Nature Genetics was launched in 1992, just as the era of positional cloning was getting underway, and a large number of genes underlying single-gene human disorders were soon identified. Such a disease-gene identification is gratifying in and of itself, as it may lead to immediate help for those affected, and certainly tells us something new about the genetics of the human organism in general. The additional hope is that it will tell us something about more common (and complex) diseases that are related. This final hope has not often been borne out, which is why the example of filaggrin, and its role in atopic disease is worth celebrating. In the space of two months, Irwin McLean and his international team of collaborators have shown that mutations in the gene underlie ichthyosis vulgaris (itself a fairly common mendelian disorder of scaly skin), and now atopic dermatitis and associated asthma. Atopic disease affects about 20% of the population in the developed world.

Although filaggrin was a good candidate, and the mutant allele is quite common, this result was a long time coming. Irwin McLean explains why.

Irwin McLean writes:

In the outermost cells of the epidermis, a huge number of structural molecules are expressed which form a barrier largely impervious to external attack from pathogens, chemicals and allergens. This barrier also prevents water loss through the skin. Many genes encoding these late-differentiation proteins are located in a dense cluster on 1q21 known as the epidermal differentiation complex (EDC). Despite the high density of candidates on 1q21, surprisingly few genes for keratinizing disorders have emerged from the EDC. One exception is filaggrin, which has long been suspected of having a role in the most common keratinizing disorder, ichthyosis vulgaris, or common dry, scaly skin.

Upon terminal differentiation of epidermal keratinocytes, the 400 kDa profilaggrin is cleaved into 10-12 identical filaggrin peptides which aggregate and condense the keratin cytoskeleton. This protein-lipid matrix is then enzymatically cross-linked to form the skin barrier. From the late 1980s onwards, a number of biochemical, immunohistochemical and genetic linkage studies pointed to a possible filaggrin defect not only in IV, but also in an IV-like mouse mutant (flaky tail), and importantly, atopic dermatitis (“eczema”). However, filaggrin mutations have not been forthcoming.

The delay lies in the unusual structure of the FLG gene which consists of two small exons followed by a huge final exon encoding 10 or more tandemly repeated copies of the ~1 kb filaggrin sequence. Exon 3 varies also in size from 12-14 kb in the population and the repeats within it are incredibly similar in sequence at the DNA level, making it very difficult to amplify by PCR and even more difficult to sequence. For added fun, the very few specific bases that one finds within each filaggrin repeat often turn out to be polymorphic.

The problems of analyzing this type of exon were first encountered by Frances Smith and myself about 10 years ago when we undertook sequencing of the plectin gene (Smith FJD et al., Nat. Genet. 13, 450-457, 1996). Plectin’s final exon is about 7 kb in size and like filaggrin, consists of ~1 kb repeats. Part of the problem with this type of gene is knowing which repeat you have sequenced. It’s very easy to be fooled into thinking you have sequenced it all when in fact you’ve just sequenced one part of it multiple times. Basically our approach consisted of long-range PCR using various polymerase mixes and buffers, a great deal of staring at lengthy sequence alignments, and then making and testing a fairly frightening number of primers.

After plectin, we said “never again”, but following a conversation we had with our dermatology collaborator in Seattle, Phil Fleckman, whilst on a bus in Utah several months ago, Frances decided to take a shot at filaggrin. We were suspicious when realized the first IV samples we got had previously been in a number of different labs in the USA and Europe who had given up on filaggrin. Undeterred, Frances applied her strategy and after considerable effort, found the R501X mutation. By analyzing a number of key families identified by Alan Irvine in Dublin, the semi-dominant inheritance in IV became apparent. Finding a second loss-of-function mutation explained most of our IV families.

We were excited to find the cause of a really common monogenic defect carried by 9-10% of European people but filaggrin had a further reward in store for our efforts (and primer bills). From a published genome scan for atopic dermatitis (AD) susceptibility genes, the EDC locus had shown linkage. Many IV patients also have AD and in our IV families, the filaggrin variants gave a significant lod score as a second trait. To investigate further, we teamed up with Somnath Mukhopadhyay and Colin Palmer in Dundee, Hans Bisgaard in Copenhagen, and Alan Irvine to analyze a number of patient cohorts. Quickly screening thousands of people for the filaggrin variants was no picnic and involved most of the lab in an incredibly intense team effort but in every case, we found an extremely significant association of these mutations both with AD, and to a lesser extent, with the form of asthma that occurs with AD.

We hypothesize that the key early event in AD is an impairment of skin barrier function, as suggested previously by Bill Cookson and others. This leads to increased epidermal water loss, leading to the dry skin seen in IV and AD. Importantly, the defective barrier may well allow continual presentation of antigens and allergens to the immune system via the skin. This leads to AD and in some cases, AD with asthma. Further work with animal models should shed further light on these pathomechanisms and provide the means to test new therapies or preventative measures.

So, have we had enough of big repetitive genes? “Never again” is what we swore, the time before.