Research ArticleIMMUNOLOGY

Natural polymorphism of Ym1 regulates pneumonitis through alternative activation of macrophages

See allHide authors and affiliations

Science Advances  21 Oct 2020:
Vol. 6, no. 43, eaba9337
DOI: 10.1126/sciadv.aba9337


We have positionally cloned the Ym1 gene, with a duplication and a promoter polymorphism, as a major regulator of inflammation. Mice with the RIIIS/J haplotype, with the absence of Ym1 expression, showed reduced susceptibility to mannan-enhanced collagen antibody–induced arthritis and to chronic arthritis induced by intranasal exposure of mannan. Depletion of lung macrophages alleviated arthritis, whereas intranasal supplement of Ym1 protein to Ym1-deficient mice reversed the disease, suggesting a key role of Ym1 for inflammatory activity by lung macrophages. Ym1-deficient mice with pneumonitis had less eosinophil infiltration, reduced production of type II cytokines and IgG1, and skewing of macrophages toward alternative activation due to enhanced STAT6 activation. Proteomics analysis connected Ym1 polymorphism with changed lipid metabolism. Induced PPAR-γ and lipid metabolism in Ym1-deficient macrophages contributed to cellular polarization. In conclusion, the natural polymorphism of Ym1 regulates alternative activation of macrophages associated with pulmonary inflammation.


Allergic asthma and other complex chronic pulmonary diseases, such as chronic obstructive pulmonary disease and chronic bronchitis, are characterized by airway obstruction and the pathological structure of inflammatory alterations in the lungs. Although many loci have been found to be associated with asthma, few results have been replicated, because of the complex and context-dependent interactions of the genetic variants with environmental factors, as well as with other genes (1). Therefore, it has been difficult to conclusively identify critical genetic and pathogenic factors in pulmonary inflammation. Candidate genes thought to be important based on human studies have been genetically modified in mouse models to understand disease mechanisms. However, most of the genes in question seem to be linked more to regulation of basic physiology than to pathogenesis per se, and humanized genes may not be physiologically regulated in the mouse model as they are in humans. An alternative approach is to use inbred experimental animals, with diseases and biological pathways mimicking humans, for the genetic analysis. The major associated loci can be identified by using genetic segregating crosses of inbred strains and then phenotyping the offspring for quantitative traits followed by genetic mapping (2). The first locus to be positionally identified was an amino acid replacement single-nucleotide variant in the Ncf1 gene (3). A single-nucleotide polymorphism (SNP) in the human Ncf1 gene and a spontaneous mutation of the Ncf1 gene in mice were subsequently shown to be significantly correlated with systemic lupus erythematosus; Ncfl variants are also likely to play a critical role in many other inflammatory diseases such as rheumatoid arthritis, multiple sclerosis, and psoriasis (49). Further studies of this newly identified pathway offer a fruitful direction for future research on human immune diseases.

Here, we have addressed another major locus identified from genetic segregating intercrosses between B10.RIII and RIIIS/J mice, using collagen-induced arthritis (CIA) and experimental autoimmune encephalomyelitis (EAE) as inflammatory diseases traits. Using these crosses, the loci Eae3/Cia5 and Eae2 located on murine chromosomes 3 and 15 were identified (10, 11). Furthermore, in a partial advanced intercross between Eae3/Cia5 and Eae2 bicongenic mice, the original loci were fine-mapped, and a 2-Mb RIIIS/J-derived congenic segment on chromosome 3 located within the Cia22 locus, including a cluster of chitinase-like genes, was identified (12). Subsequent microarray analysis of the selected genes in the chitinase gene cluster indicated Ym1 as the major differentially expressed candidate gene (13).

Ym1 [chitinase-like protein 3 (Chil3)] is a member of chitinase-like proteins (CLPs) but was originally identified as an eosinophil chemotactic factor, which is strongly induced in mice infected with tissue and gastrointestinal nematode parasites (14). It is also considered a signature marker for alternatively activated macrophages (AAMs) (15), although a potential functional role of Ym1 on macrophage polarization has not yet been addressed. A recent study reported that Ym1/2 promoted interleukin-17 (IL-17)–mediated neutrophilia and was involved in nematode killing and host damage in mice (16). It was also suggested that Ym1 induced oligodendrogenesis, and silencing Ym1 increased the severity of EAE (17). Furthermore, Ym1/2 promoted TH2 (T helper 2) cytokine expression implicated to regulate allergic inflammation by inhibiting 12/15(S)-lipoxygenase in mice (18). Ym1+Ly6Chi monocytes have been shown to be associated with the resolution of inflammation and tissue healing (19). It seems that Ym1, which lacks chitinase activity, has important but diverse functions; however, a conclusive role of Ym1 in complex diseases such as pulmonary inflammation has so far been lacking.

In this study, we have positionally identified Ym1 polymorphism and developed a congenic mouse strain with a RIIIS/J-derived haplotype that leads to reduced Ym1 expression; these findings provide an opportunity to further investigate the regulation of Ym1 and its role in the immune response. We demonstrate that the lack of expression due to mutations in the promotor of the Ym1 gene alleviates pulmonary inflammation and promotes the alternative activation of macrophages. Thus, the Ym1 gene contains a positionally identified genetic polymorphism, offering new insights into the immune system and the pathogenesis of inflammatory disorders.


Promoter polymorphism of Ym1 controls its gene expression

To address the potential of Ym1 as a candidate gene for a major locus regulating inflammatory disease, Ym1 congenic deficient mouse strain (BR.Ym1Δ mice) carrying a 2-Mb RIIIS/J fragment was generated by introgressing the congenic fragment into the B10.RIII background (Fig. 1A). The tissue distribution of Ym1 and its highly homologous gene Ym2 was studied in naïve B10.RIII mice. We found that Ym1 mRNA was strongly expressed in lung, spleen, and bone marrow but only weakly expressed in other tissues, including stomach and lymph nodes, while Ym2 mRNA was highly expressed in stomach but undetectable in the other tissues analyzed (Fig. 1B). We found that the expression of Ym1 mRNA in both lung and spleen was remarkably low in RIIIS/J-derived congenic mice compared with B10.RIII, Balb/c, and B6/NJ, as analyzed with real-time quantitative polymerase chain reaction (RT-qPCR) (Fig. 1C). In addition, we validated the serum levels of Ym1 protein of different inbred mouse strains with an enzyme-linked immunosorbent assay (ELISA), and in line with our qPCR results, we could detect Ym1 in the serum of B10.RIII mice but not in the BR.Ym1Δ littermate congenics (Fig. 1D).

Fig. 1 Polymorphisms of Ym1 lead to the variance of gene expression.

(A) Genetic map of mouse chromosome 3 and congenic fragment. Cia5, Cia21, and Cia22 loci were identified by linkage analysis and partial advanced intercross. The smallest congenic fragment covering the Ym1 gene was minimized as shown. Ym1 expression in different mouse strains was assessed (n = 3 to 5 for each strain). (B) Gene expression of Ym1 and Ym2 in wild-type B10.RIII mice was assessed by PCR with β-actin as a housekeeping gene. (C) Ym1 expression in lung and spleen of indicated strains was detected by RT-qPCR. (D) Circulating Ym1 levels in different strains were detected by ELISA. (E) Natural polymorphisms for indicated inbred and wild-derived mouse strains. Data of Ym1 expression in different strains were from the ImmGen database. N.D., no data. (F) Promoters from B10.RIII and RIIIS/J mice were cloned into pGL4.17 plasmids (named as pGL-B10.RIII and pGL-RIIIS/J). Site mutation was done using pGL-B10.RIII to convert SNP1 to SNP4 sites to RIIIS/J genotypes (named as pGL-SNP1, pGL-SNP2, pGL-SNP3, and pGL-SNP4). (G) Constructed reporter gene plasmids were transfected into HEK293T cells, and relative luciferase activities were detected. All values are expressed as means ± SEM. **P < 0.01 and ***P < 0.001.

Ym1 expression differs considerably between different inbred and wild-derived mouse strains, which might be due to natural polymorphism in their respective promotor regions. Therefore, we decided to take a closer look at their genetic differences using the online database from Ensembl together with our own sequencing data. A copy number variance was found—a 187-kbp (kilo–base pair) fragment covering the Ym1 gene (from 106,123,423 bp to 106,311,121 bp in GRCm38/mm10). C57Black strains, LP/J-, NOD-, and 129-derived strains have this duplication, but other strains including Balb/c and RIIIS/J strains do not (Fig. 1E). In general, mouse strains harboring the duplication show increased Ym1 expression (data from ImmGen), suggesting that Ym1 expression is strongly associated with this copy number variant. However, Balb/c mice without the duplication still showed higher Ym1 expression than RIIIS/J mice (Fig. 1D), indicating that there are additional polymorphisms affecting Ym1 expression. The additional effect could be explained by differences in the Ym1 promoter sequences in the different mouse strains. SNPs were found in SPRET, PWK, NZO, and RIIIS/J strains, and the RIIIS/J was found to have a distinct haplotype in the promoter region of Ym1 gene, with four SNPs compared to C57Black strains (Fig. 1E). Subsequently, we cloned Ym1 promoters of both B10.RIII and RIIIS/J mice into luciferase reporter plasmids (Fig. 1F; designated as pGL-B10.RIII and pGL-RIIIS/J), mimicking the four SNPs, respectively, by site mutation (Fig. 1F; pGL-SNP1, pGL-SNP2, pGL-SNP3, and pGL-SNP4). We carried out a dual-luciferase reporter assay to test the effect of each SNP on the regulation of Ym1 expression, and the results showed that the promoter from RIIIS/J mice had a lower transcription activity than the one from wild-type B10.RIII mice (Fig. 1G). The results also showed that it was SNP2, SNP3, and SNP4, but not SNP1, that reduced the transcription activity compared with the wild-type promoter sequence. SNP2 had the largest effect, indicating that it might be the key site controlling the Ym1 expression (Fig. 1G). Together, the natural polymorphism controls Ym1 expression.

In addition, the immune cell populations in spleen and peripheral blood from naïve Ym1 congenic deficient mice and its wild-type control B10.RIII mice, which differed in Ym1 expression, were analyzed by flow cytometry. Except the slightly decreased proportion of eosinophils in BR.Ym1Δ congenic mice, there were almost no difference between the two strains (fig. S1, A and B).

Ym1-deficient mice are protected from mannan-induced arthritis, dependent on macrophages

As Ym1 was originally identified in segregating partial advanced intercross associated with arthritis, we next determined the effect of Ym1 deficiency on arthritis pathology to confirm its disease phenotype. Collagen antibody–induced arthritis (CAIA) model, which is developed in the absence of an adaptive immune system (αβT and B cells), is induced after intravenous injection of anti-CII antibodies, leading to mild arthritis (20). Usually, an intraperitoneal injection of lipopolysaccharide (LPS) or mannan is given a few days after the initial transfer of the CII-specific antibodies to enhance the arthritis susceptibility (21, 22). We found that the Ym1 alleles had no impact on LPS-enhanced CAIA susceptibility or severity, whereas deficiency of Ym1 led to a reduction in severity of mannan-enhanced CAIA (Fig. 2A). Mannan-CAIA is a different disease than LPS-CAIA, best demonstrated by the contrasting effect in Ncf1-mutated mice, which have a lower capacity to make an oxidative burst by the NADPH (reduced form of nicotinamide adenine dinucleotide phosphate) oxidase (Nox2) complex (22). LPS-CAIA is ameliorated, whereas mannan-CAIA is markedly enhanced in Ncf1-mutated mice.

Fig. 2 Ym1 deficiency in congenic mice protects arthritis development induced by mannan through macrophages.

(A) B10.RIII and BR.Ym1Δ congenic mice (n = 11 to 18 per group) were injected intravenously with four monoclonal antibodies at day 0 and boosted with LPS (25 μg) or mannan (2 mg) at day 5 to induce CAIA. The arthritis severity was scored. BR.Ncf1* and BR.Ncf1*.Ym1Δ mice (n = 6 per group) were administrated with 20 mg of mannan intranasally (i.n.) to induce psoriasis and chronic arthritis. (B) Arthritic joint phenotype and psoriasis-like skin lesions in the hind paws. Photo credit: Wenhua Zhu, Xi’an Jiaotong University Health Science Center. (C) The arthritis and psoriasis severity was scored. The area under the arthritis scoring curve until D28 between the two strains was analyzed statistically (P < 0.05). (D) Clodronate-liposome (CL-lipo) or PBS-liposome (PBS-lipo) was treated intranasally to BR.Ncf1* mice (n = 6 per group) to deplete macrophages. Two days later, mice were administrated with mannan intranasally to induce arthritis. (E) BR.Ncf1* and BR.Ncf1*.Ym1Δ mice (n = 7 per group) were administrated with mannan intranasally with or without Ym1 protein supplement to induce arthritis. Disease severity was scored using a macroscopic scoring system. All values are expressed as means ± SEM. *P < 0.05.

Ym1 is highly expressed in lung macrophages and is likely to have an important role in regulating the immune response in the lung. We crossed our Ym1 congenic mice with Ncf1-mutated mice on B10.RIII background (BR.Ncf1*.Ym1Δ mice) and exposed the mice by treating them intranasally with mannan directly. It is well known that sterile aspergillus induces severe inflammation in the lung (23), and we chose to treat them intranasally with mannan based on our earlier observation that a single intraperitoneal injection of mannan induces an acute form of psoriasis and psoriatic arthritis (PsA) in Ncf1-mutated mouse strains (24). Intranasal exposure of mannan in Ncf1-deficient B10.RIII mice induced psoriasis and chronic development of arthritis (Fig. 2, B and C). Ym1-deficient littermate mice were completely protected against the disease, demonstrating that the presence of Ym1 is critical for the development of chronic arthritis induced by mannan (i.e., PsA) in Ncf1-deficient mice (Fig. 2, B and C). To further confirm the importance of macrophages, we used clodronate-liposome to deplete the lung-resident macrophages of BR.Ncf1* mice (fig. S2) and found that the severity of mannan-induced PsA (MIP) was alleviated significantly (Fig. 2D). In addition, to directly prove the role of Ym1 in disease pathogenesis, we tried to reestablish arthritis severity through recombinant Ym1 protein administration. Mannan was injected intranasally in BR.Ncf1*.Ym1Δ mice, which were subsequently treated with or without Ym1 protein. As expected, BR.Ncf1*.Ym1Δ mice supplemented with Ym1 protein exhibited enhanced disease severity compared to mock [phosphate-buffered saline (PBS)]–treated mice of the same strain, thereby displaying comparable severity levels with BR.Ncf1* mice (Fig. 2E). Because both CAIA and mannan-induced arthritis are independent on adaptive immunity, our results suggested that Ym1 play a role in regulating innate immune response with the involvement of macrophages.

Ym1-deficient mice with mannan intranasal treatment show less pneumonitis

To further delineate the role of Ym1 for regulation of macrophage functions, we treated mice with mannan intranasally, which primarily induces pneumonitis (Fig. 3A) and is followed by a late secondary immune response leading to arthritis. First, lung tissues from both BR.Ncf1* mice and BR.Ncf1*.Ym1Δ mice at days 0, 2, and 5 after mannan treatment were collected, and lung index (the lung weight normalized by the body weight) was calculated. BR.Ncf1* mice showed higher lung index than congenic mice (Fig. 3B). Histopathological changes of the lungs were observed by evaluating the inflammation around the airways and in interstitia. Results showed that the inflammatory cell infiltration was increased significantly at day 5 in BR.Ncf1* mice but not in BR.Ncf1*.Ym1Δ mice (Fig. 3C). BR.Ncf1*.Ym1Δ mice had mild inflammation around the airways compared with BR.Ncf1* mice (Fig. 3C).

Fig. 3 Ym1 deficiency reduces pulmonary inflammation in mannan intranasally induced arthritis mice.

(A) BR.Ncf1* and BR.Ncf1*.Ym1Δ mice (n = 5 to 8 per group) were administrated with 0.5 mg of mannan intranasally (i.n.) to induce pulmonary inflammation. Lung tissues and BALF were collected at days 0, 2, and 5. (B) Lung weight normalized by body weight was calculated. (C) Lung sections were stained by hematoxylin and eosin (H&E), and peribronchial infiltration and interstitial infiltration of inflammatory cells were evaluated. Then, total cells, eosinophils, interstitial macrophages, and alveolar macrophages from mannan intranasally treated mice (D to G) or from mannan intranasally treated mice with Ym1 protein (1 μg per mouse) supplement (H to J) were analyzed by flow cytometry. (K) mRNA expression of Il4 and Tgfb1 in lung tissues was detected by RT-qPCR. In addition, Ym1 concentration (L) in BALF at days 0, 2, and 5 was determined by ELISA. Besides, Ym1 expression in total BALF cells (M) and different types of immune cells (N) in mannan-treated mice at day 2 was analyzed by flow cytometry. The representative histograms of Ym1 in alveolar macrophages and interstitial macrophages (N) were shown. All values are expressed as means ± SEM. # indicates the comparison among different time points in each mouse strain, and * indicates the comparison between the two mouse strains. */#P < 0.05, **/##P < 0.01, and ***/###P < 0.001.

Then, the immune cells in the bronchoalveolar lavage fluid (BALF) were analyzed by flow cytometry. The number of total BALF cells, eosinophils, and infiltrated interstitial macrophages was increased after mannan treatment in BR.Ncf1* mice and showed significant difference between the two strains at day 2 (Fig. 3, D to F). The major constituent cell type of BALF cells is alveolar macrophages (25), where their proportion was significantly decreased in BR.Ncf1* mice because of immune cell infiltration, but was only slightly changed in BR.Ncf1*.Ym1Δ mice (Fig. 3G). Meanwhile, the cell population in Ym1 protein supplement experiment was also observed by flow cytometry, and we found an increase in total BALF cells and of eosinophils (Fig. 3, H and I). Thereby, the increased proportion of alveolar macrophages in BR.Ncf1*.Ym1Δ mice were reversed when the mice were treated with Ym1 protein (Fig. 3J).

Cytokine expression in the lung tissues of mannan-treated mice were also determined by RT-qPCR. Results showed that both Il4 and Tgfb1 mRNA expression were increased at day 2 in BR.Ncf1* mice (Fig. 3K). In contrast, BR.Ncf1*.Ym1Δ mice showed reduced Il4 expression, but not Tgfb1 expression, suggesting a specific regulatory effect by Ym1 on IL-4 (Fig. 3K).

In addition, the Ym1 protein level in BALF was determined by ELISA, and the result showed that the Ym1 concentration in the BALF of BR.Ncf1* mice was increased with a higher degree of inflammation, whereas no change was observed in BR.Ncf1*.Ym1Δ mice (Fig. 3L). Besides, the Ym1 expression in immune cells of BALF at day 2 was also detected by flow cytometry. BALF cells from BR.Ncf1* mice showed higher Ym1 expression than from BR.Ncf1*.Ym1Δ mice (Fig. 3M). Ym1 was mainly expressed by lung-resident macrophages, and alveolar macrophages from BR.Ncf1* mice showed higher Ym1 expression than those from BR.Ncf1*.Ym1Δ mice (Fig. 3N).

Low expression of Ym1 also alleviates antigen-induced pneumonitis

The above results showed that Ym1 deficiency weakened the innate immune response in lung triggered by mannan. Innate immune regulation contributes to the activation and maintenance of adaptive immunity. To further confirm the role of Ym1, the classical antigen-induced pulmonary inflammation (AIPI) model using ovalbumin (OVA) was induced in B10.RIII and BR.Ym1Δ mice (Fig. 4A). In the OVA group of both strains, only the wild-type B10.RIII mice showed increased interstitial infiltration of inflammatory cells, although both Ym1-deficient and wild-type mice showed increased inflammation around the airways (Fig. 4B). Besides, Ym1-deficient mice in the OVA group had decreased mucus hypersecretion compared with wild-type mice (Fig. 4C).

Fig. 4 Low expression of Ym1 alleviates the severity of AIPI.

(A) B10.RIII and BR.Ym1Δ mice were sensitized and challenged by OVA to induce AIPI (n = 10 to 13 per group). (B) Lung tissues were stained with H&E to observe pathological changes, and peribronchial infiltration and interstitial infiltration of inflammatory cells were evaluated. (C) Lung sections were stained with periodic acid–Schiff (PAS) to detect mucus production, and the integral optical density (IOD) of staining was calculated. (D) Immune cells in BALF were analyzed by flow cytometry. Serum concentrations of Ym1 (E), OVA-specific IgG1 (F), and OVA-specific IgE (G) were determined by ELISA. (H) mRNA expression of Ym1 and type II cytokines including Il4, Il5, Il10, Il13, and Tgfb1 in lung tissues was detected by RT-qPCR. All values are expressed as means ± SEM. # indicates the comparison between the OVA group and the PBS group in each mouse strain, and * indicates the comparison between the two mouse strains. */#P < 0.05, **/##P < 0.01, and ***/###P < 0.001.

Immunocytes in the BALF were analyzed by flow cytometry, and results showed that Ym1-deficient mice in the OVA group had less cell numbers, especially eosinophils and B cells, as compared with wild-type mice (Fig. 4D). Ym1-deficient mice in the OVA group had decreased levels of Ym1 protein and OVA-specific immunoglobulin G1 (IgG1) compared to wild-type mice, whereas no change in OVA-specific IgE level were seen (Fig. 4, E to G). The AIPI model is a classical type II immune disease (26), so we determined the expression of type II cytokines in the lung tissue by RT-qPCR. Wild-type mice in the OVA group showed increased Ym1, Il4, Il5, Il10, Il13, and Tgfb1 mRNA expression (Fig. 4H), whereas the Ym1-deficient mice had decreased levels of Ym1, Il4, and Il5 expression (Fig. 4H). Meanwhile, Il10 and Il13 mRNA expression in OVA-injected Ym1-deficient mice also showed decreased trend compared with wild-type mice, but Ym1 seemed to have no effect on Tgfb1 expression regulation (Fig. 4H). Combined with the above results, it is clear that Ym1 deficiency alleviated pneumonitis induced by OVA associated with a reduced IL-4 response.

Ym1 counteracts alternative activation of alveolar macrophages in pneumonitis

Ym1 is commonly used as a signature marker for alternative activation of macrophages in mice, and deficiency in Ym1 may therefore influence the macrophage polarization state during disease development. To analyze the alternative activation of alveolar macrophages, the polarization markers were determined in both mannan- and OVA-induced pulmonary inflammation. In lungs with an inflammatory response due to mannan intranasal treatment, the proportion of Arg1+MRC1+ cells in alveolar macrophages of BR.Ncf1* mice were decreased during disease but remained relatively stable in BR.Ncf1*.Ym1Δ mice (Fig. 5A). Alveolar macrophages from congenic mice showed higher Arg1 and MRC1 expression levels than those from BR.Ncf1* mice (Fig. 5A and fig. S3A). In both strains, there was no difference in inducible nitric oxide synthase (iNOS) expression (Fig. 5A). Particularly, the increased proportion of Arg1+MRC1+ cells and increased expression of Arg1 and MRC1 in alveolar macrophages from Ym1-deficient mice were reversed, when mice were treated with Ym1 protein (Fig. 5B and fig. S3B). In addition, in the AIPI model, the MRC1 expression in alveolar macrophages from Ym1-deficient mice was also higher than that from wild-type mice (Fig. 5C and fig. S3C). We conclude that Ym1 suppresses alternative activation of macrophages during the development of pneumonitis.

Fig. 5 Ym1 negatively regulates AAMs.

The polarization markers of alveolar macrophages (AM) in mannan intranasally (i.n.) treated mice (A) or mice with rYm1 protein supplement (B) and in mice of the AIPI model (C) were detected by flow cytometry. Bone marrow–derived macrophages (BMMs; M0) from B10.RIII, BR.Ym1Δ, BR.Ncf1*, and BR.Ncf1*.Ym1Δ mice (n = 3 per group) were treated with LPS and IFN-γ or IL-4 and IL-13 to polarize cells to CAMs or AAMs. (D) Ym1 mRNA expression was detected by RT-qPCR. (E) Arg1 and MRC1 expression in AAMs was detected by Western blotting, and representative blot was shown. In addition, the mRNA expression (F) of AAM markers including Ym1, Arg1, MRC1, and Fizz1 at different time points (0, 3, 6, 12, 24, and 48 hours) after IL-4 and IL-13 stimulation was detected by RT-qPCR. Protein expression (G) of Ym1, Arg1, and MRC1 and the activation of STAT6 (H) were detected by Western blotting, and the statistics are shown in fig. S6. All values are expressed as means ± SEM. In (A) to (C), # indicates the comparison in the same mouse strain, and * indicates the comparison between the two mouse strains. */#P < 0.05, **P < 0.01, and ***/###P < 0.001.

Ym1-regulating STAT6 limits the alternative activation of macrophages during polarization

To determine the pathway by which Ym1 down-regulated alternative activation, we investigated macrophage polarization in vitro. Bone marrow cells were collected from B10.RIII, BR.Ym1Δ, BR.Ncf1*, and BR.Ncf1*.Ym1Δ mice and differentiated to macrophages followed by polarization to classically activated macrophages (CAMs) using LPS and interferon-γ (IFN-γ) or AAMs using IL-4 and IL-13. Expression of polarization markers was compared between Ym1-deficient mice and their littermate wild-type controls. The Ym1 levels were high in AAMs from wild-type B10.RIII mice and BR.Ncf1* mice and low in the Ym1-deficient congenic mice (Fig. 5D). Particularly, Ym1 expression could only be induced by IL-4/IL-13, but not IL-10, suggesting a relatively unique effect of Ym1 on IL-4–induced AAMs (fig. S4A).

Subsequently, the AAM markers Arg1 and MRC1 were detected by Western blotting. As expected, AAMs derived from Ym1-deficient mice showed higher Arg1 and MRC1 expression than that from their wild-type control mice (Fig. 5E). The results underscore the importance of Ym1 for the AAM phenotype in vitro. In addition, IL-4/IL-13 could induce the increased Tgfb1 mRNA expression in macrophages derived from Ym1-deficient mice because these mice had lower basal expression of Tgfb1. However, there was no difference in Tgfb1 expression between the two strains, again suggesting the limited effect of Ym1 on transforming growth factor (TGF) regulation (fig. S4B). On the other hand, we assessed IL-6, IL-12, and IL-10 levels by ELISA in CAMs. IL-6 levels did not differ in any of the four strains, while IL-12 and IL-10 levels also showed no difference in Ncf1 wild-type strains (fig. S4, C to E). However, compared with BR.Ncf1* mice, CAMs from BR.Ncf1*.Ym1Δ congenic mice showed low IL-12 concentration but high IL-10 concentration (fig. S4, D and E), which indicated that Ym1 with a basal expression level might interact with Ncf1 to regulate CAMs, although the effect seemed to be very limited.

The role of Ym1 in AAM polarization remains poorly understood. Therefore, we next observed the changes of AAM markers comprehensively during proliferation. The mRNA expression of Ym1, Arg1, MRC1, and Fizz1 was determined by RT-qPCR at 0, 3, 6, 12, 24, and 48 hours after IL-4 and IL-13 stimulation. Ym1 expression was increased gradually during polarization in macrophages from BR.Ncf1* mice but maintained at a low level in macrophages from BR.Ncf1*.Ym1Δ mice (Fig. 5F). Both Arg1 and MRC1 expression in BR.Ncf1* macrophages were quickly induced at 3 and 6 hours but started to decrease later (Fig. 5F). Macrophages with Ym1 deficiency showed different expression patterns, in which Arg1 expression continued to increase and reached the peak at 24 hours, and MRC1 expression was also higher than that from BR.Ncf1* mice (Fig. 5F). Fizz1 expression in macrophages from both mouse strains was increased gradually, suggesting a regulatory pattern independent from Ym1 (Fig. 5F). We analyzed the protein levels of Ym1, Arg1, and MRC1 by Western blotting, and the expression of Ym1 was found to be consistent with that of mRNA (Fig. 5G). Arg1 and MRC1 protein expression in both strains was gradually induced, however, and macrophages from congenic mice showed higher expression (Fig. 5G and fig. S5A). STAT6 (signal transducer and activator of transcription 6), the most critical transcriptional factor induced by IL-4, controls the process of alternative activation of macrophages (27). Here, the activation of STAT6 signaling pathway was detected by Western blotting. The result showed that macrophages from congenic mice had higher phosphorylation level of STAT6 during the polarization process (Fig. 5H and fig. S5B). Therefore, Ym1 down-regulating STAT6 activation limits alternative activation of macrophages.

Ym1 regulates lipid metabolism and PPAR-γ to orchestrate polarization in macrophages

To comprehensively study the role of Ym1 in macrophages, protein profiles of primary macrophages from B10.RIII mice and BR.Ym1Δ congenic mice were determined by label-free mass spectrum analysis to find out the differentially expressed proteins and relevant pathways. Ninety-four proteins showed differential expression (fold change > 1.2 and P < 0.05) between the two groups in 4053 identified proteins totally. Compared with B10.RIII mice, 49 proteins were up-regulated and 45 proteins were down-regulated in congenic mice (Fig. 6A). Differentially expressed proteins were then hierarchically clustered and analyzed by the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment. Lipid metabolism–associated pathways and peroxisome proliferator–activated receptor (PPAR) signaling pathway were highlighted by the analysis (fig. S6, A to C). Relevant differentially expressed proteins included carnitine palmitoyltransferase 1 (CPT1), retinol dehydrogenase 11 (Rdh11), short-chain dehydrogenase/reductase 3 (Dhrs3), apolipoprotein A-I (ApoA1), apolipoprotein C-I (ApoC1), medium-chain–specific acyl–CoA (coenzyme A) dehydrogenase (Acadm), lysophosphatidylcholine acyltransferase 1 (Lpcat1), and phosphatidylserine synthase 1 (Ptdss1) (Fig. 6B).

Fig. 6 Ym1 regulates lipid metabolism and PPAR-γ to orchestrate macrophage polarization.

Protein profiles of thioglycollate-elicited macrophages from B10.RIII and BR.Ym1Δ mice were determined by label-free mass spectrum analysis (n = 4). (A) Volcano plot comparison showed the differentially expressed proteins in macrophages. (B) Expression of differentially expressed proteins relevant to lipid metabolism. Then, BMMs (M0) derived from BR.Ncf1* and BR.Ncf1*.Ym1Δ mice were polarized to AAMs, and the expression (C) of PPAR-α, PPAR-γ, and PPAR-δ was detected by Western blotting. Antagonists of PPAR-γ (T0070907) and CPT1 (Etomoxir) were used during AAM polarization, and expression (D) of MRC1 and Arg1 was detected by Western blotting. (E) In addition, BMMs were transfected with si-Stat6 (pool of si-Stat6-1, si-Stat6-2, and si-Stat6-3) and si-Pparg-3 siRNA to knock down STAT6 and PPAR-γ or negative control siRNA (si-NC). After AAM polarization, the expression of STAT6, PPAR-γ, MRC1, and Arg1 was detected by Western blotting, and the statistics are shown in fig. S8. (F) Illustration of the role of Ym1 on regulating alveolar macrophages in pulmonary inflammation. All values are expressed as means ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001.

Lipid metabolism pathways in which Ym1 might be involved have earlier been suggested to be involved in macrophage polarization (28). We hypothesized that the regulation of PPAR signaling pathways is involved in the alternative activation of macrophages from BR.Ncf1* and BR.Ncf1*.Ym1 mice. Using Western blotting, we could show that IL-4/IL-13 could induce the expression of PPAR-γ and PPAR-δ, and AAMs from congenic mice showed higher PPAR-γ expression (Fig. 6C and fig. S8A). In addition, AAMs from congenic mice had slightly increased PPAR-α expression, which seemed not strongly associated with polarization (Fig. 6C). Then, inhibitors of PPAR-γ (T0070907) and CPT1 (Etomoxir) were used during macrophage polarization. Results showed that both inhibitors, reducing MRC1 and Arg1 expression, could block the process of alternative activation of macrophages (Fig. 6D and fig. S8B). These results indicated that Ym1 regulates PPAR-γ and lipid metabolism in macrophages, leading to a phenotype best described as alternative activation. To further confirm the above finding, STAT6 and PPAR-γ were knocked down by small interfering RNA (siRNA) (fig. S7). RNA interference (RNAi) of both STAT6 and PPAR-γ could decrease MRC1 and Arg1 expression in AAMs (Fig. 6E and fig. S8C). Particularly, knocking down of STAT6 remarkably reduced the expression of AAM markers and PPAR-γ, suggesting that STAT6 should be located upstream of the polarization pathway. In summary, we conclude that Ym1 plays a novel role as a negative regulator of the alternative activation of macrophages contributing to inflammation, as illustrated in Fig. 6F.


To understand the pathogenic molecular mechanisms underlying inflammatory diseases, it is essential to identify the underlying causative natural polymorphisms. In this study, we have identified disease-associated Ym1 polymorphism, with conserved and widespread variation in Ym1 expression, and thereby function, in mice. An increasing body of literature points toward the role of catalytically active chitinases and chitinase-like proteins (C/CLPs) and their involvement in infection, inflammation, tissue injury, and remodeling (2935). However, besides the enzymatic activity, we still do not know how to associate C/CLPs with many immunomodulatory effects. Worse still, unlike other C/CLPs, the lack of Ym1 gene knockout mice until now has limited research on Ym1 function. But now with the low-expressing allele isolated in the congenic strain, we can initiate definitive physiological studies of the role of Ym1 in inflammatory diseases.

We identified the Ym1 gene using the CIA model but later found a stronger effect along a pathway that was induced by mannan, which is important for the induction of a range of different inflammatory disease; here, we tested psoriasis, PsA, and autoimmune arthritis in mice (22, 24). During this study, we also found that exposure of lung macrophages lacking the capacity to produce ROS (reactive oxygen species) with mannan leads to chronic relapsing arthritis. This could therefore be a model that could explain why smoking and air pollution could enhance the severity of many chronic inflammatory diseases (36). Ym1 exerted its role in regulating arthritis by modifying function in local tissue macrophages in the lung. Because Ym1 is strongly expressed in lung macrophages, we tried to observe its direct regulatory effect in the lung using both mannan-induced and classic antigen–induced pneumonitis models. Reduced pathological score, as well as decreased eosinophil infiltration, type II cytokine expression, and IgG1 production all suggest that lung inflammation is alleviated at low levels of Ym1 expression.

Previous studies using mice deficient in oxidative burst due to loss-of-function mutations in components of the NOX2 complex–like Ncf1 (p47phox) or Cyp1a (gp91phox) and certain colonies of SWISS mice indicated that these mice develop a severe crystalline macrophage or eosinophilic pneumonia characterized by crystallization of Ym1 protein in the lungs of affected animals (37, 38). This is followed by massive infiltration of the respective granulocytes displaying a hyperinflammatory phenotype, thereby producing a whole range of cytokines. In our study, mannan intranasally induced arthritis and pneumonitis in Ncf1-mutated mice suggests the interactive contribution of Ncf1 and Ym1 to the pathogenesis of disease. Further investigation of their interaction mechanism would be a fruitful direction for future research.

Ym1 is predominantly expressed by macrophages and is also commonly used as a signature marker for AAMs (15). However, little is known about the regulatory role of Ym1 in the alternative activation of macrophages. It was found that iNOS-deficient mice displayed enhanced CAM polarization (39), which gives us a new understanding of the role of polarization markers on regulating macrophages. Acidic mammalian chitinase (AMCase) was also found to be of critical importance in the control of type 2 immune responses because knocking in enzymatically inactive AMCase enhanced type 2 immune responses to inhaled house dust mites (40). A recent study also found opposing effects of Ym1 on type 2 immunity during nematode infection (41), which suggests the diverse roles of C/CLPs in regulating the immune response. In the present study, we found that Ym1-deficient macrophages showed an enhanced AAM phenotype. The increased expression of Ym1 during macrophage polarization might act as a brake to control or limit the induction of AAM markers Arg1 and MRC1. Our finding reveals a novel function of Ym1 in controlling the alternative activation of macrophages.

It has been demonstrated that CAMs favor enhanced glycolysis, whereas AAMs favor β-oxidation orchestrated by STAT6 and PPARγ-coactivator-1β (PGC-1β) (28). Specifically, STAT6 triggered by IL-4–inducing PGC-1β could play a critical role in alternative activation of macrophages. PGC-1β–regulated PPARs, such as PPAR-γ and PPAR-δ, have also been confirmed to control macrophage polarization (42, 43). We found that Ym1 deficiency increased the activation of STAT6, a finding confirmed by proteomics data indicating that Ym1 deficiency enhanced the lipid metabolism and PPAR pathways. Inhibition of STAT6 and PPAR-γ by RNAi and inhibitors could block the process of alternative activation of macrophages. On the basis of these observations, we suggest that STAT6 and PPAR-γ pathways could be one of the mechanisms by which Ym1 regulates macrophage polarization.

The sequences of Ym1 and other C/CLPs in human and mice are highly conserved, suggesting the possibility that C/CLPs share physiological and pathophysiological functions. Despite limited understanding of their mechanisms, human C/CLPs have been shown to have a strong association with pneumonitis, and chitinase inhibitors also show potential for the treatment of respiratory system diseases (4447). The identification of Ym1 polymorphism and its association with inflammation emphasize the importance of this protein and offer a new tool for understanding the regulatory mechanisms of C/CLPs in pneumonitis.

We conclude that the natural polymorphism of Ym1 controls its expression and thereby will influence the alternative activation of macrophages in the control of pneumonitis. Our results raise the possibility that Ym1 polymorphism might be correlated with differential susceptibility not only to pneumonitis but also to parasite infections. Our data identify chitinase-like genes as critical regulators of macrophages in immune responses, thus opening new possibilities for understanding and therapeutic intervention of pneumonitis and other inflammatory diseases.


Mice and ethic statement

B10.RIII H2r MHC (major histocompatibility complex) haplotype–bearing mice were originated in J. Klein’s laboratory (Max Planck Institute for Biology, Tübingen) and maintained in our breeding colony. RIIIS/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The BR.Cia22.Chitinase (henceforth denoted BR.Ym1Δ) congenic founder was obtained by recombination-assisted breeding from a partial advanced intercross previously described and subsequently backcrossed (12). To ensure that BR.Ym1Δ mice were devoid of contaminating RIIIS/J alleles, strain purity of congenic mice was assessed using a custom-made 8k Illumina SNP chip (48). No SNPs were observed between congenic and B10.RIII parental strains. Ym1 congenic mice were crossed with Ncf1m1J point mutation mice on the B10.RIII background (named as BR.Ncf1*.Ym1Δ). All above mouse strains, as well as C57BL/6N (B6N) and Balb/c, were kept under specific pathogen–free environment in the animal housing facilities of the Department for Medical Biochemistry and Biophysics, Section for Medical Inflammation Research, Karolinska Institute, Stockholm. They were housed in individually ventilated polystyrene cages containing wood shavings in a climate-controlled environment with a 14-hour light cycle and fed with standard rodent chow and water ad libitum. For all experiments, 10- to 14-week-old age- and sex-matched congenic and wild-type littermate controls have been used. All experimental procedures were approved by local ethical committees (Stockholm, Sweden; permit numbers M107/07, M109/07, N169/10, N66/10, N490/12, and N35/16).

Genomic DNA was isolated from toe tissue samples using the sodium hydroxide extraction method. Microsatellite markers were amplified using standard PCR. The size of the PCR fragments was determined using an ABI AB3730 DNA Analyzer system (Applied Biosystems, Foster City, CA, USA).

RNA quantitation using qPCR

Spleen, thymus, lymph nodes, stomach, lung, liver, heart, kidney, small intestine, colon, skin, brain, and bone marrow cells were obtained from age- and sex-matched congenic mice and wild-type counterparts after carbon dioxide asphyxiation. Total RNA from the above tissues was isolated by using TRIzol Reagent (Invitrogen), and complementary DNA (cDNA) was synthesized using the First Strand cDNA Synthesis Kit (Fermentas). RT-qPCR was performed by Agilent Stratagene Mx3005P with FastStart Universal SYBR Green Master (Roche) to assess the gene expression. The relative gene expression normalized by β-actin was calculated with the 2−ΔΔCT method. The primers used were described in table S1. In addition, PCR production of Ym1 and Ym2 in the indicated tissues of B10.RIII wild-type mice was separated and determined by DNA electrophoresis on 1.5% agarose gel.

Transcriptional activity assay of Ym1 promoter

The genomic DNA was extracted from B10.RIII and RIIIS/J mice, and the promoter of Ym1 gene, a 2074-bp DNA fragment (from −1788 to +274, the transcriptional start site was denoted as +1), was isolated and cloned to the pGL4.17 plasmid by Biomatik, which were denoted as pGL-B10.RIII and pGL-RIIIS/J, respectively. Then, the site mutations were performed using pGL-B10.RIII as the template to make the promoter containing one of four SNPs according to a commercial site mutation kit (Q5 Site-Directed Mutagenesis Kit, New England Biolabs), which were named as pGL-SNP1, pGL-SNP2, pGL-SNP3, and pGL-SNP4. The primers used for site mutation were described in table S1. All plasmids, including the pRL-TK vector (Promega, Fitchburg, USA) that served as a control, were prepared with the EZNA Endo-free Plasmid Mini Kit (Omega Bio-tek, Norcross, USA). The constructs were sequenced to prove sequence integrity before use.

Human embryonic kidney (HEK) 293T cells cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, USA) containing 10% fetal bovine serum (FBS) (Gibco, USA) were used for a dual-luciferase reporter assay. Briefly, both the pGL4.17-derived plasmids and renilla pRL-TK control vector (270 ng:30 ng per well) were transfected into HEK293T cells (5 × 105 cells per well seeded for 24 hours before transfection) using Lipofectamine 3000 (Invitrogen, Carlsbad, USA) in a 48-well culture plate. The luciferase activity was detected using the Dual-Luciferase Reporter 1000 Assay System (Promega) with a plate-reading luminometer (Synergy 2, BioTek, USA) 48 hours after transfection, and the relative luciferase activity value was achieved against the renilla luciferase control.

Collagen antibody–induced arthritis

Antibodies were produced and purified according to standard protocols already described elsewhere (49). Mice were injected intravenously with a monoclonal antibody cocktail in PBS containing the antibodies CIIC1, M2139, CIIC2, and UL1 using 1 mg per antibody and, in total, 4 mg of antibodies per mouse (49). Five days after the initial antibody transfer, the mice were challenged with intraperitoneal injection of 25 μg of LPS (Sigma-Aldrich, St. Louis, MO, USA) or 2 mg of mannan (Sigma-Aldrich, St. Louis, MO, USA). Clinical phenotyping for arthritis development was carried out by using a common macroscopic scoring system based on the number of inflamed joints in the paw defined by swelling and redness. Briefly, mouse paws were scored, giving each swollen ankle five points, whereas each swollen knuckle was given one point, resulting in a maximum score of 60 per mouse (50). Mice were scored twice or thrice per week to monitor clinical arthritis.

Mannan-induced psoriasis

Mice with Ncf1 mutation were administrated intranasally with 20 mg of mannan to induce psoriasis and PsA (24) and were scored daily for arthritis severity using the same scoring system as mentioned above. The severity of the psoriasis-like skin manifestations was monitored on a 15-score system, in which ears or each paw was evaluated by a scale ranging from 1 to 3 (1, weak skin peeling; 2, moderate skin peeling; and 3, heavy skin peeling with some hair loss) (24). For the in vivo macrophage depletion experiment, 50 μl of clodronate-liposome or control PBS-liposome (Liposoma BV, The Netherlands) was administrated intranasally to mice 2 days before arthritis induction. The depletion efficiency was assessed by flow cytometry analysis. For the in vivo Ym1 protein supplement experiment, mice were intranasally administrated with mannan to induce arthritis at day 0, and recombinant Ym1 protein (1 μg per mouse; Sino Biological Inc.) was treated intranasally at days 0, 2, and 4 after arthritis induction.

Mouse models of pulmonary inflammation

For the mannan-induced pulmonary inflammation model, BR.Ncf1* and BR.Ncf1*.Ym1Δ mice were treated intranasally with mannan (0.5 mg per mouse). Mice were sacrificed at days 0, 2, and 5. Lung weight normalized by body weight (lung index) was calculated. Lung tissues were prepared for hematoxylin and eosin (H&E) staining. Inflammatory cells in BALF were analyzed by flow cytometry.

For the induction of AIPI model, B10.RIII and BR.Ym1Δ mice were sensitized by the intraperitoneal injection of 0.2 ml of emulsion solution mixed with 10 μg (in 100 μl) of OVA (Sigma-Aldrich, St. Louis, MO, USA) and 100 μl of Imject Alum (Pierce, Thermo Fisher Scientific) at days 0 and 7. During days 14 to 20, mice were challenged with 1% OVA aerosol for 30 min per day. Control mice were sham-sensitized and exposed to the same volume of PBS. Mice were sacrificed at day 21. Lung tissues were prepared for histological analysis and cytokine expression detection. Inflammatory cells in BALF were analyzed by flow cytometry. Sera were collected for determining the concentration of Ym1- and OVA-specific IgG1 and IgE.

Lung histology

Lung tissues were stained with H&E to observe pathological changes, with periodic acid–Schiff (PAS) to detect mucus production. The scoring of lung histology was performed in a double-blind fashion. First, a scoring system to estimate the severity of peribronchial infiltration of inflammatory cells is as follows: 0, no cells; 1, a few cells; 2, half ring of cells around airway; 3, a ring of cells and one cell layer deep; 4, a ring of cells and two cell layers deep; 5, a ring of cells and more than two cell layers deep or cell aggregation. Second, to evaluate the inflammatory cell infiltration in lung interstitium, total cell numbers in interstitium were counted by using the Image-Pro Plus software. In addition, to estimate the extent of mucus production, the integral optical density (IOD) of the airways with PAS staining was analyzed by using the Image-Pro Plus software.

Flow cytometry

Single-cell suspensions from BALF cells were labeled with the following monoclonal antibodies in the respective experiments: anti-CD11b–Pacific blue, anti-Ly6G–Alexa Fluor 700, anti-CD11c–APC (allophycocyanin), anti-F4/80–BV605, anti–Siglec F–PerCP (peridinin chlorophyll protein)–Cy5.5, anti-Arg1–FITC (fluorescein isothiocyanate), anti-iNOS–PE (phycoerythrin)–Cy7, anti-MRC1–PE, anti-CD11b–APC, anti-CD11c–APC-Cy7, anti-Ly6G–Pacific blue, anti-F4/80–FITC, anti-TCRβ–Alexa Fluor 700, anti-B220–PE-Cy7, and anti-Ym1–PE. All of these antibodies were sourced from either Becton Dickinson, eBioscience, BioLegend, or Abcam. For intracellular staining, cells were fixed and permeated by using BD Cytofix/Cytoperm before intracellular antibody incubation. For Ym1 intracellular staining, cells were incubated with brefeldin A (10 μg/ml; Selleck) for 4 hours before staining. Samples were analyzed using FACS LSR II and FlowJo software (TreeStar Inc.).

Bone marrow–derived macrophages and polarization

The primary bone marrow–derived cells were isolated from B10.RIII, BR.Ym1Δ, BR.Ncf1*, and BR.Ncf1*.Ym1Δ mice and cultured in DMEM containing 10% FBS and M-CSF (macrophage colony-stimulating factor) (10 ng/ml) for 7 days to differentiate bone marrow–derived macrophages (BMMs). For macrophage polarization, LPS (100 ng/ml) and IFN-γ (10 ng/ml) were treated to BMM to polarize the CAMs, while IL-4 (20 ng/ml) and IL-13 (50 ng/ml) were treated to BMM to polarize the AAMs. BMM was also treated with IL-10 (20 ng/ml) to compare with IL-4/IL-13–treated BMM. The PPAR-γ inhibitor T0070907 (1 μM; Selleck) or the CPT1 inhibitor etomoxir (200 μM; Selleck) was used to block the alternative activation process of macrophages. The siRNA of mouse STAT6 and PPAR-γ was designed and purchased from GenePharma (Shanghai, China) (sequences described in table S2). Macrophages were transfected with 50 nM siRNA using Lipofectamine 2000 for 24 hours, followed by IL-4/IL-13 stimulation. RNA and protein were extracted from cells for gene and protein expression assay, and cultured media were collected for cytokine production assay.

Enzyme-linked immunosorbent assay

For quantification of Ym1 levels, a commercially available ELISA kit was used according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN, USA). The serum levels of OVA-specific IgG1 and IgE were also measured by ELISA. Briefly, the 96-well plates were coated for overnight reaction at 4°C with OVA (5 μg/ml) in PBS. After 1% bovine serum albumin (BSA) blocking and washing, diluted serum samples were added to the plates with 2-hour incubation at room temperature. Then, rat anti-mouse IgE-biotin (SouthernBiotech) or rat anti-mouse IgG1-biotin (Bio-Rad) antibodies were incubated, followed by streptavidin–horseradish peroxidase (HRP) incubation. Last, ABTS (Roche) was added as a substrate, and reactions were read at 405 nm using an ELISA reader (Thermo Electron Corporation). IL-6, IL-12, and IL-10 levels from the BMM culture supernatant were determined by ELISA using their capture antibodies and biotin-labeled detection antibodies (IL-6 and IL-12, eBioscience; IL-10, R&D Systems). Binding of antibodies was detected by europium-labeled streptavidin using a dissociation-enhanced lanthanide fluoroimmunoassay system (Wallac).

Western blotting

Total protein lysates from cells were extracted by using the radioimmunoprecipitation assay (RIPA) solution with a cocktail of protease and phosphatase inhibitors (Roche). The final protein concentration of each sample was determined with a BCA kit (Thermo Fisher Scientific). The supernatants (20 μg of total protein) from protein lysates were subjected to SDS–polyacrylamide gel electrophoresis (PAGE) gel (Invitrogen). The following primary antibodies were used: mouse anti-Arg1 antibody (Santa Cruz Biotechnology), rabbit anti-MRC1 antibody (Abcam), rabbit anti–phospho-STAT6 (Tyr641) antibody (Cell Signaling Technology), rabbit anti-STAT6 antibody (Cell Signaling Technology), rabbit anti–PPAR-γ antibody (Cell Signaling Technology), rabbit anti–PPAR-δ antibody (Abcam), rabbit anti–PPAR-α antibody (Abcam), and anti–β-actin antibody (Abcam) overnight. The signal was further detected by using the secondary antibody goat anti-rabbit IgG conjugated with HRP or goat anti-mouse IgG conjugated with HRP. Signal intensity was determined with the Immobilon Western Kit (Millipore).

Label-free quantification proteomics

Thioglycollate-elicited macrophages were isolated from B10.RIII mice and BR.Ym1Δ congenic mice (n = 4 per group), and the label-free quantification proteomics analysis was performed (Novogene, China). Briefly, proteins were extracted, and the concentrations were determined by Bradford protein assay (Bio-Rad, USA). After proteins were digested with Trypsin Gold (Promega), shotgun proteomics analyses were performed using an EASY-nLC 1200 UHPLC system (Thermo Fisher Scientific) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) operating in the data-dependent acquisition mode. Data were searched against the UniProt database for Mus musculus (85,165 sequences, downloaded 18 January 2019) database by the search engine Proteome Discoverer 2.2 (Thermo Fisher Scientific). For sequence annotation, information was extracted from the KEGG and GO resource.

Statistical analysis

Quantitative data were expressed as means ± SEM. The statistical analysis of differences between experimental groups was performed using one-way analysis of variance (ANOVA) or Student’s t test. A P value of less than 0.05 was considered significant.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We would like to thank C. Palestro, K. Palestro, and L. Li for excellent animal care, and we thank X. Li, Y. Jiang, F. Gao, W. Bai, P. Zhu, K. Wei, and L. Xiong for technical assistance. Funding: The work was supported by grants from the Swedish Strategic Science Foundation (SSF), the Knut and Alice Wallenberg Foundation, the Swedish Research Council, the Erling Persson Foundation, the National Natural Science Foundation of China (81671629 and 81970029), the Shaanxi Province Natural Science Foundation (2020JQ082 and 2018JM7057), and the Fundamental Research Funds for the Central Universities (xjj2018159). Author contributions: W.Z. designed the research, performed the experiments, including acquiring and analyzing data, and wrote and revised the manuscript. E.L. performed arthritis experiments, analyzed the data, and revised the manuscript. M.F. performed the arthritis experiments, analyzed the data, and revised the manuscript. M.J. performed the positional cloning experiment, analyzed the data, and revised the manuscript. P.T. performed the lung inflammation experiment and analyzed the data. L.M. performed the primary macrophage experiment, analyzed the data, and revised the manuscript. S.L. designed the research and revised the manuscript. R.H. designed the research, analyzed the data, revised the manuscript, and supervised and took the overall responsibility of the study. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article