Conflict of Interest: Diego Correa has no conflict of interest. Arnold I. Caplan (AIC) has served as a paid technology analyst for the venture capitalists that initiated the formation of BioMimetic Therapeutics, Inc (BMTI) and served as a member of the Board of Directors of BMTI from its inception until shortly after it went public in 2008. BMTI is currently commercializing the use of PDGF-BB in Orthopaedics. AIC has also served as a paid consultant and continues to serve on the BMTI Clinical Advisory Board. It is because of this firsthand experience that AIC came to know the literature involving PDGF-BB, although he has only one published manuscript involving PDGF-BB. It is only recently through research on MSCs, and now pericytes, that we put the thesis together into this manuscript. AIC declares this Conflict of Interest even though he is not an Orthopaedic Surgeon or officer or major stockholder of BMTI.
In the late 1980's, a hypothesis paradigm was generated which asserts that a marrow-derived multi-lineage precursor cell, a mesenchymal stem cell (MSC), is responsible for bone formation among other functions (Fig. 1). As a general rule, bone formation in embryos and adults is driven by the presence of vasculature; therefore, the absence of vasculature precludes bone formation and enhances the formation of cartilage.1–4 Of importance, every blood vessel, small and large, has cells of mesenchymal origin on the abluminal surface; these perivascular (mural) cells summarily are referred to as “pericytes” [reviewed in Ref.5]. It has been proposed on the basis of published studies6–11 that a high proportion of pericytes are MSCs with osteogenic potential and that the previously elusive MSC niche is indeed localized in a perivascular site.12, 13 As part of the mesengenic process (Fig. 1), MSCs are osteochondral progenitors implicated in the bone-forming process via both endochondral and intramembranous ossification.14, 15 Within these programs, MSCs participate in a number of different ways, some of which are not currently appreciated. In addition, it has been previously suggested that the sheets of secretory osteoblasts are oriented by the obligatory blood vessels which are the drivers of bone fabrication.1, 4, 16 The cell-signaling molecules controlling the reciprocal interactions between blood vessels and osteoblasts have not been clearly delineated, although vascular endothelial growth factors (VEGFs) and bone morphogenetic proteins (BMPs) have been implicated.17, 18
With the recent firm identification of the pericyte-MSC phenotypic equivalence (discussed below),6, 9, 11 new insight and new ways to understand the interactive details of bone formation, regeneration, and repair are now possible. Indeed, the action of some bioactive molecules can be better understood. In this process, platelet-derived growth factor (PDGF) is one of these important molecules that take on new significance in the vasculature-pericyte-MSC-osteoblast bidirectional interactions.19–22
PDGF has various isoforms (AA, AB, BB, CC, and DD) which signal through two distinct dimerized receptors (α and β) with different affinities. Despite these various isoforms, PDGF-BB is recognized as the universal PDGF because of its ability to bind to all known receptor isotypes23 and due to its physiological functions.24, 25 PDGF-BB/PDGFR-β signaling constitutes the principal pathway responsible for pericyte recruitment and attachment to the vasculature, their subsequent maturation and potential destabilization and detachment.26–29 Other factors have also been implicated in pericyte biology at specific steps. For instance, TGF-β is required for the differentiation of mesenchymal progenitors (MSCs) into pericytes. Angiopoietin-1 (Ang1) signaling through Tie2 receptor induces pericyte recruitment, while Ang2/Tie2 antagonizes Ang1 signaling promoting detachment and destabilization of the endothelial cell/pericyte interactions. Extracellular matrix (ECM) remodeling, matrix metalloproteinases (MMPs) and hypoxia, all conditions that promote angiogenesis, have also been associated with pericyte detachment from the vasculature. This destabilization of the vessels can be followed by either “new” angiogenesis or vascular regression depending on the presence of angiogenic factors such as VEGF. Subsequently PDGF-BB can reassemble new associations between endothelial cells and pericytes.
Because angiogenesis, osteogenesis, and mesengenesis are often studied as separate processes and as “independent” pathways, the central integrating role of molecules like PDGF has been largely underappreciated. This perspective combines these pathways to suggest that PDGF-BB, which actively participates in all three, could function as a central connector between the pathways and, thus, could be used as a powerful therapeutic agent for bone regeneration and repair.
A lineage relationship between MSCs and pericytes can be built based on the documentation recently presented by us and others, in which: MSCs can arise from pericytes isolated from various tissues; they exhibit similar markers in situ which are retained in vitro; they have similar multipotent differentiation ability and have similar self-renewal capacity.5–9, 11–13 These similarities now clearly explain the intimate connection between MSCs and vascular driven bone formation and, indeed, was the stimulus for the thesis described herein. However, the highly flexible plasticity typical of MSCs and their differentiated descendants make the dogmatic statement12 that “all” MSCs are pericytes difficult to prove, particularly in vivo. For example, it can be shown that cultured MSC-derived adipocytes filled with fat vacuoles can be induced into the osteogenic lineage.30 These adipocytes, however, do not dedifferentiate and become MSCs/pericytes and then enter the osteogenic lineage, but rather they shut down active adipogenesis and upregulate osteogenesis (i.e., transdifferentiation or plasticity).
At the tissue level, MSCs reside in perivascular locations close to sheets of osteoblasts, as a cellular component of the hematopoietic niche or as an inactive marrow stromal cell. The important point is that, in active angiogenic situations, MSCs from human marrow or from fat have the capacity to become pericytes on newly forming vessels. This perivascular positioning serves to stabilize these new vessels.31, 32 The proposal can be put forth that the ability of an MSC to become a pericyte might serve as a new functional definition of an MSC. More importantly, the fact that MSCs can be isolated from a diverse array of human tissues can now easily be explained by their identity with pericytes.6–9 Likewise, that these MSCs from different tissues (i.e., marrow, muscle, fat, skin, brain, etc.) are all slightly different emphasizes that the abluminal microenvironments from which they are extracted are themselves different.33
Alternative to the potential bidirectional conversion of MSCs into pericytes, they can follow at least two other functional routes. First, they can enter the osteoblastic differentiation program and thus generate bone matrix-secretory osteoblasts and then matrix-embedded osteocytes, representing the “preferable” mass action route during bone formation. Second, MSCs can be activated to exert trophic and immunomodulatory activities that enhance the local regenerative microenvironment.34–36 Both of these potential transitions (Fig. 2) make MSCs highly versatile cells especially during bone reparative circumstances.
The fact that PDGF-BB is a powerful chemoattractive, mitogenic and vascular docking agent implies that the injury or inflammation of blood vessels is a mechanism by which these pericytes are released as MSCs independent of the tissue in question. The management of these released MSCs is obviously a complicated, multifactorial process further complicated by disease and the decline of vascular density especially during the process of aging.
BONE FORMATION MODEL
When bone tissue breaks or is surgically injured, a rapid and active inflammatory response floods the injury zone with blood cells, platelets, monocytes, macrophages, and other cells of the inflammatory cascade.37–39 The result is that the injury site is isolated from the rest of the body, becoming avascular to insure that the local injury environment does not “disrupt” or “infect” the rest of the body. These segregation processes that occur at the bone break or injury sites eventually result in the repair blastema and outer surrounding reparative callus. These events are recapitulated when demineralized bone matrix (DBM) chips are placed in subcutaneous or intramuscular locations, where an endochondral bone development program ensues.40–43 Into this “isolated” environment both platelets and macrophages release huge quantities of bioactive molecules including PDGF, which prepare the conditions for a repair response.23 Subsequently, in mechanically stable circumstances, newly formed blood vessels invade the repair tissue, and the vessel-associated MSCs enter to form sheets of osteoblasts that fabricate bone as oriented by the invading blood vessels. A new view of this multicomponent process relies on the fact that this sequence of events involves PDGF in several steps, not only serving to stimulate local angiogenesis, but also to upregulate and position the stimulation of osteogenic events resulting in rapid bone formation (Fig. 3 and next section). In the context of bone repair and regeneration, the model put forth below suggests that the site management of PDGF, especially PDFG-BB, could provide a powerful, clinically relevant therapeutic.
This thesis is predicated upon the following facts and implications:
(A)Bone formation absolutely requires the presence of microvasculature.1–4
(B)Bone is formed by monolayer sheets of osteoblasts that secrete osteoid in coordinately oriented layers.1–4
(C)Blood vessels orient the sheets of osteoblasts such that the backs of the osteoblasts face the vessels and the secretion of osteoid is from the opposite side or front of the sheets.1–4
(D)Newly forming blood vessels form from rapidly dividing and migrating endothelial cells.26
(E)Endothelial cells secrete PDGF-BB, recruiting PDGFR-β-expressing pericyte progenitors into the area of new vessel formation.20, 26, 28, 44
(F)Newly formed blood vessels are stabilized by abluminally attached pericytes on the vascular structures.31, 32
(G)PDGF-BB binds to the pericellular matrix (PCM) or glycocalyx surrounding endothelial cells, by the unique C-terminus binding motif that interacts with heparan sulfate proteoglycans (HSPG) and other molecules. This binding motif retains PDGF-BB close to endothelial cells, creating a chemoattracting gradient that brings pericytes into contact with vascular endothelial cells and the newly forming basement membrane, which contains both HSPGs and PDGF-BB.26, 28, 44–46
(H)MSCs and pericytes share phenotypic characteristics.6–9, 12
(I)MSCs are the progenitors of bone-forming osteoblasts.15, 47
(J)During embryonic bone formation, fracture repair, or bone fusion sites, the same basic sequence of cellular events controls bone formation, regeneration, or repair.39–41
(K)The rate of bone or fracture repair is a function of both vascular density and MSC availability at both periosteal and marrow cavity surfaces22, 48, 49 and can now be best understood because of pericyte availability in the context of the vasculature.
PDGF-BB AND MSCs: THE PERSPECTIVE
In the previous section, the requirements for bone formation were described, with special emphasis on the independent roles of both PDGF and MSCs as a key signaling molecule and the building blocks of the process, respectively. The combination of these molecular and cellular elements is the basis for this perspective of the sequence of events leading to bone formation. To discuss the details of this “new” model, the critical question can be asked: Why is exogenous PDGF not needed in very young patients, both in the oral cavity and in distal bone fractures/defects? Answer: a high vascular density in young individuals brings a relatively high number of MSCs/pericytes into the injury site. In addition, platelets release PDGF-BB into this site, which is a fraction of the scale/size of those sites in the adult. These speculated differences are conceptually based on the following established phenomena: PDGF-BB is able to stimulate the expression of VEGF by pericytes (step 1 in Fig. 3)23, 50; PDGF-BB releases MSCs/pericytes from their abluminal location in active angiogenic sites23, 28, 51 (step 2 in Fig. 3), thus providing available local MSCs; these “free” MSCs rapidly divide in the presence of PDGF-BB, increasing the pool of osteochondral progenitors (step 3 in Fig. 3), which ultimately differentiate into bone-forming secretory osteoblasts when stimulated by osteogenic factors such as Wnt signaling family members and BMPs, with PDGF-BB involved in modulating the responsiveness to the latter (step 4 in Fig. 3)23, 52, 53; the non-differentiated MSCs can reposition themselves in a perivascular location by means of PDGF-BB26, 46, rapidly stabilizing the newly forming blood vessels which orient and drive efficient bone formation (step 5 in Fig. 3).31, 32, 43, 45, 54, 55 This sequence of known events can now be linked to allow the suggestion that the local presence of exogenously applied PDGF-BB could function to stimulate both angiogenesis and MSC/pericyte availability, thus promoting osteogenesis.
Consequently, the release of PDGF-BB at a fracture or bone injury site could (Fig. 3):
(I)Serve to stimulate secretion of VEGF by pericytes, bringing new endothelial cells into the angiogenic injury site.23, 50
(II)Serve to unhook the pericytes from the existing vessels and facilitate their attraction to the injury site, thus allowing some of these released “free” pericytes to become activated and functional MSCs, some of which function to structure a regenerative microenvironment, while others become lineage-progressing osteoprogenitors cells.34–36
(III)Serve as a powerful mitogen for both the pericytes and “free” activated MSCs.23, 45
(IV)Modulate the responsiveness to key osteogenic factors such as BMPs, which are responsible for the further osteoblastic differentiation of the “free” activated MSCs.23, 52, 53
(V)Serve to bring PDGFR-β-expressing MSCs/pericytes back into contact with the rapidly expanding and infiltrating microvessels and, thus stabilize vasculature positioning and formation.26, 31, 32, 43, 45, 54, 55
(VI)Serve an integral role to coordinate and link endothelial cells, pericytes, MSCs, ECM, PDGF-receptors and signaling pathways.
In sum, the above serves as a new mechanistic model to explain the role of PDGF-BB in the formation, repair, and regeneration of bone. Specifically, it proposes how PDGF-BB can organize the presence of osteoprogenitor cells at a specific required site, induce their efficient multiplication, modulate their responsiveness to osteoblastic differentiation factors and assemble/stabilize newly formed blood vessels. Importantly, these local MSCs can serve to protect this rapidly changing tissue from over-aggressive immuno-surveillance, enhance the regenerative microenvironment and inhibit scarring.
IMPLICATIONS OF THE PROPOSED MODEL
It is well established that PDGF-BB has critical effects on a variety of cells, especially those of mesodermal origin (Fig. 4). Certainly PDGF-BB is a powerful mitogen and chemoattractant for MSCs. The in vitro proliferative effect is dose-dependent, being additive to the response to the serum that is routinely added to culture medium. This serum contains PDGF, which suggests that the “basal” PDGF concentration in the serum is not saturating [55, 56 and unpublished observations]. In this case, the PDGF receptors transmit signals to the cell causing the cell to divide and affecting its directional mobility.57 In the context of this thesis, PDGF-BB also has a profound effect on the positioning and stability of pericytes as vascular abluminal cells.44, 54 Endothelial cells secrete PDGF-BB, and the pericytes' PDGF receptors (especially PDGFR-β) not only cause the movement of pericytes to the abluminal surface of blood vessels, but also foster the stabilization of the interaction between endothelial cells, the newly forming basement membrane and pericytes.
This proposed mechanism has, therefore, profound implications on bone formation, both at early embryonic stages and later in postnatal fracture repair. As described below, both processes depend on vascular development, underscoring the endothelial cell-pericyte-MSC-osteoblast relationship.
EMBRYONIC BONE FORMATION
The detailed cellular events involved in embryonic long bone formation are an important guide to understanding the details and governing principles of bone repair. These embryonic events have been best studied in chick embryos, but also in mice and humans.1–4 In the limbs, the long bones and digits form from a uniform mesoderm in the limb bud by first forming a rod of cartilage surrounded by an avascular “stacked-cell” layer of 3–4 cells. From the bottom of this stacked-cell layer next to the cartilage core, a monolayer of osteoblasts forms; these cells secrete a sheet of osteoid onto the cartilage model (Fig. 5). Capillaries outside the stacked-cell layer dive into the stacked cells to position themselves above the newly formed sheet of osteoblasts. This original invasive capillary network then organizes the differentiation of the cells at the bottom of the stacked-cell layer (now called a periosteum) to form a second sheet of secretory osteoblasts that form another layer of bone. In the embryonic chick tibia, this goes on for an additional 16 or more layers, each organized by its own invasive capillary network.
At the mid-diaphysis of the tibia (for example), the cells in the cartilage core are completely surrounded by bone.1, 2 These core chondrocytes become hypertrophic and eventually expire, presumably because of lack of nutrients and oxygen since they are totally avascular and encased by nutrient-impenetrable mineralized osteoid. As these hypertrophic chondrocytes expire, they secrete VEGF,58 which stimulates the invasion of the first bone capillaries that bring in resorptive cells, hematopoietic progenitors and MSCs. These cells eventually replace the cartilage model with the very complex, multi-component marrow tissue.1–4 Thus, the misconception that the cartilage model is replaced by bone can be restated: the cartilage model is replaced by vasculature and marrow.59
Importantly, this sequence of replacement of hypertrophic cartilage by marrow and vasculature also occurs at the growth plate, during fracture repair and during host-mediated responses to demineralized bone.40–43 In these last three cases, the invading vasculature and its complement of MSCs/pericytes respond to local cues to form sheets of osteoblasts that generate bone. In fact, as recently documented, the osteoblasts that generate new bone formation in the growth plate of mice originate from MSCs/pericytes of the invading vasculature.16 Again, the cuing comes from both the resident ECM and the resorptive cells. In this regard, the implication is that the local presence of PDGF-BB from these invading cells and the ECM play an osteogenic role.
For orthopaedic indications involving fracture repair and surgical procedures, the effects of PDGF-BB on destabilizing pericytes from their in situ abluminal location to generate local, activated, “free” MSCs is of great importance. This unloading effect of PDGF-BB has been documented in active angiogenic vascular beds (like the one in a fracture site) and not in quiescent vasculature.51 Therefore, in order to have such a cell detachment effect, the vasculature needs to be under active remodeling and facing high local concentrations of PDGF-BB, as is the situation present at the fracture site. Following injury or surgical intervention, acute inflammatory events flood the site with cells and bioactive factors, including platelets which release large quantities of PDGF-BB at the injury site.25 The suggestion is that these events alone can contribute to the release of pericytes and their subsequent transformation into activated MSCs that divide to form the cellular component of the repair blastema.16 These activated MSCs not only provide anti-inflammatory and trophic factors that enhance the regenerative microenvironment,34–36 but also serve as a site-specific progenitor reservoir for the formation of osteoblasts that will form new or bridging bone (Figs. 2 and 3).
Exogenously supplied PDGF-BB remains at the fracture or repair site for a relatively short time, since its primary function is to assist in the release of the pericytes to provide activated MSCs. Indeed, in labeling studies,60 96% of the initial PDGF-BB is gone within 96 h, yet its effects can only be seen several weeks later as an increase in the rate and volume of formed bone. Importantly, because of this initiating release and mitotic expansion of MSCs from their abluminal location, craniofacial and tendon soft tissues are also regenerated by the initiating effects of exogenously provided PDGF-BB.19, 61
Since this thesis also holds that bone formation is a vascular driven process, these MSC/pericytes and the resulting osteoblasts would also secrete VEGF and cause vascular endothelial cells to enter the reparative field.23, 50, 62, 63 These resultant newly formed capillaries would be stabilized by a portion of the MSCs as they become pericytes again.31, 32
If the bone fracture is mechanically stable, fragile blood vessels are able to span the break, and the MSCs both stabilize the blood vessels and form sheets of osteoblasts that generate osteoid which becomes calcified into trabecular bone. This repair bone eventually is reshaped as controlled by its loading dynamics. If the break is mechanically unstable, the fragile blood vessels cannot span the break and the blastema forms a core of cartilage. This cartilage core becomes surrounded by newly synthesized bone as bridged by the vascular driven callus (stacked-cells) as discussed in the Embryonic Bone Formation section. The internal chondrocytes become hypertrophic and expire, secreting VEGF.58 The outer bony callus mechanically stabilizes the break and responds to the expiring hypertrophic chondrocytes by stimulating the invasion by vasculature which brings its fresh supply of MSCs to form new trabecular bone to span the break site.
YOUNG VERSUS OLD BONE
Young healthy individuals repair bone fractures more rapidly than older individuals. This is likely due to the vascular density and scale of the bone at the fracture.21, 22 Given the proposal that PDGF-BB can affect the rate and extent of bone formation, it follows that osteoporosis and diabetes would be conditions that create challenging environments for fracture repair and presumably exogenously delivered PDGF-BB could improve the repair environment in those cases. Published studies clearly document that in rodent models of both osteoporosis and diabetes, exogenously delivered PDGF-BB improved both the rate of bone formation and the bone volume spanning the fracture.21, 22, 37 It would be useful to re-examine these rodent models to assess the accuracy of the early events predicted here. In the cases of elderly patients or distal bony sites, the vascular density is compromised and, thus, access to osteo- and vascular progenitors is greatly reduced. The local progenitors must be mitotically expanded and others attracted from other sites. In this context, PDGF-BB functions in several modes: as a mediator of pericyte-endothelial cell interaction, as a chemoattractant and as a mitogen. As the reparative field builds an extracellular matrix (ECM), the PDGF-BB binds to ECM component motifs that store the PDGF and act as an intrinsic slow-release modality that allows the different phases of the repair-regeneration process to proceed in a PDGF-mediated sequence. Thus, the influence of PDGF-BB is manifesting itself as a function of its concentration in both the soluble and bound ECM (Fig. 3).
From these considerations, it might summarily be said that exogenously supplied PDGF-BB is critical for osteogenesis; it frees the pericyte from its abluminal location bringing these MSCs into the bone repair field, causing them to rapidly divide and facilitating the orientation of stable blood vessels by the reattachment of the pericytes to provide stability to these fragile newly forming vascular structures. Additionally, PDGF-BB serves to enlarge the osteoprogenitor cell pool (through its mitogenic capacity) providing new osteoprogenitor cells to take the place of osteoblasts that naturally expire during the bone fabrication process (Osteoblasts have a half-life of 6–10 days). Certainly, the replacement of expiring osteoblasts implies a lineage and mobility in the space between the driver blood vessels and the sheets of secretory osteoblasts. One could envision the pericyte detaching from the vessel to become a “free” MSC, the MSC entering the osteogenic lineage as it progresses toward the sheet and becomes exposed to various osteogenic factors such as BMPs and Wnts23 and then the commitment to secrete osteoid as it is situated into the sheet (Fig. 2).
MSCs exhibit very complex synthesis and secretion characteristics that can be immuno-modulatory and trophic. It has been suggested36 that their name be changed to “Medicinal Signaling Cell” (MSC) since their potential therapeutic use could be as a medicinal cellular delivery vehicle for asthma, stroke, acute myocardial infarct, inflammatory bowel disease, graft versus host disease, Crohn's disease, multiple sclerosis and others. The term mesenchymal stem cell seems appropriate for the capacity of these cells to differentiate into cartilage or bone and for tissue engineering applications.15 These considerations beg the question as to whether all MSCs are equal and how multipotency, immuno-modulation and trophic capacity are related. Now that it has been shown that at least a proportion of MSCs reside in vivo as pericytes, we can start to address these issues. Moreover, the effects of certain molecules like PDGF-BB that are widely distributed throughout the body, present at sites of injury and that have several functions and receptors can now be more functionally related to the distribution and functional flexibility of MSCs. The cooperative management of PDGF-BB and MSCs would, thus, provide considerable clinical benefit if these newer considerations are correct. Clearly, the cross-talk between MSCs and endothelial cells is a determinative factor in the rate and extent of osteogenesis.64
The evolving database on the cytology and chemistry of MSCs brings with it new insight into the biological control processes governing tissue formation, repair, and regeneration. Seemingly ubiquitous and well-known molecules like PDGF-BB take on new meaning and functions given the newer MSC information. This new information also identifies PDGF-BB as an important controller of osteogenesis in repair/regeneration circumstances.
The authors thank their colleagues both in and outside of the Skeletal Research Center for their careful and thoughtful assistance in preparing this manuscript. Studies reported here were partially supported by the Virginia and David Baldwin Fund, NIH (Program Project Grant #AR053622), the State of Ohio and the US Army (AFIRM).