Targeting glutathione peroxidases : identifying environmental modulators, and screening for novel small molecule inhibitors

Abstract

Glutathione peroxidases (GPXs) are a family of selenoproteins that are critical regulators of reactive oxygen species (ROS) in the cell, specifically hydroperoxides like hydrogen peroxide (H2O2). ROS are important for normal cell signaling and are tightly controlled to promote cell growth, proliferation, and survival. Without the antioxidant activity of enzymes like GPXs, however, the oxidative burden in cells can reach a point that leads to DNA damage, carcinogenesis, and eventually cell death. Although this can be catastrophic in early development, as evidenced by knock-out studies, many cancer therapeutics function through manipulating redox balance, suggesting that targeting these enzymes could have therapeutic potential. The modified metabolism of cancer cells can result in increased hydroperoxide production, and GPXs are often overexpressed to compensate for the increased oxidative stress, as well as in cell lines resistant to chemotherapeutics. The studies comprising this thesis examine several aspects of GPX inhibition to better understand how to utilize GPX-targeting agents as potential anti-cancer therapeutics and to identify novel inhibitors for future development.

Paper I addresses the effects of environmental heavy metal exposures on the erythrocyte GPX activity in 9-year-old children from the Matlab region of Bangladesh. Samples from 100 children were initially analyzed for concentrations of selenium, arsenic, mercury, and cadmium, as well as C-reactive protein (CRP) using inductively coupled plasma mass spectrometry (ICPMS). GPX1 expression levels in lysed samples were measured using high-throughput immunoblotting, and total GPX activity was measured using a GPX activity assay in lysates. After finding only a slight positive correlation between GPX1 expression and GPX activity, trace elements and CRP levels were considered, and a multivariable-adjusted linear regression analyses was used to assess predictors of GPX activity. Arsenic and CRP levels were significantly negatively associated with active GPX, while not correlated with each other. These results suggest that independently both arsenic exposure and increased CRP levels due to inflammation can suppress GPX activity in erythrocytes.

Paper II characterized the off-target inhibition of selenoprotein thioredoxin reductases (TXNRDs) by the ferroptosis inducers, (1S, 3R)RSL3 and ML162. Identified originally in a synthetic lethal screen for compounds specifically cytotoxic to oncogenic RAS, these compounds were then found to induce an iron-dependent cell death via increased lipid peroxidation associated with GPX4-specific inhibition. However, neither compound showed inhibitory activity in biochemical assays using recombinant GPX4, but both show potent inhibition of TXNRD1 in both biochemical and cellular assays. In three cell lines with varying susceptibility to ferroptosis, the cell death induced by RSL3 differed from the cell death caused by more specific TXNRD1 inhibitors, TRi-1 and TRi-2. Specifically, while RSL3 cytotoxicity could be rescued by co-treatment with the ferroptosis suppressor, Fer-1, the cytotoxicity of the Tri compounds was not rescued. Additionally, selenium supplementation diminished the efficacy of RSL3 while the TRi compounds remained unchanged, or slightly more cytotoxic. In all, these studies indicate that the interconnectedness of TXNRD1 and ferroptosis, and furthermore TXNRD1 and GPX4, is complex, but that this important off-target effect needs to be understood to fully characterize the use of these ferroptosis inducers.

Paper III established a discovery pipeline for the identification of novel specific inhibitors of GPX1 and GPX4. GR-coupled activity assays using recombinant GPX1 and GPX4 were optimized and miniaturized to 1536-well formats and screened against 12,000 small molecules with annotated mechanisms of action. A suite of confirmational assays were used to ensure specificity: a GR counter-assay was used to identify false-positives in the primary screens; orthogonal endpoint GPX assays were used to confirm inhibitory activity; GPX2 assay was used to further probe specificity of the confirmed active compounds between isoforms; a TXNRD1 assay was used to differentiate small molecules with broad Sec-targeting activity from GPXspecific inhibition; and nano Differential Scanning Fluorimetry (DSF) was used to confirm direct binding. Interestingly, all GPX1 inhibitors identified showed crossinhibition of GPX2. Ultimately, five novel GPX1/GPX2 inhibitors, 13 GPX4 inhibitors, and 2 novel pan-GPX inhibitors. This series of assays and the resulting compounds identified provide a basis for future development of GPX-specific inhibitors.

Paper IV profiled the cytotoxicity of a library of >10,000 small molecules with annotated mechanisms of action (MOA) and >100,000 small molecule scaffolds in a diversity library in both normal and cancer cell lines. These screens revealed a low overall cytotoxicity rate of the diversity library. Importantly, cytotoxicity was assessed in four normal cell lines (HEK293, immortalized human embryonic kidney cells; NIH3T3, an embryonic mouse fibroblast cells; HaCat, immortalized human keratinocytes; and CRL-7250, a primary human foreskin fibroblast cells). The top enriched MOA categories showing broad cell killing in normal cells were proteosome inhibitors, heat shock protein 90 (HSP90) inhibitors, anaplastic lymphoma kinase (ALK) inhibitors, mammalian target of rapamycin (mTOR) inhibitors, and cyclindependent kinase (CDK) inhibitors. Cytotoxic compounds with specific activity against the human adenosarcoma cancer line, KB 3-1, that showed no activity in the normal cell lines were also highlighted. This work will be used as an additional triage step in lead-selection for chemotherapeutic development, removing compounds that show significant cytotoxicity in normal cell lines.