Molecular Glue Is Pharma’s Play for the Undruggable

For decades, most diseases were untouchable. Novartis is now betting billions on a way to hijack the cell’s own machinery to destroy them from within. It’s a profound change in how we design drugs.

Most human proteins can’t be drugged. That’s been the quiet, frustrating truth at the heart of medicine for a century. The vast majority of proteins that cause cancer, dementia, and autoimmune disorders lack a neat little pocket for a conventional drug to plug into. They are considered structurally chaotic, functionally untouchable. But Big Pharma is clumsy, not stupid. Novartis isn't the first to work on so-called molecular glues, but its recent strategy pivot signals a fundamental shift from lab curiosity to industrial-scale drug development. The idea is simple: if you can’t disable a target, mark it for deletion. It’s less like putting a lock on a door and more like telling the building’s demolition crew exactly where to aim the wrecking ball.

The wrecking ball, in this case, is the cell's own ubiquitin-proteasome system — its internal garbage disposal. A molecular glue is a small-molecule drug that doesn’t block the target protein itself. Instead, it acts as a matchmaker, inducing proximity between the problem protein and an E3 ligase, a cellular enzyme whose job is to tag garbage for removal. Once the glue brings them together, the ligase slaps a ubiquitin tag on the target protein. The proteasome, the cell's mulcher, recognizes the tag and shreds the offending protein into amino acids. The mechanism isn't new; the infamous birth defects of thalidomide were the result of it acting as an accidental, and devastating, molecular glue. The challenge now is designing them on purpose without causing catastrophic off-target effects, a multi-billion-dollar game of chemical twenty questions.

Novartis doubling down isn't just an R&D line item; it's a shot across the bow at the entire field. The real money isn't just in a single blockbuster drug, but in owning a validated platform that can generate drug candidates for scores of previously untreatable targets. This pits established giants against specialist biotechs like Arvinas and C4 Therapeutics, who pioneered related protein degradation technologies and now must compete with a rival that has near-infinite capital. A successful platform opens up an estimated 85% of the human proteome as a new frontier for intellectual property. The losers could be any company still building its pipeline around traditional inhibitors. The FDA, for its part, is navigating uncharted territory, forced to weigh the immense potential against the ghost of thalidomide and the unknown risks of deliberately hijacking a cell’s most basic functions.

Within two years, we’ll see a wave of molecular glues enter Phase 1 and 2 clinical trials, primarily in oncology and immunology. The initial readouts won't be about efficacy; they'll be about safety and specificity. Can these agents stick to just one target, or do they get messy? As the biology gets sorted, the next five years will be about scaling the discovery process with high-throughput screening and machine learning models to hunt for new glue candidates. The era of a chemist painstakingly designing one molecule at a time is over. We’re on the verge of being able to systematically target and degrade nearly any protein in the human body. The question is no longer just how we do it, but who gets to decide which parts of ourselves are designated for disposal?