Characterization of phototoxicity reactions in human and animal skin models

Phototoxicity is a relatively common phenomenon and is an adverse effect of some systemic drugs. A large number of pharmaceutical drug substances are known to carry the potential to elicit a phototoxic response (Moore 2002, Ferguson 2002). Reported cases of phototoxic responses may account for 5 to 10 % of recorded cutaneous adverse drug reactions (Selvaag 1997). As these abnormal reactions seriously limit or exclude the usage of certain drugs, it is essential to identify such liabilities early in development. Therefore, photosafety of drug candidate molecules needs to be evaluated (ICH M3 R2, 2009). Often this follows a step-wise process starting with physicochemical properties (light absorption), followed by in vitro, in vivo and/or clinical testing as suggested by each successive study. A key principle in such a tiered testing strategy is that a negative result obtained in a generally accepted and highly sensitive assay does usually not warrant further testing. In this work, we were interested in comparing the results obtained in nonclinical models (in vitro and in vivo experiments) with the clinical signs observed in human of well-known photosensitizer. The results of our investigation of phototoxicity mechanisms are presented in three research papers published in (or submitted to) peer reviewed journals representing the core of this thesis. Starting from an established standard model we used the modified oral UV-Local Lymph Node Assay (UV-LLNA or photo-LLNA) in Balb/c mice. We demonstrated the performance of this optimized modified murine local lymph node assay, adapted for phototoxicity assessment of systemically applied compounds. Several clinically phototoxic reference compounds were tested in mice using a sun light simulator to establish time- and dose-dependent profiles. In order to further investigate the phototoxic reaction, we decided to focus on one molecule. We took the example of vemurafenib, a B-Raf kinase inhibitor for the treatment of patients with unresectable or metastatic melanoma carrying the BRAFV600E mutation, commercially available since 2011. This drug was selected because the phototoxicity of this drug was discarded in preclinical testing but has been reported in clinical trial. Therefore, it became a fundamental question to understand this discrepancy. The aforementioned mouse oral UV-Local Lymph Node Assay, was used to investigate the impact of formulations, dose levels, duration of treatment and timing of irradiation. The initial studies performed as part of the non-clinical development of vemurafenib with hairless rats was compared to the current study in mice. Duration of treatment and exposure to both vemurafenib and the formal UVA dose (limited to 350 to 400 nm) was clearly exceeding the conditions we have used in our studies in mice. The most apparent difference was the spectral range of the irradiation light source (350 to 400 nm versus 320 to 700 nm). Since vemurafenib does not absorb sufficiently light above 350 nm, this difference alone can easily explain the negative earlier study result in hairless rats. To enhance our molecular understanding of phototoxicity mechanisms, an imaging technique based on matrix-assisted laser desorption/ionization mass spectrometric (MALDI-MS) was applied to samples from mouse skin and from a human 3D skin model. Using sparfloxacin as a model compound, concentration-dependent and irradiation-dependent effects could be observed in vitro. Furthermore, in the aforementioned established in vivo phototoxicity model, time- and irradiation dependent exposure to sparfloxacin in skin samples from mouse ears following oral treatment were demonstrated. Taken together, these results illustrate important lessons regarding photosafety testing. First of all, they demonstrate that the modified murine UV-LLNA is suitable to support preclinical photosafety assessment of systemically applied drug candidates. Furthermore, they highlight the impact of carefully designed in vivo phototoxicity studies. It is apparent that duration of treatment, timing of irradiation and appropriate irradiation conditions are key parameters to ensure an appropriate sensitivity