Small-Molecule Therapies for Genetic Skin Fragility

Of the known genetic skin diseases, dystrophic epidermolysis bullosa (DEB) is a prime target for therapy development. It is a monogenic, strongly disabling disorder with high unmet medical need. Lifelong generalized skin blistering as a consequence of minor mechanical friction, chronic wounds, excessive scarring, and progressive soft-tissue fibrosis are clinical hallmarks. Joint contractures and mitten deformities of the extremities cause severe disability, and the fibrosis of the skin increases tissue stiffness and favors development of aggressive squamous cell carcinoma, a feared complication of the disease.¹ For obvious reasons, many investigators have pursued curative gene-, cell-, and protein-based therapy approaches for DEB, but these have turned out to be much more challenging than anticipated and have shown very limited benefit in pilot clinical trials.² In the recent past, it has become clear that alternative approaches are urgently needed to improve the health and quality of life of severely affected patients. In line with this, Cogan et al.³ report in this issue on the use of aminoglycosides to induce readthrough of premature termination codons (PTCs) and to restore functional protein production in DEB keratinocytes and fibroblasts in vitro.

DEB is caused by mutations in the COL7A1 gene, which encodes collagen VII, the major component of anchoring fibrils, which attach the epidermis to the dermis and are pivotal for the integrity of the skin. All clinical forms of DEB—mild and severe, recessive (RDEB) and dominant—are allelic disorders caused by different mutations in the COL7A1 gene. To date, more than 700 distinct mutations have been reported, and homozygous, heterozygous, and compound heterozygous mutation constellations produce a broad spectrum of phenotypes.⁴ It is estimated that 10–25% of disease-causing COL7A1 mutations represent nonsense mutations that generate a PTC.

Cogan et al.³ treated cultured primary fibroblasts and keratinocyte cell lines derived from RDEB skin with three different aminoglycosides (geneticin, gentamicin, and paromomycin) to suppress PTC mutations and to induce collagen VII expression in the cells. The treatment resulted in the synthesis and secretion of collagen VII in both cell types, reaching 15–30% of expected normal levels. The effects were dose-dependent and reached a peak about one week after the treatment, tapering off thereafter, but collagen VII protein could still be observed in the cultures 14 days after the treatment. Functional tests demonstrated that the abnormal migratory phenotype of collagen VII–deficient cells improved after aminoglycoside treatment, and in organotypic skin-equivalent cultures constructed with drug-treated RDEB keratinocytes and fibroblasts, collagen VII was deposited at the epidermal–dermal interface, suggesting that the neosynthesized collagen may have normal structure and functions. The suppression of PTC was also demonstrated in HEK293 cells transiently transfected with constructs containing 22 known RDEB nonsense mutations.

After aminoglycoside treatment, collagen VII synthesis reached 15–30% of expected physiological levels—a positive finding. Preclinical studies in mice and genotype–phenotype correlations in humans have shown that the quantity of collagen VII and the density of anchoring fibrils in the skin determine the extent of its integrity.^1,5 We have shown that about 30–35% of physiological collagen VII levels provide reasonable skin integrity in mouse models.^5,6 These data are important from the point of view of therapy development for completely collagen VII–deficient DEB: it is reassuring that a 100% restoration of collagen VII will not be required for improvement of the phenotype.

The precise structure and functionality of collagen VII synthesized in aminoglycoside-treated RDEB cells in vitro remain elusive at present. Under physiological conditions, the collagen is expressed in and secreted by both keratinocytes and fibroblasts, and it undergoes a series of post-translational modification/maturation steps during protein folding into triple-helical collagen and supramolecular assembly into functional anchoring fibrils. Although Cogan et al.³ undertook preliminary characterization of the newly synthesized collagen VII using immunofluorescence and immunoblotting assays, future studies using powerful proteomics and structural biology techniques will have to provide accurate data on detailed structure–function relationships of the aminoglycoside-induced collagen.

It has been about 15 years since the first proof of concept for treating genetic disorders resulting from PTC mutations was reported for muscular dystrophy.⁷ Since then, the approach has been validated in vitro and in animal models, such as zebrafish or mouse, for a number of disorders including muscular dystrophies, cystic fibrosis, and xeroderma pigmentosum.^8,9,10 However, translation into human therapies has been limited by the well-known toxicity of this group of drugs or lack of efficacy.⁸

One reason for the latter may be the fact that the molecular mechanisms of translation termination and nonsense-mediated messenger RNA decay (NMD) are not yet fully understood. Although the core NMD machinery is evolutionarily highly conserved among species, and it is known that the specific nucleotide context in the environment of the PTC plays an important role, additional complexity of these processes is emerging. Different models of PTC recognition and NMD that are employed depending on the exon–intron structure of the specific genes have been presented (for review, see ref. 8). The family of collagen genes, including COL7A1, encompasses 42 large multiexon genes that are associated with a number of genetic disorders in many tissues.¹¹ So far, published data on PTC recognition and NMD associated with mutations in collagen genes support two fundamentally different mechanisms: (i) NMD dependent on the exon junction complex in the case of collagens I and VI and (ii) PTC recognition with the involvement of the 3′-untranslated region in the case of collagen X.¹² No information exists on these mechanisms in relation to COL7A1 nonsense mutations. Therefore, the applicability of the current PTC recognition and NMD models for the collagen VII gene and, consequently, corresponding therapeutic strategies remain unclear.¹²

The in vitro study by Cogan et al.³ raises cautious optimism for biologically valid treatments of RDEB. The application of specific PTC readthrough compounds targeting the individual nonsense mutation constellation of the patient holds promise for noninvasive drug therapies, which would, of course, be highly desirable for this intractable disease. Such treatments might benefit a subset of patients with RDEB carrying certain nonsense mutations—perhaps up to a tenth of the affected individuals—and improve their skin integrity and quality of life.

However, substantial hurdles lie ahead and must be overcome before clinical applications can become realistic. First, the specific molecular basis of NMD and PTC readthrough associated with COL7A1 nonsense mutations will have to be elucidated, and maximally specific aminoglycosides or aminoglycoside-like compounds identified for each individual patient. Second, given that PTC readthrough introduces an amino acid in lieu of the stop codon, the effects of the substitution must be assessed in light of the fact that single amino acid substitution mutations in collagen VII are also known to cause DEB, albeit milder phenotypes. Third, the toxicity of aminoglycosides poses a serious challenge. Because RDEB manifests at birth and progresses steadily, treatment must be initiated at an early age. Indeed, adverse effects of a long-term treatment in children may pose insurmountable problems, if no alternative molecules facilitating PTC readthrough will be available. The option of topical application of aminoglycosides may seem attractive at first glance but may still exhibit toxicity and not suffice to repair internal mucosal lesions and extracutaneous organ involvement in RDEB. Therefore, alternative systemic agents should be sought.

Several aminoglycoside derivatives or nonaminoglycoside compounds that facilitate PTC readthrough without the toxicity of aminoglycosides have been identified, and current screening procedures are likely to discover new compounds.⁸ Of these, PTC124, or ataluren, has been extensively investigated and seemed promising to stop progress of muscular dystrophies and cystic fibrosis in preclinical disease models.¹³ Not all of the phase II and III clinical trials have been conclusively evaluated (http://www.ptcbio.com/clinical_trials), but it seems that the drug has a good safety profile and improves some symptoms,¹³ although not all patients benefit from the treatment.¹⁴ Presumably, as delineated above, future therapies must be applied on an individualized basis, selecting patients with optimal mutation and PTC recognition constellations that respond to a particular drug and its mode of action. Of note, in the study by Cogan et al.,³ all DEB cells and expression constructs carrying COL7A1 nonsense mutations were refractory to PTC124 treatment. Therefore, it is likely that novel compounds with different specificities, or novel strategies to block NMD, will need to be found for clinically applicable drug therapies for DEB. Hopefully, continually increasing insights into the molecular mechanisms of PTC recognition and NMD,^8,13 in general and with collagen genes, will lead to discovery of small-molecule compounds with true therapeutic potential for DEB.