Akin to shoppers, farmers and artisans congregating at the town park every Saturday for Farmer’s market, biomolecular condensation is a phenomenon where cellular components self-organize in time and space by coming together to form concentrated communities of molecules that synergize to perform a particular biological function. Disrupting the composition, structure and function of these communities can lead to disease. Is it possible to repair the broken condensates to treat or cure diseases?
The dawn of the condensate-targeted drug discovery era started with Dewpoint’s inception 5 years ago, built on the conviction of the scientific founders, Tony Hyman (MPI, Dresden, Germany), Phil Sharp (MIT) and Rick Young (MIT, Cambridge, MA, USA), that the biomolecular condensation phenomenon they had been studying for a little less than a decade might hold the key for addressing diseases of high unmet need. At the time, this was a challenging task, given that the basic science foundation for the field was still under construction. Fundamental questions were largely unanswered. For example, (i) what are the rules that govern condensate structure and function? (ii) how do condensates impact disease? (iii) how do drugs interact with condensates?
In the past 5 years, Dewpoint has blazed the trail to convert basic scientific findings and assays into a state-of-the art drug discovery and development pipeline, which we are applying towards therapies for diseases of high unmet need. Dewpoint’s efforts have been supported by significant progress in answering the above questions by the growing scientific community, including academia, biotech and pharma (Figure 1).
In late 2018 the field was just scratching the surface on understanding the rules that make a particular biomolecule localize to a specific condensate. Since then, the field has uncovered the various types of interactions that drive phase separation, which now include hydrophobic and p-stacking, electrostatic interactions within disordered regions, along with interactions involving folded domains. We now understand that the nature of these interactions controls the material properties, architecture, and function of condensates. These insights enabled the detailed characterization of the structural organization of biomolecular condensates, such as paraspeckles and the nucleolus. Furthermore, they opened new possibilities for intelligent design of synthetic condensates with tunable properties and specific functions, such as specialized orthogonal translation and transcription. The intimate understanding of the molecular drivers of condensation – the sequence “grammar”, and the expanding experimental evidence, led to the development of curated databases of condensate-associated proteins (e.g., Dewpoint-sponsored CD-CODE, DrLLPS, PhaSePro, LLPSDB, PhaSepDB). The progress in AI/ML, combined with the mounting experimental evidence, led to the development of multiple condensation predictors (e.g., PSPredictor, ParSe v2, ParSe, LLPhyScore).
Significant progress has also been made in decoding the contributions of RNA structural features to the structure and function of condensates. I eagerly anticipate the emergence of parallel databases and predictors of condensate-associated RNAs in the upcoming years. Mapping of protein and nucleic acid components associated with condensates in healthy vs. diseased cells will be a next major milestone in understanding disease biology and developing efficacious treatments.
Neurodegenerative disease has been the poster child for condensate dysregulation causing pathology. There is mounting evidence that condensate dysregulation, which we call a condensatopathy, contributes to amyotrophic lateral sclerosis (ALS) and other similar conditions. Leveraging condensate biology to discover new therapies for ALS is an area of active interest across the drug discovery industry.
Additional compelling evidence that links the formation of an aberrant condensate to in vivo disease pathology was published in 2020. The study showed that a genetic mutation known to cause dilated cardiomyopathy led to aberrant localization of the splicing factor RBM20 in cytoplasmic condensates in a humanized model of diseased pigs. Another example of a condensatopathy emerged in 2022: the formation of abnormal nuclear condensates by the fusion oncoprotein NUP98-HOXA9– the product of a genetic aberration found in children suffering from several forms of leukemia –that drives transformation of healthy into cancerous cells. Enabled by progress in machine learning methods and systematic curation of genetic characterization of clinical samples, we learned that over a thousand fusion oncoproteins identified in oncology patients have the sequence “grammar” conducive to condensate formation. Many oncoproteins – whether due to mutations or overexpression – drive oncogenesis by disrupting the normal transcriptional programming through formation of aberrant condensates. We now understand that condensates play even more profound roles in cancer and are involved in most of the processes that are hallmarks of cancer.
A mounting body of work from investigators around the globe characterized the involvement of condensates in the immune evasion tactics of bacteria and viruses. Significant progress has also been made in characterizing condensates involved at various steps in the viral life cycle of multiple virus families, including SARS-CoV2, the culprit of the COVID-19 pandemic. A noteworthy milestone built on this work is the identification of compounds with pan-SARS activity identified in a condensate phenotypic screen.
By diverging from the traditional view of lock-and-key drug-target interactions, we can take the vantage point of drug interactions with condensates. In the past five years we’ve learned that a condensate-modifying drug (c-mod) can wield an effect on a condensate through direct interaction with one or more of its components, through modifications of an enzyme that modifies the posttranslational or posttranscriptional modifications on one or more components, by complex modulation of multiple cellular components and processes, or by selectively being concentrated or excluded from a particular condensate. We now understand that c-mods cover a broad and drug-like chemical space and known FDA-approved drugs act through condensate-modifying mechanisms.
The field has made significant strides into characterizing condensate-specific drug mechanisms. Cyclopamine – an RSV antiviral, prevents viral replication by hardening the replication condensates. icFSP1 – a compound that induces ferroptosis as an anti-cancer approach, inhibits its FSP1 target by trapping it into drug-induced condensates. The way drugs interact with condensate components depends on the preference of the drug to accumulate inside or be excluded from the condensate – referred to as partitioning. We now understand that condensates, due to their unique composition, create distinct local chemical environments. These distinct properties contribute to selective partitioning of drugs into condensates. Leveraging the power of AI/ML we can start to understand the “chemical grammar” of drug partitioning. Applications of this newly acquired knowledge and of future progress in this arena will be instrumental in optimizing c-mods for safety and efficacy.
How has Dewpoint evolved to apply and advance the basic science of condensate biology, chemistry, & biophysics into industrialized technologies for discovering and developing new therapies for patients suffering from diseases of high unmet need?
The idea of leveraging biomolecular condensates for therapeutic development was in its infancy in 2018; little more than an exciting possibility. While the questions that could be asked abounded, most of the existing tools for characterizing condensates were developed for basic science, low-throughput laboratory applications. The first step in Dewpoint’s evolution was to develop our toolbox: establish a laboratory infrastructure capable of basic science and high-throughput drug discovery, with a wide enough breadth to cover reductionist in vitro systems to disease-relevant cellular systems.
With these tools in hand, we were able to focus on more important questions: what diseases can we tackle to achieve the most meaningful impact in the patients’ lives? With emerging evidence that condensates serve as central nodes of dysfunction in multiple classes of polygenic and infectious diseases, we established a diversified portfolio that includes neurological disorders, oncology, cardiovascular, metabolic and infectious diseases. The progress towards the clinic of this diversified portfolio has been made possible through work performed internally at Dewpoint and through partnerships with leaders in the pharmaceutical industry. 5 years later, our two lead fully owned programs – colon cancer and ALS – are successfully progressing toward the clinic. We continuously advance and optimize our platform and discovery pipeline to incorporate state-of-the-art technology. Currently, we are able to screen roughly half a million compounds per month in house, and our end-to-end discovery platform is propelled by ERSAI – our AI-powered digital engine, which accelerates the path of c-mods from bench to bedside.
Onward to the next 5 years!
Design: SALIENCE Communication / Publiepress
Scientific animation: Visual Science