Intermittent treatment with parathyroid hormone (PTH) as well as a non-peptide small molecule agonist of the PTH1 receptor inhibits adipocyte differentiation in human bone marrow stromal cells

Abstract

Whereas continuous PTH infusion increases bone resorption and bone loss, intermittent PTH treatment stimulates bone formation, in part, via reactivation of quiescent bone surfaces and reducing osteoblast apoptosis. We investigated the possibility that intermittent and continuous PTH treatment also differentially regulates osteogenic and adipocytic lineage commitment of bone marrow stromal progenitor/mesenchymal stem cells (MSC). The MSC were cultured under mildly adipogenic conditions in medium supplemented with dexamethasone, insulin, isobutyl- methylxanthine and troglitazone (DIIT), and treated with 50 nM human PTH(1–34) for either 1 h/day or continuously (PTH replenished every 48 h). After 6 days, cells treated with PTH for 1 h/day retained their normal fibroblastic appearance whereas those treated continuously adopted a polygonal, irregular morphology. After 12–18 days numerous lipid vacuole and oil red O-positive adipocytes had developed in cultures treated with DIIT alone, or with DIIT and continuous PTH. In contrast, adipocyte number was reduced and alkaline phosphatase staining increased in the cultures treated with DIIT and 1 h/day PTH, indicating suppression of adipogenesis and possible promotion of early osteoblastic differentiation. Furthermore, intermittent but not continuous PTH treatment suppressed markers of differentiated adipocytes such as mRNA expression of lipoprotein lipase and PPARγ as well as glycerol 3-phosphate dehydrogenase activity. All of these effects of intermittent PTH were also produced by a 1 h/day treatment with AH3960 (30 μM), a small molecule, non-peptide agonist of the PTH1 receptor. AH3960, like PTH, activates both the cAMP and calcium signaling pathways. Treatment with the adenylyl cyclase activator forskolin for 1 h/day, mimicked the anti-adipogenic effect of intermittent PTH, whereas pretreatment with the protein kinase-A inhibitor H89 prior to intermittent PTH resulted in almost complete conversion to adipocytes. In contrast, the MAP kinase inhibitor PD 98059 failed to prevent the anti-adipocytic effect of intermittent PTH, suggesting that the inhibitory effect of PTH on adipocyte differentiation is predominantly cAMP-dependent. These results demonstrate a differential effect of PTH1 receptor agonists on the adipocytic commitment and differentiation of adult human bone marrow mesenchymal stem cells. This response may represent an additional mechanism that contributes to the overall bone anabolic action of intermittent PTH.

Keywords: Osteoporosis; Osteoprogenitor; Parathyroid hormone; Adipocyte; Bone marrow

Introduction

Investigations in both laboratory animals and humans have demonstrated that whereas pulsatile PTH treatment leads to a net stimulation of bone formation and bone mass at both trabecular and cortical sites, continuous PTH infusion prefer- entially increases net bone resorption resulting in bone loss
[1,2]. Indeed, intermittent PTH treatment of postmenopausal women with osteoporosis over a 1 to 2 year period, resulted in increased bone density at multiple sites and a significant reduction in both vertebral and non-vertebral fracture risk [3]. The exact mechanisms contributing to the differential bone response to intermittent versus continuous PTH exposure are complex and remain incompletely understood. The resorptive effect of continuous PTH is predominantly due to stimulation of osteoclast differentiation and activity via effects on osteoblast cytokine production. In particular, continuous PTH treatment results in a dramatic and sustained increase in the RANKL:OPG ratio, as shown in both rat bone in vivo [4] and in bone marrow stromal cells in vitro [5], a response not observed with intermittent PTH. The anabolic action of intermittent PTH treatment appears to result from effects on the differentiation, activity and/or survival of osteoblast lineage cells. First, intermittent PTH has been demonstrated to cause an immediate increase in mineral apposition rate in ovariectomized rats due to increased osteoblast activity and activation of bone lining cells [6,7]. Second, daily PTH administration may enhance the proliferation and differentiation of osteoprogenitor cells within bone marrow, an effect that has been observed in rats [8,9] but not in mice [10]. Third, intermittent PTH may enlarge the pool of active osteoblasts to effect a more sustained increase in bone formation by suppressing osteoblast and osteocyte apoptosis as observed in mice administered PTH by daily injection [11] but not by continuous infusion [12]. However, the importance of apoptosis to the bone anabolic response to intermittent PTH remains unclear because in rats daily PTH actually transiently increased apoptosis of osteoblast lineage cells [13], and in vitro studies suggest that the apoptotic response to the hormone depends on the stage of cell differentiation [14].

In adults, trabecular osteoblasts are derived from mesench- ymal stem cells (MSC, also known as bone marrow stromal cells) within the bone marrow. These cells are able to differentiate into osteoblasts and adipocytes, as well as other mesenchymal lineages (see [15] for a review). Evidence from studies in vitro [16–18] and in vivo [19–21] supports an inverse relationship between the extent of osteogenic and adipogenic differentiation by MSCs, and this may be responsible for the age and osteoporosis-associated decrease in trabecular bone volume and increase in bone marrow adiposity [22–24]. This has lead to the search for potential treatments of osteopenic disorders aimed at inhibiting bone marrow adipogenesis as well as by the stimulation of osteogenesis [25].

Interestingly, activation of the PTH1 receptor has been demonstrated to inhibit adipogenesis both in vitro and in vivo. Over-expression of PTH-related peptide (PTHrP) or prolonged PTHrP treatment blocks adipocyte differentiation in 3T3-L1 preadipocyte and C3H10T1/2 pluripotential mesenchymal cell lines [26,27]. Moreover, haploinsufficient PTHrP+/− mice develop low trabecular bone density and increased bone marrow adiposity, despite normal circulating PTH levels [28], and in monkeys daily PTH treatment reversed the effects of ovariectomy on the cellular composition of bone marrow with a trend towards reduced adipocyte and increased osteoblast numbers [29]. Therefore, we reasoned that an additional mechanism whereby PTH stimulates bone formation could be via an inhibition of adipocyte commitment by bone marrow osteoprogenitor cells, and that this effect may vary depending upon the mode of PTH administration. Here we demonstrate that intermittent but not continuous exposure of human MSC cultures to either PTH(1–34) or a small molecule (non-peptide) agonist of the PTH1 receptor, inhibited the expression of adipocyte phenotypic markers and adipocyte formation, while simultaneously stimulating the osteoblast-associated marker alkaline phosphatase. In addition, the anti-adipogenic effect of intermittent treatment with PTH1 agonists was shown to be cAMP-dependent.

Methods

Materials

Human PTH(1–34) was obtained from Bachem (Bachem California Inc., Torrance, CA). A two-site immunoradiometric assay (IRMA) kit for rat PTH was purchased from Immutopics (San Clemente, CA). The assay cross-reacts 100% with human PTH. Fluo-3-acetoxymethyl ester fluorescent indicator dye (Fluo-3AM) was obtained from Molecular Probes (Eugene, OR). Troglitazone was purchased from BioMol (Plymouth Meeting, PA). Glycerol-3-phosphate dehydrogenase and NADH, used in the G3PDH assay, were from Roche Applied Science (Indianapolis, IN). All other reagents were from Sigma-Aldrich Corp. (St. Louis, MO) unless otherwise stated. The small molecule, non-peptide PTH1 receptor agonist AH3960 was identified from a high-throughput compound screen conducted at GlaxoSmithKline. Stock solutions of 30 mM AH3960 were prepared in DMSO.

Culture of HEK293-PTH1R cells

HEK293-PTH1R clone 4 cells were derived from the HEK293 human embryonic kidney cell line by stable transfection with the wild-type human PTH1 receptor. Cells were cultured in EMEM medium supplemented with 2 mM L-glutamine, 10%(v/v) fetal bovine serum (FBS) and 400 μg/mL G418.

CRE reporter assay

Cyclic AMP responses to PTH1 receptor agonists, hPTH(1–34) and AH3960, were measured using a cAMP response element (CRE) reporter assay. The HEK293-PTH1R clone 4 cells cultured in 75 cm2 culture flasks were transiently transfected with 15 μg CRE-luciferase reporter plasmid (pMREVI- P2LUCIFHYGRO8) using the Lipofectamine 2000 reagent complexed in Opti- MEM1 medium (Invitrogen, Carlsbad, CA). On day 2 the transfected cells were harvested in medium containing 50 μM isobutylmethylxanthine (IBMX) and seeded into 384-well white plates at 10,000 cells/well in 20 μL medium. After overnight culture the cells were incubated with PTH agonists in 10 μL of phenol red-free DMEM/Hams-F12 medium added using a CyBioWell liquid handler (CyBio AG). Following 5-h incubation at 37°C, 10 μL/well Steady-Glo luminescent reagent (Promega Corp., Madison, WI) was added and lumines- cence measured after 10 min with a ViewLux luminometer (PerkinElmer Life Sciences Inc., Boston, MA).

Measurement of intracellular calcium using FLIPR

Changes in the intracellular Ca2+ ion concentration in response to PTH agonists was determined by the fluorescence imaging plate reader (FLIPR) assay. The HEK293-PTH1R clone 4 cells were seeded at 15,000 cells/well in poly-D-lysine coated 384-well black-walled/clear bottomed plates in growth medium. After 48 h, the medium was removed and cells loaded for 1 h at 37°C with Fluo-3AM calcium dye (4 μM in EMEM, 0.1% RIA grade BSA). After washing twice with 1× Krebs assay buffer (0.12 M NaCl, 4.6 mM KCl, 1.03 mM KH2PO4, 25 mM NaHCO3, 1 mM CaCl2, 1.1 mM MgCl2, 11 mM glucose,2.5 mM probenecid, and 20 mM HEPES, pH 7.4) the cells were incubated in buffer (50 μL/well) for 20–30 min at 37°C before stimulation with a complete concentration range of hPTH(1–34) or AH3960. The fluorescence signal was recorded in a CybioWell FLIPR384 instrument over 5 min.

Culture of human mesenchymal stem cells

Human mesenchymal stem cells (bone marrow stromal cells) were purchased from Cambrex Bio Science (Walkersville, MD), and expanded to passage 3 in mesenchymal stem cell basal medium (MSCBM) containing growth supplements (mesenchymal cell growth supplement [MCGS], gluta- mine, penicillin and streptomycin) according to the supplier’s instructions.

For experiments, cells were plated at 5,000 cells/cm2 in adipocyte maintenance medium (AMM) with supplements (MCGS, glutamine, P/S). At confluence, adipocyte differentiation was mildly induced by the addition of 10 nM dexamethasone, 1 μg/mL recombinant human insulin, 10 μg/mL IBMX and 0.5 μM of the PPARγ agonist troglitazone (abbreviated to AMM+DIIT). The culture medium was replaced every 2 days. Culture was continued for 12– 18 days as specified in the Figure legends. For the intermittent PTH treatment, cells were incubated for 1 h with hPTH(1–34) (5 to 50 nM) in AMM+DIIT medium, rinsed briefly with PBS and culture continued in AMM+DIIT. For continuous PTH treatment, hPTH(1–34) at the same concentration was added every 2 days at the time of the medium change. A 1 h treatment was chosen for intermittent stimulation because in the rat (i) PTH(1–34) infusion for 1 h/day stimulated bone formation whereas infusions longer than 2 h induced detrimental bone effects [30], and (ii) pharmacokinetic analysis has shown this to be the duration over which the serum concentration of PTH(1–34) remains above baseline endogenous PTH levels following subcutaneous injection of an anabolic regimen [31].

Histochemical staining

Cell layers were fixed overnight at room temperature with 10% neutral buffered formalin. For alkaline phosphatase enzyme, cells were rinsed 2× with PBS and then stained for 15–20 min with diazonium salt of naphthol AS-MX phosphate and fast blue BB in 0.1 M Tris–HCl, pH 9.2. Cytoplasmic lipid droplets were stained with 1.8%(w/v) Oil Red O in 60%(v/v) isopropanol for 15 min.

RNA extraction, cDNA synthesis and reverse transcriptase PCR

Isolation of RNA, reverse transcription and PCR were performed as previously described [32]. Briefly, total RNA was isolated from cells cultured in 10 cm plates using the phenol–guanidine isothiocyanate method, and purified RNA digested with RNase-free DNase1 prior to the reverse transcription of 4 μg aliquots with AMV reverse transcriptase in the presence of oligo-dT and random hexamer primers. Aliquots (1/25th) of cDNA were then amplified in duplicate in ‘hot start’ PCR with Taq Start antibody according to the manufacturer’s instructions (Clontech, Palo Alto, CA) using sequence specific primers for the following human transcripts: lipoprotein lipase (LPL), PPARγ2, glycerol 3- phosphate dehydrogenase (G3PDH), osteocalcin (OC), PTH1 receptor, Runx2 and GAPDH. The primer sequences and amplification product sizes are given in Table 1. Reactions were terminated during the linear amplification phase (24 cycles for GAPDH, and between 30 to 35 cycles for all other genes). The PCR products were visualized by electrophoresis in 1.5%(w/v) agarose gels stained with ethidium bromide.

Assay for glycerol-3-phosphate dehydrogenase activity

Cells were cultured in 10 cm plates as described above for 18 days post- confluency, at which time cell layers were lysed in an ice cold solution of 50 mM Tris (pH 7.5), 1 mM EDTA and 0.5 mM DTT and sonicated. G3PDH enzyme activity was determined in aliquots containing 1 mg protein by monitoring the oxidation of NADH in the presence of dihydroxyacetone phosphate at room temperature by absorbance at 340 nm, by the method of Kozak and Jensen as modified by Wise and Green [35]. The activity of cell lysates was compared to that of purified G3PDH as a control, after dilution of the enzyme in 300 mM triethanolamine, pH 7.4, where one unit of enzyme activity oxidizes 1 × 10−9 moles NADH/min.

Statistical analysis

Where applicable, results are presented as the mean and standard deviation of replicate treatments. Statistical significance was calculated using the non- paired Student’s t-test between treatment groups. A value of p equal to or less than 0.05 was considered to be significant, and notation is detailed in the figure legends.

Results

Characterization of AH3960

Agonist binding of the PTH1 receptor activates two major signaling systems; the adenylyl cyclase/cyclic-AMP/protein kinase A (AC/cAMP/PKA) pathway and the phospholipase-C/ protein kinase C and calcium (PLC/PKC/Ca) pathway [36]. AH3960 (dibutyl-diaminomethylene-pyrimidine-2,4,6-trione, Fig. 1) was identified from a high-throughput screen. The compound was found to behave as a weak PTH1 receptor agonist (Fig. 2) by producing a concentration-dependent stimulation of (i) the activity of a cAMP-responsive reporter gene in a human embryonic kidney cell line (HEK-293) stably expressing wild-type hPTH1R (EC50 = 1.5 μM), (ii) intracel- lular calcium mobilization in the same cell line (EC50 = 3.2 μM), and (iii) cAMP production in HEK293-PTH1R cells and ROS17/2.8 rat osteosarcoma cells (no EC50 determinable). By comparison, EC50 values for hPTH(1–34) in these assays was 0.3 nM, 1 nM and ∼ 0.5 nM, respectively, indicating that the potency of AH3960 for the PTH1 receptor relative to hPTH (1–34) was around 5000-fold lower with respect to cAMP signaling and 3000-fold lower with respect to calcium mobilization. The higher concentration of both agonists required for stimulation of calcium mobilization compared to induction of cAMP is in agreement with reports in a porcine kidney cell line for activation of PLC versus AC/cAMP [37].

PTH stability in culture medium

To verify that PTH was stable in culture for greater than 48 h, and thus ensure that PTH addition every 48 h to MSC cultures results in continuous hormone exposure, the concen- tration of PTH in medium spiked with 10 nM hPTH(1–34) was determined after 0, 1, 2, 6, 24, 48 and 72 h after addition by IRMA. Over this period there was no detectable reduction in PTH(1–34) concentration in medium incubated with or without cells at 37°C, indicating that addition of PTH every 48 h does indeed provide a continuous exposure of the cells (Fig. 3).

Effects of PTH on cellular morphology

Human MSC cultured in AMM without adipogenic supple- ments exhibited an elongated, fibroblastic morphology typical of bone marrow stromal cells, whereas those cultured under mild adipogenic conditions (AMM + DIIT) had adopted a more flattened, irregular appearance by post-confluent day 6 (Fig. 4). Initial indications of a differential effect of intermittent versus continuous hPTH(1–34) treatment was the preservation of the fibroblastic morphology by those cultures treated intermittently. In contrast, cells treated with continuous hPTH(1–34) appeared identical to those grown in AMM+DIIT.
By the second week of post-confluency, extensive lipid droplets became visible in the cytoplasm of some cells cultured in AMM+ DIIT, whereas cells with an adipocytic morphology were never observed in cultures maintained without the adipogenic inducers, demonstrating the absence of spontaneous adipocyte differentiation in the MSC population under basal conditions (Fig. 4). Moreover, similar numbers of lipid droplet- containing cells were produced in the cultures treated continuously with hPTH(1–34) in adipogenic medium com- pared to cultures in AMM+DIIT, whereas far fewer adipocytic cells developed in those cultures grown in AMM+DIIT but treated with PTH intermittently.

Effect of PTH on alkaline phosphatase enzyme expression and neutral lipid accumulation

Differentiation of MSC into mature adipocytes and osteo- blast lineage cells with either continuous or intermittent PTH treatment was also assessed by staining the cultures at 2– 3 weeks post-confluency for neutral lipid with oil red O and for alkaline phosphatase, to identify adipocytes and osteoblasts, respectively (Fig. 5). Staining with oil red verified the observed differences in cell morphology in that no adipocytes developed in cultures grown in AMM medium without the adipogenic inducers but that abundant adipocytes were formed with culture in AMM+DIIT. A comparable number of adipocytes developed in cultures grown in AMM + DIIT to those additionally supplemented with continuous PTH(1–34). Interestingly, daily 1 h treatment with hPTH(1–34) produced a marked inhibition of cells staining positively with Oil Red O (Fig. 5, compare panel C with B and D).

To determine whether continuous and intermittent PTH also differentially affected the development of the osteoblastic phenotype, parallel cultures were stained for alkaline phospha- tase (AP), a marker indicative of (but not specific for) osteoblast differentiation. It was noted that lipid-containing cells rarely stained positively for AP. The intensity and number of AP- positive cells was increased in cultures treated with DIIT compared with those cultured in AMM alone, most likely because of the glucocorticoid, suggesting that adipocytic and osteoblastic differentiation may both be induced when DIIT is present (Fig. 5, panel B). The intensity and number of AP+ cells was further increased by PTH and the effect of PTH was greater with intermittent daily administration rather than continuous exposure (Fig. 5, panels B–D). The effect of intermittent treatment with AH3960, was also examined. Daily 1-h treatment of post-confluent cultures of MSC with 30 μM AH3960 reduced adipocyte differentiation in the presence of DIIT (Fig. 5, panel E), but the effect was less dramatic compared with intermittent hPTH(1–34). The AH3960 also increased the number of cells staining positively for AP compared with AMM+DIIT, but again the effect was less pronounced than with PTH. The numbers of differentiated adipocytes for each treatment were determined from photographs of Oil Red O-stained cultures for the 8 or more experiments performed throughout the study (1 photograph of a representative field per treatment per experiment). Photographs were taken between days 8 and 18, depending on the experimental end-point. The adipocyte numbers (mean±SD) were as follows: AMM control, 0; DIIT, 12.0 ± 7.6; DIIT+ continuous hPTH(1–34), 10.1 ± 7.6; DIIT+ intermittent hPTH(1–34), 2.8 ± 3.4; and DIIT+ intermittent AH3960, 3.8 ± 3.7. Compared to DIIT alone, the adipocyte number was significantly lower for both intermittent hPTH(1–34) and intermittent AH3960 treat- ments (p < 0.05 by t-test). Together, these observations indicate that intermittent stimulation with PTH1 receptor agonists suppress commitment of MSC to the adipocyte lineage and may simultaneously promote their osteogenic development.

Effect of PTH on the expression of osteoblastic and adipocytic marker genes

The divergent effects of intermittent and continuous treatment with hPTH(1–34) on MSC differentiation were analyzed at the mRNA level for the expression of genes associated with the adipocyte and osteoblast phenotype. Total RNA was extracted on day 12 of post-confluent culture and RT- PCR performed (Fig. 6). The steady-state mRNA levels of PPARγ2, lipoprotein lipase (LPL) and glycerol-3-phosphate dehydrogenase (G3PDH), representative of both early and late adipocyte differentiation markers, were clearly induced in MSC cultures treated with AMM+DIIT compared to cells grown in AMM without adipogenic inducers, as expected. In agreement with the impaired adipocyte formation observed in cultures treated with intermittent hPTH(1–34) (50 nM) or intermittent AH3960 (30 μM), the expression level for each of these adipocyte markers was markedly reduced under these condi- tions. In contrast, the expression of PPARγ2, LPL and G3PDH remained elevated in cells treated continuously with either PTH (1–34) or AH3960.

The extent of osteoblastic differentiation was assessed by the expression of osteocalcin (OC), runx2/cbfa1 and PTH1R mRNAs (Fig. 6). Relative to the mRNA level in cells cultured with AMM+DIIT, runx2 and OC expression was not affected by either intermittent or continuous treatment with PTH(1–34) or AH3960. The level of PTH1R mRNA was, in fact, lower in cells treated intermittently compared to those treated continu- ously with hPTH(1–34) or AH3960. The expression of AP osteopontin and bone sialoprotein mRNAs was also deter- mined: the levels of AP and BSP mRNA were largely unaffected by either intermittent or continuous PTH, whereas expression of OP was weakly reduced with intermittent treatment (data not shown). Similar results for the expression levels of all of these mRNAs were also obtained using a lower PTH(1–34) concentration of 5 nM (not shown). Taken together, the gene expression results indicate that although expression of adipocyte differentiation related genes was suppressed by intermittent PTH and AH3960, the osteoblast marker genes were not reciprocally increased in a consistent manner under these conditions.

Glycerol-3-phosphate dehydrogenase (G3PDH) activity

To further verify the inhibition of adipocyte differentiation by intermittent administration of PTH or AH3960, the activity of the late adipocyte differentiation marker enzyme G3PDH was measured in cell lysates harvested on day 18 of post-confluent culture. As shown in Fig. 7, G3PDH enzyme activity was barely detectable in control cells cultured in AMM medium alone; conditions in which adipocytes do not develop. Consistent with the increased numbers of morphologically mature adipocytes seen at this time, culture in the presence of the adipogenic cocktail (AMM+DIIT) increased G3PDH activity 30-fold over controls, and cells treated continuously with PTH exhibited a 16- fold increase in activity. By comparison G3PDH activity was
increased by only 4- to 5-fold in cells cultured in AMM+DIIT and treated intermittently with 50 nM hPTH(1–34) or 30 μM AH3960. In separate experiments utilizing MSC obtained from other donors, G3PDH activity was stimulated to similar levels by both AMM+vDIIT and by AMM+ DIIT with continuous PTH treatment, whereas AMM+DIIT with intermittent PTH con- sistently gave far lower activities. The variability in the effect of continuous PTH on G3PDH activity between experiments accounts for the apparent discrepancy between effects on enzyme activity and its lack of effect on G3PDH mRNA level shown in Fig. 6.

The anti-adipogenic effect of intermittent PTH is mediated via adenyl cyclase/cAMP

The responses to PTH are mediated by multiple intracellular signals, principally the adenyl cyclase (AC)/cAMP/protein kinase A (PKA) and phospholipase-Cβ (PLCβ)/protein kinase C (PKC)/calcium pathways, although there is also cross-talk with the MAP kinase/Erk pathway. To determine which of these is responsible for the anti-adipogenic effect of intermittent PTH, cells were pre-treated with pathway-selective inhibitors for 30 min prior to the daily 1 h stimulation with 50 nM hPTH(1– 34). Pre-treatment with the MAPK inhibitor PD 98059 (10 μM) failed to prevent the inhibitory effect of intermittent PTH on adipocyte formation, whereas pre-treatment with the selective PKA inhibitor H89 (25 μM) resulted in an almost complete conversion of cells into adipocytes (Fig. 8). This latter finding implies that blockade of an intermittent PKA signal induces commitment and/or differentiation of MSC to adipocytes. Conversely, daily 1 h treatment of MSC cultures with forskolin (1 μM), a direct activator of AC, mimicked the effect of intermittent PTH and prevented adipocyte formation as shown by the markedly reduced numbers of oil red O-positive cells after 18 days post-confluent culture (Fig. 8). Thus, intermittent PKA activation by forskolin prevented adipogenesis. However, intermittent treatment with forskolin had little stimulatory effect on the intensity and number of AP+ cells, in contrast to the effect of intermittent PTH, indicating that this effect of intermittent PTH is likely mediated through a PKA-independent mechanism.

The ability of H89 pre-treatment to inhibit the anti- adipogenic effect of intermittent PTH, and of forskolin to mimic this effect of PTH, was also demonstrated at the mRNA level (Fig. 9). Whereas daily 1 h stimulation with PTH or forskolin decreased steady state mRNA levels of PPARγ2 and G3PDH, the H89 pre-treatment increased expression of these genes to levels similar to that of cells cultured in DIIT without PTH. In contrast, pre-treatment with PD 98059 did not affect the PTH-induced changes in adipocyte marker gene expression. None of the pathway inhibitors and activators altered mRNA expression of the osteoblast marker osteocalcin.

Discussion

Here we demonstrate that adipocyte formation by human MSC cultures is differentially regulated by hPTH(1–34) depending on its mode of administration: intermittent PTH in the form of daily 1-h treatment inhibited adipogenesis whereas continuous exposure had no effect. Inhibition of adipogenesis by intermittent PTH was demonstrated by a reduction of (i) cells possessing a differentiated adipocyte phenotype (oil red O staining of cytoplasmic lipid vacuoles), (ii) mRNA expression of adipocyte marker genes, and (iii) activity of the adipocyte marker enzyme G3PDH. In addition, the regulation of adipogenesis by duration of ligand exposure was mimicked by a non-peptide, small molecule PTH1 receptor agonist, AH3960. The anti-adipogenic effect of intermittent PTH is in agreement with other studies showing inhibition of adipogen- esis in both preadipocyte and mesenchymal stem cell lines by the other physiological PTH1 receptor agonist, PTHrP [26,27] and by an in vivo study of ovariectomized monkeys following an 18-month treatment with daily hPTH(1–34) [29].

Although intermittent PTH inhibited adipogenesis in the MSC cultures, a reciprocal increase in osteoblastic differentia- tion was less evident. Whereas intermittent but not continuous PTH treatment increased the number and intensity of cellular AP staining (indicating increased AP enzyme activity), the mRNA expression level of several osteoblast marker genes including osteocalcin, BSP and runx2/cbfa1 were not affected. The lack of effect on these later osteoblast phenotypic markers suggests that intermittent PTH treatment alone may be insufficient for differentiation of isolated osteoprogenitors in vitro beyond an early stage of maturation. Alternatively, treatment may have been for an insufficient duration (12 days) to obtain effects on late differentiation markers.

Another possibility is that the effects of PTH1R agonist treatment were masked or over-ridden by the presence of dex in the adipocyte induction cocktail, which is well known to stimulate AP, BSP and PTH1R gene expression but inhibit OC and OP expression in osteoblast lineage cells including MSC [38,39]. Nevertheless, our finding of elevated AP staining following intermittent PTH is in agreement with previous reports: AP and IGF-1 mRNA levels were up-regulated in primary murine bone marrow stromal cells by 6 h/day PTH treatment for 4 days but not after continuous treatment [5]; the number of total and AP+CFU-f colonies in ex vivo bone marrow cultures has been found to be increased after only 1 week of intermittent PTH treatment in rats [8,9]; and AP activity, osteocalcin gene expression and bone nodule formation by isolated fetal rat osteoblasts were either stimulated or inhibited by intermittent PTH depending upon the duration of exposure [40]. Similarly, type I collagen promoter activity in neonatal calvarial osteoblasts was increased by transient PTH but decreased by continuous PTH, which also indicates enhancement of osteogenesis in vitro by intermittent treatment [41]. Thus, although our findings do not demonstrate overt osteoblast differentiation by stimulation of MSC with inter- mittent PTH in vitro, we hypothesize that by inhibiting adipocytic commitment and/or differentiation, PTH may enable bone marrow osteoprogenitors to undergo osteogenesis driven by other factors present in the bone tissue environment in vivo.

At least two processes could be contributing to the PTH effect on adipocyte formation: prevention of adipogenesis by monopotential adipocyte progenitors or prevention of adipo- genesis by bi/multi-potential progenitors and only effects on the latter population of cells would be expected to result in inverse regulation of adipogenesis and osteogenesis. One mechanism that has been proposed for this reciprocal lineage commitment is the negative regulation of runx2 expression by PPARγ [42,43]; transcription factors essential for differentiation to the osteo- genic and adipocytic lineages, respectively. In agreement with this, heterozygous PPARγ-deficient mice exhibit increased bone mass due to the stimulation of osteoblastogenesis from bone marrow progenitors [21]. However, although we were able to clearly demonstrate inhibition of PPARγ mRNA levels after intermittent treatment of human MSC with PTH or AH3960, no changes in runx2 expression were observed.

Agonist binding of the PTH1 (PTH/PTHrP) receptor in osteoblasts leads to the activation of two principal intracellular signaling cascades: the AC/cAMP/PKA pathway and the PL-C/ PKC and calcium pathway, via coupling to the G-protein α- subunits of Gs and Gq, respectively [36]. The relative importance of these signaling pathways to the bone anabolic effect of intermittent PTH remains unclear, however [44]. In our studies, the anti-adipogenic action of intermittent PTH treat- ment in MSC cultures appeared to be mediated predominantly through the AC/cAMP pathway because intermittent treatment with forskolin (a direct stimulator of AC) also inhibited adipocyte formation and, conversely, intermittent treatment with a PKA inhibitor produced extensive adipogenesis. This finding therefore implies that intermittent stimulation of AC/ cAMP prevents commitment and/or differentiation to adipo- cytes. Our results are consistent with the ability of antisense oligodeoxynucleotides to the Gs protein α-subunit (which couples PTH1R and AC) to accelerate the differentiation of pre- adipocytes [45]. Therefore, the inhibitory effect of intermittent stimulation of AC/cAMP signaling on adipogenesis contrasts with the well-recognized ability of phosphodiesterase inhibi- tors, and other treatments generating prolonged elevation of cAMP and CREB activation, to promote adipocyte differentia- tion [46]. The involvement of other pathways activated through the PTH1 receptor in addition to the cAMP pathway cannot be discounted, however, because the inhibitory activity of PTHrP on 3T3-L1 preadipocyte differentiation was mediated through MAPK-dependent suppression of PPARγ activity [26].

Intracellular signaling pathways independent of cAMP appear to be responsible for the effects of PTH on osteoblasts and osteoprogenitors because, in our study, forskolin had little if any effect on the number of AP+ cells. Consistent with this notion are the observations that (i) enhancement of fetal rat osteoblast differentiation by intermittent PTH requires coopera- tive interaction of both cAMP/PKA and calcium/PKC pathways [40] and (ii) although PTHrP(1–31)-amide induced cAMP in cultured rat osteosarcoma cells as effectively as PTH(1–31)- amide, the former was unable to stimulate trabecular bone formation in ovariectomized rats [47]. Consequently, multiple pathways may participate in elaboration of the full bone anabolic effect of PTH.
The generation of non-peptide agonist ligands to family B G- protein coupled receptors, to which the PTH1 receptor belongs, has proven notoriously difficult. To our knowledge, our studies with AH3960 may represent the only report of a non-peptide small molecule possessing PTH agonist behavior.

Thus, we have shown that not only does AH3960 stimulate intracellular cAMP and Ca2+ release, the two major intracellular signaling pathways important
for PTH responses in osteoblasts, but it is also able to mimic the effects of PTH on more complex cell systems such as the lineage determination of osteoprogenitor cells. It should be noted that these effects require concentrations of AH3960 in the micromolar range owing to the low affinity of the molecule for the PTH1 receptor relative to PTH(1–34). However, although the ability of AH3960 to mimic many of the effects of PTH in osteoblast lineage cells may be compelling, its activity on other PTH target cells has not been investigated. More importantly perhaps, the exact mode of interaction of the molecule with the PTH1 receptor including binding site and contact residues remains to be elucidated. It is also unclear whether AH3960 behaves as a pure agonist or can additionally function as an allosteric modulator of PTH.

Intermittent and continuous PTH treatments may also exert divergent effects on bone formation and remodeling in the growing skeleton. Transgenic mice selectively expressing a constitutively active PTH1 receptor in osteoblasts, developed dramatically elevated trabecular bone volume due to increased trabecular and endosteal osteoblast activity [48] as well as delayed formation of bone marrow cavities and the appearance of marrow adipocytes [49]. Although the mutant PTH1 receptor (H223R, one of the mutations identified in Jansen’s metaphy- seal chondrodysplasia) causes ligand-independent stimulation of cAMP, its constitutive activation may still result in net intermittent signaling in vivo because the mutant, like the wild-type receptor, undergoes agonist-dependent internalization and desensitization [50].

In summary, we have demonstrated that either PTH or a non- peptide PTH1 receptor agonist can exert divergent effects on adipocyte and osteoblast commitment/differentiation in adult human osteoprogenitors in vitro depending upon the duration of exposure. While continuous treatment had little effect, inter- mittent treatment inhibited adipogenesis and may have even promoted the early phase of osteoblast differentiation. Assum- ing that these effects also occur in vivo, the prevention of osteoprogenitor differentiation to the adipocyte lineage would be expected to contribute to the overall stimulatory effect of daily treatment with PTH on trabecular bone formation, mass and strength.