Calcium-Sensing Receptor Function in the Skeleton: Alternative Interpretations
The current Research Article by Chang et al. (1) addresses a long-standing question regarding the function of the extracellular calcium-sensing receptor (CaSR) in the skeleton. It is well known that the major physiologic function of the CaSR is to regulate systemic calcium homeostasis, through its actions in the parathyroid gland to regulate parathyroid hormone (PTH) secretion and in the kidney to regulate renal calcium handling. The presence of inactivating mutations of Casr in familial hypocalciuric hypercalcemia (FHH) and neonatal severe hyperparathyroidism (NSHPT) provide evidence for this (2). However, there is ongoing debate as to whether activating mutations cause hypoparathyroidism (3) and targeted ablation of exon 5 of Casr in mice cause hyperparathyroidism similar to FHH and NSHPT (2), and the role of the CaSR in the skeleton, where it is expressed at low abundance, is unclear (4-12).
For example, it has been difficult to unequivocally establish that the CaSR is the relevant calcium receptor in osteoblasts, because the original Casr knockout mice fail to display any growth plate or bone phenotype when the concurrent hyperparathyroidism is corrected either by a “molecular parathyroidectomy” created by crossing glial cells from Gcm2-deficient mice onto the Casr null background (4) or by concomitant ablation of the PTH gene in Casr null mice (5). Moreover, osteoblasts derived from Casr null mice retain their full ability to sense extracellular calcium ex vivo (6). In addition, calcimimetics that target the CaSR and are used in the treatment of hyperparathyroidism have no reported effect on bone density.
In contrast, Chang et al. (1) generated a mouse with lox P sites flanking exon 7 of Casr, which encodes the seven transmembrane domains and C-terminal tail and used the 2.3-kb Col(I) α1 subunit promoter [2.3Col(I)-Cre] and OSX promoter (OSX-Cre) to produce osteoblast-specific deletion of Casr, and the type II Col promoter [Col(II)-Cre] to produce chondrocyte-specific deletion of Casr in mice. Curiously, all mouse models lacking the seven-transmembrane and C-terminal regions of CaSR, but continued production of the extracellular domain of the CaSR encoded by exons 2 to 6, developed a skeletal phenotype characterized by growth retardation, skeletal dysplasia, and defective mineralization. The authors interpret these results to indicate an important role for the CaSR in chondrocyte and osteoblastic function (1).
How do we reconcile the lack of a skeletal phenotype in the original Casr null mice after accounting for excess PTH (4, 5) with the profound skeletal phenotypes in the osteoblast- and chondrocyte-specific Casr knockout mice? The lack of a skeletal phenotype in the original “CaSR knockout mouse” after rescue of excess PTH signaling has been attributed to the presence of an alternatively spliced product lacking exon 5 in these mice (2, 7, 8) that is purported to remain functional in bone but not in the parathyroid gland or kidney (9); however, this in-frame deletion of 231 nucleotides of exon 5 of this alternatively spliced product results in loss of function in vitro (8, 10). Alternatively, it is more likely that the approach used to conditionally ablate CaSR produced a secreted extracellular domain of CaSR capable of acting as a tissue nonspecific, dominant-negative factor to disrupt CaSR function. If so, the skeletal abnormalities in the conditional Casr null mice might not necessarily arise from loss of CaSR function in osteoblasts or chondrocytes. In support of this possibility, Western blotting identified that deletion of Exon 7 resulted in the production of an alternatively spliced N-terminal CaSR protein [(1), Fig. 2D]. Although the authors demonstrated that the alternatively spliced product protein, ∆Exon7-CaSR, had no function in vitro [(1), fig. S1C], they failed to test the possibility that ∆Exon7-CaSR could be secreted and act as a dominant-negative factor interfering with CaSR function.
In unrelated studies investigating the potential biological functions of the extracellular domain of CaSR, we had designed a cDNA construct containing exon 2 through 5 (CASR.1392) to generate a secreted protein that contains the extracellular domain but lacks the transmembrane and C-terminal domains. In overexpression studies in HEK 293 cells, we confirmed that this truncated CASR.1392 is secreted into conditioned media, similar to the ∆Exon7-CASR construct reported by Chang et al. (1). However, to elucidate the function of CASR.1392, we tested whether the addition of purified CASR.1392 protein could disrupt extracellular Ca2+-mediated stimulation of CaSR function in HEK 293 cells transfected with wild-type CaSR and a SRE-luciferase reporter gene, as previously described (6). We found that the N-terminal CaSR recombinant protein at a concentration of 0.15 mg/ml significantly inhibited the ability of extracellular Ca2+ to stimulate CaSR-mediated signaling.
The ability of CASR.1392 to be secreted and disrupt CaSR function raises the possibility that some of the phenotypic features of the “conditional CaSR” mice reported by Chang et al. (1), might have been derived from indirect effect of circulating levels of this dominant-negative N-terminal CaSR. It would be important to reexamine this issue in a conditional knockout that deleted the entire Casr. Regardless, additional studies may be needed in the conditional knockout mice created by Dr. Chang and coworkers to exclude potential confounding effects of the ∆Exon7-CaSR. Consequently, the function of CaSR in the skeleton remains an open question.
Response to Quarles and Pi
In attempting to reconcile their observations on the lack of bone defects in generalized extracellular calcium-sensing receptor (CaSR) knockout (KO) mice rescued from severe hyperparathyroidism (HPT) with our findings of profound skeletal phenotypes in the osteoblast- and chondrocyte-specific CaSR knockout mice (1), Drs. Quarles and Pi propose that the extracellular domain of the CaSR (∆Exon7-CaSR), encoded by exons 2 to 6 of Casr, acts as a tissue nonspecific, dominant-negative factor to disrupt CaSR function in other cell types. They further propose that the skeletal abnormalities in our conditional Casr KO mice did not arise from the loss of CaSR function in osteoblasts or chondrocytes but from this dominant-negative effect. This is a circular argument.
How can the ∆Exon7-CaSR produced by osteoblasts or chondrocytes exert nonspecific dominant-negative effects "selectively" on irrelevant cell populations without having an impact on the osteoblasts or chondrocytes themselves? The scenario proposed by Quarles and Pi is valid only if there is a functionally critical CaSR already in osteoblasts or chondrocytes. This they have repeatedly denied. If the CaSR is not expressed in the latter cells, then which cell populations are producing ∆Exon7-CaSR and on which cell types are dominant-negative effects being exerted to produce bone- and chondrocyte-specific abnormalities? It is impossible to argue that there is no CaSR in bone and cartilage and at the same time to argue that circulating CaSR inhibitors can be both made by and act on these same cells.
The core of the argument put forward by Quarles and Pi is that ∆Exon7-CaSR exerts a dominant-negative effect. We also had considered this possibility before we adopted our gene KO strategy. We rigorously examined the signaling response of wild-type (WT) CaSR in HEK 293 cells and then cocultured these cells with cells expressing ∆Exon7-CaSR cDNA to determine if any secreted ∆Exon7-CaSR could affect the function of WT CaSR. The results confirmed that the presence of ∆Exon7-CaSR did not interfere with the function of WT CaSR [(1), fig. S1].
The argument presented to support a dominant-negative effect of the ∆Exon7-CaSR is based on the unpublished studies of the Quarles group on the protein encoded by exons 2 to 5 of Casr (CASR.1392). They found that the addition of recombinant CASR.1392 inhibited the ability of extracellular Ca2+ to stimulate signaling in HEK 293 cells that expressed WT CaSR. Because these data are, to our knowledge, unpublished, their reproducibility remains uncertain. Furthermore, the supraphysiological concentration (0.15 mg/ml) of the protein used by Quarles and Pi makes it difficult to determine whether the result is of any relevance in vivo, where the abundance of the receptor is far lower. This is particularly true for the conditions in our animal models. Our Western blot analyses of lysates of parathyroid glands indicated that the abundance of ∆Exon7-CaSR protein in the parathyroid-specific Casr KO mice was less than 10% of the abundance of full-length CaSR in WT mice (1). This reduced abundance of the ∆Exon7-CaSR protein is unlikely to support any kind of "strong" dominant-negative effect.
Finally, and most importantly, if the ∆Exon7-CaSR protein is indeed a dominant-negative factor acting on the endogenous CaSR, mice heterozygous for Casr (Casr+/-) in chondrocytes and bone cells should also demonstrate a clear skeletal phenotype. On the contrary, our Casr+/- mice grew and developed similarly to their WT littermates (1). Based on the reasoning above, it is unlikely that the skeletal phenotypes in our osteoblast- and chondrocyte-specific Casr KO mice are caused by putative dominant-negative effects of the ∆Exon7-CaSR as proposed by Quarles and Pi.
Science Signaling. ISSN 1937-9145 (online), 1945-0877 (print). Pre-2008: Science's STKE. ISSN 1525-8882