CBP/p300-Interacting Protein CITED1 Modulates Parathyroid Hormone Regulation of Osteoblastic Differentiation

Abstract

PTH regulates osteoblastic differentiation and activity and exerts different overall skeletal effects in vivo, depending on the schedule and dose of administration. In clonal Wt9 murine osteoblastic cells, mRNA and protein levels of CITED1 transcriptional coactivator were strongly up-regulated by human (h) PTH(1–34). Stimulation of CITED1 mRNA by PTH was transient, peaking at 4 h, concentration dependent, and blocked by actinomycin D but not cycloheximide. The stimulation was mimicked by forskolin, phorbol ester, and the cAMP-selective PTH analog [G¹,R¹⁹] hPTH (1–28) and inhibited completely by the protein kinase A inhibitor, H89 and partially by phorbol ester-induced protein kinase C depletion. Increased CITED1 expression was not maintained during persistent (24 h) PTH exposure. Cultured primary calvarial osteoblasts from neonatal homozygous or hemizygous CITED1-knockout (KO) mice achieved 2-fold greater mineralized nodule formation in comparison with wild type (WT) osteoblasts. This effect was blocked by restoration of CITED1 expression via adenoviral gene transfer. Intermittent administration of hPTH(1–34) (10 nm, for 4 h every 48 h) for 3–6 wk increased mineralization up to 2-fold over basal levels in both WT and CITED1 KO mouse calvarial cell cultures. Whereas the cAMP-selective [G¹,R¹⁹]hPTH(1–28) analog [at 100 nm, equivalent to 10 nm hPTH(1–34)] did not stimulate mineralization in WT cultures, it was twice as effective as hPTH(1–34) in CITED1 KO cultures. Thus, CITED1 negatively regulates osteoblastic differentiation in vitro and inhibits the cAMP-dependent stimulation of differentiation by intermittent PTH. We conclude also that PTH receptor signaling pathways independent of cAMP restrain osteoblastic differentiation, an effect normally obscured in the presence of CITED1 but revealed in its absence.

PTH INCREASES BONE mass when administered intermittently at appropriate doses to animals or humans, whereas continuous exposure to high concentrations of the hormone accelerates bone turnover and sustains net bone resorption and bone loss, especially at cortical sites (1,2,3,4). Both of these effects of PTH on bone metabolism are believed to be mediated by activation of PTH/PTHrP receptors (PTHR1) expressed in cells of the osteoblastic lineage, which indirectly regulate the osteoclastic response (5). More detailed understanding of the molecular mechanisms of PTH action on bone metabolism is of intense interest, and it would be particularly advantageous to learn whether it is possible to modulate the balance between the anabolic and catabolic effects of the hormone in vivo.

Analysis of osteoblastic gene regulation by PTH has provided important clues to the molecular basis of the hormone’s action on osteoblasts (6,7). In a preliminary (unpublished) DNA microarray study in a clonal osteoblastic cell line (Wt9), we identified CITED1 as a gene whose expression was strongly up-regulated by PTH. This study was directed at understanding the role of CITED1 in osteoblasts and in their functional response to PTH.

The gene for CITED1, formerly melanocyte-specific gene 1, was mapped to the X chromosome (8) and identified as a nuclear protein (9). Recent studies have shown that CITED1, as well as CITED2, CITED3, and CITED4, strongly activates transcription via the interaction of its conserved C-terminal transcriptional activating domain (CR2 domain) with CBP/p300 (10,11). CITED1 binds CBP/p300 to form a complex, which then interacts with mothers against decapentaplegic homolog-4 to enhance transcription induced by TGFβ. The binding affinity of CITED1 to CBP/p300 and enhancement of mothers against decapentaplegic homolog-mediated transcription is suppressed by heat-shock cognate protein 70, possibly via a competition mechanism (heat-shock cognate protein 70 also forms a complex with CITED1) (10,11). CITED1 also selectively binds the activation function 2 domain of the liganded estrogen receptors-α and -β and may enhance estrogen-dependent gene transcription (including up-regulation of TGFα) (12). Given that TGFs and estrogen, like PTH, are powerful regulators of bone metabolism, our finding that CITED1 is induced by PTH in osteoblasts led us to anticipate that CITED1 may play a role in controlling osteoblastic differentiation and its regulation by PTH.

Materials and Methods

Peptides and other reagents

PTH peptides used in these experiments were prepared as previously described (13). Other reagents were purchased from Sigma-Aldrich Inc. (St. Louis, MO) unless noted otherwise.

Mice

Wild-type (WT) and pregnant C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME). CITED1 knockout (KO) mice [previously generated by ablation of the second and third exons of the CITED1 gene and maintained in a mixed C57BL/6J and 129sv background (14)] were genotyped using a Lac Z-PGK-neo cassette incorporated in the targeting vector. For isolation of primary osteoblasts (see below), pups with the desired genotypes were produced by matings of heterozygous females (neo/X) with either hemizygous (neo/Y) or WT (X/Y) males. Mice, housed five or fewer per cage, were given free access to water and fed with a standard diet in a room maintained at 22 C with 60–75% humidity on a 12-h light, 12-h dark cycle. Animals were maintained in facilities operated by the Center for Comparative Medicine of the Massachusetts General Hospital in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and protocols were approved by the institution’s subcommittee on research animal care.

Cell culture

Clonal Wt9 osteoblastic cells were generated and maintained as previously described (15), MC3T3-E1 cells were purchased from American Type Culture Collection (Manassas, VA) and cultured under the same conditions as previously described (15).

We isolated and cultured calvarial osteoblasts from each neonatal mouse separately until the genotype was identified, whereupon the cells were combined into a CITED1-KO (NEO/NEO; NEO/Y) group or a WT (X/Y; X/X) group, respectively, for long-term culture. The DNA specimens used for PCR genotyping were extracted by a genomic DNA isolation kit (Qiagen Science, Germantown, MD) from tails collected during cell isolation. For WT allele amplification, forward primer (5′-AACCCCCATCCTTCAACCTG-3′) and reverse primer (5′-AACAGAATCGGTGGCTTTTT-3′) generated a 439-bp band, whereas for mutant allele (NEO insertion) detection, primers of 5′-TGGTCGAATGGGCAGGTAGC-3′ and 5′-CGCTTGGGTGGAGAGGCTAT-3′ produced a 366-bp fragment. For cell isolation, minced calvarial frontoparietal bones from 2-d-old neonatal mice were subjected to sequential collagenase (type I and II; ratio 1:3) digestion. Of six fractions generated by serial 20-min digestions, fractions 3–6 were combined and the cells were plated at 5 × 10⁴/cm² in a humidified atmosphere (95% air-5% CO₂) in αMEM supplemented with 10% fetal bovine serum (FBS) (16).

Adenovirus generation and infection

CITED1 cDNA, including a 5′ Kozak sequence, was cloned by PCR using high-fidelity Taq enzyme (Invitrogen, Carlsbad, CA), and its sequence was confirmed by dual-direction DNA sequencing. Adenoviruses designed to express either CITED1 (Ad-CITED1) or control proteins [Ad-Lac Z or adenoviruses expressing green fluorescent protein (GFP)] were generated using the BD Adeno-X Expression System 1 (CLONTECH Laboratories, Inc., Mountain View, CA). In brief, CITED1 cDNA was first cloned into pShuttle plasmid by endonuclease digestion and ligation, after which a mammalian expression cassette carrying the CITED1 cDNA was cut from the pShuttle plasmid with I-ceu/PI-Sce I and subsequently inserted into linearized adenoviral DNA using T₄ ligase. The viral DNA packaging and virus amplification were accomplished in HEK 293 cells. Purification of virus was accomplished with a virus purification kit, using a filter system that binds virus (CLONTECH). Viral titers were determined by infecting HEK293 cells with serially 10-fold diluted virus and detecting infected cells using an immunohistochemical stain for the viral hexon (Adeno-X rapid titer kit; CLONTECH). Expression of CITED1 was detected by RT-PCR and Western blot (as described below) 2 d after addition of virus to the cells. The protocol was approved by the Harvard Committee on Microbiological Safety, certified by the Environmental Health Office of the Boston Public Health Commission, consistent with National Institutes of Health guidelines.

Adenovirus infection was conducted 1 d after primary osteoblasts were plated into cell culture dishes. Briefly, cells were suspended by trypsin, counted, and plated into type I collagen-coated 24-well plates (5 × 10⁴ cells/well, for mineralization) or six-well plates (2 × 10⁵ cells/well, for quantitative PCR and Western blot experiments). After 24 h, the cells were washed twice with αMEM medium and refed with αMEM supplemented with 10% FBS. Ad-CITED1 or control virus (Ad-Lac Z) was added into each well at the concentration of 100 infectious units (ifu)/cell. After 48 h, the cells were refed with mineralization medium (αMEM plus 10% FBS, 10 mm β-glycerophosphate, and 50 μg/ml ascorbic acid) and then refed with this medium three times a week.

RNA isolation

Cells were plated into six-well plates or 10-cm dishes at 2.5 × 10⁴ cells/cm² and maintained at 37 C in αMEM supplemented with 10% FBS. The concentration of FBS was reduced to 1% when the cells reached 90% confluence, and 24 h thereafter cells were treated with human (h) PTH(1–34) or other agents, for designated time intervals. Before RNA extraction, cells were washed gently with cold PBS three times and total RNA was isolated using an RNeasy kit (Qiagen Science, Gaithersburg, MD).

Quantitative RT-PCR

Expression of CITED1 and other mRNAs was measured by two-step real-time PCR performed as previously described (15). The primers used are shown in Table 1.

Table 1.

Primers for quantitative RT-PCR

	Forward	Reverse
CITED1	ccactagctcctctggatcg	agccccttggtactggctat
CITED2	ctgcagaagctcaacaacca	ctggtttgtcccgttcatct
CITED4	gtagcacgcacctgcagtc	gaagcaatcgaactcgctct
GAPDH	tgtcgtggagtctactggtg	gcattgctgacaatcttgag
RANKL	agccgagactacggcaagta	gcgctcgaaagtacaggaac
RAMP3	tgcaccttcttccactgttg	aggttgcaccacttccagac
BSP	agggaactgaccagtgttgg	actcaacggtgctgcttttt
OC	aagcaggagggcaataaggt	gcggtcttcaagccatactg
ALP	gctgatatgagatgtcctt	gcactgccactgcctact
RUNX2	cccagccacctttacctaca	tatggagtgctgctggtctg

Western blots

Cells were lysed in radioimmunoprecipitation assay solution [1% Nonidet P-40, 1% sodium deoxycholate, 0.1% (wt/vol) sodium dodecyl sulfate, 0.15 m NaCl, 2 mm EDTA, 50 mm NaF, 0.2 mm sodium vanadate, and 4 μl/ml of protease inhibitor cocktail] by repeated passage through a 22-gauge needle. Western blots of the protein samples were conducted as described (15). Monoclonal antibody directed against CITED1 was generated as described by Yahata et al. (12). Intensity of the Western blot bands were quantified densitometrically (AlphaImager 2200; Alpha Innotech Corp., San Leandro, CA), normalized to the level of the same protein measured in control cells (set at 1.0) and presented as fold over control.

cAMP accumulation

cAMP accumulation was measured as described previously (15).

Assays of cellular differentiation

Osteoblastic cells from WT and CITED1 gene KO mice were plated in 24-well plates coated with type I collagen (BD Bioscience, Bedford, MA) at the concentration of 5 × 10⁴/well, and incubated at 37 C until 90% confluent. Treatment with PTH peptide then was conducted using a 4 h/48 h schedule, whereby cells were cultured in mineralization medium containing peptide or vehicle alone (0.1% trifluoroacetic acid) for 4 h, at which point the medium with peptide or vehicle was aspirated and cells were rinsed with αMEM twice before being refed with fresh mineralization medium and incubated for another 44 h before the procedure was repeated. After defined intervals, alkaline phosphatase (ALP) activity was assessed histochemically or enzymatically in cell lysates using methods previously described (15). Results of enzymatic assays were expressed in terms of standard units of ALP activity.

For analysis of matrix mineralization, cells plated and treated as above were fixed in 10% neutral formalin at the designated time and the presence of mineralized nodules was assessed by Alizarin Red staining. Briefly, the fixed cells were gently washed with distilled water, incubated in 0.1% Alizarin Red S/70% ethanol solution and then washed with distilled water. Alternatively, to determine the calcium content of the cultures, cell monolayers were washed in Ca²⁺- and Mg²⁺-free PBS and then incubated for 3 h in 0.2 ml of 0.6 n HCl. Extracted calcium then was measured spectrophotometrically at 612 nm after reaction with methylthymol blue (17).

Cell number and apoptosis

Primary calvarial osteoblastic cells, isolated from WT and KO neonates, were plated in type I collagen-coated 24-well plates and incubated at 37 C in αMEM +10% FBS until confluence. Cells then were exposed to cyclic intermittent 4 h/48 h treatment with PTH peptides as described above. At selected intervals (d 3, 14, and 28), cell number in each well was assessed by addition of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium and production of formazan product was measured spectrophotometrically according to the manufacturer’s instructions (Promega Corp., Madison, WI).

For measurements of apoptosis, calvarial osteoblasts were plated in 24-well plates, cultured, and treated with PTH peptides as described in table legends. Cells were detached enzymatically (ACCUMAX; Chemicon, Temecula, CA), washed once with PBS, and stained with Annexin and 7-amino-actinomycin D (Guava Nexin kit; Guava Technologies Inc., Hayward, CA) before enumeration of the percentage of early (annexin positive/7-amino-actinomycin negative) apoptotic cells by flow cytometry (Guava PCA, GTI-0300375; Guava Technologies, Inc.).

Results

CITED1 is up-regulated by hPTH(1–34) in osteoblasts

CITED1 was identified in preliminary cDNA microarray experiments (unpublished data) as a gene whose transcription was strongly up-regulated by hPTH(1–34) within 4 h in the clonal calvarial osteoblastic cell line Wt9 (15). This initial microarray result was confirmed by quantitative RT-PCR, which showed that CITED1 mRNA in Wt9 cells was increased as much as 10-fold by 4 h stimulation with hPTH(1–34) at 100 nm (Fig. 1A). Neither CITED2 nor CITED4, other members of the mammalian CITED gene family, was regulated by PTH (Fig. 1A). To determine whether this CITED1 response to PTH was characteristic of normal osteoblasts, we isolated primary calvarial osteoblasts from C57BL/6 neonatal mice. As shown in Fig. 1B, CITED1 mRNA was increased 3-fold (up to 4-fold in some experiments) in response to treatment with hPTH(1–34) (100 nm). Similar results were observed using MC3T3-E1 cells (data not shown). Western blot analysis in Wt9 cells indicated that CITED1 protein expression also was increased by 6 h of treatment with hPTH(1–34) but returned to baseline despite the presumed continued presence of hPTH(1–34) (Fig. 1, C and D). To determine whether CITED1 might be a PTH-induced primary response gene, Wt9 cells were treated with cycloheximide (CHX; 5 μg/ml) for 1 h to inhibit protein synthesis before addition of hPTH(1–34) at 30 nm. The effect on CITED1 mRNA induction was compared with that on receptor activity modifying protein-3 (RAMP3), which we and others have found to be strongly induced in osteoblasts by PTH (13,18). As shown in Fig. 2, CHX pretreatment did not inhibit the increase in CITED1 mRNA induced by 4 h of PTH treatment, whereas CHX enhanced PTH induction of RAMP3 mRNA, as expected (18). Actinomycin D (5 μg/ml), an inhibitor of RNA synthesis, completely blocked PTH-induced expression of CITED1 (and RAMP3), indicating that PTH increases CITED1 mRNA expression via an effect on transcription (Fig. 2). Taken together, these data are consistent with the possibility that CITED1 may be a primary response gene for PTH action in osteoblastic cells.

Figure 1.

CITED1 expression in osteoblasts. A, Wt9 cells were treated with 100 nm hPTH(1–34) for 4 h before extraction of RNA and subsequent measurement of CITED1, CITED2, and CITED4 mRNA expression by quantitative RT-PCR, as described in Materials and Methods. Results are expressed as fold over control. B, Primary osteoblasts isolated from neonatal C57BL6 mice were treated with 100 nm hPTH(1–34) for 4 h before RNA extraction and measurement of CITED1 mRNA expression, as in A. C, Wt9 cells were treated with 30 nm hPTH(1–34), or vehicle alone (control), for 6 h or 24 h before simultaneous extraction of proteins in radioimmunoprecipitation assay buffer and assessment of CITED1 expression by Western blot analysis. D, Intensities of CITED1 and β-actin protein bands from five experiments performed as in C were assessed densitometrically (see Materials and Methods) and are presented as fold over the corresponding 6 h control in each experiment. Data are shown as mean ± sd for groups of three cultures (A and B). *, P < 0.01 vs. control (CON).

Figure 2.

CITED1 is a primary response gene transcriptionally regulated by PTH. Wt9 cells were treated with actinomycin D (ActD; 5 μg/ml) or CHX (5 μg/ml) for 1 h before addition of hPTH(1–34) (30 nm). After 4 h of PTH exposure, total RNA was extracted from the cells and the expression levels of CITED1 and RAMP3 mRNA were measured by quantitative RT-PCR and shown as fold over control (CON). Experiments were performed three times with similar results. Data shown are mean ± sd for groups of two cultures each. a, P < 0.01 vs. control; b, P < 0.01 vs. PTH; c, P < 0.05 vs. CHX.

PTH(1–34) increases CITED1 expression in a dose- and time-dependent manner in osteoblasts

To further analyze the effect of PTH on CITED1 expression, we investigated the time- and concentration-dependence of increased CITED1 mRNA expression in response to hPTH(1–34). As shown in Fig. 3A, CITED1 mRNA expression in Wt9 cells increased 5-fold between 1 and 2 h after addition of hPTH(1–34) and peaked (10-fold) at 4 h, declining thereafter to baseline by 24 h in the continued presence of the hormone (consistent with the results of Western blot analysis, Fig. 1, C and D). Studies of PTH peptide stability in osteoblastic cultures confirmed that 80% of hPTH(1–34) bioactivity could be recovered in medium after 24 h of incubation with cells (see supplemental Fig. 1, published as supplemental data on The Endocrine Society’s Journals Online Web site at http://endo.endojournals.org). In primary murine calvarial osteoblasts, persistent (nonpulsatile) exposure to hPTH(1–34) (100 nm) for 8 d blocked the acute up-regulation of CITED1 at 4 h; in contrast, three cycles of intermittent exposure (4 h every 48 h) to the same dose of hPTH(1–34) or [G¹R¹⁹]hPTH(1–28) (G¹R¹⁹(1–28)) did not prevent the acute CITED1 response to PTH rechallenge (supplemental Fig. 2). The CITED1 mRNA response to hPTH(1–34) in Wt9 cells was concentration dependent, increasing between 1 nm and 100 nm, which induced the maximal response (Fig. 3B).

Figure 3.

PTH increases CITED1 expression in a dose- and time-dependent manner in osteoblasts. Wt9 cells were treated with 100 nm hPTH(1–34) for indicated times (A) or 4 h with hPTH(1–34) or G¹R¹⁹(1–28) at the indicated concentrations (B). Total RNA was extracted and the expression of CITED1 was measured by quantitative RT-PCR and expressed as fold over control. Experiments were performed three times with similar results. Data are shown as mean ± sd for groups of three cultures. *, P < 0.01 vs. control (CON).

Signaling pathways mediating increased CITED1 expression by PTH

Upon binding active ligand, the PTHR1 engages several parallel signaling pathways, including cAMP/protein kinase A (PKA), and activation of protein kinase C (PKC) via either phospholipase C (PLC)-dependent or PLC-independent mechanisms (15). To determine which PTHR1 signaling pathway(s) is involved in the up-regulation of CITED1 expression, we first studied the responses to PKA and PKC pathway agonists or antagonists. These were compared with those of RAMP3 mRNA, which is known to be regulated mainly by cAMP/PKA in osteoblasts (18). As shown in Fig. 4A, forskolin (10⁻⁷ m), like hPTH(1–34), strongly induced CITED1 mRNA expression, as with RAMP3 mRNA expression. Active phorbol ester (TPA; 10⁻⁷ m) also mimicked the PTH CITED1 response but not the RAMP3 response. Pretreatment for 1 h with the PKA inhibitor H89 (20 μm), entirely blocked the CITED1 mRNA response to both PTH and forskolin (Fig. 4B). Prolonged pretreatment with TPA (1 μm) for 16 h to deplete PKC significantly inhibited PTH’s effect on the expression of CITED1 (PTH = 5.5-fold, control = pre-TPA = 1-fold, pre-TPA+PTH = 2-fold) but not that of RAMP3 (Fig. 4C). Finally, G¹R¹⁹(1–28), a cAMP/PKA-selective PTH analog (15), also increased CITED1 expression, although to a slightly lower maximum and with 10-fold less potency, as expected from its comparably reduced potency for stimulation of cAMP (Fig. 3B and supplemental Fig. 3). Collectively, these data are consistent with a role for cAMP/PKA and, to a lesser extent, PKC in up-regulation of CITED1 expression by PTHR1s.

Figure 4.

Signaling pathways mediating increased CITED1 expression by PTH(1–34). Wt9 cells were exposed to the indicated agonists for 4 h, ± pretreatment with inhibitors for 1 h, before extraction of total RNA and measurement of CITED1 and RAMP3 mRNA expression by quantitative RT-PCR (see Materials and Methods). A, Cells were treated with vehicle (CON), 100 nm hPTH(1–34) PTH(1–34)), 10⁻⁷ m forskolin (FSK), or 10⁻⁷ m TPA for 4 h. B, Cells were treated with hPTH(1–34) or forskolin as in A ± pretreatment with the PKA inhibitor H89 (20 μm). C, Cells were stimulated with hPTH(1–34) as in A with or without prolonged (16 h) pretreatment with 10⁻⁶ m TPA (Pre-TPA). Experiments were performed three times with similar results. Data are shown as mean ± sd for groups of three cultures. a, P < 0.01 vs. control; b, P < 0.01 vs. PTH; c, P < 0.05 vs. Fsk; d, P < 0.01 vs. pre-TPA.

CITED1 is an inhibitor of osteoblastic mineralization

To determine whether CITED1 is involved in regulating osteoblast differentiation, we isolated primary calvarial osteoblasts from WT and CITED1 KO neonatal mice. Both cell populations demonstrated the same cAMP response to hPTH(1–34) with the same EC₅₀ (WT = 0.83 ± 0.14 nm, KO = 0.86 ± 0.21 nm; supplemental Fig. 3A). After culturing these calvarial osteoblasts in mineralization medium for 4 wk, we observed that CITED1 KO cells produced more bone nodules than WT cells and deposited roughly twice as much calcium in the matrix (Fig. 5A; also see Figs. 6 and 7). Expression of osteoblastic genes (bone sialoprotein, osteocalcin, runt-related transcription factor 2, alkaline phosphatase, and receptor activator of nuclear factor-κB ligand) was increased in KO cells as well (supplemental Fig. 4). The increased differentiation demonstrated by CITED1 KO osteoblasts seemed not to be related to altered proliferation or apoptosis (supplemental Tables 1 and 2).

Figure 5.

Regulation of osteoblastic mineralization by CITED1. WT and CITED1 KO primary calvarial osteoblasts isolated from neonatal mice (A) or CITED1 KO primary calvarial osteoblasts, either uninfected (CON) or infected with Ad-CITED1 or Ad-Lac-Z at 100 ifu/cell (B) were cultured in mineralization medium for 4 wk. Mineralized nodules in the cultures then were assessed by staining with Alizarin Red S and imaged by either plate scanning (A, top panel, and B) or microscopy (A, bottom panel, at magnification ×40). Calcium content was measured in acid extracts (see Materials and Methods). Experiments were performed three times with similar results. Calcium content (milligrams per well) is expressed as mean ± sd for groups of three cultures. *, P < 0.01 vs. WT (A) or uninfected controls (B).

Figure 6.

Role of CITED1 in regulation of osteoblastic differentiation by intermittent PTH in 3-wk cultures. Primary osteoblastic cells from WT or CITED1 KO mice were plated in type I collagen-coated 24-well plates and cultured until 90% confluent before initiation of cyclic (4 h every 48 h) PTH treatment (see Materials and Methods), using the peptides and concentrations indicated. After 3 wk, mineralization was assessed by Alizarin Red-S staining (top panel) and calcium content was measured in acid extracts of the cultures (bottom panel). Bars depict means ± se of three independent experiments. a, P < 0.01 vs. same-genotype control; b, P < 0.01 vs. G¹R¹⁹(1–28) in same genotype; c, P < 0.01 vs. WT control.

Figure 7.

Role of CITED1 in regulation of osteoblastic differentiation by intermittent PTH in 6-wk cultures. Primary osteoblastic cells were isolated and treated with PTH peptides as described in Fig. 6. Histochemical staining for ALP activity (upper photograph), staining with Alizarin Red-S (lower photograph), and measurements of ALP activity (upper graph) and calcium content (lower graph) were performed after 6 wk of PTH treatment. a, P < 0.01 vs. same-genotype control; b, P < 0.01 vs. G¹R¹⁹(1–28) in same genotype; c, P < 0.01 vs. WT control. Note interrupted ordinate scale in upper graph.

To ascertain whether this augmented differentiation observed in cultured KO calvarial osteoblasts was due to the absence of CITED1 per se, as opposed to an alteration of the cell population or other secondary effects of the absence of CITED1 in vivo, we prepared a recombinant replication-defective adenovirus encoding CITED1 (Ad-CITED1) and infected CITED1 KO cells in vitro. In preliminary experiments, the efficiency of infection was analyzed by quantitative RT-PCR and Western blotting, which showed that 2 wk after Ad-CITED1 infection (multiplicity of infection = 100 ifu/cell), the CITED1 gene was persistently expressed in CITED1 KO cells, with mRNA levels approximately 30-fold that seen in WT cells (supplemental Fig. 5A). CITED1 expression remains as high as 23-fold over controls at 4 wk after infection (supplemental Fig. 5B). Western blotting confirmed the presence of the expected 21-kDa CITED1 protein in the infected cells (supplemental Fig. 5C). In parallel experiments, primary osteoblasts were infected with adenoviruses expressing GFP, after which GFP fluorescence was expressed within 2 d and at 2 wk was strongly expressed in virtually all cells. After 5 wk, some GFP still was visible (supplemental Fig. 5D).

Virus-infected primary osteoblasts were cultured for 4 wk before Alizarin Red S staining and calcium mass measurement. Compared with control or Ad-Lac-Z-infected CITED1 KO osteoblasts, those infected with active Ad-CITED1 showed greatly reduced mineralized nodule formation and calcium deposition, comparable with that in WT controls (Fig. 5B).

Role of CITED1 in PTH regulation of osteoblastic differentiation

Because PTH acutely and transiently increased CITED1 expression and CITED1 apparently restrains osteoblastic differentiation in vitro, we sought to determine whether the regulation of osteoblastic differentiation by intermittently applied PTH in vitro might be modified in the absence of CITED1. In preliminary experiments, we determined that intermittent treatment (4 h of 48 h) of WT primary calvarial osteoblasts with 10 nm hPTH(1–34) increased mineralization and calcium deposition in the cultures, as previously reported (19). To gauge the importance of the cAMP/PKA pathway in this response, we compared the effect of 10 nm hPTH(1–34) applied intermittently with that of the cAMP/PKA-selective analog G¹R¹⁹(1–28), applied similarly but at 100 nm (to adjust for its reduced potency in generating cAMP) (15) (supplemental Fig. 3, B and C). Cultures were performed using either WT or CITED1 KO primary osteoblasts.

The results of a typical experiment are illustrated in Fig. 6. As anticipated, after 3 wk of intermittent application, hPTH(1–34) (10 nm) significantly increased bone nodule formation and augmented calcium deposition in WT cultures by about 2-fold (0.569 ± 0.010 vs. 0.249 ± 0.007 mg/well). As previously observed (15), G¹R¹⁹(1–28) (100 nm) did not consistently mimic this effect of hPTH(1–34) in WT cells (0.387 ± 0.001 mg/well). Control CITED1 KO cells again showed more advanced nodule formation and greater calcium deposition than did WT controls (∼2.5-fold: 0.602 ± 005 vs. 0.249 ± 0.007 mg/well). Intermittent hPTH(1–34) doubled the calcium deposition in CITED1 KO cultures (to 1.313 ± 0.003 mg/well), indicating that CITED1 expression is not required for the increase in osteoblastic differentiation induced by intermittent PTH in vitro. Quite remarkably, however, the absence of CITED1 uncovered striking agonism in this assay for the cAMP/PKA-selective analog G¹R¹⁹(1–28), which increased calcium deposition 4-fold (to 2.565 ± 0.143 mg/well), a treatment effect two to three times that of hPTH(1–34) itself. These differences in responsiveness to hPTH(1–34) and G¹R¹⁹(1–28) were not associated with any differences in apoptosis in these cultures (supplemental Table 2) or to differences in degradation of these two peptides (supplemental Fig. 1). Similar results were obtained in 6-wk cultures (Fig. 7), in which the augmented differentiation of control CITED1 KO cells was especially evident when assessed by ALP activity and the absence of any prodifferentiation effect of G¹R¹⁹(1–28) in WT cells, again contrasted dramatically with its very powerful effect, relative to hPTH(1–34), in CITED1 KO cultures. These results argue strongly that the lack of efficacy observed for G¹R¹⁹(1–28) in WT cultures is not due to its lower potency in activating cAMP/PKA and, furthermore, that the cAMP/PKA-mediated differentiating effect of intermittent PTH in vitro is selectively suppressed by CITED1, which, in turn, is rapidly and transiently induced by PTH in a manner at least partly dependent on cAMP/PKA signaling.

Discussion

Our results indicate that expression of the transcriptional cofactor CITED1 (mRNA and protein) is acutely and transiently increased in osteoblastic cells during brief exposure to PTH. This response is specific to CITED1 among the CITED gene family members, appears to be mediated mainly via cAMP-dependent (but partly also by PKC dependent) PTHR1 signaling, and is not inducible after persistent prior exposure to the hormone. Examination of the differentiation of primary calvarial osteoblasts over several weeks in vitro indicated that the absence of CITED1 gene expression was associated with increased osteoblastic differentiation, as assessed by expression of ALP and other osteoblastic genes and formation of mineralized nodules. This augmented differentiation was normalized upon restoration of CITED1 expression via use of an adenoviral vector in vitro, which suggests that the enhanced osteoblastic differentiation is attributable to the absence of cellular CITED1 per se and not to other systemic influences to which the CITED1 KO calvarial cells may have been exposed in vivo. This is important because CITED1-null mice obtained via matings of heterozygous females and hemizygous males may exhibit moderate intrauterine growth retardation and lower birth weight, a phenotype that is lost during further postnatal development and that has been attributed to compromised placental function (20).

The conclusion that CITED1 may constitutively suppress osteoblastic differentiation was unexpected, given the supporting role this cofactor appears to play in other systems with respect to the actions of estrogen and TGF family members and the fact that induction of its expression by PTH in vitro, like the anabolic response to PTH in vivo, is associated with intermittent rather than persistent hormone exposure. Nevertheless, our results, the first to link CITED1 with regulation of bone function, strongly suggest that CITED1 restrains the pace of osteoblastic differentiation in vitro and, furthermore, that it is not required for the stimulation of differentiation exerted by intermittently applied PTH in the murine calvarial osteoblast culture system (indeed, it powerfully dampens the PTH effect, as discussed below).

Perhaps the most surprising result of these studies was the unexpected disproportionate augmentation, in CITED1 KO osteoblast cultures, of the prodifferentiating effect of intermittently applied G¹R¹⁹(1–28). G¹R¹⁹(1–28) is a signal-selective hPTH analog that lacks structural determinants needed for effective activation by PTHR1s of PLC (N terminal serine) and PLC-independent PKC (residues 29–34 of the hPTH sequence) (21). This analog can fully activate adenylyl cyclase, albeit with reduced potency (due to impaired binding caused by the C-terminal truncation). In most systems tested, including the osteoblastic cells studied here (13) (supplemental Fig. 3), the potency of G¹R¹⁹(1–28) is reduced approximately 10-fold relative to standard hPTH(1–34). In the present studies, we found that intermittently applied G¹R¹⁹(1–28), in contrast to hPTH(1–34), did not increase osteoblastic differentiation in WT calvarial osteoblastic cultures, even when used at a 10-fold greater molar concentration (100 vs. 10 nm) needed to compensate for the impaired binding affinity and to achieve equivalent cAMP stimulation. In striking contrast, when the same experiment was performed with CITED1 KO cultures, intermittent G¹R¹⁹(1–28) stimulated ALP and calcium deposition to a 2- to 3-fold greater extent than did hPTH(1–34). Several conclusions can be drawn from these observations. First, the absence of a prodifferentiating effect of G¹R¹⁹(1–28) in WT cultures is not due to lack of efficacy in activating PTHR1s on these cells, as also suggested by direct measurements of cAMP responses in the WT vs. KO osteoblasts (supplemental Fig. 3). Rather, it appears to result from a specific inhibition by CITED1 of the cAMP/PKA-dependent differentiation-enhancing effect(s) of PTH. Second, although cAMP-dependent signaling has been regarded as the primary basis for the anabolic effect of PTH in vivo, it seems clear that the cAMP pathway alone cannot account for the effect of intermittent (4 of every 48 h) PTHR1 activation to promote osteoblastic differentiation in the culture system used here. Rather, at least in the presence of CITED1, cAMP-independent PTHR1 signaling, likely triggered by determinants in the PTH (29–34) sequence of the ligand (15), appears to play a crucial role. In contrast, the reversal in agonist potency between hPTH(1–34) and G¹R¹⁹(1–28) observed in this system in the absence of CITED1 appears to have unmasked a strong inhibitory action of some form of cAMP-independent signaling as well, i.e. a response that is engaged by hPTH(1–34) but not G¹R¹⁹(1–28).

It is important to acknowledge that the effects of intermittent PTH to promote differentiation of murine calvarial osteoblasts in vitro bear an uncertain relation to the well-described anabolic effects of the hormone that accompany its intermittent administration in vivo. Among other limitations, the in vitro calvarial-cell cultures may not accurately reflect osteoblastic responses in vivo in bones formed by endochondral development, nor can osteoblastic responses in vitro predict fully the integrated response of intact bone that contains marrow stromal cells, osteoclasts, and embedded osteocytes. Nevertheless, these results are useful in generating hypotheses that can then be tested directly in vivo and likely reflect some actions of PTH on cells of the osteoblast lineage. One such hypothesis is that CITED1 expression in bone, transiently augmented by PTHR1 activation, may be one component of an autoregulatory feedback loop that serves to restrain the extent of PTH regulation of osteoblastic function (and hence its anabolic effect) and also to differentially modulate the responsiveness of osteoblasts to distinct PTHR1 signaling mechanisms. Our in vitro findings highlight the importance of further studies of skeletal responsiveness to PTH in mice lacking CITED1 expression and point to CITED1 as a potential pharmacologic target in the treatment of bone disease.