Publication

Sotorasib is a first-in class KRAS G12C covalent inhibitor in clinical development for the treatment of tumors with
the KRAS p.G12C mutation. In the nonclinical toxicology studies of sotorasib, the kidney was identified as a
target organ of toxicity in the rat but not the dog. Renal toxicity was characterized by degeneration and necrosis
of the proximal tubular epithelium localized to the outer stripe of the outer medulla (OSOM), which suggested
that renal metabolism was involved. Here, we describe an in vivo mechanistic rat study designed to investigate
the time course of the renal toxicity and sotorasib metabolites. Renal toxicity was dose- and time-dependent,
restricted to the OSOM, and the morphologic features progressed from vacuolation and necrosis to regeneration
of tubular epithelium. The renal toxicity correlated with increases in renal biomarkers of tubular injury.
Using mass spectrometry and matrix-assisted laser desorption/ionization, a strong temporal and spatial associ –
ation between renal toxicity and mercapturate pathway metabolites was observed. The rat is reported to be
particularly susceptible to the formation of nephrotoxic metabolites via this pathway. Taken together, the data
presented here and the literature support the hypothesis that sotorasib-related renal toxicity is mediated by a
toxic metabolite derived from the mercapturate and B-lyase pathway. Our understanding of the etiology of the rat
specific renal toxicity informs the translational risk assessment for patients.

We investigated in a rabbit model, the eye distribution of topically instilled benzalkonium (BAK) chloride a commonly used preservative in eye drops using mass spectrometry imaging. Three groups of three New Zealand rabbits each were used: a control one without instillation, one receiving 0.01%BAK twice a day for 5 months and one with 0.2%BAK one drop a day for 1 month. After sacrifice, eyes were embedded and frozen in tragacanth gum. Serial cryosections were alternately deposited on glass slides for histological (hematoxylin-eosin staining) and immunohistological controls (CD45, RLA-DR and vimentin
for inflammatory cell infiltration as well as vimentin for Mu¨ller glial cell activation) and ITO or stainless steel plates for MSI experiments using Matrix-assisted laser desorption ionization time-of-flight. The MSI results were confirmed by a roundrobin study on several adjacent sections conducted in two different laboratories using different sample preparation methods, mass spectrometers and data analysis softwares. BAK was shown to penetrate healthy eyes even after a short
duration and was not only detected on the ocular surface structures, but also in deeper tissues, especially in sensitive areas involved in glaucoma pathophysiology, such as the trabecular meshwork and the optic nerve areas, as confirmed by images with histological stainings. CD45-, RLA-DR- and vimentin-positive cells increased in treated eyes. Vimentin was found only in the inner layer of retina in normal eyes and increased in all retinal layers in treated eyes, confirming an activation response to a cell stress. This ocular toxicological study confirms the presence of BAK preservative in ocular surface structures as well
as in deeper structures involved in glaucoma disease. The inflammatory cell infiltration and Mu¨ller glial cell activation confirmed the deleterious effect of BAK. Although these results were obtained in animals, they highlight the importance of the safety-first principle for the treatment of glaucoma patients.