Association of assisted reproductive technology, germline de novo mutations and congenital heart defects in a prospective birth cohort study

Abstract

Emerging evidence suggests that children conceived through assisted reproductive technology (ART) have a higher risk of congenital heart defects (CHDs) even when there is no family history. De novo mutation (DNM) is a well-known cause of sporadic congenital diseases; however, whether ART procedures increase the number of germline DNM (gDNM) has not yet been well studied. Here, we performed whole-genome sequencing of 1137 individuals from 160 families conceived through ART and 205 families conceived spontaneously. Children conceived via ART carried 4.59 more gDNMs than children conceived spontaneously, including 3.32 paternal and 1.26 maternal DNMs, after correcting for parental age at conception, cigarette smoking, alcohol drinking, and exercise behaviors. Paternal DNMs in offspring conceived via ART are characterized by C>T substitutions at CpG sites, which potentially affect protein-coding genes and are significantly associated with the increased risk of CHD. In addition, the accumulation of non-coding functional mutations was independently associated with CHD and 87.9% of the mutations were originated from the father. Among ART offspring, infertility of the father was associated with elevated paternal DNMs; usage of both recombinant and urinary follicle-stimulating hormone and high-dosage human chorionic gonadotropin trigger was associated with an increase of maternal DNMs. In sum, the increased gDNMs in offspring conceived by ART were primarily originated from fathers, indicating that ART itself may not be a major reason for the accumulation of gDNMs. Our findings emphasize the importance of evaluating the germline status of the fathers in families with the use of ART.

Fig. 1. Offspring conceived through ART showed significantly more total (left), paternal (middle), and maternal (right) DNMs than those conceived spontaneously.

ART assisted reproductive technology, ARTP assisted reproductive technology pregnancies, SP spontaneous pregnancies, DNMs, de novo mutations.

Fig. 2. The influence of ART on the mutational spectrum.

a Relative frequency of mutational classes by parental origins. b The odds ratio of proportion comparison for specific mutational classes between the ARTP and SP groups. The P value and odds ratios were adjusted by Bonferroni methods. c The association (mutation rate ratio) between specific mutation class and ARTP/SP in the multivariable models. d The odds ratio of proportion comparison for specific functional groups between the ARTP and SP groups (IG, intergenic; NG, nearby genes; IN, intron; S, synonymous; ML, missense mutations with CADD ≤ 15; MH, missense mutations with CADD scores >  15). e The proportion of C>T mutations at the CpG sites in each functional group. *P < 0.05.

Fig. 3. Structural equation models showed that paternal CpG>TpG DNMs and non-coding functional DNMs acted as independent mediators between ART and risk of CHD.

The solid lines denoted significant (P < 0.05) relationships and dashed lines non-significant. *P < 0.05. CHD congenital heart defect, OR odds ratio, CI confidence interval.