SAT0105 Preliminary analysis of genetic variants in the immune system related to the body mass index in early arthritis patients

• Published in: Annals of the rheumatic diseases Vol. 77; no. Suppl 2; p. 915
• Main Authors: Moreno, P., Montes, N., Martín, J., Carmona, D., Mora, C.M., Gomáriz, R.P., Lamana, A., Triguero, A., Ortiz, A. M., Álvaro, I.G.
• Format: Journal Article
• Language: English
• Published: London BMJ Publishing Group LTD 01.06.2018
• Subjects: Alleles; Arthritis; Association analysis; Body mass index; Epitopes; Gene frequency; Genetic analysis; Genetic diversity; Genomes; Genotyping; Histocompatibility antigen HLA; Immune system; Learning algorithms; Polymorphism; Single-nucleotide polymorphism
• ISSN: 0003-4967; 1468-2060
• DOI: 10.1136/annrheumdis-2018-eular.3575

BackgroundWe have observed in previous analyzes in our early arthritis (EA) cohort that patients with a higher body mass index (BMI) are, more frequently, ACPA negative and these patients carry, with a lower frequency, HLADRB1 alleles that encode for the shared epitope.ObjectivesTo identify SNPs (Single Nucleotide Polymorphisms) of immune system genes related to BMI in EA patients.MethodsThe 257 patients of the PEARL (Princess Early Arthritis Register Longitudinal) cohort in which high density genotyping was available (using the Immunochip array of Ilumina Inc) were included. As a previous step, those SNPs that did not meet the requirements of a genotyping call rate lower than 98%, being out of Hardy-Weinberg equilibrium (p<10-4) and minor allele frequency lower than 1% were excluded. IMPUTE v.2 was used for the genotype imputation of the SNPs that failed in the immunochip, using as reference the data of phase III of the 1000 G project. The association analysis of the remaining SNPs was made by linear regression adjusted by sex, age and study level with PLINK v1.9. Of the 1384 SNPs associated with BMI with a value of p<0.01, 250 SNPs were selected according to the lowest values of the division of p divided by the absolute value of its β coefficient. After analyzing and excluding the SNPS that were in linkage desequilibrium, the importance of the 186 resulting SNPs was quantified with the "Random Forest" and "Boosted Regression Tree" techniques using %IncMSE (Mean Decrease Accuracy).ResultsTable 1 shows the selection of the 15 SNPs that were more important in both "machine learning" techniques according to BMI. Although most of these SNPs are located in non-coding regions (intergenic or intronic), some of the genes where the SNPs belong or the neighboring genes have shown association in some GWAS (Genome-Wide Association) with a minor (BMP7) or a greater (RSPO3) BMI; and some of them have shown to have a regulatory role in the immune system in patients with RA (WDFY4, BMP7).Table 1SNPs related to BMI in our early arthritis registry.SNP Gene β Coef. [CI 95%]p rs2746187LOC728666/RSPO32.018 [0.949,3.087]2.646x10-4 rs8103026SIGLEC6/ZNF175-2.135 [-3.09,-1.18]1.724x10-5 rs2419678LOC100132349-2.001 [-2.826,-1.175]6.623x10-6 rs17842463SULT2B1-3.515 [-5.446,-1.585]3.402x10-3 rs1131878UGT2B41.337 [0.542,2.131]1.114x10-3 rs12757445CDC73/KCNT21.804 [0.737,2.871]1.057x10-3 rs1638020PTPRN2-1.329 [-2.1,-0.558]8.409x10-4 rs11019575MULTIPLE GENES: 399942, 1001313643.356 [1.309,4.307]2.939x10-4 rs72917213MEX3C/LOC7290512.319 [1.172,3.466]9.677x10-5 rs6014959BMP7-1.942 [-3.122,-0.762]1.434x10-3 rs2870662DOK5/CBLN41.455 [0.615,2.295]7.993x10-4 rs10776644WDFY4-2.159 [-3.49,-0.828]1.661x10-3 rs7800039STEAP4/ZNF804B1.564 [0.76-2.368]1.725x10-4 rs12517451ANKRD34B/DHFR1.629 [0.695,2.562]7.351x10-4 rs12722531IL2RA-3.544 [-5.713,-1.376]1.535x10-3 ConclusionsOur preliminary approach allowed us to select 15 SNPs that may have more relevance related to BMI in patients with early arthritis. However, this is a preliminary study and it is necessary to validate these results in other populations to ensure their involvement in the relationship between the BMI and EA.Disclosure of Interest:None declared