AI-Powered Condensate Therapeutics: Dewpoint Targets TDP-43 for ALS

TDP-43 pathology is present in more than 97% of all ALS cases. For decades, the protein was considered undruggable. In January 2026, the company announced it had selected a development candidate for a small molecule that directly corrects the aberrant condensates driving TDP-43 dysfunction. It was the first time anyone achieved full restoration of TDP-43 function in preclinical models.

Why condensates matter

Biomolecular condensates are membraneless organelles that form dynamically inside cells through a process called phase separation, the same physics that makes oil droplets form in water. These structures concentrate proteins and RNA into discrete compartments, enabling cells to organize biochemical reactions without a surrounding membrane.

Dysregulation of condensates has been implicated in a wide range of diseases: cancer, neurodegeneration, diabetes, cardiopulmonary disorders, and viral infections. The fundamental insight, that many diseases are driven not by a single mutated protein but by the dysfunctional behavior of protein-RNA assemblies, shifts the drug discovery model from targeting individual molecules to modulating whole biochemical compartments.

TDP-43 is a case in point. The protein normally shuttles between the nucleus and cytoplasm, regulating RNA splicing and transport. In ALS, a disease with no meaningful disease-modifying therapy for the majority of patients, TDP-43 forms pathological cytoplasmic aggregates that lose their normal nuclear function. More than 97% of ALS cases share this pathology, regardless of the underlying genetic mutation.

The AI platform behind condensate modulation

Its approach rests on an AI-powered integrated platform that combines multiomics data, high-content imaging, and machine learning to predict condensate targets and design small molecules that modulate them. The company calls these compounds c-mods (condensate-modulating drugs).

The platform works at a fundamentally different level than traditional drug discovery. Instead of looking for a molecule that fits a single protein pocket, its models predict how a compound will partition across the cell's condensate territory: which compartments it will concentrate in, which it will avoid, and what effect that redistribution will have on disease-relevant biology.

"By targeting the condensate, you can treat all of the patients regardless of the etiology of their disease," Isaac Klein, MD PhD, Chief Scientific Officer of Dewpoint, told Drug Discovery News. "This is in direct contrast to the more common strategy which is 'oh, they have a mutation, let's target that.' That's predominantly failed."

The TDP-43 development candidate, a first-in-class small molecule, was selected based on a thorough preclinical data package demonstrating restoration of TDP-43 splicing function in cellular systems, strong activity in in vivo models of neurodegeneration including neurofilament light chain (NfL) reductions, and pharmacologic properties supportive of IND-enabling studies. The molecule also showed applicability across TDP-43 proteinopathies including frontotemporal dementia and traumatic brain injury.

Second company, same thesis

Transition Bio, a Cambridge-based condensate drug discovery company, independently validated the same approach with a different technical stack. Its Condensomics platform uses droplet microfluidics, technology originating from Harvard's Weitz Lab, combined with machine learning to characterize condensates at a scale not seen before in the field.

In November 2025, Transition Bio entered a drug discovery collaboration with Voyager Therapeutics worth up to $500 million, targeting TDP-43 pathology in ALS and frontotemporal dementia. The deal structure, upfront payment plus milestones and royalties, mirrors the terms usually reserved for more mature therapeutic modalities.

"Transition Bio's molecular condensate technology uniquely identifies small molecules that aim to precisely correct the mislocalization of TDP-43 without abolishing its important functional activity," said Alfred W. Sandrock Jr., MD PhD, President and CEO of Voyager Therapeutics.

At AACR 2026, Transition Bio presented progress on its lead small-molecule inhibitor program targeting YTHDC1, a protein that regulates RNA methylation in MYC-driven cancers. The company demonstrated that its inhibitors dissolve YTHDC1 condensates in cancer cells at nanomolar potencies with exquisite selectivity, suppressing oncogenic gene expression.

Dewpoint has also selected a MYC condensate modulator development candidate, a first-in-class small molecule that disrupts MYC-driven oncogenic transcription by modulating aberrant condensates, in February 2026. The parallel advancement of TDP-43 and MYC programs suggests the platform may be generalizable across neurodegeneration and oncology.

What makes this a new modality

The pharmaceutical industry has historically worked with four therapeutic modalities: small molecules that bind protein pockets, biologics that target extracellular proteins, nucleic acid therapies, and gene therapies. Condensate modulation represents a potential fifth modality.

Unlike traditional small molecules that must fit a specific binding pocket, c-mods work by altering the physical state of a condensate, shifting it from a pathogenic solid-like state to a functional liquid-like state, or vice versa. This mechanism makes them effective against targets that lack well-defined binding sites, which includes most of the proteins implicated in neurodegeneration and many in oncology.

Its pipeline now spans more than 20 programs across oncology, neurology, and cardiometabolic disease, with strategic collaborations with Bayer and Novo Nordisk. The company closed a Series D in September 2025 extending its runway into early 2027, with the first c-mod (DPTX3186, targeting β-catenin in gastric cancer) already in clinical development.

What happens next: IND-enabling studies and clinical proof-of-concept

It plans to initiate IND-enabling studies for the TDP-43 program while continuing to engage with the ALS community on clinical trial design. If the molecule enters the clinic and demonstrates disease modification, a high bar given ALS's history of failed trials, it would validate condensate modulation as a therapeutic modality rather than an academic curiosity.

Transition Bio's timeline follows a similar arc. Its collaboration with Voyager aims to nominate a development candidate before Voyager can exercise its option for worldwide exclusive rights. The $500 million deal structure incentivizes speed.

The critical question is whether preclinical biomarker signals — NfL reductions, splicing restoration — will translate to clinical benefit. ALS has destroyed dozens of drug candidates that looked promising in animal models. But condensate modulation attacks the pathology at a more fundamental level than any prior approach: instead of clearing aggregates or inhibiting a single enzyme, it restores the normal physical state of the TDP-43 protein inside the cell.

What happens to ALS drug development if condensate modulation works

Development scenarios

Sources