Many structural anomalies have been described in association with OC, including the MAC spectrum [17]. The presence of CP, CL with or without CP (CL+/-P) seems to represent a risk of concomitant microphthalmia and/or anophthalmia [18]. Few case series studies have focused on the co-occurrence of MAC spectrum and other birth defects. Studying a cohort of 415 live births with microphthalmia/anophthalmia/coloboma, Roos et al. [19] reported that 13% presented with coloboma and OC, and 25% with microphthalmia or anophthalmia and OC.
Our study presents the investigation of genetic variants using CMA and WES, in individuals with OC-MAC in which clinical assessment was performed by experienced dysmorphologists, following the same protocol of data collection and record [9].
OC and MAC are phenotypically heterogeneous. Throughout the years, over 300 genes have been implicated in the etiology of orofacial clefts, including the IRF6 gene, which accounts for 2% of cases of SOCs, represented by Van der Woude syndrome (OMIM #119300) and is also responsible for many cases of non-syndromic clefts (NSOCs) [20]. The genes GRHL3, TBX1, TP63, and LRP6 are also important genes related to SOCs that are also described in NSOCs [21, 22]. In addition, more than 90 genes related to MAC spectrum have been identified so far, most of them transcription factor genes (SOX2, OTX2, VSX2 and PAX6), and the retinoic acid pathway genes (STRA6, RARβ, and ALDH1A3) [23]. However, genomic imbalances and sequence variants in these genes were not detected in our cohort.
Since chromosomal imbalances are seen in about 15–20% of patients with intellectual disability or multiple congenital defects [24] and are found in up to 44.4% of individuals with microphthalmia and/or anophthalmia [19], searching for genomic imbalances was the first strategy chosen for diagnosis. However, despite previous studies have demonstrated the importance of CMA in the investigation of syndromic MAC spectrum [4] and SOCs [5], none of the patients in this OC-MAC cohort presented pathogenic genomic imbalances, hence why the study was pursued with WES analysis.
Genotype-phenotype correlation
In the present study, causative variants in the CHD7 gene were detected in 2/17 cases. Patients 1 and 2 fulfilled the diagnostic criteria for the CHD7 phenotypic spectrum expansion [25]. Variants in the CHD7 gene were also detected in two individuals from a cohort of 67 trios with MAC investigated by WES [8] and this seems to be an important gene in the OC-MAC etiology.
After genotype-phenotype correlation, etiological heterogeneity among individuals of this cohort included a causative variant in exon 13 of the PTPN11 gene in patient 3, that is responsible for Noonan Syndrome with Multiple Lentigines (OMIM #163950) [26], and a causative variant in TFAP2A gene, leading to Branchiooculofacial syndrome (BOFS - OMIM #113620) (Patient 5). Patients 4 and 6 have causative variants involving genes TP63 and POMT1, respectively, with some particularities that are mentioned below.
Patient 4 met the criteria for ADULT syndrome (OMIM #103285), which is one of syndromes included in the wide clinical spectrum of TP63 [27]. The association of TP63 variants and microphthalmia or cataract, as presented in our patient, has not been described before. However, reports of two patients with glaucoma associated to anterior segment dysgenesis and TP63 variants suggest that this gene might participate in the anterior eye segment embryogenesis [28]. Therefore, it seems that the eye anomalies detected in our patient could be explained by the TP63:c.1813C > T:p.(Arg605Trp) variant and should be considered a phenotype expansion of the TP63 gene. Further functional studies and clinical examples would need to be performed and identified to verify this claim.
Despite the VUS classification following the ACMG guidelines, some individuals who had VUSes presented with a phenotype strongly indicative of the disorders associated to the mutated gene. Further functional studies could also help to identify if these variants contribute to the diagnosis.
Patient 6 presented a splice site variant in the POMT1 gene (c.987-3C > G), classified as VUS. The phenotype included microcephaly, encephalocele, microphthalmia, anophthalmia, cataract, CLP and neurodevelopmental delay. These features are considered part of the clinical phenotypic spectrum of POMT1 gene and would allow a diagnostic conclusion [29]. Therefore, this variant was considered causative, as suggested by Johnson et al. [16].
In addition, there are some variants that could not be considered the exclusive cause of the respective patient’s phenotypes – including some heterozygous variants detected in genes with an autosomal recessive mode of inheritance. However, they seem to contribute to a specific malformation observed: microphthalmia in patient 7, anophthalmia, OC and digital anomalies in patient 8, and congenital cardiac malformations and embryonic anterior axis patterning in patient 9.
Patient 7 presented a heterozygous variant in the PORCN gene that could be related with his MAC phenotype, based on a report of a family with male individuals with variants in PORCN and presenting with isolated microphthalmia [30].
In patient 8, the BMP4 variant could be related to at least part of her phenotype, including low-set ears, micrognathia, retrognathia, anophtalmia, CLP, clinodactyly and syndactyly (OMIM #607932).
In Patient 9 we found variants in GDF1 and GRHL3 genes. Variants in GDF1 cause autosomal dominant multiple cardiac congenital defects [31]. Hence, in this patient, this variant could be related with the cardiac congenital malformations observed. The GRHL3 gene was described as a cause of Van der Woude syndrome (OMIM #606713) and NSOCs, including in the Brazilian population [32]. Clinical follow-up, additional case reports and functional studies would help for a better understanding of these variant’s contributions to the phenotype.
VUSes in other genes associated with oral clefts are present in patients 11 (SOX1) and 12 (FLNB and ARHGAP29) [33]. Other VUSes, which possibly have an additive effect in a polygenic mode of inheritance, will be discussed next (patients 10, 11 and 12).
The interpretation and reclassification of VUSes have been increasingly discussed in literature. In general, the guidelines of the ACMG and AMP consider the phenotype as supporting criteria for variant classification if it is highly specific for a unique syndrome or disease [15]. However, many rare variants detected in individuals with well-known syndromes and highly specific for them were classified as VUS according to the current classification criteria [34] and there has been a growing tendency to consider the phenotype as more important evidence for variant classification [16]. This is particularly true when considering craniofacial anomalies, which comprise a group of conditions with etiological and phenotypic heterogeneity.
Ciliopathy genes in OC-MAC
Some variants detected in patients 10, 11 and 12 are in genes that encode primary cilia structures (DYNC2H1, KIAA0586, WDR34, INTU, RPGRIP1L and KIF7) [35–40], and many of these genes have already been described as causative or contributing to phenotypes which include OC and MAC (KIAA0586, INTU, RPGRIP1L, KIF7 and LMNA) [36, 38, 39, 41–43]. However, possible mechanisms involved in these complex birth defects have not been pointed out.
Patient 10 presented a likely pathogenic heterozygous variant in DYNC2H1 gene, and VUSes in KIAA0586 and WDR34 genes. The DYNC2H1 gene encodes an intraflagellar transport (IFT) protein of cilia [35]. The KIAA0586 gene is the third most frequent mutated gene detected to cause Joubert syndrome with coloboma (OMIM #616490) [36]. The WDR34 gene is described as the cause of Short-Rib Thoracic Dysplasia 11, with or without polydactyly (OMIM #615633), an autosomal recessive skeletal ciliopathy.
The INTU gene coordinates ciliogenesis in vertebrates and was recently described as having a role in Orofaciodigital syndrome XVII (OMIM #617926) and Short-Rib Thoracic Dysplasia with polydactyly 20 (OMIM #617925) in a few patients [38]. There is a description of one individual, with compound heterozygous INTU variants, with microphthalmia and median CLP among other malformations [38]. This variant was found in patient 11, who also has VUSes in CHD7, LMNA, and SOX1.
In patients 10 and 11, compound heterozygosity was excluded based on analysis of other variants or CNVs in this group of genes.
In patient 12, who has CLP and microphthalmia, two variants in cis in RPGRIP1L were detected, both classified as VUS.
The RPGRIP1L gene encodes a protein that localizes to the central body and centrosomal structures of primary cilia and the inactivation of its ortholog in murine model leads to a phenotype like Meckel syndrome (MKS – OMIM#611561) and Cerebello-oculo-renal syndrome (Joubert syndrome type B – OMIM #608091) [39]. It is also implicated in COACH syndrome 3 (OMIM #619113), which can include coloboma among phenotypic findings [41]. Three fetuses from two unrelated families with MKS and biallelic variants in RPGRIP1L were described with median CP, microphthalmia, and other ocular malformations, among other major anomalies [39].
Patient 12 also has a likely pathogenic variant in KIF7 gene. This gene participates in SHH regulation, encodes a motor protein of primary cilia [40], and is involved in NSOCs [44]. In addition, CP was also described in individuals with variants in KIF7 [45].
Disruption of primary cilia structure or their function could be important causes of OC-MAC, at least, they represent part of the mechanisms that leads to them. The hypothesis that OC-MAC phenotype is the result of a disruption in ciliopathy genes has already been discussed in literature, and demonstrated in animal models [46] and in humans [47]. Clinical and molecular findings in the three patients herein described reinforce this possibility. However, future functional studies are still needed to confirm this hypothesis.
In addition, CL+/-P and microphthalmia are both considered midline defects, since they can be associated with a disruption in anterior-posterior or left-right axis patterning [48]. It is noteworthy that, in the present cohort, cardiovascular defects were present in eight individuals out of 17 (47.06%) and, in general, the heart is the most commonly affected organ during laterality defects [48].
Digenic/oligogenic mode of inheritance
The genes DYNC2H1, LMNA, KIF7 and INTU have already been reported in patients with conditions that present a digenic mode of inheritance [38, 42, 43, 49].
The DYNC2H1 gene is involved in a digenic mode of inheritance in a report of a patient with Short-Rib Polydactyly Syndrome (SRPS – OMIM #613091) with a heterozygous variant in NEK1 gene, resulting in a premature stop codon, and a missense variant affecting a splice site in DYNC2H1 gene [49].
Digenic inheritance is also one of the most accepted explanations for the diversity of phenotypes and the variable penetrance of the LMNA:c.1930C > T:p.(Arg644Cys) variant [42, 50]. The LMNA:c.1930C > T:p.(Arg644Cys) variant seen in patient 11, who presents with anophtalmia, atypical OC, and congenital heart disease is related to many phenotypes with diverse presentations and variable penetrance [42]. Although the clinical picture is not very similar, there is a report of an individual with atypical findings of laminopathies, including microphthalmia and cataract, who inherited this variant from the unaffected mother [42].
The KIF7 gene has been described in ciliopathies either as a modifier of GLI3 and NPH1 genes [40, 45] or as having a digenic inheritance pattern with KIAA0556 gene in an individual with coloboma and CLP among other major anomalies [43]. Putoux et al. [45] carried out functional studies with morpholino and showed evidence that hypomorphic KIF7 alleles interacts in trans with Bardet Biedl syndrome loci (OMIM #209900) [45], there-fore showing the potential for KIF7 to interact with other ciliary genes, exacerbating their phenotypes.
Concerning INTU, Toriyama et al. [38] described a patient with clinical findings suggestive of SRPS probably caused by digenic inheritance, with the participation of one heterozygous variant of this gene [38].
From these results and the information found in literature, including case reports and functional studies, it is possible to consider that the variants detected in individuals 10, 11 and 12 could contribute with the phenotypes observed by means of digenic or polygenic inheritance and, maybe, also as genetic modifiers of other genes.
Role of WES for diagnosis of OC-MAC
WES detected causative variants in six individuals (35.29%) with OC-MAC in this cohort. Currently many variants are still classified in the literature and databases as VUS due to shortfall of reports on such rare phenotypes. The underrepresentation of the Brazilian population in databases is another possible bias for the interpretation of these variants. Considering the aspects mentioned above, the genotype-phenotype correlation could be inferred in another six individuals (35.29%) based on gene function, minor allele frequency in population databases and protein deleterious effects demonstrated by in silico prediction studies (detailed description and in silico prediction are available in supplementary material).
Concluding Remarks
No genomic imbalances were observed, and WES detected causative variants in 35.29% (6/17) of the patients; two of them involving the CHD7 gene. These results suggest that WES should be complimentary to CMA, even though here, in these 17 individuals, clearly it was most effective approach in the molecular investigation of OC-MAC cases. In general, the spectrum of CHD7 should be considered in the clinical evaluation of OC-MAC.
In three patients with OC-MAC, results suggest the possibility that their phenotypes are related to defects of structure and (or) function of cilia, highlighting the importance of the pathways involved in cilia in the etiology of OC-MAC. However, further functional studies are needed to corroborate this hypothesis.
Many VUS detected in our cohort could be contributing to the phenotypes investigated and we hope that the results presented here can aggregate information on the etiology of SOCs, facilitating the reclassification of these variants in the future.