Upregulated PKM2 in Macrophages Exacerbates Experimental Arthritis Via STAT1 Signaling

Upregulated PKM2 in Macrophages Exacerbates Experimental Arthritis via STAT1 Signaling

This information is current as Jing Xu, Congshan Jiang, Xipeng Wang, Manman Geng, of September 29, 2021. Yizhao Peng, Yuanxu Guo, Si Wang, Xiaowei Li, Pei Tao, Fujun Zhang, Yan Han, Qilan Ning, Wenhua Zhu, Liesu Meng and Shemin Lu J Immunol published online 5 June 2020

http://www.jimmunol.org/content/early/2020/06/04/jimmun Downloaded from ol.1901021

Supplementary http://www.jimmunol.org/content/suppl/2020/06/04/jimmunol.190102 Material 1.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 5, 2020, doi:10.4049/jimmunol.1901021 The Journal of Immunology

Upregulated PKM2 in Macrophages Exacerbates Experimental Arthritis via STAT1 Signaling

Jing Xu, Congshan Jiang, Xipeng Wang, Manman Geng, Yizhao Peng, Yuanxu Guo, Si Wang, Xiaowei Li, Pei Tao, Fujun Zhang, Yan Han, Qilan Ning, Wenhua Zhu, Liesu Meng, and Shemin Lu

Recent studies indicate that glucose metabolism is altered in rheumatoid arthritis. We hypothesize that Pkm2, as a key regulatory enzyme of glycolysis pathway, triggers the activation of macrophages (Mw), which results in proinﬂammatory cytokine production during the arthritis progress. In this study, Pkm2 was found to be overexpressed in ED1-positive Mw in spleens and synovial tissues from arthritic rats via immunoﬂuorescence, Western blotting, and quantitative RT-PCR. To reveal the role of Pkm2, Dark Agouti rats were treated with either Pkm2 enzyme inhibitor shikonin or the RNA interference plasmids of Pkm2 and negative control plasmids, respectively, via i.p. injection. Pkm2 intervention could alleviate the severity of pristane-induced arthritis in Downloaded from aspects of the macroscopic arthritis score, perimeter changes of midpaw, and the synovitis and destruction of the bone and cartilage as well as reduce the ED1 and p-Stat1–positive cell population in rat synovial tissues. Silencing Pkm2 by RNA interference in classical activated rat and mouse Mw resulted in less Tnf-a, Il-1b production via Stat1 signaling. Collectively, Pkm2 is highly expressed in ED1-positive Mw of spleens and synovial tissues from arthritic rats and promotes Mw activation via Stat1 signaling. Pkm2 might be a promising selective metabolic target molecule for rheumatoid arthritis treatment. The Journal of Immunology, 2020, 205: 000–000. http://www.jimmunol.org/

heumatoid arthritis (RA) is a systemic autoimmune dis- mononuclear phagocyte system, such as circulating monocytes ease that leads to chronic inflammation and progressive from spleen and bone marrow, can also be activated to involve in R joint destruction. One of the apparent histopathological the pathogenesis of RA (1). changes in RA joints is accumulating macrophages (Mw)infil- Recently, different approaches have highlighted underlying trating into the inflamed synovial membrane and the cartilage- mechanisms that explain RA behavior, including activation and pannus junction region. These cells of the innate immune system modifications of signaling pathways. RA metabolism has gained possess broad capacities to produce proinflammatory cytokines, attention as recent studies have pointed out modifications in RA by guest on September 29, 2021 destruct cartilage and bone, and remodel the structure of joints immune cells and fibroblast-like synoviocytes (FLS) metabolism in both the acute and chronic phases of RA. Other types of (2). Immune cells choose different metabolic pathways in differentiation and maintenance of their phenotypes. Activated CD4+ T cells in RA patients shunt glucose flux to the pentose Department of Biochemistry and Molecular Biology, School of Basic Medical Sci- phosphate pathway leading to a hyperreduced state, due in part ences, Xi’an Jiaotong University Health Science Center, Xi’an 710061, Shaanxi, to a deficiency in phosphofructokinase PFKFB3 (3). In the People’s Republic of China; and Key Laboratory of Environment and Genes Related to Diseases, Xi’an Jiaotong University, Ministry of Education, Xi’an 710061, zymosan-induced arthritis model, treatment with fructose 1,6- Shaanxi, People’s Republic of China bisphosphate (FBP), the product of PFKFB3, attenuates disease ORCIDs: 0000-0002-2532-3776 (J.X.); 0000-0002-7457-6243 (C.J.); 0000-0002- severity (4). Lymphocytes infiltrating the joints of RA patients 0718-1938 (X.L.); 0000-0002-2820-2427 (W.Z.); 0000-0003-1985-050X (L.M.); express high level of hexokinase 2 (HK2), indicating that they 0000-0002-5535-8320 (S.L.). also exist in a highly glycolytic state similar to their activation Received for publication August 22, 2019. Accepted for publication April 22, 2020. stages (5). HK2 is also overexpressed in FLS from RA patients, This work was supported by grants from the National Natural Science Foundation of and its knockout in mouse FLS could reduce disease severity China (81701619 and 81671629) and the Fundamental Research Funds for the Cen- tral Universities (Xjj2017143 and syspz2017006). (6). Inhibition of glycolysis with bromo-pyruvate greatly re- J.X. and S.L. conceived, designed, and applied the research, analyzed data, and wrote duces disease severity in the SGK mice (5), another zymosan- the paper; J.X., C.J., X.W., M.G., Y.P., Y.G., S.W., X.L., P.T., F.Z., Y.H., Q.N., W.Z., induced model of arthritis driven by Th17 cell expansion dependent and L.M. contributed to acquisition and/or analysis of data. All of the authors ap- on glycolysis (7). Inhibition of glycolysis with 2-deoxyglucose proved the final version of the paper. significantly reduces joint inflammation and the activation of both Address correspondence and reprint requests to Dr. Shemin Lu, Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences, Xi’an adaptive and innate immune cells as well as the production of Jiaotong University Health Science Center, Yanta West Road, Xi’an 710061, pathogenic autoantibodies in KBN mouse (8). Key signaling path- Shaanxi, People’s Republic of China. E-mail address: [email protected] ways that are activated by the inflamed microenvironment The online version of this article contains supplemental material. converge to adapt cell metabolism to support immune cell ac- Abbreviations used in this article: BMM, bone marrow–derived Mw; CIA, collagen- tivation, suggesting that the study of metabolic changes in RA induced arthritis; DA, Dark Agouti; FLS, fibroblast-like synoviocyte; IF, immunofluorescent; Mw, macrophage; NC, negative control; PIA, pristane-induced arthritis; and key players could potentially lead to the identification of PKM2, pyruvate kinase muscle type 2; RA, rheumatoid arthritis; RNAi, RNA inter- new therapeutic agents. ference; RT-qPCR, quantitative RT-PCR; shNC, short hairpin NC; shPkm, short Among other metabolic changes, the recent works have high- hairpin Pkm; shRNA, short hairpin RNA; siRNA, small interfering RNA. lighted a critical role of glucose metabolism in activated Mw.We Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 have previously shown that activated autophagy in Mw promotes

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1901021 2 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL pristane-induced arthritis (PIA) in rats with proinflammatory cy- Mw activation tokine production and Mw activation (9). Furthermore, a study has Rat Mw named NR8383 cell line were cultured in F-12K medium with shown increased pyruvate kinase muscle type 2 (PKM2) via TLR2 15% FBS. Bone marrow cells were isolated from DA rats and seeded at the signaling in RA patients’ FLS (10). There are four isozymes of density of 2 3 106/ml in RPMI 1640 with 10% FBS and 20 ng/ml M-CSF pyruvate kinase (PK) in mammals (PKL, PKR, PKM1, PKM2) for differentiating bone marrow–derived Mw (BMM). The other Mw cell encoded by two different genes, PKLR and PKM. The PKL and line RAW264.7 was also used to observe the loss of function of Pkm2 in activated Mw. BMMs and Mw cell lines were stimulated by 100 ng/ml LPS PKR isozymes are generated from the PKLR by differential and 100 ng/ml IFN-g to induce classical Mw activation, observe signal splicing of RNA; the PKM1 and PKM2 forms are produced from activation, and analyze downstream gene expression. In Pkm2 activation the PKM gene by differential splicing (11, 12). PKM1 is the main study, two Pkm2 activators DASA-58 (10 mM) or TEPP-46 (10 mM) were form in muscle, heart, and brain, and PKM2 is found in early fetal treated 1 h before LPS and IFN-g were added. All the cell experiments were repeated three times. tissues as well as in most cancer cells. The identification of isoform-specific contributors to elevated cell glucose metabolism Plasmid construction without compromising systemic homeostasis or normal metabolic Plasmids were constructed by the standard method, including PCR, overlap functions could make targeting cell metabolic changes a more- extension PCR, and restriction fragment ligation, and verified by se- feasible approach. In this study, we demonstrate that Pkm2 is a quencing all cloning junctions and PCR products. All the primers for pkm2, specific isoform contributing to maintaining the proinflammatory stat1, stat3, clover, and mRuby2; product sizes; and annealing temperatures microenvironment in experimental arthritis and a key regulator are depicted in Table I. The CDS fragments of clover and mRuby2 were during arthritis process. Our results show that Pkm2 expression is cloned from pcDNA3.1-mRuby2 (19) (no. 40260; Addgene) and pLenti- FoxO1-Clover (20) (67759; Addgene). elevated in the spleens and synovial tissues from the arthritic rats. The short hairpin RNA (shRNA) oligonucleotides of pkm and stat1 Downloaded from Pkm2 RNA interference (RNAi) could ameliorate disease severity genes were selected from The RNAi Consortium library database, and of experimental arthritis and decrease inflammatory cytokine ex- the details are given in Table II. The rat Pkm shRNA oligonucleotides pression in spleen. Taken together, our data suggest that Pkm2 is were selected from the online software (http://sirna.wi.mit.edu/home.php). All the shRNA oligonucleotides were synthesized by the company. The involved in Mw activation through Stat1 activation and could be a pLKOG-TRC plasmid was used to reconstruct the shRNA plasmids and potential target to guide future development of therapeutic strat- sequenced by Sangon Biotech Company. egies for RA. http://www.jimmunol.org/ RNAi Materials and Methods For the high efficiency of transfection, the reconstructed shRNA plasmid, All the reagents used in this paper were recorded in Supplemental Tables I psPAX2, and pMD2.G plasmids were extracted by E.Z.N.A. Endo-free and II. Plasmid Maxi Kit. We then transfected the shRNA plasmid together with lentivirus package plasmids into 293T cells by TurboFect Transfection Induction of arthritis Reagent. All the methods were carried out following the described pro- cedure (21). The lentivirus medium was collected through 0.45-mm filter All Dark Agouti (DA) rats were bred in a specific pathogen–free animal and then added to NR8383 cells or RAW264.7 cells. The positive cells house of the Department of Biochemistry and Molecular Biology, School were selected by 3 mg/ml puromycin for 7 d. Then the efficiency of Pkm of Basic Medical Science, Xi’an Jiaotong University Health Science shRNA was determined by Western blotting. by guest on September 29, 2021 Center. All animal experiments were approved by the Institutional Animal Both small interfering RNA (siRNA) targeting Pkm2 (59-TTCG- Ethics Committee of Xi’an Jiaotong University (No. 2017-666). All the GAGGTTTGATGAAATC-39) and negative control (NC; 59-GCGACG- animal experiment designs followed Animal Research: Reporting of AUCUGCCUAAGAUTT-39) were purchased from the company. BMMs In Vivo Experiments guidelines (13). were transfected with either Pkm2 siRNA or NC siRNA at 75 nmol/l using To establish the PIA, DA rats at the age of 8–12 wk were given a Lipofectamine 2000 Transfection Reagent according to the manufacturer’s m single intradermal injection of 300 l of pristane as described previ- guidelines. The cells were transfected for 24–48 h for further experiments. ously (14). To establish the collagen-induced arthritis (CIA), DA rats at the age of 8–12 wk were given a single intradermal injection of 300 mg Intervention of Pkm2 in vivo of type II collagen with IFA as described previously (15, 16). Eight- to twelve-week age- and sex-matched rats were arranged randomly for Sixteen gender-matched DA rats at age 8–12 wk were randomly divided each group. All the arthritic rats were checked every 2 d. A scoring into two groups, DMSO control and shikonin group. All rats were ad- system of 1–58 points per rat was used as described previously (17). In ministrated a single intradermal injection of 300 ml of pristane to induce brief, one point was given for each inflamed knuckle or toe and up to arthritis. After pristane was injected for 6 d, the rats were treated by i.p. five points was given for an affected ankle (in total 15 points for the injection of 500 ml of 5 mg/kg shikonin/DMSO in PBS every other day for hind paw and 14 points for fore paw). All scoring was performed seven times or the same volume of the DMSO in PBS according to the blinded. Rats were sacrificed according to the experiment plans, and same scheme (Fig. 4A). In this case, the maximal DMSO concentration spleens or hinds were collected for further microscope observation and would be 3.2%. All rat arthritis macroscopic arthritis scores were blindly expression analysis. recorded. All the shRNA plasmids were extracted by E.Z.N.A. Endo-free Histology and immunofluorescence Plasmid Maxi Kit and determined by Nanodrop (Thermo Fisher Sci- Histological observation was performed on the spleen or paw paraffin entific) for the concentration. The rat Pkm shRNA plasmids 1, 2, and sections from DA rats, after staining with H&E. The pathological severities 3 were mixed together to be used (sh2 and sh3 targeted sequences were were estimated by two aspects. The synovitis was valued via four items: located in Pkm2). the thickness of synovial lining layers, pannus, synovial inflammatory Thirty-two gender-matched DA rats at age 8–12 wk were divided into cells, and angiogenesis. The joint destruction was estimated via cartilage four groups: control, PIA with endotoxin-free water, PIA with NC plas- erosion, bone erosion, joint ankle loss, and change of articular structure. In mids, and PIA with Pkm shRNA plasmids. Briefly, the rats were i.p. ad- m brief, one to three points was given for each item, in total 12 points in each ministrated 200 g/rat of indicated plasmids, and 1 d later all the PIA rats m slide/rat ankle for synovitis or joint destructions. The spleen tissue and paw received a single intradermal injection of 300 l of pristane. Arthritis paraffin sections (5 mm) were used for immunofluorescent (IF) staining. development and severity evaluation were blindly recorded. The plasmids The primary Abs included ED1 (CD68) and Pkm2, or p-Stat1, and the were injected twice a week, indicated in the scheme (Fig. 4D). Twenty secondary Abs were FITC AffiniPure Goat Anti-Mouse IgG (H+L) and days after pristane injection, all the rats were sacrificed, and spleens and Cy3 AffiniPure Goat Anti-Rabbit IgG (H+L). Antifade mountant with ankles were collected for further experiments. Ankle joints were sectioned DAPI was applied to all slides. IF staining was carried out, following the and stained with H&E or IF for histopathological examination (17). described procedures with slight modification (18). The isotype controlled Western blotting staining is shown in Supplemental Fig. 2E, 2G. The IF images were captured with a confocal fluorescence microscope (FV1000; Olympus) and Total protein lysates from spleen tissues and cells were extracted by using analyzed by Fluoview FV1000 and ImageJ software. the RIPA solution with a mixture of protease and phosphatase inhibitors. The Journal of Immunology 3

The final protein concentration of each sample was determined by a Statistics BCA kit. Total proteins (20 mg) from cell lysates were separated by 6 8 or 10% SDS-PAGE gels according to standard procedures with Bio- Data were expressed as mean SEM and analyzed using SPSS software. Rad system. The primary Abs were incubated at 4˚C overnight, which Shapiro-Wilk test was employed to validate the normal distribution of data. includes anti-PKM2, anti-STAT1, anti–p-STAT1, anti-STAT3, anti–p- Statistical analysis was performed by one-way ANOVA among groups, STAT3, anti–p-P38, anti-P38 Ab, anti-ERK, anti–p-ERK, anti–p-JNK, and the Student t test was employed to analyze the significant differ- anti-JNK, anti-IRF3, rabbit anti–p-IRF3, anti-iNOS, anti–IL-1b,and ences between the two groups. The Mann-Whitney U test was used to anti–b-actin. The signal was further detected by using the secondary analyze the density of Western blotting bands only. Statistical significance , , , Ab of goat anti-rabbit/mouse IgG conjugated with HRP. Signal inten- levels are expressed as *p 0.05, **p 0.01, ***p 0.001, and , sity was determined by Supersignal West Pico Kit. Data are expressed by ****p 0.0001. showing one representative image, and the density of the bands was measured by ImageJ software and normalized to b-actin or its own total Results protein. Pkm2 expression in spleen exhibits a significant increase in Immunoprecipitation arthritic rats DA rats were used to induce arthritis by pristane or type II collagen Total protein lysates from different treated cells were extracted by using the coimmunoprecipitation lysis solution with a mixture of with IFA for PIA and CIA, respectively. The histology of ankles protease and phosphatase inhibitors. The final protein concentration from PIA or CIA and control is shown in Fig. 1A. The ankles from of each sample was determined by a BCA kit. One hundred micro- both PIA and CIA rats were highly damaged. In the meantime, the grams of total proteins from cell lysates were collected, and then 1:50 expression of Pkm2 in the synovial tissue from arthritic rats’ an- Pkm2 Ab or its isotype control were added and incubated at 4˚C overnight. The next morning, the protein A/G beads were used in the kles was investigated. The results showed that both the Pkm2 and samples followed by the protocols. Then Pkm2, Stat1, and Stat3 were ED1 expression colocalized very well in synovial tissue from Downloaded from determined. Data were expressed by showing one representative arthritic rats (Fig. 1B, 1C), which indicated that the Pkm2 ex- image. pression upregulated along with ED1 expression cells. Further- RNA isolation and quantitative RT-PCR more, the spleens from a different stage of PIA were isolated and analyzed for mRNA (Table I) and protein of Pkm2 expression at Total RNA from spleen tissues and cells was isolated with TRI Reagent different time points. The results showed that both mRNA and Solution, and cDNA was synthesized by First Strand cDNA Synthesis Kit according to the manufacturer’s instructions. Quantitative RT-PCR (RT- protein expression level of Pkm2 was upregulated from day 6 to http://www.jimmunol.org/ qPCR) was performed by using iQ5 optical system software (Bio-Rad 26 (Fig. 2A, 2B). The perimeters of hind paws and ankles from Laboratories) with Fast Start Universal SYBR Green Master (ROX) for PIA rats also increased (Supplemental Fig. 1A, 1C). Meanwhile, mRNA quantitation of all referred genes. The information of all the pri- the correlation analysis between the Pkm2 mRNA expression level mers, products, and annealing temperatures is depicted in Table I. Gene expression analyses were performed against Actb expression for rats’ cells and the parameter changes of ankles and paws showed positive or tissues or against hprt expression for RAW264.7 cell and calculated by correlations (Supplemental Fig. 1B, 1D). Immunofluorescence re- using the 22DDCt method. sults showed that Pkm2 was mainly overexpressed in ED1-positive by guest on September 29, 2021

FIGURE 1. Pkm2 expression is upregulated in ED1-positive cells in synovial tissues from experimental arthritic rats. Ankle joints from normal, PIA, and CIA rats were stained with H&E (A). The “CA” marked the cartilage area, “BN” marked the bone area, and “SY” marked the synovial tissue area in the images. The ankle synovial tissues from PIA (B) and CIA (C) rats were stained with ED1 (green) and Pkm2 (red) Abs, and the nuclei were stained with DAPI (blue). The photos were merged automatically by the confocal microscope. 4 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL

Table I. RT-qPCR and CDS clone primer

Gene Name and NCBI Annealing Accession No. Sequence (59–39) Product Size (bp) Temperature (˚C) Rat Pkm2 F59-ATTGCCCGAGAGGCAGA-39 160 60 XM_008772937.1 R 59-ACTTGGTGAGCACGATAAT-39 Rat Tnf-a F59-TCAGCCTCTTCTCATTCCTGC-39 203 60 NM_012675 R 59-TTGGTGGTTTGCTACGACGTG-39 Rat Il-1b F59-CTGTGACTCGTGGGATGATGAC-39 322 54 NM_031512 R 59-CTTCTTCTTTGGGTATTGTTTGG-39 Rat Il-17a F59-CTACCTCAACCGTTCCACTT-39 191 65 NM_001106897 R 59-ACTTCTCAGGCTCCCTCTTC-39 Rat b-actin F59-GAGGGAAATCGTGCGTGAC-39 157 60 NM_031144 R 59-GCATCGGAACCGCTCATT-39 Mus tnf-a F59-ACGTCGTAGCAAACCACCAA-39 208 60 NM_013693.3 R 59-ATAGCAAATCGGCTGACGGT-39 Mus il-1b F59-CTGTGTCTTTCCCGTGGACCT-39 167 60 NM_008361.3 R 59-TTGTTGTTCATCTCGGAGCCT-39 Mus hprt F59-GATCAGTCAACGGGGGACAT-39 183 60 NM_013556.2 R 59-GGTCCTTTTCACCAGCAAGC-39 Mus pkm2 CDS F (EcoRI) 59-CGGAATTCATGCCGAAGCCACACAGT-39 1593 60

NM_011099.3 R (XhoI) 59-CCCTCGAGAGGTACAGGCACTACACGCAT-39 Downloaded from Mus stat1 CDS F (HindIII) 59-CCCAAGCTTATGTCACAGTGGTTCGAGC-39 2264 58 NM_001205313.1 R (BamHI) 59-CGGGATCCTACTGTGCTCATCATACTGTC-39 Mus stat3 CDS F (BamHI) 59-CGGGATCCATGGCTCAGTGGAACCAG-39 2310 58 NM_213659.3 R (EcoRI) 59-CGGAATTCCATGGGGGAGGTAGCACAC-39 Clover/mRuby2 CDS F (XhoI) 59-CCGCTCGAGATGGTGTCTAaGGGCGAAGAG-39 714 60 R (ApaI) 59-CAGGGGCCCTTACTTGTACAGCTCGTCCAT-39

F, forward; NCBI, National Center for Biotechnology Information; R, reverse. http://www.jimmunol.org/ cells in day 6, 12, and 26 PIA spleen (Fig. 2C). In the spleen (Fig. 4B). In the meantime, the histologic bone and cartilage de- of PIA rats, there was no difference in the ED1-positive cell struction scores of the shPkm plasmid treatment group also sig- (Fig. 2D). But the Pkm2 expression in the ED1-positive cells was nificantly declined compared with both water treatment control increased significantly after pristane priming (Fig. 2E). During the group and shNC plasmid treatment group (Fig. 4C). The fluores- arthritis process in DA rats, the Pkm2 was found overexpressed in cence staining showed that the number of both ED1 and Pkm2 ED1-positive cells. expression double-positive cells was much less after shPkm plasmid treatment (Fig. 4D, 4E). Thus, it is clear that suppressing by guest on September 29, 2021 Intervention of Pkm2 function significantly ameliorates the function of Pkm2 could ameliorate the arthritis process in arthritis symptoms in PIA rats PIA rats. Pkm2 was known as an enzyme to catalyze pyruvate into the w phosphoenolpyruvate. To investigate the overexpressed Pkm2 Increased Pkm2 enhances M classical activation via enzyme activity function in PIA rats, a Pkm2 special inhibitor phosphorylation of Stat1 shikonin was chosen to be applied on the PIA rats. All the pristane- Because Pkm2 expression was mainly in ED1-positive cells, the injected DA rats were treated with either shikonin or DMSO Mw cell line RAW264.7 and primary rat BMMs were used for (Fig. 3A). The macroscopic arthritis scores of the shikonin treatment investigating Pkm2 function. First, the rat BMMs were treated group were significantly decreased on day 16 and 19 compared with with LPS and IFN-g for Mw classical activation. The results the DMSO group (Fig. 3B). To further investigate the other Pkm2 showed that the expression of Pkm2 was upregulated along with function in arthritic rats, RNAi were applied on the PIA rats. The treatment time increasing significantly, whereas the expression NR8383 Mw were infected by lentivirus with different Pkm2 of Pkm1 did not change at all (Fig. 5A, 5B). Meanwhile, the shRNA to determine whether the expression of Pkm2 could be mRNA of Pkm2 expression was also upregulated in pristane- or downregulated. Results from the protein expression showed that all poly(I:C)–treated Mw (Supplemental Fig. 2A, 2B). The RAW264. three shRNAs could downregulate Pkm2 efficiently (Fig. 3C). All 7 cells were used and infected by lentivirus with different Pkm the shRNA plasmids of Pkm2 and NC were extracted and then shRNA (Table II) to determine the Pkm2 function in activated applied on DA rats according to the scheme (Fig. 3D). The mac- Mw. Results from the protein expression showed that after roscopic arthritis scores of the short hairpin Pkm (shPkm) treatment knocking down the expression of Pkm2 in RAW264.7 cells, the were significantly lower than that of the short hairpin NC (shNC) expression of Il-1b was decreased dramatically after stimulation and water treatment control group from day 8 to 20 after pristane with LPS and IFN-g (Fig. 5C, 5D). Meanwhile, the RT-qPCR injection (Fig. 3E). The pictures for hind ankles and paws of PIA results revealed that the mRNA expression of Tnf-a and Il-1b rats are shown in Supplemental Fig. 2A. The rat pad perimeters from (Table I) was also significantly decreased via Pkm2 RNAi Pkm plasmid treatment group were significantly shorter than the (Fig. 5E). When the Mw were treated with Pkm2 activators, shNC group in day 8 and also shorter than the water treatment DASA-58 or TEPP-46, which could form a Pkm2 tetramer in control group in day 8, 12, 16, and 20 (Supplemental Fig. 1F). the cell, the expression of iNOS and Il-1b was also blocked For further confirming the histology changes in PIA rats, all the (Fig. 5F, 5G). hind paws and ankles were fixed and prepared for histology de- Because proinflammatory gene expression was influenced after tection (Fig. 4A). The histologic synovitis scores of shPkm plas- Pkm2 knocking down, the signal pathways of classical activation of mid treatment group significantly reduced compared with both Mw were examined. We found that only the phosphorylation of water treatment control group and shNC plasmid treatment group Stat1 and Stat3 was significantly impacted after knocking down The Journal of Immunology 5

FIGURE 2. Pkm2 overexpresses in Downloaded from ED1-positive cells in the spleen from PIA rats. The mRNA (n = 8) and protein (n = 4) expression levels of Pkm2 in spleens from different time points after pristane injection were determined by RT-qPCR and Western blotting and compared with those from http://www.jimmunol.org/ day0(A and B). Representative images of immunoﬂuorescence staining for ED1 (green), Pkm2 (red), and nucleus (blue) in spleen tissue (n =4)werese- lected from different time points (day 6, 12, and 26) after pristane injection (C). The photos were merged automatically by the confocal microscope. The ED1-

positive cells were counted in the spleen by guest on September 29, 2021 (D), and the percentages of Pkm2- positive cells in ED1-positive cells were calculated in the spleen (E). Data are presented as mean 6 SEM. *p , 0.05, ***p , 0.001, ****p , 0.0001.

the Pkm2 expression (Fig. 6A, 6B) in RAW264.7 cells, whereas (Fig. 6C, 6D) in classically activated Mw. To further investigate the MAPK signal (p-P38, p-JNK, and p-ERK1/2) and Irf3 acti- the relation between Stat1/Stat3 and Pkm2, the recombined Pkm2 vation were not changed (Supplemental Fig. 2C). To confirm and Stat1/Stat3 were cloned with clover or mRuby2 and then this phenomenon, the rat BMMs were used and transfected with transfected in 293T cells. By doing so, expressing fluorescent siRNA. The results showed that when knocking down the ex- proteins allowed us to perform live imaging, ruling out artifacts pression of Pkm2, the phosphorylation of Stat1 was also decreased often observed by immunohistochemistry due to cell fixation 6 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. PIA is alleviated by the intervention of Pkm2 in vivo. DA rats were treated with shikonin or DMSO on day 6, after pristane injection (n = 8), according to the scheme (A). The macroscopic arthritis scores were given following the criterion mentioned in the Materials and Methods.*p , 0.05 between Shikonin group and the DMSO group (B). The protein expression of Pkm2 from NR8383 cells infected with shPkm/shNC lentivirus was detected by Western blotting (C)(n = 3). The relative expression was normalized by b-actin. DA rats were treated with plasmids i.p. (shNC, ShPkm, or endotoxin- free water) according to the scheme in (D)(n = 8). The macroscopic arthritis scores were compared among groups (E). Data are presented as mean 6 SEM. An asterisk (*) represents the shPkm group versus the shNC group, and a caret (^) represents the shPKM group versus the endo-free water. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001, ^^p , 0.01, ^^^p , 0.001, ^^^^p , 0.0001. protocols. As observed in Fig. 6E, both Pkm2 and Stat1/Stat3 cells, the expression of Il-1b also decreased after stimulation protein mainly colocalized in the cytoplasm. To determine the with LPS and IFN-g (Fig. 7B, 7C). Meanwhile, the mRNA relationship between Pkm2 and Stat1/Stat3 in activated Mw, im- expression of Il-1b also significantly decreased after Stat1 munoprecipitation was performed. In Fig. 6F and 6G, we showed RNAi (Fig. 7D, 7E). that Stat1 and Stat3 were able to be coprecipitated with Pkm2 in To further confirm the downstream effect of this Pkm2-Stat1 classical activated Mw. The above-mentioned results suggested signaling in PIA spleen tissue, proinflammatory cytokine Tnf-a Pkm2 might be involved in Mw classical activation via Stat and Il-1b expressions were detected. First, the expression of Pkm2 signaling. was determined in the spleen. As with the results shown in Fig. 8A and 8B, the expression of Pkm2 significantly reduced in PIA rats’ Downregulated Pkm2 significantly decreases proinflammatory spleens after multiple injections of knockdown plasmids. The re- cytokine expression via Stat1 in PIA rats sults further showed that the mRNA expressions of Tnf-a and Il-1b To confirm knocking down Pkm2 could impact the phosphorylation were significantly decreased in the Pkm2 shRNA plasmid treatment of Stat1, fluorescence straining was performed in the Pkm group compared with the NC group (Fig. 8C, 8D). In addition, the knockdown PIA rats. The results showed that both ED1 and p-Stat1 correlation between the Pkm2 mRNA expression level and double-positive cells could be found in synovial tissues in the PIA- the proinflammatory cytokines Tnf-a and Il-1b were analyzed. shNC group, whereas fewer ED1 and p-Stat1–positive cells could The results showed that there was a positive correlation between be found after shPkm plasmid treatment (Fig. 7A). Because the Pkm2 mRNA expression and Tnf-a and Il-1b mRNA expression phosphorylation of Stat1 was influenced in ED1-positive cells, the (Fig. 8E, 8F), whereas the mRNA expression of Il-17a displayed Stat1 knockdown RAW264.7 cells were selected (Supplemental only some decreasing trend compared with the NC group Fig. 2D). Results from the protein expression showed that (Fig. 8G), and the mRNA expression of Tlr3 was not altered at all after knocking down the expression of Stat1 in RAW264.7 (Fig. 8H). All these results suggested that the Pkm2 could be a The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 4. Intervention of Pkm2 could reduce histological score in PIA rats. Representative histological images of ankle joints stained by H&E are shown (A). The “CA” marks the cartilage area, “BN” marks the bone area, and “SY” marks the synovial tissue area in the images. The histological scores of synovitis (B) and bone and cartilage destruction (C) were quantified among groups (n = 8). Representative histological images of ankle joints stained by ED1 (green), Pkm2 (red), and DAPI (blue) are shown (D). The ED1-positive cells were counted in the ankle joint (E). Data are presented as mean 6 SEM (n = 4). *p , 0.05. good target to regulate the production of proinflammatory cyto- pathways protected from autoimmune disease, and the treatment kines in arthritis development. with nonmetabolizable glucose analogue 2-deoxy-D-glucose (2-DG) could reverse cytokine and autoantibody production in an animal Discussion model of arthritis and lupus (8, 22). However, 2-DG inhibits the The concept of metabolic reprogramming to improve immuno- hexokinase, which could inhibit the entire glycolysis pathway. therapy slowly is being translated into autoimmune diseases to Thus, there is a need for finding much more specific metabolic complement current therapies. Yet, there are few data about tar- targets that are induced in activated immune cells, such as Mw. geting metabolic changes in RA. We showed in this study that Pkm, the last regulatory enzyme of glycolysis metabolism, has targeting the last regulatory enzyme of glycolysis Pkm2 could two isoforms. Whereas Pkm1 is a ubiquitously expressed en- regulate Mw activation and proinflammatory cytokine production zyme in all living cells, Pkm2 is an inducible form and only and ameliorate arthritis symptoms and hence could be a potential reportedexpressedinearlyfetaltissues as well as in most cancer selective metabolic therapy for arthritis. cells. In rats, Pkm was previously reported as Pkm itself, without Although previous works have demonstrated the role of glucose additional isoform. In this study, from our data, we show that metabolism in cell activation and function, inhibiting general the rat also has two isoforms of Pkm. Our data suggest that Pkm2 glucose metabolism is not desirable in the overall body. In an constitutes an attractive potential selective target for arthritis important proof of studies, inhibition of glycolysis metabolic therapy. 8 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL Downloaded from http://www.jimmunol.org/

FIGURE 5. Pkm2 promotes Mw activation. Overexpressed Pkm2 was accompanied by increasing iNOS and IL-1b, classical activation markers of Mw for different time points (A and B)(n = 3). The stable cell lines, Pkm2/Pkm shRNA, and NC RAW264.7 cells were stimulated with LPS and IFN-g for different time points. The IL-1b protein expression for time point treatment with LPS and IFN-g (C and D) and the mRNA expressions of Tnf-a and Il-1b (E) were detected for 6 h with LPS and IFN-g (n = 3). The IL-1b and iNOS protein expression (F and G) was detected in Pkm2 activator DADS-58 (10 mM)/TEPP-46 (10 mM) pretreated Mw and LPS and IFN-g activated Mw (n = 3). Data are presented as mean 6 SEM. *p , 0.05.

In this study, we observed that Pkm2 is not expressed or is the Pkm2 might be modified in the inflamed splenocytes. Some expressed in low amounts in normal rats’ synovial tissues and is reports also show that Pkm2 modification also participates in cell by guest on September 29, 2021 weakly expressed in normal rats’ spleens, but its expression could process. For instance, O-GlcNAcylation of Pkm2 could destabilize be induced in synovial tissues and spleens after priming the ar- tetrameric Pkm2, which could promote the Warburg effect in thritis in DA rats. We also observed that Pkm2 expression was human cancer cells (23). Pkm2 could be translocated in the nu- consistent with the increased synovial inflammation and ankle cleus and phosphorylate STAT3, thus boosting IL-6 and IL-1b swelling. In addition, according to IF data, overexpressed Pkm2 production (24, 25). This also indicated that Pkm2 may play as a was mainly colocalized with ED1-positive cells both in spleens signal molecular during inflammation. The same upregulation of and synovial tissues, indicating that Pkm2 was upregulated in Mw Pkm2 is also found in LPS-stimulated Mw (26), along with other of arthritic rats. Interestingly, Pkm2 had two different expression glucose metabolism-related genes. Interestingly, Pkm2 protein bands in rat splenocytes during the PIA development. We thought expression is found upregulated in RA synovial tissue samples

Table II. shRNA oligonucleotide primers

Oligonucleotides Primers Sequence (59–39) Mus pkm shRNA F 59-CCGGATCATTGCCGTGACTCGAAATCTCGAGATTTCGAGTCACGGCAATGATTTTTTG-39 R59-AATTCAAAAAATCATTGCCGTGACTCGAAATCTCGAGATTTCGAGTCACGGCAATGAT-39 Mus pkm2 shRNA F 59-CCGGGCCATTATCGTGCTCACCAAGTCTCTCGAGAGACTTGGTGAGCACGATAATGGCTTTTTG-39 R59-AATTCAAAAAGCCATTATCGTGCTCACCAAGTCTCTCGAGAGACTTGGTGAGCACGATAATGGC-39 Mus stat1 shRNA F 59-CCGGGCTGTTACTTTCCCAGATATTCTCGAGAATATCTGGGAAAGTAACAGCTTTTTG-39 R59-AATTCAAAAAGCTGTTACTTTCCCAGATATTCTCGAGAATATCTGGGAAAGTAACAGC-39 Mus stat1 shRNA F 59-CCGGGCCGAGAACATACCAGAGAATCTCGAGATTCTCTGGTATGTTCTCGGCTTTTTG-39 R59-AATTCAAAAAGCCGAGAACATACCAGAGAATCTCGAGATTCTCTGGTATGTTCTCGGC-39 Mus stat1 shRNA F 59-CCGGCCGAAGAACTTCACTCTCTTACTCGAGTAAGAGAGTGAAGTTCTTCGGTTTTTG-39 R59-AATTCAAAAACCGAAGAACTTCACTCTCTTACTCGAGTAAGAGAGTGAAGTTCTTCGG-39 Rat Pkm shRNA1 F 59-CCGGAAAGCTTTGCATCTGATCCCACTCGAGTGGGATCAGATGCAAAGCTTTTTTTTG-39 R59-AATTCAAAAAAAAGCTTTGCATCTGATCCCACTCGAGTGGGATCAGATGCAAAGCTTT-39 Rat Pkm shRNA2 F 59-CCGGAAGGCCTCTTATAAATGTTTACTCGAGTAAACATTTATAAGAGGCCTTTTTTTG-39 R59-AATTCAAAAAAAGGCCTCTTATAAATGTTTACTCGAGTAAACATTTATAAGAGGCCTT-39 Rat Pkm shRNA3 F 59-CCGGGCTTTGATAGTTCTGACGGAGCTCGAGCTCCGTCAGAACTATCAAAGCTTTTTG-39 R59-AATTCAAAAAGCTTTGATAGTTCTGACGGAGCTCGAGCTCCGTCAGAACTATCAAAGC-39 shNC F 59-CCGGAATTCTCCGAACGTGTCACGTCTCGAG ACGTGACACGTTCGGAGAATTTTTTTG-39 R59-AATTCAAAAAAATTCTCCGAACGTGTCACGTCTCGAGACGTGACACGTTCGGAGAATT-39 F, forward; R, reverse. The Journal of Immunology 9 Downloaded from

FIGURE 6. Pkm2 promotes Mw activation via Stat signal activation. LPS and IFN-g were used in NC and PKM knocked-down RAW264.7 cells for http://www.jimmunol.org/ different time points. The expressions of p-Stat1, p-Stat3, and total Stat1 and Stat3 (A and B) were detected by Western blotting. siRNA of Pkm2 or NC was transfected to rat BMMs for 24 h, later all the cells were treated with LPS and IFN-g for 4 h, and then the protein was collected (n = 3). The expression of p-Stat1 and total Stat1 (C and D) was detected by Western blotting. The expression of p-Stat1 was normalized by its own total protein (n = 3). The re- combination plasmids pcDNA-Pkm2-Clover, pcDNA-Stat1-mRuby2, or pcDNA-Stat3-mRuby2 were transfected in 293T cells for 48 h, and a confocal microscope was used to capture the protein expression. The pictures were merged automatically by the confocal microscope (E). LPS and IFN-g were used in RAW264.7 cells for 4 h, the protein was collected for immunoprecipitation assay (F and G), and rabbit IgG was used as an NC (n = 3). Data are presented as mean 6 SEM. *p , 0.05.

compared with osteoarthritis patients’ samples (27). We also shPKM2 treatment rats showed very limited arthritis symptoms by guest on September 29, 2021 found that Pkm2 was also expressed in RA patients’ synovial compared with the NC group. The same effect was found in ankle tissues (data not shown). More Pkm2 overexpression evidence was swelling change and also the histological score of synovitis and found in RA synovial tissues (28). In addition, the Pkm2 ex- bone cartilage damage. These macroscopic arthritis indexes were pression levels are also increased in RA FLS via TLR2 signaling also significantly improved in Pkm2 knocking down rats, sug- (10), suggesting Pkm2 might also be involved in FLS proliferation gesting a predominant role of Pkm2 in arthritis pathological and activation. Inhibiting the glycolysis pathway could also be process. In addition, the number of ED1-positive cells also de- a target in FLS for RA therapy (29). Other evidence in RA creased after the treatment, indicating that the overexpressed patient Mw points out one of the glycogen storage protein gly- Pkm2 might also involve Mw recruitment. cogen synthase kinase 3b (GSK3b) is deactivated, which suggests From our previous study, the classical activated Mw play an that the glucose metabolism in RA Mw might focus on glycolysis important role in arthritis animal model (17), and we also proved instead of TCA or glycogen storage (30). that the overexpressed Pkm2 was derived from ED1-positive cells, Ablation of Pkm2 by injecting RNAi plasmids or inhibiting the which leads us to investigate the function of Pkm2 in Mw. The Pkm2 enzyme function in PIA rats could significantly ameliorate results showed that Pkm2 rather than Pkm1 could be upregulated the clinical signs of arthritis. First, we used a commonly used Pkm2 not only in the LPS and IFN-g treatment but also be found in inhibitor (31), shikonin, to investigate the Pkm2 enzyme function pristane- or poly(I:C)–primed Mw. Pkm2 expression pattern in in arthritic rats. We found that the macroscopic arthritis score classical activated Mw showed the same trend with the expression could be decreased from day 16 to 19 by multiple injections of of Il-1b and iNos. This result suggested Pkm2 might participate in shikonin. A similar conclusion was also found in CIA mice, and PIA development via regulating Mw activation. To further un- shikonin treatment could upregulate IL-10 and TGF-b expression, derstand the function of Pkm2 in Mw, we used lentivirus to silence inhibit IL-17a, and suppress arthritis (32). However, from our the expression of Pkm2. In the Pkm2 knocked-down Mw, the results, the expression of Pkm2 could be upregulated from day 6, mRNA expression of tnf-a and il-1b decreased, and there was no which was 1 d before the onset of arthritis symptoms. Meanwhile, IL-1b protein expression detectable. Other studies performed in in the inhibitor-treated arthritis experiment, the suppressive effect cancer cells and Mw observe a similar effect. It is also reported of arthritis macroscopic arthritis score only distinguished from day that Pkm2 is a coactivator of Hif-1a (33) and participates in the 16 to 19. All these data suggested that Pkm2 might have other inflammasome activation promoting Il-1b production in Mw functions in regulating the arthritis process. To further investigate (26, 34). Pkm2 could form dimer or tetramer in the cell (35); the Pkm2 function, RNAi of Pkm2 was used. From the results, we another question was whether the tetramer or the dimer of Pkm2 could observe a significant effect on the macroscopic arthritis was involved in Mw activation. To solve this question, two Pkm2 score of arthritis from day 10, which was very close to the onset activators, DASA-58 and TEPP-46, were used. After applying the day. After that day, the macroscopic arthritis score of all the activators of Pkm2, the protein expression of Il-1b and iNOS was 10 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 7. Intervention of Pkm2 could decrease the p-Stat1 in synovial tissue of PIA rats, and knocked-down Stat1 could reduce proinflammatory cytokine expression in Mw DA rats that were treated with plasmids i.p. The ankles of the joints were collected for histological staining. Representative histological images of ankle joints stained by ED1 (green) and p-Stat1 (red) of each group were shown (A). The stable Stat1 shRNA and NC RAW264.7 cells were selected by 3 mg/ml puromycin and stimulated with LPS and IFN-g for 6 h. The IL-1b protein expression was detected by Western blotting (B and C)(n = 3). The proinflammatory cytokine Tnf-a and Il-1b mRNA expression also was detected in the same cells (D and E)(n = 3). Data are presented as mean 6 SEM. *p , 0.05. downregulated significantly, suggesting that the tetramer of Pkm2 of Tnf-a and Il-1b significantly decreased in the spleen of Pkm2 could also block the Mw activation. The same results were also shRNA plasmid treatment group, compared with the NC group. found in LPS-treated Mw (26). Because the mRNA expression of To confirm the Pkm2-related Stat1 function in Mw, the RNAi of downstream effective genes could be regulated after Pkm2 was Stat1wasalsoperformedinMw. After knocking down Stat1, the knocked down, we checked the Mw activation signaling and mRNA expression of Il-1b was also found to be significantly found that only Stat1 and Stat3 phosphorylations were influ- decreased, the same as the expression pattern in Pkm2 knocked- enced. In coronary artery disease, PKM2 regulates Mw IL-6 down Mw, suggesting that Pkm2 might regulate the proin- expression also via STAT3 phosphorylation (24). In cancer flammatory cytokine expression via phosphorylation of Stat1. cells, PKM2 could directly regulate STAT3 phosphorylation and Meanwhile, the mRNA expression of Il-17a from Pkm shRNA participate in the cell migration (36). This could also explain plasmid treatment group showed to be partially decreased. We why knocking down the expression of Pkm in PIA rats caused thought Il-17a expression might be indirectly regulated by Stat1 the declined ED1-positive cell. Confocal studies revealed not or Stat3. The same results were also found in colorectal cancer only the Pkm2 and Stat3, but also, the Pkm2 and Stat1 had a very cells [depletion of STAT3 decreased PKM2-induced Tnf-a and good colocalization, suggesting that Pkm2 might also be in- IL-1b expression (25)], also indicating that STAT3 mediated the volved in Stat1 and Stat3 phosphorylation. The immunoprecip- proinflammatory effects of PKM2. Thus, when the expression of itationassayinactivatedMw also suggested that the Pkm2 Pkm2 was knocked down in PIA rats, the phosphorylation of Stat1 correlated with Stat1 and Stat3. Our previous study showed that and Stat3 decreased, and downstream of Stat1- and Stat3-regulated Stat1 could regulate Irf1 and Tlr3 expression in PIA rats (9). In genes, especially the proinflammatory cytokines, could also be al- addition, the Ncf1 deficient mice could also spontaneously de- tered, suggesting that the expression of Pkm2 may play a crucial velop arthritis because of the aberrant Stat1 activation (37, 38). role in regulating the arthritis development. Further experiments are The related proinflammatory cytokines were also checked in needed to study the special effect of Pkm2 in regulating immune arthritic rats’ spleens, and we found that the mRNA expression signal response and the function in other cell types in arthritis. The Journal of Immunology 11 Downloaded from

FIGURE 8. Intervention of Pkm2 could reduce proinﬂammatory cytokine expression in PIA rat spleens. DA rats were treated with plasmids i.p. The mRNA (A) and protein (B) expression of Pkm2 in ShPkm/NC plasmid-treated http://www.jimmunol.org/ PIA rat spleen (n = 7 or 8). The proin- ﬂammatory cytokine Tnf-a (C)andIl-1b (D) mRNA expressions were detected in the same spleen samples. The correlation between Pkm2 and Tnf-a (E)orbetween Pkm2 and Il-1b (F) mRNA expression was analyzed. The mRNA expressions of Il-17a (G)andTlr3 (H) were detected in the same spleen samples (n = 7 or 8).

Data are presented as mean 6 SEM. by guest on September 29, 2021 *p , 0.05, **p , 0.01, ***p , 0.001.

In this study, PKM2 expression is increased in both evidence that PKM2 could have potential as a novel target for con- spleen and synovial Mw from rats with arthritis. Specifically trolling RA. induced PKM2 can activate Stat1 signaling to upregulate proinflammatory cytokine production and joint destruction Acknowledgments in arthritis development. Intervening PKM2 can dramatically The mRuby2 plasmid was reconstructed from pcDNA3.1-mRuby2, a gift mitigate the severity of experimental arthritis. The findings from Michael Lin (Addgene plasmid no. 40260). The Clover gene was support the new insights into metabolic mechanisms of Mw activation cloned from pLenti-FoxO1-Clover, a gift from Peter Rotwein (Addgene in arthritis occurrence and development and provide experimental plasmid no. 67759). 12 UPREGULATED PKM2 IN Mw EXACERBATES ARTHRITIS BY STAT1 SIGNAL

Disclosures Improving FRET dynamic range with bright green and red fluorescent proteins. Nat. Methods 9: 1005–1012. The authors have no financial conflicts of interest. 20. Gross, S. M., and P. Rotwein. 2015. Akt signaling dynamics in individual cells. J. Cell Sci. 128: 2509–2519. 21. Xu, J., W. Zhu, Y. Guo, C. Jiang, N. Hussain, L. Meng, and S. Lu. 2017. Con- References struction of conveniently screening pLKO.1-TRC vector tagged with TurboGFP. 1. Kinne, R. W., R. Braüer, B. Stuhlmu¨ller, E. Palombo-Kinne, and G. R. Burmester. Appl. Biochem. Biotechnol. 181: 699–709. 2000. Macrophages in rheumatoid arthritis. Arthritis Res. 2: 189–202. 22. Yin, Y., S. C. Choi, Z. Xu, D. J. Perry, H. Seay, B. P. Croker, E. S. Sobel, 2. McGarry, T., and U. Fearon. 2019. Cell metabolism as a potentially targetable T. M. Brusko, and L. Morel. 2015. Normalization of CD4+ T cell metabolism pathway in RA. Nat. Rev. Rheumatol. 15: 70–72. reverses lupus. Sci. Transl. Med. 7: 274ra18. 3. Yang, Z., H. Fujii, S. V. Mohan, J. J. Goronzy, and C. M. Weyand. 2013. 23. Wang, Y., J. Liu, X. Jin, D. Zhang, D. Li, F. Hao, Y. Feng, S. Gu, F. Meng, Phosphofructokinase deficiency impairs ATP generation, autophagy, and redox M. Tian, et al. 2017. O-GlcNAcylation destabilizes the active tetrameric PKM2 balance in rheumatoid arthritis T cells. J. Exp. Med. 210: 2119–2134. to promote the Warburg effect. Proc. Natl. Acad. Sci. USA 114: 13732–13737. 4. Veras, F. P., R. S. Peres, A. L. Saraiva, L. G. Pinto, P. Louzada-Junior, 24. Shirai, T., R. R. Nazarewicz, B. B. Wallis, R. E. Yanes, R. Watanabe, T. M. Cunha, J. A. Paschoal, F. Q. Cunha, and J. C. Alves-Filho. 2015. Fructose M. Hilhorst, L. Tian, D. G. Harrison, J. C. Giacomini, et al. 2016. The glycolytic 1,6-bisphosphate, a high-energy intermediate of glycolysis, attenuates experi- enzyme PKM2 bridges metabolic and inflammatory dysfunction in coronary mental arthritis by activating anti-inflammatory adenosinergic pathway. Sci. Rep. artery disease. J. Exp. Med. 213: 337–354. 5: 15171. 25. Yang, P., Z. Li, H. Li, Y. Lu, H. Wu, and Z. Li. 2015. Pyruvate kinase M2 ac- 5. Okano, T., J. Saegusa, K. Nishimura, S. Takahashi, S. Sendo, Y. Ueda, and celerates pro-inflammatory cytokine secretion and cell proliferation induced by A. Morinobu. 2017. 3-bromopyruvate ameliorate autoimmune arthritis by modu- lipopolysaccharide in colorectal cancer. Cell. Signal. 27: 1525–1532. lating Th17/Treg cell differentiation and suppressing dendritic cell activation. Sci. 26. Palsson-McDermott, E. M., A. M. Curtis, G. Goel, M. A. Lauterbach, Rep. 7: 42412. F. J. Sheedy, L. E. Gleeson, M. W. van den Bosch, S. R. Quinn, R. Domingo- 6. Bustamante, M. F., P. G. Oliveira, R. Garcia-Carbonell, A. P. Croft, J. M. Smith, Fernandez, D. G. Johnston, et al. 2015. Pyruvate kinase M2 regulates Hif-1a R. L. Serrano, E. Sanchez-Lopez, X. Liu, T. Kisseleva, N. Hay, et al. 2018. activity and IL-1b induction and is a critical determinant of the warburg effect in

Hexokinase 2 as a novel selective metabolic target for rheumatoid arthritis. Ann. LPS-activated macrophages. [Published erratum appears in 2015 Cell Metab.21: Downloaded from Rheum. Dis. 77: 1636–1643. 347.] Cell Metab. 21: 65–80. 7. Shi, L. Z., R. Wang, G. Huang, P. Vogel, G. Neale, D. R. Green, and H. Chi. 27. Li, X. J., M. Xu, X. Q. Zhao, J. N. Zhao, F. F. Chen, W. Yu, D. Y. Gao, and 2011. HIF1alpha-dependent glycolytic pathway orchestrates a metabolic B. Luo. 2013. Proteomic analysis of synovial fibroblast-like synoviocytes from checkpoint for the differentiation of TH17 and Treg cells. J. Exp. Med. 208: rheumatoid arthritis. Clin. Exp. Rheumatol. 31: 552–558. 1367–1376. 28. Biniecka, M., M. Canavan, T. McGarry, W. Gao, J. McCormick, S. Cregan, 8. Abboud, G., S. C. Choi, N. Kanda, L. Zeumer-Spataro, D. C. Roopenian, and L. Gallagher, T. Smith, J. J. Phelan, J. Ryan, et al. 2016. Dysregulated bioen- L. Morel. 2018. Inhibition of glycolysis reduces disease severity in an autoim- ergetics: a key regulator of joint inflammation. Ann. Rheum. Dis. 75: 2192–2200. mune model of rheumatoid arthritis. Front. Immunol. 9: 1973. 29. de Oliveira, P. G., M. Farinon, E. Sanchez-Lopez, S. Miyamoto, and M. Guma. 9. Zhu, W., J. Xu, C. Jiang, B. Wang, M. Geng, X. Wu, N. Hussain, N. Gao, Y. Han, 2019. Fibroblast-like synoviocytes glucose metabolism as a therapeutic target in http://www.jimmunol.org/ D. Li, et al. 2017. Pristane induces autophagy in macrophages, promoting a rheumatoid arthritis. Front. Immunol. 10: 1743. STAT1-IRF1-TLR3 pathway and arthritis. Clin. Immunol. 175: 56–68. 30. Zeisbrich, M., R. E. Yanes, H. Zhang, R. Watanabe, Y. Li, L. Brosig, J. Hong, 10. McGarry, T., M. Biniecka, W. Gao, D. Cluxton, M. Canavan, S. Wade, S. Wade, B. B. Wallis, J. C. Giacomini, T. L. Assimes, et al. 2018. Hypermetabolic L. Gallagher, C. Orr, D. J. Veale, and U. Fearon. 2017. Resolution of TLR2-induced macrophages in rheumatoid arthritis and coronary artery disease due to glycogen inflammation through manipulation of metabolic pathways in Rheumatoid Arthritis. synthase kinase 3b inactivation. Ann. Rheum. Dis. 77: 1053–1062. Sci. Rep. 7: 43165. 31. Chen, J., J. Xie, Z. Jiang, B. Wang, Y. Wang, and X. Hu. 2011. Shikonin and its 11. Yamada, K., and T. Noguchi. 1999. Regulation of pyruvate kinase M gene ex- analogs inhibit cancer cell glycolysis by targeting tumor pyruvate kinase-M2. pression. Biochem. Biophys. Res. Commun. 256: 257–262. Oncogene 30: 4297–4306. 12. Yamada, K., T. Noguchi, T. Matsuda, M. Takenaka, P. Monaci, A. Nicosia, and 32. Dai, Q., J. Li, Y. Yun, and J. Wang. 2018. Toll-like receptor 4-myeloid differ- T. Tanaka. 1990. Identification and characterization of hepatocyte-specific reg- entiation primary response gene 88 pathway is involved in the Shikonin treat- ulatory regions of the rat pyruvate kinase L gene. The synergistic effect of

ment of CIA by regulating Treg/Th17 expression. [Published erratum appears in by guest on September 29, 2021 multiple elements. J. Biol. Chem. 265: 19885–19891. 2019 Evid. Based Complement. Alternat. Med. 2019: 2198708.] Evid. Based 13. Kilkenny, C., W. J. Browne, I. C. Cuthill, M. Emerson, and D. G. Altman. 2010. Complement. Alternat. Med. 2018: 2428546. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 8: e1000412. 33. Luo, W., H. Hu, R. Chang, J. Zhong, M. Knabel, R. O’Meally, R. N. Cole, 14. Vingsbo, C., P. Sahlstrand, J. G. Brun, R. Jonsson, T. Saxne, and R. Holmdahl. A. Pandey, and G. L. Semenza. 2011. Pyruvate kinase M2 is a PHD3-stimulated 1996. Pristane-induced arthritis in rats: a new model for rheumatoid arthritis with a coactivator for hypoxia-inducible factor 1. Cell 145: 732–744. chronic disease course influenced by both major histocompatibility complex and 34. Xie, M., Y. Yu, R. Kang, S. Zhu, L. Yang, L. Zeng, X. Sun, M. Yang, non-major histocompatibility complex genes. Am. J. Pathol. 149: 1675–1683. T. R. Billiar, H. Wang, et al. 2016. PKM2-dependent glycolysis promotes 15. Bakharevski, O., A. N. Stein-Oakley, N. M. Thomson, and P. F. Ryan. 1998. NLRP3 and AIM2 inflammasome activation. Nat. Commun. 7: 13280. Collagen induced arthritis in rats. Contrasting effect of subcutaneous versus 35. Anastasiou, D., Y. Yu, W. J. Israelsen, J. K. Jiang, M. B. Boxer, B. S. Hong, intradermal inoculation of type II collagen. J. Rheumatol. 25: 1945–1952. W. Tempel, S. Dimov, M. Shen, A. Jha, et al. 2012. Pyruvate kinase M2 acti- 16. Romas, E., O. Bakharevski, D. K. Hards, V. Kartsogiannis, J. M. Quinn, P. F. Ryan, vators promote tetramer formation and suppress tumorigenesis. Nat. Chem. Biol. T. J. Martin, and M. T. Gillespie. 2000. Expression of osteoclast differentiation 8: 839–847. factor at sites of bone erosion in collagen-induced arthritis. Arthritis Rheum. 43: 36. Yang, P., Z. Li, R. Fu, H. Wu, and Z. Li. 2014. Pyruvate kinase M2 facilitates 821–826. colon cancer cell migration via the modulation of STAT3 signalling. Cell. Signal. 17. Meng, L., W. Zhu, C. Jiang, X. He, W. Hou, F. Zheng, R. Holmdahl, and S. Lu. 26: 1853–1862. 2010. Toll-like receptor 3 upregulation in macrophages participates in the ini- 37. Kelkka, T., D. Kienho¨fer, M. Hoffmann, M. Linja, K. Wing, O. Sareila, tiation and maintenance of pristane-induced arthritis in rats. Arthritis Res. Ther. M. Hultqvist, E. Laajala, Z. Chen, J. Vasconcelos, et al. 2014. Reactive oxygen 12: R103. species deficiency induces autoimmunity with type 1 interferon signature. 18. Donaldson, J. G. 2015. Immunofluorescence staining. Curr. Protoc. Cell Biol. Antioxid. Redox Signal. 21: 2231–2245. 69: 4.3.1–4.3.7. 38. Sareila, O., C. Hagert, T. Kelkka, M. Linja, B. Xu, J. Kihlberg, and R. Holmdahl. 19. Lam, A. J., F. St-Pierre, Y. Gong, J. D. Marshall, P. J. Cranfill, M. A. Baird, 2017. Reactive oxygen species regulate both priming and established arthritis, M. R. McKeown, J. Wiedenmann, M. W. Davidson, M. J. Schnitzer, et al. 2012. but with different mechanisms. Antioxid. Redox Signal. 27: 1473–1490. Supplemental material Supplemental table1. Antibodies Antibody Source Lot. No. Application Dilution Company ED1 mouse ab31630 IF 1/200 Abcam IF 1/100 Pkm2 rabbit #4053 WB 1/1000 CST IF 1/100 p-Stat1 rabbit #7649 WB 1/1000 CST Stat1 rabbit #14994 WB 1/1000 CST Stat3 rabbit #12640 WB 1/1000 CST p-Stat3 rabbit #9145 WB 1/1000 CST p-P38 rabbit #4511 WB 1/1000 CST P38 rabbit #8690 WB 1/1000 CST ERK rabbit #4695 WB 1/1000 CST p-ERK rabbit #4370 WB 1/1000 CST p-JNK rabbit #9251 WB 1/1000 CST JNK rabbit #9255 WB 1/1000 CST IRF3 rabbit #4302 WB 1/1000 CST P-IRF3 rabbit #29047 WB 1/1000 CST iNOS rabbit BA0326 WB 1/200 BOSTER IL-1β (pro-IL1β) rabbit ab9722 WB 1/1000 Abcam β-Actin (ACTB) mouse E021020-03 WB 1/2000 Earthox FITC AffiniPure Goat Anti-Mouse IgG goat E031210-01 IF 1/400 Earthox Cy3 AffiniPure Goat Anti-Rabbit IgG goat E031620-01 IF 1/400 Earthox goat anti-rabbit IgG conjugated with HRP goat ab205718 WB 1/10000 abcam goat anti-mouse IgG conjugated with HRP goat ab205719 WB 1/10000 abcam goat anti-rabbit IgG light chain conjugated with HRP goat A25022 WB 1/5000 Abbikine

Supplemental table2. Other reagents Reagents Lot. No. Application Dilution Company DAPI D9542 IF 1/100000 Sigma-Aldrich RIPA solution P0013C WB Beyotime protease and phosphatase inhibitors 11836170001 WB/IP Roche BCA kit A53225 WB/IP Thermo Scientific Co-IP lysis solution P0013 IP Beyotime protein A/G beads B23201 IP Bimake Supersignal® West Pico kit 35061 WB/IP Thermo Scientific TRI Reagent™ Solution AM9738 RT-qPCR Thermo Scientific First Strand cDNA Synthesis Kit 4897030001 RT-qPCR Roche Fast Start Universal SYBR Green Master (ROX) 4913914001 RT-qPCR Roche pristane 1921-70-6 PIA Acros Organics IFA 263910 CIA BD Difco™ F-12K medium N3520 Cell culture Sigma-Aldrich RPMI1640 670089 Cell culture Thermo Scientific FBS E600001 Cell culture Sangong M-CSF 315-02 Cell culture Peprotech LPS L4391 Cell culture Sigma-Aldrich IFN-γ 315-05 Cell culture Peprotech DASA-58 HY-19330 Cell culture MCE TEPP-46 HY-18657 Cell culture MCE shikonin S8279 PIA Selleckchem T4 DNA ligase M1801 plasmid Promega GXL DNA polymerase #R050A plasmid Takara DH5α plasmid Tiangen Restriction enzymes plasmid NEB E.Z.N.A.™ Endo-free Plasmid Maxi Kit D6926 plasmid Omega Bio-tec TurboFect™ Transfection Reagent R0531 Cell culture Thermo Scientific puromycin A1113803 Cell culture Thermo Scientific siRNA Cell culture Oligobio Lipofectamine™ 2000 Transfection Reagent 11668019 Cell culture Thermo Scientific

Supplementary Figures S. Fig.1

Supplementary Fig.1 Positive correlations between Pkm2 mRNA expression and the length changes of PIA rats ankle and paw were evaluated (n=8). DA rats were treated with a single injection of pristane. The changes in PIA rat ankle (A) and paw (C) were recorded. The correlation between Pkm2 mRNA expression and ankle perimeter change was analyzed (B). The correlation between Pkm2 mRNA expression and paw perimeter change was analyzed (D). Intervention of Pkm2 could reduce the swelling in PIA rats. The representative pictures of the hind legs and paws from PIA rats with different treatment were displayed (E). The perimeter changes of mid-paw (F) were compared among groups (n=8). Data were presented as mean ± SEM. *p < 0.05.

S. Fig.2

Supplementary Fig.2 Pkm2 was overexpressed in pristane or polyI:C treated macrophages. The mRNA expression of Pkm2 was determined in 1mM pristane (A) or polyI:C (B) treated RAW264.7 cells for 8 hours. The inflammatory related signal pathway was examined in Pkm2 knocked down and NC RAW264.7 cells (C). LPS and IFN-γ were used in control (NC) and Pkm2 and Pkm knocked down RAW264.7 cells for different time points. The expression of p-Irf3, p-P38, p-Erk, p-Jnk, and total Irf3, P38, Erk, Jnk were detected by Western blotting. The Stat1 knock-down efficiency was determined in RAW264.7 cells. The stable Stat1 shRNA and shNC RAW264.7 cells were selected by 3μg/mL puromycin. And the Stat1 protein expression was detected by Western blotting (D). Data were presented as mean ± SEM. *p < 0.05. Isotype antibodies control staining for immunofluorescence on the rat’s ankle slides. E. Isotype IgG (Rb) used as p-Stat1 or Pkm2 antibody isotype control, was stained with ED1 antibody and DAPI. F-G. Isotype IgG1 (mouse) used as ED1 antibody isotype control, was stained with DAPI and p-Stat1(B) or Pkm2(C) antibody.