GSK Kills Camlipixant: The P2X3 Cough Bet Just Went to Zero

This week the money came roaring back into biotech — seven startups queued up to go public in a single month, Lilly closing its eleventh acquisition of the year — and right in the middle of the party, a two-billion-dollar bet quietly went to zero.

GSK walked away from camlipixant, Merck put an oral cholesterol pill and a lung-cancer proof-of-concept on the board, and Novartis locked in a full approval in IgAN. Risk and reward in the same five days. Here's how to read it.

Welcome to The Pharma Closeout Weekend Edition for Sunday, July 19th. I'm Alex Mercer.

And I'm Maya Patel. It was a loaded week. Let's take it apart.

The story to sit with this week in review is the one we flagged Friday — because a couple of days later it looks bigger, not smaller. GSK ceased development of camlipixant after mixed Phase 3 data. In doing it, they wrote off most of a two-billion-dollar bet. Another drug in the P2X3 class has failed to advance.

And the two billion is the number everyone anchored on. It's the least interesting one here. GSK bought camlipixant as the great refractory chronic cough hope — the mechanism that was finally going to give these patients something. Real unmet need. P2X3 was the whole thesis.

So GSK looks at the data, and instead of grinding toward a filing, they walk. That's a clean kill.

OK, but I'd push back on reading this as a GSK stumble. The write-down is the small story. Camlipixant was the selective bet — engineered to be more P2X3-selective to dodge the taste problem that hobbled the first-in-class drug. And the mixed efficacy tells you something uncomfortable. The selectivity that fixed tolerability may have cost them the cough benefit. Same receptor biology sitting on both sides of that trade.

That's fair. And think about what that trade-off does to the whole design philosophy behind these programs. The entire pitch for a next-gen P2X3 was "we kept the efficacy, we dropped the taste loss." If the selectivity is what's killing the cough signal, then you can't engineer your way out of it — the thing you'd tune to win tolerability is the same knob that costs you the endpoint. So the read isn't "GSK bought a dud" —

It's that the whole P2X3 mechanism for chronic cough is now in serious doubt. Three failures. When you can't separate the efficacy from the on-target liability, that's not a molecule problem. It's a target problem. And a target problem is a much more expensive thing to have discovered, because every dollar the field spent chasing a cleaner molecule was spent solving the wrong problem.

Let that sit for a second, because it reframes the money. GSK didn't overpay for a bad asset — they paid a fair price for a mechanism the field believed in, and the mechanism blinked. That's a different kind of loss than a bad diligence call. You can underwrite around a weak molecule. You can't underwrite around biology that only reveals itself in a Phase 3 the size of the one it takes to prove chronic cough. The fallout lands on patients first: refractory chronic cough is right back to essentially nothing. And commercially, anyone still funding a P2X3 program is walking into a much harder capital conversation tomorrow morning.

Which is the part I'd underline. Because it's not just the P2X3 owners who feel this. Every investor who wrote a check into a mechanism-driven respiratory story now has to ask whether their thesis rests on the same kind of single-target bet that just failed three times in a row. The camlipixant kill doesn't stay in its lane.

The competitive read: if you run early-stage respiratory business development and you've got a P2X3 asset anywhere in your comp set, that mechanism just got repriced. Three swings, three misses. The burden of proof now sits entirely with anyone claiming this class works. And "burden of proof" in this context means a bigger, longer, more expensive trial than anyone wanted to run — which is exactly the kind of program that doesn't get funded in a market that's already picky about where it puts late-stage dollars.

And the open thread is whether anyone keeps swinging. Gefapixant cleared — barely. Now both follow-ons behind it are gone. So the question I'm watching: does the field pivot to a different cough target entirely, or does chronic cough go back to being the graveyard it was for a decade? Because the honest answer is that a barely-cleared first-in-class plus two dead follow-ons is not a category — it's a warning label. The strategic implication is that the next wave of cough investment, if it comes at all, has to come from a completely different mechanism, and that resets the clock by years.

Now flip the tape, because on the deal side the mood couldn't be more different. The through-line of the week was appetite. Lilly agreed to buy AtaiBeckley for $2.8 billion — its eleventh drugmaker acquisition this year, by far the most among its peers, and it plants a flag in psychedelics as a real category. And the IPO window is wide open: Latigo, BlossomHill, and Vogenx became the fifth, sixth, and seventh startups to join July's queue alone.

And the contrast with our top story is the whole point. Capital is flooding back in at the exact moment a two-billion-dollar bet just proved how fast it evaporates. That's not irrational. But it's a risk-on market with a very short memory. And short memory is the operative phrase, because the camlipixant write-down and the AtaiBeckley signing are separated by days, not quarters. The same institutions cheering one are absorbing the other.

Let me bring in Marcus Webb on the Lilly pattern, because eleven deals in a single year is not normal behavior.

And it wasn't just the buyers moving. The IPO side rhymes with that appetite — Attovia, Braveheart, and Apnimed all filed this month too. Meanwhile at the other end of the risk curve, Jasper got a lifeline: an all-stock merger with immune drugmaker Kira, plus a concurrent licensing deal with a startup called Mirador, after setbacks had wiped out most of Jasper's market value.

And that Jasper deal is the honest bookend to the week. Same market that's handing IPOs to seven companies is also quietly stitching survival mergers for the ones that stumbled. The window rewards conviction — right up until a Phase 3 reminds everyone what conviction actually costs. And notice the shape of the Jasper rescue — an all-stock merger plus a licensing deal bolted on. That's not a company being bought for its value. That's a company recombining what's left of itself into something a public market can still hold. The IPO queue and the Jasper structure are the two exits of the same risk curve.

On the regulatory side, the week was lopsided toward wins — the counterweight to camlipixant. Lead item: Novartis got full FDA approval for Fabhalta in primary IgAN, converting from accelerated approval. The way Novartis frames it, this is the first and only complement inhibitor cleared to slow kidney function decline in these patients. And the real value isn't the headline — it's what traditional approval does. It locks in the label and takes the confirmatory-data risk off the table.

That's a franchise Novartis is building brick by brick — Fabhalta keeps collecting indications. And the reason that matters commercially is that each conversion from accelerated to full approval turns a contingent asset into a certain one. Payers treat a locked label differently than a provisional one. So this isn't Novartis adding a line item — it's Novartis de-risking a franchise it can now sell against for years without a confirmatory-trial cloud hanging over it.

Then two more worth flagging fast. Celcuity got the nod for Revtorpyk in HR-positive, HER2-negative breast cancer — but shares fell nearly twenty percent, because they disclosed a launch delay and the prescribing details caught analysts off guard. The approval wasn't the milestone that moved the stock. The label and the timing were. And that's the lesson that keeps repeating: the market stopped paying for approvals a while ago. It pays for how fast you can launch and how broad the label lets you sell. Celcuity cleared the hard scientific bar and still lost a fifth of its value in a day, because the commercial read got worse, not better. And Merck won what BioPharma Dive called the first approval of an oral PCSK9 pill — a genuinely different delivery story for cholesterol.

And "different delivery" is doing a lot of work there, because the PCSK9 class already works — the barrier was always that it was an injectable in a category where patients expect a pill. An oral version doesn't just add a competitor, it potentially unlocks the primary-care prescriber who was never going to reach for an injection. That's a much larger funnel than the specialist market the injectables have been fighting over. And then there's the one we've been tracking all week — Merck's ADC, sac-TMT.

Right. The new piece is the China Phase 3, OptiTROP-Lung06. It hit its primary endpoint on progression-free survival in combination with Keytruda in front-line lung cancer, and Wall Street called it direct proof-of-concept. Merck's been calling this ADC a cornerstone. This is the data that starts to earn the word. And you have to read this against the Keytruda cliff — Merck needs the next cornerstone franchise, and an ADC that pairs with Keytruda in front-line lung is exactly the kind of asset that both extends the current franchise and builds the bridge to what comes after it.

Chemo replacement in front-line lung is not a small claim.

It isn't. Front-line is the biggest, most defended setting in oncology, and "replace the chemo" is the highest bar you can set for an ADC. A single positive PFS readout doesn't get you all the way there — but it moves the word "cornerstone" from a marketing line to a defensible thesis. There was one quieter clinical note too — Biogen's tau-targeting Alzheimer's data. It's the clearest evidence yet that a medicine like this might actually improve cognition, but here's the thread I can't get past: the higher doses underperformed, and nobody's explained why. In a field this scarred, an unexplained dose-response inversion is exactly the kind of detail that decides whether this program lives.

That's the line that keeps a development team up at night. Because in Alzheimer's, the graveyard is full of programs that had a signal and couldn't explain their own dose curve. When the higher dose does worse, you either have a mechanism you don't understand or a trial you can't trust — and both of those are the kind of thing a regulator asks about first, not last.

It should. But put the whole scorecard together and it's a yes-heavy week. Fabhalta, Revtorpyk, the PCSK9 pill — every one is a label-breadth story now, not an approval story. The approvals are done. The commercial fight is all about how wide each label runs. And that's the quiet shift in the whole industry this week: the science bar got cleared across the board, and every one of these winners now graduates into a fight that has nothing to do with whether the drug works and everything to do with who it's allowed to be sold to.

Under the radar this week — buried beneath all the IPO noise — a report that the biotech startup funding gap is actually widening. Even though first-half venture investment hit its highest level since early 2022. The money's back. It's just concentrating.

And that's the tension nobody's pricing. The window being open for Attovia and Braveheart doesn't mean the seed and Series A layer is healthy. Those are two different markets wearing the same headline. The IPO queue tells you late-stage capital is flowing. The funding gap tells you the earliest capital — the stuff that has no data yet, only a hypothesis — is getting harder to raise even while the aggregate number looks great.

Which is the layer that becomes the next decade's pipeline.

Exactly. If capital pools at the top, you get a strong class of late-stage names — and a thinning bench of the early companies that turn into everyone's future acquisition targets. That's the number to watch while the VC rebound gets its victory lap. Because a Lilly buying eleven companies a year needs a deep bench of targets to keep buying from. Starve the seed layer now and you're not feeling it this year — you're feeling it in the back half of the decade, when the serial acquirers go shopping and the shelves are thin.

Looking at the week ahead, a few things on the calendar. Novartis reports earnings Tuesday the twenty-first — and after the Fabhalta approval, the pipeline commentary is what to listen for. Thursday the twenty-third, Sanofi has an FDA action date for subcutaneous Sarclisa in multiple myeloma — the on-body injector version. And the following Tuesday, Outlook Therapeutics has a resubmission decision on its nAMD program.

And one open loop worth naming honestly. Novartis and Ionis guided pelacarsen Phase 3 topline for the near-term future. We've been tracking it — and that window has effectively closed with no readout. It hasn't landed.

So Tuesday's call gets interesting. Because Novartis walks into that earnings call with a fresh Fabhalta win in one hand and an overdue pelacarsen readout in the other — and the market will weigh the silence on the second at least as heavily as the noise on the first.

It does, because an Lp(a) outcomes trial slipping is a signal all by itself. If they address the timing, listen closely to how. If they don't address it at all — that tells you something too. Guidance that quietly lapses without comment is the kind of thing management hopes gets buried under a good-news headline. Our job is to not let it.

That wraps our Weekend Closeout — and if you want one frame for the week, it's this: biopharma is running risk-on and risk-off at the same time. Lilly's eleventh deal and a packed IPO queue on one side, camlipixant's two-billion-dollar failure on the other, with Novartis, Merck, and Celcuity stacking up approvals in between. Novartis earnings and that Sarclisa decision headline the days ahead. Subscribe now and you'll walk into this week sharp — and if this briefing saves you time, follow us on Spotify and drop a rating.

Go enjoy the rest of your Sunday while it's still yours. See you tomorrow for Novartis and the read into the week. Let's go!