Small Harvard-led trials find that BCG, the classic tuberculosis shot, triggers lasting immune changes in cerebrospinal fluid — and shifts amyloid levels in people without Alzheimer’s pathology
The research was led by Dr. Mahesh Chandra Kodali, a neuroscientist originally from Hyderabad at Harvard Medical School and Massachusetts General Hospital, as part of a team conducting two linked, one-year, open-label clinical trials in adults 55 and older.
For nearly 100 years, the Bacillus Calmette-Guérin (BCG) vaccine has done one job well: protect against tuberculosis. But a growing body of research suggests this humble, decades-old shot might have a second act — as a tool against Alzheimer’s disease. Now, researchers at Massachusetts General Hospital and Harvard Medical School have produced some of the most direct evidence yet that BCG can reach where few therapies do: the immune cells bathing the human brain and spinal cord.
The findings, published in Communications Medicine, come from two small, open-label clinical trials in older adults, some cognitively healthy and some living with Alzheimer’s-related brain changes. The results are early and far from conclusive — but they offer a tantalizing glimpse of what scientists call “trained immunity” operating inside the central nervous system for the first time in humans, alongside measurable shifts in the hallmark Alzheimer’s protein, amyloid-beta.
An old vaccine, a newer idea
BCG has a strange reputation among immunologists. Beyond guarding against TB, it seems to leave the innate immune system generally sharper — better primed to respond to threats it has never even encountered. This phenomenon, called trained immunity, involves epigenetic reprogramming of immune cells like monocytes, essentially leaving a molecular “memory” that boosts future responses.
That memory effect has been linked, in several retrospective studies, to something unexpected: lower rates of Alzheimer’s disease among bladder cancer patients who received BCG as a routine treatment. Animal studies have echoed this, showing BCG can tamp down neuroinflammation and reduce amyloid buildup in Alzheimer’s-model mice.
The problem was always translation. Just because BCG retrains immune cells in the blood doesn’t mean anything similar happens in the brain — a notoriously protected compartment shielded by the blood-brain barrier. Until now, nobody had looked directly.
Going straight to the source
To find out, the research team recruited 23 adults aged 55 and older — 11 with biomarker evidence of Alzheimer’s pathology and 12 without — through two related year-long trials registered with the FDA. Participants received two intradermal BCG injections a month apart, then were followed for a full year with repeated blood draws and, notably, lumbar punctures to sample cerebrospinal fluid (CSF) directly.
That willingness to repeatedly access CSF — an invasive and uncomfortable procedure — is what sets this study apart. It let the team watch, in real time, how immune cells living in direct contact with the brain and spinal cord responded to a peripheral vaccine.
Using cytokine assays and single-cell RNA sequencing on more than a quarter-million cells, the researchers tracked how monocytes, T cells, and other immune populations in both blood and CSF changed their behavior and gene expression over 12 months.
The brain’s immune cells learned something — but not everything
The central finding: CSF monocytes showed lasting evidence of trained immunity. When exposed months later to a bacterial toxin (LPS) in the lab, these cells mounted stronger, more sophisticated transcriptional responses than they had before vaccination — activating genes tied to cytokine signaling, cellular metabolism, and stress adaptation, with distinct waves of gene activity unfolding across early, middle, and late phases of the year.
Curiously, when researchers exposed the same CSF cells directly to heat-killed BCG — the very organism used in the vaccine — nothing happened. The trained response only showed up with a different, unrelated stimulus. That’s an important and somewhat counterintuitive detail: it suggests the brain’s immune cells weren’t simply remembering BCG itself, but had become generally more reactive and metabolically primed — the hallmark of trained immunity rather than a specific antigen memory. The team speculates this reprogramming could come from BCG-trained immune progenitor cells migrating from the bone marrow into the CNS, or from more direct epigenetic effects on resident immune cells, though the study can’t fully distinguish between these possibilities.
Meanwhile, in blood, the picture looked different — and more strongly shaped by disease status. Peripheral immune cells in people with Alzheimer’s pathology mounted more intense and prolonged inflammatory responses than those without, hinting that BCG’s effects on the body’s immune system play out differently depending on existing brain pathology.
A hint of amyloid movement
Perhaps the most eye-catching result involves amyloid-beta, the protein whose buildup in the brain is a defining feature of Alzheimer’s disease. Among participants without Alzheimer’s pathology, BCG vaccination coincided with a decline in amyloid-beta levels in cerebrospinal fluid, paired with a rise in amyloid-beta in blood plasma — pushing the ratio between the two compartments down by nearly 40 percent over the year.
This pattern is intriguing because it moves in the direction researchers would predict if the brain’s ability to clear amyloid were improving — pushing more of the protein out into peripheral circulation rather than letting it accumulate centrally. Notably, this shift appeared only in participants without established Alzheimer’s pathology. Those with more advanced disease showed no comparable amyloid changes, despite showing stronger inflammatory responses in blood — a disconnect the authors suggest may reflect the way advanced Alzheimer’s pathology can decouple immune activation from actual protein clearance.
The authors are careful — repeatedly and explicitly — to note that this is not proof BCG clears amyloid, slows cognitive decline, or treats Alzheimer’s disease. The study wasn’t designed to test clinical benefit, and the sample size (a handful of participants per subgroup) is far too small to draw firm conclusions. Cognitive test scores, tracked over the same period, showed no meaningful change in participants with Alzheimer’s pathology.
Reasons for both excitement and caution
Several details make this study worth watching rather than a Alzheimer’s breakthrough:
It’s open-label with no placebo group. Every participant knew they were getting BCG, and there was no comparison group receiving a placebo injection — a real limitation for isolating the vaccine’s specific effects from expectation or natural fluctuation.
The sample is tiny. With just 11 and 12 participants in the two comparison groups, the study is powered to generate hypotheses, not confirm them. The authors explicitly describe their statistics as exploratory and unadjusted for multiple comparisons.
Safety looked clean. Encouragingly, BCG was well tolerated across the cohort. The only vaccine-related adverse event was a single mild case of injection-site skin irritation — consistent with BCG’s long track record as one of the most widely used vaccines in the world.
It echoes a growing pattern. The results align with earlier epidemiological work linking BCG treatment in bladder cancer patients to lower dementia rates, and with a separate signal suggesting the shingles vaccine may also reduce dementia risk — hinting that immune system “tune-ups” in general may matter for brain aging, though the underlying mechanisms likely differ.
What comes next
The researchers frame their work explicitly as hypothesis-generating, not a clinical endorsement of BCG for Alzheimer’s prevention or treatment. They call for larger, placebo-controlled trials to determine whether these immune and biomarker shifts are real, reproducible, and — most importantly — translate into any meaningful difference in cognitive outcomes.
Still, the study represents something genuinely novel: the first direct human evidence that a peripheral vaccine can reprogram immune cells living in contact with the brain, without provoking the kind of runaway inflammation that drives neurodegeneration. In an era when Alzheimer’s research has focused heavily on expensive antibody infusions targeting amyloid directly, the idea that a cheap, globally available, century-old vaccine might nudge the immune system toward brain-protective behavior is the kind of finding that — cautiously — could reshape where researchers look next.
This article is based on findings published in Communications Medicine (2026). The study was funded by the National Institutes of Health, the Alzheimer’s Association, the Massachusetts Life Sciences Center, and the Cure Alzheimer’s Fund. One author discloses employment at AbbVie Inc., though the research was conducted independently at Massachusetts General Hospital.
