""We believe that iterative somatic transmission (through bioactive compounds via the
placenta and breast milk) of behavioral traits is a prominent intergenerational
and multigenerational mechanism."
"Missing heritability" and non-DNA mediated transmission of behavior across generations:
Dr. Toth’s Lab page
Mitchell et al, “Behavioural traits propagate across generations via segregated iterative-somatic and gametic epigenetic mechanisms,” Nature Communications May 2016
Dr Toth, I was fascinated by your recent paper, and I’m curious to know how you decided to examine non-genomic generational transmission of traits.
As often happens in research, the decision was based on a serendipitous finding. Several years ago a graduate student in my lab, studying the behavioral effects of a serotonin receptor mutation, compared mutant and genetically normal mice. However, in contrast to standard practice, she used two groups of genetically normal mice, one derived from mutant mothers and the other from normal mothers. Of course, we expected that mutant mice will be different from either normal control. Instead, she found that both the mutant and genetically normal offspring of mutant mothers were different from the offspring of normal mothers.
After a little “lab crisis” and additional experiments to exclude a mix-up in genotyping, we had to conclude that the offspring’s behavior was not related to their own mutation but rather to their mother’s mutation. Additional experiments showed the behavioral traits are propagated across several generations.
What’s wrong, or at least limited, with the current genomic inheritance model of psychiatric diseases and behavioral disorders?
It cannot explain the familial aggregation of many psychiatric symptoms and diagnoses. “Missing heritability” refers to this phenomenon. Non-genetic inheritance likely contributes to this unexplained fraction of heritability.
How can the germline epigenetic and extra-genomic information carry instructional errors to the next generation?
Initially it was hypothesized that epigenetic marks in gametes at certain genomic locations are exempt from genome-wide epigenetic erasure and reprogramming between generations, providing the means for the continuous transmission of ancestral information across generations. However, it has been difficult to find these putative gametic epigenetic marks, such as DNA methylation and histone modifications. An interesting recent development in this story is the identification of small RNAs in gametes that carry the ancestral information. However, the gametes do not produce the small RNAs but rather acquire them from other non-gametic i.e. “somatic” tissue during their maturation.
Extending the idea that bioactive molecules (and not chromatin per se) carry ancestral information, one can imagine the gamete-independent transfer of immune molecules and metabolites or other bioactive compounds from mother to offspring via the placenta and breast milk. Indeed, this is the predominant mechanism we found in our experiments. Since this somatic mechanism recreates the maternal condition in the F1 offspring, it is repeated in the F2 generation and beyond. We call this mechanism iterative somatic programming.
Is there a mechanistic or other difference between maternal and paternal transmission of non-genomic instruction?
If we strictly speak about gametic epigenetic transmission, the majority of germline chromatic proteins is replaced during spermatogenesis, limiting the scale of putative gametic epigenetic transmission via sperm. If we include RNA molecules and other bioactive molecules carried by the gametes, the bigger cytoplasm of the oocyte may increase the chance of transmission.
On the other hand, oocytes, once produced in the embryo, remain dormant until ovulation/fertilization, and thus less likely to be affected by various “stressors” through life. Sperm is produced through life and could be exposed to more environmental factors.
Finally, we believe that iterative somatic transmission (through bioactive compounds via the placenta and breast milk) of behavioral traits is a prominent intergenerational and multigenerational mechanism because of the wide range of possible molecules involved and because the period of environmental exposure is extended to embryonic and early postnatal development, two of the most vulnerable periods in life.
What are the most vulnerable periods for the germline during both male and female gametogenesis, and why? How can various stressors on the germline and somatic tissues across its lifecycle have aggregate effects?
As mentioned above, the male germline/sperm is potentially exposed to more environmental effects, although a dormant oocyte may also be vulnerable. Regarding the aggregation of various exposures, experiments with rodent as well as epidemiological studies in human indicate the compounding effects of early life adversity and later life stress. Consistent with this notion, history of childhood maltreatment increased the level of epigenetic marks in blood cells in individuals with adult trauma related PTSD.
Would you briefly explain the findings of your latest paper? What stressors were imposed on the gestating females?
Females had an inflammatory/autoimmune-like condition that was transmitted to their offspring and even grandoffspring, together with an anxiety-like condition in mice. This model reproduces the higher incidence of autism and other neuropsychiatric diseases, associated with maternal gestational infection and autoimmune disorders. In our experimental model, inflammation and autoimmunity were caused by a genetic manipulation, specifically by the deletion of a neurotransmitter receptor. While we do not exactly know how the receptor-loss leads to inflammation, the phenotype is highly reproducible and less variable than inflammation caused by infectious agents.
What outcomes did you see in the F1s and the F2s, in terms of behaviors, molecular findings, and brain differences?
The most prominent behavior was heightened innate fear, interpreted as anxiety-like behavior. It is the higher than normal avoidance of open spaces, similar to the avoidance of stressful situations and social interactions of individuals with various anxiety disorders. This behavioral change was accompanied by epigenetic changes in neurons in an area of the brain that has been linked to anxiety. Although a direct causative relationship between brain epigenetic and behavioral changes is difficult to establish, the F1 and F2 offspring exhibited the same behavioral and epigenetic changes while the F3 lacked both. Moreover, the epigenetic changes were localized in genes that encode synaptic proteins. This makes sense given the involvement of synaptic proteins in fear and anxiety.
How do you think anxiety and stress can alter DNA methylation at enhancers within genes relating to synaptic function?
I would pose the question the other way, that is, we believe that epigenetic changes at enhancer regions misregulate synaptic genes that in turn bring about anxiety.
Were your findings similar to those found in other recent studies, such as the one from Dr. Mansuy's lab, "Transgenerational transmission and modification of pathological traits induced by prenatal immune activation”?
Prenatal or “in utero” immune activation (i.e. by administering an agent that induces inflammation in the mother and fetus) is a common starting point of many studies, including the above mentioned and ours. However, most, including the Mansuy, studies, are primarily focused on further transmission via the paternal line, while our study follows transmission through the maternal line.
The main reason for favoring paternal transmission is that transgenerational transmission can be established following two generations, while three generations are needed for maternal transmission. Also, it is much easier to obtain sufficient amount of sperm than oocytes DNA or chromatin for epigenetic studies.
Even less research has been done regarding maternal transmission via non-gametic mechanisms, such as iterative somatic programming across generations. The advantage of studying transmission via the maternal line is that parallel somatic and gametic mechanisms can be revealed and their interactions can be studied.
What do you mean by generation-dependent penetrance?
The same phenotype can be more or less robust in consecutive generations. For example, anxiety was most pronounced in F2s, and males, in general, had stronger phenotypes. This indicates that paradoxically, certain phenotypes become more apparent as the distance from ancestors is increased.
Could you explain reasons for sex differences in behavioral outcomes, and pleiotropy?
It is known that the male fetus and neonate are more vulnerable to gestational and perinatal insults, presumably because of the effects of sex hormones on brain development. In other words, females are probably more resilient to early life adverse events. This sex difference, combined with generation differences as mentioned above, generates pleiotropy in the behavioral phenotype. Pleiotropy of psychiatric disease symptoms, even within the same family, is very noticeable, and we believe that the sex and generation dependent manifestations of non-genetically transmitted traits may explain it, at least partly.
Some human studies suggest the transmission of parental behavioural/psychiatric conditions to the grandchildren— could you explain? Some examples were increased vulnerability of adult children and grandchildren of Holocaust survivors to psychological distress and increased incidence of anxiety, depression, schizophrenia, autism and attention-deficit hyperactivity disorder in the progeny with ancestral maternal stress and infection.
These are epidemiological studies and cannot establish causation and cannot even exclude the role of genetic factors. Nevertheless, because of the difficulty to link the transmission to genetic factors, it raises the possibility that maternal and sometimes paternal psychiatric diseases and/or trauma are transmitted to the children and grandchildren.
In a human clinical context, what pregnancy exposures might concern us? Which one might we suspect could lead to serious pathology? My work is currently focused on smoking and certain drugs, including endocrine-disrupting synthetic hormones and anesthesia.
These environmental toxins are definitely a concern during gestation. I would add maternal serious infection and stress as additional factors. A Danish study showed that the incidence of autism in children is increased fourfold when the mother was hospitalized with severe infection during pregnancy. This means that a typical mild virus or bacterial infection has no significant effect, but more serious cases that require hospitalization and which are associated with more sever inflammatory responses, may increase the chance of disease. Still, this situation would be rare because most children will not be affected when exposed to significant inflammatory environment in utero.
What are some other implications of your research? What do you think it might mean for future research and public health generally?
I would again emphasize the bias towards gametic mechanisms when we think about non-genetic multigenerational transmission. This is understandable because gametes represent a plausible vehicle to transfer ancestral information. However, somatic transmission of bioactive molecules such as cytokines and hormones, and more recently metabolites, represent another pathway which may actually be more prominent in maternal transmission, given the length of possible exposure during gestation and early postnatal life. Ultimately, genetic, together with gametic and somatic non-genetic mechanisms determine the individual’s vulnerability to neuropsychiatric disease.
"We had to conclude that the offspring’s behavior was not related to their own mutation but rather to their mother’s mutation."
"Extending the idea that bioactive molecules (and not chromatin per se) carry ancestral information, one can imagine the gamete-independent transfer of immune molecules and metabolites or other bioactive compounds from mother to offspring."
"We believe that iterative somatic transmission (through bioactive compounds via the placenta and breast milk) of behavioral traits is a prominent intergenerational and multigenerational mechanism."
"Females had an inflammatory/
autoimmune-like condition that was transmitted to their offspring and even grandoffspring, together with an anxiety-like condition in mice."
"This behavioral change was accompanied by epigenetic changes in neurons in an area of the brain that has been linked to anxiety."
"We believe that epigenetic changes at enhancer regions misregulate synaptic genes that in turn bring about anxiety."
"For example, anxiety was most pronounced in F2s, and males, in general, had stronger phenotypes."
"Pleiotropy of psychiatric disease symptoms, even within the same family, is very noticeable, and we believe that the sex and generation dependent manifestations of non-genetically transmitted traits may explain it, at least partly."
"These environmental toxins are definitely a concern during gestation. I would add maternal serious infection and stress as additional factors."
"Ultimately, genetic, together with gametic and somatic non-genetic mechanisms determine the individual’s vulnerability to neuropsychiatric disease."