Background Cohen syndrome is a rare autosomal recessive inherited disorder that results from mutations of the VPS13B gene. Clinical features consist of a combination of mental retardation, facial dysmorphism, postnatal microcephaly, truncal obesity, slender extremities, joint hyperextensibility, myopia, progressive chorioretinal dystrophy, and intermittent neutropenia.
Patients and methods The aim of the study was to determine which of the above clinical features were the best indicators for the presence of VPS13B gene mutations in a series of 34 patients with suspected Cohen syndrome referred for molecular analysis of VPS13B.
Results 14 VPS13B gene mutations were identified in 12 patients, and no mutation was found in 22 patients. The presence of chorioretinal dystrophy (92% vs 32%, p=0.0023), intermittent neutropenia (92% vs 5%, p<0.001), and postnatal microcephaly (100% vs 48%, p=0.0045) was significantly higher in the group of patients with a VPS13B gene mutation compared to the group of patients without a mutation. All patients with VPS13B mutations had chorioretinal dystrophy and/or intermittent neutropenia. The Kolehmainen diagnostic criteria provided 100% sensibility and 77% specificity when applied to this series.
Conclusion From this study and a review of more than 160 genotyped cases from the literature, it is concluded that, given the large size of the gene, VPS13B screening is not indicated in the absence of chorioretinal dystrophy or neutropenia in patients aged over 5 years. The follow-up of young patients could be a satisfactory alternative unless there are some reproductive issues.
- Cohen syndrome
- VPS13B gene
- chorioretinal dystrophy
- diagnostics tests
- clinical genetics
Statistics from Altmetric.com
Cohen syndrome (CS) (OMIM 216550) is a rare autosomal recessive disorder first described in 1973, which involves a broad spectrum of clinical manifestations.1 Based on the observations of 29 Finish patients with CS, Kivitie-Kallio and Norio2 were the first to propose the essential features for CS diagnosis before the identification of the COH1 gene: (1) non-progressive mental retardation, motor clumsiness, and microcephaly; (2) typical facial features including wave shaped eyelids, short philtrum, thick hair, and low hairline; (3) childhood hypotonia and joint hyperextensibility; (4) retinochoroidal dystrophy and myopia by 5 years of age; (5) periods of isolated neutropenia. These criteria were modified by Chandler et al3 to be more applicable to young patients, when there is not yet evidence of chorioretinal dystrophy (CRD) or patients with a more heterogeneous genetic background. These authors proposed that CS could be diagnosed in the presence of at least two of the following major criteria in a child with significant learning difficulties: (1) facial gestalt, characterised by thick hair, eyebrows and eyelashes, wave shaped, downward slanting palpebral fissures, prominent, beak shaped nose, short, upturned philtrum with grimacing expression on smiling; (2) pigmentary retinopathy; (3) neutropenia (defined as <2000/mm3).
The VPS13B gene was subsequently identified on chromosome 8q22-q23. This gene is composed of 62 exons that span a genomic region of around 864 kb and encodes a putative transmembrane protein of 4022 amino acids with a complex domain structure (OMIM 607817).4 Although the exact function of VPS13B protein remains unknown, homology to the Saccharomyces cerevisiae VPS13 protein suggests a role in vesicle mediated sorting and intracellular protein trafficking.4 Since the first identification of a VPS13B gene mutation,4 more than 100 distinct VPS13B gene mutations have been identified.4–14 Following identification of the VPS13B gene, the Chandler criteria were modified by the same team, since new clinical features were noted.7 Patients were considered as having Cohen syndrome when six of the following eight criteria were fulfilled: developmental delay, microcephaly, typical facial dysmorphism, obesity and slender extremities, sociable behaviour, joint laxity, myopia/retinal degeneration, and intermittent neutropenia.7 In spite of considerable genotypic variability, positive patients fulfilled the diagnostic criteria with relative clinical homogeneity.
Given the increasing number of requests for molecular testing in patients with suspected CS and the very large size of the gene, the aim of our study was to determine which of the above clinical features are the best indicators for the presence of VPS13B gene mutations by comparing patients carrying VPS13B mutations with negative patients in a series of 34 patients referred for molecular testing of VPS13B in suspected CS. The results should help clinicians to evaluate whether VPS13B mutations are likely to be responsible for a clinical phenotype combining a number of features. We also evaluated the current diagnostic criteria.
Subjects and methods
A total of 34 patients from 29 families were ascertained for VPS13B testing and suspected CS at the molecular diagnostic laboratory of Dijon University Hospital in France. When a sample was received for VPS13B screening, a standardised comprehensive clinical form was sent to the referring physician. All blood samples received were screened for a VPS13B gene mutation, whether or not the patients presented the clinical criteria for the diagnosis of CS according to Chandler et al3 Written informed consent was obtained according to the French regulatory requirements for genetic testing.
Molecular genetic analysis
Each blood sample was processed for DNA extraction using the ‘salting-out’ method.15 PCR analysis and sequencing of exons and exon–intron boundaries of the VPS13B gene as well as amplification of gene sequences encoding exons 1 to 62 were performed. All PCR products were directly sequenced using a BigDye terminator kit and an ABI Genetic Analyser 3100 capillary sequencer according to the manufacturer's instructions (Applied Biosystems, Foster City, California, USA). Primer sequences are available on request. Corresponding reference sequences of the genomic DNA sequences of the VPS13B gene were downloaded using Ensembl Genome Browser (Accession number ENSG00000132549). The SeqScape software v2.5 package (Applied Biosystems) was used to visualise capillary trace electropherograms, for sequence assembly and alignment, and to search for comparisons with consensus and reference sequences and variations. Depending on the analysed exon, only between 20 and 30 bases inside the intervening sequences donor site and acceptor site from the consensus splice junction sequences are clearly investigated by the sequencing analysis program and subjected to alignment with the reference sequence. Mutation nomenclature16 numbering was based on the current Ensembl transcript (Ensembl Transcript ID ENST00000358544), with +1 as the A of the ATG initiation codon. Mutations leading to premature truncations were considered pathogenic. When a missense mutation was found, its absence was verified in 220 control chromosomes.
The proportion of clinical features of the CS spectrum in the group of patients with VPS13B gene mutations was compared with that in the group of patients with negative VPS13B sequencing, using Fisher's exact test. These analyses were possible because the age at molecular screening in the two groups was similar.
Patients with VPS13B mutations
VPS13B mutations were found in 12 patients (seven males and five females) originating from eight families (table 1), comprising one consanguineous Moroccan family and seven non-consanguineous French families. The age range at CS suspicion was 2.5 to 43.0 years and at genetic screening was 4.8 to 43.6 years. Except for two children aged 2.5 and 4.5 years, all patients were aged over 5 years at the time of diagnosis. The percentages of clinical features of the VPS13B spectrum are reported in table 1. All patients had mental retardation, typical or evocative CS facial gestalt, microcephaly and slender extremities with narrow hands/feet. Truncal obesity was reported in all but one patient aged 4.5 years. CRD was reported in 11/12 patients (92%), diagnosed on fundus examination in five patients and on fundus and electroretinography (ERG) in six. CRD was absent on fundus examination in a 19-year-old female, but ERG was not performed. Early signs of CRD were observed at ERG as early as 16 and 24 months in two young siblings. Myopia was present in the nine patients for whom the information was available. Neutropenia was reported in 11/12 patients. The clinical diagnosis of CS was fulfilled in all patients according to Kolehmainen et al7 criteria.
Patients without a VPS13B mutation
The 22 patients (seven males, 15 females) in whom no VPS13B gene mutation was found originated from 21 families, of which one was consanguineous. The mean age at molecular screening was 15.8±7.7 years. The percentages of clinical features of the CS spectrum are reported in table 1. In particular, CRD was observed in 6/19 ascertained patients. One patient had neutropenia. Kolehmainen's criteria for CS were fulfilled in 5/22.
VPS13B gene mutations
Table 2 summarises the mutations found in this series as well as the predicted consequence of each mutation at the protein level. A total of 14 different mutations, of which 11 have never been published, were identified in the eight families. All mutations except three resulted in premature truncation (table 2). One patient had, on the same allele on exon 56, a missense mutation c.10880C>T followed by an 18-base-pair (bp) deletion (c.10883_10900delCGAGGCAGCTTGTGCACG) leading to the deletion of six amino acids. A homozygous missense mutation (c.4907T>A, p.I1636N) was identified in the two Moroccan siblings born to consanguineous parents. The pathogenic nature of the mutation was suspected for several reasons: (1) isoleucine at position 1636 is a non-polar amino acid (whereas the mutant asparagine residue is polar); (2) the hydrophobic nature of the amino acid side chain at this position was 100% preserved following the alignment of proteins from different animal species (supplementary figure 1); (3) the cosegregation of the mutation in the family (supplementary figure 2); (4) the absence of this variant in 100 healthy controls. Only one truncating heterozygous VPS13B mutation was found in two patients, but these patients were considered as having CS in the presence of the typical clinical picture of the condition.
Significant differences were found between the group of patients with VPS13B mutations and the group of patients without VPS13B mutations. Indeed, the probability of finding CRD, neutropenia, microcephaly and myopia was higher in the group of patients with VPS13B mutations (table 1). Conversely, there was no significant difference for the presence of facial gestalt, narrow extremities and truncal obesity. In this series, the sensitivity of CDR and neutropenia was 92% for both parameters while specificity was 68% and 95%, respectively.
The hallmarks of CS include mental retardation, facial dysmorphism, chorioretinal dystrophy and neutropenia, and patients exhibit high clinical homogeneity. Following identification of the VPS13B gene, it was discovered that patients with VPS13B mutations, who did not fulfil CS clinical criteria, were exceptional.4–12 In contrast to other examples in the literature,17 the identification of the VPS13B gene has not made it possible to enlarge the clinical spectrum of CS. On the contrary, patients with suspected CS based on the presence of evocative facial gestalt but in the absence of ophthalmologic manifestations were reclassified as having Cohen-like syndrome, since no VPS13B mutations were found on either allele.8 Only Kolehmainen et al7 has given a brief description of patients negative for VPS13B mutations. The aim of this study was to compare the clinical features of patients with VPS13B mutations with those in patients without VPS13B mutations in order to give clues to the clinician on the indication for VPS13B screening according to the clinical phenotype, especially since molecular screening of VPS13B is a time consuming task. We showed that all patients with VPS13B mutations had either CRD or neutropenia. The presence of microcephaly, found in all patients with VPS13B mutations in this series, was another clinical feature that can help to distinguish between CS and differential diagnoses. Conversely, there was no significant difference for the presence of evocative dysmorphism, obesity and slender extremities. Therefore, except in young children, these features, in the absence of neutropenia or CRD, are not sufficient to suspect CS. Of note, the assessment of facial gestalt is subjective whereas CRD and neutropenia may be assessed objectively. According to the London Dysmorphology DataBase, around 80 syndromes associate mental retardation and retinopathy,18 19 while only a few syndromes associate neutropenia and mental retardation. This explains why the probability of finding a VPS13B mutation in patients with neutropenia (specificity of 95% in our series) in a context of suspected CS is even higher than it is in patients with associated CRD (specificity of 68% in our series).
We also evaluated the clinical criteria used to diagnose CS in our series. All patients with COH1 mutations fulfilled the Kolehmainen criteria, and among patients with no COH1 mutations, 5/22 fulfilled the Kolehmainen criteria. These criteria, therefore, provided 100% sensitivity and 77% specificity. These results are totally consistent with those reported by Kolehmainen et al.7 All patients with VPS13B mutations fulfilled their clinical diagnostic criteria. Out of 24 patients with no VPS13B mutation and full clinical data, only two fulfilled their clinical criteria. Among the 22 Cohen-like patients, none had neutropenia, and only one out of the seven patients with myopia and/or retinal dystrophy had the typical facial appearance of CS—although the patient did not sufficiently fulfil the other criteria for a true diagnosis of CS.
Table 3 shows the prevalence of neutropenia in previous series reporting VPS13B mutations. This prevalence was high except in the series reported by Hennies et al.6 He found only 10 patients with neutropenia in a series of 18 patients in spite of repeated haematological examinations. These results could not be assigned to any specific ethnic group or genotype–phenotype correlation. Table 3 also shows the prevalence of CRD in previous series. When considering all of the 160 patients with VPS13B mutations reported in the literature and assessed for CRD and neutropenia, CRD was found in 92% of patients aged over 5 years (86% if the age of patients was not taken into account). The difficulties in diagnosing CRD in young patients (<5 years of age) have been discussed by others.2 When CS is suspected, CRD may be diagnosed at the early stage of the disease using an electroretinogram (ERG) instead of, or in addition to, a fundus examination. Indeed, the amplitude of ERG waves may be reduced as soon as the photoreceptor cells are impaired, while in examinations of the fundus, bull's eye maculopathy or optic disc pallor with narrow vessels may be the only symptoms. Typical pigmentary lesions are observed later and only after cell death. This was clearly demonstrated by Chandler et al: the early onset of CRD was indicated using ERG in 80% of children under 5 years with suspected CS.3 20 The absence of CRD in a 19-year-old patient and its presence in two siblings aged 16 and 24 months in our series show the large variability in age at diagnosis of CRD in CS. These are other examples showing that this feature cannot be mandatory for the diagnosis of CS. Only 6% of patients had neither neutropenia nor CRD; this fell to 4% if children aged ≤5 years were excluded.
We cannot exclude the possibility of undetected VPS13B mutations and in particular genomic deletions of the VPS13B gene as recently described.13 Although this hypothesis is very likely in our two families in which only one pathogenic mutation was identified, it is unlikely when no pathogenic mutations are identified after genomic sequencing of the entire gene in non-consanguineous families.
In conclusion, our study showed that CRD and neutropenia appear to be the best predictors for the presence of VPS13B mutations. Given the large size of the gene, and since the probability of finding a VPS13B mutation in the absence of these key features is very low, screening for VPS13B mutations is not recommended unless they are present. Because of the age dependent onset of CRD, caution should be exercised in young children, in whom follow-up could be a reasonable alternative. However, VPS13B screening could be offered in such cases when the reproductive context is indicative since there is a 25% risk of recurrence.
The authors thank the patients and their families for their participation in the study. The authors also thank Philip Bastable from the ‘pôle de recherche’ of Dijon Hospital for his English review of the manuscript.
Competing interests None.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
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.