One Pill, Fifteen Years, and Thin Evidence · Morning Gist
Saturday, October 245 stories · 15 min
One Pill, Fifteen Years, and Thin Evidence
Today: a lung cancer outlier, a Medicare fight over thinly tested pain treatments, an APOE4 myelin finding, a new diet book and a vaccine-safety credibility dispute.
Picked from 39 new articles across 12 sources · arrow keys for other days
Today in 30 seconds
30s left
Janet Freeman-Daily marks 15 years of lung cancer control on crizotinib; she says gene testing and trial access still decide who benefits. story 1 ↓
Five Medicare contractors covering 37 states limit nerve-block coverage for chronic pain; a physician warns decades-old pelvic treatments will be cut off. story 2 ↓
MIT researchers found APOE4 damages myelin by suppressing Wnt signaling; blocking GSK3b restored it in human cells and mice. Preclinical only. story 3 ↓
Stanford diet-trial researcher Christopher Gardner's new book Food Sense covers protein, seed oils and sugar; Marion Nestle recommends it. story 4 ↓
Neurologist Jonathan Howard challenges NIH director Bhattacharya's new vaccine-injury center, noting he once called a 1-in-500 death risk "very moderate." story 5 ↓
Nigeria has recorded its first case in the Congo-origin Ebola outbreak; officials are checking whether it links to Kenya's first case.
1/5STAT· STAT Staff· Oct 23
Lung cancer patient marks 15 years on one targeted pill, crizotinib
Janet Freeman-Daily, whose lung cancer is driven by a rare ROS1 gene change, has had no detectable disease since starting crizotinib in a 2012 trial. Many fellow patients, including two co-founders of her advocacy group, have died.
Why it matters · It shows what targeted therapy can do at its best, and how much depends on whether patients get gene testing, a knowledgeable doctor and a trial place.
Know this first
Context from the editor, not the article · 25 sec
1 · Driver mutations in lung cancer
Some non-small cell lung cancers grow because of one faulty gene, a 'driver'. A tumor biopsy tested for these drivers (biomarker or molecular testing) can point to a pill that blocks that gene's protein.
2 · ROS1 and crizotinib
In ROS1-positive cancer, the ROS1 gene is broken and fused to another gene, so its growth signal stays on. Crizotinib blocks that signal. The FDA approved it for ROS1-positive lung cancer in 2016.
3 · Targeted drugs rarely cure
Targeted drugs usually control the cancer rather than eliminate it. Patients take them every day, often for years, and tumors usually find ways around the drug over time.
2/5STAT· Antje M. Barreveld· Oct 23
Medicare contractors limit coverage of nerve blocks for chronic pain
Starting in late October, five private contractors that run Medicare across 37 states will stop covering many nerve-block procedures for chronic pain, citing weak evidence. Pain physician Antje Barreveld argues this will hurt patients.
Why it matters · The dispute is a real test of how Medicare should handle treatments that seem to help some patients a lot but have only small, short trials behind them.
Know this first
Context from the editor, not the article · 27 sec
1 · Who decides Medicare coverage
The Centers for Medicare and Medicaid Services (CMS) writes Medicare's rules. Private regional firms called Medicare Administrative Contractors (MACs) process claims, and they can issue local coverage policies that decide which procedures get paid for in their regions.
2 · What a nerve block is
A doctor injects numbing medication around a specific nerve to block pain signals. It can show which nerve is causing the pain (diagnostic), and sometimes it relieves pain for much longer than the drug itself lasts (therapeutic).
3 · Why opioid alternatives matter
Since the opioid epidemic, doctors and regulators have pushed for non-opioid ways to treat chronic pain. Procedures, physical therapy and non-opioid drugs are usually combined, which is called multimodal care.
MIT researchers report that the APOE4 gene variant turns down a signaling pathway called Wnt in the cells that insulate nerve fibers. Fat then piles up in those cells instead of becoming insulation. Restoring the pathway brought the insulation back in lab cultures and mice.
Why it matters · APOE4 is the strongest common genetic risk factor for Alzheimer's, so any new way it harms the brain is a possible drug target. Myelin loss is a mechanism that inflammation-focused approaches miss.
Know this first
Context from the editor, not the article · 29 sec
1 · APOE and Alzheimer's risk
The APOE gene makes a protein that carries fats and cholesterol. It comes in three common versions: ε2, ε3 and ε4. One copy of ε4 raises Alzheimer's risk several-fold and two copies raise it much more. ε3 is the usual reference version.
2 · Myelin and oligodendrocytes
Myelin is a fatty sheath wrapped around axons, the long fibers that carry signals between neurons. It makes signals travel fast. In the brain, cells called oligodendrocytes build and maintain it. Myelin thins somewhat with normal aging.
3 · Wnt and GSK3b
4/5Food Politics (Marion Nestle)· Marion· Oct 23
Marion Nestle recommends Christopher Gardner's new nutrition book, Food Sense
Nutrition writer Marion Nestle recommends Food Sense, a new book by Stanford diet researcher Christopher Gardner. It offers flexible, science-based eating advice and recipes. Her post is a short endorsement, not a review.
Why it matters · Gardner has run some of the best-known head-to-head diet trials, so his take on protein, seed oils and sugar is likely to rest on evidence, not wellness trends.
Know this first
Context from the editor, not the article · 19 sec
1 · Who Christopher Gardner is
A Stanford nutrition scientist who runs randomized diet trials. His best known, DIETFITS, assigned about 600 adults to a healthy low-fat or a healthy low-carb diet for a year. Weight loss was about the same on both.
2 · Who Marion Nestle is
An emerita professor of nutrition at New York University. She writes the Food Politics blog and has spent decades criticizing how the food industry shapes dietary advice and research. Her endorsement counts in the field.
Image: Food Politics (Marion Nestle)
Marion Nestle, a nutrition professor and long-time critic of the food industry, used her weekend-reading post to recommend Food Sense: The Science and Practice of Eating Well (Simon Element, 2026, 287 pages). The author is Christopher Gardner, a Stanford researcher who runs clinical trials comparing diets.
5/5Science-Based Medicine· Jonathan Howard· Oct 23
Neurologist challenges NIH director's new vaccine-injury center, citing his 2020 COVID remarks
In an opinion piece, neurologist Jonathan Howard attacks the new vaccine-injury center at the National Institutes of Health. He argues that vaccine harms are already studied rigorously, and that its director once brushed off far larger COVID risks.
Why it matters · Federal research money and messaging on vaccine safety now come from officials with a record of doubting vaccines. That affects which questions get studied and whether anyone trusts the answers.
Know this first
Context from the editor, not the article · 27 sec
1 · Who runs vaccine policy now
Jay Bhattacharya, a Stanford health economist who co-wrote the 2020 Great Barrington Declaration against broad lockdowns, directs the National Institutes of Health (NIH). Robert F. Kennedy Jr., a longtime vaccine critic, is Health and Human Services secretary and oversees the NIH.
2 · How vaccine harms get caught
The US watches for vaccine side effects through reporting systems, large health-record databases and follow-up studies. Since 1986, a no-fault federal program has paid compensation for recognized vaccine injuries.
Freeman-Daily (second from right) with early members of the ROS1 patient group that became The ROS1ders.
Janet Freeman-Daily has had advanced lung cancer for 15 years. She has taken a daily pill, 1, since November 2012, and every scan since then has shown no evidence of disease. At last week's STAT Summit in Boston she explained how she got the drug and why most patients like her don't do as well. STAT published her prepared remarks. They are one person's account, not trial data.
A rare driver, found almost by chance
She was diagnosed in 2011, at 55, with no known risk factors. Chemotherapy and radiation failed. Molecular testing, which checks a tumor for gene changes a drug might target, came back negative. The cancer spread to her other lung.
A member of an online patient forum told her she fit the profile for a 2. In this change the ROS1 gene breaks and fuses with another gene, which leaves a growth signal switched on. It drives about 1% to 2% of 3. Her oncologist had never heard of testing for it and expected her to be on chemotherapy for the rest of a short life.
On a trip to Denver she dropped by the cancer center that had tested her tumor. The lead researcher there had since built a ROS1 test, and some of her tissue was still on file. The result was "wildly positive", and he offered her the last place in a crizotinib trial. Crizotinib is a 4: it blocks the overactive ROS1 protein. It controls her cancer but hasn't cured it, so she has to keep taking it.
An outlier, not the norm
The article gives no trial figures. From general published background: in the trials behind crizotinib's approval, the cancer typically stayed controlled for about a year and a half before resistance set in. Her 13-plus years of control make her an 5.
In 2015 she and four other ROS1 patients, all on crizotinib with no options left but chemotherapy, started what became The ROS1ders, a nonprofit for patients and caregivers. In 2017 the group helped create 12 6 (cancer cells grown from patients' tumors) that researchers now use to study drug resistance and develop new drugs. Two of the five founders, and many members, have since died.
The cost of surviving
Living on the drug isn't easy. She has had two dose reductions, takes other medicines for side effects, and is on a 7 weight-loss drug to shed 70 pounds she gained on crizotinib. Freeman-Daily argues that research tracks survival but neglects survivors' quality of life. She wants more trials to use 8, which are measures that patients rate themselves, and wants drug sponsors to "own their toxicities".
Janet Freeman-Daily has taken 1, a daily pill, since November 6, 2012. Every scan since then has shown no evidence of disease. At the STAT Summit in Boston last week she described how she got the drug and why her experience is the exception. STAT published her prepared remarks. They are one patient's account, not a study, and should be read that way.
The problem: a rare subtype that few doctors knew to test for
Freeman-Daily was a former aerospace engineer. She was diagnosed with advanced lung cancer in 2011, at 55, with no known risk factors. At the time, she notes, most people with lung cancer died within a year of diagnosis. Her doctors first thought the cancer could be cured and treated it with chemotherapy plus radiation. During a break to let her blood counts recover, a 3-inch mass grew on her collarbone.
She arranged molecular testing through a clinical trial. Molecular testing looks for specific gene changes that drive a tumor and that a drug might block. Every test came back negative. After more chemotherapy and radiation, a scan showed the cancer had spread to her other lung.
The lead she needed came from an online patient forum. Another member, who was in a trial in Boston, told her she fit the profile of patients whose cancer is driven by a 2. In this change the ROS1 gene breaks and fuses with another gene, which leaves a growth signal stuck on. It occurs in about 1% to 2% of 3, the most common type. Her local oncologist had not heard of ROS1 testing. He told her she would be on chemotherapy for the rest of her life, and that it would probably be short.
How she got the drug
Getting the right test came down to chance. While visiting family in Denver, she stopped by the cancer center that had done her earlier molecular testing to say thank you. The researcher who led that trial told her his lab had since developed and validated a ROS1 test, and that some of her tumor tissue was still on file. She let them test the leftover slides. Weeks later he called: her tumor was "wildly positive" for ROS1, and they were holding the last slot in their ROS1 crizotinib trial for her.
Crizotinib is a 4. It blocks the enzyme activity of the overactive ROS1 fusion protein and cuts off the growth signal. A rough analogy: if the fusion is a jammed accelerator, the drug wedges a block under the pedal. It works only while it stays in place, which is why Freeman-Daily writes that she is "not cured" and must keep taking the drug.
How unusual is 15 years?
The article gives no trial statistics, so this section uses general published background. In the early-phase trial that led to crizotinib's 2016 FDA approval for ROS1-positive lung cancer, roughly 70% of patients' tumors shrank. The median time before the cancer progressed was about a year and a half. Most patients eventually develop resistance, often through new mutations in ROS1 itself. Newer ROS1 drugs have been approved since, partly to treat cancers that have become resistant to crizotinib. More than 13 years of continuous control on the first drug puts Freeman-Daily at the far end of the response curve. She is an "5", and her case cannot show what a typical patient should expect.
She says as much: not all patients are offered biomarker testing, are treated by doctors who know their subtype, hear about trials or can reach them, respond to treatment, or respond for long.
What patients built
In 2015, five people with ROS1-positive cancer met at an advocacy event. All were on crizotinib, all had children at home, and none had any option left except chemotherapy. They started a global Facebook group, which became The ROS1ders, a nonprofit of patients and caregivers focused on community, education and research.
In 2017 the group co-created the ROS1 Cancer Model Project with nonprofits and researchers. At the time only a few lab models of ROS1-positive cancer existed. The project produced 12 new 6, which are cancer cells grown from patients' tumors that labs can study. These have been widely shared and used to study acquired resistance (how tumors learn to evade the drug), surface-omics (cataloguing the proteins on a cancer cell's surface, which can be drug targets) and drug development. The group also runs the PRIME Summit to teach patients about ROS1 treatment and research.
The cost of living on a drug for years
Two of the five founders and many members have died. Long-term survival also has costs that survival statistics miss. Freeman-Daily has had two dose reductions of her TKI and takes other prescriptions to manage side effects. She gained 70 pounds on crizotinib and now takes a 7 drug, the class that includes semaglutide, to lose it.
She argues that research tracks survival closely but pays little attention to survivors' quality of life. She calls for:
bringing the patient perspective into research early
using 8, meaning symptoms and function rated by patients themselves, in more trials
getting sponsors to "own their toxicities", which in practice means studying and disclosing the long-term side effects of their drugs
What the piece leaves open
The essay doesn't explain why her response has lasted so long when others' did not, and nobody may know. It gives no figures on how often ROS1 testing is now done or how widely patients can reach trials. Her case shows that a matched targeted drug can turn a cancer expected to kill within a year into a chronic, managed condition. It also shows that whether a patient gets there can depend on an online forum and a chance visit to Denver.
Remember
Rare drivers such as ROS1 (1–2% of non-small cell lung cancers) are found only if the tumor is tested for them.
Fifteen years of control on crizotinib is an outlier; resistance usually develops much sooner.
Long-term targeted therapy brings lasting side effects, such as weight gain and dose cuts, that survival figures don't capture.
Starting in late October, five 1 will restrict coverage for many peripheral nerve procedures used to diagnose and treat chronic pain. MACs are the private companies that process Medicare claims and apply its rules, and these five oversee 37 states and several U.S. territories. In a STAT opinion piece, Antje Barreveld, chief of pain medicine at Newton-Wellesley Hospital and immediate past president of the American Academy of Pain Medicine, argues the change will take effective relief away from some of the hardest-to-treat patients.
What's changing
A 2 is an injection of numbing medication around a nerve to interrupt its pain signals. It can help locate where pain is coming from, and for some patients it brings relief that lasts long after the few hours of numbness are over.
The final policy is somewhat looser than the 2025 proposal. Barreveld credits that to pushback from patients, doctors and advocates. Even so, the final version drops coverage for therapeutic blocks of the 3 and the 4, which are used for some kinds of severe pelvic pain. She describes patients for whom sitting, putting on pants or having sex was excruciating until a block helped. One woman whose pain began after gynecologic surgery will lose access to that treatment within weeks.
The contractors' case
Barreveld calls the contractors' concerns legitimate. They point to small studies, short follow-up and uncertainty about whether the benefits last. For several procedures, they concluded the evidence was too thin to support coverage. She also grants that questions about overuse and financial incentives are fair, and that Medicare shouldn't keep paying for blocks that don't help.
Her counterargument
Pain is subjective, she argues, so a treatment with a modest average effect across a trial group can still help particular individuals a great deal. Population data can miss which patients get a 5. With about 1 in 4 U.S. adults living with ongoing pain, she says even weeks or months of relief count. She also sees nerve blocks as part of 6 care: they help patients do physical therapy, use fewer medications, stay out of the ER and sometimes avoid surgery. Cutting them, she warns, ignores a lesson of the opioid epidemic, which is that people in pain need alternatives to opioids.
What she wants
She calls for larger studies with longer follow-up, federal funding to pay for them, and coverage standards that keep access open while that evidence builds. She also wants contractors to revisit the restrictions as new data come in.
Caveat
This is an advocacy piece. It relies on patient stories and doesn't cite trial data, and it concedes that the evidence base is weak.
Starting in late October, new coverage policies from five regional 1 will limit payment for many peripheral nerve procedures used to diagnose and treat chronic pain. MACs are private companies that run much of Medicare's day-to-day machinery. Together these five oversee 37 states and several U.S. territories. In an opinion piece for STAT, Antje Barreveld, a pain physician, argues the change will cut off treatments that have been used for decades. She is chief of pain medicine at Newton-Wellesley Hospital (Mass General Brigham), an associate professor at Tufts and immediate past president of the American Academy of Pain Medicine.
How Medicare coverage gets decided
The 2 writes and oversees Medicare's rules. Medicare handles too many claims to process centrally, so private contractors apply those rules to individual providers and claims. They can also issue regional policies on what gets covered. That's why a decision by five contractors, rather than a national rule, can affect millions of patients across most of the country.
What a nerve block does
A 3 delivers numbing medication around a nerve to interrupt pain signals. It has two uses. As a diagnostic tool, it tests whether a particular nerve is the source of pain: if numbing it stops the pain, you've probably found the culprit. As a therapy, it sometimes gives relief that lasts well beyond the few hours the anesthetic is active. The article doesn't explain why that longer relief happens.
What the final policy excludes
The final policy is somewhat less restrictive than the version proposed in 2025. Barreveld calls that an important response to patients, physicians and advocates who objected. But it still excludes therapeutic procedures on nerves such as the 4 and the 5. The pudendal nerve supplies the pelvic floor and genital area. The ganglion impar is a cluster of nerve cells near the tailbone. Both are targets for some forms of severe pelvic pain.
Barreveld describes the patients involved. One had pelvic pain she compared to "a hot poker," and it went away after a targeted block. Another developed chronic pain after gynecologic surgery. After a pudendal block, she said she could wear pants, sit in a chair and get through the day without being consumed by pain. Barreveld writes that this patient will lose access to the treatment within weeks. She also mentions an 87-year-old with severe bladder pain who had been passed from doctor to doctor. These are anecdotes. They show what is at stake for individual patients, but they say nothing about how often the procedures work.
The contractors' reasoning
Barreveld is direct that the contractors have a legitimate concern. In reviewing these procedures, they cited small studies, short follow-up periods and uncertainty about whether benefits last. For several procedures they judged the evidence insufficient to support coverage. She also calls fair the questions about whether the procedures are overused, whether the evidence is strong enough, and whether financial incentives drive their use. "Not every nerve block works," she writes, "and Medicare should not pay indefinitely for treatments that do not help."
Her argument: averages can hide responders
Her main point is about how evidence is gathered. Pain is subjective by definition, and patients experience it differently. So a treatment that shows limited benefit on average across a study population may still give profound relief to some individuals.
A purely illustrative example (not from the article) shows how this works. Suppose a trial of 100 patients finds a small average drop in pain scores. That result could mean everyone improved a little. It could also mean 20 patients improved dramatically and 80 barely changed. The averages look the same, but the second case suggests a group of strong responders that a better-designed study might be able to identify in advance. Barreveld argues that population-level evidence matters but can't always show who gets a 6. Her article doesn't cite data showing that such a responder group exists for these particular blocks.
She makes two more arguments. First, about 1 in 4 U.S. adults live with unrelenting pain, and for them, she says, even a few weeks or months of relief is meaningful. Second, nerve blocks are rarely used alone. They are part of 7, multidisciplinary care. In her account, they help patients take part in physical therapy, regain function, cut back on medications, avoid ER visits and hospital stays, and sometimes delay or avoid surgery. Restricting them, she argues, forgets a central lesson of the opioid epidemic: people with chronic pain need alternatives to opioids and treatment tailored to them.
What she proposes
Barreveld doesn't ask Medicare to pay without conditions. She calls for:
larger studies with longer-term patient follow-up;
federal funding to support that research;
coverage standards that recognize clinically meaningful benefit while still demanding better evidence;
contractors reconsidering the restrictions as new evidence emerges;
researchers, clinicians and specialty societies building and finding funding for the studies Medicare says it needs.
"Absence of perfect evidence should not be confused with absence of meaningful patient benefit," she writes.
How to weigh it
This is an advocacy piece by a leading figure in pain medicine. Her specialty performs these procedures, a potential conflict she raises herself in general terms. The article gives no effect sizes, trial results or utilization figures, and it doesn't say exactly which procedures stay covered. It also doesn't say whether the contractors left any route open for case-by-case approval. What it does show is a familiar tension in coverage policy. Requiring trial evidence protects patients and taxpayers from paying for ineffective care. But for conditions that vary widely between patients and have never been studied well, that requirement can also block access to treatments that some patients say have changed their lives. Barreveld's proposed middle path is to keep coverage open while funding better trials. Whether that works depends on someone actually paying for those trials.
Remember
The new restrictions come from regional contractors, not a national CMS rule, yet they cover 37 states.
Even the author agrees the evidence is weak: small trials, short follow-up, uncertain durability.
Her core claim is that average trial results can hide individual patients who respond strongly.
Wnt is a signaling pathway that controls cell growth and development. In it, the protein beta-catenin carries the signal forward. The enzyme GSK3b tags beta-catenin for destruction, which keeps the pathway switched off.
Image: Fight Aging!
The finding
MIT researchers report a new way the 1 gene variant, the biggest common genetic risk factor for Alzheimer's, may damage the brain. It disrupts 2, the fatty insulation around nerve fibers.
The cells that make myelin are called 3. In these cells, APOE4 suppresses a signaling pathway called 4. With too little Wnt signaling, the cells store fat as lipid droplets instead of using it to build myelin. When the team restored Wnt signaling, myelination recovered in human cell cultures and in mice used to model Alzheimer's.
Why it's notable
APOE makes a protein that carries fats. Most work on APOE4 has focused on how it fuels inflammation in 5, the brain's immune cells. This study points to a separate problem with how fat is handled, inside a different cell type.
Myelin already thins somewhat with normal aging, and that loss is thought to weaken communication between neurons. Reason, who writes Fight Aging!, argues it is reasonable to suspect that APOE4 makes this worse and so speeds neurodegeneration. That is a hypothesis, not a result.
How they found it
The team ran a 6 screen. CRISPR is a gene-editing tool, and here it was used to switch off every protein-coding gene, one at a time, in human APOE4 oligodendrocytes. Switching off genes that boost Wnt increased lipid droplets. Switching off genes that suppress Wnt reduced them.
The cause was an imbalance between two proteins. Beta-catenin drives Wnt signaling, and the enzyme 7 shuts it down. Compared with cells carrying the common APOE3 version, APOE4 cells had less beta-catenin and too much GSK3b.
Testing it
Boosting GSK3b with a chemical called Dif3 gave normal APOE3 cells more lipid droplets.
Blocking GSK3b with a lab inhibitor, CHIR99021, reduced droplets in APOE4 cells.
In 8 (small 3D tissue cultures) combining APOE4 oligodendrocytes with neurons, blocking GSK3b increased myelination and axon growth.
Caveats
All of this is preclinical. The summary gives no effect sizes and does not say whether the mice's memory or behavior improved. GSK3b inhibitors have been tested in Alzheimer's before, mainly to target tau, the protein that forms tangles in Alzheimer's brains. Results were disappointing (this is general background, not from the article). Turning up Wnt signaling broadly also carries risks, because the pathway drives cell growth. The researchers call GSK3b inhibition a proof of concept, not a treatment.
The finding
Researchers at MIT's Picower Institute report a second way the 1 gene variant may harm the brain. APOE4 is the strongest common genetic risk factor for Alzheimer's disease. In 2, the brain cells that make 3 (the fatty insulation around nerve fibers), APOE4 suppresses a signaling pathway called 4. Without enough Wnt signaling, the cells store fat as droplets instead of using it to build and maintain myelin. When the team restored Wnt signaling, myelination recovered. This held in human cell cultures and in mice used to model Alzheimer's. The write-up comes from Reason at Fight Aging!, summarizing MIT's press release.
Why look at myelin at all
APOE makes a protein that moves lipids (fats and cholesterol) around the body and brain. Most research on how APOE4 raises Alzheimer's risk has focused on 5, the brain's resident immune cells. In microglia, APOE4 is linked to lipid problems that fuel inflammation. That has made inflammation the leading explanation for APOE4's risk.
Myelin is another place where lipid handling matters a great deal. Myelin is mostly fat. Oligodendrocytes have to keep supplying and repairing it so signals can travel quickly along axons, the long fibers that wire neurons into networks. Myelin maintenance already slips somewhat in normal aging, and that loss is thought to contribute to age-related decline in brain function. Reason argues that if APOE4 speeds up this loss, it is reasonable to suspect it adds to faster neurodegeneration. He presents this as a hypothesis, not a result.
How they found Wnt
The key tool was a genome-wide 6 screen. CRISPR is a gene-editing method, and here it was used to switch off each protein-coding gene, one at a time, in cultured human oligodendrocytes carrying APOE4. The researchers then checked which knockouts changed the amount of lipid droplets. The results fit together in a consistent way:
Knocking out genes that promote Wnt activity increased lipid droplets.
Knocking out genes that suppress Wnt activity reduced them.
When knockouts on both sides of a pathway push the outcome in opposite directions, that is strong evidence the pathway controls the outcome. Here it pointed to Wnt as a regulator of whether oligodendrocytes store fat or use it.
The mechanism: a tug-of-war inside the cell
Wnt signaling depends on two opposing proteins. Beta-catenin carries the signal forward. 7, an enzyme, marks beta-catenin for destruction and so holds the pathway down. A rough analogy: beta-catenin is the accelerator and GSK3b is the brake. Compared with oligodendrocytes carrying the common APOE3 version, APOE4 cells had less beta-catenin and too much GSK3b. In other words, a weaker accelerator and a stronger brake, which together explain the low Wnt signaling.
Testing the brake both ways
The team then changed GSK3b activity in each direction:
In normal APOE3 oligodendrocytes, boosting GSK3b with a chemical called Dif3 increased lipid droplets. This made healthy cells look like APOE4 cells.
In APOE4 oligodendrocytes, blocking GSK3b with CHIR99021, a standard lab inhibitor, reduced lipid droplets.
In 8 (small 3D tissue cultures) that combined APOE4 oligodendrocytes with neurons, blocking GSK3b and .
The authors describe GSK3b inhibition as a proof-of-concept therapeutic strategy.
What the summary doesn't tell us
The article gives no effect sizes. It does not say how much myelin was restored, how many lipid droplets disappeared, or how big the mouse groups were. It does not name the journal or describe the mouse experiments in detail, and it reports no effects on memory or behavior. Everything here is preclinical, from cell cultures, organoids and mice. Nothing yet shows that this pathway drives myelin loss in human APOE4 carriers, or how much it adds to their risk compared with the inflammatory route through microglia.
Caveats on the drug angle
This part is general background, not from the article. GSK3b inhibitors are not new to Alzheimer's research. The enzyme has long been studied for its role in tau pathology, and inhibitors such as tideglusib, and lithium (which partly inhibits GSK3b), have been tested in people with modest or disappointing results. Wnt drives cell growth, so turning it up broadly carries theoretical risks, including cancer. CHIR99021 is a research tool, not a medicine. Any treatment would probably need to target oligodendrocytes, be dosed carefully, or be given at a particular stage of disease.
Open questions
Does restoring myelin in APOE4 mice improve memory or slow decline?
Does the same Wnt and GSK3b imbalance appear in brain tissue from human APOE4 carriers?
How does APOE4 tip the balance between beta-catenin and GSK3b in the first place?
Could imaging that measures myelin pick out APOE4 carriers who might benefit early?
Remember
APOE4 may damage the brain through myelin loss as well as through microglia-driven inflammation.
Opposite results from blocking and boosting GSK3b strongly suggest that GSK3b's control of Wnt signaling drives the fat buildup.
The work is preclinical with no reported effect sizes, and earlier human trials of GSK3b inhibitors were disappointing.
Readers should know two things up front. Nestle wrote a promotional blurb for the book, and the two authors had lunch together after book events in Marin County. Her post is an endorsement, not a critical review. It doesn't assess any claims, cite any studies or report any data from the book.
What the book argues
Her blurb calls it a "remarkably sensible and easy-to-read philosophy" of eating well and enjoying food. In her account it draws on Gardner's scientific training, his experience running trials and his cooking, since the book includes recipes.
In the excerpt she quotes, Gardner blames the confusion about diet on the food system. He says it is "really powerful and entrenched" and no longer serves the public's interests. He argues that pulling between taste, cost and convenience makes even simple advice seem complicated, and that this leaves room for people selling "miracle cures and quick fixes." He wants an approach with "a strong foundation based in science with lots of room for flexibility."
The topics
The chapter titles Nestle lists show where he lands on several current arguments:
The Diet Wars
Yes, You Probably Get Enough Protein. This pushes back on the current enthusiasm for eating more protein.
1. He defends oils like canola and soybean, which some influencers blame for chronic disease.
Plant-Based Meats—Instead of What? Whether they help depends on what they replace.
Sugar Is Sugar, and Less Is Always Better. This rejects the idea that "natural" sugars are healthier.
What to make of it
These positions largely match mainstream nutrition science. Gardner's own trials have found that diets of very different types can work about equally well when food quality is high. The post doesn't tell you how he backs up each chapter, or how firmly he separates trial evidence from observational studies. To judge that, you'd need to read the book.
Marion Nestle, a nutrition professor and long-time critic of the food industry, used her weekend-reading post to recommend a new book: Food Sense: The Science and Practice of Eating Well by Christopher Gardner (Simon Element, 2026, 287 pages). The post is short. Most of it is Nestle's own promotional blurb, one excerpt and a handful of chapter titles, so this write-up is short too.
Who's involved, and the conflict to note
Gardner is a nutrition scientist at Stanford who runs 1 of diets. Nestle is an emerita professor at New York University and one of the field's best-known voices on how industry shapes food advice.
Their relationship is out in the open. Nestle wrote the blurb printed on the book. The post also mentions that the two held book events at the same time a couple of miles apart in Marin County, California: hers at Book Passage in Corte Madera, his at Copperfields in Larkspur. They had lunch together afterward. So treat this as a colleague's endorsement, not an independent review. The post doesn't evaluate the book's claims, cite its references or report any data from it.
What Nestle says about it
Her blurb calls the book a "remarkably sensible and easy-to-read philosophy of how to eat healthfully and enjoyably." She says it rests on three things: Gardner's "extensive and deep scientific training," his "experience running clinical trials," and his cooking skills. The book includes recipes. She says he uses this background to "weigh in on your most perplexing questions about diet and health." She closes the post with "It's a great read."
Gardner's argument, in his words
The excerpt (pp. 6–7) gives the book's framing. Gardner says people care so much about diet because for decades they've been told society needs to eat better. They are, in his words, "up against a really powerful and entrenched food system that may have started with good intentions but hasn't had our best interests at heart for some time."
He names three pressures that make healthy eating hard: taste, cost and convenience. He argues that pulling between them makes "even simple and straightforward nutrition guidance seem complicated." That confusion, he writes, is "a perfect environment for people to try to sell you miracle cures and quick fixes."
His answer is an approach that is "louder than the messages you're getting from, well, everywhere." It should also provide "a strong foundation based in science with lots of room for flexibility." Flexibility is the key word. The book seems to give principles rather than one prescribed diet.
The chapter titles, and the debates behind them
Nestle lists five chapters. Each takes a side in a live public argument. The background below is general context, not material from the book.
How to weigh this
Judging by the titles, Gardner's positions fit mainstream nutrition science. They also fit the view that food quality matters more than which label your diet carries. That makes the book a likely antidote to the hype this digest avoids.
The post leaves several things open. It doesn't show how each chapter is argued or sourced. It doesn't say whether Gardner is careful to separate randomized-trial evidence from 5, which is where many nutrition claims fall apart. And it doesn't say whether the book has new analysis or mainly summarizes existing evidence for a general audience. Nestle's endorsement says the book is sensible and readable. To know how strong each recommendation is, you'd need to read the book.
Remember
Nestle wrote the book's blurb, so her post is a friendly endorsement, not an independent review.
The chapter titles push back on seed-oil fears, the high-protein fad and the idea that 'natural' sugar is healthier.
The book offers flexible principles and recipes, not a single prescribed diet.
A "1 in 500" risk means 2 deaths per 1,000 people. Comparing risks fairly means using the same scale for every harm, whether it comes from an infection or from a vaccine.
Image: Science-Based Medicine
The 1, the main US funder of medical research, has opened a new Vaccinology and Immunology Center. Its director, Jay Bhattacharya, says people who report health problems after vaccination, or long-lasting symptoms after an infection, deserve to be "evaluated with the best tools science can offer." On 2, a blog run by physicians, neurologist Jonathan Howard argues against the idea.
This is an opinion piece with no new data. It also doesn't describe what the center will do, what it will cost, or how it will run its studies.
Claim one: vaccine harms are already studied
Howard draws on pediatrician and vaccine researcher Paul Offit. Offit rejects the claim by Health Secretary Robert F. Kennedy Jr. that people reporting vaccine injuries have been "gaslit," and lists rare harms that researchers found and confirmed:
the 1976 swine flu vaccine and 3, a paralysis that climbs up from the legs
an early rotavirus vaccine and 4, a bowel blockage
a European swine flu vaccine and 5, a lifelong sleep disorder
Johnson & Johnson's COVID vaccine and rare, sometimes fatal blood clots
the mRNA COVID vaccines from Pfizer and Moderna and 6, inflammation of the heart muscle
The same methods can also clear a vaccine. After a 1998 paper claimed the measles-mumps-rubella (MMR) vaccine caused autism, more than 24 studies found no link. Howard agrees with Offit. Given what he calls Kennedy's and Bhattacharya's record of anti-vaccine disinformation, he says no reasonable person will trust what the new center finds.
Claim two: Bhattacharya's own yardstick
In October 2020, a writer asked Bhattacharya about COVID's risk to teachers. He took the example of a 40-year-old, ignored all non-fatal harms, and called a 1-in-500 risk of death "very moderate." That is 2 deaths per 1,000 people. Every vaccine harm on Offit's list is described as rare, so far less common. Howard's jab: by Bhattacharya's own standard, any side effect less deadly than 1 in 500 should be shrugged off too.
How to read it
The historical record is the useful part. It shows the existing safety system catching real problems and ruling out false ones. The 1-in-500 argument is about consistency, not evidence. The article doesn't say where the figure came from, and one person's inconsistency about risk doesn't settle whether a research center is worth funding. The questions that matter for patients are still open: what the center will study, who will review its methods, and whether its findings will shape vaccine advice or compensation.
The 1, the main US funder of medical research, has announced a new Vaccinology and Immunology Center. Its director, Jay Bhattacharya, says people who report health problems after vaccination, or long-lasting symptoms after an infection, "deserve to be heard, taken seriously, and evaluated with the best tools science can offer." On 2, a blog run by physicians who critique medical claims, neurologist and psychiatrist Jonathan Howard argues that the center fixes a problem that doesn't exist, and that its leadership has no standing to run it.
This is an opinion piece, not a study. It reports no new data. Its value is in the history it gathers and in the consistency argument it makes.
What the NIH says it is doing
The article quotes only part of Bhattacharya's statement. In that part he cites the 3, which Congress set up in 1986. The law accepted that vaccines help many people and that, like any medical treatment, they can harm some. The article does not say what the new center will actually do: its budget, staff, research plans, or whether it will see patients or only fund studies. Readers should keep that gap in mind.
Howard's first argument: vaccine harms are already taken seriously
Howard rejects the idea that people reporting vaccine injuries have been ignored. He draws on pediatrician and vaccine researcher Paul Offit, who answered the "gaslighting" charge from Robert F. Kennedy Jr., the Health and Human Services secretary, with a list of rare harms that researchers found and confirmed:
Offit's point is that the same methods can also clear a vaccine. After a 1998 British paper suggested that the measles-mumps-rubella (MMR) vaccine caused autism, more than 24 studies found no link. Parents' fears, Offit writes, were "taken seriously, and millions of dollars were spent" testing them. Howard calls this "absolutely right." He adds that, given what he calls Kennedy's and Bhattacharya's "abominable track record of anti-vaccine disinformation," no reasonable person will trust what the new center finds.
Howard's second argument: Bhattacharya's own standard
The second argument turns on a quote. In October 2020, writer David Zweig asked Bhattacharya about COVID's risk to teachers. Bhattacharya took the example of a 40-year-old, set aside all of COVID's non-fatal harms, and called a 1-in-500 risk of death "very moderate."
Howard turns that remark into a test. One in 500 means 2 deaths per 1,000 people. Every vaccine harm on Offit's list is described as rare, which means far less common than that. Howard argues that if Bhattacharya applied his own 2020 standard, he would have to dismiss as "very moderate" any vaccine side effect less dangerous than a 1-in-500 chance of dying. Howard admits this works even "in a fantasy world" where Bhattacharya is trustworthy. The argument is about consistency, not about data.
How to weigh it
The historical record is the strongest part of the piece. It shows that the existing safety system, made up of surveillance databases, follow-up studies and a compensation program, has found real harms, some of them only months after a vaccine came out. It has also ruled out claimed harms such as the MMR-autism link.
The consistency argument works as rhetoric, but it has limits. The article gives no source for the 1-in-500 figure or for how Bhattacharya reached it, so readers can't check whether it matched the evidence at the time. A weaker point is that one person being inconsistent about risk doesn't decide whether a research center is worth funding. That depends on what the center studies and how. Howard's charge of disinformation is his characterization. The article doesn't set out the evidence for it.
The piece also leaves open questions that matter more to patients. Will the center study long-term symptoms after infection, such as long COVID, as seriously as reported vaccine injuries? Who will review its methods? Will its results feed into vaccine recommendations or compensation decisions? The article answers none of these.
Remember
Rare vaccine harms, from Guillain-Barré in 1976 to myocarditis in 2021, were found by the existing safety system.
More than 24 studies tested the claimed MMR-autism link and found none. A false claim got serious study too.
The 1-in-500 argument shows inconsistency. It says nothing about what the new center will actually study.
The Diet Wars. This is the long fight over which diet is best: low-fat, low-carb, vegan, paleo and so on. Gardner has tested this directly. His 2 trial randomly assigned about 600 adults to a healthy low-fat or a healthy low-carbohydrate diet for 12 months. Average weight loss was similar in both groups. Both diets stressed whole foods and cutting refined starches and added sugars. The post doesn't say how the book uses this work.
Yes, You Probably Get Enough Protein. Wellness media and food marketing now push high protein intake hard. The title says most people in wealthy countries already eat enough.
3. Some influencers blame oils such as canola, soybean and sunflower for inflammation and chronic disease. The word "slander" shows Gardner thinks the charge is unfounded. Controlled trials that swap these oils in for saturated fat have generally lowered 4, the "bad" cholesterol linked to heart disease.
Plant-Based Meats—Instead of What? The title points to the substitution question. Whether an ultra-processed plant burger helps depends on what it replaces. Swapping it for red meat is different from swapping it for beans.
Sugar Is Sugar, and Less Is Always Better. This rejects the idea that honey, agave or "natural" sugars are meaningfully healthier than table sugar. The advice is simply to eat less sugar of any kind.
1976: the US swine flu vaccine was linked to 4, a paralysis that starts in the legs and climbs upward.
About 2000: an early rotavirus vaccine was linked to 5, where part of the bowel folds into itself and blocks it.
2010: a swine flu vaccine used in Europe was shown to cause 6, a permanent disorder of sleep and wakefulness.
2021: Johnson & Johnson's COVID vaccine was linked to rare and sometimes fatal blood clots, including clots in the brain. It uses an 7, a harmless virus that carries genetic instructions into cells.
2021: the Pfizer and Moderna mRNA vaccines were linked to rare cases of 8, inflammation of the heart muscle.
