Projects

An important reason for the limited success of genuine redox-targeting drugs in the clinic is the lack of a biomarker concept that helps to identify the cancer patients most likely to respond to ROS-inducing drugs. Using proteomics, transcriptomics, and drug sensitivity data from a large panel of non-small cell adenocarcinoma cell lines we identified a set of 15 biomarkers which accurately predict drug efficacy, both of our TXNRD1 inhibitors and related compounds, e.g. ferroptosis inducers [Samarin et al. 2023]. The gene set, which we call anti-oxidant capacity biomarkers (ACBs) and which constitutes major pathways of redox regulation and defense against oxidative stress, is tightly repressed in sensitive cells. Contrary to expectation, constitutively low ACB expression is not associated with an increased steady state level of reactive oxygen species (ROS) but a high level of nitric oxide, which is required to sustain high replication rates. ACBs are favorably expressed only in a small subset of patients in most cancer entities, with the highest percentage of patients found in leukemias and SCLC. ACBs are expressed at high levels in resistant cells due to mutations in the negative regulator of NRF2, KEAP1. Surprisingly, pharmacological inhibition of KEAP1, resulted neither in increased resistance of cancer cells to chemotherapeutics, nor to ferroptosis- or redox targeting drugs. In contrast, cells derived from normal tissues respond to pharmacological NRF2 induction by upregulating ROS scavenging enzymes and resist drug-induced apoptosis longer. This observation is fully reflected in mouse models for therapy-mediated organ damage, which show that pharmacological induction of NRF2 mitigates dose-limiting side effects. We are currently investigating whether differential regulation of ACBs in cancerous versus healthy tissue can be utilized to increase the therapeutic window of redox targeting drugs.

Histone deacetylases are a group of hydrolase enzymes which are responsible for the removal of acyl groups primarily from lysine residues on proteins, but also from other acylated amine metabolites. HDAC10, for example, is known to be a poor lysine deacetylase but an efficient polyamine (e.g. spermidine) deacetylase. Furthermore, a number of studies have highlighted HDAC10 as a potential cancer drug target: (1) High HDAC10 expression levels were found to correlate with poor clinical outcome for advanced stage 4 neuroblastoma patients who received chemotherapy. (2) Consistent with these findings, HDAC10 depletion in neuroblastoma cells interrupts autophagic flux and sensitizes cells for chemotherapy, and enforced HDAC10 expression protects neuroblastoma cells against doxorubicin treatment.

We have published the first systematic investigations into the development of HDAC10 inhibitors and have found that a number of putatively selective HDAC6 inhibitors are also very good HDAC10 inhibitors. Indeed, we found that tubastatin A, one of the best-known HDAC6 inhibitors, is an even better HDAC10 binder. More recently, we published the first fully characterized highly selective HDAC10 inhibitor chemical probes. Currently, our work is focused on the synthesis of HDAC10 PROTACs and the use of these compounds to better understand HDAC10 biology and its potential as a therapeutic target in cancer.

Kallikrein-related peptidases (KLKs) are a family of 15 secreted serine proteases that have been shown to play a role in a variety of pathological conditions. Largely investigated as biomarkers, it is now becoming increasingly clear that KLKs also have a direct role in disease progression. KLK6, for example, has been shown to promote migration and invasion of tumor cells in melanoma and colon cancers, but few reports of KLK6 inhibitors have been made in the literature.

We have recently described a novel series of depsipeptides which are covalent inhibitors of KLK6. After a thorough investigation of the SAR and MOA of these substances, we converted them into activity probes by the introduction of an alkyne handle. These probes could then be used to pull-down endogenous KLK6 from cell lines and enable a first estimation of enzymatic activity in biological samples.

Current work aims to improve these inhibitors/probes and develop a reliable diagnostic method to measure KLK6 activity in human samples.

In a repositioning project, we combined two FDA approved non-oncology drugs, aurothiomalate and disulfiram, and discovered the two react to form a new substance. Named DKFZ-608, this substance is a TXNRD1 inhibitor with potent cellular activity against selected tumor entities, including small cell lung cancer (SCLC) cells, regardless of their genetic makeup or resistance to chemotherapy. DKFZ-608 is a single-digit nanomolar TXNRD1 inhibitor that rapidly induces high ROS levels in cancer cells, leading to apoptotic cell death. We improved the solubility of DKFZ-608 without compromising selectivity or ligand efficiency [e.g. DKFZ-682, Morgen et al. 2021] and were able to demonstrate high in vivo efficacy in mouse models for SCLC. DKFZ-682 was used to develop the biomarker concept ACB, which allows to predict the efficacy of redox targeting drugs in cancer cells. Currently, we lead a joint beLAB2122 Bridge drug discovery project with Evotec and BMS, aiming to discover non-metal containing TXNRD1-inhibitors.

Methionine aminopeptidase 2 (MetAP2) is one of two redundant enzymes that cleaves the starter methionine from proteins as part of their maturation process. MetAP2 is overexpressed in many cancers, and selective inhibition of MetAP2 results in reduced vascularization and growth of tumors. Furthermore, MetAP2 has been shown to play a role in fat production and is investigated as an anti-obesity target.

Many covalent MetAP2 inhibitors are semi-synthetic substances that are derived from the natural product fumagillin. Aiming to discover novel, drug-like, covalent MetAP2 inhibitors, we have developed a synthetic route to access novel fumagillin-inspired compounds which are potent, selective, and cellularly active MetAP2 inhibitors.

We currently work on a four different PROTAC projects, including HDAC10 PROTACs, but these projects have not yet been made public.