|Home | About | Journals | Submit | Contact Us | Français|
Heterotopic ossification (HO) is a frequent complication following combat-related trauma, but the pathogenesis of traumatic HO is poorly understood. Building on our recent identification of mesenchymal progenitor cells (MPCs) in traumatically injured muscle, the goal of this study was to evaluate the osteogenic potential of the MPCs in order to assess the role of these cells in HO formation. Compared to bone marrow-derived mesenchymal stem cells (MSCs), a well-characterized population of osteoprogenitor cells, the MPCs exhibited several significant differences during osteogenic differentiation and in the expression of genes related to osteogenesis. Upon osteogenic induction, MPCs showed increased alkaline phosphatase activity, production of a mineralized matrix, and up-regulated expression of the osteoblast-associated genes CBFA1 and alkaline phosphatase. However, MPCs did not appear to reach terminal differentiation as the expression of osteocalcin was not substantially up-regulated. With the exception of a few genes, the osteogenic gene expression profile of traumatized muscle-derived MPCs was comparable to that of the MSCs after osteogenic induction. These findings indicate that traumatized muscle-derived MPCs have the potential to function as osteoprogenitor cells when exposed to the appropriate biochemical environment and are the putative osteoprogenitor cells that initiate ectopic bone formation in HO.
Heterotopic ossification (HO), characterized by the formation of mature bone in the soft tissues, is a frequent complication that occurs following orthopedic trauma. HO is prevalent in patients with severe time-of-war extremity wounds, and the incidence of HO increases by 57% in patients that sustain a polytraumatic blast injury.1 This condition presents a host of problems, including skin and soft tissue breakdown over the residual limb, suboptimal prosthetic fitting, and impaired limb function. Complicating this situation is the degree of difficulty in removing heterotopic bone after it has formed, which includes the risk of injury to neurovascular structures that are often intimately associated with HO.2 With a better understanding of the cellular and molecular events that lead to HO, it may be possible to design targeted treatment modalities to prevent these complications by preventing the growth of ectopic bone. An important step to elucidate the pathogenesis of HO is to identify and characterize the progenitor cell population responsible for the initiation of ectopic bone lesions that occur as a result of extremity trauma.
There is a widely accepted sequence of events that is assumed to occur during the pathogenesis of HO.2,3 First, progenitor cells within the muscle begin to proliferate and generate a fibroproliferative lesion rich with collagen and other matrix proteins. Some of these proliferating cells have the capability to become osteoprogenitors and can be stimulated to undergo osteogenic differentiation. Muscle tissue is conductive to bone formation, and under the appropriate signaling environment, the progenitor cells undergo endochondral or membranous ossification, which eventually leads to heterotopic bone.3 For example, the genetic disorder fibrodysplasia ossificans progressiva (FOP) is characterized by the formation of HO following minor trauma, and has been traced to a mutation in the BMP receptor ACVR1.4 The mutated receptor leads to an exaggerated response to BMP signaling and is correlated to the up-regulation of BMP-4 and a down-regulation of noggin, a BMP antagonist, in the affected tissue,3 which initiates HO formation. Although HO may occur from different etiologies, an osteoinductive signaling environment is likely a common factor in the development of ectopic bone. The cell population in muscle tissue that is responsive to osteoinductive signaling following traumatic injury has not yet been determined.
Our laboratory has recently identified a population of multilineage mesenchymal progenitor cells (MPCs) in traumatized muscle that is capable of undergoing osteogenic differentiation in vitro.5 The MPCs were isolated and enriched from the digested muscle tissue based on their high adhesion characteristics to tissue culture plastic and have a cell surface epitope profile that is characteristic of bone marrow-derived mesenchymal stem cells (MSCs).6 Their osteogenic potential distinguishes the MPCs from the other plastic-adherent populations of progenitor cells that are commonly associated with human muscle tissue, satellite cells, and myoblasts, which do not readily differentiate into osteoblasts.7 The localization of these potential osteoprogenitors in the traumatized tissue suggests their possible involvement in HO development. On the basis of this finding, the goal of this study is to evaluate the osteogenic potential of traumatized muscle-derived MPCs in order to better assess the role of these cells in HO formation. Our specific aims were: (1) to determine the ability of MPCs to form osteogenic colonies; (2) to characterize the osteogenic differentiation of MPCs; and (3) to compare the osteogenic gene expression profile of MPCs to that of bone marrow-derived MSCs, a well characterized population of osteoprogenitor cells.8–10
Traumatically injured muscle was collected with informed consent and Institutional Review Board (IRB) approval from the Walter Reed Army Medical Center (WRAMC). Patients were considered for inclusion in this study if they had sustained traumatic injury with extensive soft tissue extremity wounds (n = 14; age: 23.6 ± 2.9; sex: 100% male). Patients arrived at WRAMC approximately 96 h after injury and underwent serial debridement and irrigation procedures every 2–3 days until wound closure and definitive orthopedic treatment. Tissue for this study was collected at the second or third surgical debridement. Nine of the patients were eventually presented with symptoms of HO during routine clinical follow-up appointments.
MPCs were harvested from the healthy wound margin of the debrided traumatized muscle tissue using a previously described method.5 Approximately 200 mg of muscle was isolated from the tissue sample, minced extensively, transferred to a 50 mL conical tube containing Digestion Medium (DMEM with 0.5 mg/mL collagenase 2; Worthington Biosciences, Lakewood, NJ), and incubated at 37°C for 2 h with gentle agitation. The solution was vortexed briefly, strained using a 40 µm cell strainer, and centrifuged for 5 min at 200 × g. The cell pellet was resuspended in DMEM supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA) and 5 units/mL of penicillin, streptomycin, and fungizone (PSF; Invitrogen), plated in a T175 tissue culture flask, and incubated at 37°C. After 2 h, the cells were washed extensively with phosphate-buffered saline (PBS; Invitrogen) to remove any cells that did not adhere to the tissue culture plastic. The adherent cells were cultured in DMEM supplemented with 10% FBS and 3 units/mL of PSF and washed daily with PBS. The cells were subcultured into new flasks after tightly packed colony forming units (CFUs) were observed, and the cells were cultured in growth medium (GM: DMEM supplemented with 10% FBS and 1 unit/mL of PSF) from this point forward. Subsequent subcultures were performed when the cells were approximately 85% confluent.
Bone marrow-derived MSCs were harvested from femoral heads obtained from total hip arthroplasties using an established technique9 and with the consent of the patients (n = 8; age: 64.9 ± 6.6; sex: 50% male) following an IRB approved protocol at the University of Washington.
MPCs and MSCs obtained after the second and third passages were plated onto tissue culture plastic dishes at a density of 10,000 cells/cm2 and cultured for up to 21 days in osteogenic induction medium (OM) containing GM supplemented with 10 mM β-glycerolphosphate (Sigma, St. Louis, MO), 50 mg/mL ascorbic acid (Sigma), 10 nM 1,25-dihydroxyvitamin D3 (Biomol International, Plymouth Meeting, PA), and 0.01 mM dexamethasone (dex; Sigma).
Cultures of MPCs and MSCs were fixed with buffered 4% paraformaldehyde (FD Neurotechnologies, Ellicott City, MD) and stained for alkaline phosphatase activity using the fast blue BB kit (Sigma) or with alizarin red (Rowley Biochemical Institute, Danvers, MA) to detect mineralized matrix. To quantify alkaline phosphatase (ALP) activity, the cells were lysed using 0.1% Triton X-100 in Tris-buffered saline (TBS; Bio-Rad, Hercules, CA) followed by flash freezing. p-Nitrophenyl phosphate (pNPP; Sigma) was used as the substrate to measure ALP activity based on p-nitrophenol (pNP) release, which was measured spectrophotometrically as A405. Enzyme specific activity was calculated as a function of DNA content, which was estimated using the PicoGreen Assay kit (Invitrogen).
For gene expression analysis, cells were lysed in TRIzol (Invitrogen), homogenized using QiaShredder columns (Qiagen, Hilden, Germany), and total RNA was extracted according to the manufacturer’s protocol, purified using RNeasy Mini columns (Qiagen), and RNA concentrations were estimated on the basis of A260. The RNA samples were reverse transcribed with the SuperScript III System for quantitative reverse transcription-polymerase chain reaction (RT-PCR; Invitrogen) using oligo-dT and random hexamers. RT-PCR analysis of osteogenic gene transcription was done using Platinum Taq (Invitrogen) and with previously reported primer pairs.6 The housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), was used as a control for RNA loading of samples. PCR products were visualized electrophoretically using the DNA 1000 Labchip Series II on a Bioanalyzer (Agilent, Santa Clara, CA). Real-time RT-PCR analysis was performed using SYBR Green (Bio-Rad) and a BioRad iCycler iQ real-time PCR detection system and gene expression was normalized to GAPDH. Profiling of osteogenic gene expression was performed using RT2 PCR arrays for osteogenesis (SABiosciences, Frederick, MD).
Quantitative ALP measurements were analyzed using one-way ANOVA with multiple comparisons. Gene transcript levels measured by real-time PCR were analyzed using Student’s t-tests. Statistical significance for all tests was assigned to p < 0.05 except for gene expression levels measured using the RT-PCR array, for which statistical significance was assigned to p < 0.018 to limit the false discovery rate.
Approximately 150 million cells were harvested per gram of traumatically injured muscle tissue, and 1% of the cells remained adherent to the tissue culture plastic after the nonadherent cells were washed away. After 24 h, two of the cultures were discarded due to contamination, and the cells in the remaining 14 cultures exhibited a spindle-shaped, fibroblastic morphology, which is characteristic of the MSC phenotype (Fig. 1A,B). After 2 weeks, an estimated 4,300 colonies per gram of tissue stained positively for ALP (Fig. 1C,D).
After 21 days under osteogenic induction conditions, the MPCs demonstrated histological and histochemical evidence of differentiation into osteoblasts. The MPCs cultured in osteogenic induction medium stained positively for ALP activity (Fig. 2A). Quantitative and statistical analysis showed that the ALP activity was increased in the differentiated MPCs, which was a response similar to that of bone marrow-derived MSCs (Fig. 2B; one-way ANOVA with multiple comparisons, n = 3). However, the specific ALP activity of the MPCs decreased from day 7 to day 14, whereas that of bone marrow-derived MSCs increased during this period of differentiation. Based on DNA content, cell number in the MPC cultures increased from day 7 to day 14; a similar increase was not observed in the bone marrow-derived MSCs. By day 21, both cell populations produced a mineralized matrix that stained positively using alizarin red and contained mineralized nodules characteristic of osteoblasts (Fig. 3).
Traumatized muscle-derived MPCs also expressed an osteogenic gene expression profile characteristic of early osteoprogenitor cells. After 21 days in osteogenic induction medium, the MPCs up-regulated CBFA1 (core binding factor-α1), an osteoblast-specific transcription factor, ALP, and osteocalcin (BGLAP), an indicator of terminal differentiation into osteoblasts (Fig. 4A). The expression level of osteocalcin appeared to be lower in the MPCs compared to the bone marrow-derived MSCs, and real-time RT-PCR confirmed that the up-regulation of osteocalcin in the MPCs under osteogenic induction conditions was significantly lower than in bone marrow-derived MSCs (Fig. 4B; p < 0.001, Student’s t-test with n = 3).
The expression of a wide panel of genes associated with osteogenesis was compared between MPCs and bone marrow-derived MSCs during osteogenic differentiation using real-time PCR arrays. Prior to osteogenic induction, there were notable differences in the osteogenic-gene expression profile of the traumatized muscle-derived MPCs compared to that of bone marrow-derived MSCs (Fig. 5A; Student’s t-tests with α = 0.018 and n = 3). MPCs expressed significantly greater levels of COL15A1 (collagen type XV α1) and GDF-10 (growth and differentiation factor-10) prior to induction, although the expression levels of these genes were substantially down-regulated when the cells were induced to differentiate (Fig. 5B). Bone marrow-derived MSCs expressed significantly higher levels of VEGF-A (vascular endothelial growth factor-α), VCAM-1 (vascular cell adhesion molecule-1), and IGF-2 (insulin-like growth factor-2), and the expression of these genes did not change substantially during differentiation (Fig. 5C). The up-regulation of CBFA1, ALP, and osteocalcin in the MPCs and bone marrow-derived MSCs observed using the real-time PCR array corroborated the findings of the real-time RT-PCR assay (Fig. 5B,C).
Although HO is a known sequela of traumatic injury, and occurs frequently in patients following combat-related extremity injuries and traumatic amputation,1 little is known about the etiology of this pathological wound repair process. In this study, we have rigorously characterized the osteogenic potential of a population of MPCs isolated from within the traumatized muscle tissue.5 The morphology and cell surface epitope profile of the MPCs are similar to those of bone marrow-derived MSCs,6 osteoprogenitor cells resident within the bone marrow. We have also demonstrated that the traumatized muscle-derived MPCs are capable of forming ALP-positive colonies, which is consistent with the mesenchymal osteoprogenitor phenotype.11 Upon osteogenic induction, the MPCs increase their ALP activity and begin to generate a mineralized matrix, although unlike the bone marrow-derived MSCs, significant cell proliferation continues in the MPC population while being cultured in the osteoinduction medium. The osteogenic gene expression profile of the MPCs is characteristic of early differentiation into osteoblasts, although some significant differences are apparent between the MPCs and bone marrow-derived MSCs. Taken together, these findings provide insight into the possible involvement of the traumatized muscle-derived MPCs in the formation of HO following orthopedic trauma.
Several aspects of our study lend strength to the relevance of the findings. First, the muscle samples in this study were obtained from a patient population that has a documented predisposition to HO formation. Therefore, the MPC populations were derived from tissues that were likely to undergo HO, and our findings are directly applicable to HO formation in humans following traumatic injury. Second, we have compared the osteogenic differentiation of MPCs to bone marrow-derived MSCs, a well-characterized population of osteoprogenitor cells8–10 that served as a valid positive control. Finally, the differentiation of the MPCs into osteoblasts was evaluated based on a number of criteria, including histology, histochemistry, and gene expression analyses, with each assay performed using a different set of patients. The corroborative evidence is taken together to accurately assess the osteogenic potential of the traumatized muscle-derived MPCs.
However, there are two caveats to this study that should be noted. First, the bone marrow-derived MSCs were obtained from patients undergoing elective total hip arthroplasties, and thus were not age or sex matched to the substantially younger population of soldiers from whom we obtained the traumatized muscle samples. Although the MSCs are an adequate population to use as a positive control, any observed differences between the cell populations must take into account age-related changes12,13 and sexual dimorphism14 in the osteogenic potential of bone marrow-derived MSCs. Second, a modified alpha (α = 0.018) was used to assess statistical significance in the RT-PCR array assay and limit the false discovery rate. While comparisons of the osteogenic gene expression profile as a whole may be made between the two cell populations, the differential expression of any specific genes in the array will need to be independently verified before beginning further investigation to determine their role in osteogenic differentiation of the MPCs or ectopic bone formation in traumatized muscle.
Despite these caveats, significant differences were noted between the traumatized muscle-derived MPCs and bone marrow-derived MSCs. First, the MPCs continued to proliferate while being induced to differentiate into osteoblasts. It is not known whether the entire population is slow to shift from the proliferative state to differentiation, or if a subset of the population continues to proliferate while a second subset differentiates. There is evidence supporting the former, because histological evidence of differentiation appears homogeneous throughout the MPC cultures undergoing osteogenesis. These cells also express lower levels of osteocalcin, an osteoblastic gene that is expressed during later stages of osteogenic differentiation. Second, there are differences in the osteogenic gene expression profile between the MPCs and MSCs, which may reflect the tissue of origin for both cell types. MPCs express higher levels of COL15A1, a gene associated with muscle tissue development,15 and GDF-10, shown to be a negative regulator of osteogenesis,16 whereas the bone marrow-derived MSCs express higher levels of genes associated with bone physiology and maintenance: VEGF-A,17 VCAM-1,18 and IGF-2.19 These differences may also reflect the fact that traumatized muscle-derived MPCs are harvested from an active wound bed, where they likely participate in the process of muscle tissue repair. During osteogenic differentiation, COL15A1 and GDF-10 are substantially, albeit nonsignificantly, down-regulated, while VEGF-A, VCAM-1, and IGF-2 are similarly up-regulated, suggesting that the MPCs can assume the role of osteoprogenitors under the appropriate biological environment, in a manner similar to other populations of MSCs.20
The MPCs used in this study were harvested from traumatized muscle tissue, and when the tissue was obtained, it was not possible to determine whether the patient would eventually develop HO. However, the osteogenic potential of the MPCs does not appear to be patient specific. Using the limited clinical outcomes data, we could not make any correlation between the ability of MPCs to undergo osteogenic differentiation and the incidence of HO in our patient population. Given this apparent homogeneity of the MPCs harvested from different patients, it is reasonable to assume that the potential to differentiate into osteoblasts is an inherent property of all the MPCs in the traumatized tissue. Multipotent stem cells have previously been isolated from untraumatized muscle using immuno-selective techniques,21 and the plastic-adherent MPCs studied here may be the descendants of these stem cells that have been activated within their niche by the injury and have begun to proliferate in the tissue. It has also been hypothesized that pericytes, which occupy a perivascular niche in vivo, are the cells that exhibit an MSC phenotype in vitro22; thus the MPCs might also be activated pericytes that have entered the wound bed. These scenarios imply that migration of MPCs into the traumatized muscle is a part of the normal wound healing response, and HO occurs when pathological factors are present in the environment to encourage the expression of their osteogenic phenotype. The traumatized muscle tissue is an active wound bed, with intense inflammatory and wound healing responses. We are currently studying the cytokine expression profiles of the traumatized tissue to determine which factors are sufficient to trigger osteogenic differentiation of the MPCs and initiate the formation of ectopic bone.
In summary, the findings of this study indicate that traumatized muscle-derived MPCs are the putative osteoprogenitor cells responsible for HO following traumatic injury. Although there are notable differences between the MPCs and MSCs, which likely reflect their tissue of origin and in vivo function, both cell types demonstrate the ability to adopt the osteogenic differentiation pathway under the appropriate induction conditions. It is unclear what pathological signaling events occur in vivo to initiate the differentiation of traumatized muscle-derived MPCs into osteoblasts, and the identification of this triggering mechanism is an active area of investigation in our laboratory. Furthermore, the MPCs appear to differentiate more slowly than the bone marrow-derived MSCs, which suggests that an additional, intermediate step may be involved to encourage their terminal differentiation into osteoblasts, such as the endochondral ossification of the fibroproliferative lesions. This investigation will be followed up using an in vivo model to study the mechanism of osteogenic differentiation of MPCs that leads to ectopic bone formation in a physiological system. Nevertheless, our findings have shed substantial light on the mechanism of HO by identifying the MPCs in traumatized muscle as the osteoprogenitor cells that are likely to initiate bone formation, and may be used in future investigations to further elucidate the signaling mechanism involved in this process.
This study was supported by a grant from the Military Amputee Research Program at WRAMC (PO5-A011) and by the NIH NIAMS Intramural Research Program (Z01 AR41131). Dr. Nesti and Dr. Tuan were co-senior authors of this article.
*This article is a US Government work and, as such, is in the public domain in the United States of America.