Role of Chitinase 3-like-1 in Myelofibrosis via Fibroblast-Produced Extracellular Matrix Enhancement.

Introduction

Myelofibrosis (MF) is a crucial pathological feature of primary MF (PMF), post‐essential thrombocythemia (ET) MF, and post‐polycythemia vera (PV) MF. MF forms are associated with driver mutations, Janus kinase (JAK) 2 ( JAK2 ), calreticulin ( CALR ), and myeloproliferative leukemia ( MPL ) (thrombopoietin [TPO] receptor [TPO‐R]) (Tefferi et al. 2014 ). These driver mutations activate the JAK/signal transducers and activators of transcription signaling pathway, resulting in the growth and differentiation of a myeloproliferative neoplasm (MPN) clone. Simultaneously, TPO‐R signaling activation induces collagen production in myofibroblasts due to cytokine overproduction, such as transforming growth factor (TGF)‐β1 from platelets and megakaryocytes of MPN clones (Wagner‐Ballon et al. 2007 ). We previously revealed that TPO‐R signaling activation directly induces fibrocyte differentiation of MPN clones, causing MF in MPN (Maekawa et al. 2017 ).

Cluster of differentiation (CD) 14 + monocytes differentiate into spindle‐shaped fibroblast‐like blood cells, also known as fibrocytes, which express stromal (collagen I, procollagen, and collagen III) and hematopoietic (CD45, CD34, CD11b, and CD68) cell markers (Maekawa et al. 2017 ; Verstovsek et al. 2016 ), and are well known for their significant role in the pathologic fibrosis of various organs besides the bone marrow (BM). Recently, some studies revealed the essential role of fibrocytes in MF using animal models: murine xenograft model (Verstovsek et al. 2016 ), romiplostim (Rom, a TPO‐R agonist)‐induced murine model (Maekawa et al. 2017 ), and JAK2 V617F transgenic mouse model (Ozono et al. 2021 ). Fibrocytes generate some collagen, and murine JAK2 V617F fibrocytes expressed more TGF‐ β1 than normal fibrocytes (Ozono et al. 2021 ). However, the difference in TGF‐ β1 production between normal and JAK2 V617F human fibrocytes remains unknown. Further, the molecular contribution of fibrocytes to MF remains unclear.

Fibrocytes expressed more chitinase 3–like‐1 (CHI3L1) than monocytes, as the RNA sequencing‐based gene profiles indicated. The 18‐glycosyl hydrolase‐related molecule, CHI3L1, is a member of the enzymatically inactive chitinase‐like protein family and is known as BRP‐39 in mice and YKL‐40 in humans (Lee et al. 2011 ; Riabov et al. 2014 ). CHI3L1 is expressed in various cells, such as neutrophils, macrophages, chondrocytes, endothelial cells, and tumor cells (Lee et al. 2009 ; Bussink et al. 2007 ; Funkhouser and Aronson 2007 ), and plays crucial roles in organ immune response, inflammation, remodeling, and fibrosis (Lee et al. 2009 ; Recklies et al. 2002 ; Knudsen et al. 2006 ; Johansen 2006 ). Furthermore, higher serum CHI3L1 levels were more prominent in patients with PMF than in those with PV and ET; however, their pathological significance is unknown (Bjørn et al. 2014 ).

This study evaluated the association between CHI3L1 and human fibrocytes and measured serum CHI3L1 levels in patients with MPN. Further, the roles of CHI3L1 were confirmed using the Rom‐induced MF, CHI3L1 ‐deficient mice, and JAK2 V617F TG mouse models.

Materials and Methods

Human Subjects

This study complied with the National Defense Medical College Ethics Committee–approved protocols (Approval Numbers 2153 and 4419). This study enrolled 52 patients diagnosed with MPN between January 2017 and December 2018 and 80 patients with lymphoid tumors (plasma cell tumors or malignant lymphoma), either newly diagnosed or with recurrence, between August 2021 and August 2023. All diagnoses were established according to the 2016 WHO classification criteria (Arber et al. 2016 ), and written informed consent was obtained from all participants. In MPN patients, one pathologist and one hematologist independently evaluated the severity of BM fibrosis based on the European consensus criteria (Thiele et al. 2005 ). MF‐0 represented patients with MPN without MF, and MF‐1, ‐2, or ‐3 represented those with MPN with MF. Patients with lymphoid tumors were evaluated for MF grade, presence of splenomegaly, and proportion of fibrocyte precursor cells, and serum CHI3L1 levels were assessed (supporting Table 1 ). We obtained blood samples from all patients and 17 healthy donors to measure serum CHI3L1 levels in a cross‐sectional manner, regardless of treatment stage. Peripheral blood samples from four healthy donors and three patients with JAK2 V617F were used to culture human fibrocytes.

Human Fibrocyte Culture Assay

Density centrifugation from 10 mL of peripheral blood samples using Pancoll (PAN‐Biotech, Aidenbach, Germany) was used to purify peripheral blood mononuclear cells (PBMCs) from healthy donors. The PBMCs were cultured on 10‐cm culture dishes in 10 mL of Dulbecco's modified eagle medium (DMEM) (Wako 048‐33575, high glucose with sodium pyruvate and without additional amino acids; Wako, Osaka, Japan), with 2 mM of l ‐glutamine (Nacalai tesque, Kyoto, Japan), 20% fetal bovine serum (FBS) (Corning, Corning, NY), 100 U/mL of penicillin, and 100 μg/mL of streptomycin (Thermo Fisher Scientific, Waltham, MA). The dishes were incubated at 37°C in a humidified incubator containing 5% CO 2 . After incubation for 3 days, the supernatant was removed to deplete the floating cells in the dishes, and DMEM was added again. The culture medium was changed every 4 days.

RNA sequencing analysis of fibrocytes to investigate the changes in gene expression profiles associated with fibrocyte differentiation used three sample types: adherent cells of days 2, 9, and 12, identified as monocytes, fibrocytes (day 9), and fibrocytes (day 12), respectively.

Trypsin‐EDTA solution (0.25%; Wako, Osaka, Japan) was added on days 4, 7, and 10 after PBMCs began to emerge in culture. Cells were dislodged, and 2 × 10 5 cells/well were cultured for 48 h on a 6‐well plate (Becton, Dickinson and Company, Franklin Lakes, USA). They were recognized as fibrocytes (day 6), fibrocytes (day 9), and fibrocytes (day 12). A previous report (Pilling et al. 2009 ) revealed that an elongated spindle‐shaped adherent cell with an oval nucleus indicated differentiated fibrocytes. Each well was counted in five low‐power fields of view. Cell lysates and supernatants were collected to investigate the expression of CHI3L1 mRNA, CHI3L1 protein, and TGF‐β protein and to perform a culture assay with the human fibroblast cell line HS‐5 (species: Homo sapiens ; sex: male; tissue of origin: fibroblast; official cell line name: HS‐5, RRID: CVCL_3720; ATCC CRL‐3611, Manassas, VA, USA) (Roecklein and Torok‐Storb 1995 ).

HS‐5 cells were cultured in DMEM (ATCC 30‐2002, high glucose with sodium pyruvate, sodium bicarbonate, and amino acids) supplemented with 10% FBS, penicillin, and streptomycin. The culture medium was changed every 3 days.

RNA Sequencing Analysis of Human Monocytes and Fibrocytes

We used the RNeasy Mini Kit (QIAGEN, Hilden, Germany) to extract the total RNA from monocytes, fibrocytes (day 9), and fibrocytes (day 12), following the manufacturer's instructions. The Bioanalyzer 2100 RNA Nano 6000 Assay Kit (Agilent Technologies, CA, USA) was used to evaluate RNA integrity. We utilized 1% agarose gel to monitor RNA degradation and contamination. A Hiseq. 4000 (Illumina, San Diego, USA) generated 150‐bp paired‐end reads for total RNA sequencing. CLC Genomics Workbench 11.0 (QIAGEN) was used to statistically analyze RNAseq data, and the human hg19 reference genome and annotation data were retrieved from the Ensembl website ( http://asia.ensembl.org/info/data/ftp/index.html ). Each pairwise comparison between the three groups was analyzed for differential gene expression. A false discovery rate of < 0.05 and a log2‐fold change cut‐off of 2 were applied. The differentially expressed gene list was further filtered to include genes with absolute fragments per kilobase of exon per million mapped reads (FPKM) values of > 1 in each group to generate heatmaps. The map was drawn after extracting 25 genes that exhibited high variability.

Luminex Assay

Serum samples of patients with MPN were stored at −20°C until analysis. Human Luminex Assays (R&D Systems, Minneapolis, MN) were used to evaluate CHI3L1 serum levels, as previously described (Maekawa et al. 2019 ).

Mouse Model of MF Induced by a TPO‐R Agonist

Eight‐week‐old female C57BL/6 J mice (CLEA Japan, Tokyo, Japan) were classified into two groups: control ( n = 16) and Rom ( n = 16). Each group was further categorized into three: treated once ( n = 4), treated twice ( n = 8), and treated thrice ( n = 4) groups. The control group mice were administered saline, and the Rom group mice were subcutaneously injected with 1 mg/kg of Rom at the neck on days 1, 8, and 15. Peripheral blood was drawn by tail cutting or cardiac puncture. Femurs, spleens, and tibias were extracted on day 15 after euthanasia for the twice‐treated groups. As previously described (Maekawa et al. 2017 ), spleens were weighed, and the cells were isolated using a 40‐μL cell strainer (BD Biosciences, San Jose, USA). RPMI‐1640 (Sigma‐Aldrich, St. Louis, USA) was used to flush tibia‐derived BM cells. Spleen and BM cells were hemolyzed following the manufacturer's (BD Pharm Lyse, BD Bioscience) instructions. A standard protocol was used for the morphologic analysis of mouse femur‐derived BM samples stained with hematoxylin/eosin and silver, as previously described (Maekawa et al. 2017 ). Furthermore, mouse BM samples were immunostained with an anti‐α‐smooth muscle antigen (SMA) antibody (DAKO, Tokyo, Japan), as previously described (Maekawa et al. 2017 ).

Eight‐week‐old female C57BL/6 J mice were then classified into three groups: control ( n = 6), Rom ( n = 8), and Rom + clodronate liposomes (CLs) (Formumax, Sunnyvale, CA, USA) ( n = 6). The mice in the Rom and Rom + CLs groups were administered 1 mg/kg of Rom, and those in the control group were administered saline on days 1 and 8. The Rom + CLs group mice received 200 μL of CLs intraperitoneally on day 1 and 100 μL on days 3, 7, 11, and 14. Similarly, the control and Rom groups were administered control liposomes (Formumax). Tibial‐ and femoral‐derived BMs were extracted from all three groups after euthanasia on day 15 and treated as previously described (Maekawa et al. 2017 ).

CHI3L1 Knockout Mice Preparation

CHI3L1 −/− mice (C57BL/6 background) were previously prepared in our laboratory (Higashiyama et al. 2019 ). Rom of 1 mg/kg was administered to 8‐week‐old wild‐type female C57BL/6 ( n = 19) and CHI3L1 −/− ( n = 21) mice on days 1 and 8. Spleens, tibias, and femurs were collected after euthanasia on day 15 and treated as described above. Additionally, 1 mg/kg of Rom was administered to 8‐week‐old wild‐type female C57BL/6 ( n = 4) and CHI3L1 −/− ( n = 6) mice on days 1, 8, and 15. The spleens, tibias, and femurs were collected and treated similarly after euthanasia on day 22.

Preparation of JAK2 V617F Transgenic (TG) Mice

Dr. Shimoda (Miyazaki University, Miyazaki, Japan) kindly provided JAK2 V617F TG mice (Ozono et al. 2021 ). We created JAK2 V617F/ CHI3L1 −/− mice by crossing JAK2 V617F TG mice with CHI3L1 −/− mice. Spleens, tibias, and femurs of 12‐week‐old ( n = 5) JAK2 V617F TG mice and JAK2 V617F/ CHI3L1 −/− mice ( n = 7) were collected after euthanasia on day 15 and treated as described above. We used HALO AI (Indica Labs, Corrales, NM, USA) (Kawamura et al. Forthcoming ) to develop deep learning–based fibrosis quantification to assess fibrosis in these mouse groups.

Addition of Supernatant Obtained from Cultured Fibrocytes and Neutralizing Antibody to HS‐5

HS‐5 cells were incubated at 2 × 10 5 cells/well for 24 h in 6‐well plates. The fibrocyte supernatant (days 6, 9, and 12) was added to each plate as a culture medium. We evaluated the effect of anti‐CHI3L1 antibody (10 μg/mL, Merck Millipore, Burlington, MA, USA) (Faibish et al. 2011 ) on HS‐5 cell culture with day 12 fibrocyte supernatant. Mouse monoclonal IgG1κ was used as a negative control. The cell lysate was collected to analyze collagen mRNA expression after 48 h.

Enzyme‐Linked Immunosorbent Assay Analysis of Supernatant From Human Fibrocytes and Mouse Serum Samples

The human fibrocyte supernatant was evaluated using human CHI3L1 and TGF‐β Quantikine ELISA Kit (R&D systems), and mouse serum samples were measured using mouse CHI3L1 Quantikine ELISA Kit (R&D systems), following the manufacturer's instructions.

Analysis of mRNA Expression in Mouse and Human Samples

Total cellular RNA from human fibrocytes, mouse spleen, mouse BM, and HS‐5 cells were obtained using RNeasy Mini Kit (QIAGEN) following the manufacturer's instructions. The SuperScript TM III First‐Strand Synthesis SuperMix for qRT‐polymerase chain reaction (PCR) (Invitrogen, Carlsbad, CA) was used. The Light Cycler 480 SYBR Green I Master (Roche, Basel, Switzerland) was used to perform quantitative real‐time PCR, as described by the manufacturer. Supporting Table 2 lists the primer sequences used in this study. The mRNA values were analyzed using comparative CT (2 −ΔΔCT ) and normalized with the mRNA levels of encoding glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH).

Statistical Analysis

Table 1 shows the variables obtained from human data. The JMP Pro 12 software (SAS Institute, Cary, NC) was used for all analyses. We used Fisher's exact and Mann–Whitney U tests to compare the characteristics of patients with MPN MF(−) and MPN MF(+).

The remaining human and mouse data are expressed as means ± standard error of the mean. One‐way analysis of variance with post hoc Tukey's test was used to analyze data from more than two groups, while the Mann–Whitney U test was utilized to compare data between two groups. The JMP Pro 12 software (SAS Institute) and GraphPad Prism 10 (GraphPad Software, La Jolla, CA) were used for data analysis and plotting. All probability (P) values were two‐sided, and statistical significance was taken at p < 0.05.

Results

Human Fibrocyte Differentiation Correlates With an Increase in CHI3L1 Expression

RNA sequencing analysis identified 13 genes whose expression was significantly higher in fibrocytes than in monocytes. The CHI3L1 gene demonstrated an increased expression associated with fibrocyte differentiation (Figure 1A ). In contrast, fibrocyte differentiation decreased TGF‐β1 expression. The frequency of spindle‐shaped fibrocytes increased as culture days progressed (on days 6, 9, and 12, p < 0.01) (Figure 1B ), and CHI3L1 mRNA expression and CHI3L1 protein levels in the supernatant increased with fibrocyte differentiation ( p < 0.01) (Figure 1C,D ). CHI3L1 and TGF‐β1 levels produced by fibrocytes collected from three PV patients positive for JAK2 V617F (the JAK2 allele burden in the PB samples was 90.8%, 31.8%, and 94.2%, respectively) were comparable to those from four healthy donors (Figure 1E,F ).

Elevated CHI3L1 Serum Levels Are Associated With the Degree of MF

Table 1 shows patient characteristics and the results of data analysis. Patients with MPN were categorized into two groups: those without MF ( n = 21) and those with MF ( n = 29). The primary diseases were PV ( n = 19), ET ( n = 23), and PMF ( n = 8). The types of genetic mutations were JAK2 ( n = 28), CALR ( n = 8), MPL ( n = 1), and triple‐negative ( n = 13). Hydroxyurea was administered to 18 patients and anagrelide to 1 patient during sample collection. Eleven patients demonstrated palpable splenomegaly.

The median serum CHI3L1 levels in healthy donors, patients with MPN without MF, and patients with MPN with MF were 28.2, 29.0, and 57.3 ng/mL, respectively (Figure 2A ). Patients with MPN with MF demonstrated a higher incidence of splenomegaly ( p < 0.001) and higher serum LDH ( p = 0.013) and CHI3L1 ( p = 0.003) levels than patients with MPN without MF. CHI3L1 serum levels increased in association with MF grade (Figure 2B ). Subgroup analysis by MPN type revealed significantly higher serum CHI3L1 levels in patients with PMF than in those with PV without MF and ET. In contrast, no apparent differences were observed between other MPNs with and without MF (Figure 2C ). We created receiver operating characteristic (ROC) curves to determine the cut‐off serum CHI3L1 value for distinguishing patients with MPN with MF from those without MF. The resulting area under the curve was 0.742, and the sensitivity for identifying MF and specificity were 62.1% and 81.0%, respectively, when the cut‐off was set at a serum CHI3L1 concentration of > 35.0 ng/mL (Figure 2D ). In patients with lymphoid tumors, serum CHI3L1 levels significantly were higher in the presence of myelofibrosis (MF‐1 or higher), splenomegaly, and an increased proportion of fibrocyte precursor cells (defined > 8.4% as high) (Supporting Figure 1 ).

Elevated CHI3L1 mRNA Expression in BM of the Rom‐Induced Murine MF Model Abrogated via Clodronate Treatment

We evaluated CHI3L1 serum and mRNA expression levels in the spleen and BM using the Rom‐induced murine model of MF to confirm the dynamics of MF (Figure 3A ). Serum CHI3L1 levels in mice treated with Rom were more than twice as high as those in the control group ( p < 0.05) (Figure 3B ). CHI3L1 mRNA levels in the spleen showed no apparent difference between the control and Rom groups (Figure 3C ). In contrast, the BM of the Rom group demonstrated higher CHI3L1 mRNA expression levels than that of the control group ( p < 0.01) (Figure 3D ). As previously reported (Maekawa et al. 2017 ), CL administration eliminated monocytes, macrophages, and fibrocytes to alleviate BM fibrosis in this model. We confirmed the effect of CLs on MF and revealed that clodronate treatment abrogated the elevation of CHI3L1 mRNA expression in the BM of the Rom group ( p < 0.01) (Figure 3D ).

CHI3L1 −/− Mice Demonstrated Decreased Col3a1 mRNA Expression and Fewer Reticular Fibers in the BM

We studied the functional role of CHI3L1 in MF development using Rom‐treated wild‐type and CHI3L1 −/− mice. Silver and immunohistochemical staining with α‐SMA of the femur BM sections revealed that CHI3L1 −/− mice presented significantly lower MF grade and positive staining area compared with the wild‐type mice on day 15 (Figure 4A , Table 2 ). However, hemoglobin, platelet counts, or spleen weight demonstrated no significant differences (Supplementary Figure 2 ). Gene expression analysis of fibrosis‐related genes in the BM demonstrated significantly decreased Col3a1 and Acta2 mRNA in CHI3L1 −/− mice ( p < 0.05) compared to the wild‐type mice, with no apparent difference in Col1a1 and Fn1 mRNA expression (Figure 4B–E ). CHI3L1 −/− and wild‐type mice demonstrated similar Il‐13rα2 and Ptgdr2 mRNA expression, which are CHI3L1 receptors (Figure 4F,G ).

JAK2 V617F/ CHI3L1 −/− Mice Had a Smaller Area of Reticular Fibers Than JAK2 V617F TG Mice

We induced MF using JAK2 V617F transgenic mice to better understand the function of CHI3L1. Silver staining revealed a diffuse and dense increase in reticular fibers with extensive intersections in both JAK2 V617F TG and JAK2 V617F/ CHI3L1 −/− mice. However, silver staining revealed fewer collagen bundles, indicating an MF‐2 grade in both groups. The European consensus criteria‐based evaluation method demonstrated no obvious differences (data not shown). We used a novel deep learning–based method to quantify fibrosis (manuscript submitted), providing a more quantitative evaluation of reticular fibers between the two groups (Figure 5A ). The mean reticular fiber area was significantly smaller in the JAK2 V617F/ CHI3L1 −/− mice than in the JAK2 V617F TG mice ( p < 0.001) (Figure 5B ), while spleen weight between JAK2 V617F TG mice and JAK2 V617F/ CHI3L1 −/− mice showed no significant differences (Figure 5C ).

CHI3L1 Is Involved in the Interaction Between Fibrocytes and Human Fibroblast Cell Line HS‐5, as Evidenced by Coculture Assays

We used the human fibroblast cell line HS‐5 to assay human fibrocytes and fibroblasts by noncontact coculture to assess how CHI3L1 interacts with fibroblasts. Based on the finding that fibrocytes secrete increased CHI3L1 concentrations with culture time, we prepared three types of supernatants (days 6, 9, and 12) as a culture medium for HS‐5. Culturing HS‐5 cells with fibrocyte supernatant significantly increased the mRNA expression of COL1A1 and COL3A1 ( p < 0.05), indicating an association with fibrocyte differentiation. The CHI3L1‐neutralizing antibody abrogated these increases, indicating that CHI3L1 secreted from fibrocytes is crucial for fibroblast extracellular matrix production (Figure 6 ).

Discussion

Some studies have indicated the involvement of CHI3L1 in lung, liver, and kidney fibrosis. High serum CHI3L1 levels were associated with disease progression in patients with idiopathic pulmonary fibrosis, reflecting an attempt to diminish injury and repair and contribute to lung fibrosis (Zhou et al. 2014 ). Montgomery et al. ( 2017 ) revealed that CHI3L1 activates myofibroblasts, thereby promoting renal fibrosis in a mouse model of ischemia‐reperfusion injury. Urine CHI3L1 level was associated with a more significant estimated glomerular filtration rate decrease and progressive renal fibrosis (Puthumana et al. 2021 ). CHI3L1 is a known biomarker of the degree of liver fibrosis and extracellular matrix synthesis (Johansen et al. 2000 ) and is a profibrogenic factor that is overexpressed in the aging liver and patients with liver fibrosis (Nishimura et al. 2021 ). Liver macrophages produced CHI3L1, which inhibited their apoptosis in a mouse liver fibrosis model (Higashiyama et al. 2019 ). Meanwhile, CHI3L1 has recently been implicated in hematologic disorders, with elevated CHI3L1 expression associated with lymphoma onset, progression, severity, and poor prognosis (Wang et al. 2025 ). Furthermore, CHI3L1 concentration in the serum of multiple myeloma patients was closely correlated with disease severity and high concentration CHI3L1 predicted the progression of multiple myeloma disease (Tang et al. 2025 ).

Our results corroborated the role of CHI3L1 as a profibrotic factor in BM fibrosis. CHI3L1 promotes BM fibrosis as a malignant process and induces tissue fibrosis as part of the repair after injury in other organ fibrosis. High serum concentrations of CHI3L1 have already been reported in patients with PMF, although the cause and role in pathophysiology remain unknown (Bjørn et al. 2014 ). Our analysis revealed an association between serum CHI3L1 concentration and BM fibrosis. The study of Bjørn et al. ( 2014 ) may have not contained sufficient patients with PV and ET with MF. Based on these results, our mouse experiments revealed an increased CHI3L1 mRNA expression in the BM with the frequency of romiplostim administration and the degree of fibrosis in the Rom‐induced murine MF model. Further, CLs with fibrocyte removal suppressed CHI3L1 in BM and MF (Maekawa et al. 2017 ). Furthermore, we revealed that the presence of CHI3L1 contributes to fibrosis exacerbation in Rom‐induced murine MF models and JAK2 V617F TG mice. The association between CHI3L1 and BM fibrosis was demonstrated in mice and humans with great significance.

Conversely, there is scope for further research on the CHI3L1‐secreting cells and the mechanisms of physiological interaction. CHI3L1 is produced by various cells, including macrophages, neutrophils, chondrocytes, vascular smooth muscle cells, and hepatic stellate cells. Macrophages demonstrated a strong expression of CHI3L1, recognized as a valuable marker of human macrophage differentiation in tissue, from immunohistochemical‐staining observations in organ fibrosis (Rehli et al. 1997 ; Rehli et al. 2003 ). Therefore, organ fibrosis associated with CHI3L1 is believed to mainly rely on macrophages (Higashiyama et al. 2019 ; Montgomery et al. 2017 ; Puthumana et al. 2021 ), which are named based on their tissue of origin. In the present study, CHI3L1 expression was higher in fibrocytes than in monocytes and macrophages. Moreover, CHI3L1 expression levels increased with fibrocyte differentiation. Further, CLs, which is an eliminator of macrophages and fibrocytes in the mouse model, abrogated the elevated CHI3L1 observed in Rom‐induced murine MF models (Maekawa et al. 2017 ). Taken together, these findings indicate the involvement of fibrocytes in CHI3L1 secretion and MF progression.

Constitutive JAK/signal transducers and activators of transcription signaling cascade activation is thought to produce BM fibrosis in PMF due to JAK2 V617F and TGF‐β overproduction, which subsequently occurs (Jacobson et al. 1978 ; Rampal et al. 2014 ). Analyses of the JAK2 V617F mouse model revealed that neoplastic fibrocytes played a crucial role in PMF BM fibrosis. Their differentiation depends on TGF‐β1 in an autocrine manner (Ozono et al. 2021 ). However, we could not confirm the difference in TGF‐β1 production between normal and JAK2 V617F fibrocytes in human samples. Further, JAK2 mutation did not affect CHI3L1 production from fibrocytes. Our results revealed the role of TGF‐β1 produced by fibrocytes in BM fibrosis and indicated that they both play crucial roles as myofibroblast stimulators. This is particularly evident with an increasing number of fibrocytes. CHI3L1 was reported to activate the TGF‐β1 pathway in hepatocellular carcinoma, although its signaling pathway was unclear (Qiu et al. 2018 ). These reports indicate that CHI3L1 and TGF‐β1 stimulate myofibroblasts at least additively. However, the absence of TGF‐β1 hyperproduction in human JAK2 V617F fibrocytes raised the possibility of discrepancies between humans and mice in BM fibrosis pathophysiology.

Conclusions

Our results indicate that CHI3L1 is critically involved in fibrocyte–fibroblast interactions and exacerbates BM fibrosis by promoting extracellular matrix formation. CHI3L1 represents the missing link between fibrocytes and myofibroblasts and may serve as a novel therapeutic target for MF.

Author Contributions

Shoichiro Kato, Toshikuni Kawamura: Conceptualization; Writing—original draft; Investigation; Data curation; Methodology; Visualization; Formal analysis; Project administration. Takaaki Maekawa: Conceptualization; Writing—review and editing; Funding acquisition; Data curation; Project administration; Investigation; Validation; Methodology; Formal analysis; Visualization. Hiraku Ogata, Noriaki Tachi: Investigation; Data curation; Resources. Yukiko Osawa: Investigation; Data curation; Resources; Project administration. Shinichi Kobayashi: Supervision.

Ethics Statement

This study complied with the National Defense Medical College Ethics Committee–approved protocols (Approval Numbers 2153 and 4419).

Consent

This study enrolled 52 patients diagnosed with myeloproliferative neoplasm from January 2017 to December 2018 after obtaining written informed consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting_data_R1.

Acknowledgments

The authors thank the members of the Department of Immunology and Microbiology for their support with experimental procedures. The authors also thank Takako Sato in the Department of Defense Medicine, National Defense Medical College, for her assistance with pathological diagnoses. CHI3L1 −/− mice were kindly provided by Kengo Tomita in the Division of Gastroenterology and Hepatology, Department of Internal Medicine, National Defense Medical College. Genetic analysis of patient samples was performed by Soji Morishita, Marito Araki, and Norio Komatsu in Juntendo University Graduate School of Medicine. This work was partly supported by the Japan Society for the Promotion of Science Grant‐in‐Aid for Scientific Research No. 18K08344.

Data Availability Statement

The datasets generated during and/or analyzed in the current study are available from the corresponding author on reasonable request.

References

Arber , D. A. , A. Orazi , R. Hasserjian , et al. 2016 . “ The 2016 Revision to the World Health Organization Classification of Myeloid Neoplasms and Acute Leukemia .” Blood 127 : 2391 – 2405 . 10.1182/blood-2016-03-643544 . 27069254

Bjørn , M. E. , C. L. Andersen , M. K. Jensen , and H. C. Hasselbalch . 2014 . “ Circulating YKL‐40 in Myelofibrosis a Potential Novel Biomarker of Disease Activity and the Inflammatory State .” European Journal of Haematology 93 : 224 – 228 . 10.1111/ejh.12332 . 24689875

Bussink , A. P. , D. Speijer , J. M. F. G. Aerts , and R. G. Boot . 2007 . “ Evolution of Mammalian Chitinase(‐Like) Members of Family 18 Glycosyl Hydrolases .” Genetics 177 : 959 – 970 . 10.1534/genetics.107.075846 . 17720922 PMC2034658

Faibish , M. , R. Francescone , B. Bentley , W. Yan , and R. Shao . 2011 . “ A YKL‐40–Neutralizing Antibody Blocks Tumor Angiogenesis and Progression: A Potential Therapeutic Agent in Cancers .” Molecular Cancer Therapeutics 10 : 742 – 751 . 10.1158/1535-7163.mct-10-0868 . 21357475 PMC3091949

Funkhouser , J. D. , and N. N. Aronson . 2007 . “ Chitinase Family GH18: Evolutionary Insights From the Genomic History of a Diverse Protein Family .” BMC Evolutionary Biology 7 : 96 . 10.1186/1471-2148-7-96 . 17594485 PMC1945033

Higashiyama , M. , K. Tomita , N. Sugihara , et al. 2019 . “ Chitinase 3‐like 1 Deficiency Ameliorates Liver Fibrosis by Promoting Hepatic Macrophage Apoptosis .” Hepatology Research 49 : 1316 – 1328 . 10.1111/hepr.13396 . 31250532 PMC6916176

Jacobson , R. , A. Salo , and P. Fialkow . 1978 . “ Agnogenic Myeloid Metaplasia: A Clonal Proliferation of Hematopoietic Stem Cells With Secondary Myelofibrosis .” Blood 51 : 189 – 194 . 10.1182/blood.v51.2.189.bloodjournal512189 . 620081

Johansen , J. S. 2006 . “ Studies on Serum YKL‐40 as a Biomarker in Diseases With Inflammation, Tissue Remodelling, Fibroses and Cancer .” Danish Medical Bulletin 53 : 172 – 209 . 17087877

Johansen , J. S. , P. Christoffersen , S. Møller , et al. 2000 . “ Serum YKL‐40 Is Increased in Patients With Hepatic Fibrosis .” Journal of Hepatology 32 : 911 – 920 . 10.1016/s0168-8278(00)80095-1 . 10898311

Kawamura , T. , T. Maekawa , K. Kouzu , et al. Forthcoming. “ Sequential Analysis of Murine Myelofibrosis Models Using a Novel Deep Learning‐Based Fibrosis Quantitative Method .” eJHaem . 10.1002/jha2.70273 . PMC13108711 42037754

Knudsen , L. S. , M. Østergaard , B. Baslund , et al. 2006 . “ Plasma IL‐6, Plasma VEGF, and Serum YKL‐40: Relationship With Disease Activity and Radiographic Progression in Rheumatoid Arthritis Patients Treated With Infliximab and Methotrexate .” Scandinavian Journal of Rheumatology 35 : 489 – 491 . 10.1080/03009740600904300 . 17343261

Lee , C. G. , C. A. Da Silva , C. S. Dela Cruz , et al. 2011 . “ Role of Chitin and Chitinase/Chitinase‐Like Proteins in Inflammation, Tissue Remodeling, and Injury .” Annual Review of Physiology 73 : 479 – 501 . 10.1146/annurev-physiol-012110-142250 . PMC3864643 21054166

Lee , C. G. , D. Hartl , G. R. Lee , et al. 2009 . “ Role of Breast Regression Protein 39 (BRP‐39)/chitinase 3‐like‐1 in Th2 and IL‐13–induced Tissue Responses and Apoptosis .” Journal of Experimental Medicine 206 : 1149 – 1166 . 10.1084/jem.20081271 . 19414556 PMC2715037

Maekawa , T. , S. Kato , T. Kawamura , et al. 2019 . “ Increased SLAMF7high Monocytes in Myelofibrosis Patients Harboring JAK2V617F Provide a Therapeutic Target of Elotuzumab .” Blood 134 : 814 – 825 . 10.1182/blood.2019000051 . 31270105

Maekawa , T. , Y. Osawa , T. Izumi , et al. 2017 . “ Myeloproliferative Leukemia Protein Activation Directly Induces Fibrocyte Differentiation to Cause Myelofibrosis .” Leukemia 31 : 2709 – 2716 . 10.1038/leu.2017.112 . 28386106

Montgomery , T. A. , L. Xu , S. Mason , et al. 2017 . “ Breast Regression protein–39/chitinase 3–like 1 Promotes Renal Fibrosis After Kidney Injury via Activation of Myofibroblasts .” Journal of the American Society of Nephrology 28 : 3218 – 3226 . 10.1681/asn.2017010110 . 28679671 PMC5661290

Nishimura , N. , D. De Battista , D. R. McGivern , et al. 2021 . “ Chitinase 3‐like 1 Is a Profibrogenic Factor Overexpressed in the Aging Liver and in Patients With Liver Cirrhosis .” Proceedings of the National Academy of Sciences 118 : e2019633118 . 10.1073/pnas.2019633118 . PMC8092404 33888584

Ozono , Y. , K. Shide , T. Kameda , et al. 2021 . “ Neoplastic Fibrocytes Play An Essential Role in Bone Marrow Fibrosis in JAK2V617F‐induced Primary Myelofibrosis Mice .” Leukemia 35 : 454 – 467 . 10.1038/s41375-020-0880-3 . 32472085 PMC7862060

Pilling , D. , V. Vakil , and R. H. Gomer . 2009 . “ Improved Serum‐Free Culture Conditions for the Differentiation of Human and Murine Fibrocytes .” Journal of Immunological Methods 351 : 62 – 70 . 10.1016/j.jim.2009.09.011 . 19818792 PMC2783789

Puthumana , J. , H. Thiessen‐Philbrook , L. Xu , et al. 2021 . “ Biomarkers of Inflammation and Repair in Kidney Disease Progression .” Journal of Clinical Investigation 131 : 1 – 11 . 10.1172/jci139927 . PMC7843225 33290282

Qiu , Q.‐C. , L. Wang , S.‐S. Jin , et al. 2018 . “ CHI3L1 Promotes Tumor Progression by Activating TGF‐β Signaling Pathway in Hepatocellular Carcinoma .” Scientific Reports 8 : 15029 . 10.1038/s41598-018-33239-8 . 30301907 PMC6177412

Rampal , R. , F. Al‐Shahrour , O. Abdel‐Wahab , et al. 2014 . “ Integrated Genomic Analysis Illustrates the Central Role of JAK‐STAT Pathway Activation in Myeloproliferative Neoplasm Pathogenesis .” Blood 123 : e123 – e133 . 10.1182/blood-2014-02-554634 . 24740812 PMC4041169

Recklies , A. D. , C. White , and H. Ling . 2002 . “ The Chitinase 3‐like Protein Human Cartilage Glycoprotein 39 (hc‐gp39) Stimulates Proliferation of Human Connective‐Tissue Cells and Activates Both Extracellular Signal‐Regulated Kinase‐ and Protein Kinase B‐Mediated Signalling Pathways .” Biochemical Journal 365 : 119 – 126 . 12071845 10.1042/BJ20020075 PMC1222662

Rehli , M. , S. W. Krause , and R. Andreesen . 1997 . “ Molecular Characterization of the Gene for Human Cartilage gp‐39 (CHI3L1), a Member of the Chitinase Protein Family and Marker for Late Stages of Macrophage Differentiation .” Genomics 43 : 221 – 225 . 10.1006/geno.1997.4778 . 9244440

Rehli , M. , H.‐H. Niller , C. Ammon , et al. 2003 . “ Transcriptional Regulation of CHI3L1, a Marker Gene for Late Stages of Macrophage Differentiation .” Journal of Biological Chemistry 278 : 44058 – 44067 . 10.1074/jbc.m306792200 . 12933821

Riabov , V. , A. Gudima , N. Wang , A. Mickley , A. Orekhov , and J. Kzhyshkowska . 2014 . “ Role of Tumor Associated Macrophages in Tumor Angiogenesis and Lymphangiogenesis .” Frontiers in Physiology 5 : 75 . 10.3389/fphys.2014.00075 . 24634660 PMC3942647

Roecklein , B. , and B. Torok‐Storb . 1995 . “ Functionally Distinct Human Marrow Stromal Cell Lines Immortalized by Transduction With the Human Papilloma Virus E6/E7 Genes .” Blood 85 : 997 – 1005 . 10.1182/blood.v85.4.997.bloodjournal854997 . 7849321

Tang , L. , J. Liang , Y. Huang , et al. 2025 . “ An Analysis of the Clinical Value of CHI3L1 as a Biomarker of Multiple Myeloma Progression .” Clinical Immunology 275 : 110474 . 10.1016/j.clim.2025.110474 . 40118129

Tefferi , A. , T. L. Lasho , C. M. Finke , et al. 2014 . “ CALR vs JAK2 vs MPL‐Mutated or Triple‐Negative Myelofibrosis: Clinical, Cytogenetic and Molecular Comparisons .” Leukemia 28 : 1472 – 1477 . 10.1038/leu.2014.3 . 24402162

Thiele , J. , H. M. Kvasnicka , F. Facchetti , V. Franco , J. Van Der Walt , and A. Orazi . 2005 . “ European Consensus on Grading Bone Marrow Fibrosis and Assessment of Cellularity .” Haematologica 90 : 1128 – 1132 . 16079113

Verstovsek , S. , T. Manshouri , D. Pilling , et al. 2016 . “ Role of Neoplastic Monocyte‐Derived Fibrocytes in Primary Myelofibrosis .” Journal of Experimental Medicine 213 : 1723 – 1740 . 10.1084/jem.20160283 . 27481130 PMC4995084

Wagner‐Ballon , O. , D. F. Pisani , T. Gastinne , et al. 2007 . “ Proteasome Inhibitor Bortezomib Impairs Both Myelofibrosis and Osteosclerosis Induced by High Thrombopoietin Levels in Mice .” Blood 110 : 345 – 353 . 10.1182/blood-2006-10-054502 . 17374740

Wang , C. , L. Tang , H. Luo , et al. 2025 . “ Abnormally High Expression of CHI3L1 in Peripheral Blood Mononuclear Cells and Serum and Their Potential Diagnosis and Prediction From Lymphoma Patients .” Frontiers in Immunology 16 : 1557802 . 10.3389/fimmu.2025.1557802 . 40260255 PMC12009859

Zhou , Y. , H. Peng , H. Sun , et al. 2014 . “ Chitinase 3–like 1 Suppresses Injury and Promotes Fibroproliferative Responses in Mammalian Lung Fibrosis .” Science Translational Medicine 6 : 240 – 276 . 10.1126/scitranslmed.3007096 . PMC4340473 24920662