Biotech, in brief

Doctors at the Charite hospital in Berlin gave six people with severe rheumatoid arthritis a treatment built for blood cancer. [1][2]

Every one of the six had already failed up to eight other drugs over the previous ten years. [1][3]

Disease activity fell sharply in all six. Three ended up taking no arthritis medicine at all. [2][3]

The treatment is CAR T-cell therapy. Doctors take the patient's own T cells and re-engineer them in a lab to hunt one kind of cell. [2]

Here they were aimed at B cells, the immune cells that make antibodies. B cells wear a surface marker called CD19, which works like a name tag. [2]

Some of those B cells had learned to make antibodies against the patient's own joints, and kept restarting the inflammation. [2]

Scans showed the engineered cells reached B cells hiding in bone marrow, lymph nodes and joint tissue, where earlier drugs struggled to go. [2][3]

When B cells grew back, they were mostly new ones that had never learned the wrong lesson. The self-attacking antibodies were gone in almost every patient. [2]

Protection from old vaccines survived. Antibodies against chickenpox and tetanus were still there afterwards. [2]

All six had a mild immune reaction in the days after the infusion. No one had the severe nerve problems that CAR-T can cause, and infections were rare. [1][2]

One patient did have a serious liver-enzyme rise, which settled without lasting harm. [1]

This is six people, not six hundred. Some did not respond fully, and one relapsed after a spell with no medicine. [2]

The same technology is having a hard week off the ward: Bristol Myers Squibb just dropped the company building its automated cell factory. [4][5]

Why it matters. The next test is a larger trial in more patients, and whether the remissions hold past a year; a second relapse would say the wrong memory can rebuild itself.

Two unvaccinated residents of Lancaster County died of measles, the state's first measles deaths in 35 years. [7][8] Pennsylvania had 393 cases across 28 counties, and the US had 2,777 this year by 20 August, its worst since 1991. [7][8] The Lancaster outbreak alone has hospitalised 70 people, a third of the infected being children. [11] Health secretary Robert F. Kennedy Jr. offered federal help; Governor Josh Shapiro said his vaccine rhetoric was part of the problem. [9][10] Reuters counts 36 separate US outbreaks this year. [12]

Why it matters. Measles spreads once enough people stop being protected, and the line is a share of the population, not a personal choice.

A gene-editing treatment given once cut LDL cholesterol by about half in the highest-dose patients, and the drop held for a year. [13] The study followed only 15 people; an earlier study using a different editing method reported similar falls in 35. [13]

Why it matters. Half of people who start a daily statin stop taking it, so a one-time change is worth a lot if it turns out to be safe.

A personalised vaccine cut the odds of melanoma coming back in a 1,137-person trial, announced on 19 August. [14][15] Moderna's own president says nobody in the company's leadership believed it would work when mid-stage results first landed. [14]

Why it matters. Each dose has to be made for one patient from their own tumour, so the science working is only half the problem.

AstraZeneca and Ionis reported that eplontersen did not deliver the benefit expected in a common heart disease caused by a misfolded protein. [16] A similar drug from Alnylam had shown a benefit last year, and the miss has reopened the question of how big that benefit really is. [16]

Why it matters. A failure in one drug is now being read backwards as evidence about a different, approved one.

The FDA approved Jazz Pharmaceuticals' Ziihera for a common stomach and oesophagus cancer, taking the eligible US group from about 1,500 to roughly 12,000 people. [17] Analysts noted the label allows use in tumours with lower levels of the target protein than most trial patients had. [17]

Why it matters. The size of an approval, not the size of the trial, is what decides how many people get the drug.

Bristol Myers Squibb ended its partnership with Cellares, saying the startup's automated Cell Shuttle could not make its CAR-T therapy Breyanzi at commercial scale. [4] Cellares is laying off 100 staff by 20 October. [5]

Why it matters. It is the same technology as today's lead story, and it shows the hard part is no longer the biology.

Boston University researchers reported a way to grow helper T cells in bulk from stem cells, by flipping a signal called Notch. [6] Today's CAR-T treatments are made one patient at a time from that patient's own blood. [6]

Why it matters. It is aimed at the exact problem that just broke the Bristol Myers deal: making these cells at scale.

The state's biopharma workforce fell by about 3,600 jobs in 2025 to 113,503, the first decline since at least 2002. [19] Investors have come back, but mostly for later-stage drugs; MassBio says its biggest worry is the earliest startups with the riskiest science. [18]

Why it matters. Money returning to a sector does not mean it returns to the same place it left.

Sentivera launched with a licensing deal worth up to $1.5bn for an inflammatory disease drug from China's Haisco Pharmaceutical. [20] Haisco takes $40m in cash and about $36m in stock, a 17.5% stake. [20]

Why it matters. Licensing from China is now common enough that MassBio has started counting it as a pressure on US startups. [18]

Researchers mapped how the proteins made by 100 autism-linked genes physically interact, finding more than 1,800 connections. 87% had never been reported. [21] Disease-linked mutations did not break the proteins. They weakened or strengthened particular connections. [21]

Why it matters. Different genes produced similar effects, which suggests the shared thing is the network, not any one gene.

Two Nature Biotechnology papers reported methods for inserting long stretches of DNA at an exact spot without cutting both strands. One replaced a 2.9-kilobase stretch with up to 57.8% efficiency. [22] The other inserted up to 12.5 kilobases. [23]

Why it matters. Most gene editing so far changes single letters; whole genes need this.

University of Toronto researchers used inhaled engineered transfer RNAs to get past a premature stop signal in the gene behind cystic fibrosis, restoring the missing protein in laboratory models. [24]

Why it matters. The same kind of typo causes many rare diseases, so one delivery method could cover several. It is not yet in people.

BioXcel Therapeutics filed for Chapter 11 bankruptcy protection after running out of cash, and named Teva as the lead bidder for its brain-drug portfolio. [25] The baseline offer is $57.5m upfront plus up to $67.5m more if the drugs hit targets. [25]

Why it matters. A failed biotech's science does not disappear; it changes owner at a discount.

A 400-patient European trial called STAR-TREC tested chemotherapy plus targeted radiotherapy against radiotherapy alone. [26] Four in five in the combined group avoided major surgery for up to a year, against three in five. [26] Patients were followed for 30 months; fuller results come in 2027. [26]

Why it matters. The standard cure often costs the rectum and a permanent colostomy bag, and that price is now negotiable for some.

Two studies found women who took hormone therapy at menopause were less likely to be diagnosed with Alzheimer's, and had fewer signs of it in the brain. [27] Both were observational, meaning nobody was randomly assigned, so healthier women choosing treatment could explain it. [27]

Why it matters. This question has produced opposite answers for twenty years, and observational data cannot settle it.

Generation Lab offered journalists and celebrities early access to an injectable pair of drugs it says stops ageing spreading through the blood. [28] It will not say what the two drugs are, and is giving them out before any trial has reported. [28]

Why it matters. The claim rests on real animal science about young and old blood; the evidence offered for the product is a spreadsheet of self-reports.

Stanford researchers reshaped natural killer cells, a fast-acting immune cell, into a form that settles into tissue, and they entered solid tumours far better than the usual kind. [29] In mice they slowed melanoma and head and neck tumours, more so alongside the antibody drug cetuximab. [29]

Why it matters. These cells can be given to strangers without a reaction, so they could be made in batches rather than one patient at a time.

Britain ended sugar rationing in September 1953 and consumption nearly doubled, so babies born months apart had very different early diets. [30] Among 64,000 UK Biobank participants, those rationed longer in their first 1,000 days had lower rates of five cancers. [30] Liver cancer was about 69% lower and breast cancer 36% lower. [30]

Why it matters. You cannot randomise babies, so a policy change is the closest thing to an experiment researchers will ever get.

A Nature Medicine study of human brain tissue looked at what goes with faster cognitive decline. [31] It found smaller nerve fibres, thicker insulation around them, and more of the cells that make that insulation. [31] The pattern was the opposite of what the researchers expected. [31]

Why it matters. It points at the wiring's insulation rather than the nerve cells themselves, which is a different place to aim a drug.

The FDA pushed its verdict on Capricor's cell therapy deramiocel to 22 November after the company filed extra follow-up data and narrowed what it is asking to treat. [32] The agency called the new data a major amendment, which automatically adds three months. [32]

Why it matters. Adding evidence to an application late is not free; the clock resets.

Venture money going into women's health companies fell from $3.2bn in 2024 to about $2bn in 2025, and their share of healthcare deals dropped from 7.4%. [33] Investors and founders are now widening what counts as women's health to draw money back. [33]

Why it matters. When a category stops attracting money, the first move is often to redraw the category.

The most detailed cell-by-cell map yet of the adult human hippocampus, the brain's memory and learning region, found that people with depression produced new neurons less well. [34] Researchers have suspected this for twenty years without being able to show where the process stalls. [34]

Why it matters. It is the first look at which step is failing, which is what a drug would need to target.

About one person in five carries high levels of Lp(a), a blood particle that raises heart-attack risk and is set almost entirely by genes. [35] Three drugmakers have treatments for it in late-stage trials, with Novartis the front-runner. [35]

Why it matters. Diet and exercise do nothing for it, which is exactly why a drug would matter and why the market is large.

Mice bred without cGAS, a sensor that spots stray DNA inside cells, aged faster, were frailer and died sooner than normal mice. [36] Old sequences buried in their genome woke up and drove inflammation. [36]

Why it matters. cGAS was thought to cause age-related inflammation. In mice, removing it made things worse, not better.

Outer Bio, founded by Michael Polansky, says it can keep donated human skin alive and functioning for about a month rather than the usual week, immune cells included. [37] It reports more than 300 donors and 10,000 candidate treatments screened so far. [37]

Why it matters. Testing on living human tissue sidesteps the gap between a mouse's skin and yours.

The Dog Aging Project found that patterns of small molecules in dogs' blood tracked with longer or shorter lifespans in much the same way they do in people. [38]

Why it matters. Dogs live in our houses, breathe our air and age in about a decade, which makes them a fast read on a slow question.

A review pooling nine trials and 6,239 people found injected semaglutide cut body weight by about 12% against a dummy injection, and cut a common inflammation marker by about 41%. [40] Serious side effects were no more common than with the dummy. [40]

Why it matters. Single trials are run by the maker; pooling them is the closest thing to an independent read.

Frontier Medicines, which uses chemistry and software to attack proteins long thought unreachable by drugs, says this year is good after three bad ones and layoffs. [41] It is down to about 60 staff across Boston and San Francisco. [41]

Why it matters. It is what the funding numbers look like from inside one company: survive the lean years or the science stops.

The NIH put out a draft policy for public comment on giving people who join clinical studies a summary of what the study found. [39]

Why it matters. Most people who join a clinical trial are never told what the trial found.