Molecular diagnosis of osteogenesis imperfecta through massively parallel sequencing of 15 candidate genes

Osteogênese imperfeita (OI) é um conjunto de displasias esqueléticas hereditárias, heterogêneas dos pontos de vista clínico e genético. A maioria dos casos resulta de defeitos no colágeno tipo 1, entretanto diversos outros genes candidatos vêm sendo descritos. Não há definição de critérios mínimos para o diagnóstico clínico de OI, portanto o diagnóstico molecular vem permitir precisão diagnóstica, e melhor compreensão fisiopatológica e prognóstica. O objetivo deste estudo foi identificar o diagnóstico molecular da OI através da análise de genes candidatos por sequenciamento paralelo em larga escala (SPLE), buscando-se associar características clínicas ao achado molecular. Para tanto, as regiões codificadoras e junções íntron-éxon dos 15 genes candidatos COL1A1, COL1A2, CRTAP, P3H1, PPIB, SERPINH1, FKBP10, PLOD2, BMP1, IFITM5, SERPINF1, WNT1, TMEM38B, SP7 e CREB3L1 foram capturadas e sequenciadas por SPLE na plataforma Illumina. Variantes genéticas raras (frequência alélica 50 depth in 99% of target regions. A molecular diagnosis was obtained in 37 out of 38 cases (97%). COL1A1 variants were found in 18 cases, and COL1A2 variants in 9 cases, meaning that 71% of cases had type 1 collagen-related OI. Variants in the other candidate genes were found in 10 sporadic cases: SERPINF1 (n=2), FKBP10 (n=2), PLOD2 (n=2), IFITM5 (n=1), P3H1 (n=1), TMEM38B (n=1) and WNT1 + P3H1 (n=1). Only one CNV was found (deletion of exons 1 and 2 of TMEM38B). Among the 42 identified variants, 23 had already been reported in association to OI and 19 are novel variants. Identified COL1A1 and COL1A2 variants (56% glycine substitution) are scattered throughout the proteins, and associated to wide phenotypic variability, as demonstrated by the COL1A2 p.(Gly772Ser) variant found in three unrelated cases with different clinical presentations. Additional genetic variants that could help explain the phenotypic variability were not found. Variants in non-collagen genes were associated to severe OI (p=0.012), variants in COL1A1 to blue sclerae (p=0.009), and, amongst COL1A1 and COL1A2 variants, those resulting in glycine substitution were associated to dentinogenesis imperfecta (p=0.003). In conclusion, this study shows that the molecular diagnosis of OI through MPS is effective. In comparison to the literature, a higher proportion of non-collagen defects was found, which can be due to the OI severity of included patients or to a peculiarity of OI in Brazil, given that our population is underrepresented in molecular studies of OI. Hopefully, the diagnostic precision enabled by MPS will allow personalized treatment and follow-up of individuals with O