By Amy Bianco, Loose Connections Medical Editor

A group of researchers in Belgium led by Dr. Anne De Paepe has been working to uncover more defects to type V collagen that would be diagnostic for Ehlers-Danlos syndrome, classic type. Genetic testing currently uncovers mutations in about half of patients meeting clinical criteria for cEDS. While any significant expansion of the database of known collagen V mutations would improve a cEDS patient’s chances of receiving molecular confirmation of his or her diagnosis, Dr. De Paepe’s group goes so far in a forthcoming paper¹ as to assert that the COL5A1 and COL5A2 genes are likely to be the sole culprits behind the disorder, and that with refined diagnostic criteria clinicians should be able to confirm the diagnosis 90% of the time.

Dr. De Paepe’s group employed exhaustive biochemical and molecular analytical techniques to look for causal involvement of collagen V in a cohort of 126 patients with diagnosed or suspected cEDS. They found that 48 patients had “null” COL5A1 mutations, where one allele of the gene (we have two, one from each parent) had been rendered non-functional, while “structural” mutations throughout the COL5A1 gene were identified in 23 patients. A mutation is considered structural when it alters a gene product (protein) in a manner that changes its functioning. In addition, in two patients the outcome of the COL5A1 mutation was undetermined. Mutations in the COL5A2 gene were found 13 patients; all of them were structural.  In all, the group added 49 COL5 mutations to the 72 in the database of known cEDS variations. No other genes were implicated.

One of the objectives of the study was to look for genotype–phenotype correlations, a Holy Grail in clinical genetic studies. By documenting the exact mutation in each patient (where possible) and comparing it to that patient’s clinical profile, the group hoped to link specific molecular defects to certain symptoms. Toward this end they made two important observations. 

First, they found that, unlike the situation with some other heritable disorders of connective tissue, with cEDS it seems to make no significant clinical difference whether the causative mutation is structural or null. In osteogenesis imperfecta, which is caused by a defect in the collagen I gene, and in Vascular Type Ehlers-Danlos syndrome, where collagen III is involved, it has been shown that a null mutation produces a less severe phenotype than a structural mutation². This is because with a null mutation, the remaining allele continues to function with the result that normal collagen is produced but in a reduced amount. A structural mutation, on the other hand, can disrupt the production or function of collagen in many ways. 

Dr. De Paepe’s group was able to elucidate various complex mechanisms by which myriad structural COL5 mutations arrive at a common pathogenic pathway:  in the end, they all have the effect of reducing the amount of normal type V collagen available in the extra cellular matrix. Thus the group asserts that all of the known COL5A1/COL5A2 defects can be considered “functionally” null. They reason that this may be because collagen V co-assembles with collagen I to create fibrils, and that in this union the less abundant collagen V plays a regulatory rather than a structural role. Any defect in collagen V would perturb this regulatory function and result in a disorganized extra cellular matrix. The authors note that there is a subset of structural  COL5 mutations that appear to disrupt the normal interaction of type V collagen with other constituents in the extracellular matrix³. Work is continuing to see what effects these defects might have on phenotypes.

Thus, except for the fact that all of the patients with COL5A2 mutations were located on the more severe end of the cEDS spectrum than those with COL5A1 mutations, Dr. De Paepe’s group found no significant correlation between the type of COL5 mutation and the patient’s phenotype. It made no difference whether the mutation was structural or null — or if it were structural, what kind of alteration to the gene it was.  

For diagnostic purposes, then, it would appear that testing for loss of expression of one COL5A1 allele is fundamental. Dr. De Paepe’s group was able to confirm such loss in seven patients for whom the underlying mutation could not be found. This brought to 93 the number of patients for whom causal involvement of type V collagen could be reported, leaving 33 of the original cohort in whom no mutation was found, or in whom a null-allele was excluded, or the presence of a null COL5A1 allele could be neither confirmed or excluded.

Perhaps the most striking finding of the study — and the other major observation Dr. De Paepe’s group was able to make regarding genotype–phenotype correlation — was that the presence of all of the major diagnostic criteria for cEDS is a reliable predictor that a type V collagen mutation will be found. The patients in the study were divided into two groups based on their clinical profiles. The first group, of 102 patients, fulfilled all three of the major criteria for classic EDS:  skin hyperextensibility; skin fragility resulting in widened, atrophic and “cigarette paper” scars; and joint hypermobility. Group 2 consisted of 24 patients who met only two of the major criteria. All of the patients in group 2 were among the 33 for whom no causative defect in collagen V could be reported. This gave the authors confidence to assume that technical limitations kept them from finding mutations in the remaining six patients who met all three major criteria for cEDS. So, of the 102 patients who met all major criteria for cEDS, the authors found evidence of a type V collagen defect in 93, about 90%. The authors thus propose that the Villefranche criteria for cEDS be made more stringent to require that all three major criteria be met.

It is interesting to note that in this study all of the patients in group 2 were missing the same major criterion:  (clinically significant) dystrophic scarring. The authors point out that biochemical collagen analysis may still be necessary (or contributory) for differential diagnosis with other EDS subtypes; specifically, they suggest ruling out kyphoscoliotic EDS in patients with severe (kypho)scoliosis and vascular EDS in patients (especially young children with mild skin involvement) where easy bruising or family history might suggest it. But the biggest impact of this study is likely to be on the differential diagnosis between cEDS and Ehlers-Danlos syndrome, hypermobility type. Here the more stringent cEDS diagnostic criteria will be crucial, for there is no molecular test for hEDS. At present the two can be difficult to differentiate in patients with a mild expression of the cEDS phenotype. Many hEDS patients also have soft, velvety skin (though the authors point out that it is not as doughy as in cEDS patients); it is difficult to measure skin hyperextensivity, which many hEDS patients exhibit to some degree; and of course, joint hypermobility is the most important major criterion for hEDS. A further challenge is presented by the surprising finding in this study that in two thirds of the mutation-positive patients the mutation was shown to occur de novo; that is, it had not been passed down from a parent. Thus family history would provide no guidance in most cases. On the other hand, with the accuracy of molecular testing for cEDS at about 90%, clinicians will be much more confident in ruling it out.

EDS experts will be meeting in Belgium this fall to revise the nosology for the various EDS subtypes. If they do indeed adopt more stringent diagnostic guidelines for cEDS, we can expect to see a greater emphasis placed on skin fragility by clinicians. Poor would healing, spontaneous splitting of the skin, and the tell-tale cigarette paper scars over joints will become a sine qua non of Ehlers-Danlos syndrome, classic type.