Use of expanded carrier screening for retrospective diagnosis of two deceased siblings with Van Maldergem syndrome 2: case report

They were two male siblings with a similar set of malformations and their asymptomatic parents were first cousins. Figure 1 shows the pedigree of patients. Regarding evidence indicative to a possible genetic defect, six months after the death of the second child, parents were referred for genetic counseling. Written informed consent form for genetic analysis and publication of clinical and genetic data was obtained from the parents.

Figure 1.

All prenatal screening laboratory tests were normal for both patients and they were born via cesarean section.

Due to tracheostomy of both of them in the early weeks of life, it is not possible to rule out/in feeding difficulties. The only abnormal finding in laboratory tests of proband 1 was a high level of C-reactive protein (CRP) without evidence of infectious disease. They had respiratory distress and were under ventilator for weeks. Both the patients deceased in the second month of life due to hypoxemia and cardiac arrest.

Carrier screening of parents was carried out using WES. In brief, Twist Human Core Exome Plus Kit (Twist biosciences, San Francisco, CA, USA) was used for exome enrichment and the library was sequenced on Illumina platform (Illumina Inc., San Diego, CA, USA). Mean on-target coverages were 131X and 167X for the mother and father, respectively.

In silico pathogenicity assessments: Allele frequency of the variant was investigated in population databases of 1000genome project (http://www.1000genomes.org), gnomAD (https://gnomad.broadinstitute.org/), ExAC (data available in gnomAD browser), and our local database Iranome (http://www.iranome.ir/). Its damaging effect on gene product was assessed using online prediction tools MutationTaster (http://www.mutationtaster.org/) and Combined Annotation Dependent Depletion (CADD) (https://cadd.gs.washington.edu/) [12, 13]. Web-based splice site prediction tools NNSPLICE v0.9 (https://www.fruitfly.org/seq_tools/splice.html), SD-score (https://www.med.nagoya-u.ac.jp/neurogenetics/SD_Score/sd_score.html) and NetGen2 v2.4 (http://www.cbs.dtu.dk/services/NetGene2/) were applied for further analysis of spliceogenic effects of the variant [14, 15, 16, 17].

This study was in conformity with the Helsinki Declaration and ethical approval was not required for this study in accordance with the local/national guidelines. Written informed consent was obtained from the parent of patients for publication of the details of their medical case.

WES data analysis revealed a heterozygous intronic variant, NM_024582.6: c.7018+1G>A, in the gene FAT4 which was confirmed in parents by sanger sequencing. This variant substitutes the first nucleotide, G, of intron 6 for A, altering canonical splice site (CSS).

The variant was absent from population databases of 1000genome, gnomAD, ExAC, our local database Iranome and, as far as we know, has not been reported before. MutationTaster predicted it as the cause of the disease. In addition, the CADD Phred-score (GRCh37-v1.6 adapted especially for splicing variants) is 33 for this variant, which puts it among the top 0.1% most deleterious substitutions that happen throughout the human genome. We applied splice site prediction tools to analyze the mutant sequence and detect possible cryptic donor sites. All analyses indicated that the CSS is missed and suggested new donor sites; Table 1 shows in silico scores of the normal and mutant CSS.

Based on criteria from the ACMG guideline, this variant is classified as pathogenic [18, 19]. Applied rules include PVS1 for being a null variant in a gene for which loss of function is the mechanism of disease and PM2 for the variant is not found in the population databases.

In addition to FAT4, the parents shared heterozygous variants of other genes, which are present in Table 2.

We report two siblings with MCA, born to consanguineous parents, who deceased at the age of two months. During genetic counseling, their parents were explained about a possible familial recessive disorder and informed that in the absence of an accurate clinical diagnosis, a genotype-first diagnosis may be achieved. Therefore, WES was suggested and, due to no samples from their deceased children, it was applied for carrier screening of the parents. Also, they were informed about the advantages and disadvantages of this method. The advantages of WES in ECS, even in cases with no post-mortem samples, have been approved through previous studies. For example, in a recently published report, by this method they could identify the causal variant of a metabolic disorder in three dead siblings. However, as far as we know, in all of them there have been a definite or probable clinical diagnosis. In the current study, no preliminary clinical diagnosis or assumption were given to deceased patients, and from this point of view it could be assumed as the first report of its kind [20, 21, 22].

This approach led to identification of a number of variants shared with the parents, including a novel pathogenic heterozygous variant in the gene FAT4, NM_024582.6: c.7018+1G>A. (Please see Table 2 for the other variants). This variant is pathogenic by applied rules PVS1 and PM2. Regarding criteria PVS1, loss of function is an established pathogenicity underlying the diseases associated with FAT4 variants: (1) According to data from Simple ClinVar database (https://simple-clinvar.broadinstitute.org/), most of the pathogenic or likely pathogenic variants in the gene FAT4 are of types which have disrupting consequence on protein function, including stopgain, frameshift, and splicing variants; (2) the OMIM database indicates several animal studies which support the idea of loss of function in FAT4-related disorders [23].

Reverse phenotyping of the patients showed overlapping features of syndromes caused by FAT4 variants, HKLLS, and VMLDS, but the absence of lymphangiectasia or lymphedema, which are hallmarks of HKLLS, and the presence of tracheomalacia in both the siblings with small cysts in the kidney of proband 2 can be suggestive of VMLDS2 [1]. Table 3 shows a comparison of the clinical symptoms between the patients of this study with characteristics previously described in the literature. As mentioned before, clinical variable expressivity is common across patients with FAT4 variants, which makes the classification of patients in Van Maldergem or Hennekam syndrome challenging. An interesting case is a patient, F4–1, from Alders et al.'s [6] study, who had a splicing variant, NM_001291303.1: c.7200–2A>C, disrupting the acceptor site of intron 8. They showed by functional study that it leads to deletion of only two amino acids. Comparing the clinical symptoms of this patient with our patients shows significant clinical differences. She, F4–1, has been consistent with Hennekam syndrome, due to having the lymphatic anomalies with no tracheal or kidney abnormalities. Of course, overlapping symptoms are also seen in these three patients, such as dysmorphic faces and skeletal limb anomalies, evident in the form of camptodactyly in our patients, but in the form of syndactyly in the previous patient. She, F4–1, had also ear anomalies but without hearing problems. It is noteworthy that this pediatric patient, F4–1, has been alive at the time of report.

The Fat cadherin subfamily are highly conserved single-pass transmembrane receptors, which in humans consist of four members FAT1–4. All FAT cadherins have a large extra-cellular domain mainly consisting of cadherin repeats, which are followed by a number of EGF-like and Laminin G-like motifs [25]. Heterophilic interaction of FAT4 with DCHS1, another cadherin, is critical in planar cell polarity (PCP) and Hippo signaling pathways involved in cellular proliferation, orientation, migration, and tissue development [26].

FAT atypical cadherin 4, FAT4, is located on chromosome 4q28.1 and its transcript, NM_024582.6, contains 18 exons which is translated to a protein with 4981 amino acids [27]. The novel pathogenic splicing variant, c.7018+1 G>A, disturbs 5′ CSS at donor sequence motif of intron 6 (Figure 2). In silico analysis of mutant sequence using splice prediction tools showed that at least two cryptic splice sites inside intron 6 can be potentially activated, of which one of them is very strong (Table 1). Although clinical evidence supports our idea of abnormal FAT4 transcription with no functional product, unfortunately DNA or tissue samples from patients are not available, and due to remote location of the parents it is not possible to conduct further RNA studies.

Figure 2.

Diagnosis of rare diseases with variable expressivity and overlapping features just based on clinical evaluations is difficult, especially in infantile patients. NGS, as a powerful facility, can unravel genetic defects underlying these group of disorders. Moreover, reverse phenotyping after a genetic finding may extend the clinical assessments and anticipate diagnosis of late onset features [28]. Besides these advantages, NGS has ECS test to hundreds of conditions that makes it increasingly favorable for preconception, preimplantation, or prenatal diagnostic procedures (PND or PGD) [29].

However, next to these advantages, there are routine limitations associated with NGS methods, including the lack of full coverage for all genes or exons, absence of deep intronic regions in WES, and inability to investigate CNVs or chromosomal rearrangements. Moreover, finding variants with uncertain significance (VUS) can bring even more complexity to the situation. Furthermore, there are additional challenges for cases similar to ours: first, the heterozygous variant is identified in healthy adults and cannot be confirmed in a homozygous patient. Second, genotype–phenotype correlation is not straightforward and there is always concern around other possibilities which may be missed. This makes genetic counseling complicated, and any clinical intervention based on these findings must be advised more cautiously [22, 30]

ECS of parents is a powerful test which can reveal the genetic defect in deceased patients suspected to have a rare disease with recessive inheritance pattern. Here, exploiting this method revealed a novel pathogenic splice variant in FAT4 gene and resulted in diagnosis of VMLDS in two dead infantile siblings.