Background/Purpose

Previous studies suggested that blood lead level is associated with multiple joint symptoms. In this study, we conducted a Genome-Wide Gene-Environment Interactions analysis to search for genetic variations which may modify such associations.

Methods

Caucasians from the Johnston County Osteoarthritis Project (JoCo) with valid covariates and cleaned genotype data were included in this study. A composite score of multiple joint symptoms was calculated as the mean of summed 0-3 scores (none=0, 1=mild, 2=moderate, 3=severe) from seven joint sites including the hands, knees, hips and low back as previously reported.  Whole blood lead level was measured by inductively coupled plasma-dynamic reaction cell-mass spectrometer analysis at the Centers for Disease Control and Prevention (Atlanta, GA).  Genotyping was done using Illumina Infinium 1M-Duo bead arrays at Expression Analysis (Durham, NC). Original SNP data were imputed by software MACH using HapMap II Caucasian as the reference data. Linear models were applied across the genome-wide genetic data with adjustment for age, body mass index (BMI), sex, current alcohol use, smoking, and blood lead level. The residuals of the regression followed the normal distribution. 

Results

 Participants included 805 individuals, 63% of whom were women, with mean age 67 years (SD=10.5) and mean BMI of 30 (SD=6.4) kg/m². Log transformed blood lead level (median=0.53) was associated with multiple joint symptoms score with a p-value at 0.02. Two interactions between SNPs (rs2885880 and rs1598457) and lead significantly modified lead association with the multiple joint symptoms score (interaction P-value < 5x10E-08).  Rs2885880 is located in gene RNF121 (ring finger protein 121). Based on Gene Ontology (GO) annotations, RNF121 might be related to zinc ion binding. Rs1598457 is located in gene GUCY1A2 (Guanylate Cyclase 1, Soluble, Alpha 2). Guanylyl Cyclases (GC) are a group of enzymes whose function requires the presence of metal ions such as Mg2+ or Mn2+. An additional promising interaction between fifteen SNPs (Table) and lead were identified (interaction p-values < 1E-06). For all the identified SNPs, the main effects of those SNPs were also significant with p-values < 0.05.

Conclusion

 Two interactions between lead and SNPs were identified by the Genome-Wide Gene-Environment Interaction analysis. Those two SNPs are located in RNF121 and GUCY1A2, both related to handling of metals. These two genes may influence the impact of blood lead level upon symptoms in multiple joints.  These results will require validation in independent datasets.