Recent advances of osteoimmunology research in rheumatoid arthritis: From single-cell omics approach

Abstract

Cellular immune responses as well as generalized and periarticular bone loss are the key pathogenic features of rheumatoid arthritis (RA). Under the pathological conditions of RA, dysregulated inflammation and immune processes tightly interact with skeletal system, resulting in pathological bone damage via inhibition of bone formation or induction of bone resorption. Single-cell omics technologies are revolutionary tools in the field of modern biological research.They enable the display of the state and function of cells in various environments from a single-cell resolution, thus making it conducive to identify the dysregulated molecular mechanisms of bone destruction in RA as well as the discovery of potential therapeutic targets and biomarkers. Here, we summarize the latest findings of single-cell omics technologies in osteoimmunology research in RA. These results suggest that single-cell omics have made significant contributions to transcriptomics and dynamics of specific cells involved in bone remodeling, providing a new direction for our understanding of cellular heterogeneity in the study of osteoimmunology in RA.

Introduction

Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease. Its prevalence varies between geographic regions and is influenced by genetic factors, environmental exposures, demographics, and socioeconomics.^[1] In China, RA affects approximately 0.28–0.45% of the population.^[2,3] The most prominent clinical manifestations of RA are inflammation and progressive destruction of multiple facet joints.^[4] Cellular immune responses as well as generalized and periarticular bone loss are the key pathogenic features of RA.^[5] Bone remodeling is a dynamic process throughout life that is tightly regulated by the action of osteoclasts and osteoblasts. The balance between bone formation and resorption processes helps to maintain the skeletal strength and integrity. However, various physiological or pathological factors, such as age, menopause, smoking, overuse of glucocorticoids, and rheumatic diseases may disrupt this balance and contribute to osteopenia or osteoporosis (OP). Under the pathological conditions of RA, dysregulated inflammation and immune processes tightly interact with the skeletal system, resulting in pathological bone damage via inhibition of bone formation or induction of bone resorption.^[6,7]

The concept of extensive crosstalk between immune system and skeletal system, which is called osteoimmunology, has been studied for two decades.^[8] As a representative disease of inflammatory bone injury, RA is one of the most studied disease models in the field of osteoimmunology. Various immune and skeletal cells, cytokines, and signaling pathways, such as osteoblasts, osteoclasts, macrophages, T/B lymphocytes, the receptor activator of nuclear factor-κB ligand (RANKL)/RANK/osteoprotegerin (OPG) axis, tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), bone morphogenetic protein (BMP), interleukin family, sphingosine-1-phosphate (S1P)-S1P receptor-1 (S1PR1) signaling, and others, are involved in the same microenvironment and processes during bone erosion^[9,10] [Figure 1]. In the beginning, these studies mainly focused on the effect of immune system, especially the role of adaptive immunity, on bone homeostasis.^[8,11] The immune cells, antibodies, immune complexes, and inflammatory factors can directly or indirectly break the bone homeostasis and promote osteoclastogenesis. Recent research findings show that many cytokines and adhesion factors produced by osteoblasts and osteoclasts, such as granulocyte colony-stimulating factor (G-CSF), IL-1, IL-6, IL-7, CXCL12, and others can also in turn modulate the immune system by affecting the production, development, and maturation of various immune cells.^[12,13] The rapid development of osteoimmunology provides a new perspective on the pathogenesis of RA and brings new potions for targeted therapy of rheumatism. For example, humanized monoclonal antibodies against RANKL (Denosumab) have been successfully used to improve bone erosion progression in RA patients. In-depth understanding of the connections between bone and the immune system remains central to future research in this field. While cells are the basic structural and functional units of tissues, their complexity and heterogeneity determine the high degree of differences between various organs and systems. Therefore, it has become an important entry point for exploring the study of osteoimmunology in RA.

Figure 1: Crosstalk between immune system and skeletal system. Various immune and skeletal cells, cytokines, and signaling pathways, such as osteoblasts, osteoclasts, macrophages, T/B lymphocytes, the RANKL/RANK/OPG axis, TNF-α, IFNγ, BMP, interleukin family and S1P-S1PR1 signaling, and others, are involved in the same microenvironment and processes during bone erosion. RANKL: Receptor activator of nuclear factor-κB ligand;OPG: Osteoprotegerin; TNF-α: Tumor necrosis factor-α; IFNγ: Interferon-γ; BMP: Bone morphogenetic protein; S1P: Sphingosine-1-phosphate; S1PR1: S1P receptor-1.

In RA, dysregulated immune responses that stimulate bone erosion may be induced by only a small part of pathologically important cell population. Therefore, individual immune cells may be important for the pathogenesis and progression of the disease. Traditional technologies, such as flow cytometry, real-time PCR, Western blot, and others, help researchers to collect data based on bulk cell populations rather than individual cells, and thus fall short in comprehensively identifying cellular states and unraveling potentially novel biological discoveries in osteoimmunology. New technologies are urgently needed to define and identify cells and cytokine subpopulations involved in a series of essential biological processes in osteoimmunology, providing innovative ideas and potential therapies related to RA.

A single cell is the basic structural unit of life activities. It may display heterogeneous structure and function even in different tissues. Single-cell omics is a revolutionary tool in the field of modern biological research, comprising single-cell transcriptomics, single-cell proteomics, and single-cell epigenomics.^[14–17] They enable the display of the state and function of cells in various environments through profiling the genome, transcriptome, epigenome, and proteome from a single-cell resolution, providing a new direction for our understanding of cellular heterogeneity and phenotypes in multicellular systems. Single-cell omics has two core technical routes: one depends on nucleic acid capture amplification, reverse transcription, protein detection, and other molecular biology; and the other is cell separation by using automated high-throughput parallel technology.

Here we first introduce the development of single-cell omics technologies, especially two of the most commonly used ones: single-cell RNA sequencing (scRNA-seq) and single-cell resolution in assay for transposase-accessible chromatin using sequencing (scATAC-seq). Then we summarize some of the recent findings on the roles of various cells involved in the mechanisms of bone destruction in RA. Finally, we discuss the applications of such sequencing technologies in osteoimmunology research in RA [Tables 1 and 2; Figure 2].

Table 1 - Summaries of studies applying scRNA-seq to osteoimmunology research in RA.

Study subjects	Source of cells	Number of subjects	Cell analyzed	Findings	Method used	Date published	Reference
CIA mice and RA patients	Synovial tissue	RA patients (n = 23)	Murine CX3CR1hiLy6CintF4/8+I-A+/I-E+ macrophages (AtoMs)and human CX3CR1+HLA-DRhiCD11c+CD80-CD86+ cells	AtoMs are the osteoclast precursor-containing population in the pannus tissue, and FoxM1 constitutes a potential target for RA treatment.	scRNA-Seq	November 18, 2019	[56]
RA patients OA patients Healthy donors	Synovial tissue and blood	RA patients (n = 18) OA patients (n = 14)	HBEGF+ inflammatory macrophages	These HBEGF+ macrophages may enhance the tissue destructive capacity of fibroblasts.	scRNA-Seq	May 8, 2019	[58]
Early/active RA, treatment-refractory/active RA and RA in sustained remission, healthy donors	Synovial tissue	RA patients: naïve to treatment (n = 45), treatment-resistant (n = 31), in sustained clinical and ultrasound remission (n = 36)	MerTKpos TREM2high and MerTKposLYVE1pos macrophage	The relative proportions of these cells in RA may be a marker to predict the state of persistent remission vs. disease flare.	scRNA-seq, multiparameter Flow cytometry, immunofluorescent staining	June 29, 2020	[59]
Chinese subjects underwent hip replacement surgery	Bone marrow	n = 2	LEPRhiCD45lowBM-MSCs	These cells had the capacity to differentiate into osteocytes, chondrocytes, adipocytes, and terminal-stage quiescent cells. Some new markers for purification of human BM-MSC were identified.	scRNA-seq	October 11, 2021	[63]
Mice	Bone marrow	NA	Mesenchymal progenitors at different stages	We find a unique cell type that expresses adipocyte markers but contains no lipid droplets. They play critical roles in maintaining marrow vasculature and suppressing bone formation.	scRNA-seq	April 14, 2020	[64]
RA and OA patients	Synovial tissue	n = 51	T cells, B cells, monocytes, and fibroblasts	We identified 18 unique cell populations and defined distinct subsets of CD8+ T cells characterized by GZMK+, GZMB+, and GNLY+ phenotypes.	scRNA-seq, mass cytometry, bulk RNA-seq, and flow cytometry	May 6, 2019	[27]
RA patients	PBMC and Synovial tissue	ACPA+RAn = 10 ACPA-RA n = 10	CD45+hematopoietic cells	Up-regulation of CCL13, CCL18 and MMP3 in myeloid cell subsets of ACPA-RA; a lack of HLA-DRB5 expression and lower expression of cytotoxic and exhaustion related genes in the synovial tissues of ACPA-RA; DRB1/DRB5 conveys an increased risk of developing active disease in ACPA+RA	scRNA-seq, Immunohistoche mical staining	August 17, 2021	[87]
RA patients and healthy indifiduals	Peripheral blood	n = 4	RA-CCPPOS B cells	The expression of IL15R is enriched in citrulline-specific B cells within RA patients were capable of producing AREG	RNA-sequencing, aptamer-based SOMAscan assay, flow cytometry	February 23, 2019	[88]
RA patients	Synovial tissue	n = 5	Fibroblast	13 transcriptomically distinct clusters were revealed. Previously uncharacterized fibroblast subpopulations were identified and their spatial location within the synovium was discerned.	A 3D-printed, low-cost droplet microfluidic control instrument	February 23, 2018	[91]
RA and OA patients	Synovial tissue	NA	Synovium fibroblast	Seven fibroblast subsets with distinct surface protein phenotypes were identified.	Bulk transcriptomics of targeted subpopulations and single-cell transcriptomics	February 23, 2018	[92]
Mouse models of resolving and persistent arthritis	Synovial tissue	NA	FAPα+ synovial fibroblast	The location and function of two FAPα+ synovial fibroblasts.	Single-cell transcriptional technologies	May 29, 2019	[93]
Mice	Synovial tissue	NA	Fibroblast	Expression of the synovial lining markers gradually changes along the continuum of synovial fibroblast states. Endothelium-derived Notch signaling is a potential pathway for driving the differentiation of THY1 expressing fibroblasts.	Trajectory transcriptional analyses, single-cell RNA-seq datasets	July 2020	[94]
RA patients and mice	Synovial tissue	NA	Fibroblast	NOTCH3 signaling drives both transcriptional and spatial gradients in fibroblasts and plays an important role in inflammation.	scRNA-seq and synovial tissue organoids	April 22, 2020	[95]

BM-MSC: Bone marrow derived mesenchymal stem cell; CIA: Collagen-induced arthritis; NA: Not applicable; OA: Osteoarthritis; PBMC: Peripheral blood mononuclear cells;RA: Rheumatoid arthritis; scRNA-Seq: single cell RNA sequencing; TH1: T helper 1.

Table 2 - Application of scATAC-seq in osteoimmunology research in RA.

Study subjects	Source of cells	Number of subjects	Cell analyzed	Findings	Method used	Date published	Reference
RA patients and healthy donors	Synovial tissue	NA	FLS and CD14+ monocyte-derived macrophage	280 arthritic genes induced by TNF, 80 of which were in a tolerant state in macrophages. The regulatory elements associated with FSGs in TNF stimulated FLS showed a sustained chromatin activation state.	RNA-sequencing, H3K27ac, ChIP-seq, and ATAC-seq	April 23, 2019	[99]
RA patients	Synovium	n = 2	FLS	The proliferation and activation of FLS in RA patients were inhibited by JQ1 through inhibiting key BRD2/BRD4 superenhancer genes, downregulating multiple crucial inflammatory pathways, and altering the genome-wide occupancy of critical TFs involved in inflammatory signaling.	Transcriptional profiling, assay for transposable accessible chromatin by high throughput sequencing	December 7, 2020	[100]
Untreated RA patients and healthy individuals	Peripheral blood	RA patients (n = 7) Healthy individuals (n = 7)	PBMCs	10 key TFs contributed to RA pathogenesis via regulating the activity of MAP kinase and two genes (PTPRC and SPAG9).	scATAC-seq combined with GO and KEGG analyses	September 17, 2021	[101]
Human	Bone marrow and adipose tissue	NA	MSCs, stromal cells, and cell lines (pulmonary epithelial, endothelial hybrid, embryonic kidney epithelial, pancreatic cancer, and neuroblastoma cells)	A large and diverse transcriptional network of pro-osteogenic and antiadipogenic TFs was identified.	Machine learning algorithms	March 4, 2019	[103]
A healthy cohort aged 20–60 years	Bone marrow	NA	MSC	They found that enhancers and TFs can affect the cellular differentiation potential of aged BMSCs. Age change, in turn, can promote BMSCs' differentiation into tissues with related common traits, such as bone and adipose tissue.	ATAC-Seq, RNA-Seq, and proteomics studies were conducted, and ultimately an integrative multi-omics analysis was performed	September 5, 2021	[104]
Healthy individuals	Peripheral blood	n = 3	CD4+ T cells	Interactions of expression quantitative trait loci with target genes were demonstrated and showed complex interactions for 20% of RA associated loci.	ATAC-seq, Hi-C, Capture Hi-C, and nuclear RNA-seq5	September 2, 2020	[105]

BMSC: Bone marrow derived mesenchymal stem cell; FLS: Fibroblast-like synoviocyte; MAP: Mitogen activated protein; MSC: Mesenchymal stem cell; NA: Not applicable; PBMC: Peripheral blood mononuclear cell; RA: Rheumatoid arthritis; RNA-Seq: RNA sequencing; scATAC-seq: single-cell resolution in assay for transposase- accessible chromatin using sequencing; TFs: Transcription factors; TNF: Tumor necrosis factor.

Figure 2: Advances in RA osteoimmunology: application of scRNA-seq and scATAC-seq.^{[ 26 ]} BM-MSCs: Bone marrow derived mesenchymal stem cells; FLS: Fibroblast-like synoviocytes; RA: Rheumatoid arthritis; scATAC-seq: single-cell resolution in assay for transposase-accessible chromatin using sequencing; scRNA-seq; single-cell RNA sequencing; TFs: Transcription factors; TNF: Tumor necrosis factor; TPH: Peripheral helper T.

Recent Advances in Single-cell Omics Technology

In the last decade, the field of single-cell technologies and integrative analysis of multi-omics platforms have been evolving rapidly. Tens of different single-cell omics technologies have been developed successively, including scRNA-seq, proximity ligation assay (PLA), proximity extension assay (PEA), proximity ligation assay for RNA (PLAYR), cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq), RNA expression and protein sequencing (REAP-seq), chemical-labeling-enabled C-to-T conversion sequencing (CLEVER-seq), epigenetic landscape profiling using cytometry by time of flight (EpiTOF), nucleosome occupancy and methylome sequencing (NOMe-seq), scATAC-seq, high-throughput variant of chromosome conformation capture performed on single cells (scHi-C), single-cell chromatin overall omic-scale landscape sequencing (scCOOL-seq), single-cell methylome and transcriptome sequencing (scM&T-seq), single-cell nucleosome, methylation and transcription sequencing (scNMT-seq), single-cell triple omics sequencing (scTrio-seq), and others^[18–26] [Figure 3]; and each of these has its own strengths. With the help of these technologies, several collaborative projects have been initiated and greatly advanced research in the field of rheumatology.^[27,28]

Figure 3: Single-cell experimental platforms for omics analysis. Tens of different single-cell omics technologies have been developed to interrogate the transcriptome, epigenome, and proteome. Overlapping regions contain technologies that enable the integrative analysis of multiple omics in the same cells. CITE-seq: Cellular indexing of transcriptomes and epitopes by sequencing; CLEVER-seq: Chemical-labeling-enabled C-to-T conversion sequencing; EpiTOF: Epigenetic landscape profiling using cytometry by time of flight; NOMe-seq: Nucleosome occupancy and methylome sequencing; PEA: Proximity extension assay; PLA: Proximity ligation assay; PLAYR: Proximity ligation assay for RNA; REAP-seq: RNA expression and protein sequencing; scATAC-seq: Single-cell resolution in assay for transposase-accessible chromatin using sequencing; scCOOL-seq: Single-cell chromatin overall omic-scale landscape sequencing; scHi-C: High-throughput variant of chromosome conformation capture performed on single cells; scM&T-seq: Single-cell methylome and transcriptome sequencing; scNMT-seq: Single-cell nucleosome: methylation and transcription sequencing; scRNA-seq: Single-cell RNA sequencing; scTrio-seq: Single-cell triple omics sequencing.

scRNA-Seq is a powerful sequencing technology, which is easier to discover new types of cells in tissues, understand cell-specific markers and potential functions, as well as obtain gene expression profiles at the single-cell level. For example, by using scRNA-Seq technology, thousands of genes and variable shear points were detected in single mouse cleavage cells, which was impossible with conventional techniques.^[29] The general process of scRNA-Seq includes preparation of single cells, cell capture and lysis, reverse transcription and cDNA amplification, library construction, and sequencing, of which the two key steps are cell capture and cDNA amplification. Presently, several approaches have been proposed for amplification of cDNA: PCR-based amplification (including STRT-Seq,^[18] Smart-Seq,^[19] Smart-Seq2,^[30] etc.), T7-based in vitro transcription (IVT) amplification (such as CEL-Seq^[31] and MARS-Seq^[21]), and Phi29 DNA polymerase-based amplification (such as Phi29-mRNA amplification^[20]).^[32] Nowadays, scRNA-Seq is being widely used in many research fields, such as tumor, autoimmune disease, stem cells, neurobiology, etc.

scATAC-seq is a supplement to scRNA-seq, which provides a revolutionized method of insight into the epigenomic regulation of whole-genome gene expression profile.^[33] Tagmentation is the most crucial step of scATAC-seq. Simultaneously high-activity Tn5 transposase is employed to insert sequencing adaptors. This technique can identify both active and closed chromatin throughout the genome at the single cell level. In addition, it can also characterize the epigenome of cell subtypes within and across species. Currently, scATAC-Seq is mainly applied in hematologic diseases by analyzing human immune cells and hematopoietic stem cells. These research data have brought about many exciting discoveries, including mapping the epigenomic landscapes of hematopoietic stem cells and progenitor cells to uncover a regulatory network of immune cell differentiation during hematopoiesis,^[34] comprehensively characterizing the lineage-specific chromatin dynamics of hematopoiesis,^[35] and identifying disordered cis-elements and transelements related to malignant clonal expansion of immature T lymphocytes progenitors in T cell leukemia.^[36] So far, in rheumatic disorders, the scATAC-Seq data sets have been scare. Using bulk ATAC-seq methods, Dr. Capellini's team identified genes that control cartilage development and osteoarthritis in the course of evolution.^[37] Their work informed future scATAC-Seq studies, which have been used to analyze clinically relevant immune cells' subtypes and tissues in rheumatic diseases, like RA.

Mechanisms of Bone Resorption in RA

During vertebrate evolution, acquired immunity and the skeletal system appeared at the same stage. Since immune and bone cells develop in the same micromilieu, the relationship between them is complex and reciprocal.^[9,38] RA is a representative autoimmune-mediated arthritis with dysregulated synovitis and bone destruction. In general, chronic inflammation disrupts the physiological balance between osteoblasts and osteoclasts, resulting in excess bone resorption over bone formation. The bidirectional interplay of the immune and skeletal systems highlights RA as a wonderful model for exploring osteoimmunology. To uncover the detailed molecules and signaling mechanisms underlying bone erosion associated with RA, it is necessary to elucidate imbalanced bone metabolism in RA, especially the downstream signaling activated by RANKL–RANK interactions.

Osteoclasts are the exclusive cell type responsible for bone resorption; therefore, enhancement of their function is the key cause of bone damage in RA, which depends on a combination of increased cell number, enhanced activity, and extended lifespan. The monocyte/macrophage lineage is the main but not the only source of osteoclasts. Accumulating evidence has demonstrated that osteoclast could originate from osteoclast precursor cells, including not only bone marrow progenitors but also cells in the monocyte or dendritic cell differentiated pathways. This is mainly influenced by the signaling pathways and cytokines in the surrounding environment.^[39–43]

The RANK–RANKL axis is a key factor in osteoclasts' differentiation and survival. RANKL is a TNF superfamily member, expressed by a variety of cells, including osteoblasts, osteoclasts, synovial fibroblasts, T lymphocytes, B lymphocytes, and stromal cells.^[44] RANKL activates the signaling cascades culminating in osteoclast differentiation through its binding to the specific receptor RANK expressed on osteoclast progenitors. These signaling cascades, including TNF receptor-associated factor-6 (TRAF6), mitogen-activated protein kinases (MAPKs), and the transcription factor (TF) nuclear factor-κB (NF-κB), ultimately induced the activation of a crucial regulator of osteoclastogenesis, the TF nuclear factor of activated T cells 1 (NFATc1).^[9,45–47] On the other hand, OPG, a glycoprotein of the TNF receptor family, can competitively bind to RANKL to suppress its function and is therefore considered the major negative regulator of bone resorption.

In addition to RANKL, macrophage colony-stimulating factor (M-CSF) serves as another essential molecule for osteoclastogenesis by promoting proliferation of osteoclast precursors and activation as well as survival of osteoclasts.^[9,48] Macrophages also contribute to onset and progression of chronic inflammation in RA. Another critical component in joint, synovial fibroblasts, cause two major pathological lesions in the pathogenesis of RA, promoting inflammation and joint tissue destruction. They enhance osteoclastogenesis, degrade articular cartilage, and inhibit osteoblastic bone formation through complex interactions and mutual activation with immune cells.^[49,50]

This review focuses on the latest research findings on the mechanisms of bone erosion in RA pathogenesis, including the interplay between immune cell and bone, as well as the application of single-cell omics in these mechanisms.

Application of scRNA-Seq in Osteoimmunology Research in RA

Osteoclasts and macrophages

Osteoclasts, the only bone-resorbing cells, are myeloid lineage multinucleated giant cells originated from CD11b⁺ precursors of the monocyte-macrophage lineage. Their proliferation, survival, and differentiation are dependent on M-CSF and RANKL. Osteoclasts and osteoblasts not only play a vital role in maintaining the homeostasis of bone remodeling but also are involved in the immune activity.^[51–53] In the pathological conditions of RA, autoimmunity associated inflammation promotes full differentiation of osteoclasts to eventually destroy bone by serving as a suitable microenvironment.^[52,54,55] Identifying the origin of osteoclast precursors, their location in different tissues as well as their differentiation trajectory may help to clarify the mechanisms of bone destruction in RA. In recent years, single-cell omics has emerged as a powerful technology that is fully competent to address this need. Hasegawa et al^[56] found a subgroup of macrophages, CX3CR1^hiLy6C^intF4/8⁺I-A⁺/I-E⁺ macrophages (AtoMs), in the inflamed synovium of collagen induced arthritis (CIA) mice. This cell group could differentiate into osteoclasts with a high capacity, which was suppressed by deletion of Foxm1. These results suggested that scRNA-Seq could analyze cellular differences between osteoclast precursor cells at high resolution.

In addition to being able to differentiate into osteoclasts directly, macrophages can also induce disease initiation by affecting other cells involved in synovitis progression and bone erosion.^[57] The development of new technologies helps us to learn more about macrophage phenotypes that reside in RA synovium at single-cell resolution. In Kuo et al's^[58] study, a subpopulation of inflammatory macrophages marked by genes HBEGF were detected in RA synovium by application of single-cell platforms. The characteristics of these HBEGF⁺ macrophages were induced by fibroblasts residing in RA synovium, thus showing the effect of joint destruction. In turn, HBEGF⁺ macrophages may enhance the tissue destructive capacity of fibroblasts. In addition to their role in inducing disease development, recent studies also show that macrophages in synovial tissue are associated with disease remission and synovial homeostasis of RA. By using high-throughput technological approaches/single-cell transcriptomics, a more comprehensive picture of macrophages features was drawn. Alivernini et al^[59] identified MerTK^pos macrophage subpopulations in treatment-naïve and -resistant active RA patients. The relative proportions of these cells in RA may be a marker to predict the state of persistent remission vs. disease flare. Therefore, they may become the direction for potential treatment strategy.

Mesenchymal stromal/stem cells (MSC)

As an important part of the complex bone tissue, the detailed origins and underlying biology of osteoblasts, osteocytes, and chondrocytes are not characterized. MSCs are generally referred to as the cells that have the ability of multi-differentiation into osteoblast/osteocyte, chondrocyte, and adipocyte.^[60–62] In addition, since diminished bone forming activity by osteoblasts and increased marrow adiposity are participated in bone loss, understanding the nature of MSCs and their differentiation path have always been the hotspots. Recent single-cell studies have proposed some novel subtypes regarding cellular heterogeneity among bone marrow derived mesenchymal stem cell (BM-MSCs) in patients with osteoarthritis (OA) or OP. Wang et al^[63] are the first to systematically analyze human CD271⁺ bone marrow mononuclear cells (BM-MNCs) in OA and OP patients. In this cell population, a subgroup of LEPR^hiCD45^low BM-MSCs was identified. These BM-MSCs generally have several differentiation routes, including osteocytes, chondrocytes, adipocytes, and terminal-stage quiescent cells. Some new markers for purification of human BM-MSC were suggested when comparing their expression pattern with CD45^hi hematopoietic cells. In Zhong et al's^[64] study, a unique cell subpopulation expressing markers of adipocyte but devoid of lipid droplets was defined. They were considered as a new component of marrow adipose tissue for their crucial roles in suppressing bone formation. It is expected that in the near future, there will be single-cell studies on MSC in patients with RA.

T lymphocyte

A number of studies have uncovered the key role of T cells in RA pathogenesis.^[9,65,66] The current view is that activated CD4⁺ T cells infiltrate into the synovium, and enhance osteoclast differentiation and bone destruction, causing bone loss and systematic OP by secreting CD40L and various pro-inflammatory cytokines such as TNF-a, IL-6, IL-1, and IL-17.^[67] However, inflammatory T helper 1 (TH1) cells, previously thought to mediate the development of RA, strongly inhibit osteoclastogenesis by secreting IFNγ, a pro-inflammatory cytokine,^[68] thereby reducing bone erosion. In contrast, T cells from the synovium tissue of RA patients show the opposite depleted phenotype and produce low levels of IFNγ.^[11] Neither of these two T cells is conventionally considered an activated T cell. Therefore, it is significant to identify the subpopulation of CD4⁺ T cells inducing bone destruction in RA.

Previous studies have suggested that RANKL expressed by T cells can help to promote osteoclastogenesis, but it does not adequately explain the excessive osteoclastogenesis in RA.^[69] T-helper (Th) 17 and regulatory T-cells (Treg) have become main actors in the bone erosions in the pathogenesis of RA. An imbalance in activity and function between them may contribute to the onset of autoimmune diseases and associated bone destruction. Th17 cells are considered as the dominant T cell subgroup that induces differentiation of osteoclast precursors. They secrete IL-17 and other cytokines (such as IL-21, IL-22, and TNF) to induce proliferation of synovial fibroblasts and macrophages, resulting in expression of RANKL.^[70–73] In addition, Th17 cells were reported to play a role in not only the quantity but also the quality of antibodies in RA, which inhibit the expression of St6gal1 in B cell by producing IL-21 and IL-22 and induce the accumulation of desialylated IgG, thereby acquiring a higher ability to accelerate inflammation and osteoclastogenesis.^[74–76] Other T helper cells located in different tissues may play a role in humoral immunity and assist B lymphocytes to produce autoantibodies.^[77]

On the other hand, Tregs are major suppressors of the functions of Th17 cells.^[78,79] An important subgroup of Treg cells is the one expressing TF Forkhead box P3 (FOXP3), which is a critical factor for Treg function. These Foxp3⁺ Treg cells play an integral role in bone and immune homeostasis by affecting the differentiation and function of osteoclast.^[80] Single-cell omics may contribute to uncover the pathological mechanisms in RA. A particular type of TH17 cells, deriving from FOXP3⁺ T cells, has been found in synovial tissue of patients with active RA. Unlike conventional TH17 cells, this type of cells promotes osteoclastogenesis.^[81] Zhang et al^[27] identified T cells with different cellular states in synovial tissues of RA patients, including PDCD1⁺ peripheral helper T (TPH) cells, follicular helper T (TFH) cells, and new subsets of CD8⁺ T cells characterized by markers of GZMK⁺, GZMB⁺, and GNLY⁺. Pro-inflammatory cytokines TNF can be produced by T lymphocytes. These studies help to find new T cell subsets and their related cytokines, which may determine the osteoclast differentiation and induce inflammation in RA.

B lymphocyte

An interesting phenomenon was found, that healthy individuals who were positive for anti-citrullinated protein antibodies (ACPAs) but without arthritis lose bone mass systemically. In fact, humoral immunity, especially B lymphocytes, is involved in the triggering of bone injury.^[82–84] In the form of immune complexes, ACPAs are recognized by the Fcγ receptor or Fab fragments on osteoclast precursor cells, resulting in the secretary of TNF and IL-8, which in turn enhance osteoclast differentiation in an autocrine manner.^{[74,75,85,86]} In addition to the production of autoantibodies like ACPAs, B lymphocytes play roles in RA inflammation and bone erosion, including antigen presentation, secretion of pro-inflammatory cytokines like TNF and IL-6, and exertion of influence on bone metabolism by expressing RANKL. However, complete clarity is yet to be obtained concerning the immunologic differences between seropositive and seronegative RA, and the mechanisms by which the tolerance of ACPA-producing and RF-producing B cells is disrupted. With the help of scRNA-seq method, these differences and mechanisms were reported. Wu et al^[87] find that CCL13, CCL18, and MMP3 are upregulated in synovial myeloid cell subsets in ACPA- RA as compared with ACPA+ RA. Next, in Mahendra et al's^[88] study, they first demonstrate the expression of IL15Ra protein in B cells specific change in ACPA production. Then, they elucidated the mechanism by which the autoantibody ACPA induced bone destruction through combining amphiregulin.

Synovium fibroblast

Given synovitis is the crucial center in the pathogenesis of RA, the predominant cell component in synovium, synovial fibroblasts,^[89] are undoubtedly a research hot topic and promising therapeutic target. They are not only the main source of RANKL in arthritis^[69] but also producing inflammatory cytokines, chemokines, and matrix metallopeptidases that destroy bone and cartilage,^[90] and they also form a vicious circle of interaction with bone-damaging T cells. All of these indicate that synovial fibroblast are the key regulators of bone destruction. At present, there is a lack of clarity concerning the heterogeneity of fibroblasts phenotypes and the classification of different functional sub-populations. With the advent and development of scRNA-Seq technology, fibroblasts in synovial tissues can be deeply analyzed, thus revealing the existence of different subsets. From 2017 onward, single-cell RNA sequencing analysis was applied to identify a heterogeneous population of synovial fibroblasts from RA patients. It is shown that a subset of fibroblasts marked with CD34^-THY1⁺Podoplanin⁺cadherin^-11⁺ has been identified in the perivascular zone of the inflamed synovium, which has significant capacity of proliferation and invasiveness.^[91,92] From then, more and more studies on exploring the heterogeneity of synovial fibroblasts were performed. Two types of fibroblasts were identified to play different roles in the pathogenesis of RA by applying single-cell transcriptional technologies on cells isolated from inflamed synovium of mouse models and RA patients. One of them, FAPα⁺THY1^- destructive fibroblast, located in the synovial lining layer, was confirmed to destroy bone and cartilage selectively but had little effect on inflammation.^[93,94] Further studies applying scRNA-Seq analysis found their exact location in the synovium, their surface markers, and a potential target for therapeutic strategies, Notch3 signaling, for RA.^[95] Sublining fibroblasts have been considered a major source of pro-inflammatory cytokines and have therefore gained extremely high attention. In the recent study, Zhang et al^[27] applied several techniques, including scRNA-seq, to analyze synovial fibroblasts from RA patients. Four putative subpopulations of fibroblasts were identified: CD34⁺ sublining fibroblasts (SC-F1), HLA-DRA^hi sublining fibroblasts (SC-F2), DKK3⁺ sublining fibroblasts (SC-F3), and CD55⁺ lining fibroblasts (SC-F4). Among them, THY1(CD90)⁺HLA-DRA^hi sublining fibroblasts expanded in RA synovia and were the main source of inflammatory mediators IL-6.

With the constantly updated scRNA-seq technology, and its wide application in RA, the study of cell status, key populations of pathogenic cells, and their activation status in inflamed synovial tissue will be more in-depth. In combination with the information of single-cell multi-omics and spatial transcriptome, the pathogenesis of RA can be comprehensively analyzed from the gene, epigenetic, transcriptome, proteome, and tissue microenvironment levels, providing a theoretical basis for disease staging, discovery of new therapeutic targets, and determination of early diagnostic markers.

Application of scATAC-Seq in Osteoimmunology Research in RA

Epigenetics modification is an important aspect in RA etiology and pathogenesis research, which is mainly responsible for regulating gene expression and making a dynamic transcriptional response to inflammation effects. Accumulating evidence suggests that epigenomic and transcriptomic dysregulation is associated with the RA phenotype. In chromatin of eukaryotic genomes, some loose regions, including those containing active regulatory elements, play roles in autoimmune diseases by controlling gene activity.^[96] Advances in single-cell genomics research have made it possible to clarify the heterogeneity of cell types, providing novel strategies for the diagnosis and treatment of RA. By using ATAC-seq technology, we can analyze active regulatory elements including enhancers, promoters, and other regulatory sequences to maps. These regulatory elements can identify both active and open chromatin.^[97] Recently, ATAC-seq is applied in gene regulatory network of RA.

Synovial fibroblasts contribute to both inflammation and bone destruction in RA pathogenesis. However, this population of cells is heterogeneous in not only phenotype but also function. In a recent study, a series of methods including ATAC-seq were performed to explore this heterogeneity of synovial fibroblasts in RA. They identified six distinct synovial fibroblasts states under different conditions, suggesting that cytokine signaling in the inflamed synovium impacts synovial fibroblast status.^[98] Loh et al^[99] analyze the effect of TNF on genome-wide changes in fibroblast-like synoviocytes (FLS) in synovial tissue from RA patients as well as CD14⁺ monocyte-derived macrophage cell lines from healthy donors. They found 280 arthritic genes induced by TNF, 80 of which were in a tolerant state in macrophages. Further studies revealed that the gene regulatory elements that evade repression by FLS in TNF-induced RA showed a sustained chromatin activation state, and the phenomenon suggests that changing or targeting the chromatin states in FLS is a potential therapeutic strategy. In another study, Krishna et al^[100] found that the proliferation and activation of FLS in RA patients were inhibited by JQ1, an inhibitor of the bromodomain and extra terminal domain family. This discovery may provide new directions for the treatment of RA. Most recently, Zhang et al^[101] also demonstrated the value of using scATAC-seq in epigenetic studies in RA etiology. By using the ATAC-seq method combined with GO and KEGG analyses, they found that 10 key TFs contributed to RA pathogenesis via regulating the activity of MAP kinase and two genes (PTPRC and SPAG9), which help decode RA-related target genes and signaling pathway. As mentioned above, MSCs can differentiate into osteoblasts, chondrocytes, and adipocytes. Their differentiation routes were demonstrated in vitro to be mainly driven by a specific set of TFs by regulating a highly interconnected network of enhancers, including CEBPs, RUNX2, FOXO and PARP.^[102] Rauch et al^[103] identified a large and diverse transcriptional network of pro-osteogenic and antiadipogenic TFs. To uncover the effects of aging on TF-enhancer networks as well as mechanisms of MSC potential, Lai et al^[104] extracted primary human BMSCs from a healthy cohort aged 20–60 years for ATAC-Seq, RNA-Seq, and proteomics studies and ultimately performed an integrative multi-omics analysis. They found that enhancers and TFs can affect the cellular differentiation potential of aged BMSCs. Age change, in turn, can promote BMSCs differentiation into tissues with related common traits, such as bone and adipose tissue. Thus, it is demonstrated that the multi-omics approach provides new insights into the role of aging in bone and immune-related diseases. In another study on CD4⁺ T cells, Yang et al^[105] demonstrated interactions of expression quantitative trait loci with target genes and showed complex interactions for 20% of RA associated loci.

Therefore, this advanced technology, scATAC-seq, allows the profiling of DNA profiles in key cell types involved in RA pathogenesis, while identifying active and open chromatin. In this way, it is convenient for researchers to explore cell type-specific biological functions and biomarkers of disease, which can help to study cytokine and gene regulatory network and to better understand RA etiology.

Conclusions

Bone destruction triggered by inflammation is the most central link in the pathogenesis of RA. Osteoimmunology focusing on the crosstalk between immune and skeletal system is a hot topic in recent years, which provides a theoretical basis for the elucidation of the mechanism of bone erosion and the development of new therapies in RA. Single-cell omics is a revolutionary tool in the field of modern biological research, which has made significant contributions to transcriptomics and dynamics of specific cells involved in bone remodeling, improving our understanding of the pathophysiology of inflammatory bone erosion and helping provide valuable insights into the field of osteoimmunology. With the development of new technologies in single-cell omics, it is conducive to identify the dysregulated molecular mechanisms of bone destruction in RA as well as discover of potential therapeutic targets and biomarkers. The advances and optimization of these technologies are enabling the high-dimensional omics and our better understanding of the molecular signaling involved in the pathogenesis of rheumatic diseases.