Tuesday, March 15, 2016

Obesity and diabetes: Actual epigenetics or just IVF?

This press release that appeared in my newsfeed titled "You are what your parents ate!" caught my eye because I'm a new mom of a new human and also because I study and teach human evolution.

So I clicked on it.

And after that title primed me to think about me!, the photo further encouraged my assumption that this is really all about humans.


"You are what your parents ate!"

But it's about mice. Yes, evolution, I know, I know. We share common ancestry with mice which is why they can be good experimental models for understanding our own biology. But we have been evolving separately from mice for a combined total of over 100 million years. Evolution means we're similar, yes, but evolution also means we're different.

Bah. It's still fascinating, mice or men, womice or women! So I kept reading and learned how new mice made with IVF--that is, made of eggs and sperm from lab-induced obese and diabetic mouse parents, but born of healthy moms--inherited the metabolic troubles of their biological parents. And by inherited, we're not talking genetically, because these phenotypes are lab-induced. We're talking epigenetically. So the eggs and sperm did it, but not the genomes they carry!

This isn't so surprising if you've been following the burgeoning field of epigenetics, but it's hard to look away. This fits with how we see secular increases in human obesity and adult-onset diabetes--it can't be genomic evolution, it must be epigenetic evolution, whatever that means!

As the press release says...
"From the perspective of basic research, this study is so important because it proves for the first time that an acquired metabolic disorder can be passed on epigenetically to the offspring via oocytes and sperm- similar to the ideas of Lamarck and Darwin," said Professor ...
Whole new ways of thinking are so exciting.

Except when you remember a two-year-old piece by Bethany Brookshire (because you use it to teach a course on sex and reproduction) which explained something that suggests we may have a major experimental problem with the study above.

In IVF, the sperm gets isolated (or "washed") from the semen.

You know what happens, to mice in particular, when there's no semen? Obesity and other symptoms of metabolic syndrome! There are placental differences too. This was published in PNAS.


"Offspring of male mice without seminal fluid had bigger placentas (top right) and increased body fat (bottom right) compared with offspring of normal male mice (left images)" from The fluid part of semen plays a seminal role by Bethany Brookshire.

So I went back to look at the original paper that the press release with the donut lady was about. I wanted to see if they are aware of this potential problem with IVF and whether it explains their findings, rather than the trendy concept of epigenetics...

So even though they titled it "Epigenetic germline inheritance of diet-induced obesity and insulin resistance," I wanted to see if they at least accounted for this trouble with semen, like how it's probably important, how its absence may bring about the same phenotypes they're tracking, and how IVF doesn't use semen.

But I don't have access to Nature Genetics.

Who has access to Nature Genetics, can check out the paper, and wants to write the ending of this blog post?

Step right up! Post your work in the comments (or email me holly_dunsworth@uri.edu, and please include a pdf of the paper so I can see too) and I'll paste it right here.

Update 12:19 pm
Two very good comments below are helpful. Please read those.

I'll add that I now have the pdf of the paper (but not the Supplemental portion where all the methods live and other important information resides). This quote from the second paragraph implies they do not agree with the finding of (or have forgotten about) the phenotypic variation apparently caused by sperm washed of their seminal fluid:
"The use of IVF enabled us to ensure that any inherited phenotype was exclusively transmitted via gametes."
As the second commenter (Anonymous) pointed out below, there does not appear to be a comparison of development or behavior between any of the IVF mice and mice made by mouse sex. So there is no way to tell whether their IVF mice exhibit the same metabolic changes that the semen/semenless study found. Therefore, it is neither possible to work the semen issue into the explanation nor to rule out its effects. Seems like a missed opportunity.

Completely unrelated and inescapable... I'm a little curious about how the authors decided to visualize their data like this:


9 comments:

  1. Like almost everyone who cites Neel's 'thrifty genotype' hypothesis doesn't really understand what he said (and later clearly disavowed). But that's just a matter of faddy sloganeering, to give something a sexy sheen of evolutionary relevance.

    A more important problem that the authors don't seem to have checked for is that the parental gonads basically 'scrub' their sperm and egg genomes of epigenetic marking, except in particular regions. This is an active area of research, but raises questions about how or why specific epigenetically marked gene regions don't get scrubbed. In this case, it does seem likely that marking had survived this, however that happens. Your observation about the Brookshire paper (or its reference to the Bromfield PNAS paper) seems cogent, and while the authors of the new paper do cite Bromfield, I can't on a quick reading see if the current paper has somehow circumvented that issue.

    The authors do conclude by clearly stating that one needs to satisfy the 'gold standard' for epigenetic inheritance, which is to see the effect in a 3rd generation, when the intermediate generation has not been exposed to the obesity generating diet.

    ReplyDelete
  2. Ken's post does a great job. I'll just add that all F1 mice evaluated were generated via IVF (with washing of the sperm to remove seminal fluid). The observed differences in body weight and glucose sensitivity should not be due to differences in conception/implantation.

    The researchers do cite the Bromfield PNAS study (once) in describing how previous studies failed to account for effects by seminal fluid and maternal reproductive tract. They don't mention, however, that their F1 mice should be more obese than F1 mice that would have been generated without IVF.

    ReplyDelete
  3. Did the study pdf show public funding? Did you see some proprietary data or some other reason to not make it freely available?

    ReplyDelete
  4. The study cites two 'Hemholtz' sources, and one German public (government) agency, so at least in part it was publicly funded. Of course, Nature Genetics is a for-profit company, so the lack of availability is understandable (which doesn't mean justified).

    ReplyDelete
  5. So I'm curious to follow up on your comment about the violin plots (partly because I'm working on some graphs right now). I find them more informative than standard boxplots, but I could certainly be missing something. Do you find them misleading?

    ReplyDelete
  6. You may want to get the pdf via Sci-Hub.

    For example, this is the url of the article:
    http://www.nature.com/ng/journal/vaop/ncurrent/full/ng.3527.html#access

    Google Sci-hub and paste the url, voila.

    ReplyDelete
  7. The study was internally controlled, i.e. they had control IVF offspring versus IVF offspring from fat parents. This shows that the difference in weight between the offspring of normal parents versus offspring of fat parents cannot be solely down to effects of IVF.

    I'm not sure what your query is regarding the graph - it's a slightly more informative version of a box and whisker plot that lets you visualise the entire distribution of a data set rather than just the mean + standard error bars.

    ReplyDelete
  8. Right and I believe this is what is said above. However, by not including a non IVF control, they have not teased these issues apart. That they did not do that is, to a novice (me) a shame.

    As for your second comment, rorschach could give you better insight than I and regular readers know why this is topical.

    ReplyDelete
  9. Hi Ellen, They look like female genitalia. I pointed it out because it's a link to my real area of expertise. Something to ground me? Ask Freudschach for more.

    ReplyDelete