Newsletter · · Ashutosh Agarwal

Biogen Tau Drug Validates the Target While Leqembi Goes Fully At Home - Biotech Pipeline: Gene/Cell, Neuro & Tools - Week of July 19, 2026

Biotech pipeline podcast intelligence for the week of July 13 to 19, 2026. At the London AAIC meeting Biogen's tau drug diranersin slowed cognitive decline 26% with no dose response, the FDA cleared a fully at-home version of Leqembi, blood-based Alzheimer's tests posted strong accuracy, and the biotech M&A machine kept closing deals in record time while gene editing and tools stayed quiet.

Biotech Pipeline: Gene/Cell, Neuro & Tools

Week of July 19, 2026: Biogen Tau Drug Validates the Target While Leqembi Goes Fully At Home

Last week the podcasts kept pointing at a date on the calendar: the big Alzheimer's conference in London, where, everyone promised, we'd finally hear about the "other" Alzheimer's protein, tau. This week London delivered. The world's largest Alzheimer's meeting (AAIC) ran mid-week, and it reshaped the whole picture at once: the field's first serious human data on a tau drug (impressive on paper, messy underneath), a surprise approval that lets patients take the leading amyloid drug at home with a 15-second injection, and a growing pile of evidence that a simple blood test can now spot Alzheimer's decades before symptoms. Around all of it, the deal machine kept humming, Eli Lilly bought a psychedelics company, and a veteran dealmaker walked through why 2026's takeovers are closing in half the time they used to. Gene editing and the lab-tools names, by contrast, went quiet. Here's the week.

TL;DR

  • Tau finally gets its shot on goal, and it's complicated. At the London conference, Biogen's tau-lowering drug diranersin slowed cognitive decline by 26% versus placebo in a mid-stage trial, roughly the same benefit as today's amyloid drugs, but with a head-scratcher: the lowest dose worked best while barely touching tau, and higher doses caused confusion. Biogen's stock fell, and analysts openly asked whether an expensive final trial is worth it. (The Readout Loud, Jul 16; BioSpace, Jul 15)

  • Leqembi can now be done entirely at home, a real launch inflection. The FDA approved an under-the-skin version of Biogen/Eisai's Leqembi for starting treatment, ahead of its August deadline. Combined with the already-approved at-home maintenance shot, patients can now skip infusion centers entirely, a 15-second injection versus a one-hour IV, a clear edge over Lilly's Kisunla, which is still IV-only and holds roughly 80% of the early-treatment market. (BioSpace, Jul 15; Citeline, Jul 17)

  • Deals are closing twice as fast as they used to. Billion-dollar-plus biotech takeovers are closing in an average of 46 days in 2026 versus 112 days in 2023-2025, GSK's $10.6B buy of Nuvalent closed in 36. This week Eli Lilly added a ~$3.8B psychedelics deal to the pile. (Citeline Scrip Deals, Jul 16; Bloomberg Intelligence, Jul 16)

What's new

The tau era officially begins, with a genuinely confusing first data set. For years, Alzheimer's drugs have all gone after amyloid, the sticky plaque that builds up between brain cells. Tau is the other toxic protein, it tangles up inside nerve cells and is more tightly linked to the actual loss of thinking and memory. This week at the London conference, Biogen presented the first serious mid-stage human data on a tau-lowering drug called diranersin, and on The Readout Loud, the STAT team laid out exactly why it's both exciting and awkward. The good news: at its best dose, the drug "slowed clinical decline by 26% on a commonly used measure of cognition and function versus placebo," and "also substantially reduced levels of tau." That 26% is, roughly, in the same ballpark as what the approved amyloid drugs showed in their own trials, so tau, as a target, appears to be real.

The awkward part is the shape of the data. As the hosts put it, there are "more questions than answers," because "there's no dose response here", the most benefit came from the lowest dose tested, which also produced the least tau reduction. Higher doses, meanwhile, "were associated with sort of some states of confusional states." If a drug's whole story is "lowering tau helps," you want more tau-lowering to mean more benefit; here it was the opposite. There is a real potential upside: because diranersin goes after tau rather than amyloid, "there's a thought that it won't have the kind of brain swelling and bleeding that you saw with ARIA", the dangerous side effect that haunts the amyloid drugs. But investors weren't reassured. "Biogen's stock price fell when these data came out," and the team framed the deeper worry bluntly: is "a very risky, expensive, and time-consuming phase three study" the best use of Biogen's money, or is this "deja vu all over again" from the Aduhelm debacle, the controversial 2021 drug approval that Biogen has spent years trying to move past with cost cuts, a new CEO, and a pivot toward kidney and immunology drugs? (The Readout Loud)

Two more tau readouts landed the same week, and both were misses. Lest anyone think tau is easy, the JAMA editors walked through four Alzheimer's papers published to coincide with the conference. One tested sepirognostat, an oral drug meant to stop tau from aggregating; over about two years in early patients, it showed "no sign that sepirognostat slowed decline," and "in fact, the higher dose group seemed to decline faster." (An accompanying editorial by Rabinovici and Grill made the pointed reminder that shrinking a biomarker isn't the same as helping a patient, "clinical benefit should remain the gold standard.") So the honest scorecard on tau after this week: one mixed-but-intriguing hit (Biogen) and at least two clear failures. The idea is validated; the execution is hard. (JAMA Editors' Summary, Jul 17)

The bigger commercial news: Leqembi can now be a fully at-home drug. While the science press chewed on tau, the more immediately investable event was a surprise approval. On BioSpace, the team explained that the FDA cleared an under-the-skin (subcutaneous) version of Biogen and Eisai's Leqembi for initiation, the very start of treatment, and it came "much before its PDUFA date, which is August 24th" (the deadline had been pushed back after the agency asked for more information). Leqembi was already approved as an at-home shot for maintenance; now that the start of therapy can also be a shot, "it's now entirely can be done this way without the need for infusions." That's a real differentiator against Lilly's Kisunla, "the only other anti-amyloid Alzheimer's disease therapy that has received an approval," which is delivered only by IV.

Why it matters commercially: Kisunla currently has something like "80% of the market right now, despite... not having first market advantage", a number the hosts traced to Jefferies citing three neurologists. The reason patients and doctors have favored Kisunla is that it has "an end in sight": you take it until your plaques are cleared, then you stop, whereas Leqembi is given indefinitely. The at-home shot is analysts' bet for how Leqembi claws back share. Eisai's own Dr. Mike Irizarry, interviewed on the final day of the conference for Citeline, leaned hard on the word "optionality": after 18 months of every-two-week IVs, patients can switch to a once-weekly shot they give themselves or a care partner gives them, and the new approval means even the first doses can be two weekly injections at home. His pitch in one line: the injection "only takes about 15 seconds versus the one hour of infusion treatment," and no trip to an infusion center. He also confirmed the three-year real-world "LEADER" data showed the drug performing "very consistent with the label", reassuring for a therapy that's only been in wide use a couple of years. (BioSpace; Citeline)

The quiet revolution is the blood test. Threaded through half the week's Alzheimer's coverage was the same point from different angles: you can now detect Alzheimer's from a blood draw, cheaply, years before symptoms. On The Naked Scientists, Edinburgh neuroscientist Tara Spires-Jones explained that elevated phospho-tau (a form of the tau protein) in blood "actually reflects really accurately whether you have plaques in the brain," and can flag people who have plaques "at least 20 years... before symptom onset." The practical payoff arrived this week in her backyard: "just this week up in Scotland, they've started rolling out this blood test to GP practices." Her hope is that it becomes as emotionally routine as a cholesterol check, "you've got a touch of amyloid, we're going to give you one of these drugs."

The numbers behind the hype are getting solid. In a JAMA study of more than 2,600 people who were cognitively normal at the start, those with very high blood P-tau217 had "about a 38% absolute risk of progression to impairment within five years, compared with just 12% for those with low levels", and the marker "added strong independent prognostic value" even on top of an expensive amyloid PET scan (JAMA). And on Quest Diagnostics' own podcast, the company's neurology lead Matt Stroh described a validated test, combining P-tau217 with an amyloid ratio, hitting "91% sensitivity and 91% specificity," with the indeterminate ("can't tell") rate falling to about 7% once a genetic risk marker is added. His memorable framing: P-tau217 is "Gandalf the White" to the older P-tau181's "Gandalf the Gray", newer, "slightly more powerful." The strategic shift both episodes describe is that diagnosis is moving out of specialist memory clinics and into the family doctor's office. (Naked Scientists; Healthier World with Quest Diagnostics, Jul 13)

Deals keep closing, and fast. On the Scrip Deals podcast, Bruce Leuchter (a biotech CEO who's also done stints as a Goldman analyst and a Credit Suisse banker) and host Joe Haas put hard numbers on a trend that's become the market's mood music. There have been 36 deals worth $1 billion or more so far in 2026, and they're closing in an average of "46 days this year from time of announcement to close," versus "112 days or about 3.7 months" across 2023 through 2025, "less than half of the time." The showcase: GSK's "$10.6 billion takeout of New Valent [Nuvalent], which closed on July 15th, 36 days after it was announced." Leuchter's explanation was refreshingly concrete: a friendlier antitrust regulator ("a different posture at the FTC level"), a shift toward simple single-asset "rifle shot" deals instead of messy platform acquisitions, and, crucially, a stock recovery (the biotech index up "about 25% year-to-date") that finally let buyers and sellers agree on price. Behind all of it sits a wall of expiring patents: big pharma is staring down "somewhere in the neighborhood of 230 billion" in lost exclusivity by 2030, which they simply can't backfill with in-house research. (Citeline Scrip Deals)

Right on cue, Bloomberg Intelligence's "drug boss" Sam Fazeli broke down this week's headline transaction: Eli Lilly agreeing to pay up to $3.8 billion for a psychedelics developer, "$2.8 billion upfront cash and a billion dollars in contingent value rights" tied to two lead drugs (a nasal spray and an under-the-tongue form) getting through final trials and winning European approval. These are synthetic versions of a psychedelic for treatment-resistant depression and anxiety, a space Fazeli says is having a real "resurgence." He also flagged, almost as an aside, a development that rhymes with last week's one-and-done cholesterol story: Merck won approval for the first oral PCSK9 cholesterol pill, at $3,800 a year, which, in a world of $10,000-to-$50,000 drugs, he pointedly said he doesn't think "is a lot." (Bloomberg Intelligence)

The debate

Should Biogen actually run the tau phase 3? This was the one place the week's podcasts genuinely argued. The bull case: tau is the more fundamental target, the 26% benefit is real, and because it avoids amyloid it may sidestep the brain-swelling risk that limits Leqembi and Kisunla, a potentially safer drug getting at the deeper cause. The bear case came from the same panel, unprompted: the data has no dose-response, the best dose barely moved tau, higher doses caused confusion, and Biogen, of all companies, given Aduhelm, is the one being asked to spend years and hundreds of millions on a risky confirmatory trial. As one STAT host put it, if she were "in the C-suite at Biogen, I'm not sure that I would make the decision to go forward." There's even a self-cannibalization angle: a successful tau drug could undercut Leqembi, Biogen's own amyloid product. The honest read: the target got validated this week, but the first drug raised as many doubts as hopes, and the market voted with a lower stock price. (The Readout Loud)

Do the amyloid drugs justify their cost and hassle? The most useful reality check came not from a trading desk but from a working neurologist. On The Peter Attia Drive, Dr. Gayatri Devi, who's been using these drugs since 2021, laid out the unglamorous economics. The measured benefit is "0.3 or 0.4 points out of 18 points" on the standard cognitive scale, "a very small change." The drugs themselves run "about $26,000" a year, before administration fees that range from "$400" at an infusion center to "$10,000" in a hospital setting, plus repeated MRIs. And there's a maddening insurance wrinkle: because Leqembi's label starts at the full 10 mg/kg dose, insurers "will not reimburse patients if we start them at a lower dose and titrate them up", so patients who want the safer slow ramp "have to pay out of pocket." Her counterpoint is that outcomes improve dramatically with careful use and early treatment: in her own five-year experience she's seen roughly "4% risk" of imaging-detected brain changes but only "one case" of a patient with actual symptoms, thanks to slow titration and steroid pre-medication. The steel-man on both sides: these are marginal, expensive, high-maintenance drugs for most patients, and they may be genuinely valuable if you treat the right people early and manage them obsessively. Both things are true. (The Peter Attia Drive, Jul 13)

Read-throughs & names in play

  • Biogen (BIIB) / Eisai. The center of the week. Bull: the at-home Leqembi approval is a concrete commercial catalyst that attacks Kisunla's convenience edge, and the three-year real-world data holds up. Bear: the tau readout was messy, the stock fell, and the market is questioning management's appetite for another costly Alzheimer's gamble. Next catalyst to watch: more diranersin phase 2 data (promised), and how quickly the SubQ shot moves share. Eisai's own next act is a tau antibody, etalanotug, now in two mid-stage trials testing it on top of Leqembi, the industry's first real attempt at a combination amyloid-plus-tau regimen. (BioSpace; Citeline; The Readout Loud)

  • Eli Lilly (LLY). Still the industry's dominant force, but a slightly different look this week. Its Kisunla is the market-share leader in early Alzheimer's (and has an embargoed conference readout coming), and it just spent up to $3.8B moving into psychedelics, a reminder that its weight-loss cash lets it shop across the pipeline. Bull: leader in a growing category plus optionality from dealmaking. Bear: the at-home Leqembi shot directly targets Kisunla's main weakness. (BioSpace; Bloomberg Intelligence)

  • Roche (tau read-through). Not a data event, but a tell: the reason Tara Spires-Jones lost her voice at the conference, she joked, was "enthusiastic discussions about Trontinemab", Roche's "brain shuttle" antibody designed to get more drug across the blood-brain barrier. Eisai confirmed it's chasing the same brain-shuttle idea in its own labs. Worth tracking as the next-generation delivery race. (Naked Scientists; Citeline)

  • Blood-based diagnostics (Quest and the broader category). The clearest structural winner of the week's science. As diagnosis shifts to primary care and blood tests replace PET scans and spinal taps, whoever owns the validated assays picks up volume. Quest detailed a 91%/91% test already running on thousands of real-world samples; Scotland is piloting a test through GP offices. Bull: a cheap, scalable gate that every anti-amyloid and anti-tau therapy depends on. Bear: reimbursement and clinical-guideline adoption still lag the science. (Healthier World with Quest Diagnostics; JAMA)

  • AbbVie (ABBV) / Vertex (VRTX) / GSK, the M&A tape. The deal experts singled out AbbVie's ~$10B Apogee purchase and Vertex's Crinetics deal as "asset-centric" bolt-ons expected to close at 2026's new, faster pace, while GSK's Nuvalent close (36 days) is already the poster child. The read-through for holders of likely targets: a friendlier regulator plus a recovered biotech index means clean deals close fast, good for premiums, and a reason the whole small/mid-cap complex has re-rated. (Citeline Scrip Deals)

  • Abbott (ABT), the diagnostics-and-devices read. Not a tools name in the classic sense, but a useful health-check on the space: shares jumped ~11% (biggest one-day move since 2002) on earnings, with the continuous glucose monitor business at a ~$10B run rate still growing ~9.5%, and the newly acquired ExactScience cancer-diagnostics unit as the fresh growth engine. Bull: diversified, resilient procedure volumes. Bear: still down ~20% year-to-date, digesting a $23B acquisition. (Bloomberg Intelligence)

Gene editing: no readouts, but a big-picture reminder

The gene-editing names investors track, CRISPR Therapeutics, Intellia, Beam, Verve, had no fresh clinical data on the podcasts this week. What we did get was a genuinely worthwhile field-level conversation. On the ASGCT podcast, recorded live at the cell-and-gene-therapy meeting in Boston, pioneer Fyodor Urnov (a co-inventor of the term "gene editing") gave a state-of-the-union. His core message is that the field is now "modality agnostic", for any given patient his team screens every tool and picks the best one rather than betting on one technology. He name-checked the landscape approvingly: sickle cell can be fixed with the approved nuclease behind Vertex's Casgevy; "Beam Therapeutics has some incredibly impressive data on base editing for sickle... spectacular"; "the two clinical trials from Intellia run on CRISPR"; and newer players like Prime, Tessera and Arbor are in the mix. The through-line to this week's other big theme (funding, see below): Urnov made an impassioned case that essentially every modern medicine, from CRISPR to the mRNA vaccines to the cancer drug Keytruda, traces back to federally funded academic research, "the crown jewel of us as a species", and that the real barrier now isn't the science but cost, quoting Jennifer Doudna's line that "a $3 million cure is not a cure." (ASGCT Podcast Network, Jul 14)

The storm cloud: funding and China

Two conversations this week captured the structural anxiety underneath the good news. On Business of Biotech, industry veteran Jeremy Levin (author of a new book, Biotech in the Balance) argued that the US biotech engine faces "structural" rather than "cyclical" threats, cuts to federal science funding, turmoil at the FDA, and politics driving talent away. His most striking data point: American biotech "to date provided 70% of all new drugs to the pharmaceutical companies," yet capital is starting to flow elsewhere, he cited Bristol Myers Squibb redirecting money toward China, warning that "innovation can migrate. So can capital." He framed China less as a villain than as a competitor that made biotech a national priority 15 years ago while the US treated it as a "political piñata," and noted a bipartisan House bill that would have the Treasury review deals flowing into China. (Business Of Biotech, Jul 13)

The dollars-and-cents version came from Richard Bendis on Biotech Bytes, who called the last three months "the most challenging" of a 50-year career. He pointed to a proposed NIH budget cut "from $47 billion to $27 billion" and a consolidation "from 27 institutes to 8," and reminded listeners that "95% of the companies that have gotten things commercialized in the pharmaceutical world over the last 10 years received some public money." Compounding it: "77% of the investments in the first quarter of 2025 went into AI-related deals," starving non-AI, early-stage life-science startups of capital. It's the counterweight to the M&A euphoria, deals are flying for late-stage, de-risked assets, while the seed corn of the next decade is getting squeezed. (Biotech Bytes, Jul 16)

What changed

Last week the story was permission and one-shot cholesterol data; the tau conversation was still a promise pinned to "next month in London." This week London happened, and it moved the center of gravity firmly to Alzheimer's, with a nuance that matters: the tau approach got its first real validation (Biogen's 26%), but the drugs stumbled (Biogen's messy dose-response, plus two outright tau failures), and the most bankable news was an old amyloid drug going fully at-home. Meanwhile the M&A "race" that was an anecdote last week is now a measured trend, deals closing in 46 days versus 112, and the funding worry sharpened from vague unease into specific numbers ($47B-to-$27B for the NIH; 70% of new drugs historically from US biotech). The blood-test story also graduated from lab promise to something a Scottish GP can now order. (The Readout Loud; Citeline Scrip Deals; Biotech Bytes)