Background The clinical significance of 16p13.11 duplications remains controversial while frequently detected in patients with developmental delay (DD), intellectual deficiency (ID) or autism spectrum disorder (ASD). Previously reported patients were not or poorly characterised. The absence of consensual recommendations leads to interpretation discrepancy and makes genetic counselling challenging. This study aims to decipher the genotype–phenotype correlations to improve genetic counselling and patients’ medical care.
Methods We retrospectively analysed data from 16 013 patients referred to 12 genetic centers for DD, ID or ASD, and who had a chromosomal microarray analysis. The referring geneticists of patients for whom a 16p13.11 duplication was detected were asked to complete a questionnaire for detailed clinical and genetic data for the patients and their parents.
Results Clinical features are mainly speech delay and learning disabilities followed by ASD. A significant risk of cardiovascular disease was noted. About 90% of the patients inherited the duplication from a parent. At least one out of four parents carrying the duplication displayed a similar phenotype to the propositus. Genotype–phenotype correlations show no impact of the size of the duplicated segment on the severity of the phenotype. However, NDE1 and miR-484 seem to have an essential role in the neurocognitive phenotype.
Conclusion Our study shows that 16p13.11 microduplications are likely pathogenic when detected in the context of DD/ID/ASD and supports an essential role of NDE1 and miR-484 in the neurocognitive phenotype. Moreover, it suggests the need for cardiac evaluation and follow-up and a large study to evaluate the aortic disease risk.
- neurodevelopmental disorder
- 16p13.11 duplication
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Over the past years, chromosomal microarray analysis (CMA) replaced karyotyping as the first line test for patients presenting with intellectual deficiency (ID), developmental delay (DD) or autism spectrum disorder (ASD). This technique proved to be highly effective in identifying numerous disease-causing copy number variants (CNVs). On the other hand, it also revealed several variants of unclear (clinical) significance (VOUS). VOUS are defined as CNVs that are detected in affected patients as well as in control populations. They are usually inherited from apparently healthy parents and are associated with variable phenotypic traits making their clinical interpretation challenging.
The short arm of chromosome 16 is involved in many unbalanced structural variants, deletions and duplications, predominantly in the 16p13.11 region (figure 1). The latter is subdivided into three intervals (I, II and III), each flanked by low copy repeats (LCRs)-rich sequences also called segmental duplications (figure 1). These repetitive elements are believed to increase the risk of chromosomal rearrangements via non-allelic homologous recombination (NAHR). 16p13.11 Deletions are considered as pathogenic, whereas the significance of their reciprocal duplications is still under debate. Hannes et al considered them as rare benign variants,1 whereas other studies concluded to their clinical relevance based on their association with congenital heart defects, skeletal malformations and a wide range of neurodevelopmental disorders including ID, autism and schizophrenia.2–4 To date, <30 patients have been described in the literature most of whom are not precisely characterised from the molecular point of view.1–5
In the absence of a consensus on how to classify this CNV, its medical interpretation varies between the different genetic laboratories which can lead to misdiagnosis. This retrospective study aimed to refine the associated phenotype, assess the clinical significance of the variant and help to address the genotype–phenotype correlations. Our study contributes to highlight substantial elements to take into consideration for genetic counselling to the patients and their families and shed some light into possible mechanisms.
Patients and methods
Patients were referred to 11 Frenchand one Belgian genetic centers, for various developmental disorders raging from learning delay to ID with or without associated congenital malformations. All patients were examined by clinical geneticists and/or experienced paediatricians and benefited from a pangenomic CMA. Physicians in charge of patients for whom a 16p13.11 duplication was diagnosed were asked to fulfil a detailed questionnaire about clinical and cytogenetic data for the patients and their parents. Informed consent was obtained from all patients and/or parents for a genetic testing in concordance with the French ethical policies.
Chromosomal microarray analysis
CMA was performed using DNA from peripheral blood lymphocytes. DNA was extracted using commercial kits following manufacturer’s instructions. Distinct platforms of oligonucleotide microarrays have been used: (1) Agilent (Agilent Technologies, Santa Clara, California, USA) 44 K for Patients #1, 4, 11, 14–16, 36 and 38; 60K for Patients #5–8, 12, 13, 27, 29, 32, 44 and 45; 105K for Patients #10, 23, 34 and 37; or 180K for Patients #2, 3, 9, 17–20, 24–26, 28, 35, 39–43. (2) Illumina (Illumina, San Diego, California, USA) HumanHap 300 for patients 21 and 22, and HumanCytoSNP-12 for patients #30, 31 and 33. Genome annotation for the results is based on the Human Feb. 2009 (GRCh37/hg19) Assembly.
Confirmation methods for the detected variants
To confirm the detected duplications, we performed fluorescence in situ hybridisation, real-time PCR or quantitative multiplex PCR of short fluorescent fragments, depending on the size of the duplication. We also investigated the parental inheritance for 34 patients using the same confirmation assays that we used for the related patient.
We compared the prevalence of the duplication in our studied population with the prevalence in a control population which we estimated from the new curated variants database of the DGV website called ‘DGV gold standard variants’. We considered only duplications ≥200 kb and overlapping at least one of the duplications identified in our patients (variant identifiers: gssvG14132, gssvG14133 and gssvG14154) . We calculated the odds ratio (OR) and assessed statistical significance using the χ2 test.
We also estimated the penetrance as follows:
Penetrance = baseline risk of disease x (%CNV in affected population / %CNV in control population)
The background risk for congenital anomalies and developmental disorders was assumed to be 5.12%.6 The percentage of 16p13.11 duplications in the affected population is the prevalence calculated in our cohort. The percentage of 16p13.11 in a control population was calculated from the studies reported in the new curated database of variants in DGV as explaned above.
Clinical and epidemiological data
Detailed clinical information for each patient is provided in the online supplementary table 1 and summarised in table 1. Ages at diagnosis range from 6 months to 25 years old. The median is 8.5 years old, with 85% of patients who have been diagnosed after 3 years of age. For only one patient, the duplication was diagnosed during the prenatal period (patient #16).
The male to female sex ratio in our 16p13.11 duplication cohort is 0.96 against 1.48 in the CMA population.
In terms of clinical features, the prenatal period is usually featureless. In the perinatal period, hypotonia and feeding issues are present in 16% and 33% of the patients, respectively. Later, in early childhood, the most common features are speech delay and learning disabilities that are reported in >80% of the patients. ID is usually mild to moderate, requiring speech therapy and adequate support in school, but is also severe in one-third of the cases. Although most patients had some dysmorphic features, no notable pathognomonic craniofacial description could be drawn out of it.
As already reported, cardiac malformations were present in 23% of patients including the previously reported ones.
The prevalence of the duplication in our DD/ID/ASD cohort is 0.28% compared with 0.17% in a control population (variant identifiers: gssvG14132, gssvG14133 and gssvG14154). In a total of 16 132 control individuals, 28 have 16p13.11 duplication. The calculated OR is 1.62 (95%CI 1.01 to 2.60, p<0.05).
The penetrance in the affected population can be calculated as follows: 5.12% × (0.28%/0.17%) = 8.43% (95% CI 5.17% to 13.31%).
CMA results are described in table 2. Duplicated regions extend from 14,703,446 bp to 18,141,051 bp (hg19/GRCh37) with a size ranging from ~0.16 to 2.62 Mb. Patients’ tracks are represented in figure 1, which was generated using the DGV website (http://dgv.tcag.ca).
The examination of blood samples from the parents of 34 patients revealed that only 3 duplications occurred de novo and 31 (91%) were inherited. The duplication was transmitted from the mother in 20 cases out of the 31 and from the father in the remaining 11, showing a bias for a transmission through the mother. Clinical data were available for 23 parents who transmitted the duplication. Five of them (22%) displayed a cognition phenotype ranging from learning disabilities in their childhood to ID. This pinpoints to the need of an accurate evaluation of the carrier parent.
16p13.11 Genomic region is subdivided into three intervals (I, II and III) each flanked by LCR-rich sequences, also called segmental duplications (figure 1). Most of the duplications that we describe here, 93%, are characterised by breakpoints within those LCRs, suggesting that the duplication resulted from NAHR. All of them involve interval II. Two patients carried atypical smaller duplications (patient #2 and #23).
In some microdeletion/microduplication syndromes with incomplete penetrance and variable expressivity, a double hit model has been suggested to explain the variable phenotype. Our cohort did not show any evidence for a recurrent association with another CNV in relation with the phenotype variability.
16p13.11 Duplication is a recurrent NAHR-driven CNV that is observed in patients who are referred to genetic medical centers mainly for DD/ID or ASD. The interpretation of its clinical significance is highly challenging since the CNV is usually inherited from an apparently healthy parent. This leads to discrepancies in the interpretation of this CNV between different labs. The literature on the subject is poor, which makes difficult the establishment of consensual recommendations. We report here a comprehensive clinical and cytogenetic analysis of the largest reported cohort of patients with 16p13.11 duplication in a population of >16,000 patients who were investigated using CMA.
The most common clinical features were speech delay (88%) and learning disabilities/ID (86%) followed by ASD (67%), with a wide range of severity. Abnormal brain MRI and cardiac malformations were found in 24% and 23% of patients, respectively. Kuang et al suggested that 16p13.11 duplication confers a higher risk for thoracic aortic aneurysm.7 In our cohort, one patient displayed aortic dilatation at a very young age (patient #42) and the mother of patient #20, who also carried the duplication, died of ruptured aneurysm. Furthermore, the mother of patient #32 also died of ruptured aneurysm but we have not been able to assess parental transmission for this patient. Measurement of the aortic diameter was not systematically performed for all our patients. However, these observations call for implementing a cardiac evaluation of these patients and a close follow-up. A larger study would be of great interest.
Regarding male to female ratio, a previous study showed a male-biased effect of 16p13.11 CNVs.8 In our population, the ratio was 0.96, whereas the microarray assayed population shows a significant bias to male (1.48) that is well known in DD/ID and ASD populations. Our results demonstrate equal chances of being affected with 16p13.11 duplication regardless of gender, which is in contradiction with the study cited above. More data are needed to figure out this discrepancy.
16p13.11 Duplications are associated with variable expression and incomplete penetrance. We show that this variant has a low penetrance (8.4%), which is in the same range than the 15q11.2 duplication involving the NIPA1 gene (10.4%), for example.6 We also show its enrichment in the DD/ID/ASD population (OR=1.60, p<0.05), demonstrating that it is likely pathogenic or at least contributing to the neurocognitive phenotype. This enrichment is probably underestimated since some subjects in control populations might display mildly impaired cognitive functions as shown for 15q11.2(BP1-BP2) deletion carriers.9 In fact, the frequency of the variant varies between different control populations.
Among the classical explanations for phenotype variability, we could exclude a correlation with the size of the duplicated segment nor with its parental origin in contradiction with what Ullmann et al have suggested2 . Furthermore, to date, there is no evidence for a parental imprinting regulation in the region. Girirajan et al proposed a double-hit model in which a single variant of unknown significance predisposes to a neuropsychiatric disorder, but when associated with other chromosomal deletions or duplications, it exacerbates neurodevelopmental disorders.10 We hence collected data on associated CNVs, for each patient. Almost 30% of our patients harboured at least one additional CNV >200 kb but none of them was found recurrently. Furthermore, because microarray analysis does not allow to investigate the presence of single nucleotide variant or polymorphism we could not fully exclude the hypothesis of a second hit, as already demonstrated in the TAR syndrome in which a minimally deleted 200 kb region at chromosome band 1q21.1 associated with a mutation of RBM8A gene is responsible for the phenotype11. It is also worth noting that in two patients another chromosomal anomaly likely contributes to the phenotype. The first one is patient #31 who has a de novo 2q24.3 deletion encompassing SCN2A and the second one is patient #38 whose karyotype showed a sex chromosome anomaly (47,XXY).
The smallest duplications in our cohort (patients #2 and #23) support the involvement of NDE1 and miR-484 in the neurodevelopmental phenotype. Indeed, although these patients have the smallest duplications reported to date, they display ASD and DD/ID phenotype. The duplication of patient #23 involves the upstream region of these two genes which may contain cognate regulatory elements.
NDE1 (nudE nuclear distribution gene E homolog 1) is a candidate gene for the neurocognitive phenotype. It encodes for a protein that belongs to the nuclear distribution E family and locates at the centrosome. NDE1 plays an essential role in microtubule organisation, mitosis and neuronal migration.12 Houlihan and Feng suggested that NDE1 dosage alteration may result in secondary genomic lesions in cortical neurons underlying a large variety of developmental neurological disorders.13
Recently, a functional study revealed that the overexpression of miR-484, one of the numerous miRNA embedded in the 16p13.11 duplication region, induces a hyperactivity phenotype in mouse.14 The authors show the contribution of the Pcdh19, p rotocadherin 19, to the phenotype as a target of miR-484. Since miRNA usually have many targets, it is possible that other miR-484 targets could explain the other phenotypes associated with 16p13.11 duplication. Moreover, other miRNAs in the region might also be involved. The miRNA-mediated pathogenesis could also explain the low penetrance and phenotypic variability since the effect of the miRNA may be modulated by polymorphisms in its target regions.
Another gene lying in the common duplicated region for all patients but patient #23 is MYH11, myosin heavy chain 11, which was found mutated in familial aortic aneurysm.15–17 Kuang et al reported a significantly higher risk of aortic aneurysm and dissection in patients harbouring a 16p13.11 duplication with an OR exceeding 10.7 They showed that MYH11 was overexpressed in the aortic tissue and hypothesised, based on C. elegans data, that this could have a negative effect on the protein degradation. In our cohort, the two cases of aortic disease prompt for a larger study evaluating the level of this risk in 16p13.11 population to provide a better risk management and a complete information to physicians and families.
To conclude, 16p13.11 duplication is a recurrent CNV that is characterised by incomplete penetrance and variable expression. Our data in addition to the published one suggest that 16p13.11 duplications are likely pathogenic when detected in the context of DD/ID/ASD. This series supports an essential role of NDE1 and miR-484 in the neurocognitive phenotype. Studies analysing associations between this variant and nucleotide polymorphic variants, in particular within miR-484 targets, could be highly informative. We also recommend a close cardiac evaluation and follow-up especially of the aorta characteristics. Our study highlights substantial elements to take into consideration for an appropriate genetic counselling to the patients and their families.
The authors are grateful to the AChro-Puce network, the French national network of microarray users in the clinical setting (http://www.renapa.univ-montp1.fr/), for the publication of our collaboration call. They also warmly thank Dr Jacques Motte for his kind help in this project.
Contributors LAEK and EP: designed the study and collected the data. LAEK, SH and EP: analysed the data and wrote the manuscript. MDF, CV-D, AA, OB-B, MPC, LF, CF, MG, AG, AM-P, AM, MM-D, GL, MR, PE, AD-D, LDP, SCE, LS, CI, CF-F, PG and CL: performed the clinical evaluation of the patients. LAEK, SH, J-HC, JA, PC, SC-B, A-LM-B, NM, NLM, GP, ACT, AL, AN, CM, AR, CY, PV, DS, JMD, BB and EP: performed the genetic investigations. All authors revised and approved the final version.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Ethics approval French Ethical Board.
Provenance and peer review Not commissioned; externally peer reviewed.