Breastfeeding Leaves Subtle Epigenetic Marks on Infants’ DNA

Early-life nutrition is increasingly recognized as a critical determinant of short- and long-term health, with breastfeeding often linked to improved developmental and disease outcomes.

However, the biological mechanisms underpinning these associations remain only partially understood. Epigenetics—the molecular modifications that regulate gene activity without altering DNA sequence—has emerged as a promising explanatory mechanism. 

Dr. Doretta Caramaschi, lecturer at the University of Exeter and co-lead author of a recent large-scale epigenetic study, is working to understand the influence of breastfeeding on epigenetic modifications. Drawing on data from thousands of individuals across international cohorts, her recent research investigated how breastfeeding may shape infant biology at the molecular level. 

Technology Networks spoke with Caramaschi to discuss the rationale behind studying epigenetics, the challenges of detecting subtle biological signals, and the future directions needed to translate these findings into meaningful health insights.

Epigenetics as a bridge between early nutrition and long-term health outcomes 

What led you to focus on epigenetics as a potential mechanism for the beneficial effects of breastfeeding? 

Caramaschi explained that both epigenetic processes and breastfeeding occur during a critical window of early development, making their relationship biologically compelling. Epigenetic modifications, such as DNA methylation, play a central role in “priming” biological systems, influencing how genes are expressed later in life. Because these changes are especially dynamic in infancy, they present a plausible mechanism through which early nutritional exposures might exert lasting effects. 

Epigenetic markers can also be influenced by the environment, and from a developmental perspective, breastfeeding represents one of the most significant early-life environmental inputs. It provides not only nutrients but also bioactive compounds that may interact with the infant’s developing systems. This overlap in timing and biological significance motivated Caramaschi to explore whether breastfeeding could leave detectable epigenetic signatures. 

“Epigenetic changes and breastfeeding both happen very early in life, and in my mind, these two things were definitely connected.” — Dr. Doretta Caramaschi 

Early-life epigenetics and nutrition insights: 

• Epigenetic processes are highly active during early development. 

• Breastfeeding coincides with key windows of biological programming. 

• DNA methylation provides a measurable link between environment and gene regulation. 

Scaling epigenetic research through global consortia

Can you talk us through how you approached the meta-analysis of epigenome-wide association studies? Why is work of this scale so important? 

To robustly detect epigenetic signals, Caramaschi and her colleagues conducted a meta-analysis of epigenome-wide association studies (EWAS), combining data across multiple cohorts. This approach was essential because individual studies rarely have sufficient sample sizes to detect subtle molecular differences. 

“These large sample sizes are not available within just one study because you need blood samples from lots of people, which represents a participant burden,” said Caramaschi. “It is also costly, both in terms of laboratory material and in employing people to run the analysis.” 

The work relied on the Pregnancy and Childhood Epigenetics consortium, which brings together datasets from across the globe. By pooling resources, the team was able to analyze thousands of samples, far exceeding what any single study could achieve. This scale is critical for distinguishing true biological signals from background noise.  

Caramaschi emphasized that human studies are inherently variable. Differences in genetics, environment, and lifestyle introduce substantial heterogeneity, making it difficult to identify reliable epigenetic markers without large datasets. 

“One tiny molecular change might not have a huge effect—to detect it, you really need large sample sizes.” — Dr. Doretta Caramaschi

Meta-analysis helps overcome this challenge by increasing statistical power and enabling assessment of whether findings are reproducible across populations. 

Importance of large-scale epigenetic datasets: 

• Large cohorts are needed to detect subtle DNA methylation changes. 

• Consortia enable data-sharing across countries and populations. 

• Meta-analysis reduces noise and assesses reproducibility in human studies. 

Detecting breastfeeding-associated DNA methylation patterns 

What were the most important differences you found between children who were breastfed and those who were not? 

The study identified several sites across the genome where DNA methylation levels differed between children who were exclusively breastfed and those who were not. These differences were modest but consistent across multiple cohorts, strengthening confidence in their validity. Some of the identified sites were located within known genes, while others were in genomic regions with unclear functions. 

The annotated genes with different methylation patterns—ALAD, FNBP4, and CHFR—are related to developmental and immune processes. 

Importantly, the findings also revealed a dose–response relationship: longer durations of exclusive breastfeeding were associated with more pronounced methylation differences. This pattern suggests that the exposure itself may be influencing these molecular changes rather than unrelated factors. 

“The longer the exclusive breastfeeding, the more of these markers appeared,” explained Caramaschi. 

To confirm that these differences were not pre-existing, the researchers also analyzed samples collected at birth. They found no evidence of methylation differences at these sites prior to breastfeeding, supporting the idea that the observed changes emerge postnatally and are likely linked to feeding practices. 

Key molecular differences linked to breastfeeding: 

• Specific DNA methylation sites differ between feeding groups. 
• No differences in methylation at these sites were observed at birth, supporting a postnatal origin. 

• Effects scale with breastfeeding duration. 

Interpreting small epigenetic differences in complex biological systems 

How should we interpret the fact that the number of epigenetic changes identified between breastfed and non-breastfed babies was relatively small? 

Caramaschi cautioned that small numbers of epigenetic differences are expected in this context. Breastfeeding is just one of many factors shaping development, and its molecular effects are unlikely to produce dramatic or easily observable changes at the population level. “You wouldn’t expect many big biological differences because you can't detect by looking at a child whether or not they were breastfed,” she explained. 

Moreover, biological and environmental influences accumulate over time. As children grow, factors such as diet, parental care, and socioeconomic conditions can modify or dilute early epigenetic signals. Caramaschi noted that in the EWAS meta-analysis, epigenetic sampling was conducted years after breastfeeding had ceased, potentially obscuring its effects.  

Rather than viewing the small number of changes as a limitation, Caramaschi considers them meaningful indicators of early-life influence. “It’s exciting that we found these changes robustly across different cohorts,” she said.  

Understanding subtle epigenetic signals: 

• Breastfeeding effects are expected to be small but biologically relevant. 

• Post-infancy exposures can dilute early epigenetic signatures. 

• Epigenetic marker changes were detected robustly across studies.

Untangling breastfeeding effects from other factors 

What are some of the biggest challenges in separating the effects of breastfeeding from other factors that influence the epigenome? 

A major challenge in this field is disentangling breastfeeding effects from confounding variables. “These are population studies, so we cannot fix all of the other variables and leave only breastfeeding as our main variable; we have a lot of other factors at play,” Caramaschi said.  

She noted that socioeconomic status is a major confounding factor: “In most of the study populations, societies, and countries that I work with, socioeconomic status is very much correlated with breastfeeding.” This makes it difficult to isolate the specific contribution of feeding type to epigenetic changes.  

Caramaschi highlighted the need for more diverse datasets, particularly from populations where these associations may differ. “I haven’t found any societies with data available without a link between breastfeeding and socioeconomic status,” she said. If such data could be found, it could help clarify whether observed epigenetic changes are truly driven by breastfeeding or reflect broader social and environmental influences. 

In the study, statistical adjustments were made to account for known confounders, but residual bias may persist. This limitation is a common issue in population-based epigenetic research.  

Challenges in isolating breastfeeding effects: 

• Socioeconomic status strongly correlates with breastfeeding patterns. 

• Observational studies cannot fully eliminate confounding variables. 

• Statistical adjustments help, but cannot remove all bias. 

Future directions: Linking epigenetic markers to long-term health 

What kinds of studies or data would help us understand whether the epigenetic changes identified by your work translate into real-world health outcomes? 

Looking ahead, Caramaschi outlines three key research priorities. First, molecular-level studies are needed to determine the functional consequences of identified methylation changes. “While we’ve identified these epigenetic markers, we don’t necessarily know what these do at the cellular, physiological, or developmental level,” she said. 

This will involve laboratory experiments, in vitro systems, and computational modeling to understand how these modifications influence gene expression and cellular processes. 

Second, long-term longitudinal studies are required to assess whether these epigenetic differences persist and impact health outcomes beyond childhood. Such datasets are rare but essential for linking early-life molecular changes to disease risk or developmental trajectories. 

 

Finally, improving population diversity is critical. Current datasets are heavily skewed toward Western, white, educated populations with European ancestry, limiting generalizability. “We need to recruit more diverse populations and generate data in other parts of the world,” said Caramaschi. Expanding research to underrepresented regions, including Africa and South America, will be essential for building a more comprehensive understanding of breastfeeding’s epigenetic effects. 

Priorities for advancing epigenetic research: 

• Functional studies to determine the biological impact of methylation changes. 

• Longitudinal cohorts to track long-term health outcomes. 

• Greater inclusion of diverse, underrepresented populations. 

 

This content includes text that has been created with the assistance of generative AI and has undergone editorial review before publishing. Technology Networks' AI policy can be found here.