Influence of Aluminum on the Regulation of PTH- and 1,25(OH)2D3-Dependent Pathways in the Rat Osteosarcoma Cell Line ROS 17/2.8



The role of hormonal status in the development of aluminum (Al)-dependent renal osteodystrophy, which is characterized by reduced bone matrix deposition, still remains largely unknown. To address this question, we used the osteoblast-like osteosarcoma cell line ROS 17/2.8 to evaluate the role of Al on parathyroid hormone (PTH)- and 1,25-dihydroxyvitamin D3 (1,25(OH)2D3)-dependent activities in these cells. Al (1 μM) caused an inhibition of basal and 1,25(OH)2D3-induced alkaline phosphatase, but only at low doses (<1 nM) of the steroid. Al partly inhibited basal osteocalcin (OC) secretion in ROS cells (p < 0.001), and the dose-dependent increase in 1,25(OH)2D3-induced OC release by these cells was also reduced by 1 μM Al at low concentrations of the steroid (≤1 nM), whereas high doses of 1,25(OH)2D3 (≥5 nM) totally prevented the inhibiting effects of Al. Al also had strong inhibitory actions on PTH-dependent cAMP production by ROS cells over the concentration range tested (0.5–50 nM). This inhibitory action of Al was also observed for PTH-related peptide- (PTHrp, 50 nM) but not for Isoproterenol-dependent (100 nM) cAMP formation. To evaluate more fully the mechanism of this inhibition of cAMP formation, we investigated the effect of Al on toxin-modulated, G protein–dependent regulation of cAMP formation and on the activation of adenylate cyclase by Forskolin. Cholera toxin (CT, 10 μg/ml), applied to cells for 4 h prior to PTH challenge, enhanced cAMP production about 2-fold above PTH alone (p < 0.001), a process that was further stimulated by Al. Pertussis toxin (PT, 1 μg/ml, 4 h) did not modify basal PTH-dependent cAMP formation by ROS cells. However, PT treatment prevented the inhibitory effect of Al on cAMP formation by these cells (p < 0.025). The stimulation of adenylate cyclase by Forskolin (0.1 and 1 μM), which bypasses G protein regulation, was not modified by Al, indicating that Al does not affect adenylate cyclase directly. Northern blot analysis of PTH receptor mRNA levels showed that Al did not modify PTH receptor message in ROS cells. Likewise, Western blot analyses of G protein subunits showed that Al did not significantly alter Gs α subunit levels, in accordance with the results obtained for cAMP-dependent formation in response to CT. In contrast, Giα-1 and Giα-2 subunits were decreased by Al treatment, consistent with PT-restricted increases in cAMP formation in Al-treated ROS cells. Taken together, these results suggest that Al has multiple actions in osteoblast-like ROS cells. The effects of Al are modulated by hormonal control of the pathways investigated. Al affects 1,25(OH)2D3-regulated functions only when this steroid is low. Al has large inhibitory effects on PTH- and PTHrp-dependent cAMP formation. This last feature is related to the ability of Al to alter the G protein transducing pathway for PTH/PTHrp-dependent formation of cAMP since it does not affect adenylate cyclase activity directly and does not affect the PTH receptor message level. Thus, Al has stronger deleterious effects in osteoblast-like cells with an already compromised 1,25(OH)2D3 status and can modulate specifically PTH/PTHrp-mediated cAMP formation at the postreceptor level.


DIALYSIS PATIENTS CAN sometimes exhibit aluminum (Al)-dependent renal osteodystrophy characterized by decreased bone matrix deposition.1 The role of hormonal status in the development of this condition remains controversial, although it is known that dialysis patients exhibit low 1,25-dihydroxyvitamin D3(1,25(OH)2D3) levels due to reduced 1-α-hydroxylation of 25-hydroxyvitamin D3 by the kidney under these conditions.2 Osteocalcin (OC) is a major bone matrix protein produced by osteoblasts.3 Normal states such as adolescent growth and fracture healing, representing high activity of bone remodeling, and diseased states such as osteoporosis or Paget's disease, which reflect high bone turnover, show increased serum OC. OC has been implicated in bone matrix mineralization4 or the regulation of bone resorption.5 In renal osteodystrophy, serum OC is elevated,6 as a result either of increased turnover or decreased renal excretion.7

In uremic patients undergoing dialysis, plasma levels of Al can be above 100 μg/l (>3.7 μM), whereas levels of 0.37 μM can normally be detected in human serum.8 However, the range of “toxic” Al levels is somewhat narrow because some patients may experience toxicity at doses as low as 1 μM, hence only 3-fold above the “normal” level. A number of in vivo studies have pointed out that vitamin D3 status may be a contributing factor to the deleterious effects of Al in patients, since not all patients with high Al levels develop Al-related renal osteodystrophy.9–11 In one study of dialysis patients undergoing renal transplant, OC levels were reported to be high but Al levels were <100 μg/l in all patients except one, who also had the lowest OC levels (closest to control values).12 Hence, the modulation of OC levels in the serum of uremic patients may not be attributed solely to dialysis, the 1,25(OH)2D3 status of these patients, or their Al contamination level. More likely, a combination of all these conditions prevails in Al-related renal osteodystrophy.

Al may also affect the production of cAMP in osteoblast-like cells as it does in other cells.13–15 Limited information concerning this point is available for osteoblast-like cells. A few studies have indicated an inhibitory effect of Al on parathyroid hormone (PTH)-dependent cAMP formation by bone cells,16–18 yet the mechanisms that regulate this inhibition are still unknown. Whether Al has a general inhibitory effect on cAMP formation in response to various effectors or whether this inhibition is restricted to the PTH-dependent pathway is still not known. We studied the modulation of 1,25(OH)2D3-induced OC release, and PTH-, PTH related peptide (PTHrp)- and Isoproterenol (Iso)-dependent cAMP formation in the osteosarcoma cell line ROS 17/2.8. We observed that the deleterious effects of Al on OC secretion is related to the levels of 1,25(OH)2D3 applied, whereas the inhibitory action of Al on cAMP production is seemingly limited to PTH/PTHrp signal transduction with no effect on the action of Iso. The inhibition of PTH signal transduction is most likely at the postreceptor level and does not directly affect the adenylate cyclase system.


Culture conditions

The rat osteosarcoma cell line ROS 17/2.8 (ROS) was grown in 5% CO2 at 37°C in Ham's F12 containing 5% heat inactivated fetal calf serum (FCS) and 0.1 mg/ml kanamycin buffered with 25 mM HCO3, and 15 mM HEPES and NaOH to pH 7.4. Cells were harvested with 0.025% trypsin in calcium-magnesium free phosphate-buffered saline containing 1 mM EDTA, then seeded at a 1:4 dilution estimated by surface area. Cells reached confluency after 3 days with one medium change on day 2. On day 3, the growth medium was replaced with 2% heat-inactivated FCS containing 1 μM of an Al-citrate solution (1/1, termed Al throughout the text) or the vehicle (1 μM citrate solution) and the pH adjusted to 7.4 in both cases. In some cases, 10 μM Al was also tested.

Osteocalcin and alkaline phosphatase

After 2 day of treatment with or without Al, confluent cells were incubated on day 5 with Ham's F12 containing 2% heat inactivated FCS, 50 μg/ml ascorbic acid, 10−8 M menadione K3 (Sigma Chemical Co., St. Louis, MO, U.S.A.), in the presence of 1,25(OH)2D3 or the vehicle. At the end of the incubation, the supernatant was recuperated and frozen at −80°C. OC was determined on aliquots of these incubation media stored at −80°C using a radioimmunoassay (RIA) for measurement of rat OC. RIA kits were obtained from Biomedical Technologies Inc. (Stoughton, MA, U.S.A.). Standards curves were generated using purified rat OC. Cells obtained at the end of the incubation were solubilized, and alkaline phosphatase (ALP) activity was determined on aliquots following a procedure we previously described.19–21

Collagen production

Cells on day three were treated as above in the presence (1 or 10 μM Al) or absence of Al (control) for the last 48 h of culture. After 24 h of culture in the presence/absence of Al, the media was changed for the same in the presence of [3H]proline for the last 24 h of culture. At the end of the incubation, cells were scraped off the dishes after two washings with phosphate-buffered salt solution (PBSS) pH 7.4 and resuspended in 0.5 M acetic acid. The cells were homogenized by disruption with 10 strokes of a Potter on ice. Collagen was evaluated by the collagenase digestion procedure of Peterkofsky and Diegelman.22 Briefly, aliquots of total cell lysates were incubated in the collagenase buffer (120 mM HEPES, pH 7.4, 0.5 mM CaCl2, 2.5 mM N-ethylmaleimide, 400 μg/ml bovine serum albumin and 50 μg/ml high purity collagenase [Worthington Biochemical Corp., Freehold, NJ, U.S.A.) for 3 h at 37°C to digest collagen. After stopping the reaction with trichloracetic acid 10%, the proteins were spun down, and aliquots of the supernatant counted in an LKB counter (collagen digest: CDP). The pellets, representing proline incorporation into noncollagenous proteins, were solubilized and counted. The ratio of collagen to noncollagenous proteins is corrected for the relative abundance of proline in collagen (ratio 4.1) as compared with other proteins.23

cAMP formation

Confluent cells (in 35 mm Petri dishes) on day 3 were washed twice with PBSS, and preincubated for 10 minutes at 37°C in Ham's F12 medium without HCO3, containing 0.5% bovine serum albumin and 1 mM 3-isobutyl-1-methylxanthine. This medium was replaced with an identical medium containing PTH, and the incubation was stopped after 10 minutes with 3% perchloric acid (final concentration). The dishes were agitated for 60 minutes at 5°C to solubilize the cells. Cell homogenates were transferred to test tubes and centrifuged for 5 minutes at 1000g. cAMP was determined on the supernatant by RIA. Materials for the RIA were purchased from Immuno Nuclear Corp. Al treatment (0 or 1 μM) lasted for 4 h prior to PTH stimulation (0.5–50 nM), PTHrp (50 nM), or Iso (100 nM). Cholera toxin (CT; 10 μg/ml) or Pertussis toxin (PT; 1 μg/ml) were added together with or without Al for the last 4 h in culture. Cells were then treated as above. Stimulation by Forskolin was performed at 0.1 and 1 μM after the treatement with or without Al for 4 h and following the same procedure as described above.

Northern blot analysis of PTH receptor

Confluent ROS cells in T75 flasks (3 flasks per condition) were treated with or without Al for 2 days in Ham's F12 containing 1 or 2% heat-inactivated FCS. At the end of the incubation, cells were washed twice with PBSS and scraped into a guanidium-isothiocyanate lysing solution. Total RNA was extracted by the method of Chomczynski and Sacchi.24 RNA samples were size fractionated on 1.0% agarose gel containing 1.2 M formaldehyde and transferred to nylon membranes (Nytran, Keene, NH, U.S.A.). The membranes were UV autocross-linked, prehybridized at 42°C for 20 h, and hybridized at 42°C overnight with 106 cpm/ml of32P-labeled cDNA probes. cDNA probe for PTH receptor was generously provided by Dr. M. Levine at Johns Hopkins School of Medicine in Baltimore, Maryland. The cDNA probe for PTH receptor was labeled with [α-32P]dCTP using a random-primed DNA labeling kit (New England Nuclear, Boston, MA, U.S.A.). Filters were washed once for 60 minutes in 2× SSC/1× Denhardt's solution at 42°C, twice in 1× SSC/0.1% SDS for 30 minutes each at 65°C, and 15 minutes in 0.1× SSC/1% SDS at room temperature prior to exposure to Kodak X-Omat films for 24 h at −80°C. Hybridizing signals on the blots were analyzed quantitatively by densitometric scanning of autoradiograms. The filters were reprobed with cDNA for actin to monitor loading between samples.

Western blot analysis of G proteins

Confluent ROS cells in T25 flasks (three flasks per condition) were washed twice with PBSS and scraped into the same buffer as used for ALP determinations. The cell suspension was then diluted in 2× electrophoresis buffer made of 8% sodium dodecyl sulfate (SDS), 24% glycerol, 100 mM Trizma base pH 6.8, 4% β-mercaptoethanol and 0.02% bromophenol blue. SDS-polyacrylamide gel electrophoresis (PAGE) was performed in 10% acrylamide gel for 150 minutes at 100 V. Following SDS-PAGE, proteins were electrotransferred for 50 minutes at 150mA (constant current) in a semidry electroblot apparatus on nitrocellulose. Western blots were then performed in the presence of a polyclonal rabbit anti-Gsα subunit or a polyclonal rabbit antibody that recognizes both the Giα-1 and Giα-2 subunits (Calbiochem, La Jolla, CA, U.S.A.) at a final dilution of 1:1000. This antibody was then detected by horseradish peroxydase–coupled sheep anti-rabbit immunoglobulin G's (dilution 1:1000), and the protein bands detected by chemiluminescence using Luminol (Boehringher-Mannheim, Laval, Québec, Canada). Specific bands were then visualized using Kodak X-Omat films, and scanning densitometry was performed directly on these films.


All determinations were performed in triplicate for individual experiments, and each condition was tested four to eight times. Analyses of variance (ANOVA) were performed using a computerized program on Macintosh software where indicated. Student's t-tests were performed for the other determinations. A level of p < 0.05 was considered significant.


Alkaline phosphatase

Figure 1 shows the activity of ALP of ROS cells in response to 1,25(OH)2D3 and Al (1 μM). ALP activity was enhanced in response to increasing 1,25(OH)2D3 doses. Al caused an inhibition of this activity, dependent on 1,25(OH)2D3 doses applied. Indeed, in the absence of 1,25(OH)2D3 or in the presence of low concentrations of the steroid (<1 nM), Al partly inhibited the activity of ALP. In contrast, at a concentration of 1 nM of the steroid and higher, the inhibitory effect of Al was already curbed (Fig. 1).

Figure FIG. 1.

Effect of Al on ALP activity. Confluent ROS cells were incubated in the presence or absence of 1 μM Al with increasing doses of 1,25(OH)2D3 for 48 h. At the end of the incubation, cells were washed twice, solubilized, and ALP activity determined on aliquots. Results are the mean ± SEM of six preparations run in triplicate. p < 0.001 by ANOVA for the dose response, followed by Student's t-test for individual subsets: *p < 0.001, **p < 0.02 vs. the same treatment without Al.

Osteocalcin release

OC release by ROS cells is already high in the absence of 1,25(OH)2D3, but this activity can be stimulated about 3- to 4-fold in the presence of increasing doses of the steroid (Fig. 2). Al inhibited basal OC release by ROS cells (p < 0.001). Al also inhibited the 1,25(OH)2D3-induced increase in OC release by ROS cells (25–40%) at low doses of the steroid (0.5 and 1 nM; Fig. 2). In contrast, in the presence of higher doses of the steroid (5–10 nM), the inhibition by Al was prevented (Fig. 2).

Figure FIG. 2.

Effect of Al on OC secretion by ROS cells. Cells were treated as in Fig. 1, except that the supernatant was collected and frozen at –80°C. Aliquots of the supernatants were used to measure the amount of OC released by ROS using a RIA. Values are the mean± SEM of six experiments run in triplicate. p < 0.001 by ANOVA for the dose response, followed by Student's t-test for individual subsets: *p < 0.001, **p < 0.025 vs. the same treatment without Al.

Collagen production

Under basal conditions, collagen production by ROS cells was about 11.5 ± 0.4% ( Table 1). Collagen production was inhibited partly by 1,25(OH)2D3 (5 nM), a situation normally observed in osteoblast-like cells.17 In addition, Al at low doses (1 μM) inhibited collagen production by ROS cells, an effect that was also sustained at high doses of Al (10 μM; Table 1). The presence of 1,25(OH)2D3 and 1 μM Al further inhibited collagen production as compared with either 1,25(OH)2D3 or Al alone. Paradoxically, the addition of 10 μM Al in the presence of 1,25(OH)2D3 did not inhibit collagen production further than with 1,25-(OH)2D3 alone (Table 1).

Table Table 1. Effect of Aluminum on Collagen Production by ROS Cells
original image

Production of cAMP

Response to PTH, PTHrp, Iso, and Forskolin:

ROS cells responded to PTH with a dose-dependent increase in cAMP production. The production of cAMP was stimulated about 50-fold in the presence of 50 nM PTH (Fig. 3). This stimulation of cAMP formation was inhibited by 40–60% in the presence of 1 μM Al according to applied PTH doses (p < 0.001 by ANOVA). PTH-related peptide (50 nM) also stimulated cAMP formation in ROS cells to levels comparable to PTH, a process that was also inhibited (50%) by 1 μM Al (p < 0.001; Fig. 3). In contrast, although Iso (100 nM) fully stimulated cAMP to levels similar as for PTH and PTHrp, Al treatments had no effect on this response (Fig. 3). The exact mechanism by which Al can exert its inhibitory effect on cAMP formation in response to PTH and PTHrp was further investigated by studying the direct activation of adenylate cyclase by Forskolin. In the presence of 0.1 and 1 μM Forskolin, ROS cells stimulated cAMP production 7- and 30-fold, respectively (Fig. 4). Al had no effect on this activity, indicating that direct stimulation of adenylate cyclase prevented the deleterious effects of Al on cAMP production.

Figure FIG. 3.

Effect of Al on PTH-, PTHrp-, and Iso-induced cAMP formation by ROS cells. Confluent ROS cells were incubated for 4 h in the presence or absence of 1 μM Al prior to testing their responsiveness to these agents. PTH (0.5, 5, and 50 nM), PTHrp (50 nM), or Iso (100 nM) was applied for 10 minutes to cells treated or not with 1 μM Al. The incubation was stopped by 3% perchloric acid (final concentration), and cAMP determined by a RIA. Values are the mean ± SEM of four to eight experiments run in triplicate. p < 0.001 by ANOVA for the dose-response curve to PTH. Student's t-test for individual subsets: *p < 0.001, **p < 0.005, and ***p < 0.02 between respective controls and Al treatment.

Figure FIG. 4.

Effect of Al on Forskolin-dependent cAMP formation, and combined effect of Al and toxins on PTH-dependent cAMP formation by ROS cells. Confluent ROS cells treated or not with Al for the last 4 h of cultures were incubated with increasing doses of Forskolin for 10 minutes, or cells were also incubated for the last 4 h of culture with or without 1 μM Al, in the presence or absence of CT (10 μg/ml) or PT (1 μg/ml). At the end of this incubation, cells were challenged with 50 nM PTH. cAMP formation by ROS cells was then determined by RIA. Values are the mean ± SEM of four determinations run in triplicate. Statistical analysis by Student's t-test: *p < 0.001 vs. respective PTH treatment, and **p < 0.025 vs. PTH + Al.

Effect of CT and PT

To further discriminate the possible mechanism of action of Al on cAMP formation by osteoblast-like ROS cells, we studied its effect on toxin-dependent cAMP formation. CT stimulated the effect of PTH (50 nM) about 2-fold (p < 0.001), a process that was not inhibited by Al treatment (Fig. 4). In fact, although Al inhibited the response to PTH alone, in the presence of Al, CT, and PTH the formation of cAMP was slightly higher (but not significantly) than for CT and PTH alone. PT alone did not modify the response to PTH (50 nM) in ROS cells (Fig. 4). Nevertheless, although Al inhibited PTH-dependent cAMP formation by ROS cells, PT treatments partly prevented this inhibition by Al (Fig. 4). Hence, a protective effect of PT was observed (p < 0.025) in the presence of 1 μM Al. Indeed, cAMP formation under these conditions reached similar levels as in cells treated with PT alone in the absence of Al (Fig. 4).

Effect of Al on PTH receptor mRNA expression and G protein subunits

Al treatments did not significantly alter PTH receptor mRNA levels as detected by Northern blot analyses whether cells were treated in the presence of 1 or 2% FBS (Fig. 5). However, Western blot analyses did reveal differences in G proteins between Al-treated ROS cells and controls. Whereas Al did not significantly reduce the number of Gsα subunits in ROS cells even after 48 h of treatment, the same treatment significantly reduced the levels of Giα subunits (Fig. 6). Indeed, both the heavy and light protein bands of Gsα subunits were not modified by Al treatments: 728.9 ± 65.5 vs. 731.8 ± 125.1 and 383.8 ± 84 vs. 511.5 ± 65 (n = 3, relative densitometric units) for the heavy and light bands of control- and Al-treated cells, respectively. However, the Giα band was reduced by Al treatments: 512.5 ± 86 vs. 247.1 ± 31.8 (n = 3, relative densitometric units; p < 0.05).

Figure FIG. 5.

Northern blot analysis of PTH receptor mRNA. Total RNA was extracted from one T75 flask per condition after 2 days of treatment with or without 1 or 10 μM Al, in the presence of 1 or 2% FBS. The autoradiography of32P(ATP)-PTH probe is shown on the top portion, whereas the bottom portion of the figure shows methylene blue stain of RNA transferred to nylon filter.

Figure FIG. 6.

Western blot analysis of Gsα, Giα-1, and Giα-2 proteins in ROS cells. Total cellular proteins were solubilized from one T25 flask per condition after 2 days of treatment with or without 1 μM Al, in the presence of 2% FBS. After SDS-PAGE and electrotransfer to nitrocellulose filters, specific antibodies to Gsα, Giα-1, and Giα-2 subunits were used to probe the filters, followed by detection with a second antibody coupled to horseradish peroxidase and chemilunimescence detection with luminol. The top row shows the detection of Gsα subunits, and the bottom row shows the detection of Giα-1 and Giα-2 subunits.


The present study shows that the deleterious effects of Al on bone metabolism is pleiotropic and is related to the hormonal status of osteoblasts. The results show that vitamin D-dependent effects were inhibited by Al only when vitamin D levels were low. Indeed, both ALP and OC release were inhibited by Al only under low vitamin D3 conditions, but escaped Al inhibition as the concentration of the steroid was increased to pharmacological doses in vitro. Previous studies failed to investigate this effect and focused on the dose-response effect of Al under basal conditions or in the presence of only one concentration of 1,25(OH)2D3,15,25–30 although Coen et al.26 suggested that 1,25(OH)2D3 may have protective effects on bone exposed to Al. Hence, the deleterious effects of Al may be preceded by a compromised 1,25(OH)2D3 status, and may explain, at least in part, the variable pathological manifestations in patients showing similar plasma Al levels.

Although we observed that the maximal binding capacity of 1,25(OH)2D3 receptors is reduced by Al in the osteosarcoma cell line MG-63 cells (data not shown), in ROS cells saturating doses of 1,25(OH)2D3 were able to totally reverse the deleterious effects of Al. Kasai et al.31 have shown that Al and Al + transferrin can inhibit osteoblast-like cell growth, suggesting that Al could also exert its deleterious effects on the recruitment and/or proliferation of osteoblasts. Rodriguez et al.32 showed that Al is toxic for rat osteoblasts in vivo without any decrease in osteoblast number, whereas PTH can increase osteoblast surface without improving osteoid mineralization in the presence of Al. Our own results did not show any significant effect on cell proliferation because confluent cells were used for all assays and no significant differences in protein and/or cell number were noted at the end of incubations (data not shown).

Al may exert its toxic effect through an inhibition of bone formation since we an others17 observed that Al also inhibits collagen synthesis. The effect of Al, at doses that can be found in dialysis patients (1–10 μM), reproduced the inhibitory effect of 1,25(OH)2D3 on collagen synthesis. It is important to note, however, that the inhibition of collagen synthesis was enhanced in the presence of 1,25(OH)2D3 and low doses of Al, whereas higher doses of Al did not lead to further decrease in collagen production by ROS cells. In contrast, high doses of 1,25(OH)2D3, although inhibiting collagen synthesis alone, protected against further reduction in synthesis due to Al. Together with the above decrease in ALP activity and OC secretion, these results further indicate that 1,25(OH)2D3 may play a protective role against Al-induced toxicity in osteoblasts. Such a conclusion was also suggested previously from in vivo and in vitro studies.25,26,29,33,34

cAMP formation in ROS cells is very sensitive to Al. It has been previously suggested that very high PTH levels may have a protective effect against Al intoxication in bone cells, because DNA synthesis in bone cells is higher than normal in patients with severe secondary hyperparathyroidism despite Al loading.35 Our results do not indicate whether ROS cells could have responded to elevated PTH levels with increased DNA synthesis but suggest that Al curtails the elaboration of one class of second messengers in response to PTH in these cells. Indeeed, Al always inhibited cAMP formation in response to PTH regardless of hormone concentration, whereas Al did not totally prevent the PTH-induced, dose-dependent increase in cAMP formation. The restricted inhibitory activity of Al on cAMP formation by PTH and PTHrp, but not by Iso, indicate that Al may exert its effect via a specific receptor and/or postreceptor pathway(s). Indeed, Al was without effect on cAMP formation by Forskolin, which bypasses both the receptor and the regulatory G proteins involved in signal transduction (Gs and Gi).36

A specific effect of Al on PTH receptor levels is unlikely since PTH receptors, as assessed by Northern blot analysis, were not affected by Al in ROS cells, unless mRNA levels are not predictive of protein content in these cells. In contrast, Pun et al.18 reported altered binding of125I-PTH in osteoblast-like UMR cells in response to high doses of Al (>4 μM), reflecting reduced PTH receptor content. It may then be argued that high serum levels of PTH may counteract the inhibitory effects of Al by simply increasing total cAMP formation. The observation that saturating doses of PTH (50 nM) failed to reverse the effect of Al could argue against a direct interaction of Al and the PTH receptor, unless Al irreversibly damages the PTH receptor. This hypothesis would agree with the observations of Pun et al.18 who reported that Al inhibited both cAMP formation and PTH binding to UMR cells. However, the lowest Al dose that significantly affected UMR cells was 4 μM, four times the concentrations we showed to influence cAMP formation in ROS cells. In addition, the doses that Pun et al.18 used (4, 40, and 200 μM) are much higher than what has been found in vivo,8 and therefore their observations may not reflect the in vivo situation. The results shown here do not eliminate the possibility that Al could modulate the turnover/recycling of PTH receptors. Moreover, the deleterious effect of Al was restricted to PTH- and PTHrp-dependent cAMP formation, without any effect on Iso-dependent cAMP formation. This would then indicate that only the PTH/PTHrp cAMP signaling pathway is affected by Al, suggesting that the receptor interaction was perturbated or that specific, post-receptor interactions (PTH/PTHrp receptor-specific G proteins interaction) are the target for the deleterious effect of Al. This hypothesis would also refute the reduction in PTH binding in response to Al observed by Pun et al.18

CT, which stimulates the adenylate cyclase system by ribosylating a Gs protein, totally abolished the Al-dependent inhibition of cAMP formation. This indicates that Al might exert its inhibitory effect via this Gs protein, by modulating the GTPase activity of this protein or by competing at the site of rybosylation, in which case CT treatment could prevent Al intoxication. Al may also inhibit cAMP formation by increasing a Gi protein subunit or Gi protein activity, which would reduce the adenylate cyclase activation. The implication of a regulating Gi protein in PTH-dependent cAMP formation by ROS cells does not appear to be very likely since PT was without any significant effect on cAMP formation when cells were incubated with this toxin alone. A PT-induced enhancement of PTH-dependent cAMP formation in ROS cells has been previously reported by Abou-Samra et al.37; however, their cells were preincubated for 3 days with the toxin to obtain this result, whereas we incubated ROS cells for 4 h with the toxin, which could explain the difference between the two studies. Although PT did not alter PTH-dependent cAMP formation under our experimental conditions, it reduced the inhibitory effect of Al. This indicates that Al may reduce PTH-dependent cAMP formation via the stimulation of a Gi protein.

The combined results of CT and PT studies strongly suggest that either Al alters the ribosylation of G proteins, a process that usually stimulates Gs and inhibits Gi proteins,38 or else acts at yet another site, which prevents the activity of G proteins. Perturbated PTH receptor-mediated functions can be overcome in ROS cells by CT treatments, via direct stimulation of Gs protein subunits. Indeed, Gs and Gi proteins rybosylation can overcome Al toxicity in ROS cells. It is also noteworthy that our Western blot analyses of Gs and Gi subunits indicate no significant modifications in Gs subunits in response to Al, yet indicate a decrease in Gi subunits. Finally, a direct contribution of G proteins in the case of Al intoxication into an aberrant cell response is plausible if Al affects specific subsets of G proteins because other hormone signaling pathways such as Iso-induced cAMP formation was not perturbated by Al in ROS cells.

In conclusion, the present results indicate that the deleterious effects of Al in vitro are related to the 1,25(OH)2D3 status, which would translate in vivo with a more toxic effect of Al in patients with already compromised 1,25(OH)2D3 status. This dialysis contaminant also inhibits PTH-dependent cAMP formation in osteoblast-like cells, an effect that can be overcome via G protein modulation.


R.M. was supported by a studentship from the Guy-Bernier Research Center. D.L. is a Senior Scholar from the “Fonds de la Recherche en Santé du Québec.” We thank Mrs. Lucie Simoneau for her expert technical assistance. This work was supported in part by a grant from the Kidney Foundation of Canada to D.L. and by National Institutes of Health grant DK43423 to S.E.G.