Patients with an inherited blistering skin disorder — recessive dystrophic epidermolysis bullosa (RDEB) — often develop epidermal squamous-cell carcinoma (SCC). Functional defects in type VII collagen are responsible for RDEB, and Susana Ortiz-Urda and colleagues have disovered that SCC development in these patients is dependent on collagen-VII-mediated interactions between the tumour and its microenvironment.

Intriguingly, a small subgroup of individuals with RDEB never develop invasive SCC, and the investigators wondered why, when all patients with RDEB have flaws in the same protein, some are susceptible to this cancer and others are not. Initially, they assessed the tumorigenicity of primary keratinocytes from patients with RDEB. The cells were transformed using oncogenic RAS and the NF-κB inhibitor IκBα, and then they were transplanted into immunodeficient mice. In this assay, cells from non-RDEB individuals form epidermal tumours that appear the same as SCC. Tumours also grew from the keratinocytes from eight of the twelve patients with RDEB sampled, but not from cells from the remaining individuals with RDEB, confirming two distinct RDEB types — tumorigenic and non-tumorigenic.

The two populations could be further distinguished by the expression of collagen VII: primary keratinocytes from patients with tumorigenic RDEB express a portion of collagen VII, whereas the cells from the non-tumorigenic RDEB subset (termed RDEBNull) do not. Furthermore, expressing collagen VII in the RDEBNull cells restored their ability to form tumours. Ortiz-Urda and colleagues narrowed down the region of the protein responsible for tumour formation by expressing various fragments of collagen VII in RDEBNull cells. The sequence they identified (FNC1) restored tumorigenicity to RDEBNull cells, and antibodies against FNC1 prevented normal primary keratinocytes from forming tumours.

There are a range of neoplastic processes that FNC1 might affect — tumour-cell proliferation, survival and/or invasion — and the authors examined each of these. The FNC1 antibodies did not block tumour-cell division or trigger apoptosis. They did, however, halt invasion of normal keratinocytes, and restore the invasiveness of RDEBNull cells. Moreover, they stopped the expansion of tumours that were already established.

The FNC1 sequence contains a binding site for laminin-5, another extracellular protein that is found in SCC. Laminin-5 interacts physically with collagen VII, but the FNC1 antibodies disrupt this partnership. Moreover, it seems that laminin-5 is necessary for FNC1 to promote tumour invasion, as antibodies to laminin-5 inhibited this process.

Whether the presence of the FNC1 region in RDEB individuals is a reliable predictor that they will develop SCC has yet to be proven. But the authors suggest that as collagen VII and laminin-5 are extracellular, they will in principle be accessible to therapeutics such as antibodies.