CADASIL1 via the NOTCH3 Gene

CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) is a small-vessel disease that primarily affects the vascular smooth muscle cells in the brain. Symptoms usually begin between the second to fourth decades of life, first manifesting as migraine with aura, followed by subcortical ischemic strokes, executive dysfunction, motor disability, mood disturbance, and dementia. T2 white matter abnormalities can be detected on magnetic resonance imaging beginning from the first stages of disease, and the presence of granular osmiophilic material (GOM) detected in a skin biopsy is a pathognomonic sign of this disorder. Seizures, paralysis, sensory disturbance, parkinsonism, vertigo, and dizziness are other reported, but less frequent, features of the disorder (Hack et al. 1993. PubMed ID: 20301673). CADASIL is one of the most common cerebral small-vessel diseases, affecting roughly 2-5 people per 100,000 individuals (https://rarediseases.org/rare-diseases/cadasil/). Genotype-phenotype correlations exist for CADASIL, and knowledge of the specific genetic variant may help to predict the age of onset or severity of the disease. Some variant types predispose to milder or later-onset forms of the disease, with some cases not presenting initial symptoms until the 8^th decade. Additionally, even among family members carrying the same pathogenic variant, CADASIL can have highly variable severity and age of onset (Rutten et al. 2019. PubMed ID: 30032161). Advantages of testing for CADASIL include prognostic information, and the ability to make informed family planning decisions.

While CADASIL is by far the most common disease resulting from pathogenic NOTCH3 variants, three other extremely rare disorders have also been associated with this gene. All three are pediatric disorders. The first, lateral meningocele syndrome (Lehman syndrome) is caused by early termination variants in the last exon of NOTCH3. This disease is characterized by protrusions of arachnoid and dura through the vertebrae of the spine (lateral meningoceles), developmental delay, neurogenic bladder, paresthesia, paraparesis, and pain. This disease affects children beginning in infancy. Other common features include hypotonia, brain malformations, joint laxity, hernias, scoliosis, abnormal vertebrae, wormian bones, and reduced muscle bulk. Less common features include high nasal voice, hearing loss, cleft palate, congenital heart defects, feeding difficulties, dysphagia, and gastroesophageal reflux. A characteristic facial gestalt is observed that often includes hypertelorism, telecanthus, arched eyebrows, down-slanting palpebral fissures, ptosis, malar hypoplasia, long philtrum, thin upper lip, micrognathia, and a hypotonic appearance (Gripp et al. 2015. PubMed ID: 25394726; Ejaz et al. 1993. PubMed ID: 27336130).

Another NOTCH3-associated disorder affecting children is early-onset arteriopathy with cavitating leukoencephalopathy. This is an autosomal recessive disease with clinical features overlapping those of Sneddon syndrome and CADASIL but with greater severity and much earlier onset. Features include livedo reticularis, seizures, diffuse cavitations of the cerebral white matter, multiple lacunar infarcts, and disseminated microbleeds. Notably lacking was granular osmiophilic material (GOM) deposits in skin, as seen in CADASIL. Heterozygous parents of affected children are either asymptomatic or may have minor or subclinical symptoms of the disorder. Very few cases of this autosomal recessive disorder have been reported to date (Greisenegger et al. 2021. PubMed ID: 32980981; Pippucci et al. 2015. PubMed ID: 25870235).

The final NOTCH3-associated disorder is infantile myofibromatosis. This disease has an infantile onset and is characterized by multiple benign tumors of the perivascular myofibroblastic cells. These tumors  can occur throughout the body, but most often occur on the head, neck, or oral cavity. The severity of the disease depends on the location of the growing tumors. All NOTCH3-associated cases of infantile myofibromatosis reported to date are caused by a single missense alteration in the NOTCH3 gene (p.Leu1519Pro; Martignetti et al. 2013. PubMed ID: 23731542).

Pathogenic variants in NOTCH3 have been primarily associated with CADASIL1 but are also associated with autosomal dominant infantile myofibromatosis type 2, autosomal dominant lateral meningocele syndrome, and autosomal recessive early-onset arteriopathy with cavitating leukoencephalopathy (Greisenegger et al. 2021. PubMed ID: 32980981; Pippucci et al. 2015. PubMed ID: 25870235; Mašek and Andersson. 2017. PubMed ID: 28512196). 

CADASIL1 is an autosomal dominant disease typically caused by missense alterations in NOTCH3. It can be inherited or occur de novo (Joutel et al. 1996. PubMed ID: 8878478; Hack et al. 1993. PubMed ID: 20301673). Most CADASIL-causing variants in the NOTCH3 gene result in the gain or loss of one or more cysteine residues in the extracellular EGFr-like domains of the protein. A total of 34 EGFr-like repeats comprise the majority of the NOTCH3 receptor’s large extracellular region, each with six cysteine residues that are necessary to create disulfide bridges for correct protein folding. Pathogenic cysteine-altering variants in EGFr domains 1-6 appear to be fully penetrant and are usually associated with the classical CADASIL phenotype. However, there is variability in disease severity. Pathogenic cysteine-altering variants in EGFr domains 7-34 have a much higher frequency in the general population and can predispose to a milder small-vessel disease, possibly even displaying incomplete or at least very late-onset complete penetrance (Rutten et al. 2016. PubMed ID: 27844030; Rutten et al. 2019. PubMed ID: 30032161). The CADASIL-causing variant p.Arg544Cys has the highest population frequency of any causative variant, observed in the gnomAD database in 0.4% of alleles, which includes over 80 heterozygous individuals of unknown phenotype (gnomAD). This alteration is a founder variant in the East Asian population, and other common founder variants include p.Arg133Cys in the Finnish population, and p.Arg1006Cys in Italians (Hack et al. 1993. PubMed ID: 20301673). A few deletions, insertions, and splice site variants have been reported, most of which are also expected to result in an alteration of the cysteine residues within the extracellular EGFr-like repeats. A very small number of cysteine-sparing changes in NOTCH3 have been documented as causative for CADASIL (Muiño et al. 2017. PubMed ID: 28902129). Rarely, homozygous or compound heterozygous pathogenic NOTCH3 variants have been reported, and currently the data is inconclusive about whether these may lead to an earlier-onset or more severe form of CADASIL. Early termination changes in NOTCH3 are not an established mechanism of disease for CADASIL, and several have been documented in unaffected individuals (Rutten et al. 2013. PubMed ID: 24000151). However, unpublished evidence indicates that some (potentially domain-specific) early termination NOTCH3 variants may indeed lead to CADASIL, and additional studies are needed. To date, no large deletions or duplications have been reported as causative in this gene (Human Gene Mutation Database).

NOTCH3 encodes a transmembrane receptor protein that is a member of the NOTCH family. NOTCH signaling is involved in the regulation of various processes in both the embryo and the adult. In adult human tissues, NOTCH3 is primarily expressed in vascular smooth muscle cells and pericytes. In CADASIL1, abnormal NOTCH3 protein accumulates on the cell membrane of the vascular smooth muscle cells, resulting in the hallmark granular osmiophilic material (GOM), and a toxic gain of function effect from these accumulations is thought to underlie the mechanism of disease for CADASIL1 (Joutel et al. 2000. PubMed ID: 10712431). Mouse models of CADASIL exist and are currently being used to test novel therapeutic strategies for this disease (Liu et al. 2015. PubMed ID: 25251607; Hack et al. 1993. PubMed ID: 20301673). NOTCH3 has been cited as a nonessential gene for growth of human tissue culture cells (Online Gene Essentiality, ogee.medgenius.info).

Regarding the childhood-onset disorders associated with NOTCH3, lateral meningocele syndrome results from early termination variants in the last exon, within or upstream of the terminal PEST domain (Gripp et al. 2015. PubMed ID: 25394726; Mašek and Andersson. 2017. PubMed ID: 28512196). Pathogenic lateral meningocele syndrome variants are mostly de novo, but inherited cases have been reported. Two homozygous early termination variants have been documented as causative for the severe presentation of autosomal recessive early-onset arteriopathy with cavitating leukoencephalopathy (Greisenegger et al. 2021. PubMed ID: 32980981; Pippucci et al. 2015. PubMed ID: 25870235). Finally, a single missense alteration in NOTCH3 (p.Leu1519Pro) has been documented as causative for infantile myofibromatosis type 2 (Martignetti et al. 2013. PubMed ID: 23731542).

Pathogenic heterozygous and homozygous or compound heterozygous variants in the HTRA1 gene cause CADASIL2 and CARASIL (cerebral autosomal recessive arteriopathy with subcortical infarcts and leukoencephalopathy)—two disorders phenotypically similar to CADASIL1 (Verdura et al. 2015. PubMed ID: 26063658; Di Donato et al. 2017. PubMed ID: 28782182; Onodera et al. 1993. PubMed ID: 20437615). Pathogenic variants in the ADA2 gene have been reported as causative for Sneddon syndrome, which has phenotypic overlap with autosomal recessive NOTCH3 early-onset arteriopathy with cavitating leukoencephalopathy (Greisenegger et al. 2021. PubMed ID: 32980981).

This test is expected to identify a pathogenic NOTCH3 variant in over 90% of patients fulfilling the diagnostic criteria for CADASIL (Mizuta et al. 2017. PubMed ID: 28991717).

This test is performed using Next-Gen sequencing with additional Sanger sequencing as necessary.

This test provides full coverage of all coding exons of the NOTCH3 gene plus 10 bases of flanking noncoding DNA in all available transcripts along with other non-coding regions in which pathogenic variants have been identified at PreventionGenetics or reported elsewhere. We define full coverage as >20X NGS reads or Sanger sequencing. PGnome panels typically provide slightly increased coverage over the PGxome equivalent. PGnome sequencing panels have the added benefit of additional analysis and reporting of deep intronic regions (where applicable).

Dependent on the sequencing backbone selected for this testing, discounted reflex testing to any other similar backbone-based test is available (i.e., PGxome panel to whole PGxome; PGnome panel to whole PGnome).