Until recently, the discovery of novel targets for targeted protein degradation (TPD) has been largely serendipitous, and the lack of target knowledge was a major bottleneck in the development of small molecule degraders. New technological advances in proteomics and mass spectrometry (MS) show promise in removing this critical hurdle by systematically connecting degrader compounds and their targets, which were previously undruggable. TPD is now positioned as one of the most promising approaches to develop alternative treatments for life-threatening diseases.
Introduction
Over the past two decades, targeted protein degradation (TPD) has emerged as a therapeutic modality with the potential to tackle disease-causing proteins that have historically been highly challenging to target with conventional small molecules. TPD utilizes the waste-disposal system of the cell to selectively eliminate disease-causing proteins. Advancements in instrumentation and software have led the pharmaceutical industry to look more closely at TPD for new drug discovery possibilities because of its ability to eliminate — rather than inhibit — disease-causing proteins. Interest in the field was bolstered recently when preliminary clinical data for bavdegalutamide (ARV-110), a proteolysis-targeting chimera that flags the androgen receptor for degradation, indicated that the drug is safe and shows some efficacy in men with metastatic castration-resistant prostate cancer.1
The capabilities of TPD are centered around small molecules commonly referred to as degraders. However, although protein degraders could become blockbuster drugs in many therapeutic areas, their discovery has been largely by chance.2 Technological advances in proteomics now show capabilities to screen extensive small molecule libraries for degraders of pathogenic proteins previously considered undruggable.
Principles of Degradation-Inducing Molecules
Unlike traditional small molecule drugs that only inhibit their targets, degradation-inducing molecules deplete their targets and do not require active binding sites to exert their effect. That means entities like molecular glues, heterobifunctional degraders known as proteolysis targeting chimeras (PROTACs), monovalent degraders, and deubiquitinase (DUB) inhibitors can potentially target the large part of the human proteome that lacks active binding sites and is thus inaccessible to small molecule inhibitors. The novel therapeutic mechanism makes protein degraders promising therapeutic agents for indications of high medical need unmet by conventional medicines.3
Many degrader drugs redirect E3 ubiquitin ligases to non-native target proteins called neosubstrates. After forming a ternary complex with the target and the degrader, the E3 ligase modifies these neosubstrates by attaching multiple ubiquitin molecules. The proteasome recognizes the resulting poly-ubiquitin chains and degrades the neosubstrates into amino acids that are recycled for novel protein synthesis. Unlike conventional small molecule drugs, degrader molecules don’t just inhibit their targets. Instead, they act as catalytic agents that eliminate disease-causing proteins and their associated functions.4
Established degrader drugs that reprogram E3 ligases come mainly in two varieties: heterobifunctional molecules also known as PROTACs (PROteolysis Targeting Chimeras) or molecular glue compounds.5 PROTACs consist of a recruitment ligand for the E3 ligase and a targeting ligand (“warhead”) that binds the target protein, connected by a linker. Molecular glue degraders are smaller in size and have more favorable drug-like properties. They reshape the substrate interaction site of E3 ligases to create complementary binding sites for non-physiological targets. Degraders with alternative mechanisms of action are also being developed. Monovalent degraders induce degradation by triggering conformational changes in their target proteins rather than by repurposing an E3 ligase. DUB inhibitors effectively block removal of poly-ubiquitin chains from target proteins, causing their proteasomal degradation (Figure 1). Finally, new TPD approaches go beyond the proteasome, exploiting autophagic or lysosomal degradation pathways to eliminate extracellular targets and protein aggregates.
Figure 1. Proteomics addresses the relevant questions in degrader drug discovery by identifying candidate degrader molecules that engage (new) E3 ligases or act through different TPD mechanisms; systematically exploring the true target scope of degrader compound libraries in intact cells; mechanistically validating candidates as bona fide TPD targets in the proteome-wide context; and comprehensively monitoring degrader compound selectivity in endogenous cellular environments.
Use of TPD for Drug Discovery
Discovery efforts in targeted protein degradation start either from degrader targets or from degrader chemistries. Both strategies benefit from high-throughput deep proteomic screening, which unbiasedly analyzes the interaction between small molecules and the proteome to guide drug development. The first case applies mostly to heterobifunctional molecules like PROTACs, which are rationally designed to possess binding moieties for harmful proteins. Monitoring their selectivity against all cellular proteins provides crucial information for optimization efforts. In general, proteome-wide selectivity analysis should be performed for any degrader molecule under development for a known target.
In the second case, drug discovery originates from small molecules presumed or designed to be degraders. Testing such compounds against entire proteomes enables systematic identification of their targets. For example, molecular glue compounds, like those that bind to the E3 ligase cereblon, have been clinically validated to eliminate disease-causing proteins. With low molecular weight, favorable drug-like properties, and the ability to promote target recognition in the absence of cavities or binding pockets, molecular glues can address proteins considered undruggable by conventional small molecule drugs. Their discovery, however, has so far relied on serendipitous observations that have led to only a few dozen potential neosubstrates. Thus, the true target scope of molecular glue degraders is largely unknown, despite their enormous therapeutic potential. Only a systematic, proteomics-based discovery approach that connects degrader compounds with targets in an unbiased manner can exploit the full targeting potential of molecular glues (Figure 2).
Figure 2. The agnostic proteomics platform screens and analyzes all types of degrader molecules in high throughput (e.g., PROTACs, molecular glues, monovalent degraders, DUB inhibitors), mechanistically validates potential targets, and monitors proteome-wide selectivity during compound optimization.
Scalable Proteomics Infrastructure
Putative degrader targets found in the proteomics screens can readily be validated using a unique proteomics-based validation pipeline (NEOsphere Biotechnologies, Planegg, Germany). Besides identifying neosubstrates at unprecedented scale, the resulting data can be used for degrader library optimization and expansion, as well as for computational approaches, such as AI-based predictive modeling of novel degrader drugs. The high throughput and fast turnaround capabilities of this platform also support proteome-wide selectivity profiling in drug optimization cycles.
This proteomics-based platform comprises four major components (Figure 3):
- Cell treatment and automated sample preparation: Automated, flexibly scalable workflows provide efficient and reproducible cell treatment, cell lysis, proteolytic cleavage, and peptide purification in 96-well format. Screening is routinely performed in three replicates per treatment condition to enable powerful statistical analysis. Optimized experimental design can filter out noise and identify significant downregulation of neosubstrate candidates.
- Single-shot liquid chromatography coupled with MS: The system provides outstanding performance with unparalleled detection and quantification of up to 11,000 protein groups from a single cell line sample in a single MS analysis. It uses data-independent acquisition (DIA) in combination with trapped ion–mobility spectrometry (TIMS) (timsTOF Pro, Bruker, Billerica, MA). In DIA modes, precursor ions are isolated into windows of a pre-defined m/z-ratio and fragmented; all fragmented ions in each window are then analyzed by a high-resolution mass spectrometer. DIA proteomic analysis provides broad protein coverage, high reproducibility, and accuracy. The TIMS device accumulates and concentrates ions of a given mass and mobility, offering a unique increase in sensitivity and speed along with the additional dimension of separation. Combined with multiparameter optimization in MS sample preparation, high-throughput screening of tens of thousands of samples is possible with the highest data quality and resolution.6
- Automated data analysis and identification of novel target candidates: The platform performs MS data processing using a customized version of DIA-NN in combination with an in-house data analysis pipeline for maximum precision, data completeness, and sensitivity. DIA-NN is a leading software for DIA proteomics data processing. It exploits deep neural networks and quantification and signal correction strategies for processing DIA proteomics experiments. It also improves the identification and quantification performance in conventional DIA proteomic applications and is particularly beneficial for high-throughput applications, as it is fast and enables deep and confident proteome coverage when used in combination with fast chromatographic methods.7
- Mechanistic validation of degrader targets: Potential new targets can be validated using high-end MS-based methods. Their dependency on the ubiquitin–proteasome system (UPS) can be verified with pharmacological tools and through genetic inactivation of the respective E3 ligase. Detection of degrader-induced E3 ligase binding is done via interactomics (affinity enrichment–MS and cellular proximity labeling). To demonstrate degrader-induced modification with ubiquitin, ubiquitinomics can be used to reliably quantify up to 50,000 ubiquitination sites in single MS runs.
Figure 3. A unique proteomics-based proteomics platform systematically connects degrader compounds with their targets to unlock and explore the undruggable proteome.
Comprehensive Detection of Candidate Degrader Targets upon Treatment with E3 Ligase Targeting Compounds
To demonstrate the capabilities of this technology, colon cancer cells were treated for the indicated times with E3 ligase–modulating drugs, and single-shot MS was performed as described. Proteome coverage of up to 11,000 protein groups per sample was achieved with a data completeness of more than 99% and a median protein coefficient of variation (CV) below 10%.
Statistical data analysis revealed statistically significant up- and downregulation events of cellular proteins.
Thalidomide derivatives, such as pomalidomide, mezigdomide, or lenalidomide, are known to induce ubiquitination and proteasomal degradation of zinc-finger proteins by recruiting them to cereblon. As expected, zinc-finger target proteins expressed in the analyzed cell line were downregulated, including ZNF692, ZFP91, and ZNF827. Other neosubstrates were also degraded, such as the lenalidomide-specific target casein kinase 1a (CSNK1A1).
The sulfonamide-type molecular glue indisulam promotes interaction between DCAF15 E3 ligase and the splicing factors RBM39 and RBM23, leading to their ubiquitination and proteasome-mediated degradation. Treatment with the VHL based PROTAC degrader ACBI18 for four hours downregulated its direct cellular targets SMARCA2 and SMARCA4 together with two interacting BAF complex members, while secondary regulation events are detected eight hours later.
Figure 4. The effect of compound treatment on cellular regulation. The volcano plots depict the fold change of proteins in compound treated vs. control cells (x-axis) and the standard as a measure of precision of quantification (y-axis). Statistically significant downregulation events of cellular proteins are shown in red, upregulation events in blue.
Conclusion
Recent technological advances in proteomics promise to eliminate bottlenecks in degrader development by systematically connecting degrader compounds and their targets. The underlying concepts and strategies for undruggable target space exploration include systematic comparisons across diverse cell lines and types to maximize the accessible proteome, selection of most responsive screening models, and scoring and reviewing metrics based on individual treatment and global data assessment to distinguish likely direct degradation from secondary regulation or off-target effects. Optimized screening of both biological and chemical diversity can identify cell type-specific and rare target candidate hits, as well as the sweet spots/regions in chemical space, creating rationales to expand chemistry where more target candidates are likely encountered.
All of these new capabilities of MS-based proteomics can reveal the true target scope of targeted protein degradation and help to advance degrader drugs, to broadly position TPD as an alternative to existing therapeutic modalities.
References
- Békés, M., D.R. Langley, and C.M. Crews. “PROTAC targeted protein degraders: the past is prologue.” Rev. Drug Discov. 21: 181–200 (2022).
- Wu, T., et al. “Targeted protein degradation as a powerful research tool in basic biology and drug target discovery.” Nature Structural & Molecular Biology. 27: 605–614 (2020).
- Liena, Q., D. Han, and W. Junfeng. “Key Considerations in Targeted Protein Degradation Drug Discovery and Developmen” Frontiers in Chemistry. 10 (2022).
- Radhakrishnan, S., O. Hoff, and M.K. Muellner. “Current Challenges in Small Molecule Proximity-Inducing Compound Development for Targeted Protein Degradation Using the Ubiquitin Proteasomal System.” Molecules. 27: 8119 (2022).
- Kozicka, Zuzanna and Nicolas Holger Thomä. “Haven’t got a glue: Protein surface variation for the design of molecular glue degraders.” Cell Chemical Biolog 28:1032-1047 (2021).
- Meier, F., A.D. Brunner, M. Frank, et al. “diaPASEF: parallel accumulation–serial fragmentation combined with data-independent acquisition.” Methods. 17: 1229–1236 (2020).
- Demichev, V., et al. “DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput.” Methods 17: 41–44 (2020).
- ACBI1 was kindly provided by Boehringer Ingelheim via its open innovation platform opnMe, available at https://opnme.com.
