Mesenchymal stem cells with a pericyte-like phenotype for tissue engineering

Abstract

The existence of specific mesenchymal stem cell (MSC) subsets with pericyte-like properties in bone marrow and mobilized peripheral blood opens the possibility of harvesting these cells relatively easily. It is known that MSCs with pericyte-like properties play an important role in maintaining hematopoiesis in vivo, which is why products based on them can be used in hematology, transplantation medicine, and oncology. Therefore, the development of a technology for the precise isolation of this MSC subtype is a highly promising objective. However, for tissue engineering, e.g., the engineering of bone, cartilage, muscle, and nervous tissues, the use of this MSC subtype may be of comparable importance and potential. This significantly broadens the relevance of this problem.

Key Words: Pericyte-like phenotype; Bone marrow-derived mesenchymal stem cells; Mesenchymal stem cell-like pericytes; CD146+; Hematopoietic stem cells; Tissue engineering; Bone; Vascularization; Immunity

Core Tip: The novel cell subtypes of mesenchymal stem cells with pericyte-like properties can be isolated from bone marrow and peripheral blood. It is known that mesenchymal stem cells with pericyte-like properties play an important role in maintaining hematopoiesis in vivo, which is why products based on them can be used in hematology, transplantation medicine, and oncology. However, for tissue engineering, the use of this cell subtype may be of comparable importance and potential. This significantly broadens the relevance of this problem.

INTRODUCTION

The question of the in vivo tissue origin and identity of mesenchymal stromal cells (MSCs) has long remained a subject of debate. The classical method of MSC isolation, based on selective adhesion to culture plastic, allowed for the procurement of heterogeneous cell populations but provided no insight into their natural tissue niche. A fundamental breakthrough in this field occurred in 2008, when a publication by Crisan et al[1] radically altered the existing paradigm. The authors convincingly demonstrated that it is possible to isolate a cellular fraction from the perivascular space of various human organs that expresses canonical pericytic markers [CD146, NG2, and platelet-derived growth factor receptor beta (PDGFRβ)] in the complete absence of hematopoietic, endothelial, and myogenic determinants. Under in vitro conditions, these cells exhibited the full spectrum of attributes characteristic of MSCs: Multilineage differentiation (osteogenic, chondrogenic, and adipogenic) as well as clonal self-renewal capacity. Thus, for the first time, direct experimental evidence was obtained that pericytes and other perivascular cellular elements represent the endogenous equivalent of MSCs, with the perivascular space serving as their primary reservoir in the postnatal organism.

In subsequent years, this concept has been further developed and refined. In a comprehensive work by Crisan et al[2], the heterogeneity of the perivascular niche itself was demonstrated: In addition to classical capillary pericytes [CD146+, PDGFRβ+, NG2+, α-smooth muscle actin (αSMA)+/-], a distinct cell population (CD34+, CD146-, αSMA-) was identified in the adventitial layer of arterioles and venules. These adventitial cells also possess MSC potential and are capable of differentiation toward a pericyte-like phenotype, indicating significant plasticity and hierarchical organization within the perivascular system.

A fundamentally important step was the reconsideration of the functional role of MSCs in the context of their perivascular localization. In a conceptual review by Caplan[3], a model was proposed whereby MSCs, mobilized from the vessel wall in response to injury or danger signals, function not only as progenitor cells (“repairmen”) but also as active regulators of tissue homeostasis—acting as “sentinels" and "gatekeepers”. They monitor the microenvironment, modulate local inflammation, regulate endothelial barrier permeability, and coordinate reparative processes in close interaction with endothelial cells.

Recent research over the past two years not only fully confirms but substantially advances this paradigm. Specifically, it has been shown that pericytes/perivascular cells isolated from muscle tissue, as well as skeletal muscle-derived myoendothelial cells, are phenotypically similar to bone marrow-derived MSCs (BM-MSCs), express standard MSC surface markers, and possess multilineage differentiation potential[4]. Furthermore, a novel rare subpopulation of vasculogenic precursor cells (VPCs), uniquely expressing CD141 (thrombomodulin), has been identified from human bone marrow. This subtype of CD141+ pericyte-like cells does not carry typical surface markers of endothelial progenitor cells or activated endothelial cells (such as CD31); however, when co-cultured with human umbilical vein endothelial cells and BM-MSCs during Matrigel tube formation in vitro under inflammatory conditions, these cells induce the expression of mature endothelial markers [CD31, vascular endothelial growth factor (VEGF) receptor 2, and VE-cadherin][5]. Moreover, CD141+ VPCs are characterized by a short population doubling time (approximately 20 hours), are capable of generating an elaborate vascular network on Matrigel even under conditions deficient in exogenous angiogenic factors, and recruit BM-MSCs to this network as pericyte-like cells, maintaining this ability in the presence of tumor necrosis factor alpha[5]. It has also been demonstrated that human dental pulp stem cells, after long-term biobanking (7-8 years), fully retain MSC identity and multilineage differentiation potential, with most cells maintaining a pericyte-like phenotype characterized by the expression of angiogenic/endothelial genes (CD31, VEGF receptor 2, neuropilin-1, and angiotensin-converting enzyme) under pro-angiogenic culture conditions[6]. Collectively, these data convincingly indicate that pericyte-like MSCs are present in various tissues and can be both mobilized in situ and expanded ex vivo for regenerative medicine applications.

As noted above, in the presence of VEGF and/or under conditions mimicking the endothelial microenvironment, MSCs are capable of acquiring a pericyte-like phenotype characterized by the expression of markers such as CD31, CD146, von Willebrand factor, PDGFRβ, and α-SMA, and demonstrate the ability to form capillary-like structures both in vitro and in vivo[6-8]. Pericyte-like MSCs can independently initiate the formation of vascularized tissue, which is crucial for tissue engineering, as adequate vascularization is a critical prerequisite for the successful regeneration of bone, cartilage, muscle, and nervous tissues[6]. The significance of the pericyte-like phenotype for regenerative medicine is realized through three interrelated mechanisms.

Stabilization of the vessel wall

Pericytes, closely associated with endothelial cells, provide mechanical support to capillaries, regulate microcirculatory blood flow, and maintain the integrity of histohematic barriers. In the absence of pericytic coverage, newly formed vascular structures remain immature, are characterized by increased permeability, and undergo rapid regression[9]. Pericyte-like MSCs, by integrating into the vessel wall, effectively perform this stabilizing function.

Regulation of angiogenesis

Pericyte-like MSCs control the proliferative activity and directed migration of endothelial cells through key signaling pathways, primarily PDGFβ/PDGFRβ and transforming growth factor-β (TGF-β). They thereby prevent excessive, uncontrolled vascular growth, which is a prerequisite for the formation of stable vascular networks within tissue-engineered constructs[10]. Under the influence of pro-inflammatory cytokines (specifically, tumor necrosis factor-alpha), pericyte-like cells are capable of recruiting additional MSCs for vascular network formation, as demonstrated in the CD141+ VPC model[5].

Paracrine support

These cells secrete a wide range of bioactive molecules, including VEGF, hepatocyte growth factor, insulin-like growth factor-1, and angiopoietin-1, as well as extracellular vesicles and extracellular matrix components. This paracrine activity not only stimulates angiogenesis but also exerts immunomodulatory effects, suppresses apoptosis, promotes the survival of resident stem cells, and facilitates the recruitment of endothelial progenitor cells[3]. Notably, the long-term retention of the pericyte-like phenotype even after cryopreservation (as demonstrated for dental pulp cells[6]) makes this population particularly attractive for the development of clinically relevant cell-based products. Thus, the ability of pericyte-like MSCs to simultaneously participate in the formation of vascular structures, stabilize them, and support the microenvironment via paracrine mechanisms renders them an indispensable component of tissue-engineered constructs aimed at the regeneration of vascularized tissues.

Despite their considerable therapeutic potential, the widespread clinical application of MSCs faces several significant obstacles. Traditional MSC cultures, obtained by plastic adherence, represent an extremely heterogeneous cell population in which cells with varying proliferative, differentiation, and paracrine potentials coexist. Furthermore, MSC properties can vary significantly depending on individual donor characteristics, tissue source (bone marrow, adipose tissue, dental pulp, etc.), and in vitro expansion conditions[11,12]. This heterogeneity results in low reproducibility of experimental outcomes and hinders the transition to standardized cell-based products.

A fundamental implication arising from previous work[1-3] is that the capacity to perform pericytic functions is not inherent to the entire heterogeneous MSC population. Pericyte-like activity is a characteristic of a specific subpopulation that appears to be most closely related to its in vivo precursors—pericytes and adventitial cells[13]. Consequently, the development of clear identification criteria and isolation methods for this functionally significant subpopulation represents a pressing and as-yet incompletely resolved task.

In this context, several studies deserve particular attention[12,14]. Based on detailed phenotypic profiling, these studies convincingly demonstrate the existence of a specific MSC subpopulation characterized by the immunophenotype CD45-/CD73+/CD39+/CD146+ in bone marrow and, most importantly, in mobilized peripheral blood. The most critical aspect of this research is not merely the identification of this population in bone marrow, but the demonstration of its presence in the circulation following mobilization. This finding indirectly points to the functional significance of this MSC subset and opens up the potential for relatively non-invasive procurement of these cells for clinical applications[12]. The use of the CD45-/CD73+/CD39+/CD146+ marker combination appears to be an effective approach for isolating precisely the pericyte-like MSCs: Selection based on CD45- excludes cells of the hematopoietic lineage; expression of CD73 and CD39 is closely associated with the MSC phenotype and helps differentiate them from most mature endothelial cells; and the presence of CD146+ indicates an association with the pericytic lineage and, more broadly, with the vascular wall[14]. Thus, these studies propose a specific algorithm for the search, identification, and isolation of functionally relevant MSCs with a pericyte-like phenotype.

This review is conceived not as a formal recapitulation of existing methodological approaches, but as an attempt—together with colleagues working in related fields—to navigate the convoluted path leading from the fundamental question of MSC tissue identity to the practical isolation of their functionally significant fraction. Our goal is to critically re-evaluate the existing arsenal of strategies for isolating and characterizing pericyte-like MSCs, to systematize the accumulated experience of their application in hematology and regenerative medicine, and to outline the range of unresolved issues that must be overcome for further progress. The logic of the presentation is structured around several key themes, the sequential consideration of which will, we hope, help provide a comprehensive picture.

First, we will examine in detail the panel of surface markers that currently serves as the primary tool for identifying pericyte-like subpopulations. We will compare the classical markers (CD146, NG2, and PDGFRβ), characterized in pioneering works[1,2], with the recently introduced candidates—CD141, CD73, and CD39—and attempt to address a fundamental question: To what extent does the expression of these molecules correlate with the actual functional competence of the cells, whether it be the ability to stabilize the vessel wall, support angiogenesis, or form a hematopoietic niche? Of particular interest in this context are the recently identified CD141+ vasculogenic progenitor cell subpopulation[5], as well as the approach for enriching pericyte-like MSCs based on the CD45-/CD73+/CD39+/CD146+ combination[12,14].

Next, the discussion will move to the experimental level. We will discuss which model systems—ranging from relatively simple Matrigel tube formation assays to complex co-cultures with endothelial cells—allow for sufficiently reliable verification of the pericyte-like phenotype and assessment of the behavior of these cells under conditions approximating the inflammatory microenvironment in vivo. As shown by Blocki et al[10], not all MSCs are capable of performing pericytic functions, necessitating the use of stringent functional tests for the validation of isolated subpopulations.

We will then analyze how the accumulated fundamental knowledge is being translated into specific tissue engineering applications. Where does the use of pericyte-like MSCs already enable the resolution of vascularization challenges in volumetric implants for the repair of bone, cartilage, muscle, or nervous tissue, and where does their potential remain a subject of cautious optimism? The possible role of these cells in hematology—specifically, their ability to replicate key elements of the stromal microenvironment required for maintaining hematopoietic stem cells (HSCs) ex vivo—also warrants separate attention, echoing recent findings on the ability of muscle-derived perivascular stem cells to support hematopoietic progenitors[4].

Finally, we will not sidestep the fundamental and applied barriers that currently impede the routine clinical use of pericyte-like MSCs. Why does the standardization of isolation protocols, particularly concerning rare subpopulations circulating in peripheral blood, remain such an elusive goal? What approaches can help maintain the pericytic phenotype during culture expansion and prevent its “erosion”—a problem particularly relevant in light of data on the heterogeneity of traditional MSC cultures[11]? And perhaps the most concerning aspect: To what extent are concerns justified regarding the risk of uncontrolled angiogenesis and the potential influence of these cells on the progression of latent tumor processes? These safety considerations require the closest scrutiny when planning preclinical and clinical studies.

We invite colleagues whose interests lie in the fields of stem cell biology, vascular biology, tissue engineering, and regenerative hematology to join this examination. Our shared objective is not merely to inventory established facts, but also to chart the directions for future research that will transform the unique biological phenomenon of pericyte-like MSCs into a robust and reproducible tool for clinical medicine.

THE PHENOTYPIC LANDSCAPE OF PERICYTE-LIKE MSCS: MARKERS AND ISOLATION STRATEGIES

The development of effective isolation strategies for functionally significant pericyte-like MSCs is impossible without a thorough understanding of their phenotypic landscape. The challenge lies not simply in recognizing cells by a set of surface proteins, but in identifying combinations of markers that can reliably predict their in vivo behavior—their capacity to stabilize blood vessels, modulate inflammation, and support regenerative processes. In this section, we will systematically review the evolution of understanding regarding the phenotype of these cells: From the use of classical pericytic markers to the rational design of complex panels aimed at isolating specific functional subpopulations (Figure 1).

Figure 1 Strategy for isolation of pericyte-like mesenchymal stem cells from bone marrow and peripheral blood. The workflow for pericyte-like mesenchymal stem cell isolation and validation is shown. HSCs: Hematopoietic stem cells.

Historically, the first steps toward isolating pericyte-like MSCs were taken using three key markers: CD146 (MCAM), NG2 (CSPG4), and PDGFRβ. It was precisely the combination of these molecules, along with the absence of hematopoietic, endothelial, and myogenic markers, that enabled Crisan et al[1], in their pioneering 2008 work, to isolate for the first time cells from the perivascular space possessing the full spectrum of MSC properties. Later, this approach was confirmed in studies of other tissues. For example, human stem cells from exfoliated deciduous teeth and periodontal ligament stem cells, which also exhibit a pericytic phenotype and the ability to stimulate angiogenesis, were positive for NG2, PDGFRβ, and CD146. Similarly, myoendothelial progenitor cells isolated from human skeletal muscle also express these canonical markers[15].

α-SMA is often considered an additional marker indicating a more mature, contractile pericyte phenotype. However, its expression is highly variable and depends on the vessel type, tissue microenvironment, and cell activation state. For instance, in the capillaries of many tissues, pericytes exhibit an α-SMA-/Low phenotype, and its expression may increase under pathological conditions, such as tumor growth, indicating that the cells are acquiring properties similar to vascular smooth muscle cells[16].

The primary limitation of classical markers lies in their relatively low specificity. Large-scale studies using single-cell RNA-seq demonstrate that CD146 is expressed not only on pericytes and MSCs but also on endothelial cells and certain lymphocyte subpopulations[17]. NG2, or neural/glial antigen 2, is widely present on glial progenitor cells in the central nervous system, limiting its utility for isolating pericytes from neural tissue[18]. PDGFRβ, in turn, is a common marker for many mesenchymal cells, including fibroblasts, which also diminishes its value as a unique identifier[19].

Contemporary research is increasingly moving away from single markers toward more complex panels that enable the isolation of functionally specialized subpopulations. One notable example is the identification of a rare population of VPCs in human bone marrow, uniquely expressing CD141 (thrombomodulin)[5]. CD141+ VPCs do not express classical endothelial markers (CD31, VEGFR2, and VE-cadherin) in a quiescent state but acquire them during the formation of capillary-like structures in Matrigel. This subpopulation exhibits high proliferative activity (population doubling time of approximately 20 hours) and the capacity for vasculogenesis even under growth factor-deficient conditions[5].

Another promising combination is the CD73/CD39 axis, associated with the production of immunosuppressive adenosine. These two ectonucleotidases sequentially catalyze the dephosphorylation of ATP/ADP to adenosine, which, by binding to receptors on immune cells, exerts a potent anti-inflammatory effect[20]. In the work by Martin and Gullo[14] in 2025, it was precisely the combination of CD73+/CD39+, together with CD146+ and CD45-, that enabled the isolation from bone marrow and mobilized peripheral blood of an MSC subpopulation presumably responsible for supporting HSCs. Functionally, this subpopulation produces significantly more adenosine than unsorted MSCs, which may represent a key mechanism for modulating the hematopoietic niche microenvironment[13].

Pericyte-like MSCs have been isolated from various tissues, and their phenotypic and functional properties differ substantially. Cells derived from bone marrow exhibit heterogeneity: Classical CD146+ pericytes possess high osteogenic potential in vivo, particularly in combination with CD34+ adventitial cells[1,13], whereas the rare CD141+ VPC subpopulation from the same source is distinguished by outstanding angiogenic activity and the ability to form vascular networks even without the addition of exogenous growth factors[5]. Dental pulp stem cells are notable for retaining a pericyte-like phenotype and angiogenic potential even after long-term (7-8 years) cryopreservation[6]. Myoendothelial progenitors from skeletal muscles express, in addition to classical markers, CD56 and CD34, possess multilineage differentiation potential, and are capable of supporting HSCs in vitro[4,21]. Pericyte-like cells from human umbilical artery also exhibit remarkable plasticity, forming neurospheres and differentiating along neural lineages[22]. Adipose tissue-derived adventitial cells, conversely, are characterized by low intrinsic angiogenic potential but exert a powerful paracrine effect, stimulating osteogenesis and modulating immune responses[2,13]. Thus, the choice of cell source should be dictated by the specific clinical objective: For direct vascularization of implants, bone marrow-derived VPCs or dental pulp cells are preferable, whereas for immunomodulation and paracrine support of regeneration, the use of adipose tissue-derived adventitial cells is more justified.

The work by Martin and Gullo[14] serves as an excellent example of a rational approach to constructing a marker panel. The use of the CD45- marker reliably excludes all cells of the hematopoietic lineage that could “contaminate” the primary culture. A key element of the strategy is the CD73+/CD39+ pair. First, these markers are part of the standard MSC phenotype[2]; second, and more importantly, they select for cells with an active adenosine pathway, potentially crucial for immunosuppression and stem cell niche support[22]. Finally, the addition of the CD146+ marker imparts pericytic specificity to the panel, distinguishing these cells from fibroblasts and other stromal elements that express CD73 and CD39 but are not associated with blood vessels[9]. This panel design can be successfully adapted for isolating functionally significant subpopulations from other pericyte-rich tissues, such as dental pulp[6] or skeletal muscle[4].

THE ROLE OF PERICYTE-LIKE MSCS IN SUPPORTING HEMATOPOIESIS AND THE HSC NICHE

Maintenance of the HSC pool throughout the lifespan of an organism is impossible without a specialized microenvironment—the hematopoietic niche. For a long time, it was believed that osteoblasts, forming the so-called “endosteal niche”, played the key role in this process. However, over the past two decades, a fundamental paradigm shift has occurred: There is now no doubt that it is the perivascular space, populated by pericytes and related stromal cells, that serves as the primary reservoir of signals regulating HSC quiescence, self-renewal, and differentiation[23-26]. In this section, we will examine the evolution of concepts regarding the vascular niche, summarize the evidence for hematopoietic support by various types of pericytes, and discuss the prospects for using isolated subpopulations of pericyte-like MSCs to address one of the most pressing challenges in hematology: The efficient expansion of functional HSCs ex vivo.

The notion that perivascular cells form a niche for HSCs crystallized in the early 2010s and has since garnered substantial experimental support. Early studies demonstrated that in adult mammalian bone marrow, HSCs are predominantly localized in close proximity to sinusoidal vessels, where they contact pericytes and endothelial cells[23]. It was later established that distinct pericyte subpopulations, differing in marker expression [CD146, NG2, PDGFRβ, nestin, and leptin receptor (LepR)], are associated with either arterioles or sinusoids, and each of these subpopulations makes a unique contribution to the regulation of HSC behavior[25]. Specifically, arteriolar pericytes express NG2 and support HSC quiescence, whereas sinusoidal pericytes, expressing the LepR, actively secrete factors necessary for HSC proliferation and mobilization[26].

Recent studies have significantly deepened our understanding of the vascular niche organization. It has been shown that endothelial cells and perivascular stromal cells jointly form heterogeneous vascular niches that differ in cellular composition and secreted molecular profiles across distinct bone regions, which likely determines local variations in HSC regulation[24]. The intra-bone marrow diversity of endothelial cells directly impacts HSC development and maintenance, adding another layer of complexity to the vascular niche[27]. Moreover, the vascular niche concept is not limited to bone marrow: It was recently demonstrated that the spleen also contains perivascular niches associated with sinusoidal vessels, which are capable of supporting HSCs under conditions of extramedullary hematopoiesis. It is important to emphasize that the capacity to support hematopoiesis is not a universal property of all perivascular cells but rather a characteristic of specific, functionally specialized subpopulations. For instance, recent work has shown that a gradient of the epithelial-mesenchymal transition program, along with stemness and pericyte-like properties, determines the ability of MSCs to function as a niche supporting HSCs, and that attenuation of this program leads to impaired hematopoietic support in degenerative hematological disorders[28].

The ability to support hematopoiesis has been convincingly demonstrated for pericytes and pericyte-like MSCs from various tissue sources. Classical studies in mouse models showed that bone marrow pericytes expressing nestin and LepR are key components of the HSC niche and are essential for maintaining their pool in vivo[23,25]. These observations were subsequently extended to human cells[1,23,29].

Of particular interest are data obtained from muscle-derived perivascular cells. In a study by Yang et al[4] in 2025, it was demonstrated that CD146+ pericytes isolated from human skeletal muscle are capable of effectively supporting hematopoietic stem and progenitor cells (HSPCs) in vitro. These cells express classical pericytic markers (CD146, NG2, and PDGFRβ) and, critically, secrete key hematopoietic factors, including stem cell factor (SCF) and CXCL12, thereby creating a microenvironment conducive to the preservation of HSC stemness properties[4]. Earlier, Chen et al[21] in 2014 showed that myoendothelial cells isolated from human skeletal muscle also possess the capacity to support HSCs, indicating the existence of a specialized perivascular niche in muscle tissue that is potentially significant for extramedullary hematopoiesis[21,30,31].

Equally important results have been obtained from studies of adipose tissue pericytes. Corselli et al[29] in 2012 showed that adventitial cells (CD34+, CD146-) and pericytes (CD146+) from human adipose tissue can differentiate into MSCs and support hematopoiesis in co-culture with HSCs, with adventitial cells demonstrating particularly high paracrine activity by secreting a broad spectrum of hematopoietic cytokines.

It should be noted that hematopoietic support by pericytes is achieved through a complex array of mechanisms involving both direct cell-cell contacts and paracrine regulation. Pericytes secrete key niche factors—SCF, CXCL12, angiopoietin-1, and thrombopoietin—which collectively ensure the balance between HSC quiescence, self-renewal, and differentiation[4,23]. Furthermore, pericytes regulate the metabolic microenvironment of the niche, particularly by maintaining local hypoxia and participating in adenosine metabolism[21,32,33].

One of the most pressing issues in modern transplantation hematology remains the limited number of functional HSCs available for transplantation, especially when using umbilical cord blood or when repeated transplantations are necessary. Traditional ex vivo HSC expansion protocols based on the addition of exogenous cytokines typically lead to rapid cell differentiation and loss of long-term engraftment capacity. In this context, the use of pericyte-like MSCs as a feeder layer that mimics the natural niche microenvironment appears to be a promising alternative.

The work by Martin and Gullo[14] in 2025 proposes a specific cellular substrate for implementing this strategy: The CD45-/CD73+/CD39+/CD146+ subpopulation isolated from bone marrow and, importantly, from mobilized peripheral blood. The authors hypothesize that this particular subpopulation, due to high expression of CD73 and CD39 and, consequently, the ability to actively produce immunosuppressive adenosine, may be a key niche element responsible for supporting HSCs[13]. The rationale for this hypothesis is based on the fact that CD39 and CD73 sequentially catalyze the conversion of extracellular ATP into adenosine, which, by binding to receptors on HSCs, may promote the maintenance of their quiescence and protect them from premature differentiation[20].

The promise of using MSCs for ex vivo HSC expansion is supported by several independent studies. It has been shown that MSCs genetically modified to secrete hematopoietic cytokines (SCF, thrombopoietin, FMS-like tyrosine kinase-3-ligand, and interleukin-3) can support HSPC proliferation and preserve their stem cell characteristics even in the absence of exogenous cytokines[34]. Another approach—using “revitalized” MSCs with restored niche factor expression—achieved a seven-fold increase in functional HSC numbers in culture and protected lethally irradiated recipients after transplantation[35]. It is important to note that the efficiency of HSC expansion critically depends on the source of MSCs. Comparative studies indicate that MSCs derived from Wharton’s jelly of the umbilical cord possess particularly high potential for supporting HSPCs, especially in three-dimensional co-culture systems[36,37]. Recent work has also demonstrated that direct cell-cell contact between HSCs and MSCs significantly outperforms non-contact systems in effectiveness, underscoring the importance of recapitulating not only soluble but also contact-mediated niche signals[38,39]. Thus, the isolation and utilization of the well-characterized CD45-/CD73+/CD39+/CD146+ subpopulation proposed by Martin and Gullo[14] may represent an important step toward standardizing ex vivo HSC expansion protocols and enhancing their clinical efficacy.

Despite the obvious appeal of the hypothesis regarding the key role of the adenosine axis in HSC support, the question of whether the adenosine-producing capacity of CD45-/CD73+/CD39+/CD146+ cells is indeed critical for their niche function remains open and requires further experimental investigation. On one hand, a substantial body of evidence points to the importance of adenosine signaling in the bone marrow. It has been shown that adenosine produced by CD39+/CD73+ cells is involved in regulating HSC quiescence and mobilization. Specifically, inhibition of CD39 and CD73 by small molecules enhances HSC mobilization into the peripheral blood, suggesting a retaining role for adenosine within the niche[23,24]. Additionally, adenosine exerts potent immunosuppressive effects by suppressing T-cell activity and promoting a tolerogenic microenvironment, which may be important for protecting HSCs from immune attack[24].

On the other hand, recent studies have revealed paradoxical effects of adenosine in the context of the hematopoietic niche. In a study investigating the role of CD39 on LepR-expressing (LepR⁺) perivascular MSCs, it was unexpectedly shown that deletion of CD39 in these cells does not suppress but rather enhances the engraftment of allogeneic HSCs. Moreover, adenosine produced by both Tregs and LepR+ MSCs paradoxically inhibits niche Tregs, whereas only adenosine originating from Tregs acts as an effector molecule of niche immune privilege[39]. These findings warrant caution regarding an overly simplistic interpretation of adenosine’s role as a universal mediator of HSC support.

Therefore, to definitively answer the question of the functional significance of the CD73/CD39 axis in the study by Martin and Gullo[14], direct experiments are needed involving selective inhibition of these ectonucleotidases in the isolated cells, coupled with an assessment of their ability to support HSCs in co-culture in vitro and in transplantation models in vivo. It cannot be ruled out that hematopoietic support by these cells is achieved through a more complex set of mechanisms, including the secretion of classical hematopoietic cytokines, contact-dependent interactions, and modulation of local metabolism, with the adenosine axis being only one component of this multifaceted system. Indeed, the functional repertoire of pericyte-like MSCs is not limited to the hematopoietic niche; as schematically summarized in Figure 2, these cells exert distinct, tissue-specific mechanisms of action depending on the local microenvironment, a concept that will be further explored in the context of various tissue engineering applications in the following section.

Figure 2 Multifunctionality of pericyte-like mesenchymal stem cells in regenerative medicine. This core schematic compares functional properties of CD146+ pericytes, CD34+ adventitial cells, and classical mesenchymal stem cells, highlighting their tissue-specific mechanisms. The diagram also illustrates the functional pleiotropy of CD45-CD73+CD39+CD146+ mesenchymal stem cells. MSC: Mesenchymal stem cell; TGF: Transforming growth factor; VEGF: Vascular endothelial growth factor; BMP: Bone morphogenetic protein; PDGF: Platelet-derived growth factor; IGF: Insulin-like growth factor; SCF: Stem cell factor.

APPLICATIONS IN TISSUE ENGINEERING

The translation of fundamental knowledge about pericyte-like MSCs into the realm of tissue engineering encounters a non-trivial challenge: The same cells, depending on the tissue context, may engage fundamentally different mechanisms of action—ranging from direct differentiation to paracrine modulation of the microenvironment[2,3]. In this section, we will critically examine how these mechanisms are implemented in the regeneration of bone, cartilage, muscle, and nervous tissues[13,31,40,41], and discuss the therapeutic potential of pericyte-like MSCs in ischemic conditions[5,42]. Particular attention will be paid to data obtained by our group using a poly(3-hydroxybutyrate) (PHB) scaffold model, which demonstrates that even standard culture medium, combined with a three-dimensional microenvironment, can trigger the osteogenic program of MSCs[43,44].

Bone tissue represents a unique model for studying the functional dichotomy of pericyte-like MSCs, as their ability to simultaneously support angiogenesis and osteogenesis provides a basis for synergistic regeneration; however, distinct subpopulations of perivascular cells exhibit pronounced functional specialization. Elegant evidence of this specialization was provided by Wang et al[13] in 2019 in a comparative analysis of CD146+ pericytes and CD34+ adventitial cells isolated from adipose tissue: CD146+ pericytes induced enhanced formation of vessel-like structures in vitro and angiogenesis in vivo, whereas CD34+ adventitial cells demonstrated more pronounced paracrine-induced osteogenesis[2,29]. Crucially, it was the combined use of both subpopulations in a critical-sized calvarial defect model in NOD/SCID mice that resulted in significantly more complete bone tissue restoration compared to using each cell type separately, indicating the complementary nature of their functions[13,45].

This functional complementarity of pericytes and adventitial cells finds an explanation at the molecular level: Upon co-culture, CD146+ pericytes enhance the expression of osteogenic markers (runt-related transcription factor 2, ALP, and collagen type I alpha 1) in CD34+ adventitial cells, whereas the latter, in turn, stimulate the expression of angiogenic factors (monocyte chemoattractant protein 1, insulin-like growth factor binding-protein-1, and insulin-like growth factor binding-protein-2) in pericytes, creating a bidirectional paracrine loop critical for coupling osteogenesis and angiogenesis[13]. This concept of “osteogenic-angiogenic coupling” is currently regarded as one of the key principles for successful bone tissue engineering[43,44]. At the same time, it is necessary to consider that the osteogenic niche created by differentiating cells and osteoinductive factors may paradoxically suppress vasculogenic ingrowth, requiring additional stimulation, for instance, with PDGF-BB, to overcome this effect[26,42,46].

In the context of these findings, the results obtained by our research group on the osteogenic potential of MSCs within biopolymer scaffolds gain particular significance. We demonstrated that when MSCs are cultured in porous scaffolds made of PHB and its composite with hydroxyapatite, even in standard culture medium lacking osteoinductive supplements, an increase in alkaline phosphatase activity by 2.7- and 4.6-fold, respectively, was observed compared to cells grown on culture plastic[30,43,44]. Concurrently, a substantial increase in the expression of the phenotypic marker CD45—by 1.9- and 12.2-fold, respectively—was recorded, indicating the activation of hematopoietic regulatory programs in response to the three-dimensional microenvironment and contact with the mineralized matrix. These data suggest that the scaffold’s architecture and physicochemical properties alone are capable of triggering a cascade of differentiation events, even in the absence of exogenous growth factors[19,26,47].

Upon implantation of PHB/hydroxyapatite scaffolds seeded with MSCs into a critical-sized calvarial defect area in rats, we observed significant stimulation of bone tissue regeneration, accompanied by active vascularization of the scaffold[30,43,44]. Notably, the major volume of bone tissue formed between days 22 and 28, with the rate of osteogenesis in the presence of MSCs being 3.6-fold higher than when using acellular scaffolds[43]. These results clearly demonstrate that simultaneous stimulation of osteoinduction and vascularization—whether through the synergy of different cell subpopulations or through the creation of a favorable three-dimensional microenvironment—leads to far more efficient bone tissue repair than isolated targeting of either pathway. A promising direction in this field is the use of three-dimensional cellular systems, such as prevascularized spheroids[36,48] or triculture systems with controlled release of bioactive factors[26,47], which enable concurrent stimulation of osteogenesis, angiogenesis, and neurogenesis[44,47].

However, it must be emphasized that the efficacy of MSC-mediated bone regeneration critically depends on the immunological context. In a study by Takeshita et al[49] in 2017, it was shown that xenotransplantation of human MSCs in the form of three-dimensional cell aggregates (C-MSC) into a calvarial defect in immunocompetent mice did not result in bone regeneration and was accompanied by T-cell infiltration, whereas pre-treatment of the cells with interferon-γ (C-MSCγ), which induces indoleamine-2,3-dioxygenase expression, ensured successful defect repair[49,50]. Tellingly, in immunodeficient mice, both forms of the cell product induced comparable regeneration, directly pointing to the key role of immunomodulation in realizing the osteogenic potential of allogeneic MSCs[50,51]. This conclusion has fundamental implications for the interpretation of preclinical studies: The use of immunodeficient models, widespread in the literature, may mask the true immunogenic properties of cell products and lead to overestimation of their therapeutic efficacy[2,3,52].

When moving from bone tissue to other targets of tissue engineering, the dominant mechanism of action of pericyte-like MSCs changes substantially, reflecting the unique physiological characteristics of each tissue. In the case of cartilage tissue, which is normally avascular, the angiogenic potential of pericytes is not only unnecessary but may be counterproductive. Donnelly et al[40] in 2018 elegantly demonstrated that CD146+ pericytes isolated from remnant auricular tissues of microtia patients possess pronounced chondrogenic potential; however, precise control of TGF-β1 signaling is critically important for its realization. Adipose tissue-derived pericytes have likewise been successfully applied for cartilage repair, further supporting the chondrogenic potential of perivascular cells[53]. The use of a poly(ethyl acrylate)-based system enabling controlled presentation of TGF-β1 within a fibronectin matrix allowed for efficient chondrogenic differentiation while minimizing the risk of hypertrophy—a key adverse event in cartilage tissue engineering[40,45]. An alternative approach involves the application of mechanical stimulation to activate endogenous TGF-β within GelMA-based scaffolds, thereby inducing chondrogenesis without the addition of exogenous growth factors[54,55]. Thus, in cartilage tissue engineering, the primary value of pericyte-like MSCs lies not so much in their vascularization capacity, but rather in their chondrogenic plasticity and, potentially, in immunomodulatory properties mediated by adenosine production via the CD73/CD39 axis[12,20,23].

In skeletal muscles, pericytes perform a fundamentally different role, acting as key regulators of microcirculatory homeostasis and active participants in the regenerative process. As reviewed by Murray et al[31] in 2017, under conditions of muscle injury, pericytes contribute to the formation of a regenerative microenvironment through the secretion of trophic factors and modulation of local immune responses, and are also capable of direct differentiation into myofibers[54]. Their role in skeletal muscle regeneration appears to be centered on two key aspects: Stabilization of newly formed capillaries, which is critical for meeting the metabolic demands of regenerating tissue, and modulation of the inflammatory response via a secretome containing pro-angiogenic and immunoregulatory molecules[31,49,53]. The prospects for using MSCs and their secretome for skeletal muscle regeneration are discussed in detail in a recent review by Sandonà et al[56] in 2021, which emphasizes the importance of paracrine mechanisms and immunomodulation. Furthermore, modern biofabrication approaches, such as rotational wet spinning, allow for the creation of anisotropic microenvironments that guide myogenic differentiation of pericytes and the formation of contractile myotubes, opening new avenues for skeletal muscle tissue engineering[57,58]. Yang et al[4] in 2025 showed that muscle-derived CD146+ pericytes can support HSCs, indicating the existence of a specialized perivascular niche in muscle tissue that is potentially significant for extramedullary hematopoiesis[4,21,24].

The most unexpected and exciting data in recent years have come from studies on the role of pericytes in nervous tissue regeneration[9,59]. Sun et al[41] in 2025 demonstrated that following spinal cord injury, pericytes undergo profound changes in architecture and functional state, and that in the absence of therapeutic intervention, they constrict sensory axons on their surface, causing structural and functional impairments that hinder regeneration. However, local administration of PDGF-BB dramatically alters pericyte behavior: They form cellular bridges lined with fibronectin, which serve as a permissive substrate for axonal growth and promote the recovery of hindlimb motor function[44,60]. This phenomenon of in vivo “programming” of pericytes opens fundamentally new perspectives for tissue engineering of the central nervous system, shifting pericytes from the category of passive participants to that of active targets for therapeutic modulation[30,32,51].

Beyond classical applications in tissue engineering, pericyte-like MSCs demonstrate impressive potential in the therapy of ischemic conditions, where their angiogenic and vasculogenic properties come to the forefront[5,42,43]. Park et al[5] in 2024 identified a rare subpopulation of CD141+ VPCs in human bone marrow, which, when co-transplanted with BM-MSCs at a 2:1 ratio, promoted large artery restoration and limb salvage in a mouse model of severe critical limb ischemia. Notably, CD141+ VPCs do not express classical endothelial markers (CD31, VEGFR2, and VE-cadherin) in a quiescent state but acquire them during vascular structure formation in Matrigel, indicating their unique plasticity and capacity for vasculogenic differentiation in response to microenvironmental signals[5,44]. Complementary data were obtained by Kim et al[28] in 2026 using vascular multipotent stem cells from adipose tissue: These cells, expressing CD141, CD31, and VE-cadherin, in combination with adventitial stem cells exhibiting a pericytic phenotype (α-SMA+, transgelin+), ensured the formation of mature vascular structures and limb salvage in a critical ischemia model[28,29]. This dual-cell therapy holds promise for clinical application in severe peripheral artery disease and diabetic ulcers, where existing therapeutic options remain severely limited.

Thus, an analysis of the application of pericyte-like MSCs in various areas of tissue engineering demonstrates that their therapeutic value is determined not by a single universal mechanism, but by a tissue-specific combination of angiogenic, osteogenic, chondrogenic, myogenic, neuroprotective, and immunomodulatory properties[2,3,11]. The CD45-/CD73+/CD39+/CD146+ phenotype, characterized by Martin and Gullo[14], represents a promising cellular platform for various tissue engineering applications[12]; however, the dominant functional mechanism and, consequently, the optimal application strategy will differ substantially depending on the tissue context, requiring a tailored approach to the design of cell-based products[13,31,40]. Furthermore, the potential competition between distinct MSC differentiation programs must be considered: It is possible that active osteogenic, chondrogenic, or neurogenic differentiation could negatively impact the ability of these cells to support HSCs and their angiogenic potential, which necessitates further systematic investigation[29].

THE ROLE OF IMMUNITY AND STRESS

If the preceding sections were devoted to the accumulation of facts, we will now take the liberty of pausing to pose questions that, in our view, often remain outside the frame of optimistic preclinical reports. Do we indeed isolate the very cells that perform their function in vivo, or are we dealing with artifacts of stress and inflammation induced by our own manipulations[19,54]? How stable and predictable are the immunomodulatory properties upon which we place so much hope[23,52]? And will the plasticity of these cells not present an unwelcome surprise in the long term[55,57]? Answers to these questions are not merely of academic interest; they directly influence the choice of strategy and the interpretation of results when transitioning to clinical trials (Figure 3).

Figure 3 Donor- and microenvironment-related variables influencing pericyte-like mesenchymal stem cell functionality in tissue engineering. VEGF: Vascular endothelial growth factor; BMP: Bone morphogenetic protein; BDNF: Brain-derived neurotrophic factor; G-CSF: Granulocyte colony-stimulating fact; TNF: Tumour necrosis factor; IL: Interleukin; IFN: Interferon; SCF: Stem cell factor; TGF: Transforming growth factor; PDGF: Platelet-derived growth factor; MSC: Mesenchymal stem cell.

The notion that “pericyte-like” MSCs can be obtained from peripheral blood, particularly following granulocyte colony-stimulating factor (G-CSF) mobilization, appears highly attractive for clinical applications[14]. Indeed, G-CSF triggers a cascade of events that leads to the release into the circulation of not only hematopoietic but also stromal cells[32,61]. Several studies have shown that such mobilized MSCs possess an even more pronounced anti-inflammatory potential compared to their “sedentary” bone marrow counterparts, which has been linked to the activation of specific signaling pathways within them[20,61]. Moreover, G-CSF is capable of directly modulating the transcriptome and functional properties of MSCs, altering their migratory capacity and synthesis of extracellular matrix components[19,62].

However, this brings us to a fundamental question: Is the appearance of such cells in the bloodstream not rather an alarm signal than a scheduled mobilization event? It cannot be excluded that the G-CSF stimulation procedure itself, and the subsequent “stress” associated with entry into the circulation and exposure to hemodynamic forces, fundamentally alter the cell, endowing it with a phenotype and functions not characteristic of resident niche pericytes[19,63]. A hypothetical model illustrating how inflammatory and stress signals—including G-CSF, interleukin-1β, and catecholamines—may drive the egress of CD45-/CD73+/CD39+/CD146+ cells from the bone marrow into the bloodstream is presented in Figure 4. In this sense, “pericyte-like” MSCs in the blood may be not so much an analog of progenitor cells from the perivascular space, but rather a phenotypic “artifact” of a potent inflammatory response. To dissociate the effect of mobilization itself from a non-specific stress reaction, it would be exceedingly informative to conduct a control experiment in which donors (or animals) were administered a placebo mimicking the side effects of G-CSF, followed by a comparison of the properties of cells obtained from the blood in these two groups. This would help clarify precisely what cellular material we are dealing with and the extent to which it is suitable for regenerative medicine applications.

Figure 4 Mobilization of pericyte-like mesenchymal stem cells into the bloodstream under stress and inflammation: A hypothetical model. G-CSF: Granulocyte colony-stimulating fact; IL: Interleukin; MSC: Mesenchymal stem cell.

The immunomodulatory capacity of MSCs is their hallmark and one of the primary drivers of clinical interest in these cells[3,11]. A special place here is occupied by the adenosine axis (CD73/CD39), which is believed to be a key mechanism in the CD45-/CD73+/CD39+/CD146+ subpopulation[12,20]. Indeed, adenosine production is a powerful tool for suppressing T-cell responses[20,32]. However, it is hardly appropriate to reduce the entire immunosuppressive potency of MSCs to this single pathway alone. The arsenal of these cells is far broader: It includes prostaglandin E2, indoleamine-2,3-dioxygenase[50,51], TGFβ, and numerous other soluble and contact-dependent factors[32,54].

The question that arises here pertains to the stability and predictability of this profile. How long will cells isolated based on the CD73+/CD39+ signature retain this activity during scale-up in culture? Will they not, under the influence of ex vivo expansion conditions, begin to rely on other, perhaps less effective, mechanisms? And, more importantly, will their immunomodulatory profile not undergo dramatic changes depending on the tissue microenvironment that they encounter post-transplantation? This donor-specific and contextual variability remains one of the primary challenges for the standardization of cell-based products, and without addressing it, discussing reliable and reproducible therapy is exceedingly difficult[2,3].

Finally, perhaps the most intriguing and disconcerting aspect is the phenotypic plasticity of MSCs in vivo. In an elegant study by Hakim et al[51] in 2019, it was demonstrated that following transplantation into a spinal cord injury site, MSCs do not remain “themselves” but undergo a fundamental transformation, acquiring features characteristic of immune cells. Plasticity is not limited to immune phenotypes; pericyte-like cells from the umbilical artery can form neurospheres and differentiate into neural lineages, further illustrating the contextual reprogramming of these cells[22]. This is not merely an academic observation; it compels us to re-evaluate the very essence of the therapeutic action of MSCs. We are accustomed to thinking of them as “conductors” of regeneration, orchestrating the recipient’s cellular ensemble. But what if, in fact, they themselves become part of that ensemble, directly engaging in complex immune interactions?

On one hand, such plasticity could be beneficial, enabling cells to more effectively modulate inflammation and promote regeneration[35,51]. On the other hand, it introduces new risks. If MSCs can acquire an immune phenotype, does this not enhance their own immunogenicity? Could they, upon entering an unfavorable microenvironment (e.g., within a tumor), undergo undesirable transformation? Such concerns are particularly relevant in chronic inflammatory conditions like diabetic microangiopathies, where pericyte plasticity, while offering regenerative opportunities, may also contribute to vascular dysfunction[64]. And how, ultimately, can we track and control the fate of these cells within the patient’s body? These questions remain open, and the search for answers to them will likely shape the research agenda in cell therapy for the coming years.

CONCLUSION

Thus, the advanced work of Martin and Gullo[14] and further studies on the functional properties of the CD45-/CD73+/CD39+/CD146+ cell subset will allow the mechanisms for supporting the expansion of HSCs to be revealed and, in general, the full therapeutic potential of this MSC subtype to be unlocked. This will enable the development of both novel cell-based therapeutics and tissue-engineered products.

ACKNOWLEDGEMENTS

This work was performed within the state assignment of Lomonosov Moscow State University.