Departments of Medicine (M.J.E.) and Medical and Molecular Genetics (M.J.E., T.F.), Indiana University School of Medicine, Indianapolis, Indiana 46202
Address all correspondence and requests for reprints to: Michael J. Econs, M.D., Associate Professor of Medicine and Medical and Molecular Genetics, Director, Division of Endocrinology and Metabolism, Department of Medicine, Clinical Building, 459, 541 North Clinical Drive, Indiana University School of Medicine, Indianapolis, Indiana 46202.
Absorptive hypercalciuria is a frequent cause of nephrolithiasis, a common condition that is responsible for substantial morbidity (1). Although minimally invasive techniques have markedly reduced the morbidity of removing stones once they are formed, nephrolithiasis still presents an important clinical problem. First, these techniques do not prevent recurrence of stones. Second, most stones are spontaneously passed without therapeutic intervention, but at great pain and time lost from work. Families segregating an autosomal dominant type of absorptive hypercalciuria have been reported (2, 3). There is also a strong genetic contribution to absorptive hypercalciuria in the general population.
There are numerous successes using the positional cloning approach to identify genes for Mendelian traits as well as the promise that these techniques will lead to the identification of genes that predispose individuals to complex traits, such as osteoporosis and absorptive hypercalciuria. Clearly, it is wise to use these techniques to identify genes that are responsible for Mendelian forms of absorptive hypercalciuria, as well as those genes that predispose to nephroliathiasis in the general population. Identification of these genes will likely result in a better understanding of the pathogenesis of the disorder, an improved understanding of the role of the kidney in mineral ion homeostasis, and identification of pathways for therapeutic intervention.
In this issue of JCEM, Reed et al. (4) report the results of a mutation search in three absorptive hypercalciuria kindreds as well as a case control study in 80 patients with absorptive hypercalciuria and controls. These investigators previously performed a genome-wide linkage scan in the same three families reported in this issue of the journal (5). They found linkage to chromosome 1q23-24, with the bulk of the evidence coming from the one large kindred (5). In the current study, Reed et al. (4) report the occurrence of base pair substitutions in the soluble adenylate cyclase gene that segregate with the disease in the three kindreds. They further report that these changes were associated with absorptive hypercalciuria in a population of patients. A careful look at the evidence is necessary before concluding that these substitutions are causative for disease.
We first examine the evidence that the observed base pair substitutions in the soluble adenylate cyclase gene are responsible for autosomal dominant absorptive hypercalciuria. The investigators first examined three genes in the linked chromosome 1 region: RE2, ATP1B1, and MPZL1. No mutations or polymorphisms were detected in RE2. Two single base pair substitutions were found in ATP1B1, and one was found in MPZL1. All three were found to be single nucleotide polymorphisms (SNPs). The investigators next turned their attention to the soluble adenylate cyclase gene, which is a ubiquitously expressed gene. While the fact that the gene is expressed in most tissues tends to argue against its having a role in a Mendelian disease, whose sole manifestation is absorptive hypercalciuria, this phenomenon has clearly been observed in other disorders. For example, the LRP5 gene is ubiquitously expressed, but a missense mutation results in a bone-specific phenotype (high bone mass) whereas inactivating mutations result in osteoporosis pseudoglioma (6, 7). Using current models of calcium homeostasis, it is not immediately obvious how the soluble adenylate cyclase gene would play a role in the pathogenesis of absorptive hypercalciuria. However, one of the advantages of the positional cloning approach is that it allows the investigative team to find genes that do not fit into current models of disease pathophysiology.
Six base substitutions were found in the three chromosome 1-linked kindreds. For the most part, these changes segregated with the disease in the kindreds, as would be expected for any polymorphism in the linked region. One of these substitutions occurred in exon 7 and resulted in a change in the amino acid sequence from threonine to methionine. This exonic substitution was also identified in 4% of normal controls. In many studies, the presence of a putative disease-producing mutation in normal controls would eliminate this substitution as a causative mutation; however, because absorptive hypercalciuria is a fairly common condition, it is possible, although unlikely, that some of the controls either have absorptive hypercalciuria or are nonpenetrant carriers of the substitutions. Reed et al. (4) raise this hypothesis in their discussion. The other polymorphisms consisted of one silent exonic substitution, and the remaining were intronic substitutions. Intronic substitutions can cause Mendelian disorders by creating pseudo exons (8). However, the authors do not indicate that this is the case for absorptive hypercalciuria. Indeed, they do not postulate a mechanism to explain how these substitutions would cause disease. Thus, it is difficult to posit how the observed substitutions in the soluble adenylate cyclase gene might result in absorptive hypercalciuria.
It is, of course, possible that the identified SNPs are physically linked to an as yet unidentified mutation that causes absorptive hypercalciuria. Reed et al. (4) did not find evidence for a more obvious mutation in the soluble adenylate cyclase gene; however, the mutation could be in the regulatory elements of the gene, which are difficult to detect. Such mutations might change levels of expression. It is also possible that the observed base substitutions affect expression in another gene that is downstream of the soluble adenylate cyclase gene. In this regard, Enattah et al. (9) recently found that a base substitution in an intron in the MCM6 gene, which is upstream to the gene encoding lactase-phlorizin hydrolase, is responsible for adult type hypolactasia. The C allele, which is associated with lactose intolerance, creates a binding motif for the transcription factor AP-2, whereas the T allele disrupts this binding motif. Thus, intronic base substitutions in the MCM6 gene affect expression of the lactase-phlorizin hydrolase gene, which is downstream. It is, of course, possible that a similar mechanism explains the role of the soluble adenylate cyclase SNPs in autosomal dominant absorptive hypercalciuria. However, much work would need to be done to explore this possibility. Finally, it is also possible, and indeed likely, that the observed substitutions are linked with disease-producing mutations in a neighboring gene and defects in this other gene are responsible for absorptive hypercalciuria. Thus, the data regarding the likelihood that the soluble adenylate cyclase gene is the gene causing absorptive hypercalciuria in the chromosome 1-linked kindreds are tentative.
Reed et al. (4) sought to provide further evidence that this gene contributes to the phenotype by examining a sample of "nonfamilial" absorptive hypercalciuria patients. This approach is quite common and has been taken by many authors who have first identified the chromosomal region of a disease gene using kindreds segregating a Mendelian form of the disease, identified a candidate gene with disease-causing mutations, and then examined the role of the candidate gene in sporadic cases of disease. The hypothesis in these studies is that a subset of the sporadic cases might be due to mutations in the gene and/or functional polymorphisms in this gene might increase predisposition to the disease, whereas additional unidentified genes or environmental factors might contribute to disease risk in sporadic cases.
The reported association studies provide some evidence that the soluble adenylate cyclase gene may play a role in absorptive hypercalciuria. Support for this gene as a causative factor comes from the finding that all six base substitutions in the gene that were observed in the autosomal dominant disorder had significantly different allelic frequencies in the absorptive hypercalciuria patients as compared with the control sample. In addition, it is also promising that there was no evidence of association with the other two genes that were evaluated. However, it should be noted that only 3 polymorphisms were found for these two other genes, whereas 19 polymorphisms were studied in the soluble adenylate cyclase gene. Clearly, these data would be stronger if more polymorphisms in other genes were studied and found not to be associated with absorptive hypercalciuria, thereby adding important negative controls to this study.
As with all case control studies, several notes of caution must be raised in the interpretation of these data. First, similar to the concerns raised when the intronic substitutions were proposed to be disease producing in the three Mendelian kindreds, it is difficult to hypothesize a mechanism by which the five other SNPs reported to be associated with sporadic absorptive hypercalciuria could be disease producing. This failure to address a potential mechanism is an important limitation of the reported study.
A second potential concern is the sample of control individuals. Ideally, in a case control study, the two groups should be matched so that they differ only in their disease status. In many case control studies, the control sample is poorly matched to the affected cases and differences in allele frequencies may easily be due to factors other than disease status. In this study, the control sample is reported to be matched by race with the absorptive hypercalciuria patients; however, race is the most general form of matching. More subtle ethnic stratification or admixture, resulting in different allele frequencies in the two groups, can be a cause for spurious association results. Although population admixture can result in deviations in allele frequencies of the cases and controls from Hardy-Weinberg equilibrium, looking for deviations from Hardy-Weinberg equilibrium is an insensitive method for detecting admixture. As proposed by Pritchard and Rosenberg (10) a more sensitive method would be to perform association studies with numerous negative control genetic markers that are not thought to be linked to the phenotype. Thus, as noted above for the Mendelian disorder, additional negative controls would be extremely useful in analyzing these data. Finally, among a subset of the apparently normal volunteers, there were some who on further evaluation were hypercalciuric, suggesting that this is not an ideal control sample. Although the presence of hypercalciurics among the control sample only reduces the power to detect association, these results suggest further screening of the control sample would have been warranted.
The identification of the gene causing absorptive hypercalciuria will likely yield important new scientific hypotheses regarding not only stone formation but also, as the authors of this report suggest, candidate genes for osteoporosis. This region of chromosome 1 has previously been reported linked to peak bone mineral density (BMD) in a sample of premenopausal women (11). The identified linkage region is quite broad, and the ability to identify biologically important candidate genes will be an extremely powerful means to more rapidly identify the genes contributing to peak BMD. It will be essential to use multiple study designs to evaluate the role of candidate genes in the chromosome 1 region on peak BMD and osteoporosis. To rigorously evaluate a potential candidate gene, it will be important not to limit these studies to case control designs.
In summary, the genetics of absorptive hypercalciuria is an important topic that is very worthy of study. The soluble adenylate cyclase gene is an interesting candidate gene, and the results among both the three kindreds as well as the case control association study suggest evidence of linkage and linkage disequilibrium between the gene and absorptive hypercalciuria. However, none of the evidence provided in this report conclusively demonstrates that substitutions or mutations in this gene cause absorptive hypercalciuria. Rather, these studies suggest that additional unidentified mutations in this gene or perhaps in a gene in close physical proximity to the soluble adenylase cyclase gene are, in fact, disease producing. Alternatively, as recently reported, the SNPs observed in the soluble adenylate cyclase gene may be disease producing; however, their effect may be on a different gene.
Acknowledgments
Footnotes
This work was supported by NIH Grants P01-AG18397, K24-AR02095, and R01-AR42228.
Abbreviations: BMD, Bone mineral density; SNP, single nucleotide polymorphism.
Received January 28, 2002.
Accepted January 28, 2002.
References