Osteoporosis is a skeletal disorder characterized by compromised bone strength predisposing a person to an increased risk of fracture.1 Owing to major efforts in basic, translational, and clinical research, disease awareness and therapeutic options for postmenopausal and age-related osteoporosis have increased significantly over the last decades.2–4 However, osteoporosis not only affects postmenopausal and elderly women but also can occur in men.5 Aged males who have sustained a hip fracture display excess mortality compared with women.6 Men with osteoporosis frequently suffer from an underlying pathology such as endocrine disorders, hemato-oncologic diseases, or drug-induced bone loss.7 In case of absence of secondary causes of bone loss, male idiopathic osteoporosis (MIO) is diagnosed.8 The term male idiopathic osteoporosis appears to cover a disease of multifactorial etiology with a strong genetic background. MIO has been associated with subclinical alterations of sex hormones,9 as well as changes in the growth hormone insulin-like growth factor 1 (IGF-1) axis.10, 11 Several studies based on histomorphometry12 and explant cell cultures13 provide evidence for an underlying osteoblast defect, but the detailed pathophysiology of MIO remains to be elucidated. Expanding on the concept of osteoblast deficiency, we used iliac crest biopsies of clinically homogeneous MIO patients to detect alterations in the local expression of genes involved either in osteoblast differentiation and regulation or with an association with osteoblast activity. Second, we intended to relate WNT10B, RUNX2, OSX, Osteocalcin, RANKL, OPG, and SOST expression to histomorphometric and microstructural biopsy characteristics. Stressing the translational aspect of this study, we aimed at interpreting gene expression and microstructural findings in a clinical context.
Patients and Methods
Following study approval by the local ethics committee, MIO patients were recruited at the outpatient clinic of the Medical Department II, St Vincent Hospital Vienna, Vienna, Austria. All patients signed a written informed consent prior to any study-related procedures. Medical history details, including the previous use of antiosteoporotic drugs, current medications, lifestyle habits (eg, smoking), fractures (including trauma history), and parental fractures were recorded in detail. Inclusion criteria were adapted from major interventional trials in male osteoporosis14, 15 and consisted of either (1) a femoral neck T-score of −2 or less and a lumbar spine T-score of −1 or less or (2) a femoral neck T-score of −1 or less and at least one prevalent osteoporotic fracture (ie, at least one moderate vertebral fracture according to the Genant criteria or one low-trauma peripheral fracture).16
Exclusion criteria included the previous use of any specific antiosteoporotic substances other than vitamin D and calcium supplements (eg, bisphosphonates). Moreover, hypogonadism, thyroid disorders, any malignoma history, or the use of corticosteroids, thyroid hormones, or antiepileptic substances within the last 5 years excluded patient enrollment. Age-matched control bone samples were obtained from male sudden-death or accident victims examined at the Department of Forensic Medicine of the Ludwig-Maximilian University, Munich, Germany. The procedure was approved by the university's institutional review board. A control patient was considered eligible for bone sampling if autopsy revealed no evidence of major chronic diseases such as malignancy or severe kidney or hepatic disease. Moreover, medical records and postmortem family inquiries were not allowed to reveal prevalent fragility fractures in the donor or his relatives. The use of either antiosteoporotic drugs or medications leading to secondary osteoporosis (eg, corticosteroids) further excluded donor sampling. To protect data privacy, patients and controls received an anonymous study code that was used throughout the study and its statistical analysis.
Clinical evaluation, densitometry, and X-ray imaging
Physical examination was performed, and body weight and patient height were assessed using a stadiometer with integrated weighing scale (BWB 700; Tanita, Tokyo, Japan). Fasting blood samples were drawn from all patients between 8:00 and 10:00 a.m. Routine blood and urine analyses were performed at the Vienna-based Central Routine Laboratory of the St Vincent Group. Routine blood tests included a whole blood count and determinations of serum potassium, sodium, calcium, and phosphate, as well as parathyroid hormone (PTH), 25-hydroxyvitamin D, thyroid-stimulating hormone (TSH), kidney and liver function parameters, and total testosterone measurements. Calcium and phosphate excretion was measured from a 24-hour urine collection. Moreover, patients underwent dual-energy X-ray absorptiometry (DXA) scanning of the hip and spine (Lunar iDXA; GE Healthcare, Piscataway, NJ, USA). To complete the clinical workup, vertebral fracture status was analyzed by anteroposterior and lateral X-ray studies of the thoracic and lumbar spine. The images were read by an experienced physician trained for Genant scoring.
For each patient, 2 mL of serum was stored in 1-mL voids at −70°C for batched bone marker analyses at the Department of Laboratory Medicine of the Medical University of Vienna. Bone markers were measured using electrochemiluminescence immunoassays [ECLIA; β-CrossLaps (CTX), N-MID Osteocalcin (OCN), total N-terminal type 1 procollagen propeptide (P1NP); all Roche Diagnostics, Mannheim, Germany) on a Modular Analytics E170 device (Roche Diagnostics). Because of autopsies being the source of control bone samples, serum was unavailable from these individuals. Instead, patients were compared with preexisting in-house data from marker studies in healthy controls. The data set was age-matched (n = 152), and the normal range was defined as the 5th to 95th percentile of each parameter.
Transiliac bone biopsy
In order to perform histomorphometry and structural analyses, as well as gene expression testing, each patient underwent two parallel-oriented left transiliac bone biopsies. Using a trephine needle, all biopsies were carried out under sterile conditions at the local operating theater. At the ward, patients received an analgesic premedication (lornoxicam). Before local anesthesia, sedation was achieved using midazolam. No biopsy-related complications occurred. Both biopsy cylinders (inner diameter = 7 mm) were examined visually. The larger, more intact sample was selected for subsequent histomorphometric and structural analyses and was placed in 70% ethanol. The second biopsy was submerged immediately in RNA-Later (Ambion, Warrington, UK) and stored as instructed by the manufacturer. With a median postmortem interval of 23 hours, control samples were obtained as soon as possible during forensic autopsies. Identical to patient samples, one specimen was placed in 70% ethanol, and the other was submerged in RNA-Later. The RNA sample was stored at −20°C and shipped on dry ice to the Medical University of Vienna.
Histology and histomorphometry
Specimen processing for histology and histomorphometry was performed according to Roschger and colleagues.17, 18 Briefly, all structure biopsies were fixed in ethanol, dehydrated, and embedded in polymethyl methacrylate-(PMMA). Histologic sections were cut from the blocks and stained with Goldner stain. In order to exclude malignant causes of osteoporosis such as lymphoma or mastocytosis, a certified pathologist performed routine reading in patients.
Histomorphometry was performed by the BIOQUANT Image Analysis Corporation (Nashville, TN, USA) using the BIOQUANT OSTEO Bone Biology Research System, Version 8.40.10. For each section, 25 systematically random fields of view were imaged with a ×20 objective from within the trabecular compartment. Histomorphometric data are reported according to the standardized nomenclature.19 In addition, marrow volume (Ma.V) was calculated as tissue volume minus bone volume. Fat volume (FatV) was calculated as the total adipocyte volume within bone marrow volume.
Micro–computed tomography (µCT)
The µCT imaging system (µCT40, Scanco Medical AG, Brüttisellen, Switzerland) used in this study was equipped with a 5-µm focal-spot X-ray tube as a source. A 2D charge coupled device (CCD) coupled with a thin scintillator as a detector permitted acquisition of 206 tomographic images in parallel. The long axis of the PMMA-embedded biopsy specimen was oriented along the rotation axis of the scanner. The X-ray tube was operated at 70 kVp and 114 µA with an integration time set to 200 ms. Scans were performed at an isotropic nominal resolution of 10 µm (high-resolution mode). The image data were filtered using a Gaussian filter (σ = 1.2, support = 1) to partially suppress noise in the volume. The mineralized tissue was segmented from soft tissue by a global thresholding procedure,20 with a threshold value set to 22% of the maximum grayscale value. A special contouring algorithm was used to automatically detect the envelope of the biopsy, followed by a 3D erosion algorithm to define the trabecular bone volume of interest (VOI) within the biopsy and to exclude any cortical bone. Morphometric indices were determined for the trabecular bone compartment using a direct 3D approach21 and included bone volume (BV/TV), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), trabecular number (Tb.N), connectivity density (Conn.D), and the structure model index (SMI).
RNA extraction, cDNA, and real-time PCR
Using RNAse-free instruments, a small cube of trabecular bone (approximately 5 × 5 × 5 mm) was cut out of the middle part of the intact biopsy at a laminar flow bench. To reduce marrow content, the cube was rinsed repeatedly with RNA-Later. Together with two small steel beads, the bone sample was transferred to an RNAse-free Eppendorf tube and subsequently flash frozen in liquid nitrogen. For tissue homogenization, the cooled tubes were placed in a grinding mill (3 minutes, 30 Hz). Subsequently, RNA extraction was performed on the basis of a guanidinium thiocyanate–phenol–chloroform protocol (Trizol; Invitrogen, Carlsbad, CA, USA).22 Apart from double precipitation and centrifugation at the isopropanol stage (12,000g, 10 minutes, 4°C), the manufacturer's protocol was followed. RNA quality and quantity were checked by electrophoresis and photometry at 260 and 280 nm. The 260/280-nm absorbance ratios ranged from 1.5 to 2. cDNA was sythesized from 4 µg of total RNA using a cDNA synthesis kit (High Capacity cDNA Reverse Transcription Kit; Applied Biosystems, Warrington, UK). Real-time PCR was performed using assay-on-demand primers and probes following the manufacturer's instructions. For each reaction well, the amplification mixture (20 µL) consisted of 9 µL of cDNA (dilution 1:10), 10 µL of mastermix buffer (TaqMan Universal PCR Mastermix, Applied Biosystems, Warrington, UK), and 1 µL of probe, that is,. Wnt10b (Hs00559664_m1), RUNX2 (Hs00231692_m1), OSX (Hs00541729_m1), Osteocalcin (Hs00609452_g1), SOST (Hs00228830_m1), RANKL (Hs00243519_m1), OPG (Hs00171068_m1), and GAPDH (Hs99999905_m1). Using a thermal cycler (ABI Prism Sequence Detection System 7700; Applied Biosystems), cycler conditions were 50°C for 2 minutes and 94°C for 2 minutes, followed by 40 cycles at 94°C for 15 seconds and 60°C for 30 seconds. Amplification curves were checked visually for exponentiality, and thresholds were set at 0.15 unit for RANKL, OPG, OSX, and RUNX2. The threshold for GAPDH was 0.03 unit, whereas the threshold for WNT, Osteocalcin, and SOST was set at 0.07 unit. All experiments were performed in triplicate and were normalized to the housekeeping gene GAPDH. GAPDH was chosen as the housekeeping gene because of its repeated use in human gene expression studies addressing osteoporosis research questions.23 The results were calculated applying the ΔΔCt method and are presented as fold increase relative to GAPDH expression.
Parametric data were reported as means ± SEM. For group comparisons of parametric data, t tests were calculated. Nonparametric data were reported as median values and the 25th to 75th percentile range. Nonparametric data were compared using Kruskal-Wallis tests (PASW 18.0 for Mac; SPSS, Inc., Chicago, IL, USA). For correlation analyses, Spearman coefficients were calculated. The critical value for data significance was set at p < .05.
Nine patients with the clinical diagnosis of male idiopathic osteoporosis (MIO) were included in the study. Their mean age was 53 years, ranging between 38 and 68 years. The mean age of control donors was 58 years (33 to 78 years, n = 9, p = 0.38). All MIO patients had a positive fracture history, including both adequate and inadequate traumas. Peripheral low-trauma fractures were present in two-thirds of patients. A single patient even reported four peripheral low-trauma fractures, whereas three others had two peripheral low-trauma fractures. The remaining third had prevalent vertebral fractures confirmed by the inclusion X-ray. The highest number of vertebral fractures in a single patient was three. In addition, 56% had previous fractures following adequate traumas (eg, from skiing or other sport-related accidents). Hip fractures were not prevalent in any participant. One-third reported a positive family history, defined as the occurrence of osteoporotic fractures in either of the parents. Patients with a positive family history had a 10.5% lower bone density at the lumbar spine (p = .025), but all other parameters investigated throughout this study were similar. One-third had never smoked; two-thirds were former (50%) or current smokers (50%). The mean body mass index (BMI) was 27 kg/m2 (19.7 to 33.3 kg/m2). In control patients, the autopsy protocols did not reveal major chronic diseases (eg, cancer). Owing to the nature of sample procurement (ie, forensic autopsies), a fracture history was not available.
Densitometry, serum analyses, and bone markers
The patients' mean T-scores for total hip, femoral neck, and lumbar spine (L1–L4) were −2.32 (±0.22), −2.33 (±0.22), and −2.5 (±0.19), respectively. Based on mere densitometry, 8 of 9 men met the International Society for Clinical Densitometry (ISCD) criteria for the diagnosis of male osteoporosis (Z-score < −2.0). According to World Health Organization (WHO) criteria (T-score < −2.5), two-thirds of our patients were osteoporotic; the rest were osteopenic, with a T-score of at least −2.1 at any of the sites. Mean 25-hydroxyvitamin D levels were relatively low (32.5 ng/mL), but serum calcium, serum phosphate, parathyroid hormone (PTH), total alkaline phosphatase, and urinary calcium excretion were normal in all patients. In a single patient, a minimal increase in phosphate excretion was found, but his serum phosphate level was normal. The biopsy of that patient revealed no evidence of osteomalacia. TSH, total testosterone, serum potassium and sodium, and electrophoresis of all patients were normal. Routine blood chemistry showed no evidence of kidney dysfunction or hepatic disease. Age-matched healthy males 52 years of age, on average, served as in-house controls for marker reference ranges. Metabolic parameters and bone markers are presented in Table 1.
Table 1. Biochemical Markers of Bone Metabolism in MIO Patients
Note: Data are presented as mean ± SEM.
Serum calcium (mmol/L)
2.27 (± 0.03)
Serum phosphate (mmol/L)
0.99 (± 0.04)
Total alkaline phosphatase (U/L)
79 (± 5)
25-(OH)-vitamin D (ng/mL)
32.5 (± 5.4)
44.6 (± 4.0)
24-Hour calcium excretion (mmol)
3.7 (± 0.6)
24-Hour phosphate excretion (mmol)
24.1 (± 4.5)
46.8 (± 4.9)
19.0 (± 1.6)
0.34 (± 0.4)
Histology and histomorphometry
Neither osteomalacia nor malignant infiltration was reported from routine readings. Bone surface (BS), bone volume (BV), and bone volume fraction (BV/TV) were reduced significantly in MIO patients. Osteoid volume (OV) also was lower. In MIO biopsies, significantly fewer osteoclasts were found. The marrow parameters (Ma.V and Ma.V/TV) were inversely related to BV/TV. Fat volume within the marrow (FatV/Ma.V) did not differ between the two groups. BS and OV were associated with the serum bone-formation marker P1NP (r = 0.733, p = .025; r = 0.794, p = .011). Confirming the absence of histomorphometric evidence of osteoporosis in the control group, we found similar BV/TV values in our autopsy samples and published data on healthy males.24, 25 Histomorphometric data are given in Table 2.
Table 2. Histomorphometric Parameters
Note: Parametric data are given as mean ± SEM; nonparametric data are represented as median and 25th to 75th percentile.
1.1 (± 0.1)
1.1 (± 0.1)
12.9 (± 1.5)
23.9 (± 1.8)
87.1 (± 1.5)
76.1 (± 1.8)
34.5 (± 3.1)
32.0 (± 4.2)
Micro–computed tomography (µCT)
Three-dimensional (3D) microstructural assessment showed a strong trend toward reduced trabecular bone volume (BV/TV) in MIO patients (−29%, p = .096). MIO patients revealed a significantly reduced number of trabeculae (Tb.N), a significantly increased trabecular separation (Tb.Sp), and a significantly reduced connectivity density (Conn.D) of the microstructure compared with nonosteoporotic controls (Fig. 1). Interestingly, we found increases in trabecular separation to be associated with elevated but still normal urinary phosphate excretion (r = 0.810, p = .015). Trabecular thickness (Tb.Th) was similar in both groups. The structure model index (SMI) showed a slight increase in MIO patients. Family history did not have any influence on bone microstructure. The results of µCT analyses are given in Table 3.
Table 3. Micro–Computed Tomography (µCT)
Data are presented as mean ± SEM or as median and the 25th to 75th percentile.
0.18 (± 0.04)
0.17 (± 0.03)
The expression of WNT10B, RUNX2, and RANKL was decreased significantly in MIO patients. Osterix, Osteocalcin, and OPG expression did not differ between the two groups (Fig. 2). MIO patients also had a reduced RANKL/OPG ratio (−62%, p = .007). Despite single outliers, the statistical range of the expression profile of all genes except Osteocalcin was smaller in MIO biopsies. Considering all biopsies, WNT10B expression correlated significantly with RUNX2 (r = 0.554, p = .017), RANKL (r = 0.617, p = .006), and local Osteocalcin (r = 0.509, p = .031), as well as BV/TV (r = 0.540, p = .025). In addition, RUNX2 was associated with Osterix (r = 0.499, p = .035), RANKL (r = 0.509, p = .031), and local Osteocalcin expression (r = 0.517, p = .028). Moreover, RUNX2 correlated with PTH serum levels (r = 0.683, p = .042). Positive correlations also were found for RANKL and BV/TV (r = 0.489, p = .046), trabecular number (r = 0.506, p = .046), and osteoid volume (r = 0.519, p = .033). However, RANKL was negatively associated with Tb.Sp (r = −0.576, p = .019). Osterix and local Osteocalcin expression correlated with osteoblast number (r = 0.594, p = .012; r = 0.709, p = .001). OPG was positively associated with age (r = 0.520, p = .027). SOST expression was significantly lower in MIO patients (−97%, p = .002). In all biopsies, SOST was strongly correlated with BV/TV (r = 0.807, p = .000) and inversely related to marrow volume (Ma.V/TV; r = 0.807, p = .000). Like trabecular microstructure, gene expression was independent of family history.
To the best of our knowledge, this is the first MIO biopsy study reporting reductions of WNT10B, RUNX2, and RANKL at the tissue level. Being aware of the largely unknown and presumably multifactorial etiology of MIO, we put major effort into recruiting a clinically homogeneous patient cohort. The double-biopsy approach enabled us to assess histomorphometric data and 3D microstructure in addition to gene expression analyses. When compared with other recent MIO studies,26, 27 our patient ages were comparable. Since all participants had a history of peripheral and/or vertebral low-trauma fractures, we are confident to have examined patients with a clinically relevant disease profile.
Other research groups used explant cell cultures from iliac crest biopsies of MIO patients. In line with our study, they found decreased osteoblastic DNA synthesis, impaired cell proliferation, and lower Osteocalcin gene expression on vitamin D stimulation.26, 27 Supporting the model of osteoblast dysfunction, Pernow and colleagues reported a decreased proliferative response to exogenous parathyroid hormone–related protein [PTHrP(1–34)] but also increased basal PTHrP expression in cultured MIO osteoblasts.26 Demonstrating decreased gene expression of WNT10B and RUNX2, our PCR data point to a potential disturbance in osteoblast differentiation.
Osteoblast lineage cells arise from mesenchymal stem cells (MSCs). In the presence of specific factors, including bone morphogenetic protein 2 (BMP-2), insulin-like growth factor 1 (IGF-1), 1,25-hydroxyvitamin D, or PTH, MSCs differentiate into preosteoblasts. Differentation into osteoblasts further requires Wnt pathway–dependent intracellular accumulation of β-catenin, as well as the expression of certain osteoblast-specific genes, namely, RUNX2 and Osterix.2Wnt10b transgenic mice display major increases in bone mass and mechanical competence, as well as a resistance to aging- and ovariectomy-induced bone loss.28 Since Wnt10b seems to enhance osteoblastogenesis and to inhibit osteoblast apoptosis, the osteoblastic transcription factors runx2, dlx5, and osterix were found to be upregulated in these transgenic mice. In reverse, Wnt10b−/− mice display substantial bone loss after only a few weeks of life.29 Somewhat surprisingly, prior to these changes, young Wnt10b−/− mice exhibit a rich trabecular microarchitecture.30 Stevens and colleagues concluded that this high-to-low bone quality sequence could provide evidence for an early depletion of the osteoprogenitor pool. Pointing at the potential importance of WNT10B in human medicine, genetic association analyses revealed a relation of WNT10B polymorphisms with hip bone mineral density (BMD) in Afro-Caribbean offspring.31 Although such associations were not detectable in a cohort of Spanish postmenopausal women,32 we found reduced WNT10B expression in iliac crest biopsies of MIO patients. The potential relevance of reduced WNT10B expression in MIO bone tissue is supported by a significant correlation with decreased local trabecular bone volume (BV/TV). Positive correlations of WNT10B with RUNX2, Osteocalcin, and RANKL indicate potential inhibitory downstream effects on osteoblastic transcriptional activity, but expression profiles have been studied in tissue samples and not pure human osteoblasts. Therefore, it would be of great interest to confirm the downregulation of RUNX2, Osteocalcin, and RANKL expression by low WNT10B via cell culture experiments.
RUNX2 and Osterix are key osteogenic transcription factors that are required for osteoblastic differentiation of MSCs.33 Owing to the entire absence of osteoblasts, Runx2−/− and Osx−/− mice are lacking bone from birth.34, 35 Leading to preosteoblast formation, RUNX2 is upregulated by BMP-2, IGF-1, and Wnt signaling.36, 37 Despite its osteoinductive role in preosteoblasts and immature cells, RUNX2 was shown to inhibit the late stage of osteoblast maturation.38, 39Osterix is considered to act downstream of RUNX2.35, 40 In our study, MIO patients exhibited significantly decreased RUNX2 levels that were correlated with low Osterix, Osteocalcin, and RANKL expression. However, Osterix only displayed a trend decrease in osteoporotic patients. It has been published that Wnt signaling and RUNX2 inhibit adipogenesis.41 Moreover, elderly osteopenic and osteoporotic patients seem to exhibit a shift toward increased fat in their marrow.42, 43 Despite these reports, we did not observe such marrow changes in our study. We speculate that this could be due to the relatively young age of MIO patients enrolled.
RANKL belongs to the tumor necrosis factor (TNF) ligand superfamily and is of pivotal importance for the generation, activity, and survival of osteoclasts.44, 45 Acting as a decoy receptor, its endogenous opponent, osteoprotegerin (OPG), blocks RANKL action and thereby balances bone turnover.46 In bone, RANKL is expressed most abundantly by MSCs, osteoblasts, and T cells.47 Since osteoblast-osteoclast coupling is exerted mainly by immature cells of the osteoblast lineage, RANKL expression is inversely related to osteoblast differentiation.48 Similarly, OPG is produced by MSCs and osteoblasts. Besides PTH and 1,25-hydroxyvitamin D, RANKL expression is stimulated by interleukin 6 (IL-6)–type ligands. Wnt signaling seems to have inhibitory effects on RANKL, but Wnts enhance OPG.49 RANKL and OPG are opposedly regulated by RUNX2, which matches the reductions in RANKL expression and osteoclast number that we found in MIO patients.50–52 Increased expression of RANKL and IL-6 was reported from female hip fracture patients,53, 54 suggesting major pathophysiologic differences between MIO and postmenopausal osteoporosis.
Sclerostin, the SOST gene product, is a strong inhibitor of osteoblastogenesis and enhancer of osteoblast apoptosis.55 Specific loss-of-function mutations are associated with Van Buchem disease and sclerosteosis, two human sclerosing bone disorders.56 Contrasting the literature, SOST expression was lower in patients with osteoporosis. Further, SOST correlated positively with trabecular number and inversely with trabecular separation. The strongest correlations were found with marrow volume, a parameter that inversely depends on bone volume. Since osteocytes are the main source of sclerostin, we assume that the reduction of SOST expression is due to reduced trabecular volume. Hypothetically, the SOST reduction in MIO also could indicate an autoregulatory rescue mechanism, but these assumptions were not further tested and thus remain speculative.
In our study, MIO patients displayed a much smaller overall range of gene expression than controls. These findings overlap with a study from Balla and colleagues investigating tissue-specific gene expression in elderly postmenopausal women. These authors analyzed the expression of 96 selected genes in femoral head bone samples obtained during hip arthroplasty and found a decreased expression pattern of almost all genes.57 The mean age of the study participants was about 70 years. At that stage, postmenopausal high turnover is mostly replaced by an age-related low-bone-turnover pattern. Likewise, Marie and colleagues reported decreased DNA synthesis in cultured MIO osteoblasts.13 These findings, together with our results, provoke speculation that low-turnover osteoporosis could be a disease set off, maintained, or at least partially expressed by insufficient transcriptional activity of osteoblasts.
The histomorphometric pattern that we found in the biopsies resembled that of previous MIO studies.24, 58, 59 Low trabecular bone volume (BV/TV) was accompanied by decreased amounts of osteoid. Osteoclast number was reduced, further suggesting low bone turnover. Since high-turnover patterns were found in hypercalciuric MIO patients,24 the low number of osteoclasts matches the normocalciuria of our patients. The number of osteoblasts was similar among the two groups. Osteoporotic alterations of bone microstructure seem to be gender-dependent. Like postmenopausal women, osteoporotic men were found to lose trabecular bone volume. Trabecular thinning that is reminiscent of glucocorticoid-induced osteoporosis is considered to predominate in men. Trabecular perforations are classified as a more female than male disease characteristic.60 Although age-related changes in men do not seem to involve major connectivity loss, males with vertebral fractures display this feature.61 Using µCT for the first time in a clinically homogeneous MIO cohort, we found significant decreases in trabecular number combined with increases in trabecular separation. Distinguishing the MIO pattern from age-related changes, we also observed major losses of trabecular connectivity. However, microstructural properties of MIO patients with vertebral fractures did not differ from those of patients with other fractures (data not shown). Trabecular thickness was unchanged, and plate-to-rod transitions were evident only as a nonsignificant trend.
Further interpreting our study from a clinical point of view, it is interesting that neither gene expression nor trabecular microstructure differed in patients with or without a positive family history. Not entirely matching the low-turnover pattern observed at the tissue level, mean serum CTX levels were close to the upper limit of the age- and sex-matched normal range. Providing additional evidence for potentially inadequate osteoclast-osteoblast coupling, P1NP and osteocalcin were centered within the normal range. However, when interpreting the marker results, it should be restated that age- and sex-matched but still individually different controls were used here rather than for the rest of the study. In accordance with the presumptive osteoblast pathology observed throughout our study, serum P1NP levels were associated with osteoid volume and bone surface, which per se were decreased.
A major strength of our study is the diligent definition of inclusion and exclusion criteria, which aimed at optimization of patient homogeneity. However, since MIO is considered to be a strongly heterogeneous entity, a homogeneous group of patients is difficult to define. Another limitation of our study was the decision for static histomorphometry. We opted against tetracycline labeling because it is unknown whether gene expression would have remained unbiased. Although our control data were stable, and Kulibawa and colleagues provided evidence for adequate quality of postmortem bone RNA, control sampling from forensic autopsies may be technically challenging.62 Nevertheless, we believe that performing biopsies in healthy individuals raises ethical concerns. Finally, it remains arguable whether the iliac crest, as a non-weight-bearing site, provides representative tissue information. Similar to site-specific characteristics in bone microstructure, gene expression patterns also may depend on location and bone compartment.
In conclusion, our data strongly support the hypothesis of osteoblast dysfunction in MIO. Regarding this potential pathomechanism, osteoanabolic drugs could provide a specifically promising therapeutic option for MIO patients with low bone turnover.
JMP has received speaker honoraria from Amgen. HR serves as a paid consultant for Eli Lilly, Amgen, Roche, Novartis, Nycomed, and Servier. He is a speaker for Merck (MSD), Lilly, Servier, Roche, and Nycomed and has received educational grants/research support from Lilly and Roche. CB is a paid consultant for Roche Diagnostics. PP has received research support and/or honoraria from Amgen GmbH, Eli Lilly GmbH, Leo Pharma, Merck, Sharp and Dohme GmbH, Novartis Pharma, Nycomed Pharma, Roche Austria, Servier Austria, and Sanofi-Aventis. CM is speaker for Eli Lilly, Amgen, Nycomed, Servier, Roche, Daichi Sankyo, Novartis, and Sanofi-Aventis. He has received research grants from Roche Austria and is member of the national advisory boards for Amgen Austria, Eli Lilly Austria, and Novartis Austria. TK is stock owner of B-cube AG and a member of the board of directors. All the other authors state that they have no conflicts of interest.
We are grateful to Prof Philippe Zysset, Dr Thomas Woegerbauer, Dr Peter Varga, Mag Julia Deutschmann, Mrs Katharina Wahl, Mrs Gerda Dinst, and Mrs Daniela Gabriel for medical, technical, and logistic support. This research was supported by the Austrian Federal Bank (Grant No. 12544 to PP) and educational grants from VINFORCE, Roche Diagnostics, Roche Austria, and Eli Lilly Austria.