This work presents a study of the receptors for the gut hormones GIP, GLP-1, GLP-2, GHR and OB, in three osteoblastic cell lines. The chosen in vitro model represents different stages of osteoblastic development and has been valuable in analysing the patterns of expression of the genes of interest in different levels of osteoblastic maturity. Although investigation of cell lines may differ in some responses from naturally occurring primary cells the use of cell lines represents a suitable model to study determined characteristics in a specific stage of osteoblast maturity as an initial observation.
COL1A2, ALP and OPG mRNA expression were used as bone markers that are correlated with osteoblastic differentiation. It has been shown that type 1 collagen mRNA signal increases according to the degree of differentiation in cultured rat calvaria cells . ALP is associated with the phase of differentiation and stabilisation of mineralised matrix ; OPG mRNA expression has been found to be increased during matrix production and maturation in rat calvaria cultures . The results confirmed that TE-85 and Saos-2 had the highest levels of relative expression for COL1A2 compared to MG-63. ALP mRNA also showed its highest expression in Saos-2 and lowest in MG-63. OPG mRNA levels were similar in TE-85 and Saos-2, (2.5 and 2.6 fold respectively), and MG-63 expressed the lowest levels. This suggests that Saos-2 is the most mature; MG-63 is the least mature of the cell lines and TE-85 intermediate between the other two.
GIP-R has been described in chondrocytes, osteocytes, osteoblasts and in two of the cell lines studied here (MG-63 and Saos-2) [7, 20], however the novelty of this current work is the profile of expression for those receptors which may be evident in different stages of maturity and correlated with the maturity of the cell lines. MG-63 and Saos-2 express GHS-R but no visual differences were demonstrated in earlier qualitative PCR screening . Also GHS-R peaked three days after induction of osteoblastic differentiation of a murine cell line and the expression decreased with time, i.e. in more mature osteoblasts. In the current work GHS-R expression was greatest in the cell lines representing early osteoblasts (MG-63 and TE-85), while the most mature showed the lowest expression, in keeping with the time course of expression shown in the previous report. GHR has been shown to induce direct responses in bone cells through its receptor, in several different ways: differentiation, proliferation and/or viability [19–21].
GLP-1R had not previously been reported in human osteoblasts, or in osteoblastic cell lines like MG-63 or Saos-2 [7, 12] but lately a study detected the presence of a functional receptor in a murine cell line, independent of the cAMP linked receptor  which is different from the human one we described in this paper. We indeed confirmed that Saos-2 cells do not express the receptor we looked for. However, TE-85 and MG-63 were positive for GLP-1R. The reasons for the discrepant results in MG-63 are not clear, but an explanation for it may lay in the size difference of the amplification template: smaller templates show more efficiency when they are amplified [34, 35]. In the present work, the size of the template was much smaller (114 bp) than the size used in the previous report (695 bp), because qPCR calls for small templates. One study  showed that in a knockout animal model lacking GLP-1R, cortical BMD at tibia and lumbar spine was significantly reduced. They also established the lack of effects of GLP-1 treatment on Saos-2 cultures, which is in agreement with the current study demonstrating lack of GLP-1R.
The authors proposed that GLP-1R is essential in the control of bone resorption indirectly, since this receptor is expressed in thyroid C cells, and GLP-1 is able to stimulate calcitonin production in those cells with calcitonin receptor being able to inhibit osteoclastic bone resorption. No studies have confirmed GLP-2R presence either in osteoblastic or osteoclastic cells, and only two short reports mention its existence in bone-related cells [36, 37], in contrast, the effects on bone resorption after administration of the peptide in clinical trials have been well documented [17, 38, 39]. GPR39 is believed to be the receptor for obestatin [22, 40], but some reports have questioned that possibility [24, 41, 42]. A study demonstrated that transformed human embryonic kidney cells (293T) when transfected with a plasmid encoding human or mouse GPR39, but not GHS-R, exhibited high-affinity binding to an analogue of obestatin (monoiodobestatin) . These receptors have not been described in bone-related cells, and the only record is listed in GenBank reporting its presence in a chondrosarcoma cell line (CA749039.1). In this current work PCR reactions showed positive signals for the receptor however the relationship between this receptor, obestatin and their influence on bone should be interpreted with caution.
Regarding functional responses, Saos-2 exhibited the most significant changes following exposure to GIP in terms of increases in cell viability, ALP secretion and P1NP production, according to the level of expression for this receptor. GHR and OB also induced significantly higher levels of ALP activity in Saos-2. In addition, it has been demonstrated that GIP and GHR, are able to induce functional responses in osteoblast-like cells , or to promote bone formation [19, 21]. These observations are consistent with the findings in this study, showing responses at the concentrations from 10-12 to 10-9M. This is relevant in relation to the physiological response in vivo as these concentrations are equivalent to the plasma levels that these hormones reach after meals [38, 43, 44]. However, a cautious note on this finding is the fact that a significant change in ALP activity is an indicator that a change in bone cell activity is taking place, and not only suggestive of bone formation as it has been demonstrated that increased values are associated with bone loss in aging, osteoporosis and menopause . Levels of this enzyme decrease with therapy antiresorptive agents , but can increase again in response to anabolic treatments like teriparatide (recombinant form of PTH) .
MG-63 and TE-85, representing younger osteoblasts, showed significant higher levels of cell viability when the cell cultures were incubated in the presence of GLP-1 or GLP-2, as they showed higher expressions for these receptors. However they did not display any significant changes in terms of ALP secretion and TE-85 only responded to GHR at its highest concentration (10-9M). This lack of response might be attributable to lower ALP synthesis (as assessed by the relative level of ALP mRNA expression) rather than an absence of stimulation. No major changes were demonstrated for Saos-2, confirming the lack of the receptors in these cells.
Decreased P1NP concentrations were observed in supernatants from early osteoblasts-like (MG-63 and TE-85) after long exposure to GIP, GLP-1, GLP-2, OB in TE-85 and after GLP-1 in MG-63 cells. Short exposure did not cause significant changes and a sustained stimulation displayed a negative effect. The possibility that this pattern was caused because cell health was compromised can be rejected, as the viability assay did not show any significant cell death after long exposure to the peptides. Thus the changes in P1NP production, either by P1NP degradation or a decreased production of collagen, by the cell lines were caused by the presence of the peptides. There are no studies reporting changes in P1NP production by cells in vitro in response to these peptides and this is the first study to examine this bone marker in these cell lines. Conversely, Saos-2 cells showed significant increases of P1NP levels following GIP and OB. This is consistent with the studies that show GIP stimulates bone formation [5, 9, 48, 49]. These opposite effects in response to GIP, increased levels in Saos-2 and decreased concentrations in TE-85, can be related to other findings which after prolonged exposure to high concentrations of GIP there was downregulation of the GIP-R in Saos-2 cells . This finding may underpin the results observed in TE-85 and may account for the modest decrease of P1NP observed in MG-63 at the highest GIP concentration. On one hand GIP can promote bone formation (which is consistent with the results shown here for Saos-2 in terms of ALP and P1NP increases), but also a downregulation of the receptor might be a feasible explanation for the decreasing levels found in TE-85 after longer exposure to higher concentrations of the peptide.
It has been repeatedly reported that GLP-1R is not present in osteoblasts, but it should be noted that the published data refer mainly to the studies performed using Saos-2, which we confirmed do not express the receptor. TE-85 and MG-63 express the receptor and exhibited some significant responses to this peptide in the viability assay, and in P1NP secretion after long exposure to GLP-1. These observations support previous studies which show the role of GLP-1R in the mediation of bone resorption  and the possible association of GLP-1 treatment with the improvement in bone disorders linked to glucose intolerance . While the reports linking GLP-1 to bone activity are few, there are more studies showing effects of GLP-2 in bone, and all of them found reductions in bone resorption markers, in clinical trials [17, 38, 39, 50]. In the current work, the responses to GLP-2 were observed in TE-85 and MG-63, with significant decreased secretion of P1NP and OC. Most clinical trials have not shown any significant difference in the levels of those markers after treatment with GLP-2 [16, 39, 50]. Here, GLP-2 produced a significant response in TE-85 decreasing P1NP and in MG-63 caused a significant increase after 48 h and a significant decrease after 120 h. No regulation of the receptor has been suggested in previous reports.
GHR can induce responses from bone-related cells in vitro[19–21]. We confirmed these findings and a significant increase of ALP was found in Saos-2 and TE-85 but no significant changes were found at any time point for P1NP or OC. We regard that GHR may be an important modulator in the cells with a higher degree of differentiation, although the expression was more important in MG-63, these cells did not respond significantly to this peptide in any of the tests. In agreement with the latter experimental data, there is a clinical study reporting no significant effects of a GHR infusion on plasma CTX or P1NP when healthy and post-gastrectomy patients were studied  but an inverse relationship existed between baseline plasma GHR and CTX, suggesting that GHR may have a role regulating bone resorption.
OB caused significant ALP and P1NP increases in Saos-2, but TE-85 displayed decreased responses. The only data on OB effects is a short report where a human chondrocyte cell line, C28-I2, was studied and no changes were found in ALP production . Some contradictory results have been reported where pancreatic beta cells displayed higher levels of survival after OB treatment  and no effect on cadiomyocytes viability was observed .
These paradoxical results, also observed in some of the clinical trials, may be explained by the nature of the receptor for these ligands which belong to the G protein coupled receptors and may be desensitised after long exposure to the ligand. In the case of GIP in Saos-2 a significant increase with the lowest concentration, might be explained by changes in receptor expression. Because the receptors for the hormones studied are GPCR, they can be subjected to some desensitization or rapid attenuation of receptor sensitivity after exposure to agonists .