An unusual transfer RNA (guanine-2-)-methyltransferase from transplantable rat mammary tumors.

We have previously demonstrated that (guanine-2-)-methyltransferase activity in extracts from 9,10-dimethyl-1,2-benzanthracene-induced rat mammary tumors differs from that of nonneoplastic mammary tissue. In this report, we explore further the nature of these differences by purification and characterization of the two major transfer RNA (tRNA) (guanine-2-)-methyltransferases from transplantable mammary tumors and proliferating mammary glands from pregnant rats. The position 10-specific (guanine-2-)-methyltransferases (2mGI) from proliferating rat mammary gland and mammary tumor were found to have similar properties with respect to molecular weight, substrate specificity, and elution behavior on ion-exchange columns. In addition, no tissue-specific differences were observed when the mammary tumor and mammary gland 2mGI activities were compared with those of purified rat liver enzyme. In contrast, the position 26-specific (guanine-2-)-methyltransferase (2mGII) from mammary tumors was seen to possess properties different from both the nontumorous mammary gland and liver enzyme. The tumor 2mGII activity showed unusual elution behavior on diethylaminoethyl-Sephadex, eluting along with the 2mGI activity. A small difference in molecular weight was detected between tumor and nontumorous 2mGII activities. Examination of the tumor enzyme in comparison with the well-characterized 2mGII from rat liver indicated that the mammary tumor 2mGII methylated a broader range of tRNA substrates. In particular, mature yeast phenylalanine-specific tRNA, which is methylated in vivo at all major eukaryotic methylation sites and should not be a substrate for eukaryotic methylating enzymes in vitro, could be methylated at low levels by the tumor enzyme. Two-dimensional electrophoretic fingerprint maps of T1 RNase-digested phenylalanine-specific tRNA from Escherichia coli methylated in vitro showed the presence of a methylated oligonucleotide which could not be correlated with normal sites of methylation on the tRNA. From these results, it appears that the mammary tumor 2mGII can methylate at some unusual site(s) on the tRNA molecule.

Methyl-accepting RNA in 13762 mammary adenocarcinoma correlated with low adenine methyltransferase levels.

Methylation reactions carried out with mammalian transfer RNA (tRNA) methyltransferases and RNA prepared from the homologous source do not normally show significant incorporation of methyl groups into the tRNA. However, our studies with the transplantable mammary adenocarcinoma 13762 indicate that tRNA from this tumor can be methylated in vitro with the homologous methyltransferases at a level 10 times higher than seen when tRNA from rat liver is reacted with its own enzyme. Analysis of the methylated nucleosides formed in vitro shows that greater than 80% of the methyl groups incorporated into 13762 RNA is found as 1-methyladenosine. Examination of the tRNA methyltransferase content of adenocarcinoma 13762 indicates that this tumor possesses unusually low levels of the adenine-1 methyltransferase responsible for methylating the invariant adenine at position 58 on tRNA. The nature of the methyl-accepting RNA from 13762 tumors has been examined using highly purified adenine-1 methyltransferase prepared from rat liver. Methylation of tumor RNA eluted from polyacrylamide gels after separation by electrophoresis indicated that while methyl-accepting material is found throughout the RNA-containing region of the gel, RNAs migrating slower than the bulk of mature tRNA are particularly good substrates for adenine methyltransferase. Similarly, when 13762 RNA is first methylated by the adenine-methylating enzyme and then run on acrylamide gels, several peaks of methyl-3H are seen in the region of slow-migrating tRNA. These results indicate that the 1-methyladenine deficiency in adenocarcinoma 13762 results in the appearance of selected populations of tRNA which are substrates in vitro for adenine-1 methyltransferase. The electrophoretic mobility of the methyl-accepting RNA in 13762 adenocarcinomas suggests that at least some of these may be precursor tRNAs.

tRNA methyltransferases from rat liver. Differences in response of partially purified enzymes to polyamines and inorganic salts.

Three tRNA methyltransferases, purified from rat liver, have been compared for their activity in the presence of various amines and Mg2+. The enzymes differ with respect to the ion which permits maximal activity; they also differ with respect to the concentration of a given ion necessary for maximal activity. The methyltransferase which forms N2-methylguanine in the region between the dihydrouridine loop and the acceptor stem (2mG I), when assayed using purified tRNA as substrate, shows high activity with 3--5 mM sperimidine or 20 mM putrescine and significantly lower rates of methylation with 200--350 mM ammonium acetate or 1--10 mM magnesium acetate. The enzyme responsible for forming N2-methylguanine between the dihydrouridine and anticodon loops (2mG II) works well in the presence of 0.2--0.5 mM spermidine, 10 mM putrescine or 200--300 mM ammonium acetate and shows slightly lower activity with 1 mM magnesium acetate. The optimal conditions for assaying 1-adenine methyltransferase (1mA) with purified tRNAs are either 200--300 mM ammonium acetate or 30 mM putrescine; spermidine is slightly less effective and magnesium acetate permits less than 25% of maximal activity. The addition of 10 mM Mg2+, in combination with polyamines or NH4+, depresses slightly the activity of the guanine methyltransferases but completely abolishes the polyamine or ammonium-stimulated activity of the adenine methyltransferase. When unfractionated (Escherichia coli) tRNA is used as substrate, the concentrations of polyamines required for optimal methyltransferase activity are increased but the patterns of response of the three enzymes do not differ significantly from those obtained with purified tRNA substrates. Based on the studies with these three enzymes, unfractionated tRNA and 40 mM putrescine should provide the most reliable system for detecting methylating activity if the nature of the tRNA methyltransferase is unknown.

Mature methyl-deficient tRNA isolated from a mammary adenocarcinoma.

A transplantable rat tumor, mammary adenocarcinoma 13762, accumulates tRNA which can be methylated in vitro by mammalian tRNA (adenine-1) methyltransferase. This unusual ability of the tumor RNA to serve as substrate for a homologous tRNA methylating enzyme is correlated with unusually low levels of the A 58-specific adenine-1 methyltransferase. The nature of the methyl-accepting RNA has been examined by separating tumor tRNA on two-dimensional polyacrylamide gels. Comparisons of ethidium bromide-stained gels of tumor vs. liver tRNA show no significant quantitative differences and no accumulation of novel tRNAs or precursor tRNAs in adenocarcinoma RNA. Two-dimensional separations of tumor RNA after in vitro [14C]methylation using purified adenine-1 methyltransferase indicate that about 25% of the tRNA species are strongly methyl-accepting RNAs. Identification of six of the tRNAs separated on two-dimensional gels has been carried out by hybridization of cloned tRNA genes to Northern blots. Three of these, tRNALys3 , tRNAGln and tRNAMeti , are among the adenocarcinoma methyl-accepting RNAs. The other three RNAs, all of which are leucine-specific tRNAs, show no methyl-accepting properties. Our results suggest that low levels of a tRNA methyltransferase in the adenocarcinoma cause selected species of tRNA to escape the normal A58 methylation, resulting in the appearance of several mature tRNAs which are deficient in 1-methyladenine. The methyl-accepting tRNAs from the tumor appear as ethidium bromide-stained spots of similar intensity to those seen for RNA from rat liver; therefore, methyladenine deficiency does not seem to impair processing of these tRNAs.

Purification and characterization of two tRNA-(guanine)-methyltransferases from rat liver.

tRNA(guanine-1-)-methyltransferase (EC 2.1.1.31) and tRNA(N2-guanine)-methyltransferase I (EC 2.1.1.32) were isolated from rat liver. The (guanine-1-)-methyltransferase preparation is 6800-fold purified and is free from contaminating methyltransferases or ribonuclease. The molecular weight of (guanine-1-)-methyltransferase is 83 000. Of seven purified Escherichia coli tRNAs examined, only tRNAMetf was utilized as substrate by (guanine-1-)-methyltransferase. The methylation of tRNAMetf is maximally stimulated by 40 mM putrescine with a pH optimum of 8.0. Using E. coli K-12 tRNA, the Km for S-adenosylmethionine is 3 micrometer and Ki for S-adenosylhomocysteine is 0.11 micrometer for (guanine-1-)-methyltransferase. (N2-Guanine-)-methyltransferase is 6200-fold purified and is also free of interfering enzymes. It has a molecular weight of 69 000. E. coli tRNAPhe, tRNAVal and tRNAArg are substrates for this enzyme which introduces a methyl at the 2-amino group of the guanine at position 10 from the 5'-terminus of these tRNAs. The methylation of tRNAPhe is maximally stimulated by 100 micrometer spermidine with a pH optimum of 8.0. (N2-Guanine-)-methyltransferase has a Km for S-adenosylmethionine of 2 micrometer and a Ki for S-adenosylhomocysteine of 23 micrometer with E. coli K-12 tRNA as methyl acceptor.

S-adenosylhomocysteine analogues as inhibitors of specific tRNA methylation.

Of 17 base- or amino acid-modified analogues of S-adenosylhomocysteine, six were found to produce at least 50% inhibition of the activity of an unfractionated tRNA methyltransferase extract at concentrations of 200 micron. The inhibitory effects of these six analogues on five purified rat liver tRNA methyltransferases were examined. The purified enzymes differed greatly in their sensitivity to the analogues. Ki values for the inhibitory analogues were determined for the three most highly purified methyltransferases. The kinetic analyses indicated that inhibition is competitive for nearly all enzyme/inhibitor combinations. The Ki values for good enzyme/inhibitor pairs were in the range of 0.11--2 micron. Each analogue appears to inhibit one methylation more strongly than others; e.g. the Ki values obtained for N6-methyl-S-adenosyl-L-homocysteine are approx. 0.4 micron for guanine-1 tRNA methyltransferase, 6 micron for adenine-1 tRNA methyltransferase and 100 micron for N2-guanine tRNA methyltransferase I. Structural features which are important for inhibitory activity are presence of a terminal amino group on the amino acid and the presence of adenosine rather than any other base. Ring nitrogens, a terminal carboxyl group and conformation at the asymmetric carbon appear to be important for some but not all of the tRNA methyltransferases examined.

Osteogenic effects of bioactive glass on bone marrow stromal cells.

Bioactive glass (BG) is an effective synthetic bone graft material. BG granules of narrow size range (300-355 mum) have the ability to form new bone tissue inside excavations produced by in vivo resorption. Previously, we demonstrated that BG stimulates the differentiation of cultured osteoblast precursors if the glass surface was biomimetically modified by the formation of bone-like apatite and adsorption of serum proteins. We now report that modified BG can also increase the rate at which multipotential rat bone marrow stromal cells (rMSC) will undergo osteogenesis. BG promoted rMSC osteogenesis both when cells were plated in contact with BG and when cells were not directly in contact with the BG. Alkaline phosphatase activity, a marker of bone cell differentiation, was used as an indicator for osteogenesis. Alkaline phosphatase activity of rMSCs exposed to osteoinducers such as ascorbate, dexamethasone, and BMP-2 was enhanced in the presence of BG. The stimulatory effect of BG was more pronounced in rMSC cultures with low basal alkaline phosphatase activity than in those with higher activity. The enhanced differentiation of rMSCs was associated with both a change in rMSC morphology and altered chemical composition of the cell culture media. rMSCs cultured on BG in the presence of BMP or dexamethasone exhibited a more rounded osteoblast-like appearance as compared with cells grown on tissue culture plastic. In the presence of BG, elevated levels of calcium and silicon in the culture medium were observed throughout the 7-day culture period, suggesting a continuous dissolution of surface-modified BG and resulting release of BG dissolution products. The data suggest that both surface- and solution-mediated events play a role in the osteogenic effect of BG.

A BMP responsive transcriptional region in the chicken type X collagen gene.

Bone morphogenetic proteins (BMPs) were originally identified by their ability to induce ectopic bone formation and have been shown to promote both chondrogenesis and chondrocyte hypertrophy. BMPs have recently been found to activate a membrane serine/threonine kinase signaling mechanism in a variety of cell types, but the downstream effectors of BMP signaling in chondrocyte differentiation remain unidentified. We have previously reported that BMP-2 markedly stimulates type X collagen expression in prehypertrophic chick sternal chondrocytes, and that type X collagen mRNA levels in chondrocytes cultured under serum-free (SF) conditions are elevated 3- to 5-fold within 24 h. To better define the molecular mechanisms of induction of chondrocyte hypertrophy by BMPs, we examined the effect of BMPs on type X collagen production by 15-day chick embryo sternal chondrocytes cultured under SF conditions in the presence or absence of 30 ng/ml BMP-2, BMP-4, or BMP-7. Two populations of chondrocytes were used: one representing resting cartilage isolated from the caudal third of the sterna and the second representing prehypertrophic cartilage from the cephalic third of the sterna. BMP-2, BMP-4, and BMP-7 all effectively promoted chondrocyte maturation of cephalic sternal chondrocytes as measured by high levels of alkaline phosphatase, diminished levels of type II collagen, and induction of the hypertrophic chondrocyte-specific marker, type X collagen. To test whether BMP control of type X collagen expression occurs at the transcriptional level, we utilized plasmid constructs containing the chicken collagen X promoter and 5' flanking regions fused to a reporter gene. Constructs were transiently transfected into sternal chondrocytes cultured under SF conditions in the presence or absence of 30 ng/ml BMP-2, BMP-4, or BMP-7. A 533 bp region located 2.4-2.9 kb upstream from the type X collagen transcriptional start site was both necessary and sufficient for strong BMP responsiveness in cells destined for hypertrophy, but not in chondrocytes derived from the lower sterna.

Utilization of bone morphogenetic protein receptors during chondrocyte maturation.

Cartilage from the upper, cephalic portion of embryonic chick sternums undergoes hypertrophy, while the lower, caudal portion of the sternum remains as cartilage. Bone morphogenetic proteins (BMPs) induce type X collagen (colX) in cultured upper but not lower sternal chondrocytes (LSCs). We have examined the utilization of BMP receptors (BMPRs) by upper sternal chondrocytes (USCs) and LSCs both by analyzing receptor expression and by overexpressing mutant BMPRs. Reverse-transcription polymerase chain reaction (RT-PCR) analyses indicate that both upper and lower chondrocytes produce messenger RNA (mRNA) for all three receptors: BMPR type IA (BMPR-IA), BMPR type IB (BMPR-IB), and BMPR type II (BMPR-II). Infection of USC with retroviral vectors expressing constitutively active (CA) BMPRs showed that CA-BMPR-IB, like exogenous BMP-4, induced both colX mRNA and elevated alkaline phosphatase (AP), while CA-BMPR-IA was markedly less potent. However, expression of activated receptors in LSC cultures resulted in only minimal induction of hypertrophic markers. Consistent with the results seen for CA receptors, dominant negative (DN) BMPR-IB blocked BMP-induced hypertrophy in USCs more effectively than DN-BMPR-IA. These results imply that the major BMPR required for BMP induction of chondrocyte hypertrophy is BMPR-IB, and that difference between permanent and prehypertrophic chondrocytes is not caused by absence of receptors required for BMP signaling.

The mechanism of beta-glycerophosphate action in mineralizing chick limb-bud mesenchymal cell cultures.

Differentiating chick limb-bud mesenchymal cells plated in micromass culture form a cartilage matrix that can be mineralized in the presence of 4 mM inorganic phosphate (Pi), and 1 mM calcium. Previous studies showed that when beta-glycerophosphate (beta GP) is used in place of Pi, the mineral crystals formed are larger and differ in distribution. The present study shows that the difference in distribution is not associated with alterations in cell proliferation, protein synthesis, or with collagen, proteoglycan core protein, or alkaline phosphatase gene expression. Cultures with 2.5, 5, and 10 mM beta GP did show different levels of alkaline phosphatase activity, and in the presence of low (0.3 mM) Ca had different Pi contents (4, 6 and 9 mM, respectively), indicating that the increase in CaxP product may in part be responsible for the altered pattern of mineralization. However, cultures with beta GP in which alkaline phosphatase activity was inhibited with levamisole still had an altered mineral distribution as revealed by Fourier transform-infrared (FT-IR) microspectroscopy. The presence of a casein kinase II-like activity in the mineralizing cultures, the ability of specific inhibitors of this enzyme to block mineralization, and the known ability of beta GP to block phosphoprotein phosphatase activity suggests that altered patterns of matrix protein phosphorylation may influence mineral deposition in these cultures.

Ascorbic acid regulates multiple metabolic activities of cartilage cells.

Bones grow in length because of the activities of cartilage cells in the epiphyseal growth plate. We have examined selected events that occur in the growth cartilage by the use of cultured epiphyseal cells; we have also evaluated the influence of ascorbate on these activities. Our studies indicate that 1) ascorbate induces the expression of a unique collagen isoform, type X collagen; 2) ascorbate stimulates alkaline phosphatase activity of maturing chondrocytes; and 3) ascorbate regulates the energy status of the maturing chondrocyte. We have found that in the presence of ascorbate there is a change in oxidative activity. Thus, lactate formation is inhibited, there is an increase in the adenylate energy charge ratio, and there is an elevation in the activity of isocitrate dehydrogenase. The results of these studies point to multiple effects of vitamin C on chondrocyte maturation involving changes in protein synthesis and energy metabolism.

Importance of 1,25-dihydroxyvitamin D3 and the nonadherent cells of marrow for osteoblast differentiation from rat marrow stromal cells.

Although steroid hormones regulate mature osteoblast function, much less is known about their actions on osteoprogenitor cells. The possibility of steroid hormone regulation of early stages in osteoblast differentiation was investigated by measuring the growth and induction of the osteoblast marker enzyme alkaline phosphatase (AP) in rat bone marrow stromal cell cultures. Experiments were performed in charcoal-stripped serum; conditions which markedly impaired stromal cell growth. However, growth could be stimulated by nonadherent marrow cell-derived conditioned medium. 1,25(OH)2D3, but not dexamethasone, 17 beta-estradiol, or retinoic acid, increased both stromal cell proliferation and AP activity. The increased proliferation with 1,25(OH)2D3 was nonadherent cell-dependent. BMP-2 also increased AP levels and acted in synergy with 1,25(OH)2D3. These results suggest that (i) nonadherent marrow cells may support stromal cell development, and (ii) 1,25(OH)2D3 as well as glucocorticoids may regulate osteogenesis from the bone marrow but a similar role for estrogen is not supported.

Hypertrophic chondrocytes. The terminal stage of differentiation in the chondrogenic cell lineage?

Chondrocytes emerging in the limb or other locations during embryogenesis are currently considered terminally differentiated cells and thus represent the last stage of differentiation in the chondrogenic cell lineage. Most chondrocytes, however, undergo further major phenotypic changes during late embryogenesis and early postnatal life as they take part in the endochondral ossification process. During this process, "resting" chondrocytes first enter an active, proliferative phase and then develop into large, round hypertrophic chondrocytes with unique phenotypic traits. The question thus arises as to whether hypertrophic chondrocytes actually represent the terminal stage of differentiation in the chondrogenic lineage. To assess the developmental position of these cells along the lineage, we examined the expression of four genes encoding extracellular matrix components in chondrocytes undergoing endochondral ossification in chicken tibial growth cartilage. We found that the steady-state levels of mRNAs coding for proteoglycan core protein increased in regions of cartilage destined for endochondral ossification. Similarly, type II collagen gene expression increased markedly in proliferating chondrocytes and then returned to basal levels in hypertrophic chondrocytes. As revealed by in situ hybridization, type X collagen gene expression was undetectable in resting and early proliferating chondrocytes and was detectable in hypertrophic chondrocytes. Osteonectin synthesis appeared to characterize chondrocytes in the resting, proliferating, and hypertrophic zones of growth cartilage. The protein was scarce, however, and cell-associated in the former zones, although it was very abundant and matrix-associated in the hypertrophic zone. Clearly, the emergence of hypertrophic chondrocytes during endochondral ossification is accompanied by marked quantitative and qualitative changes in gene expression. Interestingly, these changes occur during or immediately after the period of active chondrocyte proliferation. On the premises of the cell lineage definition proposed by Holtzer, the above data suggest that the hypertrophic chondrocytes represent the terminal stage of differentiation in the chondrogenic cell lineage.

Retinoic acid is a major regulator of chondrocyte maturation and matrix mineralization.

During the process of endochondral bone formation, chondrocytes undergo a series of complex maturational changes. Our recent studies indicate that this maturational process is influenced by the vitamin A derivative retinoic acid (RA). To learn how this agent regulates chondrocyte development, we characterized matrix gene expression during maturation of cartilage cells in chick sternum. RNAs were isolated from the cephalic portion of day 13, 14, 16, 18, and 20 chick embryo sternum and analyzed via northern blots. Type II collagen RNA levels remained fairly constant during this developmental period. In contrast, expression of type X collagen and alkaline phosphatase (APase) genes was first detected at day 16, followed by that of osteonectin (ON) and osteopontin (OP). To explore the mechanisms triggering these changes, chondrocytes were isolated from the cephalic portion of day 17-18 sternum (US cells) and grown in monolayer in standard serum-containing medium. After 3 weeks in culture, most of the cells enlarged and became type X collagen-positive, but they exhibited low APase activity and contained only trace amounts of ON and OP mRNAs. Treatment of parallel 3-week-old cultures with RA (10-100 nM) rapidly increased expression of the APase, ON, and OP genes severalfold. In concert with a significant increase in APase activity, there was abundant calcium accumulation in the RA-treated cultures. Electron microscopy confirmed the formation of large matrix-associated mineral crystals and the presence of numerous matrix vesicles. The effects of RA were also studied in cultures of immature chondrocytes isolated from the caudal portion of sternum (LS cells).(ABSTRACT TRUNCATED AT 250 WORDS)

Smad-Runx interactions during chondrocyte maturation.

BACKGROUND: Intracellular signaling triggered by bone morphogenetic proteins (BMPs) results in activated Smad complexes that regulate transcription of BMP-responsive genes. However, the low specificity of Smad binding to regulatory sequences implies that additional tissue-specific transcription factors are also needed. Runx2 (Cbfal) is a transcription factor required for bone formation. We have examined the role of Smads and Runx2 in BMP induction of type X collagen, which is a marker of chondrocyte hypertrophy leading to endochondral bone formation. METHODS: Pre-hypertrophic chondrocytes from the cephalic portion of the chick embryo sternum were placed in culture in the presence or absence of rhBMP-2. Cultures were transiently transfected with DNA containing the BMP-responsive type X collagen promoter upstream of the luciferase gene. The cultures were also transfected with plasmids, causing over-expression of Smads or Runx2, or both. After 24-48 hours, cell extracts were examined for levels of luciferase expression. RESULTS: In the presence of BMP-2, chondrocytes over-expressing BMP-activated Smadl or Smad5 showed significant enhancement of luciferase production compared with that seen with BMP alone. This enhancement was not observed with over-expression of Smad2, a transforming growth factor beta (TGF-beta)-activated Smad. Overexpression of Runx2 in BMP-treated cultures increased transcriptional activity to levels similar to those seen with Smads 1 or 5. When chondrocytes were simultaneously transfected with both Runx2 and Smad 1 or 5, promoter activity was further increased, indicating that BMP-stimulated Smad activity can be augmented by increasing the levels of Runx2. CONCLUSIONS: These results implicate the skeletal tissue transcription factor Runx2 in regulation of chondrocyte hypertrophy and suggest that maximal transcription of the type X collagen gene in pre-hypertrophic chondrocytes involves interaction of BMP-stimulated Smads with Runx2. CLINICAL RELEVANCE: Many skeletal abnormalities are associated with impaired regulation of chondrocyte hypertrophy in growth plates. These studies demonstrate that both BMP-activated Smads and Runx2 levels can modulate chondrocyte transition to hypertrophy.

Disparate osteogenic response of mandible and iliac crest bone marrow stromal cells to pamidronate.

OBJECTIVE: Long-term administration of intravenous bisphosphonates like pamidronate is associated with jaw osteonecrosis but axial and appendicular bones remain unaffected. Pathogenesis of bisphosphonate-associated jaw osteonecrosis may relate to skeletal site-specific effects of bisphosphonates on osteogenic differentiation of bone marrow stromal cells (BMSCs) of orofacial and axial/appendicular bones. This study evaluated and compared skeletal site-specific osteogenic response of mandible (orofacial bone) and iliac crest (axial bone) human BMSCs to pamidronate. MATERIALS AND METHODS: Mandible and iliac crest BMSCs from six normal healthy volunteers were established in culture and tested with pamidronate to evaluate and compare cell survival, osteogenic marker alkaline phosphatase, osteoclast differentiation in co-cultures with CD34+ hematopoietic stem cells, gene expression of receptor activator of NFkappaB ligand (RANKL) and osteoprotegerin, and in vivo bone regeneration. RESULTS: Mandible BMSCs were more susceptible to pamidronate than iliac crest BMSCs based on decreased cell survival, lower alkaline phosphatase production, and structurally less organized in vivo bone regeneration. Pamidronate promoted higher RANKL gene expression and osteoclast recruitment by mandible BMSCs. CONCLUSION: Mandible and iliac crest BMSC survival and osteogenic differentiation are disparately affected by pamidronate to favor dysregulated mandible bone homeostasis.