Introduction
Osteonecrosis of the femoral head (ONFH) is a progressive and disabling chronic disease that can cause femoral head collapse, and lead to total hip replacement [
1‐
3]. ONFH is characterized by bone cell necrosis due to ischemia of the femoral head [
4]. It is estimated that there are approximately 100,000–200,000 new ONFH cases were diagnosed in China per year [
5]. The early symptoms of ONFH are often not apparent, and patients are often diagnosed at the late stages (ARCO stage III–IV). Various approaches exist for treating early ONFH, ranging from surgical interventions to preserve the hip to medication [
6‐
9]. Despite the valuable insights provided by recent studies on ONFH in guiding our clinical practices, a thorough understanding of the molecular mechanism remains elusive. Hence, there is an urgent need to investigate novel biomarkers and effective therapeutic targets for early diagnosis and treatment of ONFH patients.
N6-methyladenosine (m6A) is the most abundant epigenetic RNA modification in eukaryotes, regulating RNA metabolism in cells by affecting writers, erasers, and readers, ultimately impacting human health and disease processes [
10]. The m6A function is primarily regulated by the combination of m6A binding protein, demethylase and m6A methyltransferase [
11]. M6A is involved in various phases of RNA metabolism, including mRNA translation, degradation, and folding [
12]. Dysregulated m6A levels was associated with numerous diseases, such as tumors, neurological disorders, and metabolic conditions [
13‐
15]. Moreover, the m6A RNA methylation also regulates the differentiation of bone marrow mesenchymal stem cells (BMSCs) and involves in the disease progression associated with BMSCs abnormalities [
16]. Although studies have confirmed that m6A regulates the translation of various femoral necrosis-related mRNAs in tumor and inflammation-related diseases [
17], there is still limited studies detecting the role of m6A in ONFH. Similarly, increasing evidence has highlighted the integral role of inflammatory osteoimmunology in ONFH pathogenesis [
18‐
20]. Therefore, determining the role of m6A and the immune microenvironment in ONFH provides a novel approach for its prevention and treatment.
In this study, we explored the impact of m6A modulators in the diagnostic classification of ONFH and evaluated the degree of immune infiltration in ONFH. We screened seven key regulators from our database to assess the disease risk of ONFH and assessed the immune microenvironment of two m6A patterns of ONFH. Additionally, we conducted cross-validation of key m6A regulatory factors in ONFH using external datasets and bone samples.
Methods
Data collection
The ONFH dataset GSE123568 was acquired from the GEO database (
https://www.ncbi.nlm.nih.gov/geo/). This dataset comprises peripheral serum samples from 10 healthy individuals and 30 ONFH patients. It contained 25 m6A regulators, consisting of 15 readers, 2 erasers, and 8 writers [
21‐
23].
Construction of random forest (RF) and support vector machine (SVM) models
We conducted the RF and SVM models to evaluate the risk of ONFH. The RF algorithm was generated by the “randomForest” package in the R software. We determined the importance score > 1, thus screening 13 key regulators out of 25 m6A regulators. SVM is a useful tool for developing predictive models or classifiers [
24,
25]. We assessed the accuracy of both models through receiver operating characteristic (ROC) curve analysis, “residual plots,” and “reverse cumulative distribution of residuals.’
Nomogram model
We created a nomogram model to predict the risk of ONFH using the “RMS” and “rmda” packages in the R software. To evaluate the prediction accuracy of the nomogram model, we utilized decision curve analysis (DCA) and calibration curves. Besides, we used clinical impact curves to determine whether the predictions of the nomogram model would beneficial the patients.
Identification of two m6A isoforms and m6A gene isoforms
Based on the 25 m6A regulators, the sample data were categorized into m6A models by consistency clustering analysis. Next, we screened the differentially expressed genes (DEGs) related to m6A. After that, we classified the DEGs associated with m6A into two m6A subtypes using consensus clustering analysis. The m6A-related DEGs were implemented by R's “limma” package. For subtype classification, the consensus clustering algorithm was generated by the “ConsensusClusterPlus” package of R software.
The sample's m6A score
The m6A score was used to assess the genetic characteristics of the m6A pattern of ONFH. The m6A pattern was first determined using principal component analysis (PCA), and then the m6A score was calculated by the following formula: m6A score = 6 (PC1i + PC2i) [
26‐
28].
Functional enrichment analysis of DEGs between different M6A modes
The gene ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of DEGs between different M6A patterns were performed. Analysis of GO mainly includes biological process (BP), molecular function (MF), and cellular component (CC) domains [
29]. KEGG is a reference knowledge base for the biological interpretation of genomic sequences and other high-throughput data, which enables the identification of potential signaling pathways through KEGG enrichment analysis [
30].
Evaluation of the microenvironment for immune infiltration
In order to investigate the immune infiltration status of ONFH, we performed single-sample gene set enrichment analysis (ssGSEA) to assess the level of immune infiltration in the samples [
31]. The gene expression levels of the samples were ranked and their grades were obtained, which were recorded as ssGSEA scores. Then, we detected the degree of immune cell infiltration in each sample.
Cross-validation of the key m6A regulatory factors in ONFH
We initially confirmed the expression of 7 key m6A regulatory factors in GSE74089. Then, we collected 6 femoral head specimens from the Third Affiliated Hospital of Guangzhou University of Chinese Medicine, comprising three with ONFH and three with femoral neck fractures (FNF). Approval for this study was granted by the Ethics Committee of the Third Affiliated Hospital, and patient informed consent was obtained. The samples were obtained during total hip arthroplasty. The bone tissue was immediately frozen in liquid nitrogen, and RNA extraction followed a standardized protocol. Thermo Scientific's qRT-PCR quantified mRNA transcripts using the NanoDrop Lite and CFX96 Touch real-time PCR detection systems (BIO-RAD, CFX96, USA). Amplification conditions comprised an initial 95 °C for 10 min, succeeded by 40 cycles of 95 °C for 15 s, 60 °C for 30 s, and 60 °C for 30 s. The stability of potential candidate gene expression was evaluated using the standard comparative method (ΔΔCt), with relative target gene expression levels computed using the 2 − ΔΔCt method. Primer sequences are shown in Additional file
1: Supplementary Table 1.
Statistical analysis
We used linear regression analysis to assess the association between writers and erasers. Additionally, we employed the Wilcoxon test for comparisons between two groups and the Kruskal–Wallis test for multiple group comparisons. Statistical significance was defined as P values less than 0.05 (two-sided). All statistical analyses were conducted using R version 4.2.2.
Discussion
ONFH is a chronic degenerative disease with a complex, currently unknown pathomechanism, posing challenges in diagnosis and treatment. Growing evidence indicates that immune response and inflammation play a pivotal role in the pathogenesis of ONFH [
34‐
36]. As the most prevalent mRNA post-transcriptional modification in eukaryotic cells, m6A has garnered significant attention from researchers. The etiology of ONFH is widely recognized to be intricately associated with BMSCs, encompassing factors such as decreased cell quantity and impaired osteogenic potential. In their research, Wang et al. identified that FTO inhibits the osteogenic differentiation of bone marrow mesenchymal stem cells by mediating the demethylation of Runx2 mRNA, consequently leading to a decrease in bone mass [
37]. Another study demonstrated that the conditional knockout of METTL3 in BMSCs resulted in increased bone loss and impaired bone formation in mice [
16]. Furthermore, the down-regulation of METTL14 expression was observed in ONFH, and its up-regulation was found to enhance the proliferation and osteogenic differentiation of ONFH bone marrow mesenchymal stem cells [
17]. These investigations establish a substantial correlation between m6A methylation and the advancement of ONFH. Moreover, recent studies have unveiled its pivotal role in both innate and acquired immune responses [
38]. However, the exploration of m6A in patients with ONFH remains limited. In this study, we systematically investigate m6A's modification patterns within the ONFH immune microenvironment. We explore how m6A modifications affect immune cell infiltration, response, function, and activation pathways in ONFH. The finding will provide new insights and guidance for the early diagnosis and target therapy of ONFH.
First, we obtained the transcriptome expression matrix and clinical data of ONFH patients from the GEO database. Through differential expression analysis, we identified 13 differences in m6A gene expression between the normal group and ONFH patients. Specifically, YTHDF3, FMR1, YTHDC2, and RBM15 showed significant upregulation in ONFH samples, while ALKBH5 exhibited significant downregulation in ONFH samples. By utilizing the random forest algorithm, we screened seven key m6A regulators (CBLL1, YTHDF2, ALKBH5, RBM15B, YTHDF1, HNRNPC and YTHDF3) that are implicated in the development of ONFH. Based on these seven key risk factors, we constructed a nomogram model to predict the risk of ONFH. This model holds significance in enabling early identification of individuals at high risk for ONFH during clinical practice, and the identified m6A regulators may serve as potential new molecular targets for ONFH diagnosis or treatment. Furthermore, we validated these 7 key m6A regulators factors through external datasets and bone samples. Furthermore, the precise role and underlying regulatory mechanisms of m6A modification in ONFH remain unclear.
We classified ONFH patients based on the screened 7 key m6A regulators. Additionally, we observed expression correlations between several m6A regulators; for example, METTL14 showed a negative correlation with FTO (
P = 0.013), which revealed a modeled regulatory network for m6A. We assessed the impact of m6A modification patterns on immune cell infiltration. Our findings indicated a close association between YTHDF3 and the infiltration of inflammatory immune cells in ONFH, such as Gamma Delta cell, immature dendritic cell, and T follicular helper cell. The YTH protein family, characterized by the presence of the YTH structural domain, plays a significant role in organismal development and evolution [
39]. According to reports, YTHDF3 plays an important functions in various tumor and immune processes [
40]. Moreover, it collaborates with YTHDF1 and YTHDF2 to regulate mRNA stability and degradation, impacting cell proliferation and differentiation [
41]. Our study suggests that YTHDF3 may seve a key marker for the regulation of ONFH by immune cells particularly T follicular heloer cell. However, limited research has been conducted on the relationship between YTHDF3 and ONFH. In addition, we found that ONFH immune infiltration was predominantly influenced by immune cells such as Gmma delta cell and Th2 cell under different m6A modifications. T cells, being a key component of cell mediated acquired immunity, are closely associated with the development of ONFH [
42]. Th2 cells, direct the immune response through producing signature cytokines to control pathogens and regulate other immune cells [
43]. The elevated levels of IL15, IL21, and IL34 in ONFH patients indicate a close association between immune inflammatory responses and ONFH [
44‐
46]. Collectively, these findings suggest that both immune and m6A mediated inflammatory responses play pivotal roles in the ONFH, involving Th2-dominated immune cells and inflammatory mediators.
Our study offers novel insights and guidance for the early diagnosis and target therapy of ONFH. However, certain limitations exist in our research. First, our work in view of bioinformatics analysis, numerous theoretically effective methods require validation through additional wet experiments to confirm their accuracy. Second, the sample size of ONFH patients in our study was limited, warranting a larger cohort for further exploration. Nonetheless, our investigation underscores the substantial influence of m6A modification on ONFH's molecular mechanisms, furnishing a new outlook on comprehending its potential pathogenesis (Additional file
1).
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.