Reading passage
For more than a century, immunology rested on a tidy distinction. The innate immune system, made up of physical barriers, phagocytes and inflammatory signals, was thought to defend the body quickly but blindly, attacking anything that looked foreign. Memory, the capacity to respond more strongly to a microbe encountered before, was credited exclusively to the adaptive system, whose lymphocytes generate exquisitely specific antibodies and T cells. On this account, only the adaptive arm could learn from experience, and vaccination worked because it educated those lymphocytes in advance.
That dogma has been quietly dismantled over the past two decades. The first clues came from organisms that possess no adaptive immunity at all. Insects and plants, which lack lymphocytes, nevertheless become more resistant to a pathogen after surviving an earlier exposure. Epidemiologists then noticed something stranger in humans: children given the BCG tuberculosis vaccine in West Africa were less likely to die not only of tuberculosis but of unrelated infections such as malaria and respiratory disease. A vaccine designed to train one specific defence appeared to be sharpening many others it was never meant to touch.
Immunologists now call this phenomenon trained immunity. When certain innate cells, notably monocytes and macrophages, encounter a pathogen or a vaccine, they do not simply fight and forget. Their DNA is repackaged through epigenetic modifications, chemical marks on histone proteins that leave particular genes easier to switch on for weeks or months afterwards. The cells also rewire their metabolism, shifting towards aerobic glycolysis and cholesterol synthesis, changes that prime them to release inflammatory molecules more vigorously when challenged again. No antibodies are involved, yet the functional outcome resembles memory.
Because monocytes circulate for only a few days, the durability of trained immunity posed an obvious puzzle. The resolution came from the bone marrow, where hematopoietic stem cells give rise to every immune lineage. Experiments in mice showed that BCG reprograms these stem cells directly, biasing them towards producing myeloid cells with a heightened defensive posture. Researchers therefore distinguish central trained immunity, anchored in the marrow, from peripheral training confined to mature cells in the tissues. The central version can persist for months and perhaps years, which explains how one brief encounter can remodel long term defence.
The discovery has forced a reappraisal of what immune memory means. Classical adaptive memory is antigen specific, slow to develop and capable of lasting a lifetime; trained immunity is broad, rapid and transient. Rather than rivals, the two systems appear layered, with innate training providing a general rise in readiness while adaptive responses supply precision. Some researchers propose that the adaptive system evolved on top of an older, trainable innate foundation, and that memory in a functional sense may be a property of immunity as a whole rather than the achievement of one cellular lineage.
Not all training is beneficial. The same epigenetic machinery that heightens defence can become locked into a maladaptive state. Persistent signals from cholesterol crystals, oxidised lipids or high glucose appear to train innate cells towards chronic inflammation, implicating trained immunity in atherosclerosis, gout and neurodegenerative disease. In this framing, features of the modern lifestyle may act like an unwanted vaccine, repeatedly priming inflammatory circuits that damage the very tissues they are meant to protect. Animal studies support the idea: a high fat diet induces lasting reprogramming of myeloid cells even after a normal diet resumes.
Therapeutically, the concept cuts both ways. If innate memory can be induced deliberately, it might serve as a stopgap defence against emerging pathogens before specific vaccines exist, a strategy explored during the recent pandemic in trials of BCG among healthcare workers. Conversely, drugs that erase pathological training, such as inhibitors of the mevalonate pathway, could treat inflammatory disease at its epigenetic root instead of merely suppressing its symptoms. Several clinical programmes are now testing whether trained immunity can be boosted, dampened or reset with the precision that the adaptive system long monopolised.
Healthy scepticism remains necessary. Much of the evidence comes from animal models and observational human studies, and the durability of the effects in people is still debated. Laboratory markers of training do not always translate into clinical protection, and the heterogeneity of responses between individuals is poorly understood. Critics warn that the field risks rebranding nonspecific inflammation as memory without proving benefit. Proponents reply that the mechanistic detail now available, down to individual histone marks, distinguishes trained immunity from the vague notions of innate activation that came before it.
Whatever the disputes, the episode shows how a foundational assumption can obstruct a discovery hiding in plain sight. Reports of nonspecific vaccine effects circulated for decades before anyone thought to look for a mechanism, partly because theory insisted that none should exist. Now that the theory has changed, immunology is being rewritten to accommodate a second kind of memory, one written not in the sequences of antibodies but in the chemistry of chromatin, and the consequences are only beginning to be mapped.