Small molecule drug discovery remains the backbone of modern therapeutics, accounting for the majority of approved medicines despite rapid advances in biologics, cell therapies, and nucleic-acid–based modalities. Their enduring relevance stems from intrinsic advantages: oral bioavailability, chemical tunability, intracellular target access, scalable manufacturing, and comparatively lower production costs. From early successes such as aspirin to modern precision oncology agents, small molecules continue to shape how diseases are treated across oncology, infectious disease, cardiometabolic disorders, inflammation, and central nervous system (CNS) indications.

At the same time, small molecule drug discovery is an inherently high-risk, high-attrition endeavor. Of the thousands of compounds synthesized and screened, only a handful advance to clinical development, and fewer still achieve regulatory approval. This reality has driven the field to adopt increasingly sophisticated strategies—integrating medicinal chemistry, computational drug design, high-throughput screening, ADMET prediction, and artificial intelligence—to improve efficiency and decision-making earlier in the pipeline.

This review provides a comprehensive, end-to-end perspective on small molecule drug discovery, structured as a logical progression from foundational principles to advanced methodologies and future directions. Rather than treating drug-likeness as a standalone filter, we emphasize how physicochemical properties, ADMET behavior, and molecular design principles are continuously applied throughout the discovery workflow, shaping success or failure at every stage.

Figure 1 | Conceptual overview of small molecule drug discovery.
Small molecule drug discovery integrates disease biology, medicinal chemistry, and pharmacology to translate biological hypotheses into clinically viable therapeutics. The process spans target identification and validation, compound screening, iterative chemical optimization, and preclinical development, with drug-likeness and ADMET considerations guiding decision-making at every stage. Advances in computational chemistry, structural biology, and artificial intelligence increasingly augment traditional experimental approaches, improving efficiency and reducing attrition.

Small molecule drugs are typically defined as low molecular weight organic compounds (commonly <900 Da) capable of modulating biological targets through reversible or irreversible binding interactions. Most function by interacting with proteins—enzymes, receptors, ion channels, or transporters—although nucleic acids and lipids may also serve as targets.

Several features distinguish small molecules from larger biologic modalities:

• Chemical diversity: Small molecules occupy a vast chemical space, enabling fine-tuned optimization of potency, selectivity, and pharmacokinetics.
• Cell permeability: Their size and physicochemical properties often allow access to intracellular targets inaccessible to antibodies.
• Manufacturing scalability: Synthetic chemistry enables robust, cost-effective, and reproducible production.
• Formulation flexibility: Oral, transdermal, inhaled, and injectable formulations are feasible.

Historically, small molecules have driven transformative advances in medicine, from antibiotics and anti-inflammatory drugs to kinase inhibitors and antiviral agents. Even as biologics expand therapeutic options, small molecules remain indispensable—particularly in chronic diseases, global health, and precision oncology.

Figure 2 | The multi-stage small molecule drug discovery pipeline.
The discovery process proceeds through target identification and validation, hit identification, hit-to-lead optimization, lead optimization, and preclinical development. Although depicted sequentially, the pipeline is highly iterative, with feedback loops between chemistry, biology, and ADMET profiling. Early-stage decisions—particularly target choice and physicochemical property control—strongly influence downstream success and clinical translatability.

Small molecule drug discovery is best understood as a multi-stage, iterative process, not a linear handoff. Decisions made at early stages—especially around target selection and chemical properties—reverberate throughout development.

At a high level, the pipeline comprises:

1. Target identification and validation
2. Hit identification
3. Hit-to-lead optimization
4. Lead optimization
5. Preclinical development

Each stage integrates considerations of drug-likeness, ADMET, pharmacokinetics (PK), and pharmacodynamics (PD), progressively increasing confidence that a candidate can become a safe and effective medicine.

Target identification begins with a deep understanding of disease biology. The goal is to identify a molecular entity whose modulation produces a therapeutically meaningful effect with an acceptable safety margin.

Modern target discovery increasingly relies on systems-level data, including:

• Genomics and human genetics (e.g., disease-associated variants)
• Transcriptomics and proteomics
• Pathway and network analysis
• Clinical phenotype correlations

Targets supported by strong human genetic evidence tend to show higher clinical success rates, underscoring the value of early biological validation.

Validation seeks to establish a causal relationship between target modulation and disease outcome. Common approaches include:

• Genetic perturbation (CRISPR, siRNA, knockout/knock-in models)
• Chemical probes with known selectivity profiles
• Phenotypic rescue experiments
• Biomarker modulation in disease-relevant models

From a small molecule perspective, target druggability is assessed early. Considerations include binding site accessibility, structural information availability, and feasibility of achieving selectivity with drug-like molecules.

Figure 3 | Experimental and computational strategies for hit identification.
Multiple complementary approaches are used to identify initial chemical matter. High-throughput screening (HTS) enables rapid experimental evaluation of large compound libraries, while virtual screening prioritizes molecules using structure- or ligand-based computational methods. Fragment-based drug discovery (FBDD) identifies low-molecular-weight binders with high ligand efficiency, which are subsequently elaborated into more potent compounds. Phenotypic and natural product screening further expand accessible chemical space and biological mechanisms.

Hit identification aims to discover initial chemical matter that modulates the target with measurable activity. Hits are typically weak and unoptimized but provide a starting point for medicinal chemistry.

HTS involves screening large compound libraries—often hundreds of thousands to millions of molecules—against biochemical or cell-based assays. Its strengths include:

• Broad chemical coverage
• Direct experimental readouts
• Compatibility with diverse target classes

However, HTS hits often suffer from poor drug-likeness, assay interference, or unfavorable ADMET profiles, necessitating careful triage.

Figure 4. DNA-encoded library–based hit identification in small molecule drug discovery.
DNA-encoded libraries (DELs) consist of vast collections of small molecules individually tagged with unique DNA barcodes. Libraries are screened in pooled format against a protein target using affinity selection, followed by DNA amplification and sequencing to identify enriched binders. DEL technology enables efficient exploration of large chemical spaces but requires off-DNA resynthesis and validation to confirm binding, functional activity, and drug-like properties.

DNA-encoded libraries (DELs) have emerged as a powerful and complementary approach for hit identification in small molecule drug discovery, particularly when screening very large chemical spaces against purified protein targets. DEL technology enables the synthesis, pooling, and screening of millions to billions of small molecules in a single experiment, far exceeding the scale of traditional high-throughput screening.

In a DEL, each small molecule is covalently linked to a unique DNA barcode that records its synthetic history. Libraries are typically constructed using templated or split-and-pool combinatorial synthesis, where successive chemical building blocks are appended to a growing scaffold, with each step encoded by an additional DNA sequence. The result is a highly diverse collection of compounds, each physically attached to a DNA tag that serves as a molecular identifier.

Hit identification proceeds through affinity-based selection rather than functional screening. The pooled library is incubated with an immobilized or tagged protein target, allowing binders to associate while non-binders are washed away. Bound compounds are then recovered, and the DNA barcodes are amplified and sequenced to identify enriched chemical structures.

DELs offer several advantages in early-stage discovery:

• Unprecedented library size, enabling broad exploration of chemical space
• Low material consumption, as screening is performed at picomole to femtomole scale
• Rapid identification of binders through next-generation sequencing
• Efficient triaging of chemical series based on enrichment patterns

These features make DELs particularly attractive for targets that are difficult to address with conventional HTS, including proteins with shallow binding pockets or limited assay tractability.

Despite their power, DELs impose unique chemical and biological constraints that influence hit quality:

• DNA compatibility restricts reaction conditions, limiting accessible chemistries
• Binding-only readouts do not directly report functional activity
• False positives may arise from DNA–protein or linker-mediated interactions
• Resynthesis requirement, as hits must be synthesized off-DNA for validation

As a result, DEL-derived hits often require careful follow-up to confirm binding mode, potency, and functional relevance.

To mitigate downstream attrition, modern DEL platforms increasingly incorporate drug-likeness principles at the library design stage. This includes:

• Selecting building blocks with controlled molecular weight and lipophilicity
• Designing scaffolds compatible with Lipinski and beyond-Rule-of-Five space
• Limiting excessive polarity and rotatable bond count
• Post-selection filtering of enriched compounds using in silico ADMET models

DEL hits are typically viewed as starting points rather than lead candidates, entering the hit-to-lead phase where medicinal chemistry optimization, ADMET profiling, and structural validation are applied.

Virtual screening leverages structure-based docking, pharmacophore modeling, and ligand-based similarity searches to prioritize compounds before synthesis or testing. When paired with experimental validation, computational approaches can significantly reduce cost and cycle time.

FBDD screens low-molecular-weight fragments that bind weakly but efficiently to targets. Fragments offer:

• High ligand efficiency
• Efficient exploration of chemical space
• Strong structural insights via crystallography or NMR

Fragments are subsequently grown or linked into more potent, drug-like molecules.

Phenotypic screening focuses on functional outcomes rather than predefined targets, while natural products offer structurally complex scaffolds evolved for biological activity. Both approaches can uncover novel mechanisms but often require additional effort to optimize drug-likeness.

Hit-to-lead optimization transforms diverse hits into lead compounds with improved potency, selectivity, and basic ADMET properties.

This stage marks the first systematic application of drug-likeness principles, including:

• Molecular weight control
• Lipophilicity optimization
• Hydrogen bonding balance
• Early solubility and permeability assessment

Medicinal chemists iteratively modify chemical structures to understand how changes affect biological activity. Early SAR focuses on:

• Identifying pharmacophores
• Removing liabilities (reactive groups, PAINS)
• Improving ligand efficiency

Basic in vitro assays assess:

• Aqueous solubility
• Passive permeability (e.g., Caco-2, PAMPA)
• Metabolic stability in liver microsomes
• Early cytotoxicity

Compounds failing at this stage are deprioritized, reducing downstream attrition.

Figure 5 | Iterative optimization of potency, pharmacokinetics, and safety during lead optimization.
Lead optimization involves repeated cycles of compound design, synthesis, and testing to balance potency, selectivity, ADMET properties, and pharmacokinetic–pharmacodynamic (PK/PD) relationships. Improvements in one parameter (e.g., potency or permeability) frequently introduce liabilities in others (e.g., solubility or clearance), necessitating holistic optimization strategies. Successful leads demonstrate robust efficacy, acceptable exposure, and a favorable safety margin suitable for preclinical advancement.

Lead optimization is the most resource-intensive stage of small molecule drug discovery. Here, compounds are refined to achieve a delicate balance between efficacy, exposure, and safety.

Potency improvements must be matched with increasing selectivity to avoid off-target effects. Structure-based design, supported by crystallography or cryo-EM, enables precise tuning of binding interactions.

• Solubility and permeability are optimized in tandem, recognizing their inherent trade-offs.
• Metabolic stability is improved by reducing CYP-mediated clearance and avoiding toxic metabolites.
• Distribution considerations include plasma protein binding and tissue penetration.
• Toxicity screening expands to ion channels, off-target receptors, and genotoxicity assays.

Pharmacokinetic and pharmacodynamic data are integrated to define exposure–response relationships. Successful leads demonstrate:

• Adequate bioavailability
• Predictable dose–response
• Therapeutic windows supporting clinical dosing

Preclinical development prepares a candidate for first-in-human studies. Activities include:

• In vivo efficacy studies in disease models
• Repeat-dose toxicology in multiple species
• Safety pharmacology (cardiovascular, respiratory, CNS)
• Formulation development
• IND-enabling studies and regulatory documentation

Only a small fraction of optimized leads reaches this stage, reflecting the cumulative attrition of small molecule drug discovery.

Lipinski’s Rule of Five emerged from empirical analysis of orally active drugs and reflects fundamental biological constraints:

• Molecular weight influences absorption and diffusion.
• Hydrogen bonding capacity affects membrane permeability.
• Lipophilicity balances solubility with membrane partitioning.

These parameters collectively approximate the requirements for passive intestinal absorption, making Ro5 a useful early heuristic rather than a rigid rule.

Figure 6 | Classical and expanded chemical space in small molecule drug discovery.
Lipinski’s Rule of Five defines physicochemical boundaries commonly associated with oral bioavailability, emphasizing molecular weight, lipophilicity, and hydrogen bonding capacity. Beyond Rule of Five (bRo5) compounds extend into higher molecular weight and polarity while maintaining permeability through conformational control, intramolecular hydrogen bonding, and increased three-dimensionality. These strategies enable targeting of challenging biological interfaces while preserving acceptable pharmacokinetic behavior.

Modern drug discovery increasingly targets challenging biological interfaces, necessitating exploration beyond classical Ro5 space. bRo5 strategies rely on:

• Conformational control and 3D shape
• Intramolecular hydrogen bonding to mask polarity
• Macrocyclization and rigidity to enhance permeability

These principles expand accessible chemical space while maintaining acceptable ADMET behavior.

Figure 7 | Integration of artificial intelligence and machine learning in small molecule drug discovery.
AI and machine learning approaches are applied throughout the discovery pipeline, from target identification using omics data to virtual screening, de novo molecular design, ADMET prediction, and retrosynthetic planning. By leveraging large chemical and biological datasets, these methods accelerate hypothesis generation, compound prioritization, and decision-making, complementing experimental and expert-driven medicinal chemistry workflows.

AI and ML models are increasingly applied to:

• Target identification from omics data
• De novo molecule generation
• ADMET and toxicity prediction
• Retrosynthetic planning

While not replacing medicinal chemists, AI augments human decision-making by accelerating hypothesis generation and prioritization.

High-resolution structural data enables rational design, while molecular dynamics simulations provide insight into binding kinetics and conformational flexibility.

Small molecule drug discovery faces persistent challenges:

• Rising R&D costs and long timelines (often 10–15 years)
• High attrition rates
• Difficult targets such as protein–protein interactions

Emerging solutions include targeted protein degradation, molecular glues, drug repurposing, and personalized medicine approaches.

Small molecule drug discovery is a multidimensional optimization problem, requiring the seamless integration of biology, chemistry, and data science. Drug-likeness and ADMET are not static filters but dynamic design principles applied throughout the discovery pipeline. As technologies evolve and chemical space expands, small molecules will continue to play a central role in translating biological insight into effective therapies.

1. 1. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Experimental and Computational Approaches to Estimate Solubility and Permeability in Drug Discovery and Development Settings. Advanced Drug Delivery Reviews 1997, 23, 3–25.
   2. Waring, M. J.; Arrowsmith, J.; Leach, A. R.; Leeson, P. D.; Mandrell, S.; Owen, R. M.; Pairaudeau, G.; Pennie, W. D.; Pickett, S. D.; Wang, J.; Wallace, O.; Weir, A. An Analysis of the Attrition of Drug Candidates from Four Major Pharmaceutical Companies. Nature Reviews Drug Discovery 2015, 14, 475–486.
   3. Doak, B. C.; Over, B.; Giordanetto, F.; Kihlberg, J. Oral Druggable Space beyond the Rule of 5: Insights from Drugs and Clinical Candidates. Chemical Biology 2014, 21, 1115–1142.
   4. Paul, S. M.; Mytelka, D. S.; Dunwiddie, C. T.; Persinger, C. C.; Munos, B. H.; Lindborg, S. R.; Schacht, A. L. How to Improve R&D Productivity: The Pharmaceutical Industry’s Grand Challenge. Nature Reviews Drug Discovery 2010, 9, 203–214.
   5. Hughes, J. P.; Rees, S.; Kalindjian, S. B.; Philpott, K. Principles of Early Drug Discovery. British Journal of Pharmacology 2011, 162, 1239–1249.
   6. Walters, W. P.; Murcko, M. A. Prediction of “Drug-Likeness”. Advanced Drug Delivery Reviews 2002, 54, 255–271.
   7. Schneider, G. Automating Drug Discovery. Nature Reviews Drug Discovery 2018, 17, 97–113.

Do you have an inquiry?

Get In Touch

Related articles

Molecular glue is an approach that is rapidly reshaping the way drugs are developed. The success in targeting difficult-to-drug proteins has led to increasing interest in the use of molecular glues as therapeutics and research tools. Unlike conventional drugs that inhibit or activate protein functions directly, molecular glues operate through a unique mechanism — bringing proteins together to elicit therapeutic effects. The concept has opened new pathways for addressing previously “undruggable” targets, among others in the treatment of cancer, neurodegenerative disorders, and infectious diseases (Tomlinsson et al., 2025, Schreiber, 2024).

This class of small molecules is reshaping the principles of drug discovery by shifting the focus from traditional occupancy-driven modulation of protein activity to strategies that enhance or induce protein-protein interactions (PPIs). In a promising event-driven modality, these molecules can exploit the body’s natural protein degradation pathways to selectively degrade disease-causing proteins (targeted protein degradation, TPD), but driving interactions can also lead to functional changes in protein activity directly. 

Hence, the value of molecular glues lies in their ability to overcome the limitations of traditional drug development, offering the potential to target proteins that were once deemed impossible to target. 

The potential of such drugs is tremendous — it is estimated that only 10% of the coding genome is druggable by traditional small molecule inhibitors. The remaining 90% includes proteins involved in protein complexes or in DNA/RNA binding, which due to their function often have flat or protruding surfaces adapted to interact with other proteins or nucleic acids but not small molecules. Many proteins, such as transcription factors and scaffolding proteins tend to be intrinsically disordered, having regions that fold only in the presence of protein co-factors and are therefore normally unavailable for the binding of small molecules.

The term molecular glue refers to a type of small molecule that affect PPIs, either by inducing de novo protein-protein interactions, or by stabilizing already existing interactions, resulting in stable ternary complexes. 

Most molecular glue approaches address the interaction between a target protein and an E3 ubiquitin ligase – thereby inducing the selective degradation of the target protein via the proteasomal pathway. This targeted protein degradation (TPD) approach circumvents many limitations of traditional therapeutics and offers a powerful tool for addressing diseases, simply by removing disease-causing proteins. Molecular glues differ from other TPD modalities, such as PROTACs (Proteolysis Targeting Chimeras), by being monovalent and structurally simpler, yet equally potent in facilitating protein-protein interactions.

The ability to direct protein degradation has far-reaching implications for drug development, especially for targeting non-druggable proteins like transcription factors and scaffolding proteins that are traditionally difficult to engage with small molecule inhibitors.

The term molecular glue was first coined with the discovery of the ability of cyclosporin A and FK506 to induce the binding of cyclophilin and FKBP12, respectively, to the protein phosphatase calcineurin, thus being molecular glues (Liu et al., 1991). Notably, the discovery of the mechanism of action of cyclosporin A occurred 8 years after its approval by the FDA for its use in controlling transplant rejection. 

In general, molecular glues have been recognized through serendipitous discoveries, most notably with immunomodulatory drugs (IMiDs) such as thalidomide, lenalidomide, and pomalidomide. Initially developed for other therapeutic purposes, these compounds were later found to exert their effects by modulating E3 ligase complexes to degrade specific transcription factors like Ikaros and Aiolos (Krönke et al., 2014). This revelation triggered a paradigm shift in drug discovery, illuminating a new strategy for modulating protein functions indirectly by degrading them rather than inhibiting them.

As the potential of molecular glues was realized, efforts shifted toward identifying other drugs that could be repurposed to engage in this mechanism. For instance, the discovery that thalidomide analogs could selectively degrade specific proteins raised hopes for using molecular glues in diseases like multiple myeloma, where targeting aberrant transcriptional activity could significantly impact survival outcomes.

Over the past two decades, the role of molecular glue in drug discovery has expanded significantly, leading to a deeper understanding of the molecular underpinnings of protein degradation. With advances in proteomics, structural biology, DEL screening, and medicinal chemistry, the rational design and screening for molecular glues with enhanced specificity and therapeutic potential has become possible. One example of a successful screening for molecular glue is the use of a fluorescence polarization-based assay for discovering molecules enhancing the interaction of the oncogenic transcription factor, β-Catenin, and its cognate E3 ligase, SCFβ-TrCP, leading to degradation of β-Catenin in a cellular system (Simonetta et al., 2019), 

Key molecular glue properties include high specificity, stability, cell permeability, and the capacity to induce novel protein-protein interactions that do not naturally occur, or to strengthen naturally occurring interactions.

Molecular glues are in general characterized by their small molecular weight and monovalent binding nature. Unlike bifunctional molecules like PROTACs, which consist of two linked ligands targeting a protein and an E3 ligase, respectively, molecular glues bind a single site and induce a new protein interface that promotes the formation of a ternary complex.

The mode of action of molecular glues typically involves altering the surface properties of one protein (often an E3 ligase), thereby enhancing its affinity for a secondary protein (the target). 

Similarly, molecular glues can complement missense mutations, providing for example ‘‘molecular prosthetics’’ as novel modalities in medicine, such as those seen when the disease-causing W580S missense mutation in MALT1 can be ameliorated with the MALT1 binder MLT748. W580 stabilizes an intramolecular interaction, and MLT748 can bind in the pocket created in the W580S mutant and mediate the domain-domain interaction that is crucial for MALT1 activity (Quancard et al., 2019). 

Clinical examples like the use of lenalidomide and pomalidomide show how molecular glues work in the context of cancer. These drugs interact with the E3 ligase cereblon, facilitating the ubiquitination and subsequent degradation of the transcription factors IKZF1 and IKZF3. In this case, lenalidomide and pomalidomide mimic a PTM that marks proteins for cellular degradation (the C-terminal cyclic imide ‘‘degron,’’). This degradation is critical for the treatment of multiple myeloma, underscoring the therapeutic potential of molecular glues in oncology.

The mechanism of molecular glue-induced protein degradation operates through the ubiquitin-proteasome system (UPS), a crucial cellular pathway for maintaining protein homeostasis. Once a target protein is tagged with polyubiquitin by an E3 ligase, it is recognized and degraded by the 26S proteasome. Molecular glues exploit this pathway by reprogramming the substrate specificity of E3 ligases—turning them into tools for selective protein clearance.

This approach allows for a more complete and irreversible silencing of disease-associated proteins compared to reversible inhibition. Moreover, molecular glue-induced degradation has been shown to produce fewer off-target effects, and is associated with reduced drug resistance, making it a promising strategy in precision medicine. In some cases, this degradation may result in the activation of other cellular pathways, offering the potential for broader therapeutic benefits.

Cancer treatment has been one of the most prominent areas of molecular glue application. As mentioned above, thalidomide derivatives, for instance, have demonstrated clinical efficacy in multiple myeloma by targeting transcription factors essential for cancer cell survival. The degradation of IKZF1 and IKZF3 by lenalidomide and pomalidomide not only highlights the clinical relevance of molecular glue but also underscores its potential for targeting proteins traditionally considered undruggable (Lu et al., 2014).

The emerging generation of molecular glues aims to address a wider array of cancer types. Notably, the targeting of oncogenic proteins like BCL6 and STAT3, which are involved in the regulation of immune responses and tumor progression, is being actively explored. For example, studies on BCL6 suggest that molecular glues can cause it to multimerize, resulting in its degradation. (Slabicki, et al., (2020).

Beyond oncology, molecular glue has shown promise in the treatment of neurodegenerative diseases such as Alzheimer’s, Parkinson’s, and Huntington’s disease. 

For example, tafamidis is used to treat both senile and hereditary transthyretin (TTR) amyloidosis, which is caused by the deposition of transthyretin polymers in neuronal and cardiac tissues. It works by stabilizing TTR tetramers via PPI interface binding, thereby precluding monomerization, polymerization, amyloidosis, and subsequent disease. Tafamidis is an excellent example of how stabilizing one endogenous PPI state over another can treat pathology, a model with utility in other aggregation and fibril-related diseases.

In a broader sense, by promoting the degradation of pathogenic proteins like tau, alpha-synuclein, and huntingtin, molecular glues offer a novel therapeutic avenue that addresses the root cause of neurodegeneration rather than merely alleviating symptoms. 

Similarly, in antiviral research, molecular glues are being explored for their ability to target viral proteins or host factors essential for viral replication. Emerging studies suggest that reprogramming E3 ligases to degrade specific viral components could offer a new strategy for combating resistant strains of HIV, hepatitis viruses, and even coronaviruses (Zhong et al., 2024).

The integration of molecular glue into structure-guided drug design is revolutionizing medicinal chemistry. Advances in cryo-electron microscopy (cryo-EM), X-ray crystallography, and computational modeling are facilitating the rational design of glues that optimize the shape, charge, and hydrophobic interactions required for effective protein-protein binding.
Machine learning algorithms and AI-driven platforms are also playing a pivotal role in predicting glue-like activity and guiding compound optimization. These tools allow the prediction of novel interactions, the identification of binding hotspots, and the tailoring of molecular properties to enhance selectivity and pharmacokinetics. The result is a more efficient drug development pipeline and an expanding repertoire of molecular glues with high therapeutic potential.
Emerging trends also show the combination of molecular glue strategies with other cutting-edge technologies like CRISPR-Cas9 to create more precise and effective therapeutic interventions. The potential for precision editing of E3 ligases, combined with the power of molecular glues, could lead to more personalized treatments with minimized off-target effects.

Despite their immense promise, molecular glues face several challenges. One major limitation is the current lack of high-throughput screening platforms specifically tailored to detect novel interactions. Here, recent developments in screening technologies, such as DEL based screening for molecular glues can become of importance. Additionally, for TPD, increasing the knowledge on suitable E3 ligases with favourable tissue- or cell-type specific expression will be of utmost importance for expanding the molecular toolbox available for the development of novel therapies.

Emerging biotechnologies such as targeted protein stabilization and autophagy-based degraders (AUTACs and ATTECs) may also be integrated with molecular glue strategies to broaden their therapeutic scope.

Currently, there is a primary focus on the use of molecular glues for TPD, but it is well known, that protein–protein interactions are vital for the regulation of virtually all cellular processes, from signalling to DNA replication, and that aberrant PPIs are causing several diseases. The concept of protein degradation induced by novel protein-protein interactions has inspired a broader field of induced-proximity modulators, and recently several proximity-driven approaches have been reported, such as deubiquitinase-targeting chimeras, lysosome-targeting chimeras, phosphorylation-inducing small molecules, tricomplex inhibitors of kRAS, and more. Thus, a new era of targeting disease-causing mechanisms and proteins previously regarded as undruggable may very well arise as the methods for discovery of molecular glues and other proximity-driven pharmaceuticals are refined and more widely used.

When comparing molecular glues to traditional degraders like PROTACs, several distinct advantages and trade-offs emerge. 

PROTACs offer a modular design, with high flexibility in targeting diverse protein domains. Their bivalent nature also simplifies the discovery phase in many cases, albeit at the cost of increased synthetic complexity and larger molecular size. 

From a drug discovery perspective DELs are in general well suited for the discovery of heterobifunctional molecules (PROTAC like molecules) as the discovery of binders to POI and E3 ligase is well established. In addition, the DNA attachment site on the small molecule indicates a potential molecular position for placing the linker between binders.

Molecular glues are smaller, simpler, and often more stable, which improves cell permeability and oral bioavailability. They do not require linker optimization and, once discovered, provide a shorter route from discovery to pharmaceutically usable molecules. 

As mentioned above, the drawback is that the discovery of molecular glues is not straight-forward, although some strategies have been pursued, hereunder DEL screening approaches. 

In practice, both approaches are complementary. The integration of both strategies within the same therapeutic pipeline could ultimately maximize the efficiency and range of targeted protein degradation therapies.

References and further reading

• Békés, M., et al., (2022). PROTAC targeted protein degraders: The past is prologue. Nature Reviews Drug Discovery, 21, 181. https://doi.org/10.1038/s41573-021-00371-6
• Krönke, J., et al. (2014). Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science, 343, 30. https://doi.org/10.1126/science.1244851
• Lamb, Y.N., Deeks, E.D. (2019). Tafamidis: A Review in Transthyretin Amyloidosis with Polyneuropathy. Drugs 79, 863–874. https://doi.org/10.1007/s40265-019-01129-6
• Liu, J. ∙ Farmer, Jr., J.D. ∙ Lane, W.S. (1991) Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes Cell, 66, 807. https://doi.org/10.1016/0092-8674(91)90124-h
• Lu, G., et al. (2014). The myeloma drug lenalidomide promotes the cereblon-dependent destruction of Ikaros proteins. Science, 343, 305. https://doi.org/10.1126/science.1244917
• Mayor-Ruiz, C., et al. (2020). Rational discovery of molecular glue degraders via scalable chemical profiling. Nature Chemical Biology, 16, 1199. https://doi.org/10.1038/s41589-020-0594-x 
• Quancard, J. et al., (2019). An allosteric MALT1 inhibitor is a molecular corrector rescuing function in an immunodeficient patient. Nature Chemical Biology 15, 304. https://doi.org/10.1038/ s41589-018-0222-1
• Schreiber, S.L (2024). Molecular glues and bifunctional compounds: Therapeutic modalities based on induced proximity. Cell Chemical Biology 31, 1050. https://doi.org/10.1016/j.chembiol.2024.05.004
• Slabicki, M., et al (2020). Small-molecule-induced polymerization triggers degradation of BCL6. Nature 588, 164. https://doi.org/10.1038/s41586-020-2925-1
• Simonetta KR., et al (2019). Prospective discovery of small molecule enhancers of an E3 ligase-substrate interaction. Nature Communications 10, 1402. https://doi.org/10.1038/s41467-019-09358-9.
• Tomlinsson, A.C.A. et al., (2025). The “three body solution”: Structural insights into molecular glues. Current Opinion in Structural Biology 91, 103007. https://doi.org/10.1016/j.sbi.2025.103007
• Yoon H., et al. (2024). Induced protein degradation for therapeutics: past, present, and future J Clinical Investigation. 134: e175265. https://doi.org/10.1172/JCI175265
• Zhong G. et al., (2024) Targeted protein degradation: advances in drug discovery and clinical practice. Signal Transduct Target Therapy 9, 308. https://doi.org/10.1038/s41392-024-02004-x

Do you have an inquiry?

Get In Touch

Related articles