BoneKEy-Osteovision | Meeting Reports
Meeting report from the 2nd international workshop on the genetics of bone metabolism and disease
Omar M Albagha
Stuart H Ralston
Tracy L Stewart
DOI:10.1138/2003095
The 2nd International Workshop on the Genetics of Bone Metabolism and Disease, held February 15-18, 2003 in Davos, Switzerland, once again brought clinicians and scientists with an interest in genetics from the field of metabolic bone disease together with leading authorities in molecular genetics and linkage analysis.
John Blangero (San Antonio, TX) opened the meeting with a provocative review of current approaches to mapping complex disease genes, using family based-designs. While acknowledging that extended family pedigrees are not always available, he emphasized the advantages of family-based approaches over using unrelated individuals and of using different study designs to maximize the likelihood of finding the causal genes. Blangero specifically focused on variance component-based linkage analysis in the evaluation of quantitative trait loci (QTLs) as implemented in the sequential oligogenic linkage analysis routine program (SOLAR). (See http://www.sfbr.org for more information.) Blangero argued that the component-based algorithms for quantitative trait analysis included in SOLAR give greater power than an affection status model and allow one to control for gene-environment interactions and epistasis. He warned against focusing purely on linkage disequilibrium (LD) mapping, arguing that it can produce spurious associations because of the fact that neighboring single nucleotide polymorphisms (SNPs) could act independently. Blangero admitted, however, that at the present time, the use of variance component methods in genetic epidemiology has been used on cherry-picked QTLs and speculated that integration of disciplines and data will be required for more complex QTLs.
Lynn Jorde (Salt Lake City, UT) reiterated the benefits of combining different analysis techniques to map complex disease genes. Jorde argued in favor of using association studies and LD mapping to narrow down the large candidate regions that are often yielded by linkage-based studies. He went on to review the factors that affect the patterns of LD, discussing how locus heterogeneity and allelic heterogeneity can adversely affect the power to detect disease-causing allelic variants. Jorde also emphasized the importance of trying to minimize sources of environmental variance by careful study design and how to maximize the genetic signal through selection of extreme phenotypes. He discussed the benefits of DNA pooling to increase study power while reducing cost and encouraged researchers to explore the demographic history of their study populations to determine their usefulness for LD mapping. For example, assuming that a mutation occurs early in a founder population and that the relative effect of each locus is similar in diverse populations, younger populations are useful for the initial detection of a disease locus via LD, whereas “ancient” populations are more useful for the fine-scale mapping of the disease locus.
Wesley Beamer (Bar Harbor, ME) discussed the benefits of selective breeding of mice to focus the fine mapping of QTLs. He described the results from his own laboratory on the analysis of total femoral bone density by pQCT. Beamer has used analysis of congenic offspring of mice with extremes of bone phenotypes to resolve their fundamental genetic constituents and biological analysis of QTLs. Focusing on chromosome 1, he has identified a 20-cM region that exerts influence over femur length and femur bone density and has confirmed the presence of this QTL in another mouse strain. Beamer emphasized the importance of replicating QTL in different mouse strains and speculated that confirmation of specific QTL in several strains may help to resolve the effect of gene clusters.
Veronica van Heyningen (Edinburgh, UK) opened the session on gene structure and regulation with a fascinating review of the influence of long-range regulatory elements on gene expression, using the Pax6 gene as an example. She emphasized the need to develop new methodologies to evaluate the role of distant control elements and pointed out that unfortunately studies thus far had mostly been performed in developing (rather than adult) animals. She pointed out, however, that there clearly was a need to determine whether regulatory elements distant to genes are active in adult as well as developing organisms. If this proves to be the case, it may result in a re-investigation of the results of gene knockout studies because some regulatory elements have been found to exert influence over distances of 150 kb.
Continuing the theme of long-range control elements, Michel Georges (Liège, Belgium) gave an entertaining account of genetic imprinting as revealed by the “beautiful buttocks” (callipyge) phenotype, a type of muscular hypertrophy in sheep. Extensive breeding enabled identification of an unusual transregulation effect on gene expression. It turns out that the callipyge mutation affects a putative control element that can produce its subsequent effect only on a heterozygous background. The SNP, located in a highly interspecies-conserved region, exerts control over a suite of nearby imprinted genes, either directly or indirectly, through a corresponding RNA transcript.
John Quinn (Liverpool, UK) discussed the application of YAC transgenesis to address the role of nonexonic regulatory polymorphisms. Quinn argued that mouse models might not be the most useful tools for assessing some diseases, because some human polymorphisms do not exist in the mouse genome. He advocated the use of crossing a human allele onto a corresponding mouse knockout, creating a “tool” to better address drug design against human proteins of interest. He directed the audience to the National Center for Biotechnology Information (NCBI) website (http://www.ncbi.nih.gov/), which catalogs available BACs and YACs that would be useful for this purpose. He pointed out that the capacity to clone large YACs (150-300 kb) might help to circumvent the effects of removing regulatory elements and also enable studies of gene dose-related effect. These large YACs are capable of being inserted into endothelial cells, for example, and as yet have not shown problems of immune response beyond the initial injection.
Gerard Karsenty (Houston, TX) discussed the approach used in his laboratory to elucidate the transcriptional control of cell differentiation in chondrocyte and osteoblast lineages. Karsenty presented the case for the mouse as an exceptionally good animal model for the study of human skeletal biology because of the close concordance between skeletal phenotype in mice, where disease genes had been knocked out, and humans who had a “natural” knockout of the same gene caused by genetic disease. Karsenty also acknowledged the importance that bone histomorphometry has had in characterizing the skeletal phenotype in transgenic animals and pointed out that without this tool, we would know much less about the mechanisms by which disease genes affect bone metabolism. Karsenty then reviewed the results of a series of elegant experiments that showed that the effects of leptin, a hormone that his group had identified as a major regulator of bone mass, were mediated by a hypothalamic relay involving the sympathetic nervous system and that the modulating sympathetic activity affects bone mass and osteoblast function without affecting body weight.
Timothy Triche (Los Angeles, CA) opened the session on pharmacogenomics by detailing how his group has used microarray technology to probe the biology of osteosarcoma, addressing the need to predict clinical course from the first biopsy. The authors selected the transcriptome as a target because it is 100-fold smaller than the genome and, unlike the proteome, can be tackled at this time. They have successfully identified 10 genes that predict poor survival, three of which are strong predictors. Triche highlighted the fact that when analyzing the data, magnitude of expression is of little value because of the cascading effects of genes. One needs instead to look for statistical associations with outcome. Triche revealed that the greatest challenge to researchers is data management and analysis. Key issues include determining the robustness of gene expression clusters, determining the contribution of an individual gene within a cluster, and identifying relationships between gene clusters. He pointed out that sources of variation occurred at each stage of the processing of a sample (from biopsy to chip), necessitating replication in a series of individuals. In experiments in his own lab, results were replicated in at least 10 individuals. Many members of the audience pointed out the significant cost implications of this approach, given the high price of gene chips, and a plea to reduce this price resonated throughout the audience.
David Herrington (Wake Forest, NC) provided a stimulating account of the pharmacogenetics of the estrogen response, exploring the paradigms exposed by the recently conducted Women's Health Initiative study of hormone replacement therapy (HRT) in postmenopausal women. Herrington reviewed research in which his group had identified several SNPs in the estrogen receptor alpha gene that were associated with biological response to HRT. One of the most important was an SNP within intron 1, located at an myb binding site within the ESR1a gene. He presented data to suggest that this polymorphism affected the transcriptional response to Myb and suggested that different individuals with different alleles at this SNP may respond differently to HRT in terms of both side effects and therapeutic effects.
The session on the genetics of rare bone diseases was opened with a presentation by Jean Morissette (Quebec, Canada). Morissette reviewed the results of genetic linkage studies conducted in his laboratory, which resulted in the identification of a recurrent proline-leucine mutation at codon 392 (P392L) in the sequestosome 1 gene (SQSTM1) as a common cause of Paget's disease in the French-Canadian population. In addition, he pointed out that this and other mutations in the SQSTM1 gene have also been identified in other populations. He reported that another candidate locus for Paget's disease had been identified on chromosome 5q31, which contained several candidate genes, including members of the interleukin family and SMAD 5. He reported that although the screening of candidate genes within this region was underway, the causal gene had not yet been identified.
Wim van Hul (Antwerp, Belgium) followed with a concise review of current knowledge and recent developments in the pathogenesis of osteopetrosis and sclerosing bone dysplasias, focusing on the molecular genetic basis of these disorders. His talk illustrated the importance of studying monogenic bone diseases as a route to identifying molecules that play key roles in regulating osteoblast and osteoclast activity. Examples included the SOST gene, which van Hul's group and others have identified as the cause of sclerosteosis and Van Buchem's disease; the TGFb1 gene, which has been identified as the cause of Camurati-Engelmann disease; the CLCN7 and TCIRG1 genes, which have been identified as common causes of osteopetrosis by affecting chloride channels (CLCN7) and the proton pump (TCRIG1) in osteoclasts; and the LRP5 gene, which has been identified as the cause of high bone mass syndromes as well as osteoporosis pseudoglioma syndrome.
The genetic basis of osteoporosis was tackled by Hong-Wen Deng (Omaha, NE). Deng emphasized the importance of focusing on identification of genes associated with osteoporotic fracture, as well as just bone mineral density, but acknowledged that this was challenging because of the fact that fractures have an important environmental component. He also emphasized the need for studies to address the functional mechanisms by which genetic mutations and polymorphisms affect bone metabolism and bone mass. Deng advocated the use of the transmission/disequilibrium test (TDT) in the study of late onset diseases. He also promoted the use of exclusion mapping to remove false positives, but stressed that caution should be used regarding the assumptions involved in this analysis because this too could lead to false positives.
John Loughlin (Oxford, UK) presented a review of current progress in the genetics of osteoarthritis, where the emphasis so far has been on a combination of candidate gene studies and genome-wide scans. He reviewed the results of the genome-wide scans reported to date and noted that although 15 distinct loci have been identified, only two (11q and 16p) have been confirmed in more than one study. Using high-density mapping and sib TDT, Loughlin's group has narrowed down the results of their genome-wide scan to a 5-Mb interval on chromosome 6 and a 30-Mb interval on chromosome 2. He reported that his group has selected several positional candidate genes, including frizzled B (FRZB), from these regions of interest and presented results of preliminary studies, which have shown a possible association of FRZB with osteoarthritis in females.
In a satellite symposium on the use of bisphosphonates (BPs) in rare genetic bone diseases, Mike Rogers (Aberdeen, UK) delivered an informative review on the current understanding of the mechanism of BP action. His group has determined that nitrogen-containing BPs (N-BPs) inhibit binding of the endogenous substrates of farnesyl diphosphate (FPP) synthase, halting the process of protein prenylation. This affects downstream intracellular trafficking, disrupting osteoclast function. Identification of a protein that specifically prevents prenylation has revealed that a specific type of prenylation, geranylgeranylation, is vital for osteoclast function. These findings should identify new candidates for targeted drug discovery.
Francis Glorieux (Montreal, Quebec) discussed the use of BPs in the treatment of osteogenesis imperfecta (OI). Glorieux reviewed the molecular basis of OI, explaining that seven types have now been described, three of which are not linked to mutations in either COLIA1 or COLIA2. He reported that the initial results of treatment with BPs are good, with significant beneficial effects on bone pain and BMD. He explained, however, that fractures still occurred because of the underlying nature of the disease and that this could not be corrected completely with BPs. Glorieux mentioned that a placebo-controlled trial was currently in progress to examine the effects of alendronate in OI and that the results of this trial would be important in defining the long-term benefits of BP therapy in patients with OI.
Stuart Ralston (Aberdeen, UK) gave a well-illustrated review of rarely encountered genetic bone diseases and the role that BPs can play in their treatment. He reviewed the results of clinical experience with BPs in a variety of rare bone diseases and reported that the drugs had given fairly consistent and rapid reductions in bone pain in a variety of bone disorders. Although Ralston emphasized that clinical experience with BPs in rare bone diseases has been for the most part positive, he highlighted the need for caution when interpreting the results of uncontrolled studies and emphasized the need to establish multicenter, randomized, controlled treatment trials in these rare disorders.
On the final day, David Clayton (Cambridge, UK) delivered a critical review of current approaches to epidemiological studies of gene-environment interactions with respect to complex phenotypes. He queried the purpose of studies designed merely to detect statistical variations, preferring instead to ask whether effects seen are biologically plausible and clinically important. Clayton suggested that case-control and pedigree studies are most suited to investigating gene-disease associations, whereas cohort studies should be used to examine gene-environment interactions. He emphasized the need for well-founded hypotheses in order to consider the roles of gene and environment together. Clayton also proposed that gene identification should help focus the attention of epidemiologists on the relevant environmental factors that need to be measured. He concluded, however, that in polygenic diseases, targeting intervention at genetically susceptible subgroups is not currently viable in terms of public health policy.
John Ioannidis (Ioannina, Greece) closed the meeting with a lively critique of the use of meta-analysis to quantitate the effects of polygenes. Ioannidis pointed out that for polygenes, strong associations seem to diminish as further studies are added to the analysis. Findings in large studies are most likely to be confirmed, whereas studies that look at a large number of genes, but that find significance in relatively few, are least likely to be replicated. These findings emphasize the need for large-scale studies, but the avoidance of heterogeneity, bias, and the over-interpretation of results. Ioannidis concluded by reminding us that statistical techniques cannot compensate for poor study design.
In addition to the invited speakers, delegates were treated to a wide variety of oral and poster presentations from the submitted abstracts. Unnur Styrkarsdottir and colleagues at deCODE Genetics, Inc. (Reykjavik, Iceland) performed a genome-wide search to identify chromosomal regions that affect the genetic susceptibility to osteoporosis. Using an affected-only linkage analysis approach in 98 extended families genotyped for 1000 microsatellite markers, the authors reported that they had identified a locus on chromosome 20p12 with significant evidence for linkage (LOD 3.6). The lodscore increased to 4.9 when the authors used a denser map of markers. Using LD mapping, they found a significant association between osteoporosis and haplotypes in the BMP2 gene and osteoporosis, with odds ratios ranging between 2 to 5, depending on the skeletal site studied.
Kristina Allen and colleagues at Genome Therapeutic Corp. (Waltham, MA) presented their work on the structural and functional analysis of LRP5 mutations. LRP5 is involved in Wnt signaling, and mutations in this gene are responsible for high bone mass (HBM) trait and osteoporosis pseudoglioma syndrome. Using computer simulation and reporter gene studies, the authors created several mutations in the LRP5 gene that mimic the previously identified G171V mutation that causes the HBM phenotype. All substitutions of G171 and mutations, such as A65V, A214V, and M282V, resulted in an HBM-like phenotype, whereas E128V and G199V did not.
Serge Ferrari (Geneva, Switzerland) presented his work on the role of the LRP5 gene in peak bone mass, vertebral size, and height in Caucasians. He found a strong association between a missense substitution (V667M) and lumbar spine BMD, bone mineral content, and height. Of interest, he observed that the largest differences among V667M genotypes occurred in adult males. Weaker association was found with two other mutations in this gene.
Annalisa Frattini and colleagues (Milan, Italy) presented a clinical and molecular analysis of 89 patients with autosomal recessive osteopetrosis investigated for two genes, TCIRG1 and CLCN7. The authors found that the spectra of clinical presentations caused by the two genes differed from one another. TCIRG1-dependent defects are restricted to severe autosomal recessive osteopetrosis, with the majority of cases caused by null mutations that consist of nonsense, deletions, and splicing defects. The clinical spectrum of CLCN7 mutations is wider than that of TC1RG1 and consists of missense substitutions, which cause severe (as well as benign) autosomal dominant osteopetrosis.
Many other oral presentations and posters were also presented at the meeting, and details of these abstracts can be found in the March 2003 issue of Calcified Tissue International .
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.