Intergenerational mouse trauma
January 9, 2014 7:13 PM   Subscribe

...Ressler... and Dias wafted the scent [of acetophenone] around a small chamber, while giving small electric shocks to male mice. The animals eventually learned to associate the scent with pain, shuddering in the presence of acetophenone even without a shock. Despite never having encountered acetophenone in their lives, the offspring exhibited increased sensitivity when introduced to its smell, shuddering more markedly in its presence compared with the descendants of mice that had been conditioned to be startled by a different smell or that had gone through no such conditioning.
posted by latkes (34 comments total) 22 users marked this as a favorite
So these mice are Bene Gesserit, then?
posted by XMLicious at 7:22 PM on January 9, 2014 [2 favorites]

I can't help but chuckle at the prospect of listing "accomplished mouse traumatizer" on a resume.
posted by sonascope at 7:25 PM on January 9, 2014 [5 favorites]

Lysenko lives!
posted by spitbull at 7:32 PM on January 9, 2014 [5 favorites]

Nice work guys!
posted by ian1977 at 7:37 PM on January 9, 2014

Tracy Bale, a neuroscientist at the University of Pennsylvania in Philadelphia, says that researchers need to “determine the piece that links Dad's experience with specific signals capable of producing changes in epigenetic marks in the germ cell, and how these are maintained”.

I'd would have liked to know more about how the pups were raised. For instance, are they raised by their parents and exposed to the acetophenone in their presence, where something else the conditioned parents are doing — some behavioral or chemical cue that is not genetic — could trigger sensitivity in the offspring.
posted by Blazecock Pileon at 7:38 PM on January 9, 2014 [3 favorites]

I couldn't read the study in full but I heard about the story on This Week in Science and they were talking about how some of the baby mice were produced by in vitro fertilization (also mentioned in this link) and were actually raised on a different part of campus from their parents, but still responded with fear to the smell.
posted by latkes at 7:52 PM on January 9, 2014 [2 favorites]

Maybe this isn't the topic for you.
posted by Pogo_Fuzzybutt at 8:31 PM on January 9, 2014

I wonder if the experiment was properly blinded to prevent the experimenters from unconsciously influencing the offspring.
posted by benzenedream at 8:32 PM on January 9, 2014 [1 favorite]

[Guys, wishing physical violence on people is something that is Not Done here. Thanks. ]
posted by restless_nomad (staff) at 8:33 PM on January 9, 2014 [1 favorite]

This is quite remarkable. They even found particular epigenetic marks, hypomethylation of the gene that makes the olfactory receptor that senses acetophenone, in the sperm of the males. So somehow the hypomethylation of this gene is conducted all the way to the gametes?

I know very little of neurology, but maybe there's prior evidence for stimuli of receptors causing them to be hypomethylated, i.e. there's some regulatory mechanism that can change either neural cells? And somehow that same regulatory program gets triggered in the germ line? Maybe it happens all over the mouse's body? Such a speedy epigenetic regulatory mechanism could be very useful for adapting mice to quickly changing environments, if it really does exist.

Also remarkable that there's only two names on this paper, which is quite small for a team these days! And I can imagine that something this interesting would usually result in an entire lab chipping in to try to get it finished and published more quickly, especially with it using so many different lab and computer techniques.

I wait for replication by others, but they could have found fundamental and revolutionary biology.
I wonder if the experiment was properly blinded to prevent the experimenters from unconsciously influencing the offspring.
From the paper (behind paywall, unfortunately):
All behavior was performed in a double-blind manner and data acquired using automated computer software programs. We are grateful to S. Banerjee, R. Andero-Gali, D. Choi, J. Goodman and F. Morrison for help with ensuring double-blindness of data acquisition and analysis.
posted by Llama-Lime at 8:36 PM on January 9, 2014 [3 favorites]

I wonder if the experiment was properly blinded to prevent the experimenters from unconsciously influencing the offspring.

I'm sure it was. Some sorts of epigenetic effect has been demonstrated before. I know of some researchers that are trying to suss out the pathways in non-human primates - the big idea being able to detect and intervene earlier for people who are prone to anxiety disorders.

This is exciting stuff - but it really, really complicates things.
posted by Pogo_Fuzzybutt at 8:39 PM on January 9, 2014

Remember, "statistically significant" just means we get a p-value below a threshold that we feel comfortable rejecting the null hypothesis. In this study. We're not getting a meta-analysis of all the other experiments where mice were abused in smelly environments and the results did not reach significance.

The eagerness of the big impact factor journals to publish "breakthrough" articles leads to some unintended outcomes. XMRV 2.0?
posted by meehawl at 8:42 PM on January 9, 2014 [1 favorite]

Things like this really complicate my Little Wonder Book understanding of genetics. I recently read about hereditary transmission of functional prions, and obligate carriers of bacteria, which means that a "full" genome of some species must consist of DNA, methylation on that DNA, MTDNA, prions, plus bacterial and viral colonies.
posted by Joe in Australia at 8:47 PM on January 9, 2014

[One comment deleted. This needs not to become a thread all about the ethics of animal research, which is a fine topic for its own post and discussion.]
posted by taz (staff) at 10:18 PM on January 9, 2014

Remember, "statistically significant" just means we get a p-value below a threshold that we feel comfortable rejecting the null hypothesis. In this study. We're not getting a meta-analysis of all the other experiments where mice were abused in smelly environments and the results did not reach significance.
It's always good to hold new science in a true/false superposition, but these particular criticisms of leaning too hard on statistical significance and needing metanalysis don't apply to this type of science. There's dozens of p-values in this paper, and multiple lines of evidence pointing to something entirely new. It's not clear that a metanalysis would ever really be possible. The independent lines of evidence are far more convincing than repeating the same experiments would be (though the same experiments should definitely be repeatd). Only the first interesting thing is really reported in the summary article:
  • there's a transgenerational behavioral conditioning -- phenotype assay for startle response (and for those concerned for mouse wellbeing this amounts to a total of 15 mild shocks of quarter-second duration per father mouse, over the course of 3 days. It looks like it's just enough to cause a vocalization, but not anything like a tasering.)
  • the olfactory neurons express more of the olfactory receptor that is specific for the scent -- the researchers made transgenic mice that allow staining for exactly which cells express the olfactory receptor of interest (both imaging and quantitative comparisons of the protein's expression)
  • they show changes in the regulatory machinery for the olfactory receptor -- DNA hypomethylation at the start of genes typically leads to greater gene expression. And they see these differences not only in the fathers' sperm, but also in the childrens' sperm.
So this isn't a picture that hinges on a single test, and looking at figure 1, the initial phenomenon isn't even that strikingly statistically significant. What's far more interesting for me is the rest of the paper where they show greater expression, and changes in the regulatory signals.
posted by Llama-Lime at 10:26 PM on January 9, 2014 [2 favorites]

Do clones also show this response? Could it be an all-body DNA update? How quickly does the change show up in the sperm?
posted by five fresh fish at 10:51 PM on January 9, 2014

As the DNA that regulates the shock-scent response complex is changed in the cells that are replicating within that complex, it is also being changed in the complex that makes sperm cells?

posted by five fresh fish at 10:56 PM on January 9, 2014

I took a great class on epigenetics and neural development in college, and I recall them having found the indicators that there was a demethylation enzyme somewhere, but they hadn't found it yet. Has that changed?
posted by Punkey at 11:02 PM on January 9, 2014

Being pretty ignorant about biology, the thing I can't wrap my head around is what possible mechanism could there be that would map a specific change in the grown parent back into the appropriate part of the DNA, in a way that would then lead to the same change in the offspring.

It is as if there is a way to reversibly encode information from grown mice to sperm cells, when my very limited grasp of the mechanics says the mapping should be strictly forward only. How in hell does acetophenone know which part of the sperm cell to mess with?
posted by Dr Dracator at 11:11 PM on January 9, 2014

That, actually, is one of the primary ways that epigenetics is different from regular genetics. The Wikipedia article on epigenetics is here, but I'll try to do a more basic breakdown here.

So, let's start with the all-important Genetic Code, your DNA, the Blueprint of What Makes You You. It's called that because it contains encoded sequences for every single protein the body makes - RNA transcription enzymes look for marker tags, transcribe the gene you need into RNA, which is then carried out to ribosomes for translation into proteins. But, how does a cell know what it needs to become? Every single nucleated cell in your body (that is, every cell that isn't a red blood cell) carries not just the genes for their own function, but every gene for every cell in your body. So, how does a liver cell know how to be a liver cell? What stops a skin cell from accidentally replicating into a neuron? The answer is (mostly) epigenetics.

Okay, that's great, but what does epigenetics do and how does it work? Well, the start for both of those answers is in the name - epi-genetics, or above/over/on genetics. What is "above" the genetic level? The structure of how the genetic code is packaged. Like many things, the depiction of how DNA is held in our bodies is vastly oversimplified before major-focus college courses. DNA doesn't just float freely in the nucleus (unless you're a prokaryote that lacks a nucleus, in which case loops of DNA just kinda hang around in your endoplasm) - it's wrapped around proteins called histones like a really long piece of yarn wrapped around a series of spiked balls. This combination of DNA and histones is called a nucleosome, which forms into long chains called chromatin, and when you package a bunch of chromatin together, you get a chromosome. However, there are two different ways that chromatin can be packaged together - heterochromatin and euchromatin. Heterochromatin is tight and compact, and doesn't allow for much in the way of transcription to RNA, while euchromatin is the yarn-and-balls (or beads-on-a-string) arrangement that's nice and open and allows for the transcription enzymes to get up in there and do their thing. You can have some sections of chromatin be all tight and bunched up into heterochromatin, while further down the line everything's loose and free for transcription.

It's modifying and regulating that structure is where epigenetics does its thing. By adding chemical functional groups to the DNA or the histones, your body can regulate which genes are available or unavailable for transcription. Basically, if you're a liver cell? You don't need that shit for producing serotonin getting in your way, so you've had the area for that gene blocked off for you. That can be done by modifying the chromatin structure where the gene that makes serotonin lives by attaching things to the histones, converting it from euchromatin to locked-down heterochromatin or vise versa, and by slapping methyl (CH3-) groups onto the outside of the phosphate section of the DNA itself. Both of these make it nearly impossible for transcription enzymes to get ahold of the DNA and do their work, and effectively blocks off the gene from being transcribed. Different types of cells have different epigenetic "encoding" that determines which genes are available.

Of course, this article isn't talking about cellular differentiation, which is where epigenetics starts to really get strange. Unlike your genetic code, your epigenetic "code" can be changed over time. We have found both internal and environmental factors that will change the epigenetic "encoding" on your cells, both in specific parts of the body and globally. Epigenetic modifications caused by cigarette smoking, for example, have been shown to be a contributing factor to lung cancer, and epigenetic inactivation of tumor suppression genes has been found in a wide variety of cancers. Certain epigenetic factors are heritable as well - for example, research is indicating that at least some of the biological roots of obesity are epigenetic, and those epigenetic changes are heritable.

Most interestingly (for me, at least) and relevantly to the topc, epigenetics seems to play a large role in neural development, well into adulthood. The genes that are activated for memory formation, for example, have been found to be epigenetically regulated, and not only have traumatic mental disorders, psychotic disorders, addiction, and PTSD been shown to have an epigenetic root in the brain, but these epigenetic changes have also been shown to be heritable across generations.

And that's what the root of this study is looking for - heritable epigenetic modifications in neural development. Since, well, it's sadly easier and most effective to be horrible to experimental animals than it is to be nice to them, they traumatized the shit out of some mice, and then bred them. The offspring for two subsequent generations inherited the fear of the scent through epigenetic modifications that were passed down from parent to child. Specifically, they found that a gene - Olfr151 - was hypomethylated when compared to the control mice. Since DNA methylation suppresses gene expression, that means that the test mice produced more of the protein encoded by Olfr151 than the control mice did, which should make them more sensitive to that smell. Now, epigenetics isn't nearly as stable as genetics is, so it went away, but imagine if this was the scent of a new environmental hazard or predator - you'd want to get really good at remembering to be freaked out by that scent. Epigenetic modification allows being scared by that scent to be passed down from parent to child without having to wait for random genetic mutation to do its thing.

Now, if you really want to get weird, let's talk about drugs that can modify epigenetics. We have at our disposal an array of drugs that can selectively or non-selectively alter epigenetics - mostly histone acetylases and deacetylases (acetylation increases transcription, deacetylation decreases transcription). Cancer treatments are already being used that primarily work on altering the epigenetics of cancerous tumors, and animal studies on memory, trauma and epigenetics have already been done on lab animals. One of the papers I had to do a summary on demonstrated that traumatic mental issues in rats could be blocked by a dose of a drug that stopped the epigenetic modifications leading to the trauma - the memory survived, but the trauma simply never formed. Whether or not that's a good thing is up to you to decide, but it's something that we're on the verge of being able to demonstrate on humans. A "morning after" pill for PTSD? Epigenetic treatments to increase leptin sensitivity and help treat obesity? These are all things that might be possible.
posted by Punkey at 12:36 AM on January 10, 2014 [26 favorites]

How in hell does acetophenone know which part of the sperm cell to mess with?

Chemicals don't "know", they just react with whatever they can, wherever they can.

Your genetic library isn't made of up "open books" — it all gets squeezed into a very small space with the help of proteins called histones.

However, the books aren't all uniformly squeezed; some books remain "open" for reading, some are "closed" off. Having some books open and others unreadable is how you get some genes controlled in different parts of your body, which make different organs like your heart, brain, liver, etc.

Some genes make heart muscle, and those muscle cells have had certain areas opened up and closed down to use those genes. Other genes make nerve cells, and those nerves have had different areas opened and closed for those genes.

There are chemical changes called "histone modifications" which determine how those different areas get opened and closed. All of this is part of a larger body of biology called "chromatin remodeling" (this is what my lab works on).

There are different kinds of chemical modifications to histones that are possible. The modification mentioned in this mouse paper is methylation, or the addition of a methyl chemical group to a specific part of the histone. This changes the shape and behavior of the histone — depending on where the methyl group is placed on the histone, it will squeeze the DNA some more, hiding genes away and keeping them from being read, or it could open it up, making genes easier to read.

Acetophenone appears to selectively mold the histones that are related to "olfactory" or smell functionality that associates with fear conditioning (electric shocks). A related set of histones in sperm cells might be manipulated in a similar way, which after fertilization perhaps leads to a developmental change in the offspring, such that their olfactory system or brains are hardwired to be sensitized to the chemical, to associate the chemical with pain in the same way as their parents.

That part of the story is less clear, which is Tracy Bale's objection. There is evidence for inheritance but less (some, but less) evidence of how the sperm cell is modified, which is directly causal. There could be other chemical pathways that acetophenone is involved in, which make other changes not measured that still lead to the inherited outcome.

It's an interesting story, if it bears out. It does makes sense for mice, which have a large olfactory system, to have evolved ways to more quickly adapt to the smell of a dangerous environment, faster than simple genetic mutations can allow.
posted by Blazecock Pileon at 12:50 AM on January 10, 2014 [7 favorites]

Wow. Put me in the skeptical group. The sperm data I can kinda understand. But since maternal imprinting occurs during oocytogenesis, and oocytogenesis is complete either before or shortly after birth, how the hell do you explain the finding that "similar experiments showed that the response can also be transmitted down from the mother"??

I can't access the original paper right now, so perhaps I'm missing something.
posted by kisch mokusch at 1:06 AM on January 10, 2014

kitsch, I don't have direct access to the paper in the FPP, but here's a quote from the obesity paper I linked above on various theories for how obesity could be epigenetically passed from mother to fetus:

"The mechanisms by which nutritional challenges affect the risk of disease in later life are poorly understood. However, evidence indicates that the establishment of the epigenome can be affected by environmental factors during critical developmental periods [69]. Possible disturbances of methylation may arise during foetal development due to lack of availability of dietary methyl donors [70–72]. Potential interactions between the environment and epigenetic mechanisms mediating the expression of genes associated with increased BMI and adiposity, may also be possible as suggested for; the FTO locus is a DNA-demethylase enzyme [23], the MC4R gene which has reduced methylation following long-term exposure to a high fat diet [73], the PPARγ protein which interacts with histone acetyltransferases [58] during adipogenesis and on the effect of diet on methylation of POMC [74] and Leptin [75]."
posted by Punkey at 1:10 AM on January 10, 2014

Which means that the mice would need to be pregnant at the time of the electric shocks. If so, I would suggest that there a a few alternative explanations as to why the offspring display strange behaviour not seen in untreated controls.
posted by kisch mokusch at 1:15 AM on January 10, 2014

Not if the epigenetic changes are globally induced. Like I said above, epigenetics can be altered in a single animal, and even in a single cell.
posted by Punkey at 1:18 AM on January 10, 2014

Thanks for all the information - I'm going to read both comments more closely, but this is the hairy part I think:

Acetophenone appears to selectively mold the histones that are related to "olfactory" or smell functionality that associates with fear conditioning (electric shocks). A related set of histones in sperm cells might be manipulated in a similar way, which after fertilization perhaps leads to a developmental change in the offspring, such that their olfactory system or brains are hardwired to be sensitized to the chemical, to associate the chemical with pain in the same way as their parents.

I wouldn't expect acetophenone to just randomly happen to affect the histones in a way that would result in a fear of acetophenone itsself in the offspring - that's what I meant by "knows". It makes more sense to suppose that mice already have a generic way of encoding fear of smells in their epigenetic information, in which case one would expect this result to also be reproducible with other smells.
posted by Dr Dracator at 1:20 AM on January 10, 2014

I don't think they were implying that acetophenone is a magic epigenetic chemical, and I'd imagine this would work with any distinctive scent. Acetophenone is a pretty aromatic chemical, they probably chose it because it's a good strong and distinctive smell. I wish the Nature article published what the control mice were exposed to, or if they just zapped the shit out of them without a scent component.
posted by Punkey at 1:24 AM on January 10, 2014

Not if the epigenetic changes are globally induced. Like I said above, epigenetics can be altered in a single animal, and even in a single cell.

True, but then the mechanisms must be distinct for males vs. females. In males, the authors are arguing that the sperm is affected, whereas for females, the effect cannot occur directly on the oocyte, but must occur indirectly via epigenetic changes to somatic cells.

Same outcome. Two distinct biological pathways. Both unknown. Please forgive my skepticism!
posted by kisch mokusch at 1:27 AM on January 10, 2014

Don't worry, the Nature article isn't especially clear on this, but that's not quite what the paper itself shows. All they did was isolate the females from the equation by doing an in virtro fertilization with sperm taken from males that had been sensitized to acetophenone. Why one of the authors felt it was necessary to comment on sperm cells expressing olfactory genes is beyond me - simple epigenetic inheritance through modification of the sperm germ line is way more simple.
posted by Punkey at 1:32 AM on January 10, 2014

So -- so they basically have a Lamarckian smell reaction? Whoa.

This is crazy, amazing, and super exciting because it brings me one step closer to my dream of holding several generations of puppies up in the air to create, eventually, a Lamarckian Hover Dog. This will be my finest hour.
posted by Mrs. Pterodactyl at 3:44 AM on January 10, 2014 [6 favorites]

Events in the environment cause certain genes to express.

Gene expression is modulated by methylation.

Methylation impacts mutation (as you would expect, from crap stuck on the genome, and and you see, if you map SNP rates to methylation rates at individual loci).

Lamarck was not entirely wrong, but we pretend he was because Eugenics was indeed pretty awful.

(BTW, the SNP/Methylation link isn't even minor or subtle.)
posted by effugas at 4:45 AM on January 10, 2014

Let's be fair: Lamarck was mainly wrong. Environmental changes or changes in physiology are still, overall, much less heritable (and much less likely to be heritable) than genetic changes. And if what's described in the FPP turns out to be a real phenomenon then it would best be understood within the larger context of natural selection, not really as an "alternative."
posted by en forme de poire at 3:57 PM on January 11, 2014 [1 favorite]

Finally read the paper. They did reverse the experiment in sexually naive female mice prior to conception and still saw the same effect. My concern stands regarding differential mechanisms between male and female mice. Furthermore, the F2 data has all of the conditioned group displaying increased startle response, when I would have expected a bimodal distribution based on the predicted mendelian inheritance of the epigenetically-modified locus. In other words, F2 mice should have shown a less homogeneous phenotype. Thus, it seems that yet another biological mechanism is required to explain the F2 data (epigenetic "conversion" of maternally-derived genes by to original, paternally-derived and epigenetically-modified gene). Best to wait for the independent confirmation, I think.
posted by kisch mokusch at 4:19 AM on January 12, 2014 [1 favorite]

Comments from the pubmed page:
"The statistical tests in the paper, both for the behavioral measurements as well as for the size of the M71 glomeruli , use as n, number of samples, the number of F1 and F2 individuals. This would be fine if the individuals were actually independent samples. However, they arise from a presumably small number of FO males. The numbers of FO males are not given in the paper. This is a major concern given that there is a lot of variability in the levels of expression of olfactory receptors in these mice that might be inheritable. As an example, for Figure 1a, the authors compared 16 F1-Ace-C57 mice with 13 F1-Home-C57 mice and find a p value of 0.043, with 27 degrees of freedom . But these 29 mice could have been originated from a very small number of FO mice. The actual n that should be used for the statistics in the whole paper are not the individual number of F1 or F2, but it should be the number of founding F0, for both groups. So the actual p values are larger than reported. Without the information about the size of the F0 populations used in each figure panel, it is hard to interpret the results."
Comment 2:
"An excellent point and discussed in Lazic SE, 2013 and references therein.

Also, it is not clear why the dorsal and medial glomerulus have different n for what appears to be same animals. These discrepancies occur for all histological data (e.g. Fig4 panel (i) has n=23 and n=16 for dorsal while panel (j) has n=16 and n=19 for ventral)."
posted by cashman at 2:38 PM on February 8, 2014 [1 favorite]

« Older Volcanic Lightning in a lab   |   Designer locomotives? You bet. Newer »

This thread has been archived and is closed to new comments