Exome and Genome Sequencing of Neonates With Neurodevelopmental Disorders

Keywords:

• mutation interpretation
• return of results
• whole-genome sequencing

Genetic disorders of the nervous system, encompassing both developmental and degenerative phenotypes, have been at the forefront of gene discovery since the beginning of the molecular genetic era. Cumulatively, more than half of cloned Mendelian disease genes [101] are causally associated with either isolated disorders of the nervous system, or with syndromes with significant neuropathology. Progress is also being made in neurological disorders of complex genetic architecture, most notably schizophrenia and autism [1]. There are many reasons for the apparent overrepresentation of this phenotypic class on the morbid map of the human genome. These include our fascination with the complexity of the brain, the fact that patients with neurological defects have a higher likelihood to seek physician help and, from a biological perspective, the reduced regenerative capacity of neurons compared with other cell types. Moreover, there are more genes expressed in the developing CNS compared with most other tissues, potentially offering a larger mutational target [2,3]. Finally, neurodevelopmental traits have an overt impact on reproductive fitness, thereby allowing more efficient identification of pathogenic mutations; not only are the majority of causal alleles likely to be found in patients unique, but emerging evidence suggests that neurodevelopmental and neurocognitive disorders might have a higher incidence of de novo mutations compared with other classes of genetic diseases [4–6].

Subsequent to the demonstration that it is possible to sequence the known coding regions of the genome (the exome) [7], disease-gene discovery has hyper-accelerated. Advances in molecular methodologies and statistical analytic approaches offer the promise of transforming human genetics from a secondary tool used to substantiate a clinical hypothesis, to a primary diagnostic tool [8]. There is every indication that genetic studies of neurodevelopmental and neurodegenerative traits will be a major beneficiary of this paradigm shift in DNA sequencing and analysis. Not surprisingly, the field remains in flux: the scientific and medical communities have yet to reach consensus on how to interpret the information and how to deliver to physicians, patients and patient families meaningful data that can inform clinical management while respecting patient sensitivity and privacy.

Among the many challenges raised is the question of when to pursue whole-exome and/or whole-genome sequencing (WES/WGS) and for what purpose. Current expert recommendations for screening individual patients with otherwise unexplained developmental delay/intellectual disability, autism spectrum disorders or multiple congenital anomalies include a G-banded karyotype and chromosomal microarray. Chromosomal microarray offers a diagnostic yield of 15–20% for genetic testing of individuals with unexplained developmental delay/intellectual disability, autistic spectrum disorders or multiple congenital anomalies, while G-banded karyotyping provides answers in approximately 3% of cases, excluding Down syndrome and other recognizable chromosomal syndromes [9,10].

Much of the current WES/WGS effort is focused on either historical patient cohorts or, in the immediate clinical setting, families who have exhausted all other traditional diagnostic tools such as karyotypes, microarrays and testing for specific candidate genes known to be associated with similar phenotypes. A fundamental reason for this choice is that these large-scale sequencing approaches are considered to have the highest probability of delivering interpretable data; the sequencing of multiple historical cohorts collected with a particular phenotypic focus offers the significant advantage of identifying both recurring mutations as well as large allelic series within specific genes, thus providing rigorous evidence for causality. Likewise, sequencing of a family after a phenotype has manifested fully and has been ascertained by exhaustive clinical investigation can also inform WES/WGS data interpretation.

Waiting for multiple additional tests or only targeting individuals with fully manifested phenotypes is a reasonable approach to use of WES/WGS in adolescents and adults. However, 80% of young children (including neonates) with developmental delay/intellectual disability and/or multiple congenital anomalies do not have fully developed phenotypes and are not diagnosed by current standard approaches. Thus, there is an argument to be made for offering WES/WGS soon after the first suspicion of a genetic defect, cognizant of the fact that interpretation is more challenging and likely to require interdisciplinary teams to parse the data. Our view is that early integration of genomic information is a starting point for clinical investigation that demands ongoing evaluation of the risks and benefits of this approach. Considering that a single methodology with sufficient sensitivity and specificity remains elusive, both academic and for-profit entities have started offering WES and WGS with interpretation usually limited to genes and noncoding variants already known to be causally associated with clinical phenotypes.

Although restricting one‘s universe to known pathogenic variants can be beneficial, the imposed filters for allele frequency in populations and computational prediction of the effect of candidate mutations remain limited and can intensify both false positive and false negative data that increase anxiety, uncertainty and cost. We hold the view that more unbiased approaches, such as WES/WGS, founded on the synthesis of clinical observation, population genetics, molecular biology and functional testing of candidate variants, might offer higher quality and higher fidelity interpretive capacity. Under this premise, we formed the Task Force for Neonatal Genomics at Duke University (NC, USA) [102], a multidisciplinary unit composed of physicians, geneticists, cell biologists and ethics/policy experts charged with asking the question of whether ab initio WES/WGS of neonates and infants with congenital defects of suspected genetic etiology can expedite diagnosis and aid clinical management. Considering the aforementioned interpretive challenges, our approach mandates that, with the exception of known pathogenic mutations, the task force develop physiologically relevant assays to test the functionality of all alleles found in a patient‘s genome whose pattern of inheritance and population frequency are consistent with causality. Although our experience is still modest, with some 20 families ascertained over the course of the past year, some themes have already emerged, including the fact that some 80% of the referred cases from the neonatal intensive care unit and pediatric subspecialty clinics are driven by neurocognitive or neuroanatomical abnormalities of unknown etiology. Our initial data are encouraging: of the 15 families we have analyzed to date (from clinical assessment to complete analysis of exomes, including functionalization of alleles), each patient has required functional analysis of five to ten genes. We have been able to reach either definitive or strongly suspected diagnoses for approximately 80% of cases, exceeding what would have been possible by computational approaches alone [11,12]. Moreover, this unbiased method has the ability to identify multiple genetic defects within the same patient and, in the fullness of time, offers the promise of being able to extract genomic data that can inform both primary disease causality and modification of penetrance/expressivity of severe phenotypes. We also note that, despite the labor involved in the functional testing of a larger number of alleles, the window of this genomic intervention is measured in months, not years, and has the tangible possibility of offering focused subsequent clinical investigations that can mitigate risk and cost significantly.

There is healthy skepticism about whether WES/WGS should be implemented widely at this early stage and what data could/should be returned to physicians and families. Amidst this controversy, however, it is important to note that fundamentally, WES/WGS in fetuses and neonates is little more than an increase in resolution of current practices. Since the early days of cytogenetics (and now molecular cytogenetics), infants and young children with nondescript developmental delay and/or brain MRI abnormalities have been screened routinely by microarray analysis. WES/WGS offers a significant increase in resolution of essentially the same test, especially since emergent tools can now capture both chromosomal rearrangements and point mutations for the same effort and cost [13]. One question, therefore, is not whether to offer this investigative avenue to families, but how to extract data that are relevant to an acute clinical phenotype from hundreds of alleles that appear unique to any given patient. Another question is what to do with the information about previously published risk alleles for relatively common late-onset genetic diseases such as breast cancer and Alzheimer disease. We consider our prospective cohort to be a ‘living laboratory‘ in which to gauge familial interest in and response to this information and other secondary alleles unrelated to patients‘ acute life-threatening phenotypes.

Despite these challenges and the trepidation that can accompany them, the benefits outweigh the risks; examples are emerging in the literature wherein genomic data were not only catalysts for diagnosis, but also offered a path to hitherto unappreciated treatment options. For example, the discovery of the so-called “Lifetech twins” led to a dramatic improvement in the health of two brothers with undiagnosed dystonia [14]. Arguably, the discovery of pathogenic lesions at the beginning of life is likely to maximize the possibilities for therapeutic intervention. For example, allogeneic stem cell transplantation, although still emerging, offers a realistic therapeutic paradigm, especially if deployed early. This approach has already been successful and there is every indication that it can become one therapeutic option should genomic data and functional testing indicate loss-of-function, cell-autonomous mutations [15,16]. Likewise, other options such as gene therapy could become viable alternatives secondary to early genetic diagnosis. For example, gene therapy for retinal degenerative disorders is becoming a reality but can only be deployed meaningfully if the primary genetic lesion is known and treatment is administered prior to substantial photoreceptor loss [17].

At the very least, given that the vast majority of genetic disorders remain therapeutically intractable, WES/WGS data can empower physicians, families and the proximal community of patients by mitigating uncertainty and enabling prospective management. Every novel variant unequivocally linked to the manifestation of a severe disease phenotype represents a potential new therapeutic target. Was this not the promise of the Human Genome Project?

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.