Cornelia de Lange syndrome (CdLS) is a rare disease named after the Dutch paediatrician Cornelia de Lange, who in 1933 published the cases of two unrelated female infants who presented with very similar phenotypes.1 She called the syndrome typus degenerativus amstelodamensis, in honour of the city of Amsterdam where she worked. Before that, in 1916, Brachmann had published the case of a patient with growth retardation, hirsutism and bilateral monodactyly that has since been recognised as the first documented case of CdLS,2 which is why some publications refer to this disease as Brachmann-de Lange syndrome.

At present, the prevalence of CdLS is estimated at 1 case per 10 000 – 30 000 live birth births in the general population.3 This includes cases with atypical presentations of CdLS that were not diagnosed in previous decades when the genetic tests currently used were not yet available.

The differential diagnosis of CdLS can be complex due to the variability of its clinical manifestations and the overlap with other genetic neurodevelopmental disorders, such as Coffin-Siris syndrome, Rubinstein-Taybi syndrome, KBG syndrome, CHOPS syndrome or Wiedemann-Steiner syndrome.22–26 In 2018, a consensus group of international experts proposed clinical diagnostic criteria that paediatricians could apply easily when they encounter patients with suspected CdLS3 (Table 2). These criteria allow differentiation between classic and non-classic phenotypes and of CdLS and diseases with similar presentations. It is based on assigning a specific value to a series of cardinal features and other suggestive features and, based on the obtained score, classification as classic CdLS (score ≥ 11 points, with at least 3 cardinal features present) and non-classic CdLS (score of 9–10 points, with at least 2 cardinal features present). In addition, a score is between 4 and 8 points warrants performance of molecular testing for CdLS to make the diagnosis (Table 2).

In addition, in recent years artificial intelligence (AI) tools have been developed that facilitate the clinical diagnosis of rare diseases through the analysis of facial morphology. Latorre et al.27 tested the Face2Gene application (https://www.face2gene.com) using frontal facial photographs of individuals with CdLS. Their results showed that the software offered a high sensitivity in the facial recognition of CdLS, with a success rate of 83.7% defining success as CdLS being the disorder listed as the first prediction (Fig. 1).

In short, at the time of diagnosis (Fig. 2) it is key to take a thorough history with a standardized history-taking approach, collecting clinical information using Human Phenotype Ontology (HPO) codes (Table 1). In addition, AI tools for facial phenotyping may become increasingly useful (Fig. 1). However, the definitive clinical diagnosis, and the classification as classic or non-classic CdLS should be based on the criteria defined by Kline et al.3 (Table 2).

Figure 2.

Algorithm for the clinical diagnosis of CdLS.

(0.1MB).

Cornelia de Lange syndrome can be diagnosed by the identification of a pathogenic variant in one of the 5 causative genes associated with the cohesin complex (NIPBL, SMC1A, SMC3, RAD21, HDAC8) and, with less certainty, in 3 recently described additional genes (BRD4, ANKRD11, MAU2).28,29 The cohesin complex consists of a central ring formed by proteins SMC1A, SMC3 and RAD21 and a series of regulatory proteins such as NIPBL, the main protein involved in loading cohesins on DNA, and HDAC8, a SMC3 deacetylase that allows the release of the cohesin complex from chromatin3 (Fig. 3).

Figure 3.

Location of the proteins that cause Cornelia de Lange syndrome: NIPBL and BRD4 are responsible for nucleosome clearance in DNA so that proteins SMC1A, SMC3, RAD21 and STAG2 can form the cohesin ring. MAU2 forms a heterodimeric complex with NIPBL to help load cohesin on chromatin. Then, protein HDAC8 deacetylates SMC3, allowing the release of the cohesin ring. Protein ANKRD11 is involved in the transcription activation process.

(0.21MB).

The cohesin complex is a multifunctional complex involved in many biological processes, such as chromosome segregation, genomic stabilization and organization and, above all, regulation of gene expression. Most individuals with CdLS (more than 60%) have a pathogenic variant of NIPBL, a gene that encodes a protein that interacts with chromatin remodelers and transcription factors.30 This is the reason that CdLS is classified as a transcriptomopathy, that is, a disease caused by the global dysregulation of gene expression throughout the organism, which explains the multisystemic involvement and the phenotypic variability of this syndrome.28

Technological advances in high-throughput sequencing in recent years have allowed the identification of new causative genes, especially in patients with non-classic phenotypes. Among the salient ones are ANKRD11, which encodes a protein involved in transcriptional regulation,24 or MAU2 and BDR4, which encode proteins that interact with NIPBL29,31 (Fig. 3).

This substantial clinical and genetic heterogeneity, especially in individuals with non-classic/mild or atypical CdLS, can pose a barrier to molecular diagnosis. Therefore, the use of next-generation sequencing techniques allowing simultaneous evaluation of a large number of genes is recommended for the initial approach to genetic diagnosis (Fig. 4). In recent years, the general approach has been to use of panels that include genes known to cause CdLS and other genes associated with neurodevelopmental disorders, or clinical exome sequencing to simultaneously evaluate up to 7000 genes associated with disease. On the other hand, given the improvement of molecular technologies, the reduction in the costs of sequencing and the automation of data analysis, it is possible that in upcoming years whole-exome or even whole-genome sequencing will replace these techniques.

Figure 4.

Algorithm for the molecular diagnosis of CdLS.

(0.22MB).

If all these techniques yield negative results, performance of array comparative genomic hybridization (CGH) is recommended to assess for chromosomal anomalies or copy number variants.

The presence of somatic mosaicism in some individuals (10%–13%) may pose an additional difficulty and require analysis in specialised laboratories of other tissues, such as oral mucosa specimens or fibroblasts, and the use of highly sensitive molecular diagnostic techniques when results in blood samples turn out negative.32,33

If this still fails to yield a diagnosis, patient-parent trio whole exome sequencing is recommended. The parental specimens are then used to filter the variants identified in the index case (Fig. 4).

However, despite the significant progress made to date, there is still a considerable number of individuals in which genetic diagnosis is not achieved, which suggests the existence of causative genes and mechanisms of mutation that have yet to be identified.

Although most individuals with classic CdLS have a pathogenic variant in gene NIPBL associated with the characteristic facial morphology and, in some cases, reduction deformities in the extremities, there are also patients with non-classic phenotypes (Fig. 5). The pathogenic variants of NIPBL are uniformly distributed through the coding region, but missense variants are more frequent in the HEAT repeat domain and are associated with a milder presentation.32,34

Figure 5.

Cornelia de Lange spectrum. Cornelia de Lange syndrome can be considered an umbrella term comprehending a broad phenotypic spectrum. The diagram shows the clinical spectrum caused by changes in the NIPBL, SMC1A, SMC3, HDAC8, RAD21 and other genes by means of arrows.

(0.08MB).

Individual with pathogenic variants of the SMC1A or SMC3 genes exhibit fewer structural anomalies and less severe growth retardation compared to individuals with pathogenic variants in NIPBL; however, they have significant intellectual disability that is moderate to severe.35 Patients with pathogenic variants in the SMC1A present with slightly flatter and wider eyebrows and a long and wide nasal ridge, in some cases, they develop refractory epilepsy with an early onset.36 Patients with pathogenic variants in the SMC3 gene often present with subtle or no synophrys, and 56% have heart defects.37

Individuals with a heterozygous pathogenic variant of RAD21 usually have no major structural anomalies, and tend to present milder cognitive impairment and GER, with the latter limited to early childhood.38 They usually exhibit growth delay, minor skeletal deformities and facial features that overlap with those of CdLS.

Male patients with hemizygous pathogenic variants in the HDAC8 gene have the characteristic facial features of CdLS accompanied by delayed anterior fontanelle closure, hooding of the eyelids, orbital hypertelorism, pigmentary mosaicism, dental anomalies and, most saliently, a broad/bulbous nasal tip and a friendly personality. Growth delay is not as significant, with a lower frequency of postnatal growth retardation and microcephaly. Some of these patients may also exhibit severe reduction deformities in the extremities. In female patients, the presentation tends to be milder, but the severity is strongly associated with the X chromosome inactivation pattern.39

Most CdLS cases are sporadic, especially in cases involving genes with autosomal dominant inheritance: NIPBL, SMC3 and RAD21. In the case of pathogenic variants in X-linked genes (SMC1A and HDAC8), male patients tend to have more severe presentations compared to female patients. In these families, there are cases of parents with mild disease and the risk of transmission depends on the sex of the affected parent.3

More than 99% of patients with classic CdLS have a de novo heterozygous pathogenic variant in the NIPBL gene. This means that the risk of recurrence in future siblings is very low (1–2%) and chiefly due to the possibility of germline mosaicism in one of the parents.40 In families in which a pathogenic variant has been identified in any of the genes associated with CdLS, prenatal genetic screening could be performed to detect or rule out the presence of the variant in the foetus.

At present, there is no curative treatment for CdLS, but there are symptomatic and palliative treatments that the paediatrician must be aware of. The main medical complication in classic CdLS (associated with the NIPBL gene) is GER, which is present in most patients and must be treated early and aggressively conforming to the validated treatment protocols in the paediatric literature. In some instances, when the patient does not respond well to the standard of care, alternative measures need to be used, such as nasogastric or gastrostomy tube feeding in the first months or years of life. Some patients ultimately require surgical treatment (Nissen fundoplication). When it comes to the digestive system, it is important to keep in mind the possible presence of intestinal malrotation that could become complicated, manifesting with constant or recurrent acute abdominal pain (due to obstruction or volvulus), bringing the patient to the emergency department (usually within 2 years from birth) where, if the cause is not identified early, the outcome could be death.3

Hearing should be tested from birth (otoacoustic emission test); in the case of negative or uncertain results, auditory evoked potential testing should be performed so that, should sensorineural hearing loss be detected, hearing aid devices (audiphone or cochlear implant) can be placed as soon as possible, preferably before the age when speech starts to develop. The evaluation should also include an echocardiogram and an abdominal ultrasound examination to assess for the presence of structural anomalies in the heart and the organs in the abdominal cavity (kidneys).

As regards psychomotor development, most patients with CdLS require intervention to stimulate language development (speech-language therapy) and motor development (physical and/or occupational therapy). Some patients may need pharmacotherapy to manage hyperactivity and/or attention deficit.

In November 2023, the XII CdLS World Conference was held in Zaragoza by the Reference Centre for CdLS in collaboration with the Asociación Española del Síndrome de Cornelia de Lange (Spanish Association of CdLS) and the CdLS World Federation. The congress gathered the main experts worldwide, who shared the latest advances in the knowledge of the syndrome. In the clinical area, we ought to highlight the novel technique of sensory neuromodulation to treat refractory epilepsy in patients with CdLS and pathogenic variants of the SMC1A gene, which has proven effective in 3 out of 4 treated patients. The importance of contemplating the possibility of intestinal malrotation in any child with CdLS presenting to the emergency department with acute abdominal pain was also addressed. In respect of anaesthetic procedures in children with CdLS, presenters insisted that experience has demonstrated that they are well tolerated by patients. In the basic research section, evidence was presented confirming that the absence of cohesin gives rise to changes in DNA transcription, in addition to novel 3D models of cohesin that could help identify potential therapeutic targets in CdLS. Animal models are also being developed to start clinical trials of Ataluren for treatment of epilepsy in patients with CdLS-SMC1A. Results from single-cell sequencing studies were presented that evinced the presence of transcriptional dysregulation in the early stages of embryonal development in a mouse model. The first findings on the subject of cellular senescence in CdLS, obtained with pluripotent stem cells derived from patients with CdLS, were also presented.

Strategic Health Action programme (Acción Estratégica en Salud [AES], Instituto de Salud Carlos III, Ministry of Science, Innovation and Universities of Spain): Project PI23/01370; Government of Aragón (Diputación General): Research Group B32-20R.

The authors have no conflicts of interest to declare.