Advances in single‐cell sequencing and its application to musculoskeletal system research

Abstract In recent years, single‐cell sequencing (SCS) technologies have continued to advance with improved operating procedures and reduced cost, leading to increasing practical adoption among researchers. These emerging technologies have superior abilities to analyse cell heterogeneity at a single‐cell level, which have elevated multi‐omics research to a higher level. In some fields of research, application of SCS has enabled many valuable discoveries, and musculoskeletal system offers typical examples. This article reviews some major scientific issues and recent advances in musculoskeletal system. In addition, combined with SCS technologies, the research of cell or tissue heterogeneity in limb development and various musculoskeletal system clinical diseases also provides new possibilities for treatment strategies. Finally, this article discusses the challenges and future development potential of SCS and recommends the direction of future applications of SCS to musculoskeletal medicine.


| INTRODUC TI ON
The basic unit of an organism is cell, and multicellular lives in nature begin with a single cell. Although preliminary estimates suggest that every person is composed of at least 37.2 trillion cells, 1 a deeper understanding is still very limited concerning the functions of those cells. Thus, the Human Genome Project (HGP) was officially launched in 1990 to unravel the genetic code of all of the approximately 25,000 genes in the human body and map the human genome atlas. 2 In 2001, the publication of a working draft of the human genome was considered a milestone in HGP. 3 The Human Cell Atlas (HCA) was launched in 2007 to further describe all human cells. 4 HCA integrated the information of cell types, number, location, relationship and molecular composition that facilitates to describe the cellular basis of health and disease. Gene sequencing is considered as a highly reliable method for analysing cell genetic attribution. 5 The earliest method of gene sequencing can be traced back to in 1977, 6 where 'Sanger sequencing' was a revolutionary technology, which ushered the era of gene sequencing research. Over the last several decades, due to its accuracy, 'Sanger sequencing' has been widely incorporated to scientific investigations such as HGP and diagnosis of clinical genetic diseases. 7 However, because of its complexity and high cost, investigators have been trying to develop more efficient sequencing methods.
Next-generation sequencing (NGS) is the second revolutionary innovation of traditional gene sequencing. 8 NGS has the characteristics of high throughput and low cost of per base, also known as massively parallel sequencing (MPS). 9 Further advances in computer science and technology have enabled development of thirdgeneration gene sequencing, with the ability of high-throughput, single-molecule sequencing. It can produce genome assemblies of unprecedented quality. 7 In 2009, Tang first reported a method for investigating the mRNA transcriptome using high-throughput sequencing in a single cell. This achievement is considered the beginning of widespread application of SCS technologies in scientific research. 10 Thanks to the application of SCS, cell biology and molecular biology had made great discovery in recent years. In 2018, researchers used SCS to create dynamic maps of gene expression during early embryonic development of zebrafish and frogs. Through integrating data on time scales in minutes to hours, describing the cells one by one, and tracking the eventual formation of embryo, investigators were able to build a complete map that revealed the entire developmental process from a single cell to an entire organism. 11,12 SCS can deepen our understanding of various aspects of cell function, such as tumorigenesis, 13 nerve degeneration, 14 immunology, 15 cell differentiation 16 and gene expression. 17 Limb development is a fundamental event in musculoskeletal system. Researches reveal that mesenchymal stem cells (MSCs) can act as bone progenitor cells in bone marrow. Skeleton development begin with the migration of mesenchymal cells derived from the embryonic lineage to the sites of future bone. 18 Further, researches reveal that transcription factor SRY (sex determining region Y)-box 9 (Sox9) plays a critical role in inducing osteogenic differentiation of MSCs. [19][20][21] Moreover, many morphogenetic or growth factors, such as WNTs, Hedgehogs, Notch, VEGF, FGFs, IGF-1, TGFβ and PTHRP, have been found to be involved in the regulation of endochondral bone formation. 22 For musculoskeletal disease research, trauma, pain and limb malformation are main issues. More specifically, bone fracture, joint injury, osteoarthritis and intervertebral disc degeneration are the common clinical diseases. The analysis of cell heterogeneity and activation of specific types of stem cells have become hot issues in this field. 23 Stem cells are found in the bone marrow and periosteum. The stem cell population is made up of heterogeneous cells. Previous studies have focused on histomorphology and gene expression regulation, which skipped the cellular level. Some important information is hidden in heterogeneous cell populations. By combining SCS and lineage tracing, a unique subpopulation of periosteum stem cell is identified to contribute to bone regeneration. 24 More importantly, the contribution of periosteal stem cells to bone regeneration is higher than that of bone marrow stem cells.
In this review, we first outline the major scientific issues and recent advances on the musculoskeletal system. Second, we discuss single-cell technologies in detail, including its technical features, superiority and its application in multi-omics studies. From our perspectives, cell heterogeneity has become the focus in limb development and clinical diseases of musculoskeletal system, and the application of SCS will provide us with unprecedented understanding of these issues. Moreover, the challenges and possible future directions of single-cell technologies are also discussed.

| Overview
Musculoskeletal system consists of bone, muscle, articulation, cartilage and other connective tissue that stabilize or connect bones ( Figure 1). 25 In addition to supporting the body's weight, bone and muscle work together to keep the body in position or to produce controlled and precise movements. 26 Bones provide structural support for the entire body and also protect the internal organs. Red bone marrow in long or flat bones lacunae produces blood cells. There are many attachments of muscle to the bones, which work as levers to change the magnitude and direction of the strength produced by the muscles. 27 Articulations play an important role in movement coordination. The stability and range of different articulations movement in the human body vary greatly, depending on the capacity of the associated muscles, tendons and ligaments. Cartilage is a kind of supportive buffering connective tissue. 28 Depending on its intercellular attribution, cartilage can be divided into three types: hyaline cartilage, elastic cartilage and fibrous cartilage. These three types of cartilage have different properties and functions in different parts of the body.

| Main scientific issues
The study of the musculoskeletal system is mainly divided into musculoskeletal development and clinical diseases. Chronic pain is the most common symptoms in orthopaedic clinics. In addition to injury-induced pain, degeneration and ageing of bones, cartilage and intervertebral discs (IVDs) are the main causes of pain. In the general population, bone ageing results in loss of volume and mass, usually manifested as osteoporosis and an increased risk of fracture. 29,30 Cartilage degeneration is common in joints. Degeneration of chondrocytes in the knee joint is a typical example. Ageing decreases cartilage thickness, which will lead to osteoarthritis (OA). Loss of IVD structural integrity can result in loss of IVD height, leading to collapse and compression of the spine and clinical symptoms, such as lower back pain.
Bone marrow MSCs can be stimulated to form osteoblasts, myocytes and fibroblasts. Change in the number or activity of these cells will affect a range of musculoskeletal tissues. Older MSCs exhibit a state of irreversible growth stagnation or senescence. Expression of p53, p21 and ageing-related β-galactosidase was increased compared to that in young MSCs, perhaps due to the down-regulation of age-related osteogenic genes, including Runx2 and osteocalcin. 29 Age-related changes in sex hormones can also affect normal bone biology. For example, postmenopausal women have gradually decreased oestrogen, while osteoclasts being released in response to oestrogen inhibition, increasing overall bone resorption. 31  Under these conditions, the ECM becomes granular, cracked or torn.
The structural strength of the annulus fibrosus (AF) decrease makes it easy for the nucleus pulposus (NP) to excrete. 32,33 In addition, malformation is a common concern of musculoskeletal system. The representative disease is scoliosis, among which adolescent idiopathic scoliosis (AIS) is the most common clinical type, occurring in 0.5%-3.0% of children. 34 It is currently believed that AIS is a multi-factor disease with genetic predisposition. 35 However, the chromatin-level pathogenesis of scoliosis remains uncertain.

| Recent progress
High-throughput sequencing technologies have been widely used in musculoskeletal system. These applications have led to many dis- to locate loci were 1p36.32, 2q36.1, 18q21.33 and 10q24.32 associated with AIS in Han Chinese girls. 41 The discovery that these genes F I G U R E 1 An overview of the research issues of the musculoskeletal system are closely related to bone growth and osteoblastic differentiation provides new insights into the genetic causes of AIS.  48 and microfluidics. 49 Notably, microfluidics involves fewer step and has the advantages of high analytical sensitivity and specificity as well as high throughput. Drop-seq is a representative method to capture single cell by microfluidics. 50 This method uses barcode beads ( Figure 2). The oligonucleotides on the beads include handle primer for PCR amplification, cellular barcode UMI (unique molecular identifier) that recognizes all oligonucleotides in a single cell and oligonucleotides (Oligo dT) that capture single-cell mRNA molecules.

F I G U R E 2
Basic procedures of single-cell transcriptome sequencing (scRNA-seq) and three common single-cell sorting platform. ScRNA-seq generally involves five basic procedures: single-cell isolation, reverse transcription, cDNA amplification, library preparation and high-throughput sequencing. Drop-seq, 10X Genomics and BD Rhapsody are the most widely used single-cell sorting platforms. They use microparticle beads with sequences linked by different cell barcodes, UMI and Oligo dT. These beads can capture mRNAs with poly A tails for reverse transcription and cDNA library preparation. Finally, all transcriptome information of a single cell is obtained by high-throughput sequencing Based on Drop-seq, 10X Genomics was developed as a widely used SCS library preparation platform. 10X Genomics is similar to Drop-Seq method of capturing single cells, but it is integrated into an oil droplet-based microfluidic device that effectively captures a larger cell count with higher sensitivity. 51 In addition, BD Rhapsody is the another widely used SCS platform. Instead of microfluidic channel emits cells and collides with the outgoing magnetic beads, BD Rhapsody uses CytoSeq unique cellular panel. 52  SPLiT-seq. 54 The cell acts as a reaction chamber, immobilizing the cell or nucleus, allowing efficient sample reuse. This is a low-cost method that does not require special customized instruments and enables the use of single-cell technology in a greater range of laboratory research.

| Single-cell multi-omics
The emergence of single-cell omics has provided unprecedented  was the core basis of Frisén's study. Because ATAC-seq is more repeatable and easier to operate than traditional methods, it has become the preferred method to study chromatin accessibility.

| Limb development
Bone and muscle are the most principal organs of musculoskeletal system. In the past, researches on skeleton biology had met with several obstacles. The chemical methods used for treatment of highly cellular tissues such as brain, liver and kidney, may not be applicable to skeleton tissue, 78  Several additional studies have focused on whole limbs. Because tetrapods are good models for studying the genetic and molecular basis of vertebrate pattern formation, limb development in tetrapods has received extensive research attention. 90 The genome determines development of the limb. 91  after treated as well as normal forelimbs of salamanders, then conducted scRNA-seq on 938 extracted cells. 95 They found a cluster of regenerative cells that is characterized by a significantly higher number of mitochondria than normal limb tissue cells. A novel COL2mito subcluster is further defined as COL2 + cells, which perform energy metabolite-related functions, such as responded to oxygen levels and ATP metabolism.
Collectively, these findings demonstrate the ability of scRNA-seq to isolate populations of developing limb cells at the molecular level.
The discoveries of the trajectory of limb development also provide new insights into tissue regeneration in the musculoskeletal system.

| Osteoarthritis
OA is a chronic degenerative disease closely related to ageing and progressive joint dysfunction. 96 The main pathological change of OA is a disorder of articular cartilage homeostasis, accompanied by inflammation and degradation. 97 Understanding the role and degeneration mechanism of chondrocytes will be helpful to development of  Monocyte samples RNA-seq data suggested that genes associated with SC-M1 (IL1B + proinflammatory monocytes) are significantly up-regulated in leukocyte-rich RA samples. In contrast, the marker genes associated with SC-M2 were down-regulated in OA. In other words, the leukocyte-rich RA synovium had greater numbers of IL1B + monocytes and IFN-activated monocytes but lower numbers NUPR1 + monocytes than the OA synovium. These data suggest that cytokine activation drives the expansion of the unique mononuclear population in the synovial membrane of active RA.

| CHALLENG E S AND FUTURE DIREC TIONS
SCS is an emerging technology with great promise, but also with challenges of its application to scientific research. First, the biggest issue is cost, which can easily cost more than $1,000 per sample, including cell capture and library preparation. Second, extraction of single cells is the starting step of SCS. With stroma-rich bone tissue, it is a difficulty to isolate sufficient amounts and qualities of RNA, 115  is the preferred method for genomic DNA amplification, specifically commonly used in low DNA quantities clinical samples. 116 However, MDA has the defects of amplification bias and unbalanced genome coverage. 117 In addition, the current data algorithms used in computer bioinformatics still need further development, and new methods are needed to standardize the identification of new cell types. 118 Professional analysis of the vast amount of sequence data is also a major challenge. Despite these challenges, SCS still has excellent potential. For example, single-cell technology can be used to perform multi-layer analyses of the tumour cell genome and transcriptome, which has the potential to revolutionize methods of understanding tumour growth, and this potential exists at every stage of disease development. As we have discussed in this article, SCS has unique superiority to analyse cell heterogeneity in many contexts. 119 From this perspective, SCS can be applied to musculoskeletal research to identify cell heterogeneity between stem cells populations, 44 which can help the design of methods to induce specific stem cells to play a role in tissue repair in musculoskeletal diseases.

| CON CLUDING REMARK S
In 2018, 'development cell by cell' was named 'the top one scientific breakthrough of the year' by science. 120 SCS is the core technology of the researches. Public recognition of this growing and innovative technology shows that SCS has stimulated great changes in scientific research. Over the past decade, investigators have improved or simplified several procedures resulting in drastic reduction in costs.
Thanks to the unique advantages of SCS, it is proved to be a technology with broad applications, yielding valuable data on microorganisms, tumorigenesis, and brain or nervous system. As we have seen, this technology has also enabled great advances in the study of heterogeneity of musculoskeletal system (Table 1)

ACK N OWLED G EM ENTS
This study was supported by grants from the Nature Science

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
ZY wrote original manuscript. WJ, YC, XK, YB, ZY, YL, WC and HX reviewed and edited. CQ, SL, LF and LC involved in conceptualization, and reviewed and edited. All authors reviewed the final version of manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.

Rheumatoid arthritis
ScRNA-seq CEL-Seq2 Definition overabundant stromal and immune cell populations in RA Zhang et al. 28