Benjamin Rouget was the first to describe and give his name to a particular type of cell within the capillary basement membrane with projections protruding to enwrap the vessel wall (1). In 1923, Zimmerman renamed these Rouget cells “Pericytes.” Under this broad name, they included many forms of cells from different tissue origins located around capillaries (2). Then, research on pericytes entered a hibernation period until 20 years ago, when the pericytes’ regenerative potential started attracting translational interest (3, 4). Nonetheless, a discrepancy persists in the literature regarding pericytes vs. other cardiovascular cells. For instance, pericytes are plentiful in the heart. However, there were only 73 publications on cardiac pericytes in PubMed during 2021 (out of 687 publications retrievable using the general term “pericytes”) compared with 5,400 for cardiomyocytes and 1,150 for coronary endothelial cells. This data highlights the need for more research on this cell population.

Several teams have reported technologies for the isolation and derivation of pericytes. Culture expansion allows the generation of stocks for use as a product for cell therapy and tissue engineering. Nonetheless, the consensus on pericyte-related techniques remains far behind that of human endothelial cells (ECs) (5–7). In particular, the diversity of pericytes’ properties and antigenic expression in different organs, together with the overlapping with other perivascular mesenchymal cells, represents a significant obstacle to using one method that fits all experimental conditions. Ad hoc standard operating procedures (SOP) may better accomplish distinct scientific/medical goals. For instance, pure pericyte products are necessary for obtaining regulatory approval for patients’ treatment. Semi-purified preparations are more suitable for deciphering the transition of pericytes across different antigenic and functional phenotypes in response to environmental stimuli. The main characteristics of pericytes are illustrated in Figure 1 and relevant text sections.

This review article aims to present, compare, and discuss the experience of deriving pericytes from different organs and species. We initially report current nomenclature, functional roles, and antigenic profiles that distinguish pericytes from other perivascular cell populations. The main body of the review is then devoted to (i) documenting the pros and cons of current techniques for isolation, expansion, and characterization of bona fide pericytes, (ii) suggesting avenues for standardization and harmonization, (iii) upgrading protocols for clinical manufacturing, and (iv) prospecting the potential of novel single-cell technologies in advancing pericyte research.

Developmentally, pericytes in the cephalic region and thymus originate from the neuroectoderm. The ones found in the liver, lungs, and heart derive from mesothelium. The mesoderm gives rise to pericytes in other organs (such as kidneys, liver, and pancreas) (8–10). Intriguingly, Yamazaki et al. proposed that a subgroup of pericytes may also derive from the hematopoietic cell lineage (11). Tracing studies may help reveal additional embryological contributors to vascular pericytes.

In the adult organism, microvascular pericytes are classically located within the basement membrane in direct contact with capillary ECs in different organs. In addition, pericytes are present in the vascular adventitia, the blood vessel’s outer layer near the vasa vasorum (3, 12). The nomenclature of these intravascular pericytes is inconsistent: the term “adventitial stromal cells/CD34 + cells” refers to the in situ expression of CD34, a transmembrane glycoprotein, while “adventitial pericyte-like progenitors” denotes their plasticity and positivity for typical stem cell markers (12–15). For clarity, we will refer to these cells as adventitial pericytes.

Pericytes are essential for vasculature’s physiological stability, as exemplified by the prominent pericyte-endothelial interactions forming the blood-brain barrier (16). Reduction in pericytes abundance, pericyte detachment from capillaries, and accrual of senescence markers are associated with microvascular fragility and dysfunction, as observed in aging, ischemic disease, fibrotic and neurodegenerative disorders, diabetes, tumors, and COVID-19 (17–21). Moreover, capillary pericytes modulate blood flow in different organs. Mice with brain pericyte ablation reportedly developed an acute blood-brain barrier breakdown, severe blood flow decrease, and rapid neuron loss (22). Pericyte contraction participates in the no-reflow phenomenon, the failure of blood to reperfuse an ischemic tissue after the vascular obstruction has been removed or bypassed (23–26).

In addition to their vascular functions, pericytes exert immunomodulatory tasks, sensing inflammatory stimuli through pattern-recognition receptors, and promoting monocyte and neutrophil migration and homing via a PDGFRβ signaling mechanism (27, 28). They also express ligands interacting with activated T cell receptors during adaptive immunity responses (29).

Lastly, pericytes play critical roles in tissue restoration, promoting, and stabilizing reparative neovascularization, as reviewed in Navarro et al. (30). Accordingly, regenerative therapies using exogenous pericytes have been preclinically tested (31–37). Pericytes were delivered either as dispersed preparations or as a part of a “tissue-engineered” product to models of myocardial infarction (MI) (33, 36, 37), congenital heart disease (34, 35), limb ischemia (38), muscular dystrophy and skeletal muscle injuries (39–42), blood-brain barrier disorders (43–48), diabetic retinopathy (49–51), and kidney fibrosis (52). These preclinical studies have paved the way for future therapeutic applications in human diseases.

Like other mesenchymal cells, pericytes are positive for CD105, CD29, CD44, and CD90 and negative for CD45 and CD31. They typically express Platelet-Derived Growth Factor Receptor beta (PDGFRβ), the cell surface glycoprotein MUC18 (CD146), GTPase signaling 5 (RGS5), chondroitin sulfate proteoglycan 4 (NG2), myosin heavy chain 11 (Myh11), and the transmembrane metalloprotease CD13 (12, 53–55).

The expression of antigenic markers can vary according to organ location. Capillary and adventitial pericytes express both NG2 and PDGFRβ. However, capillary pericytes are positive for CD146, whereas adventitial pericytes are positive for CD34 and negative for CD146 (3, 12, 56, 57). Kramann et al. and Crisan et al. illustrated a difference in the antigenic profile of pericytes according to their vascular site. They classified the pericytes into two types. Capillary (type I) pericytes (NG2⁺/Desmin^–/α-SMA^–) and pre-capillary (type II) pericytes (NG2⁺/Desmin⁺/α-SMA⁺) (58, 59). Bergers et al. identified post-capillary (type III) pericytes (NG2^–/αSMA⁺) (60). Moreover, coronary artery pericytes were classified as endocardial-derived (NG2^–/ PDGFRβ⁺/αSMA^–) and epicardial-derived (NG2⁺/PDGFRβ⁺/αSMA⁺) (61, 62). Three different subgroups of aortic pericytes were identified based on similarities to fibroblasts (PDGFRβ⁺/PDGFRα⁺/CD34⁺), VSMCs (PDGFRβ⁺/αSMA⁺), and neural cells (PDGFRβ⁺/PLP1⁺/CNP⁺) (63).

Pericytes can change phenotype when transiting from a quiescent to an activated state. NG2 distinguishes topical pericyte properties such as the capacity to control vascular permeability in the brain (64–66), participation in myofibroblast differentiation during the process of cardiac and renal fibrosis (67, 68), and progenitor cell activity in the coronary microvasculature (69). The concurrent expression of CD133, CD34, and Kinase Insert Domain Receptor (KDR) identifies fetal aorta-derived pericytes’ ability to differentiate into endothelial or muscular lineages (15). Adult aortic pericytes express the stemness marker SRY (sex determining region Y)-box 2 (Sox2) (12). and microvascular cardiac pericytes express OCT3/4, GATA−4, and Homeobox protein NANOG (32). The upregulation of another pluripotency gene, Krüppel-like factor 4 (KLF4), marked the switch of perivascular cells to a less differentiated state in tumors (70).

In angiogenesis, the transition from quiescent to actively proliferating pericytes is accompanied by profound changes in the antigenic and secretory profile (71). The expression of alpha-smooth muscle actin (α-SMA) and the downregulation of protein tyrosine kinase Tie2 reportedly mark the pericyte participation in pathological angiogenesis (72–74). Using RNA-Seq transcriptomics, we unveiled new markers distinguishing the differentiation stages of human cardiac pericytes into pro-angiogenic, contractile cells. Naive pericytes express the Cell Adhesion Molecule 3 (CADM3). The expression of the angiogenesis-related Cellular Retinoic Acid–Binding Protein 2 (CRABP2) and Aquaporin 1 (AQP1) distinguished contractile pericytes from naive pericytes and coronary VSMCs (55). In addition, we reported that epigenetic modifications can influence the pericyte pro-angiogenic activity (75). Transplantation of human adventitial pericytes to a murine limb ischemia model improved perfusion recovery and muscular angiogenesis. The therapeutic effect varied among different donor lines. Whole genome screening showed that this variation was significantly associated with the DNA methylation state at the promoter and gene body levels (75).

Gubernator et al. identified two types of skeletal muscle pericytes relevant to adipogenesis and myogenesis (76). Type I pericytes’ antigenic profile (NG2⁺/Nestin^–/PDGFRα⁺) is linked to a differentiation fate to adipocytes and is associated with fat accumulation (77). Type II expressional profile (NG2⁺/Nestin⁺/PDGFRα^–) is linked to myocyte commitment and associated with skeletal muscle regeneration (78).

The microenvironment within or in proximity to the blood vessel wall, the perivascular niche, contains mesenchymal stromal cells (MSCs), fibroblasts, and pericytes. Perivascular MSCs express CD146, CD105, alkaline phosphatase (ALP)⁺, Stro-1, and VCAM1 in humans, and CD105, PDGFRα, Sca1, CD44, CD29, and VCAM1 in the mouse (79). It remains unclear whether MSCs and pericytes have a lineage relationship or represent functionally distinct mesenchymal populations that share a perivascular location. Some studies support the possibility that pericytes constitute a more primitive version of MSCs (15, 80), while others suggest that all MSCs are pericytes with some narrow exceptions (81). In line with the latter possibility, a study from Crisan demonstrated that pericytes isolated from different organs express almost all the known markers of MSCs, especially CD10, CD13, CD44, CD73, CD90, and CD105 (82).

Moreover, a strict relationship exists between fibroblasts and pericytes. Fibroblasts are mesenchymal cells in the interstitial space of all organs (83). They express mesenchymal markers such as vimentin, PDGFRα, and type I collagen, along with some tissue-specific markers, such as Transcription Factor 21 (TCF21) and periostin (84). In the brain, perivascular fibroblasts surround cortical penetrating arterioles and their branches in the arteriole-capillary transition zone. Still, they are absent in the capillary zone where pericytes prevail (85). Recent single-cell transcriptomics studies in mice showed that perivascular fibroblast-like cells in the adult mouse brain express unique markers compared with other mural cells (86–88). In the heart, the periostin-expressing myofibroblast originates from tissue-resident fibroblasts and plays critical roles in adaptive healing and fibrosis, according to genetic lineage tracing of cardiac cells (84). A study in the Zebrafish, using a combination of in vivo live imaging, genetic ablation, and CRISPR mutant analysis, demonstrated that perivascular fibroblasts play dual roles in stabilizing nascent intersegmental vessel sprouts from the dorsal aorta and later functioning as pericyte progenitors (89). Furthermore, Birbrair et al. distinguished two bona fide pericyte subtypes in the skeletal muscle interstitium of mice, naming them type-1 (Nestin-GFP^–/NG2-DsRed⁺) and type-2 (Nestin-GFP⁺/NG2-DsRed⁺) pericytes. In vitro, type-1 pericytes were profibrotic, and type-2 pericytes were myogenic. The latter fate was confirmed after transplantation of type-2 pericytes in young mice and, to a lesser extent, in older mice. Moreover, type-2 pericytes participate in normal and tumoral angiogenesis (77, 78). Therefore, transition states and shared functions between pericytes and perivascular fibroblasts may coexist in disease states.

Given that there is no single-specific pericyte marker, adopting the concept of “Identifying Package’ seems more plausible (82).

The different steps of pericyte derivation and purification are illustrated in Figure 2.

Pericytes have been isolated from different tissues.

Although human cells constitute the final therapeutic product for patients, animal-derived cells represent a suitable surrogate in preclinical safety and efficacy studies (140). Using donors and recipients from the same species, ideally from the same litter, eliminates the need for immunosuppression, which is instead required when transplanting human cells into an animal model (33). Recent technological progress makes swine-derived cells, tissues, and organs attractive for immune-compatible xenotransplantation. The cells are genetically modified to remove the immunogenic factors. Human transgenes can be expressed to suppress rejection (141).

Lack of uniformity in cell manufacturing between human and animal species might influence the interpretation and translatability of preclinical results. Therefore, robust approaches for isolation, culture, expansion, characterization, and appropriate in vivo testing of pericytes in larger animal species are of utmost importance.

Induced Pluripotent Stem Cells (iPSC) derived through reprogramming somatic cells with pluripotency factors (OCT3/4, SOX2, c-MYC, KLF4) can be differentiated to a specific cell line, including pericytes. 3D-culture systems of iPSC-derived cardiovascular cells have provided a novel in vitro model to study the interplay among different cell types and potentially develop new personalized treatments for congenital and acquired cardiac diseases (116, 143–145).

The developmental origin of cardiac pericytes is thought to be the epicardium. Differentiation of iPSCs to epicardial cells can be achieved through the modulation of Wnt/β-catenin signaling and treatment with pericyte growth factors, such as PDGF-BB (144, 146, 147). Szepes et al. applied a mesodermal induction using BMP4, VEGFA, and activation of Wnt/β-catenin signaling with further VEGFA and TGFβ inhibition. FACS staining and sorting for PDGFRβ and CD31 markers and further expansion were carried out for antigenic characterization, gene expression, and function in an endothelial tube formation assay. A tri-cultured bioartificial cardiac tissue model was generated using iPSC-derived pericytes, cardiomyocytes, and ECs to study vascularization and fibrosis in vitro (145). Faal et al. reported two protocols using mesoderm and neural crest to obtain brain pericytes from iPSCs generated from healthy subjects and Alzheimer’s disease patients. Briefly, human iPSCs were cultured on a Matrigel-coated plate in a mesodermal induction medium with growth factors. The activation of Wnt/β-catenin signaling initiated neural crest differentiation through specific medium and supplements. Mesodermal and neural crest-derived cells were differentiated in pericytes seeded on gelatin or Matrigel-coated plate. Both pericyte types were passaged, split, and kept in the pericyte medium for their characterization and functional trans-endothelial electrical resistance assay with ECs (146). Jamieson et al. generated pericytes from human iPSCs obtained from mesodermal lineage. In brief, iPSCs were cultured on a feeder layer of mouse embryonic fibroblasts and gelatin-coated plates. For the differentiation, iPSCs were detached with EDTA, strained, and seeded on a collagen-coated plate using α-MEM, FBS, and β-mercaptoethanol. Following further detachment, they were grown in EC growth medium, TGFβ inhibitor, and VEGF to promote vascular cell specification and proliferation. After an enzymatic dissociation, DMEM + FBS was added to an uncoated plate to support the pericyte selection. Generated iPSC-derived pericytes were used for developing a microvessel model of the blood-brain barrier (148).

It has been shown that fewer than one-third of biomedical papers can be reproduced, resulting in associated costs of about $28 billion each in the US only, mainly attributable to the wasted follow-on work that the initial study has inspired (149, 150). Reproducibility of results from peer-reviewed papers by researchers at pharmaceutical companies failed at rates upwards of 75% (151). This reflects the inconsistent use and communication of standardized processes to achieve efficiency, quality, and uniformity of performance. The main body of information on pericytes methodologies can be retrieved from original papers, which show large variability regarding the depth of details. The following paragraphs are devoted to suggesting improvements to current methodologies.

This review highlighted the significant efforts made by different teams to develop effective protocols of isolation and in vitro expansion of human pericytes. A consensus on these protocols and the integration of methodologies will help speed up research at the highest level of reproducibility and translatability. Moreover, additional research is needed to dissect the functional diversification of pericytes according to their zonal distribution and the surrounding microenvironment. More work is also necessary to trace pericytes in tissues during vascular remodeling. Expanding and refining the atlas of different perivascular niches at the single-cell level will clarify the enigmatic phenotype of pericytes and the relevance of expressional deviations in response to organ injury. This novel technological approach should be integrated into the dimensions of space and time, for instance, by clustering dynamic gene expression profiles from single-cell RNA-Sequencing databases and associating transcripts to functionomics analysis. Most studies on pericytes have been performed in rodent models. More research in large animal models is needed to demonstrate therapeutic utility.

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

PM was responsible for the final writing and the data presented. All authors conceived and designed the article, read, and approved the final manuscript.