Cerebral Cavernous Malformations

A number sign (#) is used with this entry because of evidence that one form of cerebral cavernous malformations (CCM1) is caused by heterozygous mutation in the KRIT1 gene (604214) on chromosome 7q21.

Description

Cerebral cavernous angiomas are relatively rare vascular malformations that may involve any part of the central nervous system. Cerebral cavernous angiomas are to be distinguished from cerebral arteriovenous malformations (106070, 108010). CCMs are venous and not demonstrable by arteriography; hence they are referred to as angiographically silent.

Capillary hemangiomas (602089) are classified as distinct from vascular malformations in that hemangiomas are benign, highly proliferative lesions involving aberrant localized growth of capillary endothelium. Hemangiomas develop shortly after birth. In contrast, vascular malformations are present from birth, tend to grow with the individual, do not regress, and show normal rates of endothelial cell turnover (Mulliken and Young, 1988).

Genetic Heterogeneity of CCM

CCM2 (603284) is caused by mutation in the CCM2/malcavernin gene (607929), and CCM3 (603285) is caused by mutation in the PDCD10 gene (609118).

Evidence suggests that a 2-hit mechanism involving biallelic germline and somatic mutations is responsible for CCM1 pathogenesis; see PATHOGENESIS and MOLECULAR GENETICS sections.

Clinical Features

Some CCMs are clinically silent, whereas others cause seizures, hemorrhage, or focal neurologic deficit. Identification of these lesions is important because surgical removal of many is relatively easy. Magnetic resonance imaging (MRI) is replacing computerized axial tomography as the diagnostic modality of choice. Bicknell et al. (1978) found 3 reports of familial incidence of CCM and added 2 from their own experience. In 1 family a woman, 2 of her sons, and 1 of her son's sons were affected; in the second family a woman and her daughter were affected. Successive generations were affected in families reported by Michael and Levin (1936), Kidd and Cumings (1947), and Clark (1970). Michael and Levin (1936) described a Swedish family in which a mother, her 2 brothers, and 3 daughters had multiple 'telangiectases' of the brain, which were likely cavernous angiomata. Convulsions and migraine attacks were observed. Autopsy in one case demonstrated calcification in the vascular lesions of the brain. Clark (1970) described cavernous angioma of the brain in a man who died in 1945 at age 27 and in his daughter who died in 1969 at age 28.

Hayman et al. (1982) examined 43 relatives in 1 kindred by cranial computed tomography (CCT) and found 15 affected with cerebral vascular angiomas. Angiography failed to detect lesions in 5 patients who had positive CCT. Expression was variable and in 2 individuals, each the parent of an affected offspring, the CCT was normal. Familial cavernous angioma should be included in the differential diagnosis of any young person with cerebrovascular impairment, seizures, intracranial calcifications, or hemorrhage. Gorlin (1985) reported an extensively affected 3-generation family.

Michels et al. (1985) stated that 19 families with 77 persons with cavernous angiomas of the central nervous system and retina have been described. They described a 3-generation family ascertained through an 8-year-old boy with seizures and 2 unexplained lesions on CT and MRI. His mother presented a year later with a seizure and similar brain lesions. Angiography and eye examination were normal. The asymptomatic grandfather had 5 intracranial lesions on MRI scan.

Mason et al. (1988) described cavernous angiomas in 10 of 22 members of a large Hispanic family. The authors commented that 2 families previously reported by them (Bicknell et al., 1978), the family reported by Hayman et al. (1982), and 5 of the 6 families reported in abstract by Rigamonti et al. (1987) were Hispanic as well.

Dobyns et al. (1987) described a family in which 4 persons from 3 generations had multiple cavernous malformations ('angiomas') of the CNS and/or retina. They found reports of 16 other families containing a total of 50 cases. Excluding the probands, 68% of the patients were symptomatic. Cutaneous vascular lesions were an inconsistent manifestation. They recommended that any patient with a vascular malformation, especially a cavernous one of the brain, spinal cord, or retina, be evaluated for the possibility of this syndrome, which they referred to as 'familial cavernous malformations of CNS and retina' (FCMCR). The authors also suggested that all first-degree relatives should undergo a full evaluation if multiple vascular malformations are detected in the index patient or if the family history is suggestive because of seizures, cutaneous vascular lesions, recognized intracranial hemorrhage, or sudden unexplained death. Presymptomatic diagnosis in affected relatives would permit genetic counseling and close monitoring to allow prompt treatment if symptoms occur. Dobyns et al. (1987) concluded that there is a second group of patients with multiple cutaneous lesions and inconsistent CNS lesions referred to as hereditary neurocutaneous angioma (106070). The vascular lesions in this group were always arteriovenous malformations and were often located in the spinal cord.

Rigamonti et al. (1988) reviewed familial occurrence, presenting signs and symptoms, and radiographic features of the disorder in 24 patients with histologically verified cerebral cavernous malformations. Eleven patients had no evidence of a heritable trait and had negative family histories. The other 13 patients were members of 6 unrelated Mexican-American families. Among 64 first-degree and second-degree relatives, 11% had seizures. MRI was performed in 16 relatives (5 of whom were asymptomatic); 14 studies showed cavernous malformations and 11 studies identified multiple lesions. MRI was far more accurate in detecting these lesions than computerized tomography or angiography. Rigamonti et al. (1988) concluded that a familial form of this disorder is particularly frequent among Mexican-Americans. Bicknell (1989) described cavernous angioma of the brainstem in a 23-year-old Hispanic woman whose mother had died of brain hemorrhage. After moving to Baltimore from the southwestern part of the United States, Rigamonti (1993) concluded that there is not an unusual frequency of the disorder among Mexican-Americans. He emphasized that cavernous angiomas are not arteriovenous malformations; they represent a honeycomb of veins. They are not demonstrated by arteriography and therefore have been referred to as angiographically silent. Epilepsy is the most frequent symptom; bleeds occur in some cases.

Steichen-Gersdorf et al. (1992) reported a family in which cavernous angiomas of the brain were documented in 6 individuals in 5 sibships of 4 generations of a family. Two brothers in the third generation were asymptomatic but showed changes on MRI. Filling-Katz et al. (1989, 1992) described a family with cavernous angiomatosis in which 2 members had terminal transverse defects at the midforearm. Multiple family members had had episodic bleeding from cavernous angiomas of the central nervous system. Two had had retinal cavernous angiomas, 1 hepatic angioma, and 2 cavernous angiomas of soft tissue; skin angiomas were frequent. Studies of the forearm in 1 of the affected individuals showed abrupt termination distal to the normal radius and ulnar heads and apparently normal blood vessels. Filling-Katz et al. (1989, 1992) suggested that acute vascular disruption is the cause and that this is related to the fundamental defect in familial cavernous angiomatosis. Corboy and Galetta (1989) described a family in which the proband had suffered for 9 years from recurrent 'acute chiasmal syndrome,' diagnosed at first as retrobulbar neuritis.

Dellemijn and Vanneste (1993) investigated 20 relatives of a 23-year-old woman with cavernous angiomatosis of the central nervous system. Studies revealed 4 additional patients with symptomatic cavernous angioma and 1 with asymptomatic cavernous angioma. The basis of the neurologic symptoms had not previously been identified in the symptomatic patients. The pedigree pattern was consistent with autosomal dominant inheritance.

Computed tomography and MRI led to reassessment of the incidence of cavernous angioma of the brain including its familial occurrence. Drigo et al. (1994) described an Italian family with multiple cavernous angiomas of the brain, sometimes in association with liver angiomas, in 10 members of 4 generations. No neurologic symptoms were detected in subjects from the first 2 generations but symptoms were found in adult age in members of the third generation; 2 fourth-generation members came under medical observation at 2.5 years of age. Symptoms included partial epileptic fits which sometimes became generalized later and were generally controlled adequately by therapy. None of the patients was mentally retarded or restricted in daily life. Because of symptomatic hepatomegaly and postmortem finding of multiple liver and brain angiomas in a member of the first generation, liver ultrasonography was performed in all members of the family with detection of liver angiomas in members of the second and third generation. Retinal angioma was detected in 1 patient.

Labauge et al. (1998) established the clinical and genetic features of hereditary cavernous angiomas in a series of 57 French families. Neuroimaging investigations confirmed the high frequency of multiple lesions in hereditary cavernous angiomas. It also showed a correlation between the number of lesions and the age of the patient, suggesting a dynamic nature for the lesions.

Among 202 KRIT1 mutation carriers from 64 families, Denier et al. (2004) found that 126 had CCM and 76 were symptom-free. Mean age at clinical onset was 29.7 years, with 55% of patients presenting with generalized and/or partial seizures, and 32% with cerebral hemorrhages. The number of lesions averaged 4.9 on T2-weighted MRI and 19.8 on gradient echo MRI. Only 5 asymptomatic mutation carriers had no detectable lesions on T2-weighted MRI and gradient echo MRI. Denier et al. (2004) found that nearly half of mutation carriers aged 50 years or more were symptom-free, demonstrating clinical and radiologic incomplete penetrance of the disease.

Waters et al. (2005) reported a patient with familial CCM1 and a history of multiple CCMs, including an acutely hemorrhagic left cerebellar CCM that was resected. He also had an angiomatous skin lesion. The patient presented with acute onset of bilateral lower extremity weakness and numbness, and inability to urinate for 36 hours. Spinal MRI showed an intramedullary mass at levels T11 to T12, consistent with a cavernous angioma and hematoma. Surgical resection was successful. Waters et al. (2005) emphasized that patients with multiple CCMs tend to have predominantly intracerebral lesions, but that malformations may occur throughout the neuroaxis, including the spinal cord.

In a comparison of clinical features between mutation carriers from 86 families with CCM1 and 25 families with CCM2, Denier et al. (2006) observed that the number of gradient-echo sequence cerebral lesions increased more rapidly with age in patients with CCM1 than in those with CCM2.

Battistini et al. (2007) reported 5 unrelated Italian families with CCM1. The mean age at symptom onset was 15.9 years (range 4 to 36). The most common presenting symptoms included seizures (67%), recurrent headache (20%), and cerebral hemorrhage (13%). Genetic analysis identified 5 different heterozygous KRIT1 mutations (see, e.g., 604214.0009). The families included 33 mutation carriers, 57.6% of whom had no symptoms. Brain MRI revealed lesions in 82.3% of symptom-free mutation carriers.

Cutaneous Involvement

Norwood and Everett (1964) reported the case of a 21-year-old black female who during pregnancy developed large hemangiomas at many sites, such as earlobe and axilla, and heart failure as a result. After delivery, the hemangiomas rapidly subsided. The patient's mother and 6-year-old son had macular hemangiomas of the face and trunk and her brother had classic Klippel-Trenaunay-Weber syndrome (149000) of the right lower extremity. Beers and Clark (1942) described a family with cutaneous hemangiomas ranging in size from a millimeter to many centimeters in diameter, in 12 persons in 3 generations. Metatarsus atavicus (second toe longer than the first toe, see 189200) was an independent dominant trait in this family.

Keret et al. (1990) described an 18-year-old male with left scrotal cavernous hemangioma. Cutaneous hemangiomata were found in 34 relatives (21 males and 13 females). Only the proband had a genital lesion. The differentiation of scrotal hemangioma from varicocele was discussed.

Hyperkeratotic cutaneous capillary-venous malformations (HCCVMs) are crimson-colored, irregularly shaped lesions, the size of which may extend to several centimeters. By light microscopy, the lesions extend into both dermis and hypodermis and are composed of dilated capillaries and blood-filled venous-like channels. The overlying epidermis is hyperkeratotic. HCCVMs have been reported to be associated with cerebral capillary malformations (CCMs) (Labauge et al., 1999; Ostlere et al., 1996). Cerebral capillary malformations resemble HCCVMs in that both are composed of abnormal capillaries and venous-like vessels. In families in which these lesions coexist, all members who have HCCVMs also have CCMs (Eerola et al., 2000).

Diagnosis

Lehnhardt et al. (2005) found that MRI with T2-weighted gradient-echo sequences was more sensitive than routine MRI with T1-weighted and T2-weighted spin-echo sequences in determining lesion number and disease extent in affected members of a 3-generation family with CCM. One patient with a single lesion on routine MRI showed an additional lesion on gradient-echo sequences only, and 2 patients showed greater extent of disease only on gradient-echo sequences.

Pathogenesis

For each of the 3 CCM genes, Pagenstecher et al. (2009) showed complete localized loss of either KRIT1 (604214), CCM2/malcavernin, or PDCD10 (609118) protein expression depending on the respective inherited mutation. Cavernous but not adjacent normal or reactive endothelial cells of known germline mutation carriers displayed immunohistochemical negativity only for the corresponding CCM protein, but stained positively for the 2 other proteins. Immunohistochemical studies demonstrated endothelial cell mosaicism as neoangiogenic vessels within caverns from a CCM1 patient and normal brain endothelium from a CCM2 patient stained positively for KRIT1 and CCM2/malcavernin, respectively. Pagenstecher et al. (2009) suggested that complete lack of CCM protein in affected endothelial cells from CCM germline mutation carriers supports a 2-hit mechanism for CCM formation.

Maddaluno et al. (2013) demonstrated that endothelial-specific disruption of the KRIT1 gene in mice induces endothelial-to-mesenchymal transition, which contributes to the development of vascular malformations. Endothelial-to-mesenchymal transition in KRIT1-ablated endothelial cells is mediated by the upregulation of endogenous bone morphogenetic protein-6 (BMP6; 112266) that, in turn, activates the transforming growth factor-beta (TGF-beta; 190180) and BMP signaling pathway. Inhibitors of the TGF-beta and BMP pathway prevented endothelial-to-mesenchymal transition both in vitro and in vivo and reduced the number and size of vascular lesions in KRIT1-deficient mice. Thus, increased TGF-beta and BMP signaling, and the consequent endothelial-to-mesenchymal transition of KRIT1-null endothelial cells, are crucial events in the onset and progression of cerebral cavernous malformation disease.

Using a neonatal mouse model of CCM disease, Zhou et al. (2016) showed that expression of the MEKK3 (602539) target genes Klf2 (602016) and Klf4 (602253), as well as Rho and ADAMTS protease activity, are increased in the endothelial cells of early CCM lesions. By contrast, Zhou et al. (2016) found no evidence of endothelial-mesenchymal transition or increased SMAD (e.g., 601595) or Wnt (see 164820) signaling during early CCM formation. Endothelial-specific loss of Map3k3 (Mekk3), Klf2, or Klf4 markedly prevents lesion formation, reverses the increase in Rho activity, and rescues lethality. Consistent with these findings in mice, Zhou et al. (2016) showed that endothelial expression of KLF2 and KLF4 is increased in human familial and sporadic CCM lesions, and that a disease-causing human CCM2 mutation is normally expressed in HEK293 cells, binds KRIT1 and PDCD10 comparably to wildtype, but abrogates the MEKK3 interaction without affecting CCM complex formation. The authors concluded that their studies identified gain of MEKK3 signaling and KLF2/4 function as causal mechanisms for CCM pathogenesis.

Tang et al. (2017) identified TLR4 (603030) and the gut microbiome as critical stimulants of CCM formation. Activation of TLR4 by gram-negative bacteria or lipopolysaccharide accelerated CCM formation, and genetic or pharmacologic blockade of TLR4 signaling prevented CCM formation in mice. Polymorphisms that increase expression of the TLR4 gene or the gene encoding its coreceptor CD14 (158120) were associated with higher CCM lesion burdens in humans. Germ-free mice were protected from CCM formation, and a single course of antibiotics permanently altered CCM susceptibility in mice. Tang et al. (2017) concluded that their studies identified unexpected roles for the microbiome and innate immune signaling in the pathogenesis of a cerebrovascular disease, as well as strategies for its treatment.

Mapping

Using linkage analysis and a set of short tandem repeat polymorphisms, Dubovsky et al. (1995) mapped a gene responsible for cerebral cavernous malformations in a large Hispanic kindred to 7q11-q22. The maximum pairwise lod score of 4.2 was obtained at zero recombination with a marker at locus D7S804. Lod scores in excess of 3.0 were obtained with 4 additional markers closely linked to D7S804. A chromosome 7q haplotype of 33 cM on the sex-averaged map was shared by all affected individuals, indicating that the gene lies between D7S502 and D7S479. Using a linkage approach in 2 extended cavernous malformation kindreds, Gunel et al. (1995) also linked cavernous malformations to 7q, specifically 7q11.2-q21. Multipoint linkage analysis yielded a maximum lod score of 6.88 with zero recombination with D7S669 and localized the gene to a 7-cM region in the interval between ELN (130160) and D7S802. This gene is symbolized CCM1 for 'cerebral cavernous malformations-1.'

Marchuk et al. (1995) likewise mapped the CCM1 gene to proximal 7q by linkage methods. In 2 families, 1 of Italian-American origin and 1 of Mexican-American origin, they found a combined maximum lod score of 3.92 at theta = 0.0 for marker D7S479. Haplotype analysis placed the locus between D7S502 proximally and D7S515 distally, an interval of approximately 41 cM. The chromosomal location distinguishes this disorder from the autosomal dominant vascular malformation syndrome (VMCM; 600195) in which lesions are primarily cutaneous; VMCM is due to mutation in a gene that maps to 9p21. Johnson et al. (1995) refined the CCM1 locus assignment to a 4-cM interval bracketed by D7S2410 and D7S689.

Gunel et al. (1996) found that 47 affected members of 14 Hispanic American kindreds shared identical alleles for up to 15 markers linked to the CCM1 gene in a short segment of proximal 7q. Ten patients with sporadic cases also shared these same alleles, indicating that they too had inherited the same mutation. Thirty-three asymptomatic carriers of the disease gene were identified, demonstrating the variability and age dependence of the development of symptoms and explaining the appearance of apparently sporadic cases. Gunel et al. (1996) concluded that virtually all cases of familial and sporadic cavernous malformation among Hispanic Americans of Mexican descent are due to the inheritance of the same mutation from a common ancestor.

Genetic Heterogeneity

Craig et al. (1998) reported analysis of linkage in 20 non-Hispanic Caucasian kindreds with familial CCM. Analyses showed linkage to 2 loci other than CCM1: CCM2 (603284) at 7p15-p13, and CCM3 (603285) at 3q25.2-q27. Multilocus analysis yielded a maximum lod score of 14.11, with 40% of kindreds linked to CCM1, 20% linked to CCM2, and 40% linked to CCM3, with highly significant evidence for linkage to 3 loci. Linkage to these 3 loci could account for inheritance of CCM in all kindreds studied. The penetrance of symptomatic disease among apparent gene carriers for kindreds linked to CCM1, CCM2, and CCM3 was 88%, 100%, and 63%, respectively. These differences were not explained by differences in age or gender of gene carriers among families, and none of the asymptomatic gene carriers in this analysis was under age 20.

Laberge et al. (1999) conducted a genetic linkage analysis on 36 French CCM1 families using 8 microsatellite markers mapping within the CCM1 interval. Admixture analysis showed that 65% of these families were linked to the CCM1 locus. Haplotype analysis of CCM1-linked families did not show any evidence for a strong founder effect.

Molecular Genetics

In 12 of 20 pedigrees with cerebral cavernous malformations, Laberge-Le Couteulx et al. (1999) identified mutations in the CCM1 gene (see, e.g., 604214.0001) that segregated with the affected phenotype. They suggested that the mutations in the CCM1 gene might result in a dominant-negative effect or a loss of function; they favored the second hypothesis. Sporadic forms of cavernous angiomas manifest as unique lesions and familial forms as multiple lesions, which evokes a Knudson double-hit mechanism and would be consistent with the need for a complete loss of CCM1 function for the appearance of cavernous angiomas. All the mutations they reported predicted truncated CCM1 proteins completely or partially devoid of the putative RAP1A-interacting region.

In 1 family in which 2 of 4 members with CCMs also had hyperkeratotic cutaneous capillary-venous malformations, Eerola et al. (2000) found a 1-basepair deletion in the KRIT1 gene (604214.0005). Another novel mutation in this gene (604214.0006) was found in a family with CCM only.

In a family in which 5 individuals had both retinal and cerebral cavernous angiomas, Laberge-Le Couteulx et al. (2002) identified a heterozygous mutation in the KRIT1/CCM1 gene (604214.0010).

In 29 families and 5 sporadic cases with CCM, Davenport et al. (2001) identified 10 novel mutations and 1 previously described mutation in the KRIT1 gene (see, e.g., 604214.0008). The high frequency of loss-of-function mutations suggested loss of a tumor suppressor mechanism. In a follow-up study, Verlaan et al. (2002) reported 7 additional novel mutations and 1 previously described mutation in the KRIT1 gene in families with CCM. In combination with the previous study, Verlaan et al. (2002) found that approximately 47% of CCM families carry KRIT1 mutations. The authors noted that the majority of mutations in the KRIT1 gene lead to a substantial alteration of the gene product, supporting a loss-of-function mechanism consistent with a tumor suppressor gene.

Cave-Riant et al. (2002) screened the KRIT1 gene in 121 unrelated CCM probands having at least 1 affected relative and/or showing multiple lesions on cerebral MRI. Fifty-two individuals (43%) were shown to have a KRIT1 mutation, and 42 distinct mutations were identified, all of which were predicted to result in premature stop codons. Cave-Riant et al. (2002) concluded that the underlying mechanism of CCM may be KRIT1 mRNA decay due to the presence of premature stop codons and KRIT1 haploinsufficiency.

Verlaan et al. (2004) identified a pathogenic mutation in the KRIT1 gene in 4 (29%) of 14 unrelated patients with sporadic CCM and multiple malformations. None of 21 unrelated patients with a single malformation had a KRIT1 mutation. Verlaan et al. (2004) concluded that genetic analysis is warranted in sporadic cases of CCM with multiple malformations. In 2 additional patients of the 14 sporadic CCM patients reported by Verlaan et al. (2004), Felbor et al. (2007) used multiplex ligation-dependent probe amplification to detect a large duplication and a large deletion, respectively, within the KRIT1 gene. Thus, 6 (42%) of the 14 sporadic patients had a KRIT1 mutation.

Revencu and Vikkula (2006) reviewed the 3 genetic forms of familial cerebral cavernous malformation identified to that time and evidence on disturbed function. They pointed to work indicating that the 3 CCM genes are expressed in neurons rather than in blood vessels. The interaction between CCM1 and CCM2, which was expected on the basis of their structure, had been proven, suggesting a common functional pathway.

Among 24 Italian families with CCM, Liquori et al. (2008) identified 5 with deletions in the CCM1 gene, including 1 complete deletion of the gene.

Through repeated cycles of amplification, subcloning, and sequencing of multiple clones per amplicon, Akers et al. (2009) identified somatic mutations that were otherwise invisible by direct sequencing of the bulk amplicon. Biallelic germline and somatic mutations were identified in CCM lesions from all 3 forms of inherited CCMs. The somatic mutations were found only in a subset of the endothelial cells lining the cavernous vessels and not in interstitial lesion cells. Although widely expressed in the different cell types of the brain, the authors also suggested a unique role for the CCM proteins in endothelial cell biology. Akers et al. (2009) suggested that CCM lesion genesis may require complete loss of function for 1 of the CCM genes.

Cau et al. (2009) identified 2 different mutations in the KRIT1 gene (see, e.g., C329X; 604214.0011) in 5 (71%) of 7 Sardinian families with CCM. Haplotype analysis of patients from 4 of the affected families indicated a founder effect for the C329X mutation.

Modifying Polymorphisms

Tang et al. (2017) found that TLR4 and CD14 expression parallels human CCM burden. They studied 830 genetic variants of 56 inflammatory and immune-related genes in 188 patients who carried a KRIT1 Q455X (604214.0004) variant and measured CCM lesion burden using MRI. Following statistical analysis, SNPs in only 2 genes, TLR4 (rs10759930) and CD14 (rs778587), were found to be significantly associated with increased CCM lesion number. Further analysis of genes in TLR4-MEKK3-KLF2/4 signaling pathways identified additional SNPs for TLR4 (rs10759931) and CD14 (rs778588) in linkage disequilibrium with those previously identified, but none in other pathway genes that associated with altered lesion burden. Tang et al. (2017) found that the SNPs in TLR4 and CD14 that are associated with increased CCM lesion number are in the 5-prime genomic region of each gene and constitute cis expression quantitative trait loci (QTLs) that positively regulate whole blood cell expression of TLR4 and CD14 in a dose-dependent manner corresponding with risk allele number. These results were corroborated using the GTEx Consortium data.