Genome sequencing of patients with sporadic congenital hydrocephalus reveals mutations of large effect size indicative of a developmental origin for the disease.
Congenital hydrocephalus (CH) is a condition present at birth in which cerebrospinal fluid (CSF) accumulates in the brain and is associated with enlarged spaces called ‘ventricles’ in the center of the brain. Although CH is often considered to be distinct from acquired forms of hydrocephalus arising from a tissue response to injury, events that occur in the fetus in utero, such as hemorrhage or infection, may result in hydrocephalus that is apparent in the neonate. Epidemiological data suggest genetic causation in up to 40% of cases of CH. However, most cases are sporadic and phenotypes vary, which poses a substantial barrier to the identification of genetic causes. In this issue of Nature Medicine, Jin et al. report exome sequencing of a cohort of people with CH that identifies mutations indicative of early developmental aberrations1 (Fig. 1).
CH can be classified as communicating — without obstruction of CSF circulation — or non-communicating — typically due to stenosis (blockage) of the cerebral aqueduct, which is a narrow channel through which CSF flows from the ventricular system to the spaces surrounding the brain. The decision of whether to undertake neurosurgical shunting of the CSF, designed to alleviate increased intracranial pressure from excessive CSF accumulation, can be a difficult one. When the decision to shunt is based on clinical presentation or brain imaging alone, the procedure may not result in reduced ventricular size or improved functional outcomes. Knowledge of the genetic underpinning of CH in a patient could enhance prognostic accuracy and substantially assist clinical decision making. Only a handful of genes have been associated with hydrocephalus in infancy2. Most of the cases of CH for which there is a genetic understanding display aqueductal stenosis, and the genes responsible include X-linked L1CAM and AP1S2, or a few genes such as MPDZ and CCDC88C that cause severe autosomal recessive conditions that are not yet well delineated.
This study expands the authors’ earlier publication of genetic causes of CH that reported analyses of whole-exome sequencing (WES) of 125 parent–patient trios and found damaging de novo mutations (DNMs) in TRIM71, SMARCC1 and PTCH1, as well as de novo duplications in PTCH1 and SHH3. Here, Jin et al. use WES combined with deep phenotyping of 381 patients with neurosurgically treated sporadic CH of unknown genetic etiology1. The authors include 232 parent–offspring trios in this study, which allows them to identify mutations occurring de novo in the offspring, along with 149 single, previously unresolved cases. In the affected patients, they find that five genes (TRIM71, SMARCC1, PIK3CA, PTEN and FOXJ1) have more protein-damaging mutations than would be expected by chance; furthermore, they find that the genes MTOR, FMN2 and PTCH1 are affected by two or more damaging protein-coding alterations each and that the gene FXYD2 has a significant burden of inherited dominant mutations. The authors also identify rare recessive genotypes in consanguineous patients.
All of the CH-associated genes that show enrichment for pathogenic DNMs in this population encode molecules with known functions as regulators of the proliferation and/or differentiation of neural stem cells. Interestingly, the authors find that the genes they identify as being associated with CH risk overlap human fetal transcriptome data identified in two studies4,5, which leads them to infer that CH-related genes fall into two major classes indicative of either impaired CSF circulation or intrinsic abnormalities in cerebral cortical construction (Fig.1). While a favorable clinical response to CSF shunting may be anticipated for people who have mutations in the former class of genes, those with mutations in the latter class of genes may not benefit from the surgical intervention, and further investigation is warranted.
The authors further investigate their study patients with mutations in the genes encoding TRIM71, a post-transcriptional regulator, and SMARCC1, a chromatin remodeler, both of which are known to have roles in the regulation of neural stem cells. They find that the human hydrocephalus phenotypes are consistent with the hydrocephalus found in mice lacking these genes6,7. Moreover, they find that patients with TRIM71 mutations have substantial loss of white-matter volume and abnormalities of the corpus callosum, along with frequent features of cranial-nerve deficits, interhemispheric cysts, hearing loss and skeletal abnormalities. Their patients with SMARCC1 mutations have aqueductal stenosis and corpus-callosum abnormalities, along with cardiac and skeletal anomalies. Thus, the authors identify two likely new autosomal dominant syndromes with variable expressivity, as defined by mutations in these genes.
The mutations that the authors find to be associated with CH occur frequently in genes encoding components of the PIK3CA–MTOR pathway, including three DNMs in unrelated patients with CH who have macrocephaly, megalencephaly, polymicrogyria and craniofacial abnormalities. A number of the patients have missense mutations in genes encoding molecules responsible for known clinical syndromes, including MCAP (megalencephaly–capillary malformation–polymicrogyria) and PHTS (PTEN hamartoma tumor syndrome) and its subtypes. The authors find that a number of the patients with CH who have DNMs in genes encoding components of the PIK3CA–MTOR pathway have phenotypes that fit criteria for these diagnoses but have never been tested. Missing such a diagnosis may one day negatively affect patient care, as inhibitors of the MTOR pathway currently used in cancer therapies are now being explored for possible utility toward improving the clinical course of patients with MCAP or PHTS. Notably, the heterogeneity of genetic etiologies and variable expressivity of CH support the routine use of WES for newly diagnosed cases. This could inform prognosis and recurrence risk and, for some, could indicate the need for increased cancer surveillance (such as in the case of those with PIK3CA or PTEN mutations). Thus, the authors’ study has made a compelling argument for the present need to carry out WES on patients with CH — these people often have no genetic analyses, or testing is limited to karyotype and/or array comparative genomic hybridization.
In their study1, Jin et al. are able to overcome challenges posed by the complexity of the mutational etiologies of CH and the relative paucity of subjects available for study. Such a comparatively sizeable, well-characterized cohort is achieved through engagement of an international collaboration of clinicians and surgeons to identify and enroll participants and collect DNA samples. 22% of the cases could be accounted for by rare mutations with large effect size, of which ~2% have recessive or transmitted dominant genotypes. Using their data, the authors estimate that around 34 genes contribute to CH via DNMs, which suggests that there are substantially more than the five genes identified with high confidence here. Indeed, their calculations suggest that another 2,500 to 5,000 patients must be studied to identify all the remaining CH-associated genes. Nevertheless, these investigators have provided an intriguing road map for the fruitful investigation of complex genetic disorders through the use of modest case numbers, especially when particular gene mutations exert strong effects on phenotype. Investigation of the intergenic regions of the genomes of these patients may one day reveal additional gene associations or at least reduce the number of new cases that must be ascertained to identify more of the estimated 34 ‘risk genes’ that lead to CH.
Jin et al. provide evidence that genetic disruption of early brain development, rather than impaired CSF dynamics, is the main source of pathogenesis for a substantial number of patients with CH1. With additional analyses that include new cases, more-precise genotype–phenotype characterization may allow individually optimal treatment regimens (neurosurgical, pharmacological, neurobehavioral, etc.). A new era is approaching for increased accuracy of diagnosis and prognosis and improved intervention strategies for patients with this complex disease.
Jin, S.C. et al. Nat. Med. https://doi.org/10.1038/s41591-020-1090-2 (2020).
Tully, H. M. & Dobyns, W. B. Eur. J. Med. Genet. 57, 359–368 (2014).
Furey, C. G. et al. Neuron 99, 302–314.e4 (2018).
Walker, R. L. et al. Cell 179, 750–771.e22 (2019).
Polioudakis, D. et al. Neuron 103, 785–801.e8 (2019).
Welte, T. et al. Genes Dev. 33, 1221–1235 (2019).
Harmacek, L. et al. Dev. Neurobiol. 74, 483–497 (2014).
The author declares no competing interests.
About this article
Cite this article
Ross, M.E. Unlocking the genetic complexity of congenital hydrocephalus. Nat Med 26, 1682–1683 (2020). https://doi.org/10.1038/s41591-020-1120-0