macrophages following Bckdha knockdown ( Figures 8H, I ). These results suggested that inhibiting BCKDH-mediated BCKA decarboxylation reduces BMDM M2 polarization.