cathepsin D off-target activity drives ocular toxicity of β-secretase inhibitors

Medicilon’s toxicology department has professional teams with rich experience in toxicology studies. We offer high-quality data and rapid turnaround period to support drug discovery and development. Our toxicological studies are conducted in various animal species. The toxicological evaluation from dose design, in-life studies to histology and pathology testing along with toxicokinetics studies are all compliant with GLP or NON-GLP standards. Our study platform is certified as one of the Shanghai Public Service Research Platforms.

Website: www.medicilon.com E-mail: Marketing@medicilon.com.cn

Inhibition of β-secretase BACE1 is considered one of the most promising approaches for treating Alzheimer’s disease. Several structurally distinct BACE1 inhibitors have been withdrawn from development after inducing ocular toxicity in animal models, but the target mediating this toxicity has not been identified. Here we use a clickable photoaffinity probe to identify cathepsin D (CatD) as a principal off-target of BACE1 inhibitors in human cells. We find that several BACE1 inhibitors blocked CatD activity in cells with much greater potency than that displayed in cell-free assays with purified protein. Through a series of exploratory toxicology studies, we show that quantifying CatD target engagement in cells with the probe is predictive of ocular toxicity in vivo. Taken together, our findings designate off-target inhibition of CatD as a principal driver of ocular toxicity for BACE1 inhibitors and more generally underscore the power of chemical proteomics for discerning mechanisms of drug action.

In the development of therapeutic agents or chemical probes, on-target efficacy is typically tested in several cellular models before time consuming and costly in vivo testing. However, in many cases selectivity is only confirmed with in vitro assays against related protein targets or a broad panel of representative enzymes and receptors. This approach can fail to accurately predict selectivity, as compound potency can change markedly in complex biological settings and pharmacology is often shared outside of gene families. It can therefore be difficult to pinpoint whether unanticipated compound effects observed in animal models are a result of poorly understood consequences of on-target engagement, interactions with uncharacterized off-targets, or selectivity assays that do not accurately represent the in vivo environment. Without a specific mechanism of action, improving compound properties or defining therapeutic windows for future in vivo studies is substantially more challenging.

Understanding the drivers of undesired compound effects has been especially challenging for small molecule inhibitors of β-secretase BACE1 (β-site APP-cleaving enzyme 1). The amyloid cascade hypothesis of Alzheimer’s disease (AD) links disease pathology to an accumulation of cerebral amyloid beta (Aβ). BACE1 initiates the production of Aβ from amyloid precursor protein (APP), and therefore blocking the activity of this enzyme is considered one of the most promising approaches for treating AD. Considerable progress towards small molecule BACE1 inhibitors has been made and several inhibitors are in clinical trials, but to date none have received FDA approval. Safety liabilities have been a major cause of BACE1 inhibitor attrition, and in particular, both Eli Lilly & Co. first generation clinical candidate LY2811376 (1), and Amgen preclinical candidate AMG-8718 (2), were withdrawn from development after exhibiting ocular toxicity in preclinical animal models. Both inhibitors were found to cause an accumulation of autofluorescent material in the retinal pigment epithelium (RPE) and subsequent retinal photoreceptor degeneration, which would be expected to eventually result in severe visual impairment. The RPE is a non-regenerating layer of cells that has a number of important physical and biochemical functions crucial to the visual cycle, including the daily recycling of shed autofluorescent photoreceptor outer segments (POS). The accumulated RPE autofluorescent material observed with either LY2811376 or AMG-8718 treatment is consistent with impaired phagolysosomal POS degradation. Notably, LY2811376 also induced this effect in mice lacking the BACE1 gene, suggesting off-target effects.

BACE1 is a member of the pepsin aspartyl protease superfamily and blockade of enzyme function is typically achieved by active site-directed inhibitors that non-covalently engage the catalytic aspartate residues. Selectivity is therefore typically evaluated against a panel of purified aspartyl proteases (for example, BACE2, cathepsin D, cathepsin E, pepsin and renin) and for the most part, inferred across the broader proteome. With the exception of the closely related enzyme BACE2, selectivity against other aspartyl proteases was reported to be >60-fold for LY2811376, and >1,000-fold for AMG-8718. Selectivity against the endolysosomal aspartyl protease cathepsin D (CatD) was considered particularly significant as this enzyme had been annotated as a key component of the POS phagocytic pathway, and CatD deficiency in mammals or in humans had been shown to cause accumulations of autofluorescent material and visual impairment. Consequently, concerns were raised as to whether inhibition of BACE1—either alone or in combination off-target effects—might play a role in ocular toxicity. BACE1 knockout (KO) mice have been described as overtly phenotypically normal, although some studies report hypomyelination of peripheral nerves in neonates and delayed remyelination following peripheral nerve injury. One study of BACE1 knockout mice identified an ocular pathology, which in that study could also be induced in vivo in wild-type mice with BACE1 inhibitor IV (ref. ). Other studies that specifically examined the retinas of BACE1 KO mice or rats did not find any ocular irregularities. BACE2 KO mice have displayed coat colour defects, but have otherwise been described as phenotypically normal. Additional potential off-targets have remained largely unexplored including several components of the POS phagocytic pathway or proteins with genetic associations to accumulation of autofluorescent material. After 2010 when LY2811376 was withdrawn from clinical development, patients in some BACE1 clinical trials have had to undergo regular ophthalmologic exams.

Here we use quantitative chemoproteomics to perform the first target agnostic search for the mechanism of BACE1 inhibitor ocular toxicity. We identify CatD as a principal off-target of BACE1 inhibitors in a human RPE cell line and demonstrate that several BACE1 inhibitors show substantially enhanced potency for CatD in live cells compared with cell-free assays utilizing purified proteins. We combine these cellular target engagement measurements with exploratory toxicology studies and exposure-response analyses to designate off-target inhibition of CatD as a principal driver of ocular toxicity for BACE1 inhibitors.