01 · Who they are nowThe biotech that argues with the dictionary.
On any given Tuesday in South San Francisco, a few dozen scientists at Frontier Medicines are doing something the textbooks once said could not be done. They are designing small molecules that latch onto proteins long written off as "undruggable" - a word the industry has used for decades to politely describe failure.
Frontier dislikes the word. It treats it the way a locksmith treats a door labeled "do not enter."
The company is 77 people, give or take. It has raised more than $315 million across four rounds. Its lead drug candidate, FMC-376, is in a Phase 1/2 trial. Its partnerships include AbbVie - the kind of deal where the upfront alone is $55 million and the back-end is described, with a straight face, as "more than a billion."
"About 85% of the proteome has been considered out of reach. Our entire pitch is that the number is wrong."
02 · The problemA door labeled "do not enter."
Cancer is, very often, a protein doing the wrong thing. Find the protein, jam it with a molecule, and a tumor may stop growing. That is the entire premise of targeted therapy, and it is the reason the pharma industry exists in roughly its current shape.
The catch: most proteins do not have a convenient pocket for a small molecule to grab. They are smooth, or flexible, or disordered, or just plain shy. Drug hunters call these targets "undruggable," which sounds technical but is mostly a confession.
KRAS - the gene most famously mutated in lung, colorectal, and pancreatic cancers - was the canonical example. For forty years, KRAS humbled everyone who tried. Then, in the last decade, a few groups found a way in through a cryptic pocket that only appears in the G12C mutant form. Drugs reached the market. Patients improved. And then, often, they relapsed.
"Undruggable was never a law of physics. It was a bookmark."
This is the gap Frontier is built for. The first generation of KRAS G12C drugs only catch the protein in its inactive ("OFF") state. Tumors learn. Resistance arrives. Patients run out of options. The company's argument is that the next generation has to do more than what the first one did - and that the rest of the proteome is full of similar, quieter problems.
03 · The betA VC reads a paper. A chemist picks up the phone.
The origin story is unfashionably specific. Chris Varma, a longtime life sciences investor, was reading academic chemistry papers in 2018 - the kind of activity that does not typically make for a good origin story - when he ran into the work of Daniel Nomura's lab at UC Berkeley. Nomura, with his colleague Roberto Zoncu, was developing chemoproteomic techniques to find new ligandable sites all over the proteome.
Varma cold-called. The three co-founded Frontier Medicines.
What they bet
That two ideas, when combined, would unlock targets that neither could touch alone. First: covalent fragment screening, which uses tiny reactive molecules to bond chemically with a protein and reveal hidden pockets. Second: chemoproteomics powered by mass spectrometry, which can tell you - across the entire proteome at once - exactly where those fragments are landing.
Add machine learning to triage the firehose of data, and you have, in theory, a way to ask the proteome a question and get back a map of new drug targets.
"You don't go looking for new biology with old maps."
A short, eventful biography
04 · The productOne platform, several wagers.
What Frontier actually sells - eventually, if everything works - is a pipeline of small molecules. What it has built in the meantime is a platform, which is a less satisfying word for "an industrial-scale way of asking the proteome where to put a drug."
The Frontier Platform
A library of covalent fragments. A mass spectrometry pipeline that can read where those fragments stick across thousands of proteins at once. Structural biology to confirm the chemistry. Machine learning to prioritize the hits. The point is not novelty in any one of those tools; it is that they have been wired together end-to-end for a single purpose.
FMC-376
The lead candidate is a dual ON/OFF inhibitor of KRAS G12C. Translation: it tries to block the mutant protein in both of its conformational states, not just the resting one. The case for this is straightforward - resistance to first-generation KRAS drugs often arrives when the protein flips into the active state the older drugs ignored. FMC-376 is now being tested in the Phase 1/2 PROSPER trial in patients with KRAS G12C-driven cancers.
Targeted Protein Degradation
The deeper bet sits behind FMC-376. Frontier is identifying novel E3 ligases - the enzymes a cell uses to mark proteins for destruction - and pairing them with covalent binders to degrade targets that cannot be inhibited the usual way. This is where the AbbVie deal lives, and where the platform's longer payoff is supposed to come from.
The money, by round
"A platform is the most expensive way to find out whether your science actually works. It is also, occasionally, the only way."
05 · The proofWho's writing checks, and why.
For a privately held biotech that has not commercialized a drug, the marks of credibility are oblique. They show up as partnership terms, repeat investors, and the willingness of strategic players to put their balance sheets next to yours.
AbbVie did the first. In December 2020, the two companies signed a global partnership covering novel E3 ligases and a slate of difficult oncology and immunology targets - the kind of deal pharma writes when it has tried in-house and would like to stop trying.
Galapagos did the second. The Belgian drugmaker joined Frontier's 2024 Series C as a strategic investor at the same moment Deerfield and Droia were leading the round. That tends to be a tell: investors put in money, partners put in money plus reputation.
06 · The missionDrug the proteins, not the press release.
Frontier's stated mission is to develop transformative medicines against the most important disease-causing proteins, including those considered undruggable. It is, on the page, the kind of sentence every biotech writes. What separates the company from its own marketing copy is that it has built a method specifically for the second clause of that sentence.
The narrative tension here is not subtle. If Frontier is wrong - if the proteome is, in fact, mostly impervious to small molecules - then a platform like this becomes an expensive way to confirm an old assumption. If it is right, the company has been quietly building a way to drug a long list of targets that everyone else has been ignoring, which is the kind of head start that takes years for others to close.
"It is the curious paradox of frontier science: the further out you go, the more carefully you have to walk."
07 · Why it matters tomorrowA pocket nobody saw, in a protein nobody could grab.
Most people will never read a chemoproteomics paper. They should still care what this company is doing.
If FMC-376 works, patients with KRAS G12C-driven cancers - lung, colorectal, pancreatic - get a drug that may keep working after the current ones stop. That is, on its own, worth the trip. If the platform works more broadly, the consequence is much larger: a steady production line of medicines against targets the industry has been ducking for decades.
Back to Tuesday in South San Francisco. The scientists are still at the bench. The mass spectrometer is still running. Somewhere in a clinic, a patient is taking an experimental pill against a mutant protein that was supposed to be untouchable. The word "undruggable" is still printed in dozens of textbooks.
Frontier is not arguing with the textbooks. It is just slowly, methodically, making them out of date.