Endocrinology and Metabolism


X-linked hypophosphatemia (XLH) is caused by inactivating mutations of PEX, an endopeptidase of uncertain function. This defect is shared by Hyp mice, the murine homologue of the human disease, in which a 3′Pex deletion has been documented. In the present study, we report that immortalized osteoblasts derived from the simian virus 40 (SV40) transgenicHyp mouse (TMOb-Hyp) have an impaired capacity to mineralize extracellular matrix in vitro. Compared with immortalized osteoblasts from the SV40 transgenic normal mouse (TMOb-Nl), osteoblast cultures from the SV40 Hyp mouse exhibit diminished 45Ca accumulation into extracellular matrix (37 ± 6 vs. 1,484 ± 68 counts ⋅ min−1 ⋅ μg protein−1) and reduced formation of mineralization nodules. Moreover, in coculture experiments, we found evidence that osteoblasts from the SV40Hyp mouse produce a diffusible factor that blocks mineralization of extracellular matrix in normal osteoblasts. Our findings indicate that abnormal PEX in osteoblasts is associated with the accumulation of a factor(s) that inhibits mineralization of extracellular matrix in vitro.

  • X-linked phosphaturia
  • osteomalacia
  • osteocalcin

x-linked hypophosphatemia (XLH) is inherited as a dominant disorder and is characterized by hypophosphatemia, growth retardation, and rickets/osteomalacia (1, 16). The genetic defect underlying XLH rickets has been identified as mutations in the PEX gene product, or the phosphate-regulating genePex with homologies to endopeptidases on the X chromosome (1, 12, 14, 15, 33, 35). TheHyp mouse, a murine homologue of XLH, also has a loss of function Pexdeletion associated with renal phosphate wasting and defects in osteoblast-mediated mineralization (2, 31). This murine homologue provides a model to study the molecular and biochemical events linkingPex mutations to phosphaturia and impaired mineralization (16, 19, 31).

The physiological function of PEX is unknown. The presence of renal phosphate wasting secondary to a mutation of this gene suggests that this endopeptidase degrades a novel phosphaturic hormone [referred to as phosphatonin (1, 17)] or inactivates a phosphate-conserving factor (18-20). Given the broad constellation of phenotypic findings characteristic of XLH, it is also possible that PEX has other actions that are independent of its effects on renal phosphate transport, including regulation of bone mineralization. Although studies of primary osteoblast cultures derived from theHyp mouse have produced inconsistent results (3, 6, 10, 13), carefully performed studies suggest that osteoblast cultures derived from Hypmice do display mineralization abnormalities when transplanted into normal mice (10) and have alterations in osteoblast gene expression that are independent of hypophosphatemia (6, 13, 25, 32, 34). Moreover,Pex is expressed at high levels in osteoblasts, and its expression is temporally associated with the formation of mineralized extracellular matrix (ECM) in cultured osteoblasts (2, 9, 14). These observations suggest that bone is a physiologically relevant site of Pexexpression and that a potential relationship exists between mutations of Pex and aberrant osteoblast-mediated mineralization. Indeed,Pex may function in osteoblasts to metabolize endogenously or exogenously synthesized factors that regulate the process of osteoblast-mediated mineralization. Accordingly, osteoblast cell lines derived fromHyp mice should display a nascent defect in osteoblast-mediated mineralization, ifPex plays a role in mineralization that is independent of hypophosphatemia.

In the present investigation, we characterized the maturational profile of immortalized osteoblasts derived from SV40 transgenic normal andHyp mice, and we confirmed thatPex abnormalities are associated with osteoblast dysfunction and impaired mineralization in vitro. Moreover, we found that osteoblast cultures derived from theHyp mice produce a diffusible factor that inhibits normal mineralization in coculture experiments. Our studies support the hypothesis that abnormalities ofPex function inHyp mouse osteoblasts and that attendant accumulation of putative endogenously synthesized substrates of the gene product lead to impaired mineralization in XLH.



α-Minimum essential medium (α-MEM), DMEM/F-12, penicillin-streptomycin-amphotericin (antibiotic-antimycotic) solution, Hanks’ balanced salt solution (HBSS), and Trizol Reagent for single-step isolation of total RNA from cells were obtained from GIBCO (Grand Island, NY). Fetal bovine serum (FBS) was obtained from Hyclone Laboratories (Logan, UT). Pronase E, ascorbic acid, β-glycerophosphate, BSA,p-nitrophenol, diethanolamine, andp-nitrophenolphosphate used for alkaline phosphatase assay were purchased from Sigma (St. Louis, MO). [3H]thymidine,45CaCl2, and [α-32P]dCTP were purchased from Du Pont-NEN (Boston, MA). Bio-Rad reagent for protein assay was obtained from Bio-Rad Laboratories (Hercules, CA).

Isolation and culture of immortalized osteoblasts and clonal osteoblast cell lines from normal and Hyp mouse calvaria.

Mice were maintained and used in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals, prepared by the Institute on Laboratory Animal Resources, National Research Council (DHHS Publ. NIH 86–23, 1985), and by guidelines established by the Institutional Animal Care and Use Committee of Duke University. We established immortalized osteoblast cell lines from calvaria obtained from male normal and Hyp mice transgenic for the large T antigen of simian virus 40 (SV40).

These transgenic animals were created by mating C57BL6J males, heterozygous for the SV40 large T antigen, with femaleHyp mice, as previously described (21). Progeny containing the SV40 transgene were identified by PCR amplification of a ∼500-bp product of SV40 from individual genomic DNA, using forward primer 5′-CAGAGCAGAATTGTGGAGTGG-3′ and reverse primer 5′-GGACAAACCACAACTAGAATGCAGTG-3′. The normal andHyp mice littermates were distinguished among the SV40-positive mice on the basis of serum phosphorus values that were measured by colorimetric techniques on a Roche COBRAS MIRA-S.

We used a nonenzymatic method for obtaining the initial osteoblast cell lines (11). A fragment of the frontal and/or parietal bone from a single calvaria was aseptically removed from a 6- to 7-day-old mouse. Suture lines and endosteum were dissected away, and the bone fragment was placed in a culture dish. One or two metal strips were positioned on the endocranial surface and incubated for 3–4 days in DMEM/F-12 media containing 10% (vol/vol) FBS and 1% (vol/vol) antibiotic-antimycotic solution until the outgrowth of osteoblasts. The metal strips were removed, and the cells were allowed to grow until ∼60% confluent. The cells were scraped from the dish, transferred to a flask, and propagated by incubation in DMEM/F-12 containing 10% FBS and 50 μg/ml ascorbic acid in 5% CO2 at 37°C. We designated these cell lines as TMOb, for transgenic mouse osteoblasts. This technique produced both low and high alkaline phosphatase-producing cell lines. We selected for detailed analysis TMOb lines from normal and Hyp mice (TMOb-Nl and TMOb-Hyp, respectively) that displayed an osteoblast maturational sequence as evidenced by comparable postconfluent levels of alkaline phosphatase activity. These cell lines also maintained their osteoblast phenotype after repeated passages.

To control for potential cell heterogeneity of our immortalized cell lines, we also generated clonal osteoblast cell lines to create more homogeneous osteoblast cultures. For selection of clonal normal andHyp mouse osteoblasts, 100 cells from either normal or Hyp TMObs were plated on a 100-mm tissue culture dish and grown at low density to permit isolation of individual colonies. After 10 days of culture, colonies were isolated using trypsin-EDTA-saturated filter paper discs (2–3 mm) to lift individual colonies from the culture plates (8). After ∼3–5 min of trypsinization, the filter paper discs were removed and placed directly into individual wells of a 24-well culture dish and colonies were grown to near confluence. Clonal cells were then transferred to T-75 flasks and expanded for characterization.

We studied immortalized clonal osteoblasts rather than primary osteoblasts because the phenotype of clonal immortalized osteoblasts from Hyp mice is more likely to be due to intrinsic abnormalities related to thePex mutation than to theHyp mouse milieu, can be studied after multiple passages, and is more uniform and reproducible.

All stock cultures of TMObs were grown in α-MEM [containing 10% (vol/vol) FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml)] in a humidified atmosphere of 10% CO2-90% air at 37°C and were passed at a frequency sufficient to maintain subconfluence. Until the time of study, cells were subcultured every 3–5 days with 0.001% (wt/vol) pronase to achieve cell detachment. For studies characterizing the temporal sequence of osteoblast maturation, we plated either 40,000 cells into 35-mm-diameter multiwell dishes or 100,000 cells into 100-mm plates. TMOb-N1 and TMOb-Hyp cells were grown for periods of up to 14 days in α-MEM containing 10% FBS (vol/vol) supplemented with 5 mM β-glycerophosphate and 25 μg/ml of ascorbic acid, with media being replaced every 3 days.

Assay of cell replication.

We determined cell number at the various time points by direct counting with a hemocytometer. At the completion of the incubation period, cells were harvested by removing the media, washing twice with HBSS, and treating for 5 min with 0.25% trypsin-1 mM EDTA to achieve cell detachment. DNA synthesis was measured by determining TCA-precipitable radioactivity after a 3-h pulse with [3H]thymidine (1.5 μCi/ml), as previously described (24).

Alkaline phosphatase activity.

We analyzed alkaline phosphatase in cell layers by colorimetric assay of enzyme activity with the substratep-nitrophenolphosphate, as previously reported (24).

Mineralization assays.

The time course of mineralization was measured by radioactive calcium accumulation within the cell layer and matrix, as previously described (4, 24). Cells were incubated for 48 h in medium containing 0.5 μCi/ml of45CaCl2at the indicated times after seeding. Subsequently, the cell layers were harvested and digested in 0.1 N NaOH, and aliquots were counted by liquid scintillation spectroscopy or analyzed for total protein by the Bio-Rad protein assay (24).

ECM was isolated from TMOb-Nl and TMOb-Hyp osteoblasts after 14 days of culture in differentiation medium consisting of α-MEM containing 10% FBS, 25 μmol/ml ascorbate, and 5 mM β-glycerophosphate. Matrix was prepared by lysing cells with deuterated H2O and rinsing in 0.5% Triton X-100. After the ECM was extracted with 0.5 M EDTA and washed with water, samples of each preparation were taken for hydroxyproline (5,27).

The formation of in vitro mineralization nodules was determined by alizarin red-S histochemical staining (29). Cells were fixed for 24 h in 1:1:1.5 solution of 10% Formalin, methanol, and water; the fixative was removed; and the fixed cells and matrices were stained for 15 min with a 2% (wt/vol) solution of alizarin red-S at pH 4.0. The stained samples were washed three times with water and then air dried.

RT-PCR analysis.

We isolated total cellular RNA by a single-step method using Trizol reagent, as previously described (23). RNA samples were pretreated with DNase to remove any contaminating DNA and were quantified by absorbance at 260 nm. To identify Pex expression in TMObs, we performed RT-PCR using RNA derived from each cell line with the following primers: exon 1, M-5F (5′-TTCTGATGGAAGCAGAAACAGGGA-3′) and exon 8, M+930R (5′-GGGAATCATAGCGCTGAGTTCTGA-3′) to amplify the 5′ end of Pex; and exon 7, M+786F (5′-TAATAGCTCTCGAGCTGAACATGA-3′) and exon 20, M+1983R (5′-TATCCATTTCCTGTAAGCCC-3′) to amplify the 3′ end of Pex. To define thePex deletion break point, we used reverse primers to exon 15, M+1619R (5′-AAAGGCATTGACTGTTGTTG-3′) or to exon 16, M+1680R (5′-AAAGAAAGGCTTCTGCAGCT-3′) in combination with M+786F. One microgram of total RNA was reverse-transcribed into cDNA using the reverse primer. The RT reaction was incubated at 42°C for 1 h in 20 μl of 5 mM MgCl2, 1 × PCR buffer (Life Technologies, New York), 1 mM dNTP, 0.75 μM reverse primer, 20 units of RNase inhibitor, and 50 units of reverse transcriptase (Life Technologies). The conditions of PCR were 2 min at 94°C, followed by 38 cycles of 94°C for 1 min, 60°C for 1 min, 72°C for 1–2 min, and 72°C for 10 min for final extension. Samples without reverse transcriptase treatment were analyzed as controls. All predicted products were separated by agarose gel electrophoresis and stained with ethidium bromide. The 5′ and 3′ fragments were cloned into pCR 2.1 (Invitrogen) and confirmed as Pex by direct sequencing.

In addition, we performed RT-PCR to characterize osteoblast gene expression in clonal TMObs derived fromHyp and normal mice. We used the following primer sets to amplify osteopontin (mop-F 5′-ACACTTTCACTCCAATCGTCC-3′ andmop-R 5′-TGCCCTTTCCGTTGTTGTCC-3′), osteocalcin (moc-F 5′-CAAGTCCCACACAGCAGCTT-3′ andmoc-R 5′-AAAGCCGAGCTGCCAGAGTT-3′), and α1(I) procollagen (m±1-F 5′-TCTCCACTCTTCTAGTTCCT-3′ andm±1-R 5′-TTGGGTCATTTCCACATGC-3′). We used mouse β-actin primers (mActinF 5′-GTGGGCCGCTCTAGG CAC CA-3′, mActinR 5′-CGGTTGGCCTTA GGGTTCAGGGGG G-3′) to amplify a 245-bp fragment as a control for the amount and integrity of RNA in the PCR reactions. Gel-separated products were blotted on Nytran membrane (Schleicher & Schuell, Keene, NH) and immobilized on the membrane by ultraviolet (UV) cross-linking with a Stratalinker (Stratagene, La Jolla, CA). In some studies, the identity of the bands generated by PCR was confirmed by hybridization with radiolabeledPex and β-actin cDNA probes.

Northern blot hybridizations.

Northern analysis was carried out as described (14). Briefly, ≤20 μg of total RNA were electrophoresed on a 1.2% formaldehyde agarose gel and transferred to Nytran membrane (Schleicher & Schuell), and the RNA was immobilized on the membrane by UV cross-linking with a Stratalinker (Stratagene). The blot was hybridized overnight at 42°C in the prehybridization solution containing 10% dextran sulfate and 2 × 106counts ⋅ min−1 ⋅ ml−1of the random-labeled mouse osteocalcin and 28S probe. The blot was washed twice for 1 min at room temperature in a solution containing 2× standard sodium citrate (SSC) and 0.1% SDS, followed by washing two additional times for 15 min at 50°C in a solution containing 0.1× SSC and 0.1% SDS. The blot was air dried, and the bands were visualized by autoradiography.

Southern blotting.

For Southern blot analysis, genomic DNA (∼10 μg) from normal andHyp mouse osteoblasts was digested with EcoR I. The digested DNAs were electrophoresed on 0.7% agarose gel and blotted to nylon membranes (Schliecher & Schuell) by alkaline transfer. Hybridizations were generally performed in a hybridization buffer containing 1.5× SSPE (15 mM Na H2PO4, pH 7.4, 225 mM NaCl, and 1.5 mM EDTA), 1% SDS, and 10% dextran sulfate at 65°C overnight. A probe containing mousePex exons 7–22 was labeled by random hexamer priming, and washing was done in 0.1× SSC, 0.1% SDS at 65°C for 15 min.

Coculture experiments.

Coculture experiments were performed using a 6-well culture plate (Becton-Dickinson, Franklin Lakes, NJ) that contained a 10-cm2 lower plate well size and a 4.2-cm2 upper well insert that incorporated polyethylene terephthalate track-etched membrane (pore size 3 μm) to permit diffusion of soluble factors into a lower well. We plated TMOb-Nl and TMOb-Hyp cells in either the lower or upper well at an initial density of 40,000 cells per well to achieve coculture. Controls consisted of coculture of TMOb-Nl with TMOb-Nl and TMOb-Hyp with TMOb-Hyp cells. After 14 days in medium containing ascorbic acid and β-glycerophosphate as described above, mineralization was assessed by alizarin red-S staining and quantified by modification of previously described methods (30). Briefly, the stained matrix was washed with water and PBS, the dye was diluted with 10% (wt/vol) cetylpyridinium chloride, and the alizarin red-S was quantified at 562 nm.


We evaluated differences between groups by one-way analysis of variance (29). All values are expressed as means ± SE. All computations were performed using the Statgraphic statistical graphics system (STSC, Rockville, MD).


Characterization of the Pex mutation in Hyp mouse osteoblasts.

To confirm the presence of a Pexmutation in Hyp mice, we performed Southern blot analysis of genomic DNA using aPex cDNA probe (Fig.1 A). Consistent with previous observations, we identified a 3′ deletion of the Pex gene beyond exon 15 (31). Accordingly, in TMOb-Hypcells, RT-PCR amplification of the 3′ end ofPex, with primers designed to amplify the gene segment extending to exon 19, failed to produce an RNA product, whereas this region was amplified in TMOb-Nl cells (Fig.1 B, top). In contrast, using primers designed to amplify exons in the 5′ end ofPex, we identified the predicted-size band from normal as well as from Hypmouse osteoblasts. However, in four separate experiments,Pex was in lower abundance in the mutant cells (Fig. 1 B, middle). Additional RT-PCR studies with reverse primers to sequences in exons 15 and 16, in combination with an upstream 5′ primer (Fig.1 C), further defined the deletion break point. Consistent with a deletion break point between exons 15 and 16, primer pairs, including exon 15, amplified the predictedPex transcript, whereas no product was obtained using exon 16 primers in Hypmouse osteoblasts.

Fig. 1.

Expression of a phosphate-regulating gene with homologies to endopeptidases on the X chromosome (Pex) in normal andHyp mouse osteoblasts.A: Southern blot analysis of genomic DNA derived from normal (NL) and Hypmouse osteoblasts. Genomic DNA was digested byEcoR I and hybridized with32P-labeledPex cDNA probe, as described inmethods. A size calibration was obtained by simultaneous loading of λ-DNA/Hind III fragments and a 1-kb DNA ladder. Bands <4.0 kb [corresponding to exons 16–22 (40)] are absent in Hyp mouse, consistent with deletion of 3′ end ofPex gene.B: RT/PCR amplification ofPex transcripts using 5′ and 3′ specific primer pairs. To amplify 5′ and 3′ ends of Pex, respective primer pairs M-5F/ M+930R and M+786/M+1983R were used (seemethods) with total RNA derived from normal and Hyp mouse osteoblasts cultured for 10 days. Amplified products were transferred to nylon membranes and probed with a radiolabeledPex cDNA probe. As a control, β-actin was amplified and identified by hybridization. The 3′ end of Pex could not be amplified from RNA derived from Hyp mouse osteoblasts, and the 5′ end was present in low abundance. Both 5′ and 3′ ends of Pexcould be amplified in normal osteoblasts.C: RT-PCR detection of break points inHyp mutantPex transcripts. Using the forward primer M+786F in combination with exon 15 primer M+1619R and exon 16 primer M+1680R and RNA from normal andHyp mouse osteoblasts, we found evidence for a deletion beyond exon 15.

Phenotype characteristics of immortalized osteoblast cultures.

In subsequent studies, we examined whether the immortalized cells exhibited a temporal sequence of maturation characterized by an initial period of replication and subsequent postmitotic expression of osteoblastic characteristics. Similar to primary cultures (22) and other established cell lines (24), both normal andHyp mouse osteoblasts underwent an initial period of rapid cell proliferation that was characterized by increments in cell number (Fig.2 A) and high levels of DNA synthesis (Fig.2 C). Additionally, in both cell lines we observed a disproportionate increase in protein content relative to cell number after day 10of culture (Fig. 2 B), which corresponded to confluence of the cultures, a concordant decrement in the growth rate, and the formation of collagenous ECM (24). However, there was no significant difference between normal andHyp mouse osteoblasts with regard to parameters of cell growth and protein content.

Fig. 2.

Temporal changes in osteoblast phenotype during culture of normal andHyp mouse-derived osteoblasts. Cell no. (A), protein content (B), DNA synthesis (C), and alkaline phosphatase (ALP) activities (D) were assessed at different time points, as described inmethods. Cell nos. increased rapidly and progressively during initial period of culture, corresponding to a high rate of DNA synthesis. By day 10, replication rate diminished and DNA synthesis decreased, consistent with near growth arrest. Protein concentration also increased as a function of culture duration and showed a disproportionate increase between days 10 and14, consistent with formation of extracellular matrix (ECM). In addition, there is a time-dependent upregulation of ALP activity, consistent with expression of the osteoblast phenotype in postmitotic mature osteoblasts. Data are means ± SE of 3 separate experiments. Values sharing the same letter superscript are not significantly different atP = 0.001.

During the period of rapid cell growth, both TMOb-Nl and TMOb-Hyp cells expressed low levels of alkaline phosphatase (Fig. 2 D), consistent with their immature, preosteoblastic state. As anticipated, however, downregulation of replication was associated with a significant increase in the expression of alkaline phosphatase activity in both TMOb-Nl and TMOb-Hyp cells (Fig. 2 D), although by 14 days of culture, activity was slightly greater in normal cells. Similarly, the process of osteoblast maturation in the immortalized cells was marked by the absence of osteocalcin transcripts in 4-day-old cultures but with high levels of osteocalcin in 14-day-old cultures (data not shown). In concert, we found that Pexexpression increased in TMOb- Nl and TMOb-Hyp cells as a function of culture duration, a temporal increase corresponding to the osteoblast maturational stage (Fig. 3). Thus both immortalized osteoblast cell lines retain their capacity in vitro to undergo a normal temporal upregulation of osteoblast-related gene expression. Collectively, these observations indicate that the immortalized cells represent an excellent in vitro model system in which to study the bone mineralization defect in XLH.

Fig. 3.

Effect of osteoblast development onPex expression in normal andHyp mice osteoblasts. Total RNA was isolated at different time points, andPex transcripts were identified by RT-PCR with M-5F and M+930R to amplify a region extending from exon 1 to exon 8. β-Actin was amplified as a control. Transcripts were transferred to Nytran membrane, blotted with specific radiolabeled probes, and identified by autoradiography. Both transgenic mouse osteoblasts (TMOb)-Nl and TMOb-Hypexhibit a developmental stage-dependent upregulation ofPex transcripts. As in Fig. 1,Pex transcripts were lower inHyp than in normal mice osteoblasts.

Impaired mineralization in TMOb-Hyp osteoblast cultures.

In ensuing experiments, we assessed mineralization in normal andHyp mouse osteoblasts by use of45Ca incorporation and alizarin red-S histochemical staining. In immature TMOb-Nl cells, we observed the absence of mineralization (data not shown), whereas marked increments in 45Ca incorporation (Fig.4 A) that corresponded to the presence of alizarin red-S-stained mineralization nodules were observed in these cells by day 14 of culture (Fig.4 B). In contrast, matureHyp mouse osteoblasts exhibited significantly less 45Ca incorporation (Fig. 4 A) after 14 days of culture. Moreover, alizarin red-S staining revealed only ill-defined patches with limited dye uptake and the absence of discrete mineralization nodules (Fig. 4 B), consistent with impaired mineralization. The impaired mineralization was not related to differences in the amount of collagen produced in the normal and Hyp mice osteoblast cultures. In this regard, hydroxyproline content was similar between TMOb-Nl and TMOb-Hyp cell culture-derived matrix (0.125 ± 0.001 vs. 0.126 ± 0.001 mg/mg dry wt).

Fig. 4.

Mineralization defect in Hyp mouse osteoblasts. A:45Ca incorporation. Mineralization was quantified in normal and Hyposteoblasts by radioactive calcium accumulation within cell layer and ECM. Corresponding to developmental upregulation of osteoblast phenotype, there was an increase in ECM mineralization byday 14 of culture in TMOb-Nl but not in TMOb-Hyp. Values are means ± SE of 3 separate determinations. Values sharing the same letter superscript are not significantly different atP = 0.0001.B: in vitro mineralization nodule formation. Histochemical staining with alizarin red-S was performed to assess mineralized nodule formation, as described inmethods. Panel I depicts representative 100-mm plates derived from 14-day-old normal (NL) and Hyposteoblasts stained for mineralized nodules with alizarin red-S.Panel II is a high-powered (×100) magnification, showing well-defined mineralization nodules in TMOb-Nl and indiscrete foci of stain in TMOb-Hyp osteoblasts.

Persistence of defective mineralization in clonal osteoblasts derived from TMOb-Hyp cell cultures.

In addition, we showed that the impaired mineralization inHyp mouse-derived osteoblasts was not attributable to differences in cellular composition of the cultures, because clonal cell lines obtained from the parent TMOb cultures displayed identical results (Fig. 5). In this regard, clonal osteoblasts obtained from normal TMOb cultures exhibited maturation-dependent mineralization (Fig. 5,A andB) in association with increments in alkaline phosphatase activity (Fig.5 C) and normalPex expression (Fig.5 D). In contrast, clonal osteoblasts obtained from TMOb-Hyp cultures manifest impaired mineralization (Fig. 5,A andB) in association with the 3′Pex deletion (Fig.5 D) and significantly greater alkaline phosphatase activity compared with normal clonal osteoblasts (Fig. 5 C). Moreover, we could identify no differences in osteopontin, osteocalcin, and type I collagen mRNA expression between clonal osteoblasts derived fromHyp and normal mice (Fig.5 D). These findings suggest that a nascent defect in osteoblast-mediated mineralization is a characteristic of osteoblasts with thePex deletion.

Fig. 5.

Characterization of clonal osteoblast cell lines derived from TMOb cultures. Osteoblasts were subcloned from parent TMObs, as described inmethods.A: histochemical staining of mineralization nodules. Normal and Hypclonal osteoblast cell lines were cultured for 14 days and stained with alizarin red-S. Similar to corresponding parent cell lines (Fig. 4), clonal Hyp osteoblasts failed to form mineralization nodules, whereas clonal osteoblasts from normal TMObs formed abundant mineralized nodules.B: quantification of mineralization. Alizarin red-S stain was extracted with 10% cetylpyridinium chloride and quantified as described inmethods. ClonalHyp osteoblasts had significantly lower alizarin red-S accumulation at day 14 of culture compared with clonal normal osteoblasts.C: ALP activity. ClonalHyp osteoblasts displayed degrees of osteoblast maturation similar to normal osteoblasts. BothHyp and normal cell lines displayed a culture duration-dependent increase in ALP activity.D: mRNA phenotype analysis of clonal osteoblasts. Mouse specific primers were used to RT-PCR amplify osteopontin (OP), osteocalcin (OC), and α1(I) procollagen from clonal osteoblasts derived from TMOb-Nl and TMOb-Hyp parent cell lines. Predicted-size products for OP (239 bp), OC (370 bp), and α1(I) procollagen (268 bp) were expressed at similar abundance in normal andHyp clonal osteoblasts, whereas the 3′ end of Pex was found only in normal osteoblasts, consistent with our findings in parent cell lines (Fig. 1). β-Actin served as a control for relative mRNA abundance. Numeric values represent means ± SE of 3 separate determinations. Values sharing the same letter superscript are not significantly different at P < 0.01.

Transfer of the Hyp mouse phenotype in coculture experiments between TMOb-Nl and TMOb-Hyp.

To examine whether the abnormal mineralization in TMOb-Hyp cells is due to production of a factor(s) that inhibits mineralization, we cocultured TMOb-Hyp and TMOb-Nl cell lines separated by a semipermeable membrane. TMOb-Hyp cells displayed abnormal mineralization, whether cocultured with TMOb-Nl or TMOb-Hyp cells (Fig.6 B). In contrast, coculture of TMOb-Hyp with TMOb-Nl cells inhibited the mineralization of the normal osteoblasts, as evidenced by a failure to form discrete mineralization nodules (Fig.6 A) and significant reductions in alizarin red-S staining (Fig. 6 B). Identical results were obtained in three replicative studies, consistent with the production of factors capable of inhibiting normal mineralization by TMOb-Hyp cells.

Fig. 6.

Coculture experiments comparing normal andHyp mouse osteoblasts. TMOb-Nl and TMOb-Hyp osteoblasts separated by a semipermeable membrane were cocultured for 14 days.A: histochemical staining of mineralization nodules with alizarin red-S. Normal osteoblasts cocultured with normal osteoblasts formed mineralized nodules, and normal osteoblasts cocultured with Hypmouse osteoblasts did not mineralize.B: measurement of alizarin red-S staining, which was extracted with 10% cetylpyridinium chloride and quantified as described in methods. TMOb-Hyp osteoblasts had lower alizarin red-S accumulation than TMOb-Nl, consistent with impaired mineralization. TMOb-Hyp, however, inhibited mineralization of cocultured normal osteoblasts, consistent with the production/accumulation of a soluble inhibitor of mineralization. Values are means ± SE of 3 separate determinations. Values sharing the same letter superscript are not significantly different at P < 0.01.


The bone mineralization defect in XLH may be due to inadequate circulating levels of mineral and/or hormonal/metabolic factors that influence osteoblast function or to nascent defects in osteoblast function that impair the mineralization process. Our studies indicate that the abnormal mineralization inHyp mice is due, at least in part, to an intrinsic osteoblastic defect associated with abnormalPex function. In this regard, we found that TMOb-Hyp cells manifest a 3′ Pex deletion (Fig. 1) and, in a setting remote from the in vivo Hypmouse environment, fail to mineralize under culture conditions supporting mineralization in normal osteoblasts (Figs. 4 and 5). More importantly, we found that the Hypmouse osteoblasts produce a factor(s) that is capable of regulating the mineralization of ECM. To this end, the mineralization defect observed in TMOb-Hyp cell lines is transferable to normal osteoblasts in coculture experiments (Fig. 6). Such production of a mineralization inhibitor clearly represents a nascent defect in the osteoblasts from Hypmice.

Because a physiologically relevant site of PEX expression is the osteoblast, it appears likely that production of this mineralization inhibitor is the result of the primary genetic abnormality underlying XLH, namely inactivating mutations of PEX. Indeed, dysfunction of the gene product may result in failure to degrade an endogenously synthesized but undefined inhibitor of mineralization that is a substrate of Pex. The alternate possibility, that Pex fails to activate a novel mineralization-promoting factor, is inconsistent with our coculture experiments in which theHyp phenotype predominates (Fig. 6). In any case, further studies are necessary to identify the putativePex substrates produced by osteoblasts and to determine their relationship to the osteoblast-synthesized factor(s). In these investigations, efforts to discriminate whether the mineralization inhibitor represents phosphatonin (17) or an additional putative PEX substrate will be essential.

Although Pex substrates appear to be present in osteoblasts expressing the 3′Pex deletion, the mechanism whereby the accumulated Pex substrate causes the mineralization defect remains unknown. The impaired mineralization might be a direct consequence of a Pexsubstrate or might result from a multistepped cascade linking thePex mutation and the accumulation of its substrate with impaired mineralization. Several observations suggest that a downstream event, rather than the putativePex substrate, may be the mineralization inhibitor. In this regard, provided thatPex in the normal cells is not saturated and is located extracellularly (issues that require confirmation), the Pex endopeptidase in the normal cocultured cells should degrade any diffusible substrates, precluding a negative effect on mineralization. Given the results of our coculture experiments (Fig. 6), it is more likely that impaired mineralization results from a downstream kinase cascade that is regulated by the Pex substrate. Consistent with this possibility, additional studies have identified reductions in casein kinase and decreased phosphorylation of matrix proteins in Hyp mouse osteoblasts (16,25).

The possible coproduction of Pex and its substrate in osteoblasts is supported by several studies in which an endopeptidase and its substrate are found in the same cell (35). However, when the identity of the substrate is determined, in situ and immunohistochemical studies will be necessary to establish its precise cellular localization. In any event, our study establishes thatPex effects on bone are likely mediated by its metabolism of local factors derived from cells that are within the osteoblast lineage or coisolated with osteoblasts from calvaria.

The current investigations also clarify the nature of thePex mutation inHyp mice. We found that TMObs derived from Hyp mice have a 3′ deletion of Pex (Fig. 1). Similar to prior Southern analysis of genomic DNA (31), we identified the absence of bands corresponding to the 3′ end of thePex gene inHyp mice (Fig.1 A) and localized the site of the deletion between exons 15 and 16 by RT-PCR (Fig.1 C). This deletion predicts the production of a protein lacking a portion of the extracellular domain containing the putative catalytic sites; consequently, this is likely to result in loss of Pex function. We were unable to identify the putative intronic sequence or retained 3′ end of the Pex transcript in TMOb-Hyp cells, as reported by Beck et al. (2). The reason for this apparent discrepancy is not clear but could be due to differences related to PCR conditions, lower abundance of the truncated message, and/or differences related to amplification from contaminating genomic DNA. Regardless, we found thatPex expressed a truncated 5′ transcript, albeit at lower levels compared with normal TMOb cells (Fig. 1 B). Lower levels ofPex expression inHyp mice osteoblasts suggest that the 3′ deletion may result in additional abnormalities of message stability. The possibility that message instability may also be clinically relevant is supported by the recent identification in certain families with XLH of mutations in the 5′- and 3′-untranslated regions of PEX that may be important in stabilizing messenger RNA (7).

Many questions remain regarding the pathogenesis of XLH, despite the identification of the PEX/Pex gene. Our results add to the growing body of evidence supporting the concept that osteoblastic cells are a physiologically relevant site ofPex expression and have significant implications regarding our understanding of the pathogenesis of the mineralization defect in XLH and Hypmice. Further studies will be needed to determine the specific molecular abnormalities of ECM that are responsible for the impaired mineralization and whether these abnormalities are due to the accumulation of a Pex substrate itself or the downstream consequence of thePex substrate. Our cell culture system also will permit molecular targeting and direct manipulation ofPex expression to prove a cause-and-effect relationship betweenPex and the osteoblast phenotype. In turn, unraveling the pathogenesis of XLH and the function ofPex in osteoblasts may provide insights into novel factors that regulate bone mineralization.


We thank Suzanne Ellett for secretarial assistance in the preparation of this manuscript.


  • Address for reprint requests: L. D. Quarles, Dept. of Medicine, PO Box 3036 DUMC, Durham, NC 27710.

  • This work was supported in part by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grants RO1-AR-37308 and RO1-AR-43468 (to L. D. Quarles) and R01-AR-27032 (to M. K. Drezner).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract