Secreted protein TNA: a promising biomarker for understanding the adipose-bone axis and its impact on bone metabolism

Abstract

Background

Osteoporosis (OP) is a systemic bone disease characterized by reduced bone mass and deterioration of bone microstructure, leading to increased bone fragility. Platelets can take up and release cytokines, and a high platelet count has been associated with low bone density. Obesity is strongly associated with OP, and adipose tissue can influence platelet function by secreting adipokines. However, the biological relationship between these factors remains unclear.

Methods

We conducted differential analysis to identify OP platelet-related plasma proteins. And, making comprehensive analysis, including functional enrichment, protein-protein interaction network analysis, and Friends analysis. The key protein, Tetranectin (TNA/CLEC3B), was identified through screening. Then, we analyzed TNA’s potential roles in osteogenic and adipogenic differentiation using multiple RNA-seq data sets and validated its effect on osteoclast differentiation and bone resorption function through in vitro experiments.

Results

Six OP-platelet-related proteins were identified via differential analysis. Then, we screened the key protein TNA, which was found to be highly expressed in adipose tissue. RNA-seq data suggested that TNA may promote early osteoblast differentiation. In vitro experiments showed that knockdown of TNA expression significantly increased the expression of osteoclast markers, thereby promoting osteoclast differentiation and bone resorption.

Conclusions

We identified TNA as a secreted protein that inhibits osteoclast differentiation and bone resorption. While, it potentially promoted early osteoblast differentiation from bioinformatic results. TNA may play a role in bone metabolism through the adipose-bone axis.

Introduction

Osteoporosis (OP) is the most common bone disease, affecting an estimated 200 million people worldwide [1]. It is a systemic metabolic bone disease characterized by reduced bone mass and microstructural deterioration, which increase bone fragility and fracture risk, potentially leading to serious complications [2]. Bone mineral density (BMD) is the primary diagnostic indicator for OP. However, this diagnostic method is typically employed only after an individual with OP exhibits symptoms, which may delay preventive and therapeutic interventions. Therefore, identifying new hematological biomarkers is of great importance in OP.

Blood and bone cells coexist in close proximity within the bone marrow microenvironment. Platelets, the second most common type of blood cell, are small and short-lived (7–10 days) cell fragments derived from mature megakaryocytes [3]. In addition, platelets play critical roles in various pathophysiological processes, including inflammation, tissue repair, tumor growth, and metastasis [4]. Studies have shown that high platelet counts are associated with low BMD [5, 6]. In terms of clinical relevance, a cohort study of 1,010 men from the MrOS study in Gothenburg, Sweden, showed that high platelet counts were associated with lower BMD at all sites, including total hip BMD (r = − 0.11, p = 0.003) [6]. Additionally, Song found that the platelet/lymphocyte ratio (PLR) is significantly elevated in patients with OP exhibiting fragility fractures, and the ROC diagnostic curve suggests a high diagnostic value (AUC = 0.835) [7]. Regarding mechanistic relevance, firstly, platelet count is associated with inflammation, and previous studies have shown that inflammation is associated with OP [8]. Platelets store serotonin [9], whose counts are positively correlated with serotonin levels. Although the role of serotonin in bone metabolism is markedly complex, studies have shown that high serum serotonin levels are associated with decreased bone mass [10]. Therefore, the platelet count clear correlation with lower BMD values may, in part, be explained by high levels of serotonin and inflammation, although other hormones or cytokines are likely involved. Secondly, data show that the in vitro activation of platelets by prostaglandins and nuclear factor kappa B ligand (RANKL) predominantly induces osteoclast formation [11]. This process may also involve the provision of TGF-β as a source to activate osteoclast formation signaling pathways [12]. Other studies have shown that megakaryocytes can enhance osteoblast proliferation while inhibiting osteoclast formation [13]. Additionally, platelets perform many of their emerging pathophysiological functions by directly interacting with other cells or by storing and releasing cytokines [14]. Adipose tissue can also influence platelet function through the secretion of adipokines, ultimately leading to platelet activation [15].

Tetranectin (TNA) is a calcium-binding protein encoded by the C-Type Lectin Domain Family 3 Member B (CLEC3B) gene. The C-terminal end of the TNA protein monomer contains a glycan recognition structural domain, known as the carbohydrate recognition domain (CRD), which is approximately 130 AA in length. This domain is characteristic of the C-type lectin superfamily and can recognize plasminogen (Plg) to promote the production of plasmin [16]. TNA has been studied in various diseases. In hepatocellular carcinoma, TNA expression has been significantly correlated with tumor purity and immune cell infiltration levels, suggesting a complex interaction between TNA expression and the immune microenvironment in cancer [17]. The specific role of TNA in the regulation of bone remodeling has also been previously reported. Wewer’s study identified the potential role of TNA as a bone matrix protein. The expression of TNA is temporally and spatially consistent with mineralization both in vivo and in vitro. TNA is highly expressed in the newly formed reticular bone of newborn mice, and its overexpression leads to increased tumorigenic osteoid formation in pheochromocytoma cells in vivo, suggesting that TNA may be involved in the process of osteogenesis and mineralization [18]. In addition, TNA is an adipogenic serum protein. Seulgi Go’s study found that the adipogenic function of TNA is mediated by enhancing mitotic clonal expansion via ERK signaling [19].

Figure 1 illustrated the flowchart of our study, which investigated the role of platelet-associated plasma proteins in the development of OP. We firstly identified the key plasma protein TNA and analyzed its expression across various human tissues, discovering that it may mediate bone metabolism through the adipose-bone axis. In conjunction with RNA-seq data from different stem cell sources, our findings suggest that TNA may promote osteogenic differentiation. However, this finding was based on bioinformatics data and needs further experimental verification. Furthermore, in vitro cell experiments showed that TNA inhibits osteoclast differentiation and resorption.

Fig. 1

The flowchart of this study

Materials and methods

Data acquisition and processing

We recruited 18 patients with primary OP and 18 individuals with normal bone mass between March 2018 and February 2019 at Honghui Hospital, Xi’an Jiaotong University, People’s Republic of China. Plasma samples from all 36 participants were subjected to proteomic sequencing. Each participant provided informed consent, and the study was approved by the Institutional Review Board of Honghui Hospital, Xi’an Jiaotong University (project number: 2018-22). Detailed descriptions of the proteomic data and the inclusion and exclusion criteria are available in our previously published study [20]. Additionally, plasma proteomic data from 43 patients with obesity, both before and after weight loss, were obtained from Geyer’s study [21]. In addition, we downloaded 300 platelet-related genes from Li’s study (Supplementary Table 1) [22].

RNA-seq data from wild-type (WT) MSCs derived from induced pluripotent stem cells (iPSCs) differentiated into osteoblasts at five time points (0, 7, 14, and 17 days) were obtained from GSE102732 in the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). Using the Nature Genetics data [23] from GSE113253, we obtained RNA-seq data of bone marrow mesenchymal stromal cells (BMSCs) and adipose-derived stem cells (ADSCs) for osteogenic and adipogenic differentiation at five time points: 4 h, 1 day, 3 days, 7 days, and 14 days. On the basis of Robert’s study [24], we acquired ADSCs and performed RNA-seq following osteogenic and adipogenic differentiation induction for 24 h. The RNA-seq raw files in fastq format were obtained from the functional genomics data collection (ArrayExpress repository, https://www.ebi.ac.uk/biostudies/arrayexpress) under the number E-MTAB-6298, running FastQC (version 0.10) to assess read quality, generating Indexes, aligning to genome with STAR-aligner (version 2.7), summarizing gene counts with featureCounts (version 2.0.1), importing gene counts into R/RStudio, and using DESeq2 (version 1.10) to find significant genes.

Analysis of differentially expressed proteins (DEPs)

We performed a differential analysis of plasma proteomic data between patients with OP and healthy individuals. The mean protein values were log2-transformed and visualized using a volcano plot (|Log₂FC| > 0.5, p-value < 0.05). We used the “Retrieve/IDmapping” function to perform the id transformation based on Uniprot database (https://www.uniprot.org/id-mapping), which is the most extensive and informative protein database and is the first choice for querying protein functions. Moreover, 66 DEPs, identified after gene ID mapping, were intersected with 300 PLTs and displayed in a Venn diagram.

Gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis

After performing differential analysis, we identified six platelet-related OP genes. Then, the “clusterProfile” package in R was used to perform GO and KEGG enrichment analyses to elucidate the potential pathogenic mechanisms and biological pathways associated with these genes. The results of the GO and KEGG analyses are presented using bar graphs. In addition, the pairwise similarity of enriched terms was calculated using the Jaccard similarity index (JC). The “ggplot2,” “igraph,” and “ggraph” packages were employed for EMAP visualization of the similarity results.

Protein-protein interaction (PPI) and friends analysis

The online STRING database (http://string-db.org) is widely used for constructing PPI networks and scoring interactions between target proteins. We conducted a PPI network analysis of six proteins to explore their interactions using a confidence cutoff of > 0.4. The hub proteins within the PPI network were identified and scored using the “Cytohubba” plugin in Cytoscape. In addition, we performed Friends analysis using the “GOSemSim” package to identify key genes by constructing a gene interaction network and assessing the importance of each gene through network topology analysis.

Human protein atlas (HPA) database analysis

Immunohistochemical (IHC) data for A2M, COL1A1, and TNA in adipose tissue were obtained from the HPA database (http://www.proteinatlas.org) [25]. In addition, we analyzed the expression levels (nTPM) in human tissues using the Genotype-Tissue Expression (GTEx) database.

The DeepTMHMM tool

DeepTMHMM (https://dtu.biolib.com/DeepTMHMM) is a deep learning model for transmembrane topology prediction and classification and is currently the most complete and best-performing method for the prediction of the topology of both alpha-helical and beta-barrel transmembrane proteins [26]. We used DeepTMHMM to analyze the TNA protein to determine whether it was a classical secretory protein and whether it had transmembrane domains.

Single-cell RNA-seq (scRNA-seq) data

On the basis of Zhong’s study [27], we obtained scRNA-seq data for BMSCs from the Single Cell Portal database (https://singlecell.broadinstitute.org/single_cell). We analyzed the expression of TNA in each cell cluster, including early mesenchymal progenitors (EMPs, cluster 1), intermediate mesenchymal progenitors (IMPs, cluster 2), late mesenchymal progenitors (LMPs, cluster 3), osteoblasts (cluster 4), osteocytes (cluster 5), lineage-committed progenitors (LCPs, cluster 6), adipocytes (cluster 7), and chondrocytes (clusters 8 and 9).

Cell cultures and transfection

HEK293 cell lines, obtained from the American Type Culture Collection (ATCC), were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM; Invitrogen), supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. For the isolation and culture of bone marrow–derived monocyte-macrophages (BMMs), mouse bone marrow cells were harvested from the femurs and tibias by flushing the bones with sterile phosphate-buffered saline (PBS). The harvested bone marrow cells were then passed through a 70 μm cell strainer to remove debris, followed by centrifugation at 300 x g for 5 min. After discarding the supernatant, the cell pellet was resuspended in red blood cell lysis buffer and incubated for 2 min to lyse erythrocytes, followed by another round of centrifugation. The resulting cells were then resuspended in α-modified Minimal Essential Medium (α-MEM; Invitrogen), supplemented with 10% FBS and 10 ng/mL macrophage colony-stimulating factor (M-CSF; PeproTech), and cultured for 48 h. After incubation, nonadherent cells, which represent the BMM population, were carefully collected and prepared for subsequent experiments.

Lentiviral packaging and infection

For the lentivirus-mediated knockdown assay, sense and anti-sense short hairpin DNA (shDNA) oligonucleotides targeting the mouse TNA gene (5ʹ-gccttacagactgtgtgcctg-3ʹ) were annealed and ligated into the pLKO.1 vector (Addgene 10878). Next, 293T cells at approximately 70% confluency were transfected with 1 μg of recombinant pLKO.1 vector, 750 ng of psPAX2 (Addgene 12260), and 250 ng of pMD2.G (Addgene 12259) using Lipofectamine 2000 (Invitrogen). Twelve hours post-transfection, the medium was replaced. After an additional 36 h, the culture medium containing the virus was harvested and used to infect BMMs at 50% confluency. Following 48 h of infection, the BMMs were selected with 1 μg/mL puromycin for at least 3 days before proceeding with subsequent experiments.

In vitro osteoclastogenesis

Bone marrow–derived monocyte-macrophages (BMMs) were plated at 1 × 10⁵ cells/cm² and cultured in a medium containing 10 ng/mL M-CSF and 50 ng/mL sRANKL (PeproTech). Osteoclast cultures on plastic were terminated at 5 days and on bone at 6 days. The medium was refreshed on days 2 and 4, with the start of induction designated as day 0.

Tartrate-resistant acid phosphatase (TRAP) staining assay

Cells were fixed and stained for TRAP activity after 5 days in culture, using a commercial kit (Sigma 387-A) according to the product instructions. ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used to quantify the proportion of TRAP-positive cells.

Pit-formation assay

6 days after the initiation of induction, bone slices were incubated in 0.5 N NaOH for 30 s and cells scraped off using a cotton swab, then incubated with 20 mg/mL peroxidase-conjugated wheat germ agglutinin (Sigma) in PBS for 30 min, washed with PBS three times, and exposed to 3,30-Diaminobenzidine tablets (Sigma; D4168) for 15 min before washing. BioQuant OSTEO 2010 (BioQuant Image Analysis Corporation, Nashville, TN, USA) was used to quantify pit area. ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used to quantify the area of bone resorption.

Real-time polymerase chain reaction (rt-qPCR)

Total RNA was extracted using TRIzol reagent (Invitrogen), and cDNA was synthesized from 2 μg of total RNA using the SuperScript II First-Strand Synthesis System (Invitrogen). Quantitative real-time PCR was conducted in a 25 μl mixture containing 12.5 μl SYBR Green PCR Master Mix (Takara), 200 nM of each primer, and 5 μL of cDNA. A CFX96 Sequence Detection System (Bio-Rad) was used, employing the comparative threshold cycle method for relative quantification. GAPDH was used as an internal control. Primer sequences are listed in Table 1.

Table 1 Primer sequences for rt-qPCR

Statistics analysis

All statistical analyses were conducted using R software (version 4.2.2) and GraphPad Prism 9. A p-value of < 0.05, determined by an unpaired Student’s t-test, was considered statistically significant between the two groups. Data are presented as means ± standard errors of the mean (SEM) unless otherwise specified. The results are representative of more than three independent experiments. Pearson correlation analysis was conducted. All statistical tests were two-sided, and the level of statistical significance was set at a p-value of < 0.05.

Results

Identification of differentially expressed platelet-related proteins in OP

Figure 2A illustrates our collection of 36 samples, comprising 18 patients with primary OP and 18 individuals with normal bone mass. We identified platelet-related OP genes through differential and intersection analyses. Differential analysis of plasma proteome data was conducted using volcano plots (Fig. 2B). The results indicated that 19 plasma proteins were highly upregulated in the OP group, whereas 112 proteins were downregulated (|Log₂FC| > 0.5, p < 0.05). Following protein and gene ID conversion, a Venn plot indicated that six genes—A2M, TF, PF4, SPP2, TNA, and COL1A1—were associated with platelet-related OP proteins (Fig. 2C). Furthermore, a map was generated to illustrate the differential expression of these six key proteins between the two groups. Except for A2M, the expression levels of the other five proteins were significantly higher in the OP group (Fig. 2D).

Fig. 2

Differentially expressed platelet-related genes in osteoporosis. (A) OP plasma proteome data and platelet-related gene list acquisition and flow chart. (B) Differential analysis of plasma proteome between OP and normal group by volcano plot (|Log₂FC| > 0.5, p-value < 0.05) (OP = 18 samples, Normal = 18 samples). (C) Venn diagram of the differential plasma proteins and the 300 platelet related genes. (D) Differential expression level of six key genes (FC = OP/Normal)

Functional enrichment of six key proteins

Functional enrichment analysis was conducted for these six proteins. The results of the GO (Fig. 3A) and KEGG (Fig. 3B) enrichment analyses are presented using bar graphs, with detailed information provided in Table 2. The significantly enriched biological process (BP) terms identified included the following: humoral immune response, bone remodeling, response to nutrients, response to corticosteroids, biomineralization, positive regulation of bone resorption, regulation of macrophage differentiation, regulation of bone remodeling, and positive regulation of the canonical Wnt signaling pathway. The significantly enriched cellular component (CC) terms identified were secretory granule lumen and endoplasmic reticulum lumen. The significantly enriched molecular function (MF) terms identified were protease binding and chemokine activity. In addition, the significantly enriched KEGG terms identified were ferroptosis, mineral absorption, complement and coagulation cascades, extracellular matrix (ECM)-receptor interaction, AGE-RAGE signaling pathway in diabetic complications, viral protein interaction with cytokines and cytokine receptors, and protein digestion and absorption.

Table 2 GO and KEGG terms of six platelet related osteoporosis genes

Fig. 3

Functional enrichment analysis of six key genes. (A) GO enrichment analysis, including BP, CC, and MF. (B) KEGG enrichment analysis. (C) Similarity analysis between related terms in GO enrichment analysis. (D) Similarity analysis between related passways in KEGG enrichment analysis. (p < 0.05)

Moreover, EMAP plots showed the similarity between individual functional terms, with line thickness representing the degree of pairwise similarity between the terms. Thicker lines indicate greater similarity (Fig. 3C, D). We observed a high similarity between the terms “positive regulation of bone resorption” and “regulation of bone remodeling,” as well as with “response to corticosteroids,” " response to nutrients,” and “protease binding.”

PPI network and identification of hub proteins

The PPI network of the six proteins was analyzed using the STRING database (scores > 0.4) (Fig. 4A). Subsequently, the “CytoHubba” plug-in in Cytoscape software was used to rank the importance analysis, revealing that A2M and CLEC3B exhibited the highest scores and rankings (Fig. 4B). Additionally, Friends analysis identified A2M, COL1A1, and CLEC3B as having the highest importance scores among the six genes (Fig. 4C). The genes were ranked in descending order based on their average similarity to other genes, with the top-ranked genes showing the greatest similarity to other genes and being identified as key genes.

Fig. 4

Protein-Protein Interaction network construction and hub gene identification. (A) The PPI network by STRING database (scores > 0.4). (B) Protein scores were calculated using the Cytohubba plug-in. (C) Importance scores between 6 genes are calculated using Friends analysis

Expression levels of Hub Proteins and identification of secretion type

The expression levels (nTPM) of A2M, COL1A1, and TNA in human tissues were analyzed using GTEx data from the HPA database. The results indicated that the expression of TNA was highest in adipose tissue. Additionally, immunostaining analysis of the three proteins in adipose tissue revealed that A2M and TNA were moderately expressed, whereas the COL1A1 protein was expressed at low levels (Fig. 5A) based on HPA database.

Fig. 5

Identification of hub gene expression in adipose tissue and analysis of secretion type. (A) Expression of A2M, COL1A1, CLEC3B in tissues from the GTEx database and immunohistochemistry in adipose tissue from HPA database. (B) Flow chart of screening obesity and osteoporosis co-related key gene. (C) Venn plot between plasma proteome in 43 obese patients before and after weight loss, and 3 hub genes. (D) Clec3b expression level in brain, muscle, subcutaneous adipose and visceral adipose of WT mice (2 samples). (E) Identification the secretion mode of TNA protein by DeepTMHMM database

The flow chart for determining the intersection between the DEPs in the plasma proteome of 43 patients with obesity and the three hub genes is presented in Fig. 5B. Unexpectedly, only TNA was identified as a key protein (Fig. 5C). We performed RT-qPCR analysis of CLEC3B in four mouse tissues: brain, muscle, subcutaneous adipose, and visceral adipose tissue (two samples from WT mice). The results showed that CLEC3B expression was high in adipose tissue, but statistical analysis could not be made due to the lack of samples. (Fig. 5D).

In addition, the type of TNA protein was identified using the DeepTMHMM tool, revealing a signal peptide sequence at its N-terminus: MELWGAYLLLCLFSLLTOVTT. No transmembrane domains were predicted, suggesting that TNA is a classically secreted protein (Fig. 5E).

CLEC3B expression in MSC subsets

Through scRNA-seq data of the bone marrow microenvironment (BMM), we found that CLEC3B expression was significantly higher in the cluster 1 subset, which corresponds to EMPs. Additionally, we observed an increase in CLEC3B expression in a small number of IMP cells (Fig. 6A–C).

Fig. 6

CLEC3B expression in mesenchymal stem cell subsets, and its possible roles. (A) Cell subsets map of Bone Marrow Microenvironment. (B) Expression of CLEC3B in different cell subsets. (C) Expression of CLEC3B in different cell subsets by violin plot. (D) Expression of CLEC3B in osteogenic differentiation of BMSCs (GSE102732, each group = 2 duplicate samples). (E) Expression of CLEC3B in osteogenic differentiation of BMSCs from NG dataset (GSE113253, each group = 3 duplicate samples). (F) Expression of CLEC3B in osteogenic differentiation of ADSCs from NG dataset (GSE113253, each group = 3 duplicate samples). (G) Expression of CLEC3B in 24 h osteogenic differentiation of ADSCs (E-MTAB-6298, each group = 3 duplicate samples). (H) Expression of CLEC3B in adipogenic differentiation of BMSCs from NG dataset (GSE113253, each group = 3 duplicate samples). (I) Expression of CLEC3B in adipogenic differentiation of ADSCs from NG dataset (GSE113253, each group = 3 duplicate samples). (J) Expression of CLEC3B in 24 h adipogenic differentiation of ADSCs (E-MTAB-6298, each group = 3 duplicate samples)

Osteoblast differentiation was assessed at multiple time points using BMSCs. The expression of CLEC3B was significantly upregulated over time (Fig. 6D). However, in another data set, CLEC3B expression decreased abruptly on day 14 (Fig. 6E; Supplementary Table 2). During the osteogenic induction of ADSCs (adipose tissue [AT]), CLEC3B expression was significantly upregulated on days 3 and 7 but downregulated on day 14 (Fig. 6F; Supplementary Table 3). Additionally, in another data set, CLEC3B expression was significantly elevated within 24 h of osteogenic differentiation induction (Fig. 6G). No significant change was observed in the expression of CLEC3B during the adipogenic differentiation of BMSCs (Fig. 6H; Supplementary Table 2). In addition, no significant difference was observed in the adipogenic differentiation of ADSCs (Fig. 6I, J; Supplementary Table 3).

TNA regulates osteoclast differentiation and function in vitro

We specifically examined osteoclast (OC) genesis by inducing differentiation in primary BMMs, assessing the impact of TNA expression modulation. To evaluate TNA’s role in this process, we transduced primary BMMs with lentivirus expressing short hairpin RNA (shRNA) against TNA or a control shRNA. Our findings revealed that TNA deficiency in BMMs led to the upregulation of mRNAs encoding osteoclastic markers, including Itgb3 (which encodes Integrin β3, a crucial receptor on the osteoclast surface), Acp5 (encoding tartrate-resistant acid phosphatase, an enzyme abundantly secreted during osteoclast activation), Dcstamp (encoding dendritic cell-specific transmembrane protein, essential for osteoclast fusion) [28,29,30], and Ctsk (encoding cathepsin K, a lysosomal protease highly expressed in osteoclasts) (Fig. 7A).

Fig. 7

TNA knockdown promotes osteoclast differentiation and activity in BMMs. (A) Knockdown of TNA leads to an increase in the mRNA levels of osteoclastic markers in BMMs. BMMs were induced with 10 ng/mL M-CSF and 50 ng/mL RANKL for 5 days before being harvested for mRNA analysis. (B) TNA deficiency promotes the formation of mature osteoclasts in BMMs. BMMs were either non-transduced (control) or transduced with TNA-inhibiting shRNA and subsequently induced to differentiate with 10 ng/mL M-CSF and 50 ng/mL RANKL for 5 days. The cells were then fixed and stained with TRAP solution (Sigma 387-A). The black arrows point to the osteoclasts. Scale bar = 100 μm. TRAP-positive multinucleated cells were counted. (C) TNA deficiency increases osteoclast-mediated bone resorption in BMMs. Control and shTNA BMMs were cultured on bone slices with 10 ng/mL M-CSF and 50 ng/mL RANKL for 6 days. The cells were removed, and resorption pits were stained with peroxidase-conjugated wheat germ agglutinin. Scale bar = 100 μm. The area of the pits was measured

Further analysis of TNA’s effect on osteoclast differentiation and function showed a significant increase in TRAP-positive cells and enhanced bone resorption in TNA-inhibited BMMs (Fig. 7B). The pit formation assay results demonstrated that TNA-inhibited cells produced a greater total area of bone resorption pits, indicating that the inhibition of TNA expression enhances overall bone resorption. (Fig. 7C). These results are representative of more than three independent experiments.

Discussion

In this study, we first identified TNA as a key osteoporotic platelet-associated plasma protein. Notably, we found that TNA is significantly overexpressed in adipose tissue and is a secreted protein. Employing multi-RNA-seq data. In vitro experiments showed that knockdown of TNA expression significantly increases the expression of osteoclast markers and enhances resorption functions.

The MeOS cohort had demonstrated an association between high platelet count and low BMD [6]; however, the underlying mechanism remained unclear. Platelets can mediate the uptake and release of cytokines. In our research, we utilized plasma proteomics to explore the link between OP related to platelet related proteins and identified TNA as a key plasma protein, which may provide a potential explanation. TNA is a secreted protein released from the platelets of patients with stable angina pectoris (SAP) [31]. This finding is consistent with our hypothesis regarding the secretion pattern of TNA, but further investigation is required to determine the source of TNA secretion. Certain cancer cells or cells associated with cancerous tissue have been demonstrated to be capable of expressing TNA and secreting it into the ECM. From structural analysis, we also found TNA to be an exocrine protein. However, the level of TNA present in the serum of patients with cancer is markedly reduced, and the relationship between these two is not always straightforward [17, 32,33,34,35,36,37,38,39]. We observed a similar pattern where TNA is highly expressed in adipose tissue but relatively low in the plasma of patients with obesity. This suggests a discrepancy between the tissue expression of TNA and its expression in circulating blood.

The bone-adipose axis hypothesis has been extensively reviewed elsewhere [40]. Obesity, a major global health problem, is strongly associated with various chronic bone metabolic disorders, such as OP. When adipose tissue reaches its maximum energy storage capacity, adipokines such as Endolipin, Omentin-1, Chemerin, Lipocalin 2, vaspin, retinol-binding protein-4 (RBP-4), Nesfatin-1, Apelin, and apolipoproteins (APNs) are released, which can influence the progression of bone metabolic diseases [41]. Adipose tissue can also affect platelet function through the secretion of adipokines, ultimately leading to platelet activation [16]. TNA has also been identified as an adipogenic serum protein that enhances adipogenesis [42, 43]. We found that TNA was highly expressed in adipose tissue, and functional enrichment analysis revealed a possible involvement in processes such as ECM receptor interaction. However, the specific role of TNA in bone metabolism has been poorly studied.

The root cause of OP is the imbalance between osteoclast-mediated bone resorption and osteoblast-mediated osteogenic homeostasis. The expression of TNA has been shown to significantly increase during the mineralization phase of bone development [18, 44]. This finding aligns with our observation that TNA expression levels gradually increase during the osteogenic differentiation of stem cells. Notably, we also observed a decrease in TNA expression at later stages of osteogenic differentiation, suggesting that TNA may inhibit mineralized nodules at these advanced stages. While impaired osteogenic mineralization can contribute to OP, the high expression of TNA in the plasma of patients with OP may be a potential contributing factor.

Osteoclast differentiation is regulated by Ca²⁺ shock and its downstream signals [40]. Specifically, intracellular Ca²⁺ regulates osteoclast bone resorption through the CaMKIV-CREB and Cn-NFATc1 pathways [45]. TNA can also bind to Ca²⁺ and polysaccharide sulfate. The three Ca²⁺ binding sites of TNA are located in the CRD, and TNA binds to Plg K4 only in the absence of Ca²⁺. This suggests a competitive relationship between Ca²⁺ and Plg K4, which may serve as a switch for TNA binding to Plg K4 [46]. The binding of Ca²⁺ to regulate downstream responses may be an important function of TNA. Furthermore, we showed that knockdown of TNA expression significantly increased the expression of osteoclast markers and enhanced the resorption function of osteoclasts by in vitro experiments.

We explored the potential role of TNA and found that it may influence the differentiation and function of osteoblasts and osteoclasts through the adipose-bone axis. However, our study has some limitations. First, the sample size of the OP plasma proteome we analyzed was limited to 18 samples, necessitating the collection of additional samples for validation. Second, the induction of osteoblast differentiation by TNA was based solely on bioinformatic data, without corresponding experimental validation. Finally, while our study offers a new perspective, the specific mechanism of TNA remains unclear and requires further investigation. In particular, how to mediate the bone resorption function of osteoclasts and need to construct gene knockout mice.