Contribution
A design that tags one specific parental experience and follows it into offspring conceived afterwards. Because each odour activates a known, separate receptor, the paper can ask whether offspring respond to the odour the parent learned, not merely whether they are more anxious. It finds a double dissociation in behaviour, a matching change in the anatomy of the nose and olfactory bulb, persistence to the second generation and after IVF, and a methylation difference at the receptor gene in sperm.
Key insight
What passes is cue-specific and structural, and the authors decline to call it fear. Offspring of acetophenone-trained fathers had more acetophenone-responsive neurons and detected the odour at lower concentrations. The authors posit that the enlarged neural representation “may underlie” the sensitivity (Results, “F0 olfactory experience affects F1 neuroanatomy”). Startle to the odour is read “as an enhanced behavioral sensitivity to acetophenone, not necessarily an increase in fear” (Results, “Transgenerational olfactory sensitivity”). Their closing reading is that generations “inherit information about the salience of specific stimuli in ancestral environments” (Discussion).
Assumptions
- Odour-potentiated startle reads out sensitivity. The test was built for fear (fear-potentiated startle), and is used here for detection. Valence would need an appetitive-conditioning arm, which was not run.
- Acetophenone and propanol engage separate receptors. M71 neurons respond to acetophenone; propanol does not activate them, and the two odours’ glomerular maps do not overlap (from an online atlas and prior work). Olfr6 is taken as an unrelated control receptor.
- Shock is matched across arms. F0-Prop males were shocked as F0-Ace males were, so a general paternal stress effect should not produce an odour-specific result. Many comparisons use F0-Prop, not home-cage males, as the control for this reason.
- Lower CpG methylation means more transcription, so hypomethylated Olfr151 in sperm is taken as a candidate mark for more M71 neurons in the next generation.
- The F2 is a full generation removed. The authors grant that some hold true transgenerational inheritance needs F3 (Skinner 2008), and answer that F2 here are conceived from a father who was himself never exposed (Discussion).
Key results
- Design. F0: 2-month-old, sexually and odour-naive males, C57Bl/6J or M71-LacZ (whose M71 neurons stain blue). Home cage, or three days of odour (10 s, 10% acetophenone or propanol) co-terminating with a 0.25 s, 0.4 mA foot shock, five trials a day. No F0 male was excluded; all were mated with naive females. All tested offspring were 2-month-old males. Behaviour was double-blind and automated; anatomy used cohorts never behaviourally tested, so they stayed odour-naive (Online Methods; Results).
- F1 behaviour, startle (Fig. 1).
- C57: F1-Ace (n = 16) startled more to acetophenone than F1-Home (n = 13), t27 = 2.123, P = 0.043; no difference to propanol.
- M71-LacZ: F1-Home 11, F1-Ace 13, F1-Prop 9. To acetophenone, F2,30 = 6.874, P = 0.003, F1-Ace above both others. To propanol, F2,26 = 4.541, P = 0.020, F1-Prop above F1-Ace. A double dissociation.
- F1 behaviour, detection threshold (Fig. 2; n = 16 per group). F1-Ace avoided acetophenone at 0.03%, where F1-Prop did not; both did at 0.06% (P = 0.005, Bonferroni). F1-Prop avoided propanol at 0.003%, F1-Ace only at 0.006% (P = 0.0005).
- No general effects. No difference in anxiety on the elevated plus maze, nor in acquiring, consolidating or extinguishing auditory fear (Supplementary Figs. 2–3, not seen).
- F1 anatomy (Fig. 3). Dorsal M71 glomerular area, F2,91 = 15.53, P < 0.0001 (F1-Home n = 38, F1-Ace 38, F1-Prop 18); medial, F2,84 = 31.68, P < 0.0001. M71 neurons in the epithelium, F2,13 = 18.80, P = 0.0001 (n = 6, 6, 4). F1-Ace exceeded both controls each time. The unit of the glomerular n (animals, bulbs or glomeruli) is not stated in this copy.
- F2 (Fig. 4a–h). F2-Ace (n = 12) startled more to acetophenone than F2-Prop (n = 8), t18 = 2.664, P = 0.0158; F2-Prop more to propanol, t17 = 2.302, P = 0.0343. Dorsal glomeruli, t13 = 5.926, P < 0.0001 (n = 8 vs 7); medial, t14 = 4.44, P = 0.0006 (n = 10 vs 6).
- IVF (Fig. 4i–j). Sperm taken 10 days after conditioning, IVF done blind in a separate facility, embryos carried by CD-1 surrogates. Larger M71 glomeruli in F1-Ace-IVF: dorsal t37 = 4.083 (n = 16 vs 23), medial t33 = 5.880 (n = 19 vs 16), both P < 0.001. No behaviour was tested.
- Maternal line and cross-fostering (Fig. 5). Females were conditioned
before mating, never while pregnant, and pups were cross-fostered from
postnatal day 1.
- Offspring of acetophenone-trained mothers startled more to it than F1-Home (n = 16 vs 13), t27 = 2.362, P = 0.0256.
- Fostered by home-cage mothers, they still exceeded home-cage pups fostered by trained mothers: ANOVA F3,18 = 6.874, P = 0.0011, fostered comparison P < 0.05.
- Anatomy, with propanol-trained mothers as controls, held after fostering: dorsal F3,14 = 17.52, P < 0.0001; medial F3,15 = 5.933, P < 0.01. Group sizes were 3 to 8.
- Methylation in sperm (Fig. 6; bisulfite sequencing, done blind by
Active Motif).
- F0: Olfr151 less methylated in F0-Ace (n = 12) than F0-Prop (n = 10), t16 = 2.344, P = 0.0323; one CpG at its 3′ end, P = 0.003 (Bonferroni). Olfr6 did not differ (P = 0.54).
- F1 sperm: Olfr151 less methylated in F1-Ace (n = 4) than F1-Prop (n = 4), t14 = 2.763, P = 0.0153; two CpGs, P = 0.002.
- F1 and F2 olfactory epithelium: no difference (Supplementary Fig. 6).
- Sperm histone marks (acetyl-H3, H3K27me3) at the M71 locus: no difference (Supplementary Fig. 7).
- Effect sizes are not reported, nor are group means in the text. From t and the group sizes, my approximations of Cohen’s d: F1 startle 0.8; F2 startle 1.2; IVF dorsal glomeruli 1.3. These are computed here, not the authors’.
- Mechanism, by the authors’ account speculative. Odorants might reach the blood and activate odorant receptors expressed on sperm. With 13 days between the first conditioning day and breeding, the sperm used were precursors during conditioning, but “we cannot and do not claim to know what fraction” of sperm carry the information (Discussion).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Offspring conceived after a father’s odour–shock conditioning are more behaviourally sensitive to that odour and not to another | moderate: a double dissociation across two odours and two strains, with groups of 9 to 16 and P values from .0005 to .043 | Figs. 1–2 |
| C2 | The sensitivity is accompanied by more neurons and larger glomeruli for the conditioned odour’s receptor | moderate to strong: large, consistent anatomical differences across F1, F2, IVF and fostered offspring, blind quantitation | Figs. 3–5 |
| C3 | The effects persist to F2 and are carried by gametes rather than social transmission | moderate: F2 behaviour and anatomy; IVF anatomy only; cross-fostering only on the maternal line, with groups of 3 to 8 | Figs. 4–5 |
| C4 | Hypomethylation of Olfr151 in sperm may be the inherited mark | weak: one modest difference (P = .03) in F0 and in four-versus-four F1, absent in the tissue where the receptor is expressed, with no causal test | Fig. 6, Supplementary Figs. 6–7 |
| C5 | Generations inherit information about the salience of specific ancestral stimuli, which may bear on human phobia, anxiety and PTSD | weak: an interpretation and a translational suggestion, not tested | Discussion |
Concepts
- F0, F1, F2: the conditioned generation, their offspring, and the offspring of those offspring. Here every F1 and F2 tested was naive.
- transgenerational versus intergenerational: the paper uses the first for effects in generations never directly exposed, and notes the argument that a parent’s exposure may reach the F1 embryo and F2 germ line directly, so that F3 is the strict test.
- odour-potentiated startle (OPS): startle to a noise burst with the odour present, relative to startle alone.
- M71 / Olfr151: an odorant receptor, and its gene, activated by acetophenone. Each olfactory neuron expresses one receptor, and neurons with the same receptor converge on a glomerulus in the bulb.
- CpG hypomethylation: fewer methyl groups on cytosines followed by guanine, usually associated with more transcription.
Connections
- The question it is filed for (QUESTION-297). It is the paper behind the question’s “germline” route. Its own controls are aimed at separating that route from the social one, which in mice means maternal care and the mother’s exposure to a stressed mate.
- Reviews and critiques held but unread. Bohacek and Mansuy (LIT-1738) review this kind of work; Heard and Martienssen (LIT-1745), Horsthemke (LIT-1762) and Bird (LIT-1708) argue that germline reprogramming makes such transmission unlikely in mammals. Heard and Martienssen’s manuscript (PMC4020004) was searched for this paper and does not cite it. None of these was read for this NOTE, so the critique of Dias and Ressler is not assessed here.
- Memory through metamorphosis (QUESTION-302). Blackiston et al. (LIT-1680) trained caterpillars to avoid an odour; this paper trained fathers to fear one. The question links them. The link is looser than it looks: the moth keeps an aversion, while these offspring show greater sensitivity, which the authors decline to call aversion.
- Inherited non-genetic state (THEORY-179). A bacterial cell’s transient inherited state is the simplest relative: something about a parent’s condition, not its genes, sets the offspring’s.
Bearing on the record
- The owner’s thesis (CLAIM-491). The question’s framing (the
biological routes need no self; what passes is a state, not an
identification) is borne out here, and more sharply than the framing
says. These are inferences; the paper is about mice and says nothing of
selves.
- What passes is not even a state of fear. It is a tuning of the nose, more receptors for one odour, with a lower detection threshold. The authors refuse to say it carries the parent’s valence.
- Nothing in it requires the offspring to represent the parent, the event or a group. The design’s controls exist to exclude the social route, and the cross-fostering result is that rearing by a trained mother does not transfer the effect.
- So the paper bears on the identity route only negatively. If an ancestor’s injury becomes part of a descendant’s self, it does so by a route this kind of inheritance does not supply.
- The triviality objection (CLAIM-492). Not addressed.
- QUESTION-297. It should be cited as evidence for transmission of a cue-specific sensitivity, not of trauma or fear. That narrows what the germline route can carry on the best animal evidence the record holds.
- QUESTION-302. It weakens the analogy the question draws. If the reported offspring effect in swallowtails were real, this paper suggests that what would pass is heightened sensitivity to the odour, not the learned avoidance. That is an inference from one mouse paper.
- A dated note is proposed for QUESTION-297 (section 3). No dated note on CLAIM-491: the paper does not touch the thesis except by showing a route on which it is idle.
- No THEORY filed. One lab’s result, with a speculative mechanism and an unread critique.
- No ML instruction.
Limitations
- Small groups. Behaviour 8 to 16 per group; cross-fostering 3 to 8; F1 sperm methylation 4 per group.
- IVF tested anatomy only, so the cleanest exclusion of social and maternal routes for the father’s line has no behavioural result.
- Cross-fostering was on the maternal line. For the paternal line the social route is excluded by F2 and IVF, not by fostering.
- The methylation statistics may not be per animal. F0 has n = 12 and 10 but t16; F1 has n = 4 and 4 but t14. Those degrees of freedom fit the number of CpG sites queried (nine in F0, eight in F1) times two groups, minus two. If so, sites, not mice, were the unit. This is an inference from the numbers; the text does not say.
- Methylation was absent where it would act. No difference in F1 or F2 olfactory epithelium, and no histone difference in sperm. The authors propose other marks, untested.
- Only males were tested as offspring.
- No mechanism for how an odour learned by the father reaches sperm.
- Supplementary figures not seen, so the anxiety, auditory, MOE methylation and ChIP nulls are taken from the text.
- Oddities and typos.
- Fig. 1d reports F2,26 for groups of 11, 13 and 9 (33 mice, which gives F2,30 as in Fig. 1c). Fig. 4b reports t17 for 8 and 12 mice.
- Fig. 5b’s F3,18 = 6.874 repeats exactly Fig. 1c’s F2,30 = 6.874.
- The Methods define the OPS score as the odour trial’s startle subtracted from the leader’s, which would make potentiation negative; the order of subtraction is probably reversed in the text.
- “epidydymis” for “epididymis” (twice, N-ChIP methods); “phophorylated” for “phosphorylated”; “permit … or repress … to transcription”.
- Reference 54 (SAMtools) gives Bioinformatics volume “52”; it is 25.
Open questions
- Does the effect replicate in another laboratory, with per-animal methylation statistics and behaviour in IVF offspring?
- Does conditioning with an appetitive odour pass on the same sensitivity? If so, what passes is salience without valence, as the authors suspect.
- Is the effect present in F3, a generation whose germ line was never inside an exposed parent?