1. Pathophysiology of vascular anomalies
(For recent reviews on ettiopathogenesis of vascular anomalies, see Brouillard & Vikkula, Clin Genet 2005 and Hum Mol Genet 2007 ; Revencu & Vikkula, J Med Genet 2007 ; Limaye et al, Hum Mol Genet 2009)
Venous malformations
Our work was initiated with a large family, in which multifocal cutaneous venous malformations (VMCM) were inherited as an autosomal dominant disorder. With a genome-wide linkage approach we identified that the causative gene must lie on the short arm of chromosome 9, and subsequently discovered that the mutated gene was TIE2/TEK. This gene encodes an endothelial cell specific receptor tyrosine kinase, and the affected individuals carried a single nucleotide change, which leads to an R849W substitution in the intracellular kinase domain of the receptor.
Somatic mutations in the gene TEK profoundly affect TIE2 activation and cause common, sporadic venous malformations. Here, normal (left) and abnormal (right) endothelial cell.
We demonstrated that when overexpressed the mutant receptor is overphosphorylated by 4-6 fold compared to the wild-type (Vikkula et al., Cell 1996). The same mutation was identified in two families, and since then, we have identified 12 additional families with a TIE2 mutation (Wouters et al., submitted). These data demonstrate the crucial role of TIE2 signaling for angiogenesis in man, and pinpoint it as target for drug development.
With a genome-wide scan, we identified another locus for hereditary venous lesions on chromosome 1p (Boon et al., Am J Hum Genet 1999). The patients had more painful, gobblestone appearing lesions, called glomuvenous malformations (GVM). By positional cloning, using yeast and bacterial artificial chromosomes, and by identification of haplotype sharing, we identified the mutated gene, which we named glomulin (Brouillard et al., Genomics 2000; Irrthum et al., Eur J Hum Genet 2001; Brouillard et al., Am J Hum Genet 2002). Most mutations caused premature stop codons and were predicted to cause loss of glomulin function. Although the encoded protein was completely unknown, we have unravelled that it is an intracellular protein specifically expressed in vascular smooth muscle cells (McIntyre et al., Gene Express Patterns 2004). We have also generated glomulin knock-out mice (Brouillard et al., unpublished). The heterozygotes are phenotypically normal, whereas the homozygotes die early during embryogenesis. A conditional glomulin knock-out is now being generated to better dissect glomulin function beyond the time-point of lethality. Importantly, the identification of the pathophysiological basis of VMCM and GVM allowed us to study genotype-phenotype correlation, and to establish criteria for diagnostics and management (Boon et al., Arch Dermatol 2004).
Somatic mutations
In 1994, when we mapped the 9p21 locus for VMCM, we hypothesized that the variation in size, number and localization of the multifocal lesions may follow Knudson’s double-hit hypothesis for retinoblastoma. On this basis, we predicted that a somatic second-hit is needed for lesions to develop. We have been able to test this in one resected glomuvenous malformation and recently in one VMCM. In the GVM, a 5 bp deletion leading to a premature stop codon was identified on the allele, which does not carry the inherited mutation. Thus, locally in the lesion, there is a complete loss of function of glomulin (Brouillard et al., Am J Hum Genet 2002). Similarly, in the one VMCM-tissue, we have identified a second-hit (Limaye et al., Nat Genet 2009). This was an in-frame deletion of 129 bp affecting the extracellular ligand-binding domain of the receptor. In vitro, the over-expressed mutant receptor did not get phosphorylated even after ligand induction, and it was unable to locate on cell membranes. Thus, the mutation likely causes loss of function. As the inherited mutation causes gain of function, we hypothesize that the role of the second-hit in VMCM is to remove a protective effect of the wild-type allele (Limaye et al., Nat Genet 2009).
These findings led us to study the pathophysiology of non-inherited venous malformations as well. They account for 94% of venous anomalies, the remaining 6% being accounted for by the inherited VMCMs and GVMs. We hypothesized that these much more common sporadic VMs may be caused by somatic mutations in TIE2, even in the absence of a “predisposing” germline mutation. By screening lesion-derived DNA from 57 patients, we detected a mutation in 30/62 lesions. This accounts to almost 50% of the patients, and is thus, at the population level, the most frequent cause of vascular anomalies so far identified (n=28/57)(Limaye et al., Nat Genet, 2009). The mutations included seven novel missense mutations, which were not present in the blood of the patients. They were also absent in 89 control tissues (p=9.006e-15) demonstrating that they are not common, non-associated somatic changes.
Lymphedema
Our work on lymphedema started with a family with autosomal dominant primary lymphedema (Milroy disease), the cause of which we were able to map to chromosome 5q. The identified mutation in the intracellular kinase domain of VEGFR3, causes loss of receptor phosphorylation (Irrthum et al., Am J Hum Genet, 2000). We have subsequently demonstrated that mutations in this same gene can also cause sporadic lymphedema, due to lowered penetrance of the inherited mutation or by de novo mutations (Ghalamkarpour et al., Clinical Genetics 2006). Moreover, 2/12 sporadic hydrops fetalis cases were due to de novo VEGFR3 mutations (Ghalamkarpour et al., J Pediatr 2009). Finally, with a family of cousin-parents we demonstrated that some VEGFR3 mutations are recessive (Ghalamkarpour et al., J Med Genet 2009). These data have clarified the classification of primary lymphedema, and preclinical studies have elsewhere been started to use VEGFR3 ligands to ameliorate lymphatic dysfunction in VEGFR3 animal models.
With our continued collection of DNA and tissue samples, as well as primary cell lines, we have also been able to study a syndromic form of lymphedema, called HTL (for hypotrichosis-lymphedema-telangiectasia). We found that a similar combination of signs and symptoms existed in a mouse called the ragged mouse. Since the murine genetic locus was known, we specifically tested if the syntenic locus in the human genome would show linkage in our largest HTL-family. As this was the case, we screened the candidate gene SOX18, and identified mutations in all three available families (Irrthum et al., Am J Hum Genet, 2003). Two of the mutations were recessive, and altered the DNA-binding domain of this transcription factor, whereas the third one was dominant, and caused a premature stop codon in the transactivating domain. These studies put SOX18 in the forefront of lymphatic biology and it has since been shown to be a major controller of lymphatic endothelial differentiation.
Capillary malformations
Our interdisciplinary studies have also led to recognize a disorder, previously unknown to the medical literature. We named it CM-AVM for capillary malformation-arteriovenous malformation (Eerola et al, Am J Hum Genet, 2003). The identification was based on a whole-genome linkage analysis of a number of families in which cutaneous capillary malformations were present in more than 2 individuals. We discovered that some of the families linked to 5q had mutations in the RASA1 gene, which encodes a RASGTPase. The loss-of-function mutations were present only in families in which the affected individuals had atypical capillary malformations and some had also fast-flow malformations (Eerola et al, Am J Hum Genet, 2003). So far, we have been able to identify a hundred families (Revencu et al., Hum Mutat, 2008; and unpublished). It has become clear that 30% of the patients have fast-flow lesions and 80% of those are in the head and neck region (Limaye et al., Hum Mol Genet 2009). The clinical spectrum includes Parkes Weber syndrome and vein of Galen aneurysmal malformation. Moreover, some patients have specific neurologic tumors. CM-AVM is also likely to be as frequent as hereditary haemorrhagic telangiectasia (HHT), a well-known Mendelian disease. Since the fast-flow lesions can be difficult to treat, and the RASA1 mutations lead to increased Ras activity, small molecule inhibitors may prove useful as therapy in the future.
To know more... (pdf chapter of the last de Duve Institute report)
Publications >
Next project >
