Lars Bygren and Marcus Pembrey on transgenerational responses to nutrition

Posted by Biome on 27th March 2014 - 2 Comments

Heritability has been the focus of scientific research through the ages, with renewed interest following the advent of genomics and epigenetics. A key question yet to be comprehensively answered is whether environmental conditions experienced by one generation can impact subsequent generations and lead to certain predispositions, whether through epigenetic or other mechanisms. In a study in BMC Genetics, Lars Bygren from the Karolinska Institute, Sweden, Marcus Pembrey from University College London, UK, and colleagues reveal that changes in paternal grandmothers’ early food supply influences the cardiovascular health of their granddaughters. Here Bygren and Pembrey explain more about transgenerational responses and the effect of nutrition in this context.


Could you explain what is meant by the term ‘transgenerational response’?

In general terms, biological effects are questions of homeostasis and adaptation to the environment. In medical ecology this is studied as the interaction between man and environment. According to epidemiology the agent – chemical, physiological or biological -  interacts with man and provokes resilience to the disturbed homeostasis, all under the influence of confounders in the environment. Agents are at the focus. The basic mechanisms behind transgenerational responses are worlds of science in themselves. Much interest is now paid to epigenetics, epigenetics in interaction with other genetics. There could be still other mechanisms such as the general principle of prions´ defensive and adverse effects.

In our studies, we use ‘transgenerational effect’ or ‘transgenerational response’ as a general term for the phenomenon where experience or environmental exposure in one generation produces altered development, health or survival in the next generation(s). The word ‘response’ is used when features (such as a particular exposure sensitive period in development) suggest a pre-evolved transmission mechanism is being induced by the initial parental/ancestral exposure.

Importantly the term ‘transgenerational response’ does not imply any particular molecular mechanism (such as transgenerational epigenetic inheritance), or particular pattern of outcomes across the generations – it includes, but is not limited to, ‘inheritance of acquired characteristics.’ The outcomes in descendants can be harmful or beneficial.


There have been some prominent studies looking at the effects of very poor nutrition on offspring health, particularly from events during World War II. Could you explain what makes these datasets so valuable?

These studies are valuable in demonstrating the intra-generational effects of poor early nutrition and furthermore revealing some important concomitant molecular changes and disease outcomes. They demonstrate the exposure of the foetus and outcomes such as schizophrenia, depression and symptoms of metabolic syndrome. Effects have often been found to be minor or absent. For example, after the long Leningrad Siege during WW2 and the 1866-1867 total crop failures in Finland. Analyses of changes in nutrition per se extending to three to four generations could give us insights into the inheritance of disease and health, and the question of prevention and of therapy.

One of the challenges in the study of human transgenerational responses is to reduce confounding due to genetic or cultural inheritance.  A clear cut, well documented, switch from adequate to very poor nutrition for a few months and then back to adequate nutrition creates a natural experiment. This was the situation in parts of the Netherlands  due to the blockades in 1944. One has ‘controls’ who were conceived  after the famine, or one can compare exposure at different periods of early development of the mother on her own child’s development, since there is emerging evidence of exposure sensitive periods for inducing transgenerational responses.


What evidence is there that the effects of poor nutrition in a parent can be seen beyond the first generation? Do you think these effects diminish through the generations?

In plant and animal research it has been demonstrated that adverse and enriched environments have responses over three to seven non-exposed generations. In some experiments the effects have diminished over the generations but the opposite cannot be ruled out. It might be relevant to mention that DNA methylation is mutagenic.

The human evidence linking food supply with transgenerational responses affecting  grandchild mortality comes from the historical studies of the Överkalix population in Sweden. A good food supply in mid childhood (but not adolescence) of the paternal grandfather was associated with increased mortality rate in his grandsons (but not granddaughters). There is not enough human data to comment on whether or not the transgenerational effects diminish down the generations. There are rodent nutrition experiments where some effects do not diminish, but other features disappear. It is a complex picture.


In general, do you think greater transgenerational effects of nutrition are experienced by male or female offspring, and imparted by male or female parents?

In general transgenerational sequels of nutrition occur in many lineages and in summation are experienced quite equally between sexes or genders even if there are some differences in what parents can transmit to the next generation. Paternal line study and maternal line study are both fruitful but the combination more so. It is too early to make any generalisations, but all combinations are seen if you include animal experiments as well. Sex-specific effects and transmissions are one of the striking features of both human observations and animal experiments demonstrating transgenerational responses. Whilst the limited human data are compatible with X or Y segregation there are other possibilities. Certainly there are mouse (social stress, not nutrition) experiments that are incompatible with X or Y transmission with unaffected males transmitting the anxiety outcome to their female but not male offspring; and these unaffected male offspring likewise transmitting the anxiety outcome to their female but not male offspring.


How certain can we be that these effects are due directly to nutrition, and not to other environmental or social factors that are perhaps linked to nutrition?

This is a good point. Nutritional crises are most often linked to personal, family and society burdens. For the coming studies it will be important to discern the attributions of the two general factors. Famines are likely to be associated with stress which can induce transgenerational responses at least in experimental animals. This type of confounding can only be resolved with many more human transgenerational studies including studies of stressful events not associated with famine.


What biological mechanisms do you think underlie these transgenerational effects?

A transgenerational response involves at least three steps; some form of biological ‘capture’ of the exposure event with an enduring change in the germ line (or its integral support cells), the actual molecular signal transmitted to the zygote, and the mechanism for translating this signal into altered embryonic development of the offspring. Of course, females have more routes than males through which exposure-induced  physiological or metabolic changes can impact on the next generation. There are the trans-placental and breast feeding routes for metabolic signals, and likely maternal effects via the egg cytoplasm including mitochondrial inheritance.

Candidate mechanisms that are possibilities for males  – and should be considered for females also – include the following: Epigenetic changes induced by the exposure, such as histone modifications or specific DNA methylation profiles, being present and maintained in sperm with the potential to influence the zygote. Alternatively the transmitted signal could be an altered profile of non-coding RNAs carried by the sperm. More than one mechanism is likely to be employed in the chain of events that make up a transgenerational response. For example, the exposure might induce epigenetic changes that in turn alter the non-coding RNA profile of the sperm, which on reaching the zygote may influence development through epigenetic changes to DNA methylation. It is not impossible that seminal fluid carries biological information that can impact on the next generation.


Are you aware of any prospective studies investigating transgenerational effects? What’s next for your own research in this field?

Human studies on transgenerational effects can be retrospective and prospective – equally fruitful from pure statistical viewpoints. The retrospective have the advantage of having data from many of the long human generations and often find large variations in independent variables. The prospective will include few generations but have much better data quality. A good research environment for the former is to be found in northern Europe with its public population registries.

The transgenerational responses to loss of parents in childhood for all born during the last decades in Sweden will be followed, the Överkalix study enlarged and reproduced in another Swedish setting. In the US the first Adventist study in the 1960s is now repeated and the Dutch Hunger Winter cohorts in Leiden and Amsterdam will sooner or later be followed up to further generations.

The Avon Longitudinal Study of Parents and Children (ALSPAC) has demonstrated transgenerational responses to paternal smoking onset in mid childhood, or to the ALSPAC mother or father being exposed in their foetal life by the grandmothers smoking in pregnancy.  The growth and development of these ALSPAC ‘children’ have been followed since 1991/2. Whilst there is follow up of the next generation, it will be a long while before a sufficient number of babies are born to allow meaningful analysis.  ALSPAC plans to look at the transgenerational effects of stressful events in the early life of ALSPAC parents.


Questions from Simon Harold (@sid_or_simon), Senior Executive Editor for the BMC Series.


More about the author(s)

Lars Bygren, Professor, Karolinksa Institute, Sweden.

Lars Bygren, Professor, Karolinksa Institute, Sweden.

Lars Bygren is a preventative health specialist in the Department of Biosciences and Nutrition at the Karolinksa Institute, Sweden. He was previously Professor and Head of the Department of Social Medicine at Umeå University, Sweden.

Marcus Pembrey, Emeritus Professor of Paediatric Genetics, University College London, UK

Marcus Pembrey, Emeritus Professor of Paediatric Genetics, University College London, UK





Marcus Pembrey is an Emeritus Professor of Paediatric Genetics at the Institute of Child Health at University College London, UK, a Fellow of the Academy of Medical Sciences, UK, and co-founded the International Federation of Human Genetic Societies. In addition to his research into rare genetic syndromes, Pembrey helped establish the Avon Longitudinal Study of Parents and Children (ALSPAC) with Jean Golding.

Research article

Change in paternal grandmothers´ early food supply influenced cardiovascular mortality of the female grandchildren

Bygren LO, Tinghög P, Carstensen J, Edvinsson S, Kaati G, Pembrey ME and Sjöström M
BMC Genetics 2014, 15:12

Go to article >>
  • jvkohl

    Re: “It might be relevant to mention that DNA methylation is mutagenic.”

    It might be more relevant to place that comment into the context of ecological variation and nutrient-dependent pheromone-controlled ecological adaptations. I’m not sure how to interpret the comment outside the context of biophysically constrained adaptations.

    Are the authors suggesting the DNA methylation-induced mutations, which perturb protein folding, may be involved in natural selection that leads to biodiversity? Is there a model for that?

    Kohl, JV (2013) Nutrient–dependent / pheromone–controlled adaptive evolution: a model Socioaffective Neuroscience & Psychology, 3: 20553

  • Pingback: Remaking Local Biologies in an Epigenetic Time | Somatosphere