SEARCH

SEARCH BY CITATION

Keywords:

  • OSTEOPETROSIS;
  • OSTEOPETRORICKETS;
  • OSTEOMALACIA;
  • CLCN7;
  • TCIRG1

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Osteopetrosis is an inherited disorder of impaired bone resorption, with the most commonly affected genes being CLCN7 and TCIRG1, encoding the Cl/H+ exchanger CLC-7 and the a3 subunit of the vacuolar H+-ATPase, respectively. We and others have previously shown that the disease is frequently accompanied by osteomalacia, and that this additional pathology is also found in Tcirg1-deficient oc/oc mice. The remaining question was whether osteoid enrichment is specifically associated with TCIRG1 inactivation, or whether CLCN7 mutations would also cause skeletal mineralization defects. Here we describe a complete osteologic assessment of one family carrying a novel mutation in CLCN7 (D145G), which impairs the activation and relaxation kinetics of the CLC-7 ion transporter. The two siblings carrying the mutation in the homozygous state displayed high bone mass, increased serum levels of bone formation markers, but no impairment of calcium homeostasis when compared to the other family members. Most importantly, however, undecalcified processing of an iliac crest biopsy from one of the affected children clearly demonstrated a pathological increase of trabecular bone mass, but no signs of osteomalacia. Given the potential relevance of these findings we additionally performed undecalcified histology of iliac crest biopsies from seven additional cases with osteopetrosis caused by a mutation in TNFRSF11A (n = 1), CLCN7 (n = 3), or TCIRG1 (n = 3). Here we observed that all cases with TCIRG1-dependent osteopetrosis displayed severe osteoid accumulation and decreased calcium content within the mineralized matrix. In contrast, there was no detectable bone mineralization defect in the cases with TNFRSF11A-dependent or CLCN7-dependent osteopetrosis. Taken together, our analysis demonstrates that CLCN7 and TCIRG1 mutations differentially affect bone matrix mineralization, and that there is a need to modify the current classification of osteopetrosis. © 2014 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Osteopetrosis (OP) is a high bone mass disorder characterized by defective bone resorption.[1] Based on the analysis of genetically modified mouse models and on the identification of genes whose mutational malfunction causes OP in humans, it is possible to discriminate between two forms of the disease.[2] Osteoclast-poor OP is characterized by an impaired ability of hematopoietic precursor cells to differentiate into functional osteoclasts and can be caused by inactivating mutations in the genes encoding either receptor activator of NF-κB ligand (RANKL) (TNFSF11) or its receptor RANK (TNFRSF11A).[3] Osteoclast-rich OP is characterized by presence of high numbers of nonfunctional osteoclasts and can be caused by inactivating mutations in the genes TCIRG1, CLCN7, OSTM1, CA2, PLEKHM1, or SNX10.[2, 4] With the exception of CLCN7, in which heterozygous mutations can cause a less severe form of OP, all other forms (including osteoclast-poor OP) are inherited in an autosomal recessive fashion, and within this group of patients, TCIRG1 is the most commonly affected gene, accounting for roughly 50% of the cases.[5-7]

The protein encoded by the TCIRG1 gene is a subunit of the vacuolar proton pump involved in the process of extracellular acidification, whereby osteoclasts dissolve the mineralized phase of the bone matrix.[8] Tcirg1 was initially identified in mice as an osteoclast-specific gene, whose targeted deletion causes osteoclast-rich OP.[9] A deletion within the murine Tcirg1 gene was also found in a spontaneously derived osteopetrotic mouse model termed oc/oc, which is important, because these mice were originally described to display additional defects of skeletal mineralization.[10, 11] We have previously shown that these defects are the consequence of hypocalcemia resulting from a combined acidification impairment of osteoclasts and gastric parietal cells.[12] In a retrospective analysis of nondecalcified bone biopsies from 21 patients diagnosed with OP, we have further observed that matrix mineralization was intact in 11 cases, whereas 10 cases were characterized by pathological accumulation of nonmineralized osteoid.[12] However, because we were not able to retrieve genotype data from the respective individuals, it was impossible to conclude that skeletal mineralization defects would specifically be caused by TCIRG1 inactivation.

Another protein required for acidification of the resorption lacunae is the anion transporter CLC-7, encoded by the CLNC7 gene.[13] CLC-7 is primarily localized to the lysosomal compartment, where it interacts with OSTM1 to mediate Cl/H+ exchange.[14-16] Although a recent study of the skeletal defects in Clcn7-deficient mice revealed that there was no severe osteoid accumulation comparable to oc/oc mice,[17] it is still unknown whether CLCN7 mutations are associated with osteoid accumulation in osteopetrotic individuals. In this regard it is also important to state that homozygous CLCN7 mutations can give rise to severe forms of infantile malignant OP with accompanying neurological deficits, but also to intermediate forms of OP, in which the affected children often suffer from pathological fractures, yet survive until adulthood without hematopoietic stem cell transfer.[18] Here we describe an osteologic assessment of one such a case of intermediate CLCN7-dependent OP, in which we did not observe a pathological osteoid enrichment following undecalcified histology of an iliac crest biopsy. Based on this potentially relevant finding we additionally analyzed bone biopsies from individuals with infantile malignant OP carrying mutations of TNFRSF11A (n = 1), CLCN7 (n = 3), or TCIRG1 (n = 3), and detected osteomalacia only in the cases with TCIRG1 mutations.

Patients and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Initial case

A 9-year-old girl presented with a history of fractures after minor trauma at the University Medical Center Hamburg. At 133-cm-tall, she was at the 42.7% percentile, and there was no retardation of growth or intellectual development in the patient's history. She showed no neurologic symptoms, especially no alteration of visual acuity and sense of hearing. The first fracture occurred at the tibia, when she was 3 years old. In the following years she complained about atraumatic fractures of the metatarsal bones and ribs. We performed dual-energy X-ray absorptiometry (DXA) to determine the bone mineral density of the patient, her parents, her sister, and her two brothers. Moreover, blood samples were drawn from all family members (with the exception of the older brother) for determination of serum parameters and genetic status, which was performed in the Departments of Clinical Chemistry and Human Genetics, respectively, according to standard protocols. All procedures were performed for diagnostic purposes; yet informed consent was obtained from all subjects and/or their parents.

Subcellular localization of CLC-7 D145G

Expression constructs for OSTM1-GFP in the pEGFP-N3 vector and CLC-7 in the pcDNA3.1(+) vector have been described.[15, 19] The D145G mutation was introduced by PCR with primers carrying this mutation, and the construct was confirmed by sequencing the complete open reading frame (ORF). Plasmid DNA encoding the respective CLC-7 construct and OSTM1-GFP (with a 50-fold excess of the former) was transfected into HeLa cells using FuGENE6 (Roche, Indianapolis, IN, USA). After 30 hours at 37°C, the cells were fixed with 4% paraformaldehyde for 15 minutes. After permeabilization with 0.1% saponin, the cells were immune-labeled with primary antibodies for CLC-7 (polyclonal rabbit 7N4B)[13] and for LAMP-2 (mouse monoclonal H4B4, from the DSHB) and subsequently with AlexaFluor-coupled secondary antibodies in PBS/3% bovine serum albumin [BSA]/0.05% saponin. Images were acquired with an LSM510 microscope equipped with a 63× 1.4 NA lens (Carl Zeiss, Jena, Germany).

Two-electrode voltage-clamp analysis

CLC-7/OSTM1 currents upon expression in Xenopus oocytes were measured as described.[15] Defolliculated oocytes were injected with complementary RNA (cRNA) encoding plasma membrane–targeted CLC-7 and OSTM1 (23 ng each) transcribed with the mMessage Machine kit (Ambion) from linearized pTLN vector. After 3 days of incubation at 17°C, currents were measured in ND96 saline (96 mM NaCl, 2 mM K-gluconate, 1.8 mM Ca-gluconate, 1 mM Mg-gluconate, 5 mM HEPES, pH 7.5) at room temperature by standard two-electrode voltage-clamp with a TurboTEC amplifier (npi electronic, Tamm, Germany) and pClamp10.2 software (Molecular Devices). To determine the rate constants of the voltage-dependent activation, the currents of the first 250 ms of depolarization to +80 mV were fitted to a single-exponential function.

Bone biopsies

An iliac crest bone biopsy was taken from the initial patient described above according to the standard procedure described by Bordier and colleagues.[20] An age-matched control biopsy was taken from a skeletal-intact donor at autopsy. We further obtained needle biopsies (Jamshidi needle diameter 3.05 mm, length of the probe at least 6 to 7 mm) from the iliac crest of seven additional individuals. Three of them carried homozygous mutations of the CLCN7 gene, three of them mutations of the TCIRG1 gene, and one patient had a homozygous mutation of the TNFRSF11A gene (Table 1). Samples were fixed overnight at 4°C in 3.7% PBS-buffered formaldehyde, dehydrated in ascending concentrations of ethanol, and embedded in methylmethacrylate as described previously.[21] Sections of 5 µm thickness were cut using a Microtec rotation microtome (Techno-Med, Munich, Germany) and stained according to standard protocols by von Kossa/van Gieson, Goldner, and toluidine blue procedures as described.[21] Parameters of static histomorphometry were quantified according to the ASBMR standards using the Osteo-Measure histomorphometry system (Osteometrics, Atlanta, GA, USA) connected to a Zeiss microscope (Carl Zeiss, Jena, Germany).[22]

Table 1. Genetic Status of Osteopetrotic Individuals Included in the Histological Analysis of Iliac Crest Biopsies
PatientAge (years)SexGeneExon/intronMutation DNA levelMutation protein levelStatus
13MCLCN7Exon 1c.139C>Tp.Gln47TerHomozygous
22MCLCN7Exon 4c.296A>Gp.Tyr99CysHomozygous
31MCLCN7Exon 17c.1531G>Cp.Ala511ProHomozygous
41MTCIRG1Exon 7; intron 19c.647G>A; c.2415–1G>Tp.Trp216Ter; –Compound heterozygous
54FTCIRG1Exon 9c.1007delTp.Leu336Argfs*10Homozygous
61FTCIRG1Intron 2; exon 19c.116+1G>A; c.2282delG–; p.Gly761Alafs*22Compound heterozygous
712FTNFRSF11AExon 4c.328dupCp.Arg110Profs*52Homozygous

Quantitative backscattered electron imaging

The methylmethacrylate-embedded specimens were ground coplanar, polished, and carbon-coated. Measurements of the backscattered electron (LEO 435 VP; Leo Electron Microscopy Ltd., UK) were carried out at an accelerating voltage 20 kV, an electron beam of 580 pA, and a working distance of 20 mm (BSE Detector, Type 202; K.E. Developments Ltd., UK). The backscattered signal was calibrated using carbon and aluminum standards (MAC Consultants Ltd., UK) as described.[23] Gray-level histograms were obtained (Image J 1.42; National Institute of Health, Bethesda, MD, USA) and transferred into Ca weight as described.[21, 23] For statistical analysis the expected value (Ca mean) and the standard deviation (Ca width) of the calcium distribution were evaluated as parameters for the mean calcium content and the heterogeneity of the calcium distribution within the mineralized bone matrix.

Statistical analysis

Results are presented as bar graphs, indicating mean ± SD. Statistical analysis were performed with the two-tailed Student's t test for unpaired data. Significance of all analysis was set at p < 0.05.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Clinical assessment of a patient with autosomal recessive osteopetrosis

A 9-year-old girl presented with a history of fractures at the University Medical Center Hamburg. Radiography of the skull revealed thickening of the vault and loss of the mandibular angle (Fig. 1A). In the spine we observed a sandwich appearance (“rugger jersey spine”)[24] with abnormally dense bone at the vertebral endplates (Fig. 1B). The pelvis showed diffusely increased bone density with hyperdense areas at the iliac wings and the acetabulum, whereas the femoral neck contained areas of hyperdense bone with irregularly shaped growth plates (Fig. 1C). Radiography at the proximal tibial metaphysis revealed a pathologic enlargement, indicating a lack of normal metaphyseal modeling, which was also confirmed by the presence of band-like sclerosis in the distal metaphysis (Fig. 1D). In addition, the phalangeal and metatarsal metaphyses displayed a band-like sclerosis with interspersed lines of normal bone density (Fig. 1E, F). We next performed dual-energy X-ray absorptiometry (DXA) to determine the bone mineral density of the patient, her parents, her sister, and her two brothers (Fig. 1G). Here we observed normal bone mass in all individuals except the patient and her younger brother, for whom the Z-scores were +10.3 and +4.4, respectively.

image

Figure 1. Radiographic analysis of a case with CLCN7-dependent OP. (A) Anterior-posterior and lateral radiographs of the skull showing thickening of the vault. (B) Anterior-posterior and lateral radiographs of the spine showing a sandwich appearance (“Rugger Jersey Spine”) with abnormally dense vertebral endplates. (C) Anterior-posterior radiograph of the pelvis, where we observed a diffusely increased bone density with marked osteosclerosis at iliac wings and acetabulum. (D) Radiographs of the tibia revealing clubbed proximal metaphyses and band-like sclerosis in the distal metaphysis. (E, F) Radiographs of the phalangeal and metatarsal metaphyses displaying band-like sclerosis with interspersed lines of normal bone density. (G) Pedigree of the family with squares representing males and circles representing females. The ages and the Z-scores measured by DXA bone densitometry are indicated on the right for the parents, the sister, and the two brothers of the patient (female, 9 years old). OP = osteopetrosis; DXA = dual-energy X-ray absorptiometry.

Download figure to PowerPoint

Identification of a novel CLCN7 mutation in the affected individual

Because these findings suggested an autosomal recessive high bone mass disorder, we performed genomewide homozygosity mapping and found a candidate locus on chromosome 16p13 (Fig. 2A). Because this region included the CLCN7 gene, we performed direct sequencing of the CLCN7 exons and detected a homozygous missense mutation (c.434A > G, p.D145G) in exon 5, which segregated with the disease phenotype (Fig. 2B). Sequence homology searches revealed that D145 is highly conserved between species and within the family of chloride channels (Fig. 2C). When compared to the S. typhimurium CLC transporter as a reference,[25] D145G is localized within intramembrane helix B in close proximity to G203 (Fig. 2D), the mutation of which causes autosomal-recessive intermediate osteopetrosis.[26]

image

Figure 2. Identification of a novel OP-causing mutation in CLC-7. (A) Results of a haplotype analysis of a candidate locus on chromosome 16p13 between markers rs2107321 and rs40129. (B) Sequencing revealed a homozygous missense mutation (c.434A > G, p.D145G) in exon 5 of CLCN7 segregating with high bone mass. (C) D145 is highly conserved between species (upper panel) and within the family of chloride channels (lower panel; regions of high homology are indicated by yellow bars). (D) Location of p.D145 in chain-b of the S. typhimurium ClC channel and a second missense mutation at position p.G203 previously reported to cause autosomal recessive OP. OP = osteopetrosis; EcClC = E. coli ClC; StClC = Salmonella typhimurium ClC.

Download figure to PowerPoint

We next aimed at scrutinizing the dysfunction of CLC-7 D145G. Upon heterologous expression, this mutant localized correctly to lysosomes (Fig. 3A). OSTM1, the obligate β-subunit of CLC-7,[15, 16] was co-trafficking with CLC-7 D145G, thus demonstrating that the interaction between the CLC-7 and OSTM1 is not abolished by the D145G mutation. We next examined whether the D145G mutation alters CLC-7–mediated ion transport, similar to an equivalent mutation (D136G) in the paralogous CLC-1, which causes recessive myotonia congenita and alters the voltage-dependent gating of this plasma membrane chloride channel.[27, 28] Therefore, we inserted the D145G mutation into a partially plasma membrane–localized CLC-7 with disrupted amino-terminal sorting motifs.[19] Coexpression of this construct with OSTM1 in Xenopus oocytes yielded outwardly rectifying currents similar to those evoked by wild-type CLC-7 (Fig. 3B). Although the voltage dependence of CLC-7/OSTM1-mediated currents was unaltered by the D145G mutation, the gating kinetics of CLC-7/OSTM1 were moderately accelerated (rate constants for the activation at +80 mV: 265.5 ± 40.5 ms for wild-type [mean ± SEM for 10 oocytes] versus 165.8 ± 19.1 ms for D145G [7 oocytes]) (Fig. 3C).

image

Figure 3. Subcellular localization and biophysical characterization of the CLC-7 (D145G) mutant. (A) Subcellular localization of CLC-7 D145G (immunolabeled, yellow) coexpressed with an OSTM1-GFP fusion protein (turquoise) in HeLa cells. Both proteins localize to lysosomes immunostained for LAMP-2 (magenta). (B) Representative current traces from two-electrode voltage-clamp measurements of Xenopus oocytes coexpressing a partially plasma membrane–localized CLC-7, without (“WT”) or with the D145G mutation, and OSTM1. The voltage was clamped from −80 to +80 mV in 2-second steps of 20 mV followed by a 0.5-second hyperpolarization at −80 mV, with a holding potential at −30 mV (clamp protocol shown in the inset). Both “WT” and D145G yielded time-dependent outwardly rectifying currents. Note that the current/voltage relation was not altered by the D145G mutation, whereas both activation and deactivation kinetics were accelerated by the mutation. (C) Moderately accelerated activation kinetics of CLC-7/OSTM1 by the D145G mutation. Rate constants for polarization to +80 mV were determined for 10 (“WT”) and 7 (D145G) oocytes from three independent batches. Bars represent mean ± SD. Asterisks indicate statistically significant differences between mutant and wild-type CLC-7/OSTM1.

Download figure to PowerPoint

The homozygous CLC-7 mutation D145G does not impair calcium homeostasis and bone matrix mineralization

To analyze whether the identified CLCN7 mutation is associated with disturbed calcium homeostasis, we measured the serum levels of calcium, parathyroid hormone (PTH), phosphate, and 25-hydroxyvitamin-D3 (25-OH-D3) in comparison to the parents, the unaffected sister (all carrying the mutation in the heterozygous state), and the affected brother (carrying the mutation in the homozygous state). We did not detect any abnormalities specific to the two affected children, albeit PTH and 25-OH-D3 levels were found to be highest in the initially identified patient (Fig. 4A). Interestingly, however, markers of bone formation (osteocalcin and bone-specific alkaline phosphatase) were found to be elevated in the two affected children, not only compared to the parents, but also compared to the unaffected sister (Fig. 4B). Although these data are in line with the hypothesis that there is a secondary increase of bone formation in individuals with osteoclast-rich OP,[29] the major conclusion from the serum analysis was that basal calcium homeostasis was unaffected by the D145G mutation of CLC-7.

image

Figure 4. The homozygous CLC-7 mutation D145G does not affect basal calcium homeostasis. (A) Serum analysis of the patient (9 years old), her affected brother (17 years old), her parents (43 and 44 years old), and her unaffected sister (18 years old), demonstrating intact calcium homeostasis in the two children carrying the D145G mutation on both chromosomes (black bars). (B) Serum bone turnover markers indicating increased bone formation in the two patients.

Download figure to PowerPoint

Because we have previously observed pathologic accumulation of nonmineralized osteoid in nongenotyped patients with OP,[12] the most relevant information was, however, obtained by analyzing an iliac crest bone biopsy from the initial patient described above. Using undecalcified histology followed by von Kossa/van Gieson staining, we were able to confirm the expected increase of the trabecular bone volume in the patient's biopsy, which was compared to an iliac crest biopsy of an age-matched control individual (Fig. 5A). In contrast, although osteoid seams were present in the patient's biopsy, there were no signs of impaired bone matrix mineralization (Fig. 5B). This was confirmed by histomorphometric quantification (Fig. 5C); we did not observe a pathological increase of osteoid thickness, thereby ruling out osteomalacia in addition to the increased trabecular bone volume. In contrast, the osteoid volume per bone volume was increased in the patient's biopsy, but this was readily explained by an increased osteoid surface, indicative of excessive bone formation. Likewise, compared to the control biopsy and to the reference range reported by others,[30] we further observed increased surface indices not only for osteoclasts, but also for osteoblasts. Taken together, these data demonstrated that the identified homozygous CLC-7 mutation may trigger increased bone formation, but does not cause impaired bone mineralization.

image

Figure 5. The homozygous CLC-7 mutation D145G does not cause osteomalacia. (A) Von Kossa/van Gieson staining of undecalcified sections from iliac crest biopsies (Bars = 500 µm) demonstrating high bone mass in the case of CLCN7-dependent OP in left panels. (B) Higher magnification of the same samples to demonstrating the absence of osteomalacia (Bars = 100 µm). (C) Histomorphometric quantification of BV/TV, OV/BV, O.Th, OS/BS, OcS/BS, and ObS/BS. OP = osteopetrosis; BV/TV = bone volume per tissue volume; OV/BV = osteoid volume per bone volume; O.Th = osteoid thickness; OS/BS = osteoid surface per bone surface; OcS/BS = osteoclast surface per bone surface; ObS/BS = osteoblast surface per bone surface.

Download figure to PowerPoint

Histologic comparison of CLCN7-dependent and TCIRG1-dependent OP

Based on these potentially important observations we went on to analyze biopsies from additional individuals with diagnosed osteopetrosis, three of them carrying mutations of the CLCN7 gene, three others carrying mutations of the TCIRG1 gene, and one case with TNFRSF11A-dependent OP (Table 1). Undecalcified histology followed by von Kossa/van Gieson staining allowed a clear discrimination between the groups, because all cases with TCIRG1 mutation were characterized by severe osteoid accumulation, whereas the other cases did not display this additional pathology (Fig. 6A). Likewise, whereas histomorphometric quantification revealed high trabecular bone mass in all cases, osteoid parameters were found to be in the pathological range only in the cases with TCIRG1 mutation (Fig. 6B). Here we detected significant differences compared to the cases with CLCN7 mutation, not only for the osteoid volume per volume, but also for osteoid thickness and osteoid surface per bone surface. Because some previous histological studies of bone biopsies from individuals with OP of unknown genotype have reported the presence of an abnormal bone matrix,[31-33] it is important to state further that the osteoid we found in the three cases of TCIRG1-dependent OP not only contained osteocytes, but also a regular distribution of collagen fibers as assessed by polarized light microscopy (Fig. 6C).

image

Figure 6. Undecalcified histology of bone biopsies from osteopetrotic individuals with defined gene mutations. (A) Von Kossa/van Gieson staining of representative undecalcified sections from cases with CLCN7-dependent and TCIRG1-dependent OP (Bars = 100 µm). Mineralized bone matrix is stained in black, nonmineralized osteoid is stained in red. (B) Quantification of bone volume per tissue volume and osteoid parameters in one case with a TNFRSF11A/RANK mutation, and in three cases each with CLCN7 or TCIRG1 mutations. Bars represent mean ± SD. Asterisks indicate statistically significant differences between CLCN7-dependent and TCIRG1-dependent OP. (C) Representative images of osteoid in TCIRG1-dependent OP (Bars = 50 µm). Goldner staining (left) reveals osteoblasts located on the nonmineralized matrix (black arrows), some of them becoming embedded as osteocytes (white arrows). Polarized light microscopy (right) reveals the presence of collagen fibrils within the nonmineralized matrix (white arrows). OP = osteopetrosis; RANK = receptor activator of NF-κB.

Download figure to PowerPoint

In contrast, we did not observe significant differences between the groups in terms of osteoclast surface per bone surface or osteoblast surface per bone surface (Fig. 7A). Interestingly, however, all cases with CLCN7 or TCIRG1 mutation displayed a pathological enrichment of fibrous tissue in the bone marrow, in contrast to the case with RANK inactivation, in which osteoclasts and marrow fibrosis were not detectable. Likewise, bone marrow cells were specifically found in the biopsy from the individual with TNFRSF11A-dependent OP, thereby highlighting a potentially important difference between osteoclast-rich and osteoclast-poor OP (Fig. 7B). To analyze the calcium content within the mineralized bone matrix, we finally applied quantitative backscattered electron imaging (Fig. 8A).[34] When we compared the cases with CLCN7-dependent and TCIRG1-dependent OP, we observed a significantly lower calcium content (Ca mean) in the latter cases, whereas the heterogeneity of calcium distribution (Ca width) was not different between the two groups (Fig. 8B). Taken together, these analyses demonstrate that CLCN7 and TCIRG1 mutations differentially affect bone matrix mineralization in osteopetrotic individuals.

image

Figure 7. Cellular histomorphometry of bone biopsies from osteopetrotic individuals with defined gene mutations. (A) Quantification of OcS/BS and FT/BS. Bars represent mean ± SD. Asterisks indicate statistically significant differences between CLCN7-dependent and TCIRG1-dependent OP. (B) Representative images of biopsies from cases TNFRSF11A/RANK-dependent and CLCN7-dependent OP (Bars = 50 µm). Goldner staining reveals that bone marrow was present in one case of osteoclast-poor OP (left), whereas the bone marrow was fully replaced by fibrous tissues in osteoclast-rich OP (middle). The right panel shows the presence of osteoclasts in the case of CLCN7-dependent OP. OcS/BS = osteoclast surface per bone surface; FT/BS = fibrous tissue per bone surface; OP = osteopetrosis; RANK = receptor activator of NF-κB.

Download figure to PowerPoint

image

Figure 8. Quantitative backscattered electron on bone biopsies from osteopetrotic individuals with defined gene mutations. (A) Quantitative backscattered electron images expressed by pseudo-colors (Bars = 100 µm). Highly mineralized bone is represented by brightly colored pixels, whereas lower mineralized bone areas are predominant in darker colors. (B) Quantification of the mean calcium content (Ca mean) and the heterogeneity of the calcium content (Ca width) within the mineralized bone matrix. All bars represent mean ± SD. Asterisks indicate statistically significant differences between CLCN7-dependent and TCIRG1-dependent OP. OP = osteopetrosis.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

We and others have previously reported that OP is frequently accompanied by defects of skeletal mineralization.[11, 12, 35, 36] To understand the underlying mechanisms of this pathology we have previously performed a comparative analysis of Src−/− and oc/oc mice. Both models displayed OP, but only Tcirg1-deficient oc/oc mice exhibited an additional skeletal mineralization defect. This latter phenotype was not caused by an intrinsic defect of osteoblasts and/or chondrocytes, in which Tcirg1 was not expressed. Instead, only oc/oc but not Src−/− mice displayed severe hypocalcemia, and feeding a high-calcium liquid diet to oc/oc mice corrected their skeletal mineralization defects. We were further able to demonstrate that Tcirg1 expression in parietal cells contributes to gastric acid production, and that Src−/− mice additionally lacking the gene Cckbr, which is required for intact gastric acidification,[37] do not only display OP, but also display impaired calcium homeostasis and mineralization defects.[12] Taken together, these data strongly suggested that hypocalcemia in oc/oc mice is caused by a combined acidification defect of osteoclasts and parietal cells. Likewise, we could show the presence of the TCIRG1 protein in human stomach sections and diagnosed increased gastric pH in one individual carrying a homozygous mutation (R56W) of the TCIRG1 gene. Most importantly, our histomorphometric analysis of an undecalcified bone biopsy from this patient revealed the coexistence of OP and osteomalacia.[12] The remaining question, however, was whether this specific pathology is only found in individuals carrying TCIRG1 mutations, or if osteomalacia is a more generalized feature of OP, independent of the affected gene.[12]

To address this question we have first analyzed one case of CLCN7-dependent OP, in which we have identified a novel disease-causing mutation (CLC-7 D145G). We were able to demonstrate that the mutant protein is correctly targeted to its lysosomal destination, unlike it has been reported for some of the other osteopetrosis-causing CLC-7 mutations.[15, 38] Interestingly, an equivalent mutation (D136G) in human CLCN1 causes myotonia congenita, and this mutation drastically changes the voltage-dependence of CLC-1, rendering it able to be activated by hyperpolarization, instead of depolarization.[27, 28] Although it was unlikely that the D145 residue is acting as a voltage-sensor,[39] we analyzed whether the D145G mutation changes the voltage-dependence of CLC-7. When studied in the background of the partially plasma membrane–targeted CLC-7, however, the only significant change was an acceleration of the activation and relaxation kinetics. Such changes in gating kinetics were observed previously for other CLCN7 mutations underlying human OP,[15] further supporting that the D145G mutation is causative for the phenotype of our patient. It remains to be elucidated, however, whether the alterations in gating kinetics of CLC-7 D145G can alone explain the severity of the observed OP, or whether the in vivo expression levels are reduced by the mutation, as has been previously shown for another accelerating CLC-7 mutant.[13] Finally, it is important to emphasize that the homozygous D145G mutation was associated with a milder clinical expression, when compared to the additional cases of TCIRG1-dependent OP. Although it is evident that inherited diseases are generally variable in terms of clinical outcome, which is also true for CLCN7-dependent OP,[18] this observation is in full agreement with the fact that the D145G mutation does not cause a full CLC-7 inactivation.

Regardless of the underlying mechanism, however, it was important to demonstrate that the D145G mutation was only associated with high bone mass, but not with impaired calcium homeostasis and/or bone matrix mineralization. Triggered by these initial findings, we decided to analyze undecalcified bone biopsies from additional individuals with defined gene mutations. Although it is certainly required to continue with these analyses by including more cases, we truly believe that the phenotypic differences between CLCN7-dependent and TCIRG1-dependent OP are already validated through the results obtained so far. In fact, although we did not detect pathological osteoid accumulation in three additional biopsies from individuals with CLCN7-dependent OP, all three biopsies from individuals with TCIRG1-dependent OP were characterized by severe osteoid accumulation. Histomorphometry further revealed that this mineralization defect can be classified as osteomalacia, because not only the osteoid surface, but also the osteoid thickness was pathologically increased. That bone mineralization defects are specifically associated with TCIRG1 inactivation was further underscored by quantitative backscattered electron imaging (qBEI), a method analyzing calcium content and distribution within the mineralized phase.

Despite this striking phenotypic divergence between TCIRG1-dependent and CLCN7-dependent OP, we did not observe significant differences in terms of osteoblast or osteoclast numbers between these two groups. Regarding this, we would like to emphasize that we also obtained relevant information by analyzing the biopsy from a single case of osteoclast-poor OP caused by RANK inactivation. Here we did not observe pathological osteoid accumulation, either, and as expected there were no osteoclasts present in the biopsy of this patient. Importantly, however, although the bone marrow of all cases with TCIRG1-dependent and CLCN7-dependent OP was mostly replaced by fibrotic tissue, there was no detectable marrow fibrosis in the case with TNFRSF11A/RANK-dependent OP. In our opinion this latter finding is also potentially relevant, as it implies that a high osteoclast number is a critical determinant of marrow fibrosis, which likely contributes to the life-threatening hematopoietic complications in osteopetrotic individuals.[40] In this regard it is extremely important to state that the case of TNFRSF11A/RANK-dependent OP described here, despite having the same degree of bone mass increase, had a milder clinical expression in terms of immunological complications compared to the other cases. Because this is in agreement with previous reports on osteoclast-poor OP,[2, 41] it will be extremely informative to continue with a systematic histological comparison of bone biopsies from individuals with defined gene mutations. In fact, if it is true that a high osteoclast number is related to marrow fibrosis, and the same could apply for Paget's disease of bone,[42] it is not only important to understand the molecular mechanism behind such an association, it might even be therapeutically relevant for individuals with osteoclast-rich OP. More specifically, if marrow fibrosis is a major factor contributing to their hematopoietic deficits, it might be possible to reduce this pathology by anti-RANKL treatment. Although it certainly appears paradoxical to treat a disease of impaired bone resorption with an antiresorptive approach, this could be tested in preclinical settings (ie, mouse models), if a continued histological study of osteoclast-poor OP would confirm our findings from one case.

Finally, we suggest a modification of the current classification of OP, which is either based on the mode of inheritance or on the absence or presence of osteoclasts. Although TCIRG1-dependent OP certainly belongs to the autosomal recessive and osteoclast-rich variant of the disease, there is in our opinion ample evidence for an additional pathology associated with TCIRG1-deficiency, and we suggest the term “osteopetromalacia” (osteopetrosis and osteomalacia) to describe this specific disorder. Besides this nosology issue, however, there is also a clinical impact of our findings, and there are several open questions for future research. First, it might be required to provide calcium supplementation to TCIRG1-deficient individuals prior to their treatment by hematopoietic stem cell transplantation (HSCT), because it is likely that the donor-derived osteoclasts will preferentially resorb the mineralized bone matrix, thereby enhancing the severity of the osteomalacia. Second, if TCIRG1-deficient individuals generally display impaired gastric acidification, an issue that remains to be clarified, they may be at risk to develop secondary hyperparathyroidism after HSCT, because hypochlorhydria will probably persist.[12] And third, because TCIRG1-deficient individuals might also display defects of dentin and enamel mineralization, another issue that remains to be clarified, they could have higher risk for dental infections and jaw osteomyelitis, which are frequent complications in osteopetrotic individuals.[23, 43-45] Based on all these arguments we believe that it is important to continue with ongoing studies regarding genotype-phenotype correlations in OP, and these should ideally include a determination of skeletal mineralization by the use of undecalcified histology. Moreover, it may be required, at least for individuals with TCIRG1 mutations, to modify the current concepts of OP diagnosis, treatment, and post-HSCT analysis, which have recently been summarized.[46]

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

This work was supported by grants from the Deutsche Forschungsgemeinschaft (AM 103/14-2 and AM 103/18-2), from the Deutsche Zentrum für Luft- und Raumfahrt within the framework of the E-Rare (01GM0816), BIOASSET (01EC1005A) and Osteopath (01EC1006F TP2), and by TELETHON (GGP09018).

Authors' roles: Study design: FB, IK, UK, ADF, TJ, AT, AS, TS and MA; Study conduct: FB, IK, TS, JZ, KT, CFL, FTB, RS, CS, UK, ADF, UK, TJ, AT, AS, TS; Data analysis: FB, IK, JZ, JMP, TK, MH, TS, MA; Drafting Manuscript: FB, IK, MA and TS; Revising Manuscript: FB, IK, UK, ADF, AT, AS, TS and MA.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References