Epigenome-wide Association Study Shows Differential DNA Methylation of MDC1, KLF9, and CUTA in Autoimmune Thyroid Disease

CONTEXT

Autoimmune thyroid disease (AITD) includes Graves disease (GD) and Hashimoto disease (HD), which often run in the same family. AITD etiology is incompletely understood: Genetic factors may account for up to 75% of phenotypic variance, whereas epigenetic effects (including DNA methylation [DNAm]) may contribute to the remaining variance (eg, why some individuals develop GD and others HD).

OBJECTIVE

This work aimed to identify differentially methylated positions (DMPs) and differentially methylated regions (DMRs) comparing GD to HD.

METHODS

Whole-blood DNAm was measured across the genome using the Infinium MethylationEPIC array in 32 Australian patients with GD and 30 with HD (discovery cohort) and 32 Danish patients with GD and 32 with HD (replication cohort). Linear mixed models were used to test for differences in quantile-normalized β values of DNAm between GD and HD and data were later meta-analyzed. Comb-p software was used to identify DMRs.

RESULTS

We identified epigenome-wide significant differences (P < 9E-8) and replicated (P < .05) 2 DMPs between GD and HD (cg06315208 within MDC1 and cg00049440 within KLF9). We identified and replicated a DMR within CUTA (5 CpGs at 6p21.32). We also identified 64 DMPs and 137 DMRs in the meta-analysis.

CONCLUSION

Our study reveals differences in DNAm between GD and HD, which may help explain why some people develop GD and others HD and provide a link to environmental risk factors. Additional research is needed to advance understanding of the role of DNAm in AITD and investigate its prognostic and therapeutic potential.

Body mass index stratified meta-analysis of genome-wide association studies of polycystic ovary syndrome in women of European ancestry

BACKGROUND

Polycystic ovary syndrome (PCOS) is a complex multifactorial disorder with a substantial genetic component. However, the clinical manifestations of PCOS are heterogeneous with notable differences between lean and obese women, implying a different pathophysiology manifesting in differential body mass index (BMI). We performed a meta-analysis of genome-wide association study (GWAS) data from six well-characterised cohorts, using a case-control study design stratified by BMI, aiming to identify genetic variants associated with lean and overweight/obese PCOS subtypes.

RESULTS

The study comprised 254,588 women (5,937 cases and 248,651 controls) from individual studies performed in Australia, Estonia, Finland, the Netherlands and United States of America, and separated according to three BMI stratifications (lean, overweight and obese). Genome-wide association analyses were performed for each stratification within each cohort, with the data for each BMI group meta-analysed using METAL software. Almost half of the total study population (47%, n = 119,584) were of lean BMI (≤ 25 kg/m2). Two genome-wide significant loci were identified for lean PCOS, led by rs12000707 within DENND1A (P = 1.55 × 10-12) and rs2228260 within XBP1 (P = 3.68 × 10-8). One additional locus, LINC02905, was highlighted as significantly associated with lean PCOS through gene-based analyses (P = 1.76 × 10-6). There were no significant loci observed for the overweight or obese sub-strata when analysed separately, however, when these strata were combined, an association signal led by rs569675099 within DENND1A reached genome-wide significance (P = 3.22 × 10-9) and a gene-based association was identified with ERBB4 (P = 1.59 × 10-6). Nineteen of 28 signals identified in previous GWAS, were replicated with consistent allelic effect in the lean stratum. There were less replicated signals in the overweight and obese groups, and only 4 SNPs were replicated in each of the three BMI strata.

CONCLUSIONS

Genetic variation at the XBP1, LINC02905 and ERBB4 loci were associated with PCOS within unique BMI strata, while DENND1A demonstrated associations across multiple strata, providing evidence of both distinct and shared genetic features between lean and overweight/obese PCOS-affected women. This study demonstrated that PCOS-affected women with contrasting body weight are not only phenotypically distinct but also show variation in genetic architecture; lean PCOS women typically display elevated gonadotrophin ratios, lower insulin resistance, higher androgen levels, including adrenal androgens, and more favourable lipid profiles. Overall, these findings add to the growing body of evidence supporting a genetic basis for PCOS as well as differences in genetic patterns relevant to PCOS BMI-subtype.

Age-dependent genetic regulation of osteoarthritis: independent effects of immune system genes

OBJECTIVES

Osteoarthritis (OA) is a joint disease with a heritable component. Genetic loci identified via genome-wide association studies (GWAS) account for an estimated 26.3% of the disease trait variance in humans. Currently, there is no method for predicting the onset or progression of OA. We describe the first use of the Collaborative Cross (CC), a powerful genetic resource, to investigate knee OA in mice, with follow-up targeted multi-omics analysis of homologous regions of the human genome.

METHODS

We histologically screened 275 mice for knee OA and conducted quantitative trait locus (QTL) mapping in the complete cohort (> 8 months) and the younger onset sub-cohort (8-12 months). Multi-omic analysis of human genetic datasets was conducted to investigate significant loci.

RESULTS

We observed a range of OA phenotypes. QTL mapping identified a genome-wide significant locus on mouse chromosome 19 containing Glis3, the human equivalent of which has been identified as associated with OA in recent GWAS. Mapping the younger onset sub-cohort identified a genome-wide significant locus on chromosome 17. Multi-omic analysis of the homologous region of the human genome (6p21.32) indicated the presence of pleiotropic effects on the expression of the HLA - DPB2 gene and knee OA development risk, potentially mediated through the effects on DNA methylation.

CONCLUSIONS

The significant associations at the 6p21.32 locus in human datasets highlight the value of the CC model of spontaneous OA that we have developed and lend support for an immune role in the disease. Our results in mice also add to the accumulating evidence of a role for Glis3 in OA.

Leveraging osteoclast genetic regulatory data to identify genes with a role in osteoarthritis

There has been a growing interest in the role of the subchondral bone and its resident osteoclasts in the progression of osteoarthritis (OA). A recent genome-wide association study (GWAS) identified 100 independent association signals for OA traits. Most of these signals are led by noncoding variants, suggesting that genetic regulatory effects may drive many of the associations. We have generated a unique human osteoclast-like cell-specific expression quantitative trait locus (eQTL) resource for studying the genetics of bone disease. Considering the potential role of osteoclasts in the pathogenesis of OA, we performed an integrative analysis of this dataset with the recently published OA GWAS results. Summary data-based Mendelian randomization (SMR) and colocalization analyses identified 38 genes with a potential role in OA, including some that have been implicated in Mendelian diseases with joint/skeletal abnormalities, such as BICRA, EIF6, CHST3, and FBN2. Several OA GWAS signals demonstrated colocalization with more than one eQTL peak, including at 19q13.32 (hip OA with BCAM, PRKD2, and BICRA eQTL). We also identified a number of eQTL signals colocalizing with more than one OA trait, including FAM53A, GCAT, HMGN1, MGAT4A, RRP7BP, and TRIOBP. An SMR analysis identified 3 loci with evidence of pleiotropic effects on OA-risk and gene expression: LINC01481, CPNE1, and EIF6. Both CPNE1 and EIF6 are located at 20q11.22, a locus harboring 2 other strong OA candidate genes, GDF5 and UQCC1, suggesting the presence of an OA-risk gene cluster. In summary, we have used our osteoclast-specific eQTL dataset to identify genes potentially involved with the pathogenesis of OA.

Bone Trans-omics: Integrating Omics to Unveil Mechanistic Molecular Networks Regulating Bone Biology and Disease

PURPOSE OF REVIEW

Recent advancements in "omics" technologies and bioinformatics have afforded researchers new tools to study bone biology in an unbiased and holistic way. The purpose of this review is to highlight recent studies integrating multi-omics data gathered from multiple molecular layers (i.e.; trans-omics) to reveal new molecular mechanisms that regulate bone biology and underpin skeletal diseases.

RECENT FINDINGS

Bone biologists have traditionally relied on single-omics technologies (genomics, transcriptomics, proteomics, and metabolomics) to profile measureable differences (both qualitative and quantitative) of individual molecular layers for biological discovery and to investigate mechanisms of disease. Recently, literature has grown on the implementation of integrative multi-omics to study bone biology, which combines computational and informatics support to connect multiple layers of data derived from individual "omic" platforms. This emerging discipline termed "trans-omics" has enabled bone biologists to identify and construct detailed molecular networks, unveiling new pathways and unexpected interactions that have advanced our mechanistic understanding of bone biology and disease. While the era of trans-omics is poised to revolutionize our capacity to answer more complex and diverse questions pertinent to bone pathobiology, it also brings new challenges that are inherent when trying to connect "Big Data" sets. A concerted effort between bone biologists and interdisciplinary scientists will undoubtedly be needed to extract physiologically and clinically meaningful data from bone trans-omics in order to advance its implementation in the field.

Bone mineral density loci specific to the skull portray potential pleiotropic effects on craniosynostosis

Skull bone mineral density (SK-BMD) provides a suitable trait for the discovery of key genes in bone biology, particularly to intramembranous ossification, not captured at other skeletal sites. We perform a genome-wide association meta-analysis (n ~ 43,800) of SK-BMD, identifying 59 loci, collectively explaining 12.5% of the trait variance. Association signals cluster within gene-sets involved in skeletal development and osteoporosis. Among the four novel loci (ZIC1, PRKAR1A, AZIN1/ATP6V1C1, GLRX3), there are factors implicated in intramembranous ossification and as we show, inherent to craniosynostosis processes. Functional follow-up in zebrafish confirms the importance of ZIC1 on cranial suture patterning. Likewise, we observe abnormal cranial bone initiation that culminates in ectopic sutures and reduced BMD in mosaic atp6v1c1 knockouts. Mosaic prkar1a knockouts present asymmetric bone growth and, conversely, elevated BMD. In light of this evidence linking SK-BMD loci to craniofacial abnormalities, our study provides new insight into the pathophysiology, diagnosis and treatment of skeletal diseases.

Identification of Differentially Expressed Genes and Molecular Pathways Involved in Osteoclastogenesis Using RNA-seq

Osteoporosis is a disease that is characterised by reduced bone mineral density (BMD) and can be exacerbated by the excessive bone resorption of osteoclasts (OCs). Bioinformatic methods, including functional enrichment and network analysis, can provide information about the underlying molecular mechanisms that participate in the progression of osteoporosis. In this study, we harvested human OC-like cells differentiated in culture and their precursor peripheral blood mononuclear cells (PBMCs) and characterised the transcriptome of the two cell types using RNA-sequencing in order to identify differentially expressed genes. Differential gene expression analysis was performed in RStudio using the edgeR package. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to identify enriched GO terms and signalling pathways, with inter-connected regions characterised using protein-protein interaction analysis. In this study, we identified 3201 differentially expressed genes using a 5% false discovery rate; 1834 genes were upregulated, whereas 1367 genes were downregulated. We confirmed a significant upregulation of several well-established OC genes including CTSK, DCSTAMP, ACP5, MMP9, ITGB3, and ATP6V0D2. The GO analysis suggested that upregulated genes are involved in cell division, cell migration, and cell adhesion, while the KEGG pathway analysis highlighted oxidative phosphorylation, glycolysis and gluconeogenesis, lysosome, and focal adhesion pathways. This study provides new information about changes in gene expression and highlights key biological pathways involved in osteoclastogenesis.

Multi-ancestry genome-wide association analyses improve resolution of genes and pathways influencing lung function and chronic obstructive pulmonary disease risk

Lung-function impairment underlies chronic obstructive pulmonary disease (COPD) and predicts mortality. In the largest multi-ancestry genome-wide association meta-analysis of lung function to date, comprising 580,869 participants, we identified 1,020 independent association signals implicating 559 genes supported by ≥2 criteria from a systematic variant-to-gene mapping framework. These genes were enriched in 29 pathways. Individual variants showed heterogeneity across ancestries, age and smoking groups, and collectively as a genetic risk score showed strong association with COPD across ancestry groups. We undertook phenome-wide association studies for selected associated variants as well as trait and pathway-specific genetic risk scores to infer possible consequences of intervening in pathways underlying lung function. We highlight new putative causal variants, genes, proteins and pathways, including those targeted by existing drugs. These findings bring us closer to understanding the mechanisms underlying lung function and COPD, and should inform functional genomics experiments and potentially future COPD therapies.

Functional Assessment of Calcium-Sensing Receptor Variants Confirms Familial Hypocalciuric Hypercalcaemia

Context

In the clinic it is important to differentiate primary hyperparathyroidism (PHPT) from the more benign, inherited disorder, familial hypocalciuric hypercalcemia (FHH). Since the conditions may sometimes overlap biochemically, identification of calcium-sensing receptor (CASR) gene variants causative of FHH (but not PHPT) is the most decisive diagnostic aid. When novel variants are identified, bioinformatics and functional assessment are required to establish pathogenicity.

Objective

We identified 3 novel CASR transmembrane domain missense variants, Thr699Asn, Arg701Gly, and Thr808Pro, in 3 probands provisionally diagnosed with FHH and examined the variants using bioinformatics and functional analysis.

Methods

Bioinformatics assessment utilized wANNOVAR software. For functional characterization, each variant was cloned into a mammalian expression vector; wild-type and variant receptors were transfected into HEK293 cells, and their expression and cellular localization were assessed by Western blotting and confocal immunofluorescence, respectively. Receptor activation in HEK293 cells was determined using an IP-One ELISA assay following stimulation with Ca⁺⁺ ions.

Results

Bioinformatics analysis of the variants was unable to definitively assign pathogenicity. Compared with wild-type receptor, all variants demonstrated impaired expression of mature receptor reaching the cell surface and diminished activation at physiologically relevant Ca⁺⁺ concentrations.

Conclusion

Three CASR missense variants identified in probands provisionally diagnosed with FHH result in receptor inactivation and are therefore likely causative of FHH. Inactivation may be due to inadequate processing/trafficking of mature receptor and/or conformational changes induced by the variants affecting receptor signaling. This study demonstrates the value of functional studies in assessing genetic variants identified in hypercalcemic patients.

Molecular structure, expression, and the emerging role of Siglec-15 in skeletal biology and cancer

Siglec-15, a Siglec family member and type-1 transmembrane protein, is expressed mainly in human macrophages and dendritic cells. It is comprised of a lysine-containing transmembrane domain, two extracellular immunoglobulin (Ig)-like domains and a short cytoplasmic domain. Siglec-15 is highly conserved in vertebrates and acts as an immunoreceptor. It exerts diverse functions on osteoclast physiology as well as the tumor microenvironment. Siglec-15 interacts with adapter protein DAP12 - Syk signaling pathway to regulate the RANKL/RANK-mediated PI3K, AKT, and ERK signaling pathways during osteoclast formation in vitro. Consistently, the lack of the Siglec-15 gene in mice leads to impaired osteoclast activity and osteopetrosis in vivo. In addition, Siglec-15 is expressed by tumor-associated macrophages (TAMs) and regulates the tumor microenvironment by activating the SYK/MAPK signaling pathway. Interestingly, Siglec-15 shares sequence homology to programmed death-ligand 1 (PD-L1) and has a potential immune-regulatory role in cancer immunology. Thus, Siglec-15 might also represent an alternative target for the treatment of cancers that do not respond to anti-PD-L1/PD-1 immunotherapy. Understanding the role of Siglec-15 in osteoclastogenesis and the tumor microenvironment will help us to develop new treatments for bone disorders and cancer.

Genome-wide analysis of thyroid function in Australian adolescents highlights SERPINA7 and NCOA3

OBJECTIVE

Genetic factors underpin the narrow intraindividual variability of thyroid function, although precise contributions of environmental vs genetic factors remain uncertain. We sought to clarify the heritability of thyroid function traits and thyroid peroxidase antibody (TPOAb) positivity and identify single nucleotide polymorphisms (SNPs) contributing to the trait variance.

METHODS

Heritability of thyroid-stimulating hormone (TSH), free T4 (fT4), free T3 (fT3) and TPOAb in a cohort of 2854 euthyroid, dizygous and monozygous twins (age range 11.9-16.9 years) from the Brisbane Longitudinal Twin Study (BLTS) was assessed using structural equation modelling. A genome-wide analysis was conducted on 2832 of these individuals across 7 522 526 SNPs as well as gene-based association analyses. Replication analysis of the association results was performed in the Raine Study (n = 1115) followed by meta-analysis to maximise power for discovery.

RESULTS

Heritability of thyroid function parameters in the BLTS was 70.8% (95% CI: 66.7-74.9%) for TSH, 67.5% (59.8-75.3%) for fT4, 59.7% (54.4-65.0%) for fT3 and 48.8% (40.6-56.9%) for TPOAb. The genome-wide association study (GWAS) in the discovery cohort identified a novel association between rs2026401 upstream of NCOA3 and TPOAb. GWAS meta-analysis found associations between TPOAb and rs445219, also near NCOA3, and fT3 and rs12687280 near SERPINA7. Gene-based association analysis highlighted SERPINA7 for fT3 and NPAS3 for fT4.

CONCLUSION

Our findings resolve former contention regarding heritability estimates of thyroid function traits and TPOAb positivity. GWAS and gene-based association analysis identified variants accounting for a component of this heritability.

Characterisation of genetic regulatory effects for osteoporosis risk variants in human osteoclasts

Background

Osteoporosis is a complex disease with a strong genetic contribution. A recently published genome-wide association study (GWAS) for estimated bone mineral density (eBMD) identified 1103 independent genome-wide significant association signals. Most of these variants are non-coding, suggesting that regulatory effects may drive many of the associations. To identify genes with a role in osteoporosis, we integrate the eBMD GWAS association results with those from our previous osteoclast expression quantitative trait locus (eQTL) dataset.

Results

We identify sixty-nine significant cis-eQTL effects for eBMD GWAS variants after correction for multiple testing. We detect co-localisation of eBMD GWAS and osteoclast eQTL association signals for 21 of the 69 loci, implicating a number of genes including CCR5, ZBTB38, CPE, GNA12, RIPK3, IQGAP1 and FLCN. Summary-data-based Mendelian Randomisation analysis of the eBMD GWAS and osteoclast eQTL datasets identifies significant associations for 53 genes, with TULP4 presenting as a strong candidate for pleiotropic effects on eBMD and gene expression in osteoclasts. By performing analysis using the GARFIELD software, we demonstrate significant enrichment of osteoporosis risk variants among high-confidence osteoclast eQTL across multiple GWAS P value thresholds. Mice lacking one of the genes of interest, the apoptosis/necroptosis gene RIPK3, show disturbed bone micro-architecture and increased osteoclast number, highlighting a new biological pathway relevant to osteoporosis.

Conclusion

We utilise a unique osteoclast eQTL dataset to identify a number of potential effector genes for osteoporosis risk variants, which will help focus functional studies in this area.

Advanced Genetic Approaches in Discovery and Characterization of Genes Involved With Osteoporosis in Mouse and Human

Osteoporosis is a complex condition with contributions from, and interactions between, multiple genetic loci and environmental factors. This review summarizes key advances in the application of genetic approaches for the identification of osteoporosis susceptibility genes. Genome-wide linkage analysis (GWLA) is the classical approach for identification of genes that cause monogenic diseases; however, it has shown limited success for complex diseases like osteoporosis. In contrast, genome-wide association studies (GWAS) have successfully identified over 200 osteoporosis susceptibility loci with genome-wide significance, and have provided most of the candidate genes identified to date. Phenome-wide association studies (PheWAS) apply a phenotype-to-genotype approach which can be used to complement GWAS. PheWAS is capable of characterizing the association between osteoporosis and uncommon and rare genetic variants. Another alternative approach, whole genome sequencing (WGS), will enable the discovery of uncommon and rare genetic variants in osteoporosis. Meta-analysis with increasing statistical power can offer greater confidence in gene searching through the analysis of combined results across genetic studies. Recently, new approaches to gene discovery include animal phenotype based models such as the Collaborative Cross and ENU mutagenesis. Site-directed mutagenesis and genome editing tools such as CRISPR/Cas9, TALENs and ZNFs have been used in functional analysis of candidate genes in vitro and in vivo. These resources are revolutionizing the identification of osteoporosis susceptibility genes through the use of genetically defined inbred mouse libraries, which are screened for bone phenotypes that are then correlated with known genetic variation. Identification of osteoporosis-related susceptibility genes by genetic approaches enables further characterization of gene function in animal models, with the ultimate aim being the identification of novel therapeutic targets for osteoporosis.

Identification of Novel Loci Associated With Hip Shape: A Meta-Analysis of Genomewide Association Studies

We aimed to report the first genomewide association study (GWAS) meta-analysis of dual-energy X-ray absorptiometry (DXA)-derived hip shape, which is thought to be related to the risk of both hip osteoarthritis and hip fracture. Ten hip shape modes (HSMs) were derived by statistical shape modeling using SHAPE software, from hip DXA scans in the Avon Longitudinal Study of Parents and Children (ALSPAC; adult females), TwinsUK (mixed sex), Framingham Osteoporosis Study (FOS; mixed), Osteoporotic Fractures in Men study (MrOS), and Study of Osteoporotic Fractures (SOF; females) (total N = 15,934). Associations were adjusted for age, sex, and ancestry. Five genomewide significant (p < 5 × 10-9 , adjusted for 10 independent outcomes) single-nucleotide polymorphisms (SNPs) were associated with HSM1, and three SNPs with HSM2. One SNP, in high linkage disequilibrium with rs2158915 associated with HSM1, was associated with HSM5 at genomewide significance. In a look-up of previous GWASs, three of the identified SNPs were associated with hip osteoarthritis, one with hip fracture, and five with height. Seven SNPs were within 200 kb of genes involved in endochondral bone formation, namely SOX9, PTHrP, RUNX1, NKX3-2, FGFR4, DICER1, and HHIP. The SNP adjacent to DICER1 also showed osteoblast cis-regulatory activity of GSC, in which mutations have previously been reported to cause hip dysplasia. For three of the lead SNPs, SNPs in high LD (r2  > 0.5) were identified, which intersected with open chromatin sites as detected by ATAC-seq performed on embryonic mouse proximal femora. In conclusion, we identified eight SNPs independently associated with hip shape, most of which were associated with height and/or mapped close to endochondral bone formation genes, consistent with a contribution of processes involved in limb growth to hip shape and pathological sequelae. These findings raise the possibility that genetic studies of hip shape might help in understanding potential pathways involved in hip osteoarthritis and hip fracture. © 2018 The Authors. Journal of Bone and Mineral Research Published by Wiley Periodicals, Inc.

Expression Quantitative Trait Locus Study of Bone Mineral Density GWAS Variants in Human Osteoclasts

Osteoporosis is a complex disease with a strong genetic component. Genomewide association studies (GWAS) have been very successful at identifying common genetic variants associated with bone parameters. A recently published study documented the results of the largest GWAS for bone mineral density (BMD) performed to date (n = 142,487), identifying 307 conditionally independent single-nucleotide polymorphisms (SNPs) as associated with estimated BMD (eBMD) at the genomewide significance level. The vast majority of these variants are non-coding SNPs. Expression quantitative trait locus (eQTL) studies using disease-specific cell types have increasingly been integrated with the results from GWAS to identify genes through which the observed GWAS associations are likely mediated. We generated a unique human osteoclast-specific eQTL data set using cells differentiated in vitro from 158 participants. We then used this resource to characterize the 307 recently identified BMD GWAS SNPs for association with nearby genes (±500 kb). After correction for multiple testing, 24 variants were found to be significantly associated with the expression of 32 genes in the osteoclast-like cells. Bioinformatics analysis suggested that these variants and those in strong linkage disequilibrium with them are enriched in regulatory regions. Several of the eQTL associations identified are relevant to genes that present strongly as having a role in bone, particularly IQGAP1, CYP19A1, CTNNB1, and COL6A3. Supporting evidence for many of the associations was obtained from publicly available eQTL data sets. We have also generated strong evidence for the presence of a regulatory region on chromosome 15q21.2 relevant to both the GLDN and CYP19A1 genes. In conclusion, we have generated a unique osteoclast-specific eQTL resource and have used this to identify 32 eQTL associations for recently identified BMD GWAS loci, which should inform functional studies of osteoclast biology. © 2018 American Society for Bone and Mineral Research.

Life-Course Genome-wide Association Study Meta-analysis of Total Body BMD and Assessment of Age-Specific Effects

Bone mineral density (BMD) assessed by DXA is used to evaluate bone health. In children, total body (TB) measurements are commonly used; in older individuals, BMD at the lumbar spine (LS) and femoral neck (FN) is used to diagnose osteoporosis. To date, genetic variants in more than 60 loci have been identified as associated with BMD. To investigate the genetic determinants of TB-BMD variation along the life course and test for age-specific effects, we performed a meta-analysis of 30 genome-wide association studies (GWASs) of TB-BMD including 66,628 individuals overall and divided across five age strata, each spanning 15 years. We identified variants associated with TB-BMD at 80 loci, of which 36 have not been previously identified; overall, they explain approximately 10% of the TB-BMD variance when combining all age groups and influence the risk of fracture. Pathway and enrichment analysis of the association signals showed clustering within gene sets implicated in the regulation of cell growth and SMAD proteins, overexpressed in the musculoskeletal system, and enriched in enhancer and promoter regions. These findings reveal TB-BMD as a relevant trait for genetic studies of osteoporosis, enabling the identification of variants and pathways influencing different bone compartments. Only variants in ESR1 and close proximity to RANKL showed a clear effect dependency on age. This most likely indicates that the majority of genetic variants identified influence BMD early in life and that their effect can be captured throughout the life course.

Genome-wide association study meta-analysis for quantitative ultrasound parameters of bone identifies five novel loci for broadband ultrasound attenuation

Osteoporosis is a common and debilitating bone disease that is characterised by low bone mineral density, typically assessed using dual-energy X-ray absorptiometry. Quantitative ultrasound (QUS), commonly utilising the two parameters velocity of sound (VOS) and broadband ultrasound attenuation (BUA), is an alternative technology used to assess bone properties at peripheral skeletal sites. The genetic influence on the bone qualities assessed by QUS remains an under-studied area. We performed a comprehensive genome-wide association study (GWAS) including low-frequency variants (minor allele frequency ≥0.005) for BUA and VOS using a discovery population of individuals with whole-genome sequence (WGS) data from the UK10K project (n = 1268). These results were then meta-analysed with those from two deeply imputed GWAS replication cohorts (n = 1610 and 13 749). In the gender-combined analysis, we identified eight loci associated with BUA and five with VOS at the genome-wide significance level, including three novel loci for BUA at 8p23.1 (PPP1R3B), 11q23.1 (LOC387810) and 22q11.21 (SEPT5) (P = 2.4 × 10-8 to 1.6 × 10-9). Gene-based association testing in the gender-combined dataset revealed eight loci associated with BUA and seven with VOS after correction for multiple testing, with one novel locus for BUA at FAM167A (8p23.1) (P = 1.4 × 10-6). An additional novel locus for BUA was seen in the male-specific analysis at DEFB103B (8p23.1) (P = 1.8 × 10-6). Fracture analysis revealed significant associations between variation at the WNT16 and RSPO3 loci and fracture risk (P = 0.004 and 4.0 × 10-4, respectively). In conclusion, by performing a large GWAS meta-analysis for QUS parameters of bone using a combination of WGS and deeply imputed genotype data, we have identified five novel genetic loci associated with BUA.

Genome-wide association study using family-based cohorts identifies the WLS and CCDC170/ESR1 loci as associated with bone mineral density

Background

Osteoporosis is a common and debilitating bone disease that is characterised by a low bone mineral density (BMD), a highly heritable trait. Genome-wide association studies (GWAS) have proven to be very successful in identifying common genetic variants associated with BMD adjusted for age, gender and weight, however a large portion of the genetic variance for this trait remains unexplained. There is evidence to suggest significant genetic correlation between body size traits and BMD. It has also recently been suggested that unintended bias can be introduced as a result of adjusting a phenotype for a correlated trait. We performed a GWAS meta-analysis in two populations (total n = 6,696) using BMD data adjusted for only age and gender, in an attempt to identify genetic variants associated with BMD including those that may have potential pleiotropic effects on BMD and body size traits.

Results

We observed a single variant, rs2566752, associated with spine BMD at the genome-wide significance level in the meta-analysis (P = 3.36 × 10⁻⁰⁹). Logistic regression analysis also revealed an association between rs2566752 and fracture rate in one of our study cohorts (P = 0.017, n = 5,654). This is an intronic variant located in the wntless Wnt ligand secretion mediator (WLS) gene (1p31.3), a known BMD locus which encodes an integral component of the Wnt ligand secretion pathway. Bioinformatics analyses of variants in moderate LD with rs2566752 produced strong evidence for a regulatory role for the variants rs72670452, rs17130567 and rs1430738. Expression quantitative trait locus (eQTL) analysis suggested that the variants rs12568456 and rs17130567 are associated with expression of the WLS gene in whole blood, cerebellum and temporal cortex brain tissue (P = 0.034–1.19 × 10⁻²³). Gene-wide association testing using the VErsatile Gene-based Association Study 2 (VEGAS2) software revealed associations between the coiled-coil domain containing 170 (CCDC170) gene, located adjacent to the oestrogen receptor 1 (ESR1) gene, and BMD at the spine, femoral neck and total hip sites (P = 1.0 × 10⁻⁰⁶, 2.0 × 10⁻⁰⁶ and 2.0 × 10⁻⁰⁶ respectively).

Conclusions

Genetic variation at the WLS and CCDC170/ESR1 loci were found to be significantly associated with BMD adjusted for only age and gender at the genome-wide level in this meta-analysis.

Influence of ARHGEF3 and RHOA knockdown on ACTA2 and other genes in osteoblasts and osteoclasts

Osteoporosis is a common bone disease that has a strong genetic component. Genome-wide linkage studies have identified the chromosomal region 3p14-p22 as a quantitative trait locus for bone mineral density (BMD). We have previously identified associations between variation in two related genes located in 3p14-p22, ARHGEF3 and RHOA, and BMD in women. In this study we performed knockdown of these genes using small interfering RNA (siRNA) in human osteoblast-like and osteoclast-like cells in culture, with subsequent microarray analysis to identify genes differentially regulated from a list of 264 candidate genes. Validation of selected findings was then carried out in additional human cell lines/cultures using quantitative real-time PCR (qRT-PCR). The qRT-PCR results showed significant down-regulation of the ACTA2 gene, encoding the cytoskeletal protein alpha 2 actin, in response to RHOA knockdown in both osteoblast-like (P<0.001) and osteoclast-like cells (P = 0.002). RHOA knockdown also caused up-regulation of the PTH1R gene, encoding the parathyroid hormone 1 receptor, in Saos-2 osteoblast-like cells (P<0.001). Other findings included down-regulation of the TNFRSF11B gene, encoding osteoprotegerin, in response to ARHGEF3 knockdown in the Saos-2 and hFOB 1.19 osteoblast-like cells (P = 0.003–0.02), and down-regulation of ARHGDIA, encoding the Rho GDP dissociation inhibitor alpha, in response to RHOA knockdown in osteoclast-like cells (P<0.001). These studies identify ARHGEF3 and RHOA as potential regulators of a number of genes in bone cells, including TNFRSF11B, ARHGDIA, PTH1R and ACTA2, with influences on the latter evident in both osteoblast-like and osteoclast-like cells. This adds further evidence to previous studies suggesting a role for the ARHGEF3 and RHOA genes in bone metabolism.

Conditional testing of multiple variants associated with bone mineral density in the FLNB gene region suggests that they represent a single association signal

BACKGROUND

Low bone mineral density (BMD) is a primary risk factor for osteoporosis and is a highly heritable trait, but appears to be influenced by many genes. Genome-wide linkage studies have highlighted the chromosomal region 3p14-p22 as a quantitative trait locus for BMD (LOD 1.1 - 3.5). The FLNB gene, which is thought to have a role in cytoskeletal actin dynamics, is located within this chromosomal region and presents as a strong candidate for BMD regulation. We have previously identified significant associations between four SNPs in the FLNB gene and BMD in women. We have also previously identified associations between five SNPs located 5' of the transcription start site (TSS) and in intron 1 of the FLNB gene and expression of FLNB mRNA in osteoblasts in vitro. The latter five SNPs were genotyped in this study to test for association with BMD parameters in a family-based population of 769 Caucasian women.

RESULTS

Using FBAT, significant associations were seen for femoral neck BMD Z-score with the SNPs rs11720285, rs11130605 and rs9809315 (P = 0.004 - 0.043). These three SNPs were also found to be significantly associated with total hip BMD Z-score (P = 0.014 - 0.026). We then combined the genotype data for these three SNPs with the four SNPs we previously identified as associated with BMD and performed a conditional analysis to determine whether they represent multiple independent associations with BMD. The results from this analysis suggested that these variants represent a single association signal.

CONCLUSIONS

The SNPs identified in our studies as associated with BMD appear to be part of a single association signal between the FLNB gene and BMD in our data. FLNB is one of several genes located in 3p14-p22 that has been identified as significantly associated with BMD in Caucasian women.

Significant association between common polymorphisms in the aromatase gene CYP19A1 and bone mineral density in postmenopausal women

17β-Estradiol is important in maintaining bone structure, and regulation of its synthesis plays an important role in the development of postmenopausal osteoporosis. We and others have demonstrated associations between variation in the CYP19A1 gene (encoding aromatase) and areal bone mineral density (aBMD) phenotypes in women. In the present study 33 tag polymorphisms were genotyped across the CYP19A1 gene in a population of 1,185 Caucasian postmenopausal women to test the association between sequence variations, total DXA hip aBMD, and circulating 17β-estradiol levels. An in silico bioinformatics analysis was performed for single nucleotide polymorphisms (SNPs) associated with aBMD to identify putative functional effects, while linkage disequilibrium analysis of these SNPs was undertaken with previously published sequence variants. Five SNPs located in the central third of the gene were strongly associated with total-hip aBMD after adjustment for age (P = 0.006-0.013). A haplotype analysis of these five SNPs revealed an association between the haplotype C-G-G-G-C and increased aBMD (P = 0.008) and the haplotype A-A-A-A-A and a decreased aBMD (P = 0.021). The haplotype frequency was 9.0% for C-G-G-G-C and 15.4% for A-A-A-A-A, with the variation in mean total-hip aBMD explained by the haplotype analyses being 5% and 7%, respectively. None of these polymorphisms was significantly associated with circulating 17β-estradiol levels. In conclusion, common genetic variations within the CYP19A1 gene are significantly associated with aBMD in postmenopausal Caucasian women.

Homozygous deletion of the UGT2B17 gene is not associated with osteoporosis risk in elderly Caucasian women

Previously, homozygous deletion of the UGT2B17 gene has shown association with hip fracture. Using a high-throughput qRT-PCR assay, we genotyped UGT2B17 copy number variation (CNV) in 1,347 elderly Caucasian women and examined for effects on bone phenotypes. We found no evidence of association between UGT2B17 CNV and osteoporosis risk in this population.

INTRODUCTION

Genetic studies of osteoporosis commonly examine SNPs in candidate genes or whole genome analyses, but insertions and deletions of DNA, collectively called CNV, also comprise a large amount of the genetic variability between individuals. Previously, homozygous deletion of the UGT2B17 gene in CNV 4q13.2, which encodes an enzyme that mediates the glucuronidation of steroid hormones, has shown association with the risk of hip fracture.

METHODS

We used a quantitative real-time PCR assay for genotyping the UGT2B17 CNV in a well-characterized population study of 1,347 Caucasian women aged 75.2 ± 2.7 years (mean ± SD), to assess the effect of the CNV on bone mass density (BMD) at the total hip site and osteoporosis risk.

RESULTS

The UGT2B17 CNV distribution was consistent with the expected Hardy-Weinberg distribution and not different from frequencies previously reported in a Caucasian population. Data from ANCOVA of age- and weight-adjusted BMD for UGT2B17 CNV genotype showed no significant difference between genotype groups. Individuals with homozygous or heterozygous deletion of the UGT2B17 gene showed no increased risk of incident fragility fracture.

CONCLUSIONS

These data suggest that quantitative real-time PCR is a rapid and efficient technique for determination of candidate CNVs, including the UGT2B17 CNV; however, we found no evidence of an effect of UGT2B17 CNV on osteoporosis risk in elderly Caucasian women.

Impact of Neuritin 1 (NRN1) polymorphisms on fluid intelligence in schizophrenia

Neuritin 1 (NRN1), an activity-regulated gene with multiple roles in neurodevelopment and synaptic plasticity, is located within the 6p24-p25 interval on chromosome 6, previously identified as linked to a subtype of schizophrenia (SZ) characterized by pervasive cognitive deficit (CD). We have tested the effect of NRN1 sequence variation on susceptibility to SZ and on general cognitive ability in patients and non-psychiatric control subjects by re-sequencing the coding regions of NRN1 and its flanking sequences, and genotyping 19 single-nucleotide polymorphisms (SNPs) in 336 SZ patients and 172 healthy control individuals. All participants completed comprehensive neurocognitive assessment, including tests estimating premorbid/prior IQ and current IQ. Logistic regression analyses found no significant association for any of the 19 SNPs with SZ or its CD subtype. However, linear regression analysis gave significant association (P = 0.024 and P = 0.027 after correction for multiple testing) for polymorphisms rs1475157 and rs9405890 with current IQ in the patient group. In SZ, the rs1475157-rs9405890 haplotypes revealed a highly significant association with the abstraction component of current ("fluid") intelligence (P = 0.0014), and with percentage loss of IQ points between premorbid and current intelligence (P = 0.0041). Results in the control group were not significant after correction. This is the first analysis of association between variation in NRN1 and SZ. The findings suggest a role of NRN1 as a modifier of cognitive functioning in SZ, with implications for future research into the impact of the environment on the development and maintenance of "fluid" intelligence.

Collaborative meta-analysis: associations of 150 candidate genes with osteoporosis and osteoporotic fracture

BACKGROUND

Osteoporosis is a highly heritable trait. Many candidate genes have been proposed as being involved in regulating bone mineral density (BMD). Few of these findings have been replicated in independent studies.

OBJECTIVE

To assess the relationship between BMD and fracture and all common single-nucleotide polymorphisms (SNPs) in previously proposed osteoporosis candidate genes.

DESIGN

Large-scale meta-analysis of genome-wide association data.

SETTING

5 international, multicenter, population-based studies.

PARTICIPANTS

Data on BMD were obtained from 19 195 participants (14 277 women) from 5 populations of European origin. Data on fracture were obtained from a prospective cohort (n = 5974) from the Netherlands.

MEASUREMENTS

Systematic literature review using the Human Genome Epidemiology Navigator identified autosomal genes previously evaluated for association with osteoporosis. We explored the common SNPs arising from the haplotype map of the human genome (HapMap) across all these genes. BMD at the femoral neck and lumbar spine was measured by dual-energy x-ray absorptiometry. Fractures were defined as clinically apparent, site-specific, validated nonvertebral and vertebral low-energy fractures.

RESULTS

150 candidate genes were identified and 36 016 SNPs in these loci were assessed. SNPs from 9 gene loci (ESR1, LRP4, ITGA1, LRP5, SOST, SPP1, TNFRSF11A, TNFRSF11B, and TNFSF11) were associated with BMD at either site. For most genes, no SNP was statistically significant. For statistically significant SNPs (n = 241), effect sizes ranged from 0.04 to 0.18 SD per allele. SNPs from the LRP5, SOST, SPP1, and TNFRSF11A loci were significantly associated with fracture risk; odds ratios ranged from 1.13 to 1.43 per allele. These effects on fracture were partially independent of BMD at SPP1 and SOST.

LIMITATION

Only common polymorphisms in linkage disequilibrium with SNPs in HapMap could be assessed, and previously reported associations for SNPs in some candidate genes could not be excluded.

CONCLUSION

In this large-scale collaborative genome-wide meta-analysis, 9 of 150 candidate genes were associated with regulation of BMD, 4 of which also significantly affected risk for fracture. However, most candidate genes had no consistent association with BMD.

Bone structural effects of variation in the TNFRSF1B gene encoding the tumor necrosis factor receptor 2

The 1p36 region of the human genome has been identified as containing a QTL for BMD in multiple studies. We analysed the TNFRSF1B gene from this region, which encodes the TNF receptor 2, in two large population-based cohorts. Our results suggest that variation in TNFRSF1B is associated with BMD.

INTRODUCTION

The TNFRSF1B gene, encoding the TNF receptor 2, is a strong positional and functional candidate gene for impaired bone structure through the role that TNF has in bone cells. The aims of this study were to evaluate the role of variations in the TNFRSF1B gene on bone structure and osteoporotic fracture risk in postmenopausal women.

METHODS

Six SNPs in TNFRSF1B were analysed in a cohort of 1,190 postmenopausal Australian women, three of which were also genotyped in an independent cohort of 811 UK postmenopausal women. Differences in phenotypic means for genotype groups were examined using one-way ANOVA and ANCOVA.

RESULTS

Significant associations were seen for IVS1+5580A>G with BMD and QUS parameters in the Australian population (P = 0.008 - 0.034) and with hip BMD parameters in the UK population (P = 0.005 - 0.029). Significant associations were also observed between IVS1+6528G>A and hip BMD parameters in the UK cohort (P = 0.0002 - 0.003). We then combined the data from the two cohorts and observed significant associations between both IVS1+5580A>G and IVS1+6528G>A and hip BMD parameters (P = 0.002 - 0.033).

CONCLUSIONS

Genetic variation in TNFRSF1B plays a role in the determination of bone structure in Caucasian postmenopausal women, possibly through effects on osteoblast and osteoclast differentiation.

Bone mineral density, osteoporosis, and osteoporotic fractures: a genome-wide association study

BACKGROUND

Osteoporosis is diagnosed by the measurement of bone mineral density, which is a highly heritable and multifactorial trait. We aimed to identify genetic loci that are associated with bone mineral density.

METHODS

In this genome-wide association study, we identified the most promising of 314 075 single nucleotide polymorphisms (SNPs) in 2094 women in a UK study. We then tested these SNPs for replication in 6463 people from three other cohorts in western Europe. We also investigated allelic expression in lymphoblast cell lines. We tested the association between the replicated SNPs and osteoporotic fractures with data from two studies.

FINDINGS

We identified genome-wide evidence for an association between bone mineral density and two SNPs (p<5x10(-8)). The SNPs were rs4355801, on chromosome 8, near to the TNFRSF11B (osteoprotegerin) gene, and rs3736228, on chromosome 11 in the LRP5 (lipoprotein-receptor-related protein) gene. A non-synonymous SNP in the LRP5 gene was associated with decreased bone mineral density (rs3736228, p=6.3x10(-12) for lumbar spine and p=1.9x10(-4) for femoral neck) and an increased risk of both osteoporotic fractures (odds ratio [OR] 1.3, 95% CI 1.09-1.52, p=0.002) and osteoporosis (OR 1.3, 1.08-1.63, p=0.008). Three SNPs near the TNFRSF11B gene were associated with decreased bone mineral density (top SNP, rs4355801: p=7.6x10(-10) for lumbar spine and p=3.3x10(-8) for femoral neck) and increased risk of osteoporosis (OR 1.2, 95% CI 1.01-1.42, p=0.038). For carriers of the risk allele at rs4355801, expression of TNFRSF11B in lymphoblast cell lines was halved (p=3.0x10(-6)). 1883 (22%) of 8557 people were at least heterozygous for these risk alleles, and these alleles had a cumulative association with bone mineral density (trend p=2.3x10(-17)). The presence of both risk alleles increased the risk of osteoporotic fractures (OR 1.3, 1.08-1.63, p=0.006) and this effect was independent of bone mineral density.

INTERPRETATION

Two gene variants of key biological proteins increase the risk of osteoporosis and osteoporotic fracture. The combined effect of these risk alleles on fractures is similar to that of most well-replicated environmental risk factors, and they are present in more than one in five white people, suggesting a potential role in screening.

Polymorphisms in ALOX12, but not ALOX15, are significantly associated with BMD in postmenopausal women

The murine arachidonate 15-lipoxygenase gene (Alox15) has recently been identified as a negative regulator of peak bone mineral density (BMD). The human ALOX15 gene shares significant sequence homology with the murine Alox15 gene; however, the human arachidonate 12-lipoxygenase gene (ALOX12) is functionally more similar to the mouse gene. Multiple single-nucleotide polymorphisms (SNPs) in the human ALOX15 and ALOX12 genes have previously been reported to be significantly associated with BMD in humans. On the basis of these data, we carried out our own investigation of the human ALOX15 and ALOX12 genes and their relationship with hip and spine BMD parameters. The study population consisted of 779 postmenopausal women with a mean (+/- standard deviation) age of 62.5 +/- 5.9 years at BMD measurement and was recruited from a single large general practice in Chingford, northeast London. Three SNPs from ALOX15 and five from ALOX12 were analyzed. None of the SNPs that we analyzed in ALOX15 were significantly associated with any of the BMD parameters or fracture data. However, we found that three SNPs from ALOX12, all previously associated with spine BMD in women, were significantly associated with spine and various hip BMD parameters in our cohort (P = 0.029-0.049). In conclusion, we found no association between polymorphism in ALOX15 and BMD phenotypes but were able to replicate previous findings that genetic variation in ALOX12 seems to play a role in determining bone structure in Caucasian women.

Polymorphism of the follistatin gene in polycystic ovary syndrome

Follistatin has been reported as a candidate gene for polycystic ovary syndrome (PCOS) from linkage and association studies. Acting to regulate the development of ovarian follicles and as an antagonist to aromatase activity, alterations in follistatin function or expression may result in key features of PCOS such as reduced serum FSH, impaired ovarian follicle development and augmented ovarian androgen production. We investigated polymorphisms in the FST gene to determine if genetic variation is associated with susceptibility to PCOS or key phenotypic features of PCOS patients in a case-control association study. One hundred and seventy-three PCOS patients of Caucasian descent (mean age 30.0 +/- 4.8 years), conforming to the NIH diagnostic criteria, were recruited from a clinical practice database and 107 normal ovulating women (mean age 38.8 +/- 13.4 years) were recruited from the general community as control subjects. Morphometric data, biochemistry and genomic DNA were collected from study subjects and genotyping was performed on seven Single nucleotide polymorphisms (SNPs) in the FST gene region. Allele frequencies of the SNPs were rs1423560 G/C (0.99/0.01), rs3797297 C/A (0.80/0.20), rs11745088 C/G (0.98/0.02), rs3203788 A/T (0.98/0.02) and rs1062809 G/C (1.00/-), rs1127760 A/T (0.98/0.02) and rs1127761 A/T (0.98/0.02), and these were not significantly different between the PCOS and control groups (P < 0.05). Statistical analysis revealed significant associations between the SNP rs3797297 and sex hormone-binding globulin (P = 0.04) and free androgen index (FAI) (P < 0.01). We conclude that FST is not a susceptibility locus for PCOS; however, the SNP rs3797297 from FST gene was associated with androgenic markers for PCOS and may be of importance in the hyperandrogenaemia of the disease.

No associations between OPG gene polymorphisms or serum levels and measures of osteoporosis in elderly Australian women

Bone mass is the single most important risk factor for osteoporotic fractures in the elderly and is mainly influenced by genetic factors accounting for 40-75% of the inter-individual variation. Critical for the bone remodeling process is the balance between the newly discovered members of the tumor necrosis factor ligand and receptor superfamilies, osteoprotegerin (OPG) and receptor activator of nuclear factor-kappaB ligand, which mediate the effects of many upstream regulators of bone metabolism. In the present study, we evaluated the impact of sequence variations in the OPG gene on bone mass, bone-related biochemistry including serum OPG and fracture frequency in elderly Australian women. A total of 1101 women were genotyped for 3 different single nucleotide polymorphisms (SNP) within the OPG gene (G1181C, T950C and A163G). The effects of these SNPs and serum OPG on calcaneal quantitative ultrasound measurements, osteodensitometry of the hip and bone-related biochemistry were examined. We found no significant relationship between sequence variations in the OPG gene or serum OPG and bone mass, bone-related biochemistry or fracture frequency. Our findings confirm some recent publications investigating the same SNPs but diverge from others, indicating that generalization of the relationships found in this type of study must be done with caution and signify the importance of determining associations between polymorphisms and osteoporosis in different ethnic groups.

Klotho gene polymorphisms are associated with osteocalcin levels but not bone density of aged postmenopausal women

Osteoporosis is known to have a strong genetic basis. It has been proposed that polymorphisms within the KL (klotho) gene have a significant effect on aging, in particular, the osteoblast defect of aging. The association between polymorphisms within this gene and biochemical markers of bone formation and resorption, bone structure, and fracture rates was studied in 1,190 postmenopausal women with a mean age of 75 years. Genotyping of these polymorphic sites was carried out using Matrix-Assisted Laser Desorption Ionization--Time of Flight (MALDI-ToF) mass spectrometry. The G allele of SNP c.1775G>A was associated with a lower osteocalcin level than the A allele (P = 0.004) in a codominant model. SNPs C-387T and IVS1+8262c>t both showed nonsignificant associations with osteocalcin (P values of 0.063 and 0.068, respectively), but a haplotype analysis of 2 of 5 haplotypes of the three SNPs with a frequency greater than 4% revealed a significant association with osteocalcin (P = 0.036). None of the individual polymorphisms or haplotypes analyzed showed any associations with a marker of bone resorption the deoxypyridinoline creatinine ratio, bone structure, or fracture data. Therefore, the G polymorphism within the c.1775G>A SNP site and a haplotype including this are associated with a reduced osteoblast product osteocalcin. These data suggest that variation in the KL gene product affects osteoblast activity independent of osteoclast activity but that this defect does not result in an effect on bone structure in this population, perhaps because of "rescue" by other genetic or environmental factors in this population.