Unveiling heterogeneity in MSCs: exploring marker-based strategies for defining MSC subpopulations

Mesenchymal stem/stromal cells (MSCs) represent a heterogeneous cell population distributed throughout various tissues, demonstrating remarkable adaptability to microenvironmental cues and holding immense promise for disease treatment. However, the inherent diversity within MSCs often leads to variability in therapeutic outcomes, posing challenges for clinical applications. To address this heterogeneity, purification of MSC subpopulations through marker-based isolation has emerged as a promising approach to ensure consistent therapeutic efficacy. In this review, we discussed the reported markers of MSCs, encompassing those developed through candidate marker strategies and high-throughput approaches, with the aim of explore viable strategies for addressing the heterogeneity of MSCs and illuminate prospective research directions in this field.


Introduction
MSCs (mesenchymal stem/stromal cells) are heterogenous cell populations, residing in various tissues (such as bone marrow, umbilical cord, teeth, adipose, and so on).Differing from other types of adult stem cells or terminal differentiated cells, the main function of MSCs is sensing and responding to micro-environmental disturbances.Due to their innate characteristics and functions, they have multiple ways to respond to micro-environmental changes, such as extracellular matrix modification, recruiting other cells (the immune cells, for example), secreting small factors with various functions (immune modulation and regeneration, for example) [1].It is wellknown that many diseases are resulting from micro-environment dysfunctions.Therefore, the MSCs have been intensively and extensively applied in treating different kinds of diseases.Both pre-clinical and clinical investigations have shown that the MSCs hold great promise in developing one new therapeutic approach for treating many kinds of diseases [1][2][3][4][5].
Soon after the first demonstration of MSCs, its therapeutic applications have been investigated for decades.Unfortunately, in contrast to the rapid growth of clinical trials, few of them eventually have been developed as applicable therapeutic products.In addition to other factors inducing the therapeutic inconsistency of MSCs, cell heterogeneity is one tough challenge in their way to achieve the expected clinical outcomes [1, 2,4,6,7].
The heterogeneity of MSCs is reflected in different levels, such as the molecular levels (transcriptomics, proteomics, secretomics, and epigenomics), and the function levels (tri-lineage differentiation potentials, immunomodulatory capabilities, and regenerative activities) [2,7,8].The heterogeneity of MSCs could be induced by various factors including the donor conditions (age, gender, health condition, genetic background, and so on), tissue origin, and the strategies to isolate and expand the MSCs (digestion enzyme, matrix protein, cell culture medium, passage number, and so on) [2,6,7,[9][10][11][12] (Fig. 1).The causes of MSC heterogeneity have been extensively described in the preceding reviews, and we will not delve into them further.Pluripotent stem cell derived MSC, which can avoid the heterogeneity induced by the aforementioned factors, is also an important category of MSCs [13,14].However, we will focus exclusively on MSCs derived from somatic cells under natural conditions in this review.
Among different strategies to reduce the heterogeneity and improve the therapeutic consistency of MSCs, purifying the homogenous MSC subpopulations is suggested to yield more consistent clinical outcomes [6].MSC subpopulations refer to distinct groups or subsets within the broader MSC population that are identified based on specific characteristics or markers.These characteristics can include surface protein expressions, functional properties, gene expression profiles, or responsiveness to different environmental cues.According to the minimal criteria for defining MSCs, stated by the International Society for Cellular Therapy in 2006 [15], 55 MSC markers have been identified so far from different tissues and species (Fig. 2, Table 1).
Function enrichment by GO (Gene Ontology) analysis indicates that these MSC markers mainly regulate the process of leukocyte migration, wound healing, cell chemotaxis, and so on (Fig. 3A).Although some markers are involved in multiple functions, some of them are also cross-interacted in a network way (Fig. 3B).KEGG (Kyoto Encyclopedia of Genes and Genomes) also indicates that these MSC markers are mainly involved in the signal pathways in PI3K-AKT, adhesion, and so on (Fig. 4A).Similar to the GO analysis, some markers regulated multiple pathways (Fig. 4B) and they are crossinteracted (Fig. 4C).Most of these MSC markers are localized on the cell membrane, which is suitable for cell purification with FACS (fluorescence-activated cell sorting) and MACS (magnetic-activated cell sorting), while some of them are also intracellularly or extracellularly localized (Table 2).Normally, MSCs enriched with specific makers have functional advantages (Table 2).However, in some cases, these enriched MSCs also have some disadvantages (Table 2).
There are various strategies available for biomarker discovery, and among them, two classic approaches stand out: the candidate biomarker strategy and the highthroughput screening strategy.The candidate biomarker strategy is based on existing biological knowledge, where one or more molecules or features possibly related to a specific disease or biological process are selected as candidate biomarkers.These candidates are then experimentally validated for their expression levels or variations under different conditions.This strategy relies heavily on a profound comprehensive understanding of medical domains and relevant biological processes [16].In contrast, the high-throughput screening strategy employs techniques like genomics, transcriptomics, proteomics, lipidomics, and metabolomics to simultaneously analyze a large number of molecules and features.Through these techniques, it becomes feasible to detect thousands of molecules, facilitating the comparison of differences between heterogenous cell populations.Notably, this approach allows for the identification of biomarkers associated with specific functions without being reliant on prior knowledge [16][17][18][19][20]

MSC Heterogeneity
Fig. 1 Illustration of factors inducing MSC heterogeneity and potential solutions.The MSC heterogeneity results from various factors, including donor conditions (age, gender, health conditions), tissue origin (bone, fat, placenta/umbilical cord, teeth), and the methods employed for isolating (plastic adherence, MACS, FACS) and expanding MSCs (2D, 3D bioreactor, 3D matrix).To address MSC heterogeneity and enhance their therapeutic stability, three primary strategies are currently employed.These strategies encompass standardizing the MSC production procedures and purifying MSC subpopulations by markers.MSC mesenchymal stem/stromal cell, MACS magnetic-activated cell sorting, FACS fluorescence-activated cell sorting, 2D 2 dimensional, 3D 3 dimensional these strategies can be combined to enhance the comprehensive development of potential biomarkers.This integrated approach harnesses the biological knowledge of the candidate biomarker strategy while utilizing the technical capabilities of the high-throughput strategy to discover biomarkers associated with distinct functional subgroups in a more comprehensive and precise manner [16,[21][22][23][24].In the realm of MSCs, with the use of prior knowledge and high-throughput technologies such as single-cell RNA sequencing (scRNA-seq), specific markers related to different functional subsets of mesenchymal stem cells can be more comprehensively and accurately mined [17][18][19][20][21][22][23][24].
Therefore, in the current review, we would discuss the MSC markers that have been identified so far.Furthermore, based on the identification approaches, these markers have been categorized into two groups: the 1st generation of MSC markers, which has been identified by the candidate biomarker strategy; and the 2nd generation of MSC makers, which has been identified by highthroughput screening approaches (Table 2).

Techniques of identifying mesenchymal stem cell subpopulations
In most studies reviewed in this paper, Flow cytometry and fluorescence-activated cell sorting (FACS) is predominantly utilized for sorting MSC subpopulations.FACS are the primary methods for identifying MSC subpopulations, celebrated for their precision and versatility in scientific research.These technologies use fluorescently labeled antibodies to target specific surface markers, allowing researchers to conduct multiparameter analyses [25,26].This facilitates simultaneous assessment of various markers and functional properties within MSC populations, aiding in the identification and isolation of distinct subpopulations based on differential expression of markers.Such detailed analysis provides crucial insights into MSC heterogeneity.
Another vital technique, immunomagnetic cell sorting, utilizes magnetic beads tagged with antibodies targeting specific surface markers for selective isolation of MSC subpopulations [27,28].This method ensures high specificity and efficiency, essential for distinguishing and harvesting functionally diverse MSC subsets.
Additionally, functional assays are integral for understanding the biological characteristics of MSC subpopulations.Immunomodulatory assays, for instance, involve co-culture setups with immune cells to evaluate MSCs' effects on immune cell proliferation, activation, and cytokine production [29][30][31][32][33].These studies highlight the potential therapeutic uses of distinct MSC subsets in treating immune-related conditions.Differentiation assays, including those for osteogenic, adipogenic, and chondrogenic pathways, further elucidate the multilineage potential of MSC subpopulations, critical for identifying suitable cell sources for tissue engineering and regenerative medicine.Gene expression profiling, through techniques such as RNA sequencing, provides deep insights into the transcriptomic landscapes that define specific functional states or lineage commitments within MSC populations [17-19, 23, 34].These analyses help pinpoint molecular signatures characteristic of unique MSC subsets, enhancing our understanding of their heterogeneity.
Together, these techniques not only facilitate a comprehensive analysis of MSC heterogeneity but also specialize in pinpointing distinct MSC subpopulations.By employing these advanced methodologies, researchers can effectively characterize the diverse functional capacities and biological properties inherent to each subpopulation, significantly enhancing the precision of mesenchymal stem cell-based therapeutic strategies and the development of personalized regenerative medicine.

Immune suppression related markers
Although the MSCs have been widely investigated in the animal models of different diseases, the only approved clinical product of MSCs is for GVHD (Graft Versus Host Disease) treatment in clinics [35,36], because of their immune suppression capabilities.The immune modulation activity is one of those important contributors to the therapeutic effects of MSCs [1].

Extracellular secreted modulators
It has been demonstrated that TNFAIP6 (Tumor Necrosis Factor Alpha-Induced Protein 6) is a potential cell marker for mouse MSCs, irrespective of tissue origin and laboratory origin, with higher immune suppression activities and improved therapeutic effects [12].However, the membrane expression level of TNFAIP6 is significantly lower than its cytoplasm level [12].Indeed, TNFAIP6, also known as TSG6, is a small secreted protein with extracellular matrix remodeling and immunomodulation functions [37].On the other hand, the importance of these secreted modulators, such as the TNFAIP6 having been characterized as one efficacy predictor of MSCs in treating inflammation in vivo [38], makes it necessary to develop novel strategies to purifying these MSC subpopulations for improving their therapeutic effects.

Other immune regulators
CD200 is an immune suppressor and promotes peripheral immune tolerance [59,60].Its immune suppression function works through binding to its receptor CD200R, which then activates multiple pathways, such as MAPK-ERK, p38 MAPK, and JNK, via Dok and p120-RasGAP [61], resulting in upregulating the downstream effectors including IDO (indoleamine-2,3-dioxygenase), TGF-β, and IL-10 [59].A higher expression level of CD200 in MSCs correlates with enhanced immune suppression activities in vitro and in vivo [62].CD200 expressed on MSCs recognizes and binds to its receptor CD200R, which is expressed on myeloid progenitors, resulting in myeloid differentiation inhibition and immune suppression [63].CD200 + MSCs have much higher levels of colony-forming activity [51].However, it has been demonstrated that the expression of CD200 is undetectable in MSCs derived from umbilical cord blood [64], or very low in MSCs derived from adipose [65].In contrast, MSCs derived from the umbilical cord express higher levels of CD200 [65].Interestingly, the pro-inflammatory cytokine IFN-γ upregulates the expression of CD200 in MSCs derived from bone marrow but not adipose or umbilical cord [65].
BST2 (bone marrow stromal cell antigen 2), also known as CD317, is a type of transmembrane glycoprotein involved in virus reproduction suppression and immune regulation [66].Using the hTERT immortalized human bone marrow MSC colonies, it has been demonstrated that the MSCs from the CD317 + colony have increased cell areas and up-regulated mRNA levels of immunosuppressive genes than the CD317 − MSCs in vitro [67].Furthermore, CD317 + bone marrow-derived MSCs have better regeneration capabilities than the CD317 − MSCs [68].However, fresh CD317 − MSCs isolated from human bone marrow have better immune suppression activities but not CD317 + MSCs [68].However, our unpublished data show that CD317 + MSCs isolated from the human umbilical cord and expanded with chemically defined media have better immune suppression capabilities (unpublished data).
CD274, also known as PD-L1 (programmed death ligand 1), is a type I transmembrane protein and is widely expressed on multiple types of cells, such as lymphocytes [69].Its expression can be induced by pro-inflammatory cytokines, such as interferon-γ (IFN-γ), TNF-α, and IL-17 [69,70].And it has strong immune suppression activities through binding to its receptor PD-1 [71].It has been demonstrated that PD-L1 is expressed in MSCs [70,72,73].PD-L1 + MSCs have enhanced immune suppression activities and improved therapeutic effects on the collagen-induced mouse model of arthritis [74].
MX1, for 'myxovirus resistance' , is the gene responsible for virus immunity and an important component of interferon pathway [75].It has been demonstrated that Mx1 + MSCs are clonogenic at the single-cell level and have tri-differentiation abilities [32].Although its antivirus mechanism remains unsolved, the Mx1 + MSCs might also have immune regulatory functions.

Cell adhesion related markers
In addition to the important role of MSCs in modulating immune responses [1], another critical function is regulating cell adhesion, including both the cell adhesion and migration of MSCs, as well as the recruitment and adhesion of other types of cells, such as lymphocytes.

Mediating cell migration
CD44 is an important adhesion molecule involved in recruiting immune cells or stem cells into the inflammatory or injured tissues, via interacting with hyaluronic acid (HA), which is expressed in the injured/inflammatory sites [76,77].Their interactions induce conformational changes of CD44, recruit adaptor proteins, and lead to cytoskeletal rearrangement, resulting in the activation of various signaling pathways that involve cell growth, adhesion, and migration [76,77].In addition, CD44 also functions as a co-receptor to regulate the activities of other receptors, such as VEGFR, EGFR, FGFR and PDGFR [78].CD44 is widely expressed in multiple types of cells, including MSCs, and it also contributes to MSC recruitment [79,80].Its expression level is further induced by PDGF [79].The migration and adhesion of MSCs depend on CD44-HA (hyaluronic acid) interaction [79,80].Therefore, CD44 is a potentially important cell surface marker for MSC purification [81].However, later investigations indicate that freshly isolated mouse/human MSCs derived from bone marrow express very low levels of CD44 [82,83].MSCs show enrichment in the CD44 − fractions, as evidenced by their marker expression, colony-forming capacity, and in vitro differentiation abilities [82,83].Interestingly, CD44 is gradually up-regulated during cell expansion, even for the CD44 − fractions of MSCs [82,83].Thus, the CD44 expression levels after in vitro expansion, may not reflect their original cell identity [82].The CD44 + MSCs have enhanced colony-forming capacity and differentiation abilities [84].MCAM (melanoma cell adhesion molecule), also known as CD146, is involved in cell-ECM (extracellular matrix) interactions [85,86].Upregulation of CD146 could switch cell-cell adhesion to cell-ECM adhesion by interacting with its ligands in the ECM, preparing cells for migration and invasion by secreting related cytokines and proteins [85,86].CD146 is expressed in many cell types, especially in those cells constituting blood vessels, such as endothelial cells [86] and MSCs [87][88][89][90][91].
SDC2 (Syndecan-2), also known as CD362, is a type of transmembrane heparan sulfate proteoglycan, involved in modulating cell adhesion, proliferation, migration, and apoptosis through its interactions with the extracellular matrix and various proteins, such as proteases and cytokines.These interactions induce downstream pathway activations through intracellular protein partners [104].CD362 is mainly expressed in MSCs [104].CD362 + MSCs have enhanced colony forming, immune suppression and regeneration activities [105][106][107].Furthermore, both Phase 1 and Phase 2 clinical studies show that CD362 + MSCs are safe, feasible, and effective in treating COVID-19 infections [108].

Mediating lymphocyte adhesion
VCAM1 (vascular cell adhesion molecule 1), also known as CD106, mediates cell-cell adhesion and plays an important role in mediating the rolling, adhesion, and migration of circulating lymphocytes on the endothelium under inflammatory conditions [109][110][111].The CD106 is induced by pro-inflammatory cytokines in MSCs [112], and is involved in the immune suppression function of MSCs [113].CD106 + MSCs derived from placenta and umbilical cord have stronger abilities to suppress immune responses [112,114] and better pro-angiogenic activity, with enhanced promoting endothelial cell proliferation and migration [28,115].Furthermore, CD106 + MSCs have enhanced homing capacity [28,112].ITGA1 (integrin subunit alpha 1) is identified in the very late stage of activated T cells.ITGA1 is the major component of the ECM by binding to collagens (mainly collagen I and IV) and laminin, supporting the migration and activation of leukocytes, such as T cells, NK cells, NKT cells, and monocytes, especially the long-term activated or resident T cells [116,117].The ITGA1 has been proposed as an MSC marker for human bone marrow [118][119][120].The ITGA1 + MSCs have stronger colonyforming activity [118,119].
CD9, also known as MRP1 (motility related protein-1), is widely expressed in many cell types, including MSCs and lymphocytes, and is involved in regulating cell migration and invasion through integrin receptors [121,122].It has been demonstrated that CD9 is involved in the recognition and binding between MSCs and lymphocytes [123].CD9 + human MSCs have higher NOS (nitric oxide synthase) expression, proliferation rate, colony formation ratio, and stronger cell adhesion capability, resulting in better engraftment and improved therapeutic effects in the mouse model of hindlimb ischemia [124,125].

Other adhesion molecules
THY1, also known as CD90, is a small membrane protein located in the lipid raft [126].Although CD90 does not have an intracellular domain, it is involved in cell adhesion, migration, proliferation, and apoptosis through modulating the cell-cell and cell-matrix interactions via binding to its ligands, such as integrins, syndecan, CD90 and CD97 [126,127].CD90 has been identified as an important marker for MSCs from different species and tissues [128][129][130][131][132], and could be a potential marker for predicting the immune suppressive function of MSCs [133,134].Later studies also indicate that CD90 + MSCs have a faster proliferation rate and better differentiation capabilities [135][136][137].However, CD90 is also expressed in the fibroblasts, which might induce fibrosis [138].

Ephrin receptors
The Ephrin receptors (EphA and EphB), which can be recognized by ephrin ligands, play an important role in modulating multiple cellular functions, such as the selfrenewal and differentiation of stem cells [153][154][155][156][157][158].Proteomics studies indicate that EphA2 is expressed in the MSCs from human bone marrow and umbilical cord, and regulates the functions of MSCs [159,160].Follow-up studies showed that EphA2 could be a cell surface marker to distinguish MSCs from fibroblasts [161].Furthermore, EphA7 + MSCs proliferate faster and have higher levels of colony formation and differentiation capabilities [162].And EphB2 + MSCs have improved intestinal homing abilities and promoted the intestinal stem cell regeneration [31].It has been demonstrated that Eph/ephrin pathway is also involved in the cell migration of MSCs [163][164][165][166], and is essential for suppressing the proliferation of activated T cells by MSCs [167].

PDGFR
PDGFR (platelet-derived growth factor receptor), including PDGFRA and PDGFRB (also known as CD140α and CD140β, respectively), plays an important role in embryonic development and organogenesis, particularly in regulating the proliferation, migration, and differentiation of MSCs in various organs [168][169][170][171].Although both CD140α and CD140β have been identified as MSC markers [91,168,169], their investigation also indicates that CD140α is the negative selection marker for human MSCs derived from bone marrow, which differs from mouse MSCs [172].

Wnt pathway
FZD9, also known as CD349, is a receptor for Wnt ligands and activates β-catenin signaling pathway, which is involved in embryonic development and stem cell renewal [173,174].It has been demonstrated that CD349 is expressed in MSCs from both bone marrow and placenta, and proposed as a feasible marker for MSC isolation [175,176].Although both CD349 + and CD349 − MSCs show similar levels of MSC marker expression and differentiation abilities, the CD349 − MSCs have better neovascularization abilities than the CD349 + MSCs [177].
ROR2 is a tyrosine kinase-like orphan receptor, which can be activated by Wnt5a and regulate the tissue polarity and cell movement through downstream WNT/ PCP (planar cell polarity) signaling pathway [178,179].It has been demonstrated that ROR2 + MSCs derived from human bone marrow have enhanced chondrogenic differentiation efficiency [30].

Others
ALDH (aldehyde dehydrogenase) belongs to the metabolic enzyme family, which is involved in regulating glycolysis/gluconeogenesis and the detoxification of aldehydes via oxidation [180,181].It plays an important role in cell survival, proliferation, differentiation, and has been characterized as a classical stem cell marker [180,181].In human adipose tissues, the ALDH High MSCs represent a more primitive subpopulation than the ALDH Low MSCs, from the perspectives of cell proliferation and tri-differentiation capabilities [182,183].
STRO-1 can bind to an uncharacterized cell surface antigen, and identify around 10% of mononuclear cells in the human bone marrow [184].Purified STRO-1 + cells from human bone marrow have higher levels of colonyforming activity, proliferative rate, multilineage differentiation capabilities, and immune suppression activities by expressing higher levels of immune inhibitory factors (IL-8, LIF, IDO, HLA-G, VCAM1, TGF-β, and IL-10) [185], suggesting that STRO-1 is a potential MSC marker [184,186].Later study showed that STRO-1 + MSCs have better homing activities than STRO-1 − MSCs in the bone marrow, spleen, muscle, liver and kidney, while STRO-1 − MSCs are more prone to be trapped in the lung [187].
TLX1, also known as Hoxa11, belongs to Hox gene family which is essential for patterning during embryonic development.It has been demonstrated that the Hoxa11lineage marked (Hoxa11-eGFP) could identify the multipotent MSCs in the mouse bone marrow [197].Hoxa11 + MSCs have better colony forming potentials and tri-lineage differentiation abilities [198,199].
ISLR, also known as Meflin, is a glycoprotein (cell membrane located or secreted) with anti-fibrosis functions through interacting with BMP7 (bone morphogenetic protein 7) and inhibiting TGF-β pathway and myofibroblast differentiation [209].It has been demonstrated that Meflin is one MSC marker, and its expression positively correlates with its differentiation efficiency [210,211].
Sca-1 (stem cell antigen-1) has been characterized as a common marker in multiple types of stem cells, such as hematopoietic stem cells and MSCs [212,213].It has been demonstrated that mouse MSCs derived from bone marrow and ear express high levels of Sca-1 [214][215][216].
The SDF1-CXCR4 is the major pathway responsible for cell recruitment and retention [254,255].CXCR4 is expressed in human MSCs and contributes to the MSCs homing process [256][257][258][259][260]. For example, in the mouse model of osteogenesis imperfecta, the human MSCs migrate into the bone marrow through the SDF1-CXCR4 pathway and reduce the fracture rate [261].Furthermore, in the rat model of ischemic brain injuries, rat MSCs migrate into the injured sites of the brain and show therapeutic effects via the SDF1-CXCR4 pathway [262,263].Although the expression level of CXCR4 is high in MSCs, few the on the cell surface [256,261].However, Honczarenko et al. have demonstrated that the surface expression of CXCR4 is up to around 43% [264], indicating that some factors might induce the cell surface expression of CXCR4, such as culture conditions, stimuli, and passage numbers [265].Indeed, the expression level of CXCR4 decreases during passaging [264,266] and aging [267].The cell membrane localization of CXCR4 is induced by cytokine stimulation (such as SDF-1) [256,261,268] or 3D culture conditions [269].
Some other MSC markers have also been demonstrated, such as LepR (Leptin receptor) [270][271][272][273], CD34 [274][275][276], and TNFRSF10D [277].However, their functions in MSCs remain unclear.Purifying MSC subpopulation with one single maker has many disadvantages (Table 2).Therefore, the combination of multiple markers is a promising strategy to improve the efficiency and efficacy of MSC subpopulation purification.It has been demonstrated that the PODXL hi /ITGA6 hi MSCs have better activities of colony formation, differentiation, proliferation, homing activities, regeneration activity, and anti-apoptotic activities [152,[278][279][280][281][282].The PDGFR + Sca-1 + MSCs could differentiate into both mesenchymal and endothelial at single-cell level with enhanced self-renewal and multipotency abilities [129], and the CD146 + PDGFRβ + MSCs have higher levels of colony-forming activities [91].Combining PDGFRα and other markers, such as Ly6a, Sca-1, and CD51, would further enrich the MSC subpopulation with enhanced colony-forming and differentiation activities [129,243,283,284].On the other hand, identifying novel MSC markers with novel high-throughput technologies is also critical for both MSC subpopulation purification and understanding the heterogeneity of MSCs.

2nd generation of MSC maker identification-high-throughput approach
The emergence and development of high-throughput technologies (genomics, transcriptomics, proteomics, lipidomics, metabolomics, and so on) have revolutionized various fields of life sciences [285][286][287].These highthroughput technologies have not only expedited the pace of research but also transformed our understanding of life itself by providing a comprehensive and intricate view of biological systems.Since the introduction of Illumina's Solexa sequencing technology in 2005, a new era has been heralded by paving the way for high-throughput technologies [288].This pioneering approach, built upon parallel sequencing principles, enables the simultaneous analysis of millions of DNA fragments, dramatically boosting sequencing efficiency.Notably, this breakthrough laid the foundation for subsequent advancements, with other platforms such as 454 Life Sciences, Ion Torrent, and PacBio also contributing to the progress of high-throughput technologies [285].
The impact of high-throughput technology extends far beyond genomics, reverberating profoundly across various domains of omics research.In the realm of transcriptomics, we can now unravel intricate gene regulatory networks by simultaneously analyzing the expression of thousands of genes.Technologies like RNA-Seq have empowered scientists to assess gene expression patterns across different conditions, tissues, or developmental stages, shedding light on cellular processes and signaling pathways [286].High-throughput mass spectrometry techniques in proteomics offer a swift and comprehensive understanding of protein-protein interactions, modifications, and functions within cells.These methods allow researchers to identify and quantify proteins in complex samples, revealing insights into cellular processes, biomarker discovery, and disease mechanisms [287].In lipidomics, mass spectrometry-based methods have enabled the comprehensive analysis of lipid molecules in biological samples, uncovering lipid profiles associated with health and disease [289].Similarly, metabolomics, utilizing high-throughput mass spectrometry and nuclear magnetic resonance (NMR) techniques, offers insights into the global metabolite composition of cells or organisms, contributing to our understanding of metabolic pathways and disease biomarkers [290].Epigenomics, focusing on epigenetic modifications like DNA methylation and histone modifications, benefits from highthroughput techniques such as DNA methylation arrays and next-generation sequencing.These tools provide a genome-wide view of epigenetic modifications, aiding in deciphering their roles in gene regulation, development, and disease [291].
Among these high-throughput technologies, single-cell RNA sequencing technology (scRNA-seq) is a significant innovation in the field of MSCs that has sparked widespread interest in recent years [17][18][19][20][21][22][23][24].By deciphering the gene expression of each individual cell within a cell population, this technique reveals the astonishing complexity of cellular diversity and heterogeneity, bringing about a revolutionary breakthrough in cellular biology research [292,293].Distinct from traditional bulk RNA sequencing methods, scRNA-seq can precisely analyze cell function and types, regardless of sample heterogeneity [293][294][295].
Several novel MSC markers have been discovered since the application of scRNA-seq technology in the MSC field, such as the LRRC75A + MSCs with enhanced VEGF production [23]; the CMKLR1 + MSCs with improved immune suppression capabilities [19]; the F3 + and S100A9 + MSCs with better regenerative activities [17,18,34].
Among these novel MSC markers identified by scRNA-seq, the CMKLR1 + subpopulation with enhanced immune suppression capabilities [19] has been investigated in detail.The CMKLR1 (Chemokinelike receptor 1), also known as CCRL2 (chemokine C-C motif receptor-like 2), is the transmembrane receptor for chemoattractant chemerin, involved in recruiting and migrating of lymphocytes and immune suppression via its ligand resolvin E1, an important anti-inflammatory mediator [296].Furthermore, it has been demonstrated that the CMKLR1 + MSCs have better osteogenic differentiation potential and weaker adipogenic differentiation potentials than the CMKLR1 − MSCs [19].Indeed, the CMKLR1 pathway regulates the differentiation balance between the osteoblastogenic and adipogenic MSCs [297].However, their data also indicate that inhibiting the CMKLR1 pathway promotes the osteoblastogenic differentiation of MSCs and suppresses the adipogenic differentiation of the mouse MSCs [297].Whether the controversial data resulting from species differences needs further investigation [298][299][300].
The applications of scRNA-seq not only promote the identification of novel MSC markers, but also uncover new potential functions of MSCs.The MSC marker F3 [17,18], also known as CD142 or thromboplastin, is a transmembrane glycoprotein and a receptor for coagulation factors, which is involved in platelet activation and coagulation development after tissue injury [301,302].The discovery of F3 in MSCs might indicate that MSCs play an important role in blood clot formation at the site of injury.
The extracellular matrix modification function of MSCs is well-known [303,304].However, the extracellular matrix microenvironment is a highly complex and dynamic biological component and is critical for the functions of MSCs, including the immune modulation function and stem cell characteristics [3,305,306].Identifying new extracellular matrix-related MSC markers, such as Serpinf1 [22] and HMMR [20], would enhance our understanding of MSC biology in greater depth and breadth..
In addition, scRNA-seq is a powerful strategy for investigating the heterogeneity of MSCs.Purifying a homogenous MSC subpopulation is proposed to have improved therapeutic advantages [17,19,23,307].However, it has been demonstrated that the Gli1 + MSCs are still heterogenous, as revealed by scRNA-seq [24].Furthermore, scRNA-seq can also uncover the diversity of functions and interactions among different MSC subpopulations.Two major MSC subpopulations (Lgr5 + and Lgr6 + ) residing in the mouse lung have completely different functions, uncovered by scRNAseq analysis [21].Lgr6 + MSCs support the airway differentiation, while the Lgr5 + MSCs promote alveolar differentiation [21].In the human umbilical cord, four different MSC subpopulations (proliferative, nichesupporting, metabolism-related, and biofunctional MSCs) have been revealed by scRNA-sequencing [17].

High-throughput techniques for purifying MSC subpopulations
Throughout the developmental trajectory of scRNAseq, various innovative platforms have emerged, each catering to different research needs based on their unique principles and features.The 10 × Genomics Chromium system is one widely used platform.It employs droplet technology to combine individual cells with specific molecular barcode particles, enabling high-throughput cell capture and transcriptome sequencing [308].Similarly, Drop-seq utilizes droplet technology to encapsulate cells and molecular barcode beads in droplets, providing a cost-effective option for large-scale cell sequencing [308].For studies requiring more accurate and comprehensive gene expression information, SMART-seq2 is an ideal choice.Its principle involves introducing specific sample labels after reverse transcription of RNA, allowing individual processing and sequencing of each cell's RNA for deeper insights [309].For large-scale sample processing, CEL-seq2 proves to be a powerful selection, utilizing cellspecific molecular barcode primers to provide unique identification for each cell [310].Additionally, the C1 platform, also known as Fluidigm C1, is an advanced single-cell analysis technology platform.It combines microfluidics technology and real-time fluorescence PCR technology, enabling high-throughput capture, processing, and analysis of individual cells.The workflow of the C1 platform includes key steps such as cell capture, lysis, reverse transcription, and amplification, resulting in high-quality single-cell transcriptome data.The C1 platform can be applied to various types of cell analysis, offering crucial support for cellular biology research [311].
Apart from the aforementioned platforms, other unique single-cell sequencing platforms continue to advance the field of cell analysis.For instance, inDrop, a platform similar to Drop-seq, utilizes microfluidic chips for cell capture, boasting high-throughput performance.Its distinctive design involves encapsulating cells and molecular barcode beads together in droplets, enhancing efficiency and accuracy in cell capture and analysis [308].Moreover, sci-ATAC-seq is another notable platform that not only focuses on single-cell gene expression but also integrates transcriptome and chromatin accessibility information, providing researchers with more comprehensive data [312].SPLiT-seq, a high-throughput singlecell sequencing technology, simultaneously captures the transcriptomes of thousands of cells.Through specialized fragmentation tags, cellular lysates are split into multiple fragments, each containing a cell-specific molecular barcode.This tag design enables the concurrent amplification of RNA fragments from multiple cells in a single reaction, achieving high-throughput cell capture and sequencing [313].
The exploration of the majority of these markers (S100A9, F3, LRRC75A, SERPINF1, CMKLR1, GL1) in the context of scRNA-seq applications has primarily relied on the 10 × Genomics Chromium system [17][18][19][22][23][24], while the CD168 identification was conducted with the C1 platform [20] and the interaction between Lgr5 + and Lgr6 + MSCs were carried out by using the SMART-seq2 technology [21].The reliability of these three platforms for developing novel MSC markers has been successfully validated.However, other single-cell sequencing platforms have yet to be applied in the MSC marker field so far.Their respective unique advantages, however, suggest they still hold immense potential for the development of new and effective MSC markers.
Beyond accelerating the pace of analysis, these technologies facilitate the simultaneous analysis of expansive datasets, laying bare the intricate network of molecular mechanisms and relationships that drive biological systems [285].With the wide application of high-throughput technologies in biomedicine, we also had a deeper understanding of the complexity of biological systems and sought to go beyond the limitations of single omics.The rise of high-throughput technologies not only accelerated data generation but also paved the way for the emergence of multi-omics.This approach, fueled by the copious data generated, marries different omics layers (genomics, transcriptomics, proteomics, and metabolomics) into a comprehensive narrative of biological intricacies.By merging diverse omics datasets, multi-omics integration offers a more comprehensive biological context, enhancing accuracy and facilitating meaningful interpretation of findings [314,315].Notably, multi-omics technologies have already been applicated in MSCs [316][317][318][319]. Gao et al. demonstrated the utility of multi-omics analysis in understanding the immunosuppressive efficacy of MSCs, shedding light on cellular senescence and PD-L1 expression through single-cell transcriptome and proteomic data analysis [319].Their findings underscore the potential of multi-omics approaches in discovering new effective MSC markers.This indicates that multiomics is a feasible strategy to find new effective MSC subpopulations.
In the clinical translation of marker-based sorting, personalized therapeutic interventions are becoming increasingly feasible.However, further clinical studies are needed to validate the efficacy and safety of these approaches.Overall, leveraging the unique properties of marker-sorted MSC subpopulations holds great promise for advancing regenerative medicine and immunotherapy, offering tailored treatments for diverse medical conditions.

Conclusions and perspectives
Although the therapeutic applications of the MSCs have great promises, challenges still need to be overcome [320,321].And the heterogeneity of MSCs constitutes one of those important barriers before their clinical application [6,7].Through bioinformatic analysis of the RNA-seq data from different labs and tissues, it is shown that the isolation and expansion procedures induce more heterogeneity than the tissue origin [12].It should be noted that purifying and expanding the MSCs in vitro is a kind of stress similar to tissue damage in vivo, which might affect the molecular pathways and functions of MSCs [322].Indeed, the expanded MSCs in vitro are very different from their counterpart in vivo [323].The MSC expansion strategy would select the cell population which could adapt to these stimuli and stresses [322], indicating the necessity of standardizing the MSC processing procedures and developing a full chemical defined medium [6,[10][11][12].Therefore, selecting the suitable MSC subpopulations with specific markers based on their functions and applications is necessary and mandatory [6,7].
So far, the quest for identifying markers of MSCs has been incessant.The emergence of advanced highthroughput multi-omics techniques offers a promising avenue for discovering novel markers.In this review, numerous MSC subpopulations identified through marker-based sorting have demonstrated significant therapeutic efficacy in animal models.These subpopulations mainly exhibit enhanced therapeutic effects through their potent immunosuppressive capabilities, which have been validated across various animal models of inflammation [38,68,74,112,114].Additionally, some subpopulations possess superior homing and regenerative properties, contributing to tissue repair in the rat model of myocardial infarction and the mouse model of hindlimb ischemia [31,33,124,125].These findings offer promising directions for future therapeutic applications of MSCs.Furthermore, ongoing advances in understanding and manipulating the properties of MSC subpopulations hold great promise for the development of more targeted and effective therapies in regenerative medicine and immune modulation.
However, upon evaluating the majority of currently developed markers, a trend becomes apparent: many subpopulations that are sorted tend to revert to an unsorted state after multiple generations of in vitro proliferation.For instance, during the isolation of MSCs using markers like MCAM, CD9, CXCR4, and STRO-1, their expression diminishes upon subsequent in vitro expansion and cultivation [100,124,264,266,324].This situation might indicate that the sole reliance on biomarkers cannot purify consistent and stable subpopulations of MSCs, and eventually achieve successful applications in clinical medicine.
Under diverse physiological or pathological conditions, MSCs exhibit various forms of plasticity, including alterations in morphology, surface markers, secretion, differentiation, proliferation, migration, and apoptotic potential [325].This plasticity is intimately linked to the microenvironment surrounding MSCs, where physical, chemical, and biological factors impact MSCs' functions through distinct mechanisms [325].These mechanisms might involve critical processes like signaling pathway modulation and cellular reprogramming, ultimately influencing MSCs' capabilities [325].Illustrating the immunomodulatory role of MSCs exemplifies this phenomenon.During the acute phase or relapse of inflammation, effector T cells secrete pro-inflammatory cytokines, including IFN-γ, TNF, IL-1, and IL-7.These pro-inflammatory cytokines stimulate MSCs to produce substantial amounts of IDO (indoleamine 2,3-dioxygenase) and chemokines.Chemokines serve to attract activated T cells toward MSCs.The elevated concentration of IDO metabolites stemming from this process directly inhibits T cells, resulting in an overall attenuation of the immune response and promotion of tissue repair [1,326].On the contrary, in chronic inflammation or during remission, the concentration of anti-inflammatory cytokines, such as TGF-β, increases while pro-inflammatory cytokines decline.Consequently, the production of IDO by MSCs drops below the immunosuppressive threshold.Despite the continued expression of chemokines albeit at lower levels, recruited T cells are not restricted, thus exacerbating the inflammatory immune response [1,326].
As previously highlighted, MSCs are characterized by their exceptional plasticity.The exclusive focus on purifying MSC subpopulations could potentially impose certain limitations.A more intricate strategy revolves around carefully shaping the extracellular environment of MSCs through deliberate in vitro cultivation, a process terming 'MSC education' .The objective of this educational initiative is to tap into the inherent variability present within the cell population, steering it towards a consistent manifestation of the intended functions.Across a spectrum of models, diverse categories of educated MSCs have unveiled a range of distinctive functionalities [327][328][329].For instance, when BM-MSCs are exposed to WNT5a secreted by gastric cancer cells, a noticeable upregulation of α-SMA expression and an amplified capacity for driving tumorigenesis have been observed [327].Furthermore, the exosomes released by MSCs primed with neonatal serum have proven capable of expediting the healing of cutaneous wounds by actively stimulating angiogenesis [328].Notably, MSCs that have undergone a process of education through exposure to chemotherapy have emerged as critical mediators in facilitating communication between MSCs and tumor-initiating cells within specific tumor contexts.This communication is achieved through the selective secretion of cytokines and/or chemokines [329].In the realm of immune regulation, differently educated MSCs can even exhibit contrasting functionalities.Waterman et al. found that MSCs educated by T-cell signaling (referred to as MSC-I) are primarily geared toward producing proinflammatory factors, while MSCs educated by TLR3 signaling (referred to as MSC-II) predominantly express immune-suppressive factors [330].Similarly, MSCs educated by immune factors such as IFN-γ and TNF-α also demonstrate enhanced immune-suppressive capabilities [1,331].
To sum up, the development of MSC markers, bolstered by high-throughput techniques, holds substantial potential.Looking at the broader field of MSC research, in addressing the challenge of inconsistent therapeutic efficacy due to MSC heterogeneity, MSC education also presents a viable avenue alongside MSC markers.

Fig. 3
Fig. 3 GO analysis of MSC markers.The bioinformatic analysis of GO enrichment of MSC markers was performed with Dotplotting (A), Cnetplotting (B).GO, gene ontology.MSC mesenchymal stem/stromal cell

Fig. 4
Fig. 4 KEGG analysis of MSC markers.The bioinformatic analysis of KEGG enrichment of MSC markers was performed with Dotplotting (A), Cnetplotting (B), and Emapplotting (C).Bioinformatic analysis was conducted with package 'enrichplot' in R. KEGG, kyoto encyclopedia of genes and genomes.MSC mesenchymal stem/stromal cell

Purification By Markers Standardize Procedure Donors Sources Cell isolation Cell culture
. It's worth noting that

Table 1
Basic information for MSC markers

Table 2
Pros and cons of MSC markers