The scientific backstory begins, fittingly, with the flu. Decades ago, epidemiologists noticed that mothers who contracted influenza during the second or third trimesters of pregnancy reported a higher incidence of psychiatric and neurodevelopmental disorders in their children. At the time, the observation was little more than a statistical curiosity. But when researchers later gained access to archived maternal blood samples, a clearer picture emerged: the risk appeared to track not with the virus itself but with the mother’s immune response to it. Chief among the suspects was interleukin-6, or IL-6, a signaling protein that stokes inflammation as part of the body’s defense against pathogens. Elevated IL-6 in the maternal bloodstream, it turned out, was the common thread. That discovery allowed scientists to build reliable rodent models of the phenomenon, in which a pregnant animal’s immune system is activated without any actual infection, reproducing the elevated risk of neurodevelopmental disorders in offspring.

What those models had largely lacked, however, was an epigenetic account of the damage. Most prior characterization of maternal immune activation in rodents was behavioral or electrophysiological—describing how the animals moved, responded to stimuli, or fired neuronal signals—without explaining what had changed inside the cells to produce those deficits. That gap is precisely what the Salk team, led by co-corresponding authors Margarita Behrens and Joseph Ecker, set out to close. Epigenetics refers to the layer of chemical tags and structural modifications that sits atop the base genetic sequence and determines which genes are turned on or off in a given cell. Unlike the genome itself, which is fixed at conception, the epigenome is malleable, making it a plausible target for environmental insults such as the inflammatory storm of maternal illness. If prenatal infection reshapes the developing brain, the epigenome is where that reshaping should leave its fingerprints.

To capture those fingerprints, the researchers used the field’s standard model of maternal immune activation: an injection of viral mimetic Poly(I:C), a synthetic double-stranded RNA that tricks the immune system into believing it has encountered an influenza virus. The treatment produces the same IL-6-driven inflammatory response as a genuine viral infection without introducing a pathogen. Using this model—referred to in the study as PIC-MIA—the team tracked epigenetic and transcriptional changes in the frontal cortex of mouse offspring from mid-gestation through two weeks after birth, comparing animals from immune-activated pregnancies with those from healthy pregnancies. The frontal cortex was the logical hunting ground: it is the brain region most associated with higher cognitive functions, and its development spans the entire window of vulnerability examined in the study.

The scale of the differences was striking. Across the developing cortex, the researchers documented thousands of epigenetic and gene-expression differences between the two groups. But the changes were not distributed evenly across all cell types. They concentrated, with remarkable specificity, in deep-layer neurons—a population of cortical cells that project long-range connections to other brain regions and are known to be generated during precisely the developmental window when maternal immune activation strikes. These cells form the structural backbone of cortical circuitry, and disruptions to their maturation ripple outward into the wiring of the entire brain.

At the molecular level, the team found that methylation patterns—the placement of small chemical methyl groups on the DNA—were especially altered in genomic regions that govern deep-layer neuron identity. The most consequential changes clustered at the binding sites of Tbr1, a transcription factor that serves as a master regulator of the developing brain, directing immature cells along the path to becoming deep-layer cortical neurons. Here the researchers encountered a genuine surprise. In the offspring of immune-activated mothers, Tbr1 was actually more abundant, and its binding sites were more heavily methylated than normal. Yet the genomic regions that Tbr1 typically regulates were downregulated, not upregulated. The explanation, the data suggest, is that the excess methylation was physically blocking Tbr1 from doing its job. The cell had produced more of its master architect, but methyl groups were sealing off the blueprints that architect needed to read. Deep-layer neuron development, deprived of proper Tbr1 direction, proceeded abnormally.

The implications deepened when the team cross-referenced their findings with the SFARI Gene Database, the reference catalog of autism-associated genomic alterations maintained for the autism research community. Among the high-confidence entries in that database—the genes scientists are most certain are genuinely linked to autism spectrum disorder—roughly twenty-five percent were dysregulated in the Salk dataset. In other words, the epigenetic disruptions caused by maternal immune activation were not scattered randomly across the genome; they landed preferentially on genes already known to matter for autism. That convergence between an environmental risk factor and a genetic risk landscape is exactly the kind of mechanistic bridge the field has been searching for, suggesting a route by which prenatal illness and inherited vulnerability may converge on the same developmental programs.

To confirm that the molecular changes had functional consequences, the researchers performed electrophysiological recordings on deep-layer neurons after the animals were born. The recordings confirmed what the epigenomic data predicted: the maturation of these neurons was measurably impaired in offspring of PIC-MIA pregnancies. The electrical properties that define healthy, fully developed cortical neurons were disrupted, demonstrating that the altered methylation and gene expression translated into altered neuronal function—a defect in the brain’s hardware, not merely its chemical profile.

The findings also carry an important caveat about determinism, one the researchers are careful to emphasize. Maternal infection changes the odds of neurodevelopmental outcomes; it does not seal them. “Infection changes the odds of whether neurodevelopment is affected—not everyone who gets sick during pregnancy is going to definitively have a child with a neurodevelopmental disorder,” says Joseph Ecker, professor and Salk International Council Chair in Genetics at Salk and a Howard Hughes Medical Institute investigator. This framing matters both scientifically and for public communication. The study identifies a risk-modifying mechanism, not an inevitability, and understanding the mechanism opens the door to interventions that could shift those odds back in the child’s favor—whether through maternal therapeutics that dampen harmful inflammation or through fetal interventions that protect vulnerable developmental programs.

Much remains unknown. The researchers still do not know exactly when during brain development the epigenetic damage occurs, or which stretch of pregnancy represents the window of greatest vulnerability to severe illness. “We are closer now to understanding the consequences of maternal infection, but this is only just the beginning of the story,” Ecker says. Behrens, for her part, frames the work as the payoff of a decade of methodical groundwork. “It’s just the tip of the iceberg. We are distilling all these things that we have been doing for ten years—analyzing epigenomes for years and years to get to the point that we can ask these questions. Now we can approach questions with more detail. It’s going to be a lot of fun moving forward.” As single-cell epigenomic technologies continue to mature, the ability to ask where and when specific molecular events derail specific neuronal lineages will only sharpen, bringing the field closer to preventive strategies for a class of disorders that currently affects millions of families.

What makes the study resonate beyond the laboratory is its unifying logic. A mother’s feverish immune response, a chemical tag on DNA, a transcription factor locked out of its binding sites, a neuron that fails to mature, a brain whose circuitry carries the imprint into adulthood—these are not separate stories but a single causal chain, now traced end to end in a mammalian model. The work demonstrates the lasting impact of prenatal immune challenges on offspring health and adds a mechanistic foundation to a decades-old epidemiological observation. In doing so, it transforms a correlation once glimpsed in flu-season statistics into a testable, targetable biological pathway—one that may ultimately guide the development of maternal or fetal therapeutics designed to protect the developing brain from the collateral damage of the immune system’s own defense.

Mother-to-child HIV transmission occurs when the virus passes from an infected mother to her infant during pregnancy, delivery, or breastfeeding. Without intervention, HIV can cross biological barriers or reach the infant through infected blood and breast milk. Antiretroviral treatment, viral-load monitoring, planned delivery when appropriate, and carefully managed infant feeding have reduced transmission dramatically. As a result, increasing numbers of children are growing up HIV-free despite having experienced exposure to the virus and, in many cases, to antiretroviral medicines before birth.

The new review addresses an important scientific distinction between HIV infection and HIV exposure. Children living with HIV may experience direct effects of viral replication, inflammation, immune activation, and treatment. HIV-exposed but uninfected children do not carry the virus, yet their development may still be influenced by the biological and social conditions surrounding maternal infection. These can include prenatal inflammation, maternal illness, nutritional problems, socioeconomic stress, premature birth, and exposure to medications during critical stages of fetal brain development.

Neurodevelopmental disorders are not a single condition. They include difficulties affecting cognition, language, motor coordination, attention, learning, behavior, communication, and social interaction. Brain development begins before birth and continues through childhood, relying on tightly regulated processes such as neuronal migration, synapse formation, myelination, and the maturation of neural circuits. Disruptions to these processes may not become apparent immediately, which is why long-term developmental follow-up is important even when an infant appears healthy at birth.

In a systematic review, investigators search the scientific literature using predefined criteria, assess the quality of eligible studies, and compare results across different populations and research settings. A meta-analysis then combines quantitative findings when the studies are sufficiently comparable. This approach can increase statistical power, but it cannot eliminate differences in how researchers define HIV exposure, measure development, select comparison groups, or account for factors such as prematurity and maternal health.

The authors’ work brings together evidence on the relationship between perinatal HIV exposure and neurodevelopmental outcomes. The central issue is whether children exposed to HIV show higher rates of developmental difficulties than children who were neither exposed to HIV nor antiretroviral therapy. Any association identified in such research must be interpreted carefully: it may reflect a biological effect of maternal infection, treatment exposure, complications of pregnancy, or broader environmental circumstances rather than a single causal pathway.

Several mechanisms could plausibly connect maternal HIV infection with later developmental differences. Even when antiretroviral therapy suppresses the virus, chronic immune activation and inflammatory signaling may persist. Inflammation can influence the placenta, alter fetal immune development, and affect the chemical environment in which the brain forms. Maternal HIV may also be associated with anemia, co-infections, depression, inadequate nutrition, and reduced access to prenatal care. Each of these factors can independently influence early brain development, making it difficult to isolate the contribution of HIV exposure itself.

Antiretroviral drugs add another layer of scientific complexity. These medicines are essential for protecting the mother’s health and preventing transmission, and their benefits are overwhelmingly established. At the same time, researchers continue to study how different drug combinations might interact with fetal or infant development. Exposure does not necessarily mean harm, and an observed association cannot automatically be attributed to medication. The timing of exposure, drug concentration, genetic susceptibility, placental transport, and the health consequences of untreated maternal HIV all need to be considered together.

The findings have practical implications for pediatric care and public health. They do not weaken the case for preventing mother-to-child transmission; rather, they support a broader model of care in which HIV-exposed infants receive both infection surveillance and developmental monitoring. Early screening for language, motor, cognitive, behavioral, and social difficulties can allow families to access speech therapy, educational support, developmental pediatrics, and other interventions before problems become entrenched. Future studies will need larger cohorts, standardized developmental assessments, longer follow-up, and comparison groups that account for maternal health, treatment history, birth complications, and social conditions.

The review also highlights how the success of HIV prevention has changed the scientific questions facing medicine. The priority is no longer only whether a child becomes infected, but also how prenatal exposure to maternal infection, treatment, inflammation, and environmental adversity may shape health across the life course. Understanding these relationships can help clinicians balance the extraordinary benefits of antiretroviral therapy with more precise developmental surveillance. For families affected by HIV, the goal is a complete picture of child health—one that includes freedom from infection as well as the opportunity for healthy neurological and developmental growth.

Subject of Research: The relationship between mother-to-child HIV exposure and neurodevelopmental disorders in children.

Article Title: Mother-to-child HIV exposure and its relation with neurodevelopmental disorders: a systematic review and meta-analysis.

Article References: Dantas, C.E.F., Lins, G.A.d.S.L., Soares, A.G.A. et al. Mother-to-child HIV exposure and its relation with neurodevelopmental disorders: a systematic review and meta-analysis. Pediatric Research (2026). https://doi.org/10.1038/s41390-026-05252-5

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05252-5

Keywords: HIV, mother-to-child transmission, HIV-exposed uninfected children, neurodevelopment, developmental disorders, antiretroviral therapy, prenatal exposure, pediatric health, systematic review, meta-analysis