Authors: Morgan N. McLuckey, Erik A. Imel, Monica M. Forbes-Amrhein
Categories: Article
Source: Pediatric radiology
Authors: Morgan N. McLuckey, Erik A. Imel, Monica M. Forbes-Amrhein
Osteopetrosis describes several types of rare sclerosing bone dysplasias of varying clinical and radiographic severity. The classic autosomal dominant subtype emerges most often in adolescence but can present from infancy through adulthood. The autosomal recessive osteopetrosis, or “malignant infantile osteopetrosis”, presents in infancy with a grimmer prognosis, though the autosomal dominant forms (often mislabeled as “benign”) actually can have life-threatening consequences as well. Often osteopetrosis is detected due to skeletal findings on radiographs performed to evaluate injury or as an incidental finding during evaluation for illness. Given the varied phenotypic severity and presentations at different ages, radiologists play an integral role in the care of these patients both in diagnosis and in clinical evaluation and monitoring. A deeper understanding of the underlying genetic basis of the disease can aid in the radiologist in diagnosis and in anticipation of unique complications. An overview of current clinical management is also discussed.
Osteopetrosis is a rare genetic disorder characterized by increased bone mass and density combined with fragility due to dysfunctional osteoclast formation and/or function. First recorded by German radiologist Heinrich Albers-Schönberg in 1904, the term “osteopetrosis” (‘osteo’ meaning bone and ‘petros’ meaning stone), was introduced by Karshner in 1926 to describe the characteristic appearance of the skeleton on radiographs [1, 2]. The term osteopetrosis now applies to a group of sclerosing bone dysplasias caused by osteoclast dysfunction [3, 4]. Varying genetic mutations and incomplete penetrance confer markedly heterogeneous clinical characteristics and disease course. To better understand how different genetic mutations produce the clinical manifestations of osteopetrosis, a brief discussion of osteoclast formation and function is warranted.
Bone is comprised of both organic (collagen and other non-collagenous proteins) and inorganic (predominantly calcium hydroxyapatite) components. Growth (modeling), repair (remodeling), and mineral homeostasis are accomplished by strategic resorption of these elements by osteoclasts and coordinated new bone formation by osteoblasts [5, 6]. Osteoclasts are multinucleated cells formed from fusion of monocyte lineage progenitor cells. These mononuclear osteoclast precursors are stimulated by two cytokines, macrophage-colony stimulating factor (M-CSF) and receptor activator of nuclear factor-κB ligand (RANKL). The binding of RANKL to its receptor, receptor activator of nuclear factor-κB (RANK), provides key signals for osteoclast differentiation and bone-resorbing function of mature osteoclasts (Fig. 1). RANKL is a type 2 membrane protein, which belongs to the tumor necrosis factor (TNF) superfamily which is encoded by the gene TNFRSF11A. RANK is a type 1 transmembrane protein that also belongs to the TNF superfamily and is encoded by the gene TNFSF11 [3].
Normal osteoclast function relies on formation of the ruffled border, which facilitates bone resorption via acidification and enzymatic degradation of the matrix. The osteoclast is secured to the bone surface, forming an isolated microenvironment between the osteoclast and bone surface. Complex intracellular systems begin trafficking late endosomes and lysosomes, bearing machinery such as the vacuolar-type proton pump, ATPase (V-ATPase), and CIC-7 chloride channel to the ruffled border [5]. The V-ATPase acidifies the extracellular space through transfer of intracellular protons into the lacuna. A supply of intracellular protons is derived from a reaction catalyzed by carbonic anhydrase. The ClC-7 chloride channel functions to maintain electroneutrality of the osteoclast and acidification of the extracellular space [7]. Lysosomal proteases such as cathepsin K and matrix metalloproteases (MMP) are deposited to further degrade the organic matrix (Fig. 1). The degraded matrix is then recycled by vesicles.
To date, numerous genes critical to the function of mature osteoclasts have been isolated. T cell immune regulator 1 (TCIRG1) encodes a crucial subunit of the V-ATPase supplying acid equivalents. Chloride channel 7 (CLCN7) encodes CIC-7, and osteopetrosis-associated transmembrane protein 1 (OSTM1) encodes a subunit critical in the stabilization of ClC-7. The genes sorting nexin 10 (SNX10) and pleckstrin homology domain-containing protein, family m, member 1 (PLEKHM1) are crucial to endosomal and lysosomal trafficking pathways, enabling normal ruffled border formation. Protons for acidification would not be available without carbonic anhydrase, encoded by the CA2 gene [3, 8]. An X-linked form is caused by variants in the IKBKG gene which encodes NEMO, the regulatory subunit of IKK complex, which activates the NF-kB transcription factor.
The most classic manifestation of osteopetrosis is the autosomal dominant subtype (ADO), often eponymously referred to as “Albers-Schönberg disease”. ADO was historically regarded as having two separate subtypes (ADO1 and ADO2); however, genetic analysis has revealed that the disease previously termed ADO1 is caused by osteoblast dysfunction (due to variants in LRP5), rather than osteoclast dysfunction, and therefore does not reflect a true osteopetrosis. In addition, ADO1 is not accompanied by fragility. ADO1 is now categorized within the osteosclerotic, osteoblast-driven bone disorders [4]. Thus, in our manuscript any reference made to ADO refers to ADO2 [9, 10]. While difficult to estimate due to rarity, the incidence of ADO is approximately 20,000 [11]. ADO results primarily from a heterozygous mutation in the CLCN7 gene causing impaired osteoclast function, but with normal or elevated osteoclast numbers on bone biopsy (consequently, osteoclast-rich osteopetrosis). Recently, variants in TCIRG1 have also been demonstrated to cause a phenotype in a dominant manner [12].
Variable penetrance results in a wide range of disease severity and ages at presentation in ADO, affecting both children and adults. The disease may present anywhere from infancy to adulthood with mean age at diagnosis 15 years. Approximately one third of individuals will be asymptomatic carriers with no radiographic findings of osteopetrosis; the other two-thirds will show varying disease manifestations [9]. Symptoms and radiographic findings will tend to progress over time (see Fig. 2) [9, 13]. Those with earlier manifestations of the disease tend to have a more severe phenotype (see Fig. 3) [9].
Autosomal recessive osteopetrosis (ARO; also known as “malignant infantile osteopetrosis”) is rare and results from biallelic mutations in a variety of genes regulating either osteoclast function (osteoclast-rich osteopetroses) or formation (osteoclast-poor osteopetroses) (see Fig. 1). Estimated incidence is 250,000, but is higher in some populations due to consanguinity or geographic isolation [14, 15]. In contrast with ADO, ARO will typically present in the first year of life. Most ARO subtypes will cause severe disease and present with diffuse osteosclerosis in infancy; however, more intermediate forms of the disease do exist [3]. While morbidity from fractures and skeletal deformity complicate the disease course, mortality most often results from loss of functional bone marrow, resulting in cytopenias and infection.
Osteoclast-poor disease is characterized by lack of osteoclast formation from undifferentiated monocyte-lineage cells due to mutations in either TNFSF11 or TNFSF11A. In contrast, mutations that inhibit normal function of the ruffled border impair the osteoclasts’ ability to acidify and resorb bone, leading to increased osteoclast production and hence osteoclast-rich osteopetrosis. There are many genes whose variants result in osteoclast-rich ARO. TCIRG1 mutations account for approximately 70% of cases and is one of the most severe forms of the disease [3]. A biallelic mutation in CLCN7 (as opposed to the heterozygous ADO mutation) results in another of the most severe clinical manifestations of osteopetrosis and accounts for approximately 15% of ARO cases [3]. An additional subtype with severe disease phenotype is OSTM1 mutation, which results in a failure of ClC-7 chloride channel formation. Mutations in PLEKHM1 and SNX10 result in more intermediate and variable manifestations, respectively. Osteopetrosis with ectodermal dysplasia and immune defect (OLEDAID) is an extremely rare variant of X-linked recessive osteopetrosis (IKBKG gene) that affects immune cells in addition to osteoclasts, both cell types being of hematopoietic origin [4, 16].
The radiologist plays a key role in both the initial diagnosis of osteopetrosis as well as in the assessment of complications. Demonstration of classic radiographic features can make the diagnosis [17], and imaging findings may even be the first evidence of the disease. Therefore, it is incumbent upon the radiologist to suggest osteopetrosis when classic radiographic features are encountered, as well as to suggest referral to an endocrinologist to begin the appropriate clinical workup. Classic findings include diffuse sclerosis of the axial and appendicular skeleton, sclerosis of the skull base (see Fig. 4a–b) and vertebral bodies (“sandwich” or “anvil” vertebrae), widening of the metaphyses of long bones (Erlenmeyer flask deformity), “bone-in-bone” or “endobone” appearance, and banded metaphyses [14]. The degree of sclerosis of the vertebral body endplates can vary and at the extreme end result in an anvil-shaped appearance on the lateral radiographs, or diffuse high density throughout the vertebral body (see Fig. 5) [18].
These imaging features can be understood by osteoclast dysfunction as the underlying cause of disease. Suboptimal osteoclastic resorption of calcified cartilage at the primary spongiosa of the metaphysis and persistence of primitive woven bone (over the stronger mature lamellar bone) accounts for the overly dense, homogeneous radiographic appearance. Impaired osteoclast function also limits repair of bone microdamage and contributes to brittle bone quality [6, 19]. During growth, long bones are modeled to optimize their load-bearing-to-weight ratio by tapering towards to diaphysis. Deficient periosteal resorption at the metaphyses results in decreased/absent diaphyseal tapering, described as the Erlenmeyer flask deformity (see Fig. 3) [20]. Likewise, faulty endosteal resorption accounts for increased cortical thickness and insufficient medullary space. Another characteristic radiographic finding is the “bone-in-bone” pattern found at secondary ossification centers often most visible in the pelvis and in the bones of the wrist, hands, ankles and feet, but seen variably in other bones as well (see Fig. 4c). The pathophysiology is still not clear; however, it has been suggested to reflect the outer dense zone of provisional calcification underlying the less dense post-natal metaphyseal bone, and central inherently dense natal bone [21]. Another hypothesis, proposed in the setting of ADO, is that there are periods of better versus worse osteoclast resorption leading to this appearance. This hypothesis is supported by the fact that in ADO due to CLCN7 variants, some individuals have apparently normal skeletons radiographically (‘carriers’ of the dominant disease) and normal osteoclast resorption on culture [9, 22]. Banded metaphyses are another characteristic feature of osteopetrosis, which may be related to the inability of osteoclasts to mobilize calcium from the bone. The serum calcium in these patients is heavily dependent on calcium absorption from the gut, however variable availability of absorbed calcium may result in variable density of new bone growth in osteopetrosis, manifested as banded metaphyses [21]. In some cases, this may result in an appearance of osteopetrorickets, where there is insufficient mineral available to the growth plate, despite high mineral content in the older parts of the bone.
The radiographic spectrum of disease is broad, particularly in the case of ADO where clinical phenotypes are quite variable. Some clinical practitioners use dual-energy x-ray absorptiometry (DXA) to assess the increased bone mineral density (BMD) seen in patients with osteopetrosis. DXA can characterize disease severity and monitor response to treatment with z-scores generally +5 or greater in patients with osteopetrosis (see Fig. 6) [18, 23]; however elevated BMD is not specific to osteopetrosis. Other current research is directed at using quantitative CT to characterize disease severity and track disease over time [18, 24].
Most clinical symptoms are related to the accumulation of poor-quality bone. There is increased risk of fracture from minor trauma or repetitive stress due to the fragile bone, and resultant deformity due to poor repair (Fig. 7). Abnormal mineral homeostasis is due to the inability to mobilize calcium from the bones, which is performed by osteoclasts. Deficiency of the medullary space will cause hematologic dysfunction including anemia, thrombocytopenia, leukopenia, extramedullary hematopoiesis, and hepatosplenomegaly, and can be life-threatening in either the recessive or dominant forms. There is an increased risk of osteomyelitis (Fig. 8) especially of the jaw or after surgical procedures, as well as dental abnormalities [9]. Foraminal narrowing in the skull base can causes blindness, deafness, and other nerve palsies (Fig. 9) [25]. Routine CT and ophthalmology follow-up is crucial to assess optic canal narrowing. Vision loss will typically begin to occur while still growing in patients less than 18 years old; past this age, it is unlikely to occur [9]. Other findings include macrocephaly and frontal bossing, hydrocephalus, and otologic abnormalities [26, 27].
Several ARO subtypes will manifest additional unique findings. TCIRG1, the most common ARO subtype, results in reduced gastric acidity due to the failure of the V-ATPase to excrete intracellular protons into the stomach, which ultimately reduces gut calcium absorption. Hypocalcemia leads to elevated parathyroid hormone levels to mobilize the calcium stores from the bone. Unfortunately, the malfunctioning osteoclasts have suboptimal results. Thus, the persistent hyperparathyroidism and subsequent hypophosphatemia will result in rickets. These patients will typically have concurrent imaging manifestations of rickets and osteopetrosis (sometimes termed “osteopetrorickets”) (Fig. 10) [28]. Biallelic CLCN7 or OSTM1 mutations are known as the “neuropathic” subtypes of osteopetrosis. Approximately half of these patients suffer from an ultimately fatal primary neurodegeneration. Symptoms may include hyperreflexia, clonus, irritability, and seizures. Global cerebral atrophy may be seen on imaging [29] ( Fig. 11).
A suspicion of osteopetrosis, either clinically or by detection of features on radiograph, should prompt the radiologist to recommend a radiographic skeletal survey using an imaging protocol for the evaluation of dysplasia. Demonstration of the classic radiographic features is sufficient for diagnosis, however additional testing may be appropriate in certain settings. Genetic testing is generally recommended after diagnosis in order to distinguish forms of the disease with unique complications to anticipate (such as rickets in ARO with TCIRG1 mutation or neurodegeneration in ARO with CLCN7 and OSTM1 mutations) [17]. Presence of some mutations may also determine eligibility for certain treatments. Bone biopsy is usually unnecessary but may prove useful in the setting of indeterminant genetics to identify osteoclast-rich versus osteoclast-poor disease, which has implications for treatment [17].
Limited treatments are available for osteopetrosis. Treatment is largely supportive and aimed at managing complications such as fracture and infection. Calcium and vitamin D supplementation is first-line therapy for hypocalcemia and secondary hyperparathyroidism. Red blood cell transfusion is routinely given for symptomatic anemia [17]. Interferon-gamma is approved in severe ARO [17] and increases bone resorption, increases hematopoiesis, and improves leukocyte function [30]. However, two groups recently demonstrated a lack of response when using interferon-gamma for ADO as well as poor tolerability of this regimen [31, 32]. While high dose calcitriol was previously suggested as a therapy to stimulate osteoclast resorption, animal model data indicated that this was also not beneficial to the bone, and treated mice actually had higher bone mass [33].
Hematopoietic stem cell transplant (HSCT) provides deficient patients with a source for healthy osteoclast population and is the only therapy that has been showed to reverse progression of the disease [34, 35]. Primary indication for transplant is bone marrow failure and secondary indications include focal neurologic deficits and unremitting pain. Transplant is usually performed before one year of age [17]. While there is high mortality risk in ARO, there is also high mortality risk from complications of HSCT in osteopetrosis [36]. Patients who undergo successful transplant will see radiologic improvement of disease. However, some genetic mutations are a contraindication to HSCT. Patients with biallelic CLCN7 and OSTM1 mutations are not candidates for transplant due to severe CNS manifestations that convey poor neurologic prognosis independent of transplant. HSCT is contraindicated with the TNFSF11 mutation due to lack of marrow response to transplant [3]. Active investigation is underway to optimize HSCT protocols to improve treatment of the most severe forms of osteopetrosis. Non-genotoxic myeloablation aims to reduce complications of conditioning for HSCT, as an example.
Corticosteroids are a potential second-line therapy in ARO when HSCT is not possible, though benefits are not uniform and there is potential for adverse effects [17]. Unfortunately, long term glucocorticoid treatment also causes decreased bone turnover. RANK-L replacement therapies, being explored in mice, could provide a treatment option for osteoclast-poor (TNFSF11) osteopetrosis. Other pre-clinical research centers on in utero HSCT, viral gene therapy, and use of corrected inducible pluripotent stem cells [37].
Osteopetrosis is an osteoclastic disease with clear genetic basis. The age and severity at presentation are highly variable. ADO may present anywhere from infancy to adulthood though most often manifests in adolescence. ARO always manifests in infancy and has an almost exclusively severe phenotype; unique genetic subtypes can complicate the disease course with symptoms such as neurodegeneration. Radiologists played a pivotal role in first describing the disease and remain critical in the recognition of the radiographic features and detection of complications such as fracture, infection, and foraminal narrowing. Treatment options for osteopetrosis unfortunately remainlimited, though HSCT is a potentially curative option for some of the most limiting forms of the disease.