Pericytes for Therapeutic Bone Repair

Abstract

Besides seminal functions in angiogenesis and blood pressure regulation, microvascular pericytes possess a latent tissue regenerative potential that can be revealed in culture following transition into mesenchymal stem cells. Endowed with robust osteogenic potential, pericytes and other related perivascular cells extracted from adipose tissue represent a potent and abundant cell source for refined bone tissue engineering and improved cell therapies of fractures and other bone defects. The use of diverse bone formation assays in vivo, which include mouse muscle pocket osteogenesis and calvaria replenishment, rat and dog spine fusion, and rat non-union fracture healing, has confirmed the superiority of purified perivascular cells for skeletal (re)generation. As a surprising observation though, despite strong endogenous bone-forming potential, perivascular cells drive bone regeneration essentially indirectly, via recruitment by secreted factors of local osteo-progenitors.

Introduction

Pericytes are stellate cells in close contact with endothelial cells and embedded within a basal lamina, which form a discontinuous layer in capillaries (<10 μm diameter), and continuous one around microvessels (diameter 10–100 μm) [1]. First described in 1873 by C. Rouget as, visionarily, contractile cells regulating blood flow, it is Zimmermann who coined the term pericyte in 1923 to describe cells, also known as mural cells, structurally supporting the vasculature [2]. Promotion of angiogenesis, blood vessel diameter regulation [3], and maintenance of vascular integrity and permeability [4] are the main functions attributed to pericytes, through direct cell contact and communication.

As early as the 1970s were pericytes suggested to be also involved in tissue regeneration [5]. It was, however, not before the first decade of this century that definitive experimental evidence was gained that pericytes are native ancestors of mesenchymal stem cells (MSC), the existence of which had been previously documented exclusively in long-term cultures of vascularized organs [6]. Pericytes can differentiate into chondrocytes, adipocytes and osteocytes, regardless of their tissue of origin [6–8], as well as skeletal and cardiac muscle [6, 9], and myofibroblasts at the origin of pathologic fibrosis [10, 11]. Pericytes also support hematopoiesis [12–16] and can modulate immune-inflammatory reactions [17].

Among the potential uses of pericytes/progenitor cells for tissue engineering, the application to bone tissue is most commonly studied [18]. The bone is a richly vascularized organ, and the reaction to bone injury includes the processes of osteogenesis and vasculogenesis that go hand in hand. The theory that mural cells participate in endogenous bone tissue repair has long been posited. Before the advent of cell lineage tracing, the use of intravascular dyes that label mural cells suggested that pericytes participate in osteochondral repair [1, 19]. Later studies using smooth muscle actin (SMA) reporter animals also suggested that endogenous mural cells give rise to bone cells after fracture [20]. SMA is a non-specific marker which labels some pericytes, smooth muscle, and fibroblasts/myofibroblasts. Therefore, to our knowledge, the direct participation of bone-associated pericytes in repair has never been definitely shown. Nevertheless, these observations of the reparative potential of endogenous SMA+ cells, combined with the known mesenchymal progenitor cell properties of human pericytes [6], gave impetus for the use of exogenous pericytes for bone tissue repair. The osteogenic potential of human pericytes and other perivascular cells has been examined in both ectopic and orthotopic models. These findings are briefly reviewed below.

Identification and Purification of Perivascular Cells for Bone Repair

The possibility of using human pericytes/perivascular progenitor cells to speed bone repair was made possible by prior studies that used the cell surface marker CD146, also known as Mel-CAM (melanoma cell adhesion molecule), for the identification and purification of pericytes ([21, 22]; see also [23] for a review). Of note, CD146 expression is by no means specific to pericytes, and as a heterophilic cell-cell adhesion molecule, it is often upregulated when diverse cell types adopt a location on the outside aspect of the endothelial cell [24]. CD146 is also expressed by endothelial cells [25], vascular smooth muscle cells (VSMC) [26], fractions of lymphocytes [27], and tumor cells [28]. Therefore fluorescence-activated cell sorting for a combination of cell surface markers (CD146+CD34-CD31-CD45-) is most commonly used by our group to identify a pericyte among uncultured stromal populations [6, 29, 30]. The potential non-specific identity of CD146+ progenitors has led some investigators to favor the term tissue-specific progenitor cells over pericytes. A presumably analogous CD146+ progenitor cell can also be identified among culture-derived cell populations [22]. CD146+ pericytes/progenitors have been examined for their bone-forming potential alone [31] or in combination with other perivascular mesenchymal progenitor cells derived from the tunica adventitia and typified by CD34 expression and absence of other endothelial cell and pericyte markers (termed adventitial cells or adventitial progenitor cells) [29, 32]. When CD146+ pericytes and CD34+ adventitial progenitor cells are used in combination, they are most commonly referred to as perivascular stem cells or perivascular stromal cells (PSCs) [29], referring to their shared perivascular location. At least in the context of bone tissue engineering, PSCs are most commonly derived from subcutaneous adipose tissue [33]. The rationale for adipose derivation is based principally on the easy access and dispensability of this tissue depot. Once in culture, PSCs are able to undergo differentiation toward multiple mesenchymal lineages under appropriate culture conditions (Fig. 3.1), including osteoblastogenic (Fig. 3.1a–c) and adipo-cytic cell fates (Fig. 3.1d). Importantly the umbrella term of PSC is used, despite the clear understanding that these perivascular progenitor cells differ in their location and cellular morphology within the vascular wall, markers for in situ detection, frequency within different tissues, and gene network profiles [34]. The functional relevance of cellular differences between CD146+ pericytes and CD34+ adventitial progenitor cells for bone repair outcomes is as yet not known.

Fig. 3.1.

Differentiation of human perivascular stem cells in vitro and stimulation of an osteochondrogenic program in vivo. (a, b) PSCs are a multipotent progenitor cell type in vitro. (a) Human PSCs were cultured in the presence of osteogenic differentiation medium. Frank confluent mineralization was observed among PSC under inductive culture conditions (Alizarin red staining shown). (b) Conversely, intracellular lipid accumulation can be visualized within PSC under appropriate adipogenic conditions (Oil red O staining shown). (c, d) PSC implantation in a rat spinal fusion model induces a combination of intramembranous and endochondral bone formation. (c) Woven bone formation, and prominent bone lining osteoblasts in areas of intramembranous bone formation by PSC. (d) Chondrocyte hypertrophy and mineralization in areas of endochondral bone formation induced by PSC. Scale bar: 25 um

We have described methods general to all human organs: fresh tissues undergo mechanical and enzymatic digestion prior to cell isolation and immunolabeling for FACS, with at least one pericyte, endothelial, and hematopoietic cell marker. Sorted populations are seeded in endothelial growth medium 2 (EGM-2) in gelatin-coated plates and passaged using 20% FCS-supplemented Dulbecco’s modified Eagle’s medium [6]. The immunophenotype CD146+CD34-CD31-CD45-CD56- successfully isolates pericytes in multiple tissues including the skeletal muscle, bone marrow, white adipose tissue, placenta, pancreas, umbilical cord, heart, kidneys, infrapatellar fat pad, and liver [6, 9, 14, 18, 29, 30, 35–38, 39–41]. Of note, the same immunophenotype has been used to isolate pericytes from other mammalian species, including dog [42], sheep [43], and horse [44, 45], offering large animal models of perivascular cell-mediated tissue regeneration.

Pericytes and Ectopic Bone Formation

Animal models of ectopic bone formation have been used to confirm the capacity for in vivo osteogenic differentiation of implanted human pericytes. Human adipose tissue (AT)-derived CD146+ pericytes have been observed to directly ossify when implanted in a SCID (severe combined immunodeficiency) mouse muscle pouch [23]. Inconspicuous bone is produced when AT pericytes are implanted on a collagen sponge carrier, which represents a relatively inert substance with little osteoinductive properties [29]. In contrast, when AT pericytes are implanted intramuscularly using an osteoinductive demineralized bone matrix (DBM) carrier, robust bone formation is observed [29]. In somewhat similar observations, other groups have shown that CD146+ AT-derived progenitor cells do not form significant bone when implanted in a subcutaneous ossicle model [22]. In contrast, bone-associated (bone marrow or periosteum) CD146+ progenitor cells drive robust bone formation in the same subcutaneous ossicle model [22]. Head-to-head comparisons of AT-derived pericytes and adventitial cells from the same patient sample have been performed [23]. Both perivascular cell types induce vascularized ectopic bone and without substantive differences in the degree of bone formation [23]. This pilot study demonstrated that pericytes and adventitial cells have a similar bone-forming potential and laid the framework for later studies in which these two cell populations were combined. Next, experiments have been performed in which uncultured AT-derived PSCs (combined pericytes + adventitial cells) were implanted intramuscularly and compared with an unsorted/uncultured stromal population from the same patient’s adipose sample (termed stromal vascular fraction, SVF) [29]. Here, a DBM scaffold was again used. Results showed that independent of cell number, AT-derived PSC led to more robust intramuscular ossification in comparison to SVF from the same patient sample using quantitative metrics of bone formation by micro-computed tomography, histomorphometry, and select immunohistochemical markers of the bone [29]. Increased bone formation among AT-PSC implants was accompanied by a significant increase in vascularity of the implant site, accompanied by increased elaboration of VEGF (vascular endothelial growth factor) [38]. Ectopic bone formation induced by AT-PSC was also associated with an altered inflammatory milieu within the early wound environment [17]. Overall, these studies showed that AT-derived pericytes or AT-derived adventitial cells either alone or combined result in significant ectopic bone formation. Moreover, and for the first time, it was observed that these FACS-purified cell populations outcompete unpurified stromal cell populations from the same patient sample in terms of bone-forming efficacy.

Pericytes in Calvarial Defect Regeneration

The extent to which AT-derived PSC can induce bone repair was first examined in a mouse calvarial defect model [33]. Here, equal numbers of unpurified SVF or PSC from the same patient’s adipose tissue were implanted in a non-healing, circular, full-thickness calvarial defect of the parietal bone. Cells were implanted on a hydroxyapatite-coated polymeric scaffold for an additional osteoinductive effect. Similar to intramuscular implants, radiographic and histologic analysis showed AT-derived PSC led to a significant increase in bone regenerate at the defect site over an 8-week time course. In comparison, unpurified SVF from the same patient had no statistically appreciable benefit in comparison to a scaffold without cells. In this xenograft model, sparse but present human-specific antigens were detectable within the healing bone defect. Again, and in similarity to intramuscular studies, bone defect vascularity was significantly increased with PSC treatment. Thus, across both ectopic and bone repair models, AT-derived human PSCs have conserved features upon transplantation, including pro-osteogenic/pro-vasculogenic effects of a greater magnitude than unpurified stromal cell fractions. Whether these findings correlated with the enrichment of osteoinductive PSC, or conversely the elimination of an inhibitory cell type within the heterogeneous stroma of SVF, is still a matter of conjecture.

Pericytes in Spinal Fusion

Spinal fusion represents a more functionally demanding environment for a bone graft substitute and represents an assay for the production of contiguous and biome-chanically sound bone tissue. The use of AT-derived human PSC as a cellular therapy for bone grafting has been validated in a rat posterolateral lumbar spinal fusion model. In these studies, human AT-PSC implantation was performed across three cell densities in rats, using a DBM scaffold as a moldable carrier. PSC demonstrated a dose-dependent increase in ossification, increase in bone deposition, increase in measurements of bone strength, and complete fusion between lumbar bone segments in all rats [46]. In this model, both intramembranous (Fig. 3.1c) and endochondral bone formation (Fig. 3.1d) was spurred on by PSC implantation. Like in other studies, new bone regenerate was observed to be a product of both direct osteodifferentiation and host osteoblastogenesis. Like the calvarial defect model, paracrine-mediated bone formation of rat origin predominated [46]. In follow-up studies, Lee et al. extended these observations to rats rendered osteoporotic by ovariectomy. Here, increased numbers of implanted human AT-PSC were required to surmount the hormonal changes of estrogen withdrawal [47].

Pericytes for Non-union Fracture Healing

Atrophic non-union is associated with biological failure of fracture healing. Animal studies have shown the vascular ingrowth within atrophic non-union is much reduced at early timepoints [48]. In combination with the observation, the mesenchymal progenitor cell content within fibrous non-unions is reduced, and the proliferative and osteogenic capabilities of these non-union derived cells are likewise reduced [49]. CD146+ AT pericytes were examined in a well-established model of rat tibial atrophic non-union [48, 50]. Human AT pericytes were percutaneously injected 3 weeks after the establishment of fibrous non-union. Results showed that pericyte injection increased fracture callus size and increased mineralization, eventually resulting in increased bone union [50]. Like in other models, the efficacy of pericyte treatment was primarily a paracrine phenomenon, and in fact species-specific immunohistochemistry failed to later identity residual human cells. These data suggest that at least in the inhospitable microenvironment of atrophic nonunion, the benefit of pericytes primarily resides in their trophic abilities.

Discussion

Pericytes have crossed the limits of vascular biology and entered the field of regenerative medicine via their mesenchymal stem cell-cultured progeny. Advantages of using conventional MSCs include the simplicity of the derivation method and possibility to obtain large numbers of cells. On the negative side, MSCs are the cultured product of a heterogeneous mixture of unseparated cells, and in vitro growth involves cell exposure to animal proteins, hence chances of xenogeneic immunization, and entails risks of bacterial contamination and genetic instability. There have been occasional reports of MSC malignant transformation [51]; principally, it is increasingly accepted that MSC recruitment to the tumor stroma can favor cancer development [52]. For all these reasons, it might be beneficial to use purified, non-cultured perivascular cells in place of culture-derived MSCs for cell therapies. Bone repair has been the first envisioned therapeutic use of pericytes and adventitial perivascular cells. Bone structure is relatively simple, and targeted interventions, such as non-union fracture reduction or spine fusion, are usually not life-threatening, providing convenient models in which to gain a proof-of-concept demonstration of the therapeutic usability of perivascular presumptive MSCs. Importantly, PSCs also appear to represent a reliable source of autologous therapeutic cells, regardless of age, gender, and body mass index [30]. Experimentally, as described in this article, pericytes and adventitial cells purified from human or canine adipose tissue exhibited dramatic bone-forming potential in all autologous and xenogeneic in vivo assays performed, including calvarial regeneration and muscle pouch osteogenesis in mice, spine fusion in rats and dogs, and non-union fracture repair in rats. In these tests, PSCs performed at least as well as conventional MSCs are significantly better than the plain stromal vascular fraction. The bone produced following PSC transplantation was histologically normal and mechanically competent. These data illustrate the propensity of perivascular cells to differentiate along the bone cell lineage: culturing human adipose tissue-derived pericytes on a hard hydrogel substrate was sufficient to induce osteogenesis [53], and transcriptome analysis in single adventitial cells revealed expression of genes associated with osteogenic commitment and differentiation [34], which may have an important significance in cardiovascular pathology since adventitial progenitor cells have been shown in the mouse to drive blood vessel calcification, also known as arteriosclerosis [54]. However, even though PSCs are clearly endowed with strong osteogenic potential, a paradoxical yet recurrent observation is that over time little chimerism can be detected in newly developed bone following xenogeneic PSC transplantation, suggesting these perivascular progenitors merely mediate bone formation by recruiting local osteogenic cells and reinforcing the growing belief that MSCs and related tissue regenerative cells function largely via trophic/chemotactic factor secretion [55]. Do pericytes and adventitial cells, which all contribute to MSC cultures and are arranged along blood vessels as a hierarchy of regenerative cells [34], play distinct roles as either osteoblastic progenitors or trophic secretory cells during osteogenesis? This important question is currently under investigation in experiments where either perivascular cell subset or the combination of the two is administered in the same injury setting.

Although recognized in all tissues with canonical markers and characteristic perivascular distribution, pericytes and adventitial cells represent heterogeneous cell populations which also exhibit organ-restricted anatomic, phenotypic, and functional specializations, the complexity of which is being gradually uncovered [11]. Regarding bone formation, we have recently identified novel surface markers which typify PSC subsets endowed with higher osteogenic potential (Ding, Meyers et al., unpublished results), as was already recently done for pro-fibrotic ability [10] and chondrogenic capacity [56]. Ongoing studies will converge to explain the bone healing effect of pericytes and other regenerative perivascular cells, both natively in situ and following purification and transplantation, and contribute to the development of a refined therapeutic product (Table 3.1).

Table 3.1. Summary of in vivo orthopedic models of PSC application.

Model	Species/strain	Human cell used
Intramuscular implant	SCID mouse	AT pericyte, AT-adventitial cell, AT-PSC
Calvarial bone defect	SCID mouse	AT-PSC
Spinal fusion	Athymic rat	AT-PSC
Non-union	SD rat	AT pericyte