Extracellular pyruvate kinase M2 promotes osteoclastogenesis and is associated with radiographic progression in early rheumatoid arthritis

Extracellular PKM2 (exPKM2) levels have been reported to be increased in several cancers and inflammatory diseases, including rheumatoid arthritis (RA). This study aimed to investigate the association of circulating exPKM2 levels with radiographic progression in RA patients and the effect of exPKM2 on osteoclastogenesis. Plasma and synovial fluid exPKM2 levels were significantly elevated in RA patients. Plasma exPKM2 levels were correlated with RA disease activity and were an independent predictor for radiographic progression in RA patients with a disease duration of ≤ 12 months. CD14+ monocytes but not RA fibroblast-like synoviocytes secreted PKM2 upon stimulation with inflammatory mediators. Recombinant PKM2 (rPKM2) increased the formation of tartrate-resistant acid phosphatase (TRAP)-positive multinuclear cells and resorption pit in osteoclast precursors, dose-dependently, even in the absence of receptor activator of nuclear factor-kappa B ligand (RANKL). rPKM2 treatment upregulated the expression of dendrocyte-expressed seven transmembrane protein (DC-STAMP) and MMP-9 via the ERK pathway. Although rPKM2 did not directly bind to RAW264.7 cells, extracellular application of pyruvate, the end-product of PKM2, showed effects similar to those seen in rPKM2-induced osteoclastogenesis. These results suggest that exPKM2 is a potential regulator of RA-related joint damage and a novel biomarker for subsequent radiographic progression in patients with early-stage RA.

www.nature.com/scientificreports/ Pyruvate kinase (PK) is a key rate-limiting enzyme that converts phosphoenolpyruvate to pyruvate in the final step of glycolysis. Increased expression of PK accelerates glycolysis. Among the four PK isoforms (PKL, PKR, PKM1, and PKM2), PKM2 is abundant in proliferative cells such as stem cells, tumor cells, and stromal cells in chronic inflammatory tissues such as RA joints [11][12][13][14][15][16] . Interestingly, PKM2 can be released into the extracellular space; PKM2 levels are elevated in the blood or feces of patients with cancer and chronic inflammatory diseases 11,17,18 . Extracellular PKM2 (exPKM2) can induce the proliferation and migration of cancer cells and promote angiogenesis via integrin β1 [19][20][21] . However, the role of exPKM2 in RA pathogenesis has not been fully elucidated. In this study, we investigated the expression levels of PKM2 in the synovial tissue, synovial fluid (SF), and plasma of RA patients. Additionally, we evaluated the clinical implications of plasma exPKM2 levels and the effect of exPKM2 on osteoclastogenesis.

Methods
Study subjects. RA was diagnosed according to the 1987 American Rheumatism Association criteria 22 .
Synovial tissues were obtained from 15 patients (12 RA and 3 OA patients) undergoing joint-replacement surgery or synovectomy. RA-FLSs were isolated from the synovium of 7 RA patients. The synovial tissues of 5 RA and 3 OA patients were fixed with 4% buffered paraformaldehyde and embedded in paraffin. SF samples were collected from a different group of 25 RA and 5 OA patients.
We consecutively enrolled 139 patients with RA (age 54.3 ± 11.9 years, 120 females) and 47 sex-and agematched healthy donors (3:1 ratio; age 54. 4  Clinical and radiographic assessment. Clinical and laboratory data were collected at the time of blood sampling. RA disease activity was assessed according to the DAS28-ESR 23 . The patients were categorized based on the RA type as follows: active RA, DAS28-ESR > 3.2; remission, DAS28-ESR < 2.6; and early RA, disease duration ≤ 12 months. Radiographs of the hands and feet were taken at the time of blood sampling and after a mean of 26.8 ± 16.6 months. Radiographic damage was assessed by two blinded trained investigators (YJH and HWK) using the modified Sharp/van der Heijde score (mSHS) 24 . The interobserver intraclass correlation coefficient (ICC) was 0.979 (95% confidence interval 0.974-0.984), and the smallest detectable change in mSHS was 1.59 during the follow-up period. Erosive disease was defined as an mSHS erosion score ≥ 1 at baseline 25 . Radiographic progression was defined as ΔmSHS ≥ 1 unit/year, while erosive or narrowing disease progression was defined as ∆mSHS ≥ 1 unit/year in the corresponding subscore as previously described 26 . Cell lines and reagents. The cell lines and reagents used in this study are described in the Supplementary Methods (Additional File 2). Cryopreserved peripheral blood mononuclear cells (PBMCs) were thawed, and CD3 + , CD14 + , and CD19 + cells were isolated using a magnetic bead-based method (Miltenyi Biotec, Miltenyi Biotec, Auburn, CA, USA), according to the manufacturer's instructions.
Immunohistochemistry and double immunofluorescence (IF) staining. After the paraffinembedded synovial tissue was deparaffinized, immunohistochemistry analysis of intracellular PKM2 was performed following a previously reported protocol with minor modifications 27 . The sections were stained with anti-PKM2 antibodies (1:1600) for 3 h at room temperature. The number of PKM2-immunostained cells was manually counted among synovial lining and sublining stromal cells. At least 3 high-power fields without an area of lymphoid aggregates were randomly selected, and a minimum of 1000 cells were analyzed on each slide.
Osteoclastogenesis and osteoclast activity assay. RAW264.7 cells were seeded in 48-well plates (1 × 10 3 cells/well) and cultured in α-modified minimum essential medium (MEM) containing 10% FBS in the presence of recombinant PKM2 (rPKM2), pyruvate, or receptor activator of nuclear factor kappa-Β ligand (RANKL). Tartrate-resistant acid phosphatase (TRAP) staining was performed after 5 days, and TRAP-positive multinucleated cells were counted under a light microscope. Osteoclast activity was assessed by measuring the area of resorption pits using a calcium phosphate-coated 24-well plate as previously described 30  www.nature.com/scientificreports/ of the pit formation assay methods are provided in the Supplementary Methods (Additional File 2). CD14 + monocytes were seeded at 1.5 × 10 5 cells/well in complete MEM, and on the next day, RANKL or rPKM2 was added. The culture medium was replaced every 4 days and multinucleated TRAP-positive cells were counted after 21 days.
Reverse transcription-polymerase chain reaction (RT-PCR) and immunoblotting. The cellular expression levels of genes and proteins were evaluated by RT-PCR and immunoblotting, respectively, using specific primers and antibodies. The PCR primer pairs are shown in the Supplementary Methods (Additional File 2).
Measurement of circulating PKM2 and proinflammatory cytokine levels and pyruvate kinase activity. The exPKM2 levels were measured using commercial enzyme-linked immunosorbent assay (ELISA) kits, following the manufacturer's instructions. The serum levels of TNF-α, IL-6, and VEGF were analyzed with a Luminex 100 system (Luminex, Austin, TX, USA) using a magnetic bead-based immunoassay (R&D Systems, Minneapolis, MN, USA). To evaluate the enzymatic activity of rPKM2, a Pyruvate Assay Kit (Abcam, Cambridge, MA, USA) was used according to the manufacturer's instructions. All measurements were performed in duplicates.
Statistical analysis. Data are presented as the mean ± standard deviation or median [interquartile ranges].
Continuous variables were compared using the Mann-Whitney U test or Kruskal-Wallis test. Categorical variables were compared using the Chi-square test or Fisher's exact test as appropriate. Bivariate correlations were analyzed using Spearman's correlation coefficient. Binary logistic regression analysis (stepwise forward regression) was performed to identify independent variables associated with radiographic progression. P or corrected p values < 0.05 were considered statistically significant. All data were analyzed using SPSS Statistics for Windows version 25 (IBM Corp., Armonk, NY, USA).

Ethics approval and consent to participate. This study was approved by the Institutional Review
Board (IRB No. B-0905/075-013) and was performed according to the recommendations of the Declaration of Helsinki. Informed written consent was obtained from all participants.

Results
Increased expression of PKM2 in RA synovial tissues. Immunostaining analysis revealed that intracellular PKM2 was expressed in the synovium of RA and osteoarthritis (OA). PKM2-positive stromal cells were distributed in both the lining and sublining layers ( Fig. 1). Mononuclear inflammatory cells and vascular endothelial cells also stained positive for PKM2. As expected, the RA synovium had a significantly higher immunosignal intensity (Fig. 1a) and a higher fraction of PKM2-positive stromal cells than the OA synovium (mean 72.9 ± 6.3% vs. 47.5 ± 8.7%, p = 3.06 × 10 -6 ; Fig. 1b Supplementary Fig. 1a). exPKM2 levels were positively correlated with the numbers of white blood cells including polymorphs and monocytes/macrophages, in the SF samples ( Fig. 2b-d, all p < 0.001).
Phospho-ERK levels were upregulated upon treatment with rPKM2 for 3 days, and the increase was dose dependent (p = 0.029; Fig. 4c,d). However, the expression levels of nuclear factor of activated T cells 1 (NFATc1) and p38 mitogen-activated protein (MAP) kinase were not significantly affected by rPKM2 treatment for 5 days (Supplementary Fig. 2c). Treatment with rPKM2 significantly increased dendrocyte-expressed seven Extracellular pyruvate promotes osteoclastogenesis. Reportedly, exPKM2 can activate the epidermal growth factor receptor (EGFR) signaling pathway in cancer cells 19,20 , and EGFR ligands, including EGF and transforming growth factor-α, can enhance bone resorption 31 . However, flow cytometry analysis revealed that rPKM2 did not directly bind the RAW264.7 cell surface and that RAW264.7 cells did not express EGFR (data not shown). It was reported that the addition of 1-2 mM pyruvate significantly increased RANKL-induced osteoclastogenesis in RAW264.7 cells and murine bone marrow cells 32 . Because rPKM2 was able to actively convert phosphoenolpyruvate to pyruvate (Fig. 5a), we investigated whether pyruvate can also induce the expression of phospho-ERK and DC-STAMP even without RANKL, as exPKM2 could. When RAW264.7 cells were stimulated with 0.001 to 5 mM pyruvate, pyruvate significantly increased the levels of phospho-ERK after 72 h in a dose-dependent manner (p = 0.007, Fig. 5b). Additionally, pyruvate treatment upregulated the expression of DC-STAMP and MMP-9 transcripts in RAW264.7 cells (Fig. 5c,d). Conclusively, similar to exPKM2, 0.001 to 0.1 mM pyruvate significantly augmented the formation of TRAP-positive multinucleated cells (p = 0.0012 by Kruskal-Wallis test, Fig. 5e).

Discussion
In this study, PKM2 was upregulated in the synovial fluid, plasma and synovial tissues of RA patients. Additionally, exPKM2 was released mainly from activated macrophages and induced osteoclastogenesis via the ERK pathway in a dose-dependent manner, even in the absence of RANKL. Increased plasma PKM2 levels were associated with higher disease activity and radiographic progression, especially in early RA. Overall, exPKM2 reflects the inflammatory burden and actively contributes to RA-related joint damage. Cytosolic PKM2 is known to be upregulated during cellular growth and to be involved in metabolic reprogramming in various malignancies. Cytosolic PKM2 plays a pivotal role in glycolysis via reversible conversion between the tetramer and dimer forms. Tetrameric PKM2 catalyzes PEP to pyruvate with the production of ATP, whereas dimeric PKM2 has low catalytic activity and increases the production of glycolytic intermediates  (PMN, c), or macrophages/monocytes (d). ρ coefficients were calculated by the Spearman method. (e) exPKM2 levels were significantly increased when CD14 + peripheral blood monocytes were stimulated with 10 ng/mL of TNF-α or 50 ng/mL of IL-6 for 24 h (n = 4; † p = 0.029 by the Mann-Whitney U test), but not RA-FLSs (n = 3). LPS (100 ng/mL) also tended to increase exPKM2 levels from CD14 + monocytes in the media. www.nature.com/scientificreports/ shunted into the pentose phosphate pathway and nucleotide synthesis. In addition, nonmetabolic functions of cytosolic PKM2 are involved in cancer cell growth/survival, stemness, metastasis, or angiogenesis 11 . Dimeric cytosolic PKM2 can be translocated to the nucleus. Nuclear PKM2 can regulate the transcription of genes targeted by hypoxia-inducible factor 1 (HIF-1) or nuclear EGFR 11 . Furthermore, nuclear PKM2 was reported to regulate the production of TNF-α, IL-1β, MMP-2, and MMP-9 in colon cancer cells and to activate inflammasomes 29,33 . Based on these findings, the function or regulation of intracellular PKM2 has been widely researched and has been an attractive target for cancer therapy. Interestingly, it has been reported that PKM2 is released from cancer cells. This exPKM2 facilitates tumor angiogenesis, cancer cell migration via the phosphoinositide-3-kinase/protein kinase B (PI3K/Akt) and Wnt/βcatenin pathways, and proliferation via EGFR activation 20,34 . Indeed, exPKM2 was detected in serum or stool samples from cancer patients 19 . Moreover, exPKM2 levels were reported to be elevated in inflammatory disorders such as Crohn's disease or RA 12,14,16,17 . The present study also revealed that exPKM2 levels were significantly elevated in the joint fluid and blood of RA patients.
In the study cohort, plasma exPKM2 levels were significantly increased and correlated with disease activity indices in RA patients. Strikingly, the baseline exPKM2 levels were associated with radiographic progression in the early RA patients (Table 1) but not in the total RA patients. The significant association of exPKM2 with radiographic progression only in early RA patients may be related to a ceiling effect in radiographic progression 35 or result from therapeutic intervention. The early RA subgroup included significantly more patients not taking antirheumatic drugs than the nonearly RA subgroup at baseline; antirheumatic drugs could negatively affect RA disease activity and subsequent radiographic progression. Regardless of the causes behind the above findings, the association between exPKM2 levels and radiographic progression in the present study suggests that exPKM2 is (e) Among early RA patients (disease duration < 12 months; n = 54), plasma exPKM2 levels were significantly increased in those with radiographic progression (RP), erosive disease progression (EDP), or narrowing disease progression (NDP). Data were plotted as box-and-whisker plots. **p = 0.017; † † p = 0.0017; § § p = 0.024 by the Mann-Whitney U test.   Supplementary Fig. 3a,b. (e,f) When RAW264.7 cells were treated with RANKL or rPKM2, DC-STAMP and MMP-9 mRNA levels were significantly increased ( † p = 4.11 × 10 -5 ; n = 3). Data with error bars were expressed as mean ± SD. www.nature.com/scientificreports/ a novel biomarker predicting subsequent radiographic progression in early RA patients and high exPKM2 levels underlines the need to control RA activity aggressively and tightly. Initially, we expected that exPKM2 might be released from activated RA-FLSs, which exhibit the characteristics of cancer cells, including high metabolic activity, migration and invasion into surrounding tissue, since vimentin-expressing RA-FLSs showed the highest PKM2 expression (Fig. 1). However, monocytes/macrophages secreted more PKM2 than RA-FLSs or lymphocytes. The cellular source of exPKM2 could explain the significant correlation between exPKM2 and the number of monocytes/macrophages in SF. Moreover, since osteoclasts are derived from the monocyte/macrophage lineage and osteoclast generation is critical for joint destruction, we studied the autocrine or paracrine effect of exPKM2 on osteoclastogenesis. We observed for the first time that extracellular rPKM2 increased formation of TRAP-positive multinucleated cells from human CD14 + monocytes and RAW264.7 cells in a dose-dependent manner. Interestingly, the formation of TRAP-positive multinucleated cells was also enhanced upon treatment with a low dose of rPKM2 in the presence of suboptimal RANKL concentrations. However, rPKM2-induced osteoclasts exhibited lower bone resorptive activity than RANKLinduced osteoclasts, suggesting that exPKM2 alone cannot induce fully functional mature osteoclasts. In the present study, rPKM2 upregulated the ERK-dependent expression of DC-STAMP in osteoclast progenitors, and the ERK inhibitor U0126 inhibited rPKM2-induced osteoclastogenesis. DC-STAMP plays an essential role in the fusion of mononuclear osteoclasts, and it can induce the expression of TRAP 36 . The ERK pathway also mediates RANKL-and GM-CSF-induced fusion of osteoclast precursors and DC-STAMP expression 37,38 .
rPKM2 increases phospho-EGFR levels in breast cancer cells 20 , and EGFR increases osteoclast differentiation and survival through its interaction with RANKL 39 . However, we did not observe that nonactivated RAW264.7 cells expressed EGFR or that rPKM2 directly bound to their cell surface. Furthermore, the enhanced levels of phospho-ERK after 3 days of stimulation with rPKM2 indicated that exPKM2 could not act directly through a cell www.nature.com/scientificreports/ surface receptor. PKM2 catalyzes the final glycolytic step, where PEP is dephosphorylated into pyruvate. Pyruvate can be present in the synovial fluid; pyruvate was detected at concentrations from 0.13 to 0.28 mM in four RA patients in the study of Treuhaft et al. 1 . Additionally, a recent study showed that PEP can be detected in the circulation, and its levels are significantly higher in early RA patients with a therapeutic response to methotrexate 40 . Furthermore, extracellular pyruvate is reported to augment RANKL-induced osteoclastogenesis 32 . Based on the above findings, we investigated the effect of extracellular pyruvate on osteoclastogenesis. The present study finally revealed that the effect of extracellular pyruvate on osteoclastogenesis is in line with that of PKM2. Therefore, exPKM2 secreted from activated monocytes/macrophages can enhance RANKL-independent osteoclastogenesis via catalytic conversion of extracellular PEP to pyruvate. Although our study provides a novel role of exPKM2 in RA, this study has several limitations. First, plasma PKM2 levels as a prognostic biomarker should be confirmed in large studies. Second, the precise mechanism by which exPKM2 activates and alters intracellular signaling needs further investigation. Concerning the effect of extracellular pyruvate, the clinical implication of PEP or pyruvate levels also needs further study. Finally, therapeutic inhibition of exPKM2 needs to be tested in an in vivo model of inflammatory arthritis.

Conclusions
exPKM2 levels are increased in the circulation and synovial fluid of RA patients. They are positively associated with RA disease activity and are a significant independent risk factor for radiographic progression in early RA patients. exPMK2 can promote osteoclast differentiation in the early phase via extracellular pyruvate formation and the ERK signaling pathway. Therefore, exPMK2 could be a potential player in RA-related joint damage and a novel biomarker for subsequent radiographic progression in early RA patients.

Data availability
All data generated or analyzed during this study are included in this published article and supplementary information files.