Regenxbio Plunges 26% as FDA Halts Hunter Gene Therapy & Capricor Duchenne Decision Slips to Nov. 22 on Narrower Label

Regenxbio lost a quarter of its value today — and the clinical hold wasn't what did it. One sentence buried in the release did.

Meanwhile Roche is paying up front for muscle-sparing obesity, Merck and AstraZeneca both lost on Medicare price negotiation, and Capricor's Duchenne decision moved — with a much narrower label attached. Let's get into it.

Welcome to The Pharma Closeout for Monday, August 24th. I'm Alex Mercer.

And I'm Maya Patel.

Regenxbio fell twenty-six percent to seven ninety-six today. And the headline everyone ran with was the FDA clinical hold on RGX-121 — the Hunter syndrome gene therapy. That's not what moved the stock. What moved the stock was the company saying it does not expect to resubmit the BLA in the near term. A hold is a pause. That sentence is a program coming off the board.

And look at where the finding came from, because it inverts the usual story. This wasn't a new patient. It wasn't a new dose. These are boys treated three to six years ago, and the signal turned up because Regenxbio voluntarily expanded its MRI protocol a few months back — after the January hold on RGX-111, when a patient on the MPS I program developed a CNS tumor. They built the surveillance. The surveillance found something.

Give people the year, because the whiplash is the point. January, FDA holds both programs. February, rejection — reviewers didn't buy the surrogate endpoint. Regenxbio appeals, gets a Type A meeting, and by June the agency has dropped its demand for more patients and an untreated control arm and agreed to review a filing on existing data. Third-quarter resubmission was on the calendar three weeks ago.

Small nodule or small cystic mass on spine MRI in five participants. Investigators called them nonserious. Radiologists believe they're likely benign. All five are clinically well, with overall stability to improvement on neurocognitive and neurobehavioral assessments.

Then explain to me why that kills a filing.

Because spine MRI isn't done in this disease. Not in practice, not in trials. So nobody knows the background rate of these findings in untreated boys with Hunter syndrome. Regenxbio has a numerator and no denominator. One of the investigators on the release, Roberto Giugliani, makes exactly that argument — that asymptomatic findings like these may be inherent to the disease itself. He may well be right. But "may be inherent" is not a benefit-risk case you walk into a BLA review with, and the company clearly knows it.

Here's my pushback. The reaction still looks disproportionate to the medicine. Five asymptomatic patients, likely benign, all doing well — and the market takes out a quarter of the company in one session.

If the finding were the whole story, I'd be with you. It isn't. This asset was one filing away from a decision, and the filing just left the calendar with nothing replacing it. You're not repricing a safety signal. You're repricing a near-term revenue line to zero and attaching an open-ended timeline to whatever's left. Twenty-six percent isn't panic — it's arithmetic.

So the honest synthesis is that the hold is the mechanism and the abandonment is the event, and the market priced the event. Which throws everything onto the rest of the portfolio. The CEO spent his entire quote pointing away from Hunter — Duchenne and retinal, different capsid, different routes of administration. Duchenne BLA still planned this quarter. Wet AMD pivotal topline in the fourth quarter with AbbVie. Partner NS Pharma is working the imaging data alongside them.

And that's the sentence I'd flag for anyone else running an AAV program. Regenxbio's position is that these findings are unique and limited to Hunter syndrome. Maybe. But the capsid defense only holds as long as nobody else goes looking with a spine MRI.

Which is the uncomfortable part.

It is. Enhanced imaging is how this was found. So every sponsor delivering gene therapy into the CNS now has a live decision — do you run the expanded protocol and risk generating your own finding, or do you wait until the FDA makes that decision for you? There's no clean answer, and Regenxbio just demonstrated what the downside of the virtuous option looks like.

The competitive read: today's real event isn't the stock. It's that voluntary surveillance was just shown to be capable of stopping your own filing. Every CNS gene therapy sponsor has to price that trade this week.

And the thread nobody can close yet is a document that hasn't arrived. Regenxbio says it still hasn't received the full clinical hold letter. What the agency asks for in writing — more imaging, longer follow-up, or a natural-history comparison that doesn't exist for this disease — is what separates a delay from an ending.

Shifting to the deal side, and this one we flagged Friday as talk — today it's signed. Roche's Genentech takes global rights outside South Korea to Hanmi's HM17321 for a hundred and ninety million dollars up front, up to two point three billion with milestones and tiered royalties. Hanmi finishes the Phase 1, Genentech drives from Phase 2 on.

A UCN2 analog. Non-incretin — and the entire pitch is preserving lean mass while you lose fat.

Which tells you exactly where Roche thinks it's losing. They have the Carmot assets, they have Zealand's petrelintide, and they're still behind Lilly and Novo on incretins. You don't catch a leader on their own axis. You buy a mechanism nobody else owns and change what the axis is.

The sharper way to read the price is the stage, not the strategy. The body-composition case for this molecule is preclinical — including a primate head-to-head against Wegovy — and none of it has been replicated in a controlled human trial. Roche is buying a hypothesis at Phase 1, and the milestone-weighted structure says both sides know that.

So what would move it from hypothesis to thesis?

The endpoint Genentech picks in Phase 2. If lean mass sits as a secondary, this is a cheap hedge on a fashionable idea. If they actually power the study for it, Roche is trying to change what "quality weight loss" means on a label — and that's a much bigger swing than the upfront implies.

On the regulatory side, today was the FDA's action date for Leqembi Iqlik — Biogen and Eisai's once-weekly subcutaneous lecanemab, this time as the starting dose for early Alzheimer's. This date only exists because the agency deemed additional submitted material a major amendment in the spring and pushed the review out.

Subcutaneous maintenance is already approved though. What does moving it to the starting dose actually unlock?

It takes the infusion chair out of the front end of therapy, which is precisely where the access bottleneck lives. But the thing I'd read first isn't the approval — it's the ARIA language. The maintenance label never had to answer what weekly subcutaneous dosing does to imaging abnormality rates during titration. A starting-dose label does. That paragraph decides how many community neurologists are willing to initiate, and initiation is the whole commercial question here.

Also worth noting on the regulatory front — the courts went two-for-two against pharma today. A D.C. judge rejected Merck's challenge to Medicare price negotiation after it was filed, on First and Fifth Amendment grounds, ruling that companies remain free to decide whether to participate. And a Maryland federal court threw out AstraZeneca's suit over how CMS grouped its drugs.

Merck filed first over Januvia. What that ruling settles is that "voluntary" is now the frame the courts are comfortable with — and once that's the frame, every remaining constitutional theory hits the same wall.

Which means what for anyone modeling this?

That the industry's leverage on pricing has left the courtroom. Whatever gets won from here gets won in the negotiation cycle itself — commercially, drug by drug — not on appeal.

Before we look ahead — Capricor. We touched this yesterday; today we have the specifics. The FDA had an August 22nd deadline on deramiocel, the Duchenne cell therapy. That decision now lands November 22nd, because Capricor amended its application with follow-up data supporting what it calls a refined indication focused on upper limb function.

And that's the move worth understanding. The original filing targeted a cardiac complication of Duchenne — an advisory committee voted against it there. The same panel was notably more receptive to the upper-limb findings. So Capricor didn't argue with the cardiomyopathy verdict. It rewrote the label it was asking for.

Same drug, narrower ask, later date. The FDA called the new data a major amendment.

It's a live test of a strategy the field keeps circling — when the broad indication fails, refile against the population where the effect is cleanest. It can absolutely work. It also permanently caps the commercial ceiling, because upper limb is a far smaller label than cardiac. Capricor is trading market size for a yes.

With an activist in the room. Kaos Capital is publicly pressing for what it's calling a governance reset.

Which is the part November doesn't resolve. That date settles the drug. It doesn't settle who's running the company when the answer arrives.

Looking ahead — the Regenxbio catalysts that survived today are now carrying the entire equity story: the Duchenne BLA the company still says goes in this quarter, and wet AMD pivotal topline in the fourth quarter. Two shots, no cushion behind them.

And I want the hold letter. Regenxbio says it hasn't received it yet. Imaging cadence, follow-up duration, or a natural-history comparison that doesn't exist for spine MRI in this disease — whichever the agency asks for tells you whether RGX-121 has a path or just a file number.

What stays with me is that Regenxbio built the surveillance that found this. That was the right instinct, and it cost them the filing — the field is going to argue about that trade for a long time. I'll be reading the hold letter the day it surfaces.

And that is your Pharma Closeout for Monday, August 24th — Regenxbio's twenty-six percent day, Roche buying a non-incretin bet from Hanmi, Merck and AstraZeneca losing on the IRA, Leqembi Iqlik's action date, and Capricor buying time with a narrower label. Follow the show on Spotify, Apple Podcasts, or wherever you listen — tomorrow's briefing lands on its own.