When we think about biological inheritance, we usually think about DNA. We inherit a particular sequence of genetic letters from our parents, containing variants that influence everything from eye colour to height, cholesterol metabolism and susceptibility to certain diseases. Apart from the occasional mutation, that sequence is remarkably stable, which makes inheritance feel reassuringly straightforward: genes are passed from one generation to the next and the environment begins doing its work afterwards.

Over the last couple of decades, that tidy distinction has become considerably less tidy. Research in developmental biology suggests that the environment experienced by parents, and particularly the metabolic environment surrounding the developing egg, sperm, embryo and fetus, can influence how genes are regulated in offspring without changing the underlying DNA sequence itself. Nutrition, obesity, metabolic disease, famine, toxic exposures and stress have all been investigated as possible influences on these epigenetic processes.

This has produced some extraordinary headlines. Diet can supposedly “rewrite your genes”, parental lifestyles can be inherited by children and the consequences of what we eat today may apparently be transmitted through several generations. There is real biology underneath these claims, but there is also an impressive amount of exaggeration.

A new study published in Nature Metabolism gives us an unusually interesting opportunity to separate the two. Researchers studying diet-induced maternal obesity identified specific changes in DNA methylation within mouse oocytes — the cells that develop into eggs — at genes involved in metabolic regulation. They then did something far more informative than simply observing that the epigenetic marks were different: they experimentally recreated selected methylation changes and examined what happened to the offspring.

The manipulated offspring developed metabolic characteristics resembling some of those associated with maternal obesity.

That moves the experiment beyond correlation and towards causality. It does not demonstrate that an unhealthy diet permanently rewrites human eggs, nor that obesity in a woman condemns her children or grandchildren to metabolic disease. What it does demonstrate is that, in an experimental mammalian model, specific epigenetic information carried within an oocyte can influence metabolic physiology in the next generation.

That is a fascinating result, but understanding why requires getting underneath one of the most misunderstood words in modern health science: epigenetics.

Your DNA sequence is only part of the biological instruction manual

Almost every cell in the human body contains essentially the same DNA sequence. A liver cell and a neuron possess broadly the same genes, yet they behave completely differently because different parts of that genome are active in each cell.

Gene regulation is therefore fundamental to biology. A cell needs mechanisms that determine which genes are accessible, which are actively transcribed and which remain relatively silent.

Epigenetics describes molecular processes capable of influencing gene activity without altering the underlying DNA sequence. These include DNA methylation, modifications to the histone proteins around which DNA is packaged, chromatin organisation and several forms of non-coding RNA.

DNA methylation is probably the best known of these mechanisms. It usually involves adding a methyl group to cytosine bases at particular locations within DNA, frequently where cytosine sits beside guanine in a sequence known as a CpG site. Depending upon where methylation occurs and the surrounding regulatory context, it can influence whether transcriptional machinery can access a gene and therefore affect gene expression.

This is sometimes described as though methylation simply switches genes on or off. Biology is rarely that obliging. The relationship depends upon where methylation occurs, the gene involved, the cell type and the wider chromatin environment. Nevertheless, patterns of DNA methylation form an important part of the regulatory architecture controlling cellular behaviour.

Environmental conditions can influence those patterns. Nutritional status, inflammation, ageing, smoking and metabolic disease have all been associated with differences in DNA methylation.

The difficult question is whether any of those environmentally induced differences can survive reproduction and influence the next generation.

Most epigenetic marks are deliberately erased

This is where the popular story about inherited epigenetic changes encounters an important biological obstacle.

During the formation of sperm and eggs, and again shortly after fertilisation, the mammalian genome undergoes extensive epigenetic reprogramming. Much of the existing DNA methylation landscape is erased and subsequently rebuilt.

There are good reasons for this. An egg or sperm cell needs to contribute genetic material capable of producing an entirely new organism rather than carrying forward the specialised epigenetic identity of the cells from which it developed.

This reprogramming acts as a powerful barrier against indiscriminately transmitting acquired epigenetic changes between generations.

It is not absolute. Certain regions of the genome, including imprinted genes, are deliberately regulated in ways that retain parent-of-origin information. Some methylation patterns and other epigenetic signals may also escape or be re-established after reprogramming.

But the existence of this extensive reset means that claims about inherited epigenetic changes require unusually strong evidence.

Finding that a person's blood cells contain a particular methylation pattern after an environmental exposure does not mean that the same pattern has entered their germ cells. Finding it in a germ cell does not necessarily mean it survives fertilisation. Even if an association appears in offspring, that does not automatically demonstrate that an epigenetic mark was responsible.

There are several other routes through which parents can influence the health of their children.

The womb is an environment, not merely a delivery system

Maternal influences are particularly difficult to interpret because an offspring developing inside the uterus is directly exposed to the mother's physiology.

Glucose, lipids, hormones, inflammatory mediators, placental function and nutrient availability all contribute to the intrauterine environment. Maternal obesity, diabetes and severe undernutrition can alter that environment and influence fetal development.

This is usually described as developmental programming and forms part of the broader Developmental Origins of Health and Disease framework.

The concept has strong biological foundations. Developing tissues are remarkably plastic. Nutritional and hormonal signals during critical developmental periods can influence organ structure, endocrine regulation and metabolic function in ways that persist after birth.

If a child born to a mother with obesity later develops greater adiposity or insulin resistance, however, we cannot immediately conclude that an epigenetic change in the egg was inherited.

The effect might have arisen during pregnancy. It might reflect shared genetic susceptibility. Mother and child may share diet, socioeconomic circumstances and other environmental exposures after birth. All of these can create similarities between generations without requiring germline epigenetic inheritance.

This is why animal experiments become particularly valuable. Researchers can control mating, diet and environmental conditions in ways that would be impossible or unethical in humans, then manipulate specific molecular mechanisms to determine whether they actually cause the phenotype being studied.

The new Nature Metabolism experiment did exactly that.

What did the researchers find?

The researchers used a mouse model of diet-induced maternal obesity and examined DNA methylation within oocytes.

They identified changes at several metabolically important genes, including Hnf1a, Thra and Pdk4. These genes are involved in processes relevant to glucose and lipid metabolism, thyroid-hormone signalling and cellular fuel utilisation.

HNF1A is a transcription factor with an important role in metabolic regulation, particularly within the liver and pancreas. Rare pathogenic variants in the human HNF1A gene can cause maturity-onset diabetes of the young, illustrating just how biologically important this pathway can be.

THRA encodes thyroid hormone receptor alpha, through which thyroid hormones influence development and energy metabolism, while PDK4 regulates pyruvate dehydrogenase activity and therefore helps determine whether cells favour glucose oxidation or shift towards greater reliance on fatty-acid-derived fuels.

The presence of altered methylation at these genes in oocytes from obese mice was interesting, but by itself it would still have represented an association.

Maternal obesity could alter hundreds or thousands of biological processes simultaneously. A methylation difference might simply accompany the metabolic disturbance without causing anything important in the offspring.

The researchers therefore used targeted epigenetic editing.

Epigenetic editing turns an association into an experiment

Modern molecular tools allow researchers to target epigenetic machinery to specific regions of DNA.

The principle resembles CRISPR gene editing, but rather than cutting the DNA sequence and changing the genetic code, modified systems can be used to alter epigenetic marks at selected genomic locations.

This provides an extraordinarily useful experimental tool. If researchers suspect that methylation at a particular site contributes to a phenotype, they can reproduce that methylation change without recreating the entire environmental exposure that originally produced it.

In the new study, targeted modification of methylation at the identified loci in mouse oocytes was sufficient to produce metabolic changes in the resulting offspring. The offspring displayed alterations in glucose and lipid metabolism resembling aspects of the phenotype associated with maternal obesity.

That substantially strengthens the argument for causality.

Instead of simply showing that maternal obesity, altered oocyte methylation and offspring metabolic dysfunction occurred together, the researchers demonstrated that changing selected epigenetic marks could itself alter offspring physiology.

This is the part of the study that deserves attention. It provides a mechanistic bridge between the metabolic state of one generation, information carried within the oocyte and metabolic characteristics appearing in the next.

The human component makes the finding more intriguing, but not definitive

The researchers also examined human oocytes and identified methylation differences at some of the same metabolically relevant genes in oocytes from women with obesity.

This creates an important connection between the animal experiment and human biology. The molecular signatures observed in mice were not entirely peculiar to an artificial laboratory model.

However, the strength of evidence is very different on either side of that bridge.

In mice, researchers could manipulate the epigenetic marks and observe consequences in offspring. In humans, they could observe methylation differences in oocytes associated with obesity.

They could not ethically alter those human oocytes, create pregnancies and determine whether the resulting children developed metabolic abnormalities.

The causal component therefore remains an animal finding.

That distinction is crucial because it is very easy for a headline to collapse the entire study into the statement that “obesity in women epigenetically programs metabolic disease into their children”.

The study does not establish that.

It demonstrates a plausible mechanism in mice and identifies related molecular observations in human oocytes. That is considerably more interesting than a purely observational association, but it is still not proof that the same causal chain operates with the same magnitude in humans.

Intergenerational and transgenerational inheritance are not the same thing

Another source of confusion is the language of generations.

If a pregnant woman experiences an environmental exposure, three biological generations can be directly exposed simultaneously. The woman herself is exposed, the fetus developing inside her is exposed, and if that fetus is female, the developing germ cells inside the fetus — which could eventually become the woman's grandchildren — are also present during the exposure.

This makes maternal transmission particularly tricky to classify.

An effect observed in the child is generally described as intergenerational because that child was directly exposed during fetal development. Even an effect observed in the grandchild may potentially reflect exposure of the germ cells that were present inside the developing female fetus.

For a maternal exposure during pregnancy, evidence of true transgenerational inheritance usually requires demonstrating effects in a generation that was not itself directly exposed.

Paternal studies have a slightly different structure because sperm can carry information into the embryo without exposing the offspring to the father's internal metabolic environment during gestation.

These definitions may sound pedantic, but they matter enormously when evaluating claims that an environmental exposure has been inherited across generations.

A study demonstrating an effect in the immediate offspring has not automatically demonstrated transgenerational inheritance.

Could effects really persist for several generations?

Animal studies suggest that some environmentally associated phenotypes can persist beyond the first generation.

Experiments involving nutritional exposures, endocrine-disrupting chemicals and metabolic disturbances have reported effects in descendants that were not directly exposed to the original condition. Proposed mechanisms include DNA methylation, histone modifications, sperm RNAs and alterations in germ-cell development.

Some studies of paternal diet have been particularly influential. Changes in sperm small RNAs have been associated with altered metabolic phenotypes in offspring, and experimentally injecting selected sperm RNAs into embryos has reproduced aspects of those effects in mice.

This suggests that biological information beyond the DNA sequence can, under experimental conditions, travel through the germline.

The crucial phrase is under experimental conditions.

Demonstrating that a mechanism is possible in mice does not establish how important it is in normal human populations.

Human reproduction involves enormous genetic, behavioural and environmental variation, while generation times make controlled multigenerational experiments impossible. Researchers therefore have to rely heavily on natural experiments and observational cohorts.

Those studies can be fascinating, but they are much harder to interpret.

Human famine studies provide some of the most famous evidence

One of the best-known natural experiments comes from the Dutch Hunger Winter of 1944–45.

During the final months of the Second World War, a German blockade and severe winter created an acute famine in the western Netherlands. Because the period was geographically and temporally well documented, researchers have been able to follow people who were exposed to famine during different stages of fetal development.

Decades later, prenatal famine exposure was associated with differences in adult metabolic health, including greater risks of obesity, glucose intolerance and cardiovascular disease depending upon the timing of exposure.

Researchers also identified persistent differences in DNA methylation at certain genes, including the imprinted IGF2 locus, more than six decades after prenatal exposure.

This is extraordinary evidence that the nutritional environment during early development can leave molecular signatures lasting for much of a human lifetime.

It is sometimes cited as proof of transgenerational epigenetic inheritance.

That interpretation goes too far.

The individuals studied were directly exposed to famine while developing in the uterus. Their altered health and epigenetic profiles therefore demonstrate developmental programming, not necessarily germline transmission of an acquired epigenetic state.

There have been attempts to identify effects in subsequent generations, but separating epigenetic inheritance from shared genetics, family environment, socioeconomic factors and parental physiology becomes increasingly difficult.

The human evidence for developmental programming is therefore considerably stronger than the evidence for true transgenerational epigenetic inheritance.

Historical studies are fascinating but extremely difficult to interpret

Another frequently cited example comes from Överkalix in northern Sweden.

Researchers used historical records to examine food availability experienced by grandparents and reported associations with mortality and disease risk in subsequent generations. Some effects appeared to depend upon whether exposure occurred during particular developmental periods and whether transmission occurred through the paternal or maternal line.

These findings generated enormous interest because they seemed to suggest that nutritional abundance or scarcity experienced by grandparents could influence grandchildren who had never experienced those conditions themselves.

The problem is that historical observational studies contain countless potential confounders. Families share genes, occupations, geography, wealth, social conditions and dietary patterns. Sample sizes can be small, multiple comparisons increase the possibility of chance findings and historical estimates of food availability are necessarily imperfect.

The studies are valuable for generating hypotheses.

They cannot provide the kind of mechanistic causal evidence achievable in laboratory animals.

This is a recurring theme throughout epigenetic inheritance research. The strongest mechanistic evidence comes from animals, while the most directly relevant human evidence is inevitably less experimentally controlled.

Fathers are part of the story too

Discussions about developmental health have historically concentrated heavily on mothers, partly because the maternal environment directly supports fetal development.

That focus can become scientifically and socially problematic if every adverse childhood outcome is eventually traced back to something the mother ate, weighed or experienced before pregnancy.

Paternal metabolic health can also influence offspring biology.

Obesity, diet, smoking and other exposures have been associated with changes in sperm DNA methylation, chromatin organisation and small non-coding RNAs. Animal studies have shown that paternal nutritional exposures can alter offspring metabolism even when the father contributes nothing to the pregnancy beyond sperm.

This is important because sperm provide a cleaner experimental route for investigating germline inheritance. There is no paternal uterus, placenta or pregnancy metabolism to confuse the interpretation.

Again, animal evidence is much stronger than human evidence, but the broader principle is becoming difficult to dismiss: the biological state of both parents around conception may influence reproductive cells in ways extending beyond DNA sequence alone.

That does not mean prospective parents need to achieve metabolic perfection before conceiving. Human reproduction would have ended fairly early in our evolutionary history if that were a requirement.

It means reproductive biology is more environmentally responsive than the traditional gene-versus-environment model suggested.

Nutrition can influence epigenetic chemistry directly

There is another reason nutrition and epigenetics are so frequently discussed together.

Methyl groups do not appear from nowhere.

One-carbon metabolism provides methyl groups used in numerous biochemical reactions, including DNA methylation. Nutrients including folate, vitamin B12, choline, methionine and vitamin B6 participate in interconnected pathways involving the transfer of one-carbon units and the production of S-adenosylmethionine, a major methyl donor.

This provides an obvious biochemical route through which nutritional status can interact with epigenetic regulation.

However, the existence of that pathway does not mean eating more methyl-donor nutrients simply increases beneficial DNA methylation.

Methylation is highly site-specific. Increased methylation in one genomic region may reduce expression of a gene, while methylation elsewhere can have different regulatory consequences. More methylation is not inherently better, just as less methylation is not inherently healthier.

The genome is not a wall waiting for us to throw methyl groups at it.

Nutritional adequacy matters enormously during conception and pregnancy, with folate providing perhaps the clearest example because of its established role in preventing neural-tube defects. But using the language of epigenetics to imply that high-dose methyl-donor supplements can deliberately optimise the genome of future children moves considerably beyond the evidence.

Obesity itself is not a single exposure

The new maternal-obesity study also raises another important question: what exactly is producing the epigenetic change?

Obesity is a phenotype, not a single biochemical condition.

Two people with the same BMI can differ dramatically in insulin sensitivity, liver fat, inflammatory status, lipid profile, dietary quality, physical activity and adipose-tissue distribution. Maternal obesity can coexist with normal glucose regulation in one person and severe insulin resistance in another.

In animal experiments, diet-induced obesity also packages several exposures together. The animal consumes a particular experimental diet, gains adipose tissue and develops metabolic changes. Any of these components could contribute to alterations in the oocyte.

The new study strengthens the case that selected methylation changes can mediate downstream effects, but it does not mean adipose tissue itself is necessarily the initiating signal.

Future work will need to determine which aspects of maternal metabolic dysfunction influence germ-cell epigenetics and whether improving metabolic health before conception can reverse or prevent those changes.

That would have considerably greater clinical value than simply identifying obesity as a risk marker.

This is where the risk of mother-blaming becomes very real

Research into maternal nutrition and offspring health needs particularly careful communication.

Pregnancy already arrives with an impressive catalogue of things women are told they must do perfectly. Adding the suggestion that an imperfect diet or higher body weight might epigenetically damage several future generations would be both scientifically unjustified and spectacularly unhelpful.

Population-level risk is not individual destiny.

Even where maternal obesity is associated with greater metabolic risk in offspring, genetics, paternal biology, pregnancy physiology, childhood environment, diet, physical activity, socioeconomic conditions and countless later-life exposures continue to influence what ultimately happens.

Epigenetic regulation is also dynamic. The entire concept depends upon biological responsiveness to environmental conditions. It would be peculiar to describe epigenetic marks as environmentally modifiable when discussing how they arise and then treat them as irreversible destiny once a child is born.

The responsible interpretation is therefore not that mothers carry sole responsibility for programming future generations. It is that the metabolic health of parents and the nutritional environment surrounding early development may be biologically important enough to deserve greater support before and during pregnancy.

That shifts the emphasis from blame to prevention.

Preconception health may deserve considerably more attention

Healthcare systems tend to become intensely interested in nutrition once pregnancy has already begun.

There is a strong argument for beginning earlier.

Eggs and sperm develop and mature before conception. Maternal metabolic health influences the environment into which an embryo will implant, while paternal metabolic status may influence sperm quality and molecular characteristics.

Smoking cessation, appropriate folate intake, management of diabetes, improving dietary quality, addressing severe obesity where appropriate, correcting nutritional deficiencies and supporting physical activity before conception are already sensible for numerous established reasons.

Epigenetic research potentially adds another biological layer to those recommendations without requiring us to invent an entirely new preconception lifestyle.

This is an important point because emerging science often gets turned into exotic interventions when the practical implications are much less glamorous.

We do not currently have evidence that prospective parents need commercial epigenetic testing, methylation supplements or a special “gene-resetting” diet.

Supporting metabolic health before conception is a considerably more defensible place to start.

Could improving health reverse harmful epigenetic patterns?

This may ultimately become one of the most interesting questions in the field.

If metabolic dysfunction can influence germ-cell epigenetics, can improving metabolic health reverse those changes before conception?

Animal research suggests that at least some environmentally induced epigenetic alterations are modifiable. Exercise, dietary change and improvements in metabolic state can influence epigenetic patterns in somatic tissues, while interventions before conception have shown effects on offspring outcomes in some experimental models.

Human evidence is much more limited.

Weight loss, exercise and dietary interventions can alter DNA methylation in tissues such as skeletal muscle and adipose tissue, demonstrating that the human epigenome remains responsive throughout life. Whether similar interventions reliably remodel human oocytes or sperm in ways that improve offspring health is far less certain.

This is where the next generation of research needs to go.

Demonstrating that an adverse metabolic environment leaves a molecular mark is scientifically interesting. Demonstrating that the mark can be prevented or reversed would be clinically transformative.

Epigenetics does not overthrow genetics

The enthusiasm surrounding epigenetics occasionally produces the impression that we have discovered a completely new inheritance system that renders traditional genetics almost irrelevant.

That is not remotely the case.

DNA sequence remains fundamental. Genetic variants influence obesity, type 2 diabetes, lipid metabolism, cardiovascular disease and countless other traits. Many of the proteins controlling epigenetic regulation are themselves encoded by genes.

Epigenetics adds another regulatory layer through which cells interpret and respond to developmental and environmental information.

The interaction between the two is what makes the biology interesting.

A genetic variant may alter susceptibility to an environmental exposure. An epigenetic change may modify expression of a genetically determined pathway. Nutritional status may influence both through metabolic intermediates and signalling systems.

Nature and nurture were never particularly interested in being separated into opposing camps.

What should we actually take from this research?

The most important implication is not that every meal becomes a biological message to our grandchildren.

The stronger conclusion is that reproductive cells and developing organisms are responsive to the metabolic environments in which they exist. In experimental animals, some of that information can be carried through epigenetic mechanisms and influence offspring physiology. The new Nature Metabolism study strengthens this argument considerably by demonstrating that targeted alteration of specific methylation marks in mouse oocytes can reproduce aspects of metabolic dysfunction in offspring.

Human research adds plausibility. Obesity has been associated with related methylation differences in human oocytes, while historical famine studies demonstrate that early nutritional exposures can leave molecular and metabolic signatures detectable decades later.

What remains uncertain is the extent to which acquired epigenetic information is transmitted through the human germline, how often it survives the extensive epigenetic reprogramming occurring during reproduction and how much it contributes to disease compared with genetics, pregnancy physiology and shared family environment.

Those uncertainties matter because the phrase “epigenetic inheritance” can easily imply a degree of permanence that the evidence does not support.

If anything, epigenetics demonstrates how biologically responsive we are to our environment.

We inherit more than a DNA sequence, but we do not inherit destiny

The traditional view of inheritance was wonderfully simple: DNA passes from parent to child while lifestyle and environment begin exerting their influence afterwards. Modern developmental biology has shown that the boundary is much more porous.

The nutritional and metabolic environment surrounding reproduction can influence developing cells before an individual has taken their first breath. Some of those influences involve epigenetic mechanisms capable of altering gene regulation without changing DNA sequence. In animal models, selected signals can be carried through germ cells and influence metabolism in subsequent generations.

That deserves to change how seriously we take preconception and early-life health.

It does not justify telling people that today's dessert will give their grandchildren diabetes.

The new research is exciting precisely because it gives us a plausible molecular mechanism rather than another vague association. Specific methylation changes appeared in oocytes during maternal obesity, related signatures were identified in human oocytes, and experimentally reproducing selected changes in mice altered metabolic physiology in the next generation.

The next challenge is determining how much of this biology operates in humans, which exposures matter most and, crucially, whether improving metabolic health can modify the signal before it is transmitted.

If those questions can be answered, epigenetic inheritance may eventually give us something considerably more useful than another reason to worry about our genes. It may identify periods of life when improving nutrition and metabolic health produces benefits extending beyond the individual receiving the intervention.

That possibility is extraordinary enough without turning it into science fiction.

References

Heijmans, B.T., Tobi, E.W., Stein, A.D., Putter, H., Blauw, G.J., Susser, E.S., Slagboom, P.E. and Lumey, L.H. (2008) ‘Persistent epigenetic differences associated with prenatal exposure to famine in humans’, Proceedings of the National Academy of Sciences of the United States of America, 105(44), pp. 17046–17049. doi:10.1073/pnas.0806560105.

Huypens, P., Sass, S., Wu, M., Dyckhoff, D., Tschöp, M., Theis, F., Marschall, S., Hrabě de Angelis, M. and Beckers, J. (2016) ‘Epigenetic germline inheritance of diet-induced obesity and insulin resistance’, Nature Genetics, 48, pp. 497–499. doi:10.1038/ng.3527.

Ng, S.-F., Lin, R.C.Y., Laybutt, D.R., Barres, R., Owens, J.A. and Morris, M.J. (2010) ‘Chronic high-fat diet in fathers programs β-cell dysfunction in female rat offspring’, Nature, 467, pp. 963–966. doi:10.1038/nature09491.

Radford, E.J., Ito, M., Shi, H., Corish, J.A., Yamazawa, K., Isganaitis, E., Seisenberger, S., Hore, T.A., Reik, W., Erkek, S., Peters, A.H.F.M., Patti, M.-E. and Ferguson-Smith, A.C. (2014) ‘In utero effects. In utero undernourishment perturbs the adult sperm methylome and intergenerational metabolism’, Science, 345(6198), 1255903. doi:10.1126/science.1255903.

Skvortsova, K., Iovino, N. and Bogdanović, O. (2018) ‘Functions and mechanisms of epigenetic inheritance in animals’, Nature Reviews Molecular Cell Biology, 19, pp. 774–790. doi:10.1038/s41580-018-0074-2.

Tobi, E.W., Lumey, L.H., Talens, R.P., Kremer, D., Putter, H., Stein, A.D., Slagboom, P.E. and Heijmans, B.T. (2009) ‘DNA methylation differences after exposure to prenatal famine are common and timing- and sex-specific’, Human Molecular Genetics, 18(21), pp. 4046–4053. doi:10.1093/hmg/ddp353.

Veenendaal, M.V.E., Painter, R.C., de Rooij, S.R., Bossuyt, P.M.M., van der Post, J.A.M., Gluckman, P.D., Hanson, M.A. and Roseboom, T.J. (2013) ‘Transgenerational effects of prenatal exposure to the 1944–45 Dutch famine’, BJOG: An International Journal of Obstetrics & Gynaecology, 120(5), pp. 548–553. doi:10.1111/1471-0528.12136.