Biotech, in brief

Doubt about vaccines spreading

The US medical research agency is preparing a clinic for people who believe a vaccine made them ill. [3] A poll of 1,143 US adults found 76% call the measles shot safe for children, against 84% in 2020. [2]

Big drugmakers cutting programmes

Novartis ended its ALS drug after a 251-person trial missed every measure it was set. [4] Novo, the Danish maker of Wegovy, cancelled a deal worth up to $285m for a once-a-month weight-loss injection. [7] Eli Lilly stopped an early obesity drug that did not meet its own bar. [8]

Money moving to delivery and data

Novartis paid $125m for a technology that carries large drug molecules into the brain. [5] The Mayo Clinic and Thermo Fisher set up a company to gather molecular data from people before they fall ill. [15] Samsung Biologics signed a $262m contract to manufacture another firm's medicines until 2033. [16]

Side effects found after approval

A team at the University of Pennsylvania read 400,000 posts and found period changes and chills that weight-loss drug labels do not list. [14] The planned US study of illness after vaccination has almost no healthy vaccinated people to compare its patients against. [3]

One pressure seen across the day runs through Biotech too.

• Diseases outrunning the people sent to stop themHigh ↑ · seen in 2 beats→

1. 1Measles has killed more than 1,000 children in Bangladesh, and 83% of confirmed cases never had the vaccine
2. 2A US clinic for people who say a vaccine hurt them
3. 3Novartis ends its ALS drug after a failed trial
4. 4Novartis buys a way across the brain's barrier
5. 5Novo drops a monthly weight-loss injection deal

Measles has killed more than 1,000 children in Bangladesh since the outbreak surged in March. [1]

Watch whether Bangladesh sends health workers into the districts and age groups the first campaign missed, and whether the daily count of deaths starts to fall.

Also seen in Food.

The United States National Institutes of Health, that country's main medical research agency, is preparing a research programme for people who believe a vaccine harmed them.

It could open in mid-October in Bethesda, Maryland. [3]

The study, called MOSAIC, would collect clinical and genetic data on people who developed lung or nerve problems after a vaccination, and would be run by the agency's own staff. [3]

It was proposed last year by a US commission chaired by Robert F Kennedy Jr, the health secretary and a longtime critic of vaccines. [3]

A patient advocate who developed lasting symptoms after a COVID vaccination says the work should have happened years ago, and that being lumped in with anti-vaccine campaigners adds insult to injury. [3]

A vaccine researcher at Vanderbilt University says she wants honest answers about side effects, and does not want people who know nothing about science choosing which studies get funded. [3]

Almost nobody can check whether a vaccine is safe, so most people decide by trusting or doubting the person holding the needle.

the same step of judging the messenger, from both directions: patients say the health system dismissed their symptoms, and scientists say the answer will be believed or dismissed depending on who runs the study

people reported symptoms to each other on an online forum instead of to a doctor or a regulator, which is where trust had gone

officials cannot inspect every carcass, so the protection depends on families believing an instruction that takes away food they rely on

People have trusted the wrong source and been harmed, and people reporting real side effects have been waved away by the institutions asking them to trust.

A reader who has never read a trial report is relying on strangers exactly as the parents in a Bangladeshi clinic are, and mostly it works, which is why it is invisible until it fails.

Novartis, the Swiss drugmaker, has stopped developing VHB937 for ALS, the disease that destroys the nerve cells controlling movement, after a 251-person trial missed its main and secondary measures. [4]

The drug was meant to stabilise a protein called TREM2 so that the brain's own immune cells protect nerve cells. [4]

Two other companies' TREM2 drugs failed before it, in 2024 and last year. [4]

Novartis is still testing the same molecule in Alzheimer's disease, where the evidence for the target is stronger. [4]

It lands in a bad few weeks for Novartis, which paused cell-therapy trials after three deaths and saw two late-stage drugs fail. [4]

Analysts read the string of failures as pressure to buy growth from outside instead of growing it in its own labs. [4]

Novartis will pay Sironax, a company near Boston, $125m for full ownership of its technology for carrying drugs into the brain. [5][6]

The blood-brain barrier is a filter that keeps toxins out of the brain and keeps most large drug molecules out with them, which has blocked treatments for Alzheimer's, Parkinson's and ALS. [5][6]

Novartis took an option on the platform 14 months ago and has now exercised it. [5]

Sironax keeps the right to use the technology for three of its own early-stage drugs and will spend the money on them. [6]

Novartis has 14 neuroscience programmes and its own head of neuroscience says no single trick will solve the barrier, so it is buying a toolbox. [5]

For Sironax, selling the platform pays for the medicines it wants to keep.

Novo, the Danish maker of Wegovy, has terminated a partnership signed in November 2024 with Ascendis Pharma that was worth up to $285m in payments. [7]

The lead project was a weight-loss injection given once a month instead of once a week, built on an Ascendis technology that links a drug to a carrier that lets go slowly. [7]

Ascendis said on Monday it will now develop the monthly version of semaglutide, the ingredient in Wegovy, on its own. [7] Novo would not explain the decision. [7]

Novo lost a bidding war for a rival monthly drug last year and its pipeline is thin after two more trials of another drug were dropped. [7]

A partner walking away leaves the smaller company holding both the cost and the upside.

Eli Lilly has terminated the first human trial of LY4064912, an experimental obesity drug, after 77 volunteers, saying it did not meet the company's bar for going further. [8]

The trial ran in California, Texas and Singapore and included both healthy people and people with obesity. [8] Lilly would not say how the drug was meant to work. [8]

The company recently stopped publishing its list of early-stage projects at all, removing them from its website and its results. [8]

Lilly already leads the obesity market and has a triple-hormone drug that produced 22.6% weight loss in a late-stage trial, so the bar an early candidate must clear is set by its own products. [8]

Johnson & Johnson is in talks to sell its orthopaedics business to the investment firm Apollo in a deal worth about $20bn, Bloomberg reported. [9]

The unit makes hip, knee and shoulder implants and surgical instruments, and took $9.3bn in sales last year, barely above the $9.2bn of 2024. [9]

J&J said last October it wanted to separate the business within 18 to 24 months. [9]

It bought two of the businesses inside it for $21bn and $3.5bn, in 2011 and 1998. [9]

Novartis, Sanofi, GSK, Merck and Pfizer have all cut loose units that are not medicines. [9]

A slow-growing implant business is worth more to a buyer that wants steady cash than to a drugmaker judged on new drugs.

Connect Biopharma's drug rademikibart cut the rate of treatment failure by 66% over 28 days in a mid-stage asthma trial, but the result was not statistically significant, meaning it could have happened by chance. [11]

The company said fewer patients got worse than it had expected, which left too few cases for the difference to count as real. [11]

It pointed instead to a breathing test that did improve. [11] Its shares fell 35% to $1.12 on Tuesday morning. [11]

Results from the same drug in the lung disease COPD are due at the end of the month, and analysts asked whether those patients overlap with the ones just tested. [11]

The drug aims at the same target as Dupixent, one of the biggest-selling medicines in the world.

A University of Pennsylvania team used artificial intelligence to read more than 400,000 posts from nearly 70,000 people on Reddit, an online discussion site, about weight-loss and diabetes drugs including Ozempic, Wegovy and Mounjaro. [14]

Nearly 4% of users mentioned changes to their periods, and many described chills and hot flushes, symptoms not prominent in the drugs' trials or official information. [14]

The work was published in Nature Health. [14] The authors say it shows an association and cannot prove the drugs caused anything. [14]

Trials are built to catch the most dangerous effects in a few thousand people, so what bothers patients once millions take a drug arrives later and from the patients themselves. [14]

One of the authors says online groups work like a neighbourhood grapevine: people swap notes with each other rather than telling a doctor or a regulator. [14]

Researchers at Virginia Tech engineered a harmless gut bacterium, E coli, to carry and release viruses that kill Salmonella Typhimurium, a bacterium that causes food poisoning. [12][13]

Viruses that infect bacteria are called phages, and they do not infect people or animals. [13]

Mice given the engineered bacterium before infection kept the salmonella from taking hold and survived more often. [12]

The team also inserted a salmonella gene so the target bacterium would not recognise the virus as foreign and block it. [13]

Phage treatments usually arrive after an infection is established, by which time the bacteria are hidden in the gut lining and out of reach. [13]

Putting the phage in first is a different idea, and it has only been shown in mice.

Sling Therapeutics raised $123m to run a final-stage trial of linsitinib, a pill for thyroid eye disease, in which the tissue behind the eye swells and pushes the eyeball forward. [10]

The first patients have been dosed in a 120-person study measuring the bulge over 24 weeks. [10]

The drug blocks a receptor for a growth signal that sits on cells behind the eye and on immune cells. [10][24]

Astellas first developed it as a cancer drug before Sling licensed it in 2022. [10]

The approved treatments for the disease are antibodies given through a drip, and Sling's chief executive says a pill could reach patients who refuse infusions. [10]

A growth-signal receptor is involved in many diseases at once, which is why a shelved cancer drug can come back aimed at an eye. [24]

The Mayo Clinic, a large US hospital group, and Thermo Fisher Scientific, which makes laboratory equipment, have formed a company called Precure to collect molecular and clinical data from people before a disease is diagnosed. [15]

The aim is to find the earliest biological signals of cancer, heart and metabolic disease, brain disorders and immune disease. [15]

Mayo says any money it earns will go back into patient care, research and teaching. [15]

Diagnosis today mostly starts when somebody feels ill, and the years before that are the part nobody has measured at scale. [15]

A hospital group holds the patients and the records; an equipment maker holds the machines, and neither can do this alone.

Samsung Biologics, which makes medicines under contract for other companies, signed a $262m agreement with an unnamed European drugmaker, running at its Songdo site in South Korea until 2033. [16]

The company says its cumulative contract value now passes $21.9bn. [16] It operates 845,000 litres of manufacturing capacity in Korea and the United States, including a Maryland plant bought in March. [16]

It has also made an all-cash offer for PolyPeptide Group to move into peptide medicines. [16]

Most new medicines are not made by the company whose name is on the box. [16]

When drugmakers cut their own programmes, the factories they rent keep filling with someone else's products.

Zambia's disaster management unit says 46 hippos, eight elephants, five buffaloes and a crocodile have died of anthrax in the Munyamadzi game area since July. [17]

Anthrax is caused by a bacterium whose spores survive in soil for years and mainly kills grazing animals; people catch it from infected animal products. [17]

Twelve people are being treated and nobody has died. [17] The national coordinator said local people had eaten the carcasses because that is what they depend on for food. [17]

Zambia's 2023 outbreak reached nine of its ten provinces and killed four people, and in 2011 five died after eating hippo meat. [17]

The advice not to eat animals that died suddenly has to be followed by people who have little else to eat.

Researchers at the Sylvester Comprehensive Cancer Center in Miami report in JCI Insight that blocking a receptor called IL1RAP breaks up the network of cells that shields pancreatic tumours from treatment. [18]

IL1RAP is a shared relay that several inflammatory signals use, so blocking it interferes with many messages at once. [18]

In animal and laboratory experiments the tumours had fewer immune-suppressing cells, less scar tissue and more active cancer-killing T cells. [18]

A trial is planned in patients whose tumours can still be removed, giving the treatment before surgery. [18]

Pancreatic tumours survive partly because of the cells around them, which are inflamed and yet hold the immune system back. [18]

These results are preclinical, and the planned trial would be the first test in people.

A neural network trained on heart traces from 173,396 hospital patients was tested on 4,370 more with suspected heart attacks, in a study from University Hospital Basel. [19]

For the type of heart attack that shows a clear signature on the trace, it scored 0.96 on a standard accuracy measure against 0.89 for doctors reading the same traces. [19]

On the harder type, without that signature, it reached 0.81, and it often missed one blocked-artery subtype. [19]

A standard heart trace takes minutes and is available in any emergency department, so the test is not the scarce part. [19]

Reading it quickly and correctly is, and that is what the model does well for one kind of attack and less well for another.

RSV is a virus that is a leading cause of babies being admitted to hospital.

Researchers measured how nearly every possible mutation in one of its surface proteins affects the antibodies given to protect infants. [20]

Nirsevimab, the most-used of those antibodies, prevents roughly 75 to 80% of RSV hospital admissions. [20]

The work explains why resistance shows up more often in one of the virus's two subtypes and builds a list to check new samples against. [20]

Resistant strains exist but are rare so far. [20]

One earlier antibody failed its final trial because circulating strains had escaped it. [20] Knowing in advance which mutations matter turns surveillance from guesswork into a lookup.

A team in Taiwan reports that acrolein, a reactive chemical that builds up under stress inside cells, changes a protein called thioredoxin 1 in neutrophils, the body's most common white blood cells. [21]

The altered cells switch to a state that suppresses the immune system rather than attacking the tumour, and they block the T cells that kill cancer. [21]

In mice, blocking the pathway restored T cells inside the tumour and slowed the growth of glioblastoma, the most aggressive common brain cancer. [21]

Patients with more of these cells lived less long. [21]

Glioblastoma resists treatment partly because it turns the immune cells around it into guards for the tumour. [21]

This is a mouse and laboratory result, and no drug against the pathway has been tested in people.

The US Food and Drug Administration has allowed Medexus Pharma to import Hexatrione, a steroid injection for joints made in France by Ethypharm, because the US-approved equivalent, Aristospan, is in short supply. [22]

The imported product is not approved in the United States and no other company is allowed to bring it in. [22]

It comes in a snap-open glass ampoule rather than a vial, and it carries no barcode, so hospitals must enter it into their systems by hand. [22]

A shortage of a decades-old cheap steroid is fixed by importing a foreign version and asking staff to work around the differences. [22]

The missing barcode is the kind of detail that causes medication errors, which is why the letter spells out the alternative checks. [22]

Chris Klomp, nominated as deputy secretary of the US Department of Health and Human Services, was questioned over two days by two Senate committees about the power he already holds. [23]

Klomp currently runs Medicare, the US health insurance programme for older people, and helped negotiate the deals the Trump administration struck with drugmakers over prices. [23]

He has also run searches for senior appointments. [23] Senators focused on how he would use that authority as second-in-command to health secretary Robert F Kennedy Jr. [23]

Klomp already runs the health insurance programme for older Americans and helped make the price deals with drugmakers. [23]

The job he is up for would add staffing and the department's day-to-day running to that, under a secretary who has rewritten US vaccine policy. [23]