Discussion
Main Findings
In the present study, we found that abnormal CNVs are associated with
several fetal structural malformations among stillborn fetuses. A
pathogenic CNV deletion of 1q21.1 involving 46 genes (e.g.,CHD1L ) and a duplication of 21q22.13 involving 4 genes
(SIM2 , CLDN14 , CHAF1B , HLCS ) were associated
with a skeletal and cardiac defect, respectively.
Interpretation
In light of a previous large study that examined all
fetuses,5 our findings of CNVs in gene-rich regions
that are associated with fetal cardiac and skeletal anomalies in
stillborn fetuses is noteworthy. Donnelly et. al. showed that
chromosomal microarray analysis remained most informative in
particularly identifying fetuses with cardiac and craniofacial
anomalies, and most CNVs occurred in those with cardiac
anomalies.5 In general, fetuses with anomalies in more
than one system had higher frequency of abnormal CNVs compared with
fetuses without anomalies (13.0%; p-value<0.001), suggesting
that chromosomal microarray analysis identifies anomalies dependent on
the type and number of organ systems involved.
Cardiac abnormalities are among the most common congenital
malformations, associated with stillbirth13 and
present in an estimated 0.8% of live births.8Skeletal defects involve a heterogeneous group of bone abnormalities
resulting in abnormal growth and shape of the fetal
skeleton.24 Skeletal defects in the first trimester
are associated with nuchal translucent, and in lethal dysplasias, bone
shortening may be obvious as early as 11 weeks’
gestation.24 Among fetuses with perinatal diagnosis of
ventricular septal defects, most common cardiac defects, chromosomal
microarray was particularly effective in identifying pathogenic
CNVs.3 Thus, gene may potentially play a role in the
etiology of fetal ventricular septal defects.
Deletion of 1q21.1 (4.0 Mb in size) and duplication of 21q22.13 (500 kb
in size) CNVs were previously described as pathogenic abnormalities of
stillbirth in the same study.4 However, these CNVs
were not described in the context of fetal skeletal or cardiac defects.
Expression analyses of CHD1L (chromodomain helicase DNA binding
protein 1 like), among 46 genes implicated in 1q21.1 region in our
study, showed that it promotes neuronal differentiation of human
embryonic cells. This suggested that CHD1L plays an important
role in the nervous system development.25 In addition,
a pathogenic variant in CHD1L is also associated with short
stature.26 In the adult population, distal variant
CNVs of 1q21.1 were reported to be associated with cerebral and
cognitive alterations.27 Furthermore, high prevalence
of neurodevelopmental disorders are associated with CNVs at
1q21.28 Specifically, the CNV microdeletion and
microduplication at 1q21 were enriched in patients with schizophrenia,
intellectual disabilities and autism spectrum
disorder.28-30 These studies highlight the
pathophysiological mechanisms of CNVs in 1q21 in neurodevelopmental
disorders.
Limitations and strengths
While our study has the potential to provide high-quality data of
chromosomal abnormalities in fetuses suspected to have isolated cardiac
and skeletal malformations, several limitations exist. First,
microarray-based analyses may be limited to detect truly balanced
rearrangements of chromosomes. The large chromosomal regions identified
span several genes, and the Affy500K SNP Array is limited to detect
genetic abnormalities at higher resolution, e.g., specific mutations in
genes for clinical application. Up-to-date microarray platforms would be
sensitive to detect CNVs at greater resolution (e.g., >150
kb), and with increased sequencing our knowledge of the pathogenicity of
variants may change.31 Next-generation sequencing of
cases will improve diagnostic yield of genes in fetal structural
malformations.32 Further, due to lack of paternal DNA,
we were not able to distinguish inherited from newly occurring CNVs in
the placenta or fetus, which will be important future areas of
investigation for identifying causative of genetic abnormalities.
Finally, we have limited sample size to detect associations of CNVs with
other fetal structural malformations. As resolution for determination of
pathogenicity of CNVs is improved, CNVs that are thought to be benign
may eventually be designated as pathogenic, increasing the sample size
of stillborn fetuses identified as having abnormal CNVs in our study.
Several strengths of our study deserve mention. The population included
in our study is geographically, ethnically and racially diverse,
potentially ensuring the generalizability of our findings. Moreover,
participants provided consent for a complete evaluation, including fetal
standardized postmortem examination, placental pathological analysis,
and maternal-fetal testing. These study design features provided careful
phenotyping of stillbirth in our study, ensured accurate ascertainment
of fetal structural malformations, and maximized the validity of the
present analysis.
Conclusion
Among stillborn fetuses, specific CNVs involving several genes were
associated with fetal structural malformations and warrant further
investigation. These data improve genotype-phenotype databases,
potentially informing more precise understanding of genetic etiologies
of human developmental disorders.6 The data may also
inform counseling and care surrounding pregnancies affected by fetal
structural malformations at risk for stillbirth.