Peter vd Akker graduates cum laude on epidemolysis bullosa

Peter van den Akker graduated cum laude on 6 November 2013.

The topic of this thesis is the heritable blistering disorder dystrophic epidermolysis bullosa (DEB). DEB is one of the four major types of epidermolysis bullosa (EB), a heterogeneous group of disorders that involve skin blistering after minor trauma or friction and that can be accompanied by a variety of extracutaneous manifestations. DEB is caused by mutations in the COL7A1 gene. This gene has 118 exons and encodes the non-fibrillar type VII collagen a 1 precursor chain (pro- a 1(VII) procollagen peptide), three of which assemble into procollagen VII homotrimers. After excretion from the basal keratinocyte, the major source of type VII collagen production, two of these homotrimers align in an antiparallel way to form mature type VII collagen dimers. Ultimately, large numbers of these dimers aggregate to form the anchoring fibrils. Anchoring fibrils conclude the hemidesmosomal epidermal-dermal adhesion complex at the dermal side and are involved in tightly anchoring of the epidermis to the underlying dermis. Defective type VII collagen leads to a weakened epidermal attachment and, consequently, to skin fragility and a DEB phenotype.

DEB can be inherited either autosomal recessively (RDEB) or dominantly (DDEB). The DEB phenotype covers a spectrum from the most severe RDEB, severe generalized (RDEB-sev gen) subtype to the mildest DDEB, nails only (DDEB-na) subtype. RDEB-sev gen is characterized by neonatal-onset generalized blistering, extensive scarring that results in ‘mitten’ deformities of the hands and feet due to fusion (pseudosyndactyly) of the digits in childhood, nail dystrophies, mucosal erosions, growth retardation, and a life expectancy that is reduced to 30-40 years due to the exceptionally high risk of squamous cell carcinomas. The other subtypes are usually milder, and DDEB is usually less severe than RDEB.

Previous research has shown that the precise effect of a COL7A1 mutation (i.e. the genotype) on protein function determines which subtypes of DEB a patient could develop (i.e. the phenotype), which has resulted in general genotype-phenotype correlation concepts. Mutations reducing the amount of wild-type type VII collagen act recessively and generally cause the severest RDEB phenotypes of the DEB spectrum, whereas missense and splice-site mutations that alter the structure of the protein can act either recessively or dominantly and cause milder phenotypes. DDEB is usually caused by a specific class of missense mutations, the glycine substitution, where an invariable glycine residue of one of the many collagenous Gly-Xaa-Yaa triplet repeats is mutated. Not all glycine substitutions, however, act dominantly and some can only cause RDEB. There are other exceptions to these general genotype-phenotype correlation rules and, moreover, for some of the rarer phenotypes the genotype-phenotype correlation is only now beginning to emerge. Many aspects of the pathogenesis of DEB remain elusive.

The aims of this thesis

The general aim of this thesis was to expand the understanding of the pathogenesis of DEB by exploring the Dutch EB database of the Groningen Centre for Blistering Diseases (head Prof. M.F. Jonkman). This is the EB Center of Expertise in the Netherlands and most of the Dutch EB patients have visited it at least once. Information about their clinical phenotype, family history, results of immunofluorescence and electron microscopy studies on skin biopsies, as well as results of mutation analysis, have been stored in the database since 1989. Per May 2013, the Dutch EB database contained data on 346 different families and is thus a valuable source for EB research.

To achieve this aim, my project defined three major objectives. Part One of this thesis covers the first two objectives. The first objective was to assess the genotype-phenotype correlation for the major DEB subtypes in the Dutch population. A full understanding of the genotype-phenotype correlation is essential for accurate phenotype prediction, optimized and personalized genetic counseling, further unraveling the pathogenesis, substantiating the role of genetic and non-genetic disease-modifying factors, and selecting patients for clinical trials.

The second objective of my thesis was to build a public, online database containing all available clinical and molecular information of published and unpublished DEB patients. Studying the genotype-phenotype correlation provides important insights into the disease, but these can only be utilized optimally if the information coming from such studies is collected in an online database that is accessible to all medical professionals and researchers worldwide.

Part Two of this thesis covers my third objective: to substantiate the disease-modifying role of somatic mosaicism in DEB. Somatic mosaicism refers to the co-existence of mutant and wild-type cell populations in the body, which results in a mosaic disease distribution. In ‘forward’ mosaicism, the germline is wild-type and a pathogenic mutation has arisen during embryonic development, which leads to a segmental disease phenotype. In ‘revertant’ mosaicism the germline is mutant, but in a portion of cells the effect of the mutations is corrected by a somatic event, which leads to normal protein production and a healthy skin patch. Both forward and revertant mosaicism can thus be regarded as disease-modifiers.

Introductory chapters 1 to 3

Before the three objectives are worked out, this thesis starts with three general, introductory chapters, in which I present and explain its context. Chapter 1 introduces basic concepts about the skin, its development and structure, and the hemidesmosomal adhesion complex. Next, it zooms in on EB as the group of heritable disorders that are caused by genetic defects in any of the genes encoding the molecules of this adhesion protein complex. The focus of this chapter is on the clinical and molecular aspects of DEB, its causative gene and the mutations therein, and the known genotype-phenotype correlation concepts. This chapter subsequently describes in detail the phenomena of forward and revertant mosaicism in general and in EB particularly. Throughout this chapter, the questions that formed the basis for the studies described in this thesis are introduced. The chapter concludes with the aims and outline of this thesis.

Chapter 2 places the topics described in this thesis in a historical perspective. It summarizes the major historical landmarks and achievements in DEB research, from the very first publications on DEB in the late 19^th century to the discovery of the COL7A1 gene as the disease-causing gene in the early 1990s. This chapter explains many of the names and concepts frequently found in the DEB literature and throughout this thesis. It shows how much work needed to be done to reveal the cause and pathogenesis of DEB, to reach consensus about the classification of the disease subtypes, and to enable correct disease diagnosis and prenatal testing. And it highlights that there is still much to do!

To study the pathogenesis of any genetic disease, it is crucial that the causative mutations have been identified in the families under study. For COL7A1 only a few mutations are specific to certain populations owing to founder effects and only a few mutational hotspots exist. Most mutations are, however, unique to families and are scattered throughout the entire 118 exons of the gene. This emphasizes the need for a sensitive, reliable, and efficient mutation scanning strategy. We therefore developed a conformation-sensitive capillary electrophoresis (CSCE) system for COL7A1 mutation scanning.

In chapter 3, the design and analytical validation of this technique for use as a COL7A1 mutation pre-screening tool in a diagnostic molecular genetics service laboratory is described. The CSCE technique is based on the principle of heteroduplex formation when PCR-amplified DNA fragments containing heterozygous sequence changes are slowly re-annealed. These fragments have different migration characteristics than wild-type fragments and can be distinguished on a multi-capillary automated sequencer if fluorescently labeled. Validation of the CSCE system was performed by analysis of 29 known COL7A1 sequence changes, covering 33% of amplicons. After optimization of the conditions, all 29 sequence changes were detected by the system, irrespective of the length or CG-content of amplicons and position of sequence changes, reflecting an analytical sensitivity of 90.2%-100% (95% confidence interval). We concluded that this CSCE system is a rapid, reliable, cost-effective, and highly sensitive way of scanning for COL7A1 mutations in a molecular genetics service laboratory. When we developed the CSCE system, it was a state-of-the art technique fulfilling all the requirements of modern-day DNA-diagnostics. Nowadays, however, in the next-generation sequencing era, it is justified to question whether there is still room for such pre-screening techniques.

>The full thesis will be available from http://dissertations.ub.rug.nl/faculties/medicine/2013/ by the end of November.