Introduction
Heterotopic ossification (HO) is the formation of bone outside of the normal skeleton. This process may occur in the skin, muscles, and soft tissues and can cause significant morbidity. Nonhereditary heterotopic ossification (NHO) is a common consequence of musculoskeletal trauma and surgery. NHO has also been reported following spinal cord injury, traumatic brain injury, and severe burns.
1-5 NHO has been reported to occur in up to 90% of patients following hip arthroplasty and acetabular fractures, and in up to 53% of patients with neurologic injury.
3,6 This condition causes significant morbidity as a result of joint contractures, loss of soft tissue, and chronic pain.
2-4 Histologically, NHO is a process that demonstrates temporospatial evolution. Early lesions demonstrate a nodular fasciitis (NF)-like fibroblastic proliferation followed by gradual bone production and osseous maturation over time.
1,2,4 Similar to bone morphogenesis
7 or fracture repair,
8 NHO may form through endochondral or intramembranous mechanisms.
9 Reported frequencies of endochondral versus intramembranous ossification in NHO are highly variable.
1,2,4,5,10Progressive osseous heteroplasia (POH) is a rare genetic form of HO characterized by predominantly intramembranous ossification (OMIM: 166350).
11-17 POH typically presents in infancy with dermal ossification that progresses to involve the subcutaneous and deep skeletal muscle and connective tissues.
11-22 Ossification may progress to ankylosis of the joints causing focal growth retardation of involved limbs.
17,21 POH is caused by inactivating mutations in the
GNAS gene, which encodes the alpha subunit of the stimulatory G protein (G
sα) of adenylyl cyclase,
11,13,14,18,23-32 and has been associated with autosomal dominant
17,18,21,32 inheritance with the mutation most frequently occurring on the paternally inherited allele,
16,18,27,28,30,31 although many cases are sporadic.
11,12,17,18 The disorder exists on a spectrum of
GNAS mutation-associated hereditary HO disorders, including Albright hereditary osteodystrophy.
11,13,17,18,23,25,27 Disease progression typically slows during adulthood.
11,12 Patients have a normal level of intelligence and do not have significant congenital skeletal abnormalities or endocrine abnormalities.
12,17Fibrodysplasia ossificans progressiva (FOP) is another rare genetic form of HO that presents in childhood caused by activating mutations in the
ACVR1 gene (OMIM: 135100). In contrast with POH, FOP is characterized by predominantly endochondral ossification within muscle and deep connective tissues and spares the skin.
11,16,17,33 Patients with FOP usually have congenital malformations of the great toes and may have other congenital skeletal anomalies. In the absence of injury, HO generally arises bilaterally in a cranial to caudal, proximal to distal progression, and lesions are preceded by inflammatory tumor-like swellings.
11-13,17,19,33 The disease begins in early childhood and continues to progress through adulthood, with most patients wheelchair-bound by the second decade of life. The median lifespan for FOP patients is approximately 40 years, and patients typically succumb to thoracic insufficiency syndromes.
33-37 Most cases of FOP are sporadic; however, some families show autosomal dominant inheritance. Both de novo and inherited cases are caused by activating mutations in the
ACVR1 gene, which encodes activin A receptor type 1/activin-like kinase 2, a bone morphogenic protein type 1 receptor.
1-3,33In a prior study, we characterized the vascular patterning of NHO.
1 We found that heterotopic bone formation and maturation is coupled with dynamic temporal vascular changes.
1 In this article, we report on the vascular patterning of NHO in comparison with inherited forms of HO (including both POH and FOP). We especially sought to characterize the differences in vascular modeling across the various disease processes and across the histologic spectrum of HO.
Materials and Methods
Identification of HO Samples
Six cases of POH, 3 cases of FOP, and 5 cases of NHO were identified in our surgical pathology archives (Johns Hopkins University and University of Pennsylvania). Cases were obtained under institutional review board approval with a waiver of informed consent. All material was coded so as to protect the confidentiality of personal health information. Cases of POH and FOP were diagnosed by clinical features, followed by detection of GNAS or ACVR1 mutational analysis if available. Two skeletal pathologists independently performed a review to verify the diagnostic accuracy of all cases of NHO (AWJ, EM).
Image Acquisition and Categorization of HO Samples
Histologic images of hematoxylin and eosin–stained sections were taken of each sample, with the goal of encompassing distinct areas of each lesion. Depending on the size of the specimen, between 2 and 19 images of each case were taken at 10× magnification. A total of 131 images were analyzed (42 images of POH samples, 43 images of FOP samples, and 46 images of NHO). Next, the images were examined (by ADW and AWJ) and categorized based on their predominant histologic appearance. The following 4 phasic subcategories were employed as per our prior study
1:
1.
NF-like areas: sheets of spindled fibroblastic to myofibroblastic cells with minimal to no bone formation
2.
Cartilaginous areas: areas of prominent cartilaginous differentiation with minimal to no bone formation
3.
HO with woven bone: comprised predominantly of trabeculae of woven bone often with prominent osteoblastic rimming
4.
HO with lamellar bone: comprised predominantly of trabeculae of lamellar bone
Vascular Histomorphometric Analysis of HO Samples
Each image was analyzed to determine the following factors: blood vessel number (total blood vessel number per 10× field), blood vessel area (total vascular area per 10× field), blood vessel density (blood vessel area/tissue area per 10× field), blood vessel size (mean number of pixel units per blood vessel in each 10× field), and blood vessel wall thickness (mean blood vessel wall thickness per 10× field). Blood vessel number was counted manually. All other calculations were performed in ImageJ 1.52a (National Institutes of Health, Bethesda, MD). Vessel area was determined using the wand (tracing) tool in ImageJ in conjunction with the measurement log function to measure the area in pixel units. Density was calculated by dividing vessel area by total photograph tissue area, as determined using ImageJ. Vessel thickness was measured in pixel units using the “straight” tool in ImageJ. Vascular indices were analyzed by diagnostic category (POH, FOP, or NHO) as well as by phasic subcategory of HO (NF-like, HO with woven bone, HO with lamellar bone).
Statistical Analyses
Statistical analyses were performed using Microsoft Excel Version 16.16.3 (181015). Analysis of variance single factor and 2-tailed t test analyses were performed for each histomorphometric parameter (blood vessel number, blood vessel area, blood vessel density, blood vessel size, and blood vessel wall thickness) for each histologic phase in all samples of POH, FOP, and NHO. Analysis of variance single factor and 2-tailed t test analyses were also performed for each disease subtype (POH, FOP, and NHO) to compare each histomorphometric parameter across disease types. Results were charted on box and whisker plots in Microsoft Excel. *P < .05 and **P < .01 were considered statistically significant.
Results
Patient Demographics of POH, FOP, and NHO
A total of 6 cases of POH, 3 cases of FOP, and 5 cases of NHO were examined in this study. Demographics are summarized below and in
Table 1. The size of specimens ranged from 1.5 to 5 cm in maximum dimension.
Among cases of POH, 4 patients were female and 2 were male, with an average age of 16.2 years (range = 2-41 years). The majority of cases of POH involved the trunk and upper extremity (n = 5), and sites included the hand (n = 2), triceps (n = 1), shoulder (n = 1), neck (n = 1), and lower back (n = 1). In 1 case of POH, specimens from 2 different sites were received (shoulder and hand). Anatomic site of origin was not recorded in 1 case of POH.
In the 3 cases of FOP, 2 patients were female and 1 was male. The average age at time of FOP biopsy was 6 years (range = 2-9 years). Patient demographics and the site of each lesion are further delineated in
Table 1.
Among cases of NHO, patient samples were evenly distributed by gender (2 females, 2 males, 1 not recorded), and the average patient age was 32.0 years (range = 29-38 years). NHO samples were most commonly from the lower extremity (n = 3 samples including 1 sample from distal thigh, periarticular soft tissue of knee, and soft tissue of foot), and 1 case involved the upper extremity (periarticular soft tissue of shoulder). Anatomic site of origin was not recorded in 1 case of NHO. For cases of NHO, limited additional clinical information was available. Two cases had no known history of local trauma or other inciting event. One case had a deep venous thrombosis in the affected extremity. Two patients had no available clinical data.
Histologic Features of POH, FOP, and NHO
Typical histologic features of HO were examined in cases of NHO, POH, and FOP, using our previously defined HO histologic subcategories,
1 which included the following:
1.
NF-like areas: These areas showed sheets and fascicles of spindled fibroblastic/myofibroblastic cells, similar to the pattern observed in NF or granulation tissue (
Figure 1A-C). Minimal to no bone formation was observed in these areas. Blood vessels in these areas were numerous, slender, elongated, and thin-walled, consistent with capillaries. NF-like areas were observed in 1/6 (16.7%) cases of POH, 3/3 (100%) cases of FOP, and 3/5 (60%) cases of NHO.
2.
Cartilaginous areas: These areas showed prominent cartilaginous differentiation or endochondral ossification (
Figure 1D-F). Very few blood vessels were observed within purely cartilaginous areas. Areas of frank cartilage were observed in 1/6 (16.7%) of cases of POH and 3/3 (100%) cases of FOP. No cases of NHO showed areas that met these criteria. One case of NHO showed foci of late endochondral ossification.
3.
HO with woven bone: In these areas, lesional tissue is comprised predominantly of woven bone with prominent osteoblastic rimming (
Figure 2A-C). Blood vessels were predominantly thin-walled, with morphologic similarity to venules or bone marrow sinusoids. Areas of woven bone were observed in 5/6 (83.3%) of cases of POH, 1 case (33.3%) of FOP, and 3/5 (60%) of cases of NHO.
4.
HO with lamellar bone: These areas predominantly contained thickened trabeculae of mature lamellar bone (
Figure 2D-F). A range of blood vessel appearance was seen in these areas, from dilated and ectatic thin-walled vessels again resembling bone marrow sinusoids, to smaller, thick-walled vessels resembling arterioles. Lamellar bone was seen in 4/6 (66.7%) of POH cases, 1 case (33%) of FOP, and 3/5 (60%) of NHO cases.
Each case had at least some bone matrix. No cases included features of malignancy, such as nuclear hyperchromasia, pleomorphism or significant atypia, or atypical mitotic figures.
Vascular Histomorphometry Across the Histologic Stages of HO
As described above, similar phasic changes in vasculature seemed to be present in both genetic and NHO. For this reason, vascular indices were first examined combining all samples independent of genetic versus nongenetic etiologies (
Table 2). Here, 3 subcategories as previously described were included: NF-like areas or “NF-like,” HO with predominantly woven bone or “Woven,” and HO with predominantly lamellar bone or “Lamellar.” Too few cartilaginous areas were obtained for analysis.
Results recapitulated the temporal dynamic changes that occur in the vasculature of HO obtained in prior studies of only NHO.
1 Blood vessel number was significantly highest in NF-like areas and reduced in areas of mature lamellar bone (
Table 2). Conversely, blood vessel size and area were lowest in NF-like areas and were highest in areas of mature lamellar bone. In the case of each vascular metric, areas of woven bone showed mean values that were between that of NF-like areas and lamellar bone. Blood vessel thickness also showed dynamic changes over different histologic appearances of HO, with increased vessel wall thickness in areas of mature lamellar bone. Thus, when all samples of HO were examined irrespective of etiology, quantitative phasic changes in vasculature accompany changes in bony “maturation.” These quantitative metrics represent a shift from the slender and numerous capillary-type vessels in most immature HO, to a combination of more developed vessels in more mature bony lesions, including larger bone marrow sinusoid-type vessels along with thick-walled arterioles.
Vascular Histomorphometry in POH, FOP, and NHO
Vascular indices were next examined between genetic and NHO types across each histologic subcategory (
Table 3). NF-like areas were first examined in POH, FOP, and NHO. Despite variability, a nonsignificant trend toward increased blood vessel numbers was observed among POH and FOP in comparison to NHO. Blood vessel size among NF-like areas was similar across genetic and NHO. A similar nonsignificant trend was also observed for blood vessel area. Blood vessel wall thickness showed an increase among genetic versus NHO, which achieved statistical significance. Two-way statistical comparisons found a significant difference between blood vessel wall thickness among POH as compared with NHO (*
P = .0125). Thus, when examining the most immature areas of HO, a trend toward increased vessel numbers, increased total vascular area, and increased vessel wall thickness were all observed among cases of genetic HO.
Vascular indices were next compared for genetic and NHO among areas with predominantly woven bone. Significant differences were observed between genetic and NHO across all metrics. Vessel numbers per high-power field were significantly higher among FOP and POH in comparison to NHO. Vessel size and total vessel area were likewise significantly increased in both POH and FOP as compared with NHO samples. Vascular wall thickness was similarly highest in POH and FOP as compared with NHO samples. When comparing vasculature between genetic etiologies, slight differences were observed between POH and FOP. However, no vascular histomorphometric measurements achieved statistical significance when directly comparing POH and FOP. Images obtained from FOP samples trended toward increased number of vessels, while POH samples trended toward increased vessel size, area, and thickness.
Vascular indices were next compared for genetic and NHO among areas with predominantly lamellar bone. FOP samples had only rare examples of lamellar bone, and so statistical analysis between FOP and other groups was not performed. Overall the vascular differences between POH and NHO were no longer observed in areas of mature lamellar bone. Blood vessel numbers, size, total area, and wall thickness showed no significant difference between POH and NHO. Although comparisons are limited by small sample numbers, areas of FOP with lamellar bone showed higher vessel numbers, size, and total area.
Discussion
Heterotopic bone has diverse etiologies, and in this study, we sought to directly compare vascular patterns among HO with and without an underlying genetic driver. POH and FOP are rare genetic diseases that share a commonality, but have important differences in their site of bone formation, frequency of endochondral bone, studied signaling cascades, and even prognosis. Our qualitative and quantitative data in vascular patterning suggest that despite these differences, both genetic forms of HO demonstrate overall increased vascularity in comparison to NHO. Both genetic and NHO disorders show temporospatial variation in vascularity. To our knowledge, this is the first study comparing the vascular histomorphometric features of genetic and NHO.
The histology of heterotopic bone has been well described elsewhere, progressing from an early inflammatory phases through mature bone, which resembles normal skeletal elements. Overall and in this study, HO from all etiologies showed some a similar phasic disease process. NF-like areas were similar in appearance in both genetic forms (POH and FOP) and NHO. In contrast, the frequency and overall amount of cartilage varied greatly between genetic forms of HO as well as compared with NHO. POH has been characteristically described as having predominantly intramembranous ossification,
2,11-14,17,23,26 and in agreement with this, we observed frank cartilage in only 1 of 6 cases of POH. In contrast, FOP is routinely described as predominantly an endochondral process,
38 and all 3 cases in our files showed prominent cartilage. Interestingly, in the present study and in agreement with our clinical experience, the frequency and amount of cartilage in cases of FOP greatly exceeded that of NHO.
Despite these differences, conserved features in vascularity were seen across HO types. As we observed in prior studies,
1 early NF-like lesions showed the highest number of small caliber blood vessels. As HO bone maturation occurs, vessels are less frequent but dilated in size. While the overall patterning of blood vessels was similar in POH, FOP, and NHO, notable differences in the degree of the vascularity were present. Among NF-like areas, vascular quantification showed a slight increase in vascularity within both forms of genetic HO as compared with NHO. These differences were most obvious in areas of woven bone, where genetic forms of HO both evidenced significantly increased vascularity in comparison to NHO. Most striking for the authors were foci of POH (as in
Figure 2A), in which blood vessels within a given area actually eclipsed the bone itself in prominence. In our clinical experience, this prominence of vascularity is not seen in NHO at any stage in its evolution. Finally, these differences in vascularity between genetic and nonhereditary were not as prominent in areas of mature lamellar bone. In our experience, overtime human HO maturation progresses toward an appearance that resembles native skeletal elements. It seems that those distinguishing histologic features of genetic versus NHO (whether osseous, cartilaginous, or vascular) are lost with maturation of the lesion.
Studies have shown that osteogenesis is closely tied to vascularization.
3,39-45 Hypoxia-inducible factor 1α (Hif-1α) regulates responses to low oxygen availability and is necessary for normal osteogenesis.
44,45 Hif-1α causes an increase in vascular endothelial growth factor (VEGF) within osteoblasts, stimulating the angiogenesis that drives osteogenesis.
45 VEGF expression has been shown in areas of ectopic ossification,
39-41 suggesting that angiogenesis is also necessary for HO. Studies show that mice with increased Hif-1α activation in osteoblasts have markedly increased vascularity and increased bone production, and that those lacking Hif-1α have impaired angiogenesis and bone formation. This clinical study in subtypes of human HO provides yet more evidence of the proximal linkage between bone and vasculogenesis. Of note, experimental models of HO suggest that the lymphatic system may also be an important environmental contributor to the formation of HO.
46 In this context, surgical lymph node excision decreased HO formation in a trauma-induced mouse model.
46 An interesting future study would include the characterization of lymphatic ingrowth and distribution across human HO samples.
Current therapies for HO are largely ineffective, and HO remains a debilitating spectrum of diseases. Nonsteroidal anti-inflammatory drugs have been used to prevent NHO; however, these carry a risk of long-bone nonunion and gastrointestinal bleeding.
3 Radiation has been used effectively as preventative therapy following hip arthroplasty,
3 although this is a less feasible therapy in patients with genetic HO who may have progressive and widespread lesions. Bisphosphonates have been used by some, but their efficacy remains unclear. Lesions often recur following surgical excision, leading to soft tissue loss and increased morbidity.
2,3,11,12 In conjunction with previous studies showing the close relationship with angiogenesis, our findings suggest that antiangiogenic therapeutic agents that target vasculogenic growth factors, particularly Hif-1α or VEGF, may be useful in the treatment of HO. Neutralizing antibodies to VEGF (Bevacizumab, Avastin) have been studied in numerous examples of pathologic angiogenesis
47 and may be efficacious in HO; however, future experimental studies to assess targeting of VEGF signaling in HO are warranted. Such agents could lead to a critical reduction in vasculogenesis that is necessary for new heterotopic bone formation.
In summary, both genetic and nonhereditary forms of HO show temporospatial variation in vascularity. Both POH and FOP demonstrate a comparatively richer vascular environment than NHO. Further studies are necessary to explore the relationship between angiogenesis and osteogenesis in HO, and to determine whether targeting angiogenesis may be of therapeutic benefit.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: AWJ was supported by the NIH/NIAMS (R01 AR070773, K08 AR068316), NIH/NIDCR (R21 DE027922), Department of Defense (W81XWH-18-1-0121, W81XWH-18-1-0336, W81XWH-18-10613), the American Cancer Society (Research Scholar Grant, RSG-18-027-01-CSM), the Orthopaedic Research and Education Foundation with funding provided by the Musculoskeletal Transplant Foundation, the Maryland Stem Cell Research Foundation, and the Musculoskeletal Transplant Foundation. EMS was supported by the Progressive Osseous Heteroplasia Association, the International FOP Association, and the Center for Research in FOP and Related Disorders. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or Department of Defense.