Canet Luz M, Sánchez-Maldonado Jose M, Cáliz Rafael, Rodríguez-Ramos Ana, Lupiañez Carmen B, Canhão Helena, Martínez-Bueno Manuel, Escudero Alejandro, Segura-Catena Juana, Sorensen Signe B, Hetland Merete L, Soto-Pino María José, Ferrer Miguel A, García Antonio, Glintborg Bente, Filipescu Ileana, Pérez-Pampin Eva, González-Utrilla Alfonso, Nevot Miguel Ángel López, Conesa-Zamora Pablo, Broeder Alfons den, De Vita Salvatore, Jacobsen Sven Erik Hobe, Collantes-Estevez Eduardo, Quartuccio Luca, Canzian Federico, Fonseca João E, Coenen Marieke J H, Andersen Vibeke, Sainz Juan
Genomic Oncology Area, GENYO Centre for Genomics and Oncological Research, Pfizer / University of Granada / Andalusian Regional Government, PTS Granada, Granada, Spain.
Rheumatology Department, Virgen de las Nieves University Hospital, Granada, Spain.
Pharmacogenomics J. 2019 Feb;19(1):83-96. doi: 10.1038/s41397-018-0057-x. Epub 2018 Oct 5.
The aim of this case-control study was to evaluate whether 47 single-nucleotide polymorphisms (SNPs) in steroid hormone-related genes are associated with the risk of RA and anti-TNF drug response. We conducted a case-control study in 3 European populations including 2936 RA patients and 2197 healthy controls. Of those, a total of 1985 RA patients were treated with anti-TNF blockers. The association of potentially interesting markers in the discovery population was validated through meta-analysis with data from DREAM and DANBIO registries. Although none of the selected variants had a relevant role in modulating RA risk, the meta-analysis of the linear regression data with those from the DREAM and DANBIO registries showed a significant correlation of the CYP3A4 and CYP2C9 variants with changes in DAS28 after the administration of anti-TNF drugs (P = 0.00074 and P = 0.006, respectively). An overall haplotype analysis also showed that the ESR2 haplotype significantly associated with a reduced chance of having poor response to anti-TNF drugs (P = 0.0009). Finally, a ROC curve analysis confirmed that a model built with eight steroid hormone-related variants significantly improved the ability to predict drug response compared with the reference model including demographic and clinical variables (AUC = 0.633 vs. AUC = 0.556; P = 1.52 × 10). These data together with those reporting that the CYP3A4 and ESR2 SNPs correlate with the expression of TRIM4 and ESR2 mRNAs in PBMCs (ranging from P = 1.98 × 10 to P = 2.0 × 10), and that the CYP2C9 SNP modulates the efficiency of multiple drugs, suggest that steroid hormone-related genes may have a role in determining the response to anti-TNF drugs.KEY POINTS• Polymorphisms within the CYP3A4 and CYP2C9 loci correlate with changes in DAS28 after treatment with anti-TNF drugs.• A haplotype including eQTL SNPs within the ESR2 gene associates with better response to anti-TNF drugs.• A genetic model built with eight steroid hormone-related variants significantly improved the ability to predict drug response.
本病例对照研究的目的是评估类固醇激素相关基因中的47个单核苷酸多态性（SNP）是否与类风湿关节炎（RA）风险及抗TNF药物反应相关。我们在3个欧洲人群中开展了病例对照研究，包括2936例RA患者和2197名健康对照。其中，共有1985例RA患者接受了抗TNF阻滞剂治疗。通过与DREAM和DANBIO注册研究的数据进行荟萃分析，验证了发现人群中潜在有趣标记物的关联性。虽然所选变体均未在调节RA风险中发挥相关作用，但对线性回归数据与DREAM和DANBIO注册研究的数据进行的荟萃分析显示，CYP3A4和CYP2C9变体与抗TNF药物给药后DAS28的变化存在显著相关性（P分别为0.00074和0.006）。总体单倍型分析还显示，ESR2单倍型与抗TNF药物反应不佳的可能性降低显著相关（P = 0.0009）。最后，ROC曲线分析证实，与包含人口统计学和临床变量的参考模型相比，由8个类固醇激素相关变体构建的模型显著提高了预测药物反应的能力（AUC = 0.633对AUC = 0.556；P = 1.52×10）。这些数据以及报告CYP3A4和ESR2 SNP与PBMC中TRIM4和ESR2 mRNA表达相关（P范围为1.98×10至2.0×10），且CYP2C9 SNP调节多种药物效率的数据表明，类固醇激素相关基因可能在决定抗TNF药物反应中起作用。
关键点
• CYP3A4和CYP2C9基因座内的多态性与抗TNF药物治疗后DAS28的变化相关。
• 包含ESR2基因内eQTL SNP的单倍型与对抗TNF药物的更好反应相关。
• 由8个类固醇激素相关变体构建的遗传模型显著提高了预测药物反应的能力。
