Contribution
A six-page survey of the evidence that what parents live through reaches their offspring through the germline and changes their behaviour. It covers rodent models of trauma, stress, drugs and in-utero insults; the newer finding that good environments are transmitted too; human studies of genocide and war; and the candidate carriers, chiefly sperm RNA. It is the proponents’ side, written by the laboratory behind the most-cited mouse trauma model (MSUS). It states its open problems plainly, and it does not engage the sceptics.
Key insight
The authors’ strongest evidence is causal and specific to RNA. Injecting RNA from the sperm of males traumatized as pups into fertilized control oocytes “recapitulated trauma-induced behavioral and metabolic symptoms in the resulting progeny” (p. 99). The same move works for enrichment: sperm RNA from enriched males reproduced enhanced plasticity and memory, and blocking miR-212/132 prevented it. So in mice something in sperm RNA can carry a parent’s experience to a child. What the review cannot say is how the experience gets into the sperm, or how it gets past the child: “it is not clear how the phenotypes associated with transgenerational inheritance are further passed on beyond the first generation” (p. 99).
Assumptions
- “True biological inheritance” means the germline (p. 96). Rearing, maternal factors and body fluids passed at birth or in milk can also transmit phenotypes; the review sets them aside by definition.
- A loose definition of the generations (p. 96). “Transmission from a directly exposed parent to the offspring is typically called intergenerational …, while transmission across several generations is called transgenerational.” Nothing says which generations were exposed as germ cells or fetuses.
- Stimuli are “purely emotional/psychological” and begin in the brain (p. 99). The problem is then a bridge from brain to germ cell.
Key results
- Definitions and the germline test (pp. 96–97).
- The definition above. A “hallmark” of epigenetic inheritance is a specific signature in the exposed animal’s germline, “generally accompanied by similar changes in somatic cells” of parent and offspring.
- Germline dependence has been shown by artificial insemination, IVF, and transfer of sperm RNA into fertilized control oocytes.
- Negative traits in rodents (p. 97; Table 1, p. 98).
- MSUS (unpredictable maternal separation combined with unpredictable maternal stress, applied to pups): depressive-like symptoms, risk-taking and antisocial behaviour “up to the third or even the fourth generation”; impaired object-recognition memory and synaptic plasticity; altered serotonergic circuits; metabolic changes in brain and blood. Transmission “through both fathers … and mothers”.
- Chronic social instability in adults: anxiety and social deficits “across three generations”.
- Odour fear conditioning in adult males (Dias and Ressler): “sensitivity to the same odor” in offspring and grand-offspring.
- Drugs: offspring of cocaine-addicted males have memory and plasticity deficits, rescued by D-serine in the hippocampus; four weeks of alcohol in female rats before conception exaggerate offspring stress hormone responses; paternal ethanol vapour lowers offspring ethanol preference.
- In-utero exposure, which the review says excludes “any social factors after birth”: valproic acid causes autism-like behaviour “across three generations”; immune activation causes despair, altered sociability and fear “across several generations”.
- Table 1 gives eight rodent studies with the number of generations affected: 3 for MSUS, social instability, odour conditioning and preconception corticosterone; 2 for the rest.
- Positive traits in rodents (pp. 98–99).
- Two weeks of enrichment from postnatal day 15 in males improves offspring long-term potentiation and memory; four weeks before conception gives offspring resistance to depression and a lower cortisol response to acute stress.
- Enrichment corrects inherited deficits in a Ras-GRF knockout and co-morbid anxiety in a genetic epilepsy model.
- Enrichment of MSUS fathers “shortly after chronic trauma” reverses their altered coping and “preserves their progeny from being affected”, with correction of DNA methylation at the glucocorticoid receptor gene in fathers’ sperm and offspring hippocampus (Gapp et al. 2016).
- The same trauma can pass on advantages: MSUS offspring show more behavioural flexibility and better goal-directed behaviour. The authors suggest this “may have important evolutionary implications” (p. 99).
- Humans (p. 97; Table 2, p. 98).
- Increased psychopathology “in the offspring and grand-offspring of holocaust survivors”, with FKBP5 methylation as “a possible epigenetic signature”.
- Genocide in Cambodia and Rwanda: hyper-arousal, anxiety and depression in descendants. Bosnian war veterans with PTSD: neurodevelopmental delays and emotional deficits in offspring. Finnish wartime evacuees: daughters’ psychiatric hospitalization.
- Sperm of men with adverse childhood experiences shows miRNA changes like those in stressed mice and their offspring. The annotation gives n = 28 men, with miR-449 and miR-34 inversely correlated with ACE scores (ref. 32).
- The authors’ verdict: germline dependence “is not established in humans but initial results indicate that it is likely the case”; the evidence is “concrete in both rodents … and humans”.
- Table 2’s own title: studies “of intergenerational … inheritance … in humans”.
- The FKBP5 annotation (ref. 27): Holocaust survivors n = 32, their offspring n = 22, comparison parents n = 8 and offspring n = 8. Methylation was higher in survivors and lower in their offspring.
- Mechanisms (p. 99).
- Carriers from brain to germline: circulating hormones, cytokines, circulating DNA or RNA. Maturing sperm acquire ncRNAs and proteins from epididymosomes, which “could bridge the brain to the germ-line”.
- Of DNA methylation, histone modifications and ncRNAs, “ncRNAs are the best described candidates today”.
- Causal tests: sperm RNA from MSUS males (Gapp et al. 2014); nine miRNAs altered by adult stress reproduce the effect on corticosterone (Rodgers et al. 2015); sperm RNA from enriched males, blocked by miR-212/132 inhibitors (Benito et al. 2018). The annotation to ref. 44 adds sperm tRNA fragments from high-fat-diet males causing metabolic disorder (Chen et al. 2016).
- Open problems (p. 99): the link from stimulus to germ-cell signature “remains unknown”; how effects pass beyond the first generation is unclear; more direct causal proof and “reversal” studies are needed; mechanisms may differ by stimulus, timing, severity, sex and time of analysis; human body fluids (blood, saliva, sperm, milk) should be assessed against “direct or ancestral exposures”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | In rodents, parental trauma, stress, drugs and in-utero insults pass behavioural phenotypes to offspring, in some models for three or four generations | moderate as a summary of many studies; no effect sizes, no replication record, and the generation counts’ convention is unstated | p. 97; Table 1 |
| C2 | In rodents the transmission depends on the germline, and sperm RNA can carry it | moderate: IVF and a handful of injection experiments, three of them on behaviour or stress hormones | pp. 96–97, 99 |
| C3 | Positive parental experience (enrichment) is transmitted too, and paternal enrichment can prevent the transmission of trauma’s effects | weak to moderate: a few studies; the prevention result is one study from the authors’ laboratory | pp. 98–99 |
| C4 | Human descendants of genocide and war show similar effects, and a germline route is likely | weak: by the review’s own table these are intergenerational; nothing in them separates the germline from rearing; the one sperm study (n = 28) has no offspring | p. 97; Table 2 |
| C5 | Non-coding RNAs are the best candidate carriers; circulating factors and epididymosomes may link brain to sperm | weak as to the bridge (a hypothesis, “remains unknown”); moderate as to ncRNA in sperm | p. 99 |
| C6 | Traits transmitted after trauma can be advantageous, with possible evolutionary implications | weak: two findings in one model, and a cited review | p. 99 |
Concepts
- intergenerational epigenetic inheritance (the review’s sense): transmission from a directly exposed parent to the offspring.
- transgenerational epigenetic inheritance (TEI) (the review’s sense): transmission across several generations. Looser than Heard and Martienssen’s, which requires a generation never exposed.
- MSUS: unpredictable maternal separation combined with unpredictable maternal stress, an early postnatal model of chronic trauma in mice.
- environmental enrichment (EE): housing with more space, objects and social or physical stimulation, used as a positive parental exposure.
- epididymosomes: extracellular vesicles in the epididymis that pass ncRNAs and proteins to maturing sperm.
- reversal study: correcting an inherited epigenetic change, by the environment or by molecular means, to test whether it causes the phenotype.
Connections
- Bohacek and Mansuy 2015 (LIT-1738), unread and Unavailable. Same senior author, four years earlier. This review cites it as the review of germline-dependent transmission and of the assisted-reproduction tests (ref. 3), and as a source for the MSUS row of Table 1.
- Dias and Ressler (LIT-1746, NOTE-1281). Reported faithfully in the text, as “sensitivity”, not fear. Table 1’s cell for it adds “CpG hypomethylation in the Olfr151 gene in sperm and olfactory cortex”. The record’s reading found the sperm result, no difference in F1 or F2 olfactory epithelium, and no reported cortical test.
- The three sceptics (LIT-1745, LIT-1762, LIT-1708;
NOTE-1254, NOTE-1265, NOTE-1264). None is cited. Germline
reprogramming, which is their main argument, is not mentioned. Weismann
appears only in the title of ref. 5. The two sides use the same
evidence differently:
- Horsthemke lists Gapp et al.’s sperm-RNA injection as technically sound but unconfirmed; here it is the “pioneering” causal link.
- Bird says the FKBP5 study rests on “a sample of only 32 people”. This review’s annotation gives 32 survivors, 22 offspring and 16 comparison subjects.
- Heard and Martienssen’s F2/F3 rule is absent. Without it, “three generations” cannot be checked against the rule (see Limitations).
- Yehuda et al. (LIT-1692). The FKBP5 study, the review’s lead human case, cited with its numbers.
- The Holocaust meta-analyses (LIT-1718, LIT-1770). Not cited. The review’s “offspring and grand-offspring” claim is the claim those meta-analyses test in non-clinical samples and find no general support for.
- Costa et al. (LIT-1696, NOTE-1255). Not cited. It is the record’s read human case on the paternal line (sons of prisoners of war), and its untested “grateful nation” line is used below.
Bearing on the record
- The question (QUESTION-297), germline route. The record now has
the proponents’ case from the proponents. These are inferences from the
text, not the authors’ claims.
- On mice the case is real but bounded. Sperm RNA carries something from father to child, and the authors call the next step, beyond the first generation, unclear. That is the step reprogramming is said to block. The two sides agree on more than their tone suggests: the intergenerational effect is supported, the transgenerational mechanism is not.
- On humans the review concedes the sceptics’ point in its own table. Table 2 calls every human study intergenerational. Its text says the germline route “is not established in humans”. Of the four tabled studies, the titles of two (refs. 28 and 29) name parental styles and family violence, which are the question’s parenting route. A proponents’ table of human epigenetic inheritance thus includes studies framed by their own authors around rearing. (From the titles; those studies are not read.)
- The best design is still not available in humans. The review’s suggested next step is to measure marks in body fluids against “direct or ancestral exposures”. That is correlational and does not meet Horsthemke’s causal step. The question’s first test is still unanswered from either side.
- Enrichment adds something new to the question. If paternal enrichment after trauma prevents its transmission in mice, then what passes depends on the father’s environment after the event, not only on the event. That is a mouse result from one study.
- The identity route. The review says nothing about it. On its account what reaches the germline is a bodily state: hormones, cytokines, circulating RNA. There is no place in it for what an event means to the person or the group. One possibility (an inference, untested) joins the routes: if an event’s meaning changes the stress response, and the stress response feeds the germline, then a group’s narrative could set even the biological route. Costa et al.’s “grateful nation” reading is the record’s only hint of this, and it is untested.
- The owner’s thesis (CLAIM-491). Indirect at most. Nothing transmitted here involves a self, an ancestor represented, or a group. The triviality objection (CLAIM-492) is not addressed. No note on the claim is proposed.
- Proposed dated note on QUESTION-297 (below, section 3).
- No THEORY filed.
- No ML instruction.
Limitations
- Not the 2015 review. Bohacek and Mansuy’s Nature Reviews Genetics article is unread, and this six-page piece cannot stand for its detail.
- No engagement with critics. Germline reprogramming, replication, confounding and the F2/F3 rule are not discussed. “Conflict of interest statement: Nothing declared” (p. 100). The authors review a model, MSUS, that is their laboratory’s own. Mansuy is a named author of three of its nine references (12, 15, 16); the others’ author lists are truncated. The proponents here are one school, as the sceptics are.
- The generation counts are ambiguous. Dias and Ressler tested sons and grandsons, yet Table 1 lists three generations affected. So the counts seem to include the exposed generation. If so, “three generations” after an in-utero exposure (valproic acid) would reach only the exposed mother’s grandchildren, an intergenerational case by Heard and Martienssen’s rule. The review does not say how it counts. This is an inference from one row.
- “Excluding any social factors after birth” (p. 97) holds for the exposure, not for rearing by exposed mothers or affected offspring.
- The human claims outrun their citations. “Offspring and grand-offspring of holocaust survivors” cites refs. 24–26. All three titles concern offspring only; none names grandchildren. Table 1’s Dias and Ressler cell names “olfactory cortex”, which the paper, as read by the record, does not report.
- No effect sizes or sample sizes in the text, except in three reference annotations.
- Typos.
- “inter- of transgnerational” (p. 97).
- “intergenerational oepigenetic inheritance” (Table 2 title, p. 98).
- “Increase psychiatric hospitalizations” (Table 2).
- “brain-derived neurotropic factor” for neurotrophic (p. 97).
- “adversive” for aversive (ref. 36 annotation, p. 101).
- “the geneFKBP5” (ref. 27 annotation, p. 100).
- “both genders” for both sexes (p. 99).
- Ref. 4 is dated 2018 and marked “(in press)”.
Open questions
- Does the MSUS effect in the third and fourth generations survive the F2/F3 rule, counted from the exposed pups’ germ cells?
- Has the sperm-RNA injection result been reproduced outside the laboratories that reported it?
- Do the men in the ACE sperm study (n = 28) pass anything to their children? The study has no offspring.
- In humans, does a father’s later environment, as enrichment does in mice, change what reaches his children?