So the second example we're going to think about in terms of a change in human

diet, potentially having an outcome in the offspring is the Överkalix data.

So, Överkalix is a relatively small region in Northern Sweden, and it's

a quite an isolated region where they've been frequently over the, over the years,

over the generations subject to to famine.

But they also have extremely good records of harvest and food process, so they know

when there were periods of famine, and they know when there were periods of

feast. And so, this is the sort of thing, if we

think back to the Dutch famine, where we would like to know in another separate

study, that something similar could happen.

In this case the diet, the information's not quite the same as in the Dutch

famine. So, I'll explain how it differs.

In this case is a correlation between increased grandparental food

supply and the grand children's

longevity. So, if you have too much food,

if the grandparent had too much food or a surplus of food, this has a negative

impact on the grandchildren. So, this is not exactly the same at all

as for what we're seeing for the Dutch famine.

In this case I also found sensitive periods, but the sensitive periods

weren't necessarily the periconceptional period, but rather where the periods that

are termed the slow growth periods. So, in grandfathers, this happens from

9 to 12 years of age, and in grandmothers, it happens in two periods.

First of all, from eight to ten years of age, and then also in their fetal or

infant life for the grandmothers. So, these aren't quite the same sorts of

periods as we thought about for the Dutch famine.

However, these are a period when you still are having germ cell maturation

occurring, at least for these, eight and 9 to 12 years and eight to ten years

in grandfathers or grandmothers. So, they found two grandparental effects

that we'll think about in turn, using these pedigrees that we've got here.

These grandparental effects were sex specific.

So, in other words it depended on the gender of the grandchildren.

So, if you can imagine yourself being in this generation down in the bottom here

the third generation. If you are the grandsons, here they're

shown as the squares, then your paternal grandfather, in other words your father's

father. So, if we track back his food supply

during these slow growth periods seems to alter your longevity.

So, if your paternal grandfather, if you're a son and your paternal

grandfather had too much food then you're more likely to get diabetes or have

chronic vascular disease and have a shortened lifespan.

By contrast, if you're now a granddaughter down here, if you thought, again think

about your paternal grandmother this time.

So, your dad's mother. Then, if your dad's mother had, an

increased food supply, then we could again link these sort of sex specific

effects grandparentally. So, these effects are obviously quite

complicated. First of all, they're happening through

the grandparents and second of all they're sex specific, but they are

further complicated by the fact that there also sex specific parental effects.

So, in other words, just from this generation here.

So, this data, while it's absolutely tantalising that it might be that our

grandparents' food supply, our grandmother in my case or a grandfather

if it was, in the case of a male, might alter our own epigenetic makeup.

Clearly this is a very interesting concept and something we'd like to know

about. But it's quite difficult to understand

because of these sex-specific parental effects.

And also because although in the Överkalix data they have several cohorts,

because as I said this population was exposed to feast or famine many times

throughout the years and they have very good records for all of these times.

They don't necessarily see these effects in every cohort.

This is something that comes out of a lot of a human epidemiological studies that

we need to consider. So, normally what happens with human

epidemiological studies is that you would like to have whatever effect that you see

in your cohort replicated in a different cohort somewhere else.

This really gives people some confidence that what you're seeing might be a more

generalisable effect. At the moment for this Överkalix studies,

while they are extremely interesting, they haven't been replicated

anywhere else yet. Although, there are ongoing studies in

Britain and other countries to see if they can find these similar

effects. And in contrast to the Dutch famine, at

the moment there's no molecular data that goes with these Överkalix studies.

So, while there's a tantalising example, this is a tantalising

example, we really don't have any evidence to say that this is certainly an

epigenetic event, rather than perhaps say, being a genetic difference.

So, I guess what we want to think about then in terms of how these general sorts

of findings is, do grandparental effects when you have an effect through two

generations. So, you've got the grandparents,

the parents and the children. Do these actually mean transgenerational

epigenetic inheritance through the gametes. Is one synonymous with the other?

Well, what we know with these human studies in particular is that we don't

have any molecular data to say that the effects being passed are only through

epigenotype, rather than perhaps being due to genetic changes.

So, in lower organisms, in mice or in rats where can control the environment

we can control the experiments we are performing.

We can control for genetics, because we have inbred strains of mice that are all

genetically identical. We can also control for maternal

behaviours because we can take pups from one set of parents and foster them over

to a different mother. And we can also control for the timing of

effect. So we can control when the particular

experience of famine or feast was experienced for in an animal model.

But in these cases in this human data we don't have very much control for any of

these things. We also need to consider, whenever

we think about grandparental effect in general, the exposure of the germ cells

to that particular altered environment. In this case the altered diet.

So, this is just a simple picture if we consider a pregnant mother and she's

pregnant with a baby and when that baby is in utero at mid gestation it's already

developing the primordial germ cells for the next generation.

So, if this pregnant mother experiences a change in diet, which could influence

epigenetic control, then she of course, is being exposed as the mom.

Her baby, which is the second generation is being exposed.

But importantly, we need to remember that there are also the primordial germ cells

in that baby that are already developing and that will create that next

generation. And so grandmaternal effects could be

explained through an alter diet in the grandmother, exposing all of these

generations at once. So, actually for the Överkalix data, you

remember I said there, the sensitive period was the slow growth period of

about eight to ten years in the grandmother's or around the time they

were born. This is, it wasn't a period when the

grandmother was already pregnant with the next generation.

So, probably this is not case for the Överkalix data but is something we need

to think about in general for environmental grandmaternal

effects. Other than grandmaternal effects in

Överkalix data, there were also grandpaternal effects.

So, we had an effect where the grandfather's food supply then through

his son only not through daughters had an effect on the grandsons.

So, this passage down the male line only is also something that we need to

consider what the appropriate controls would be or how this effect might be

mediated. Now if you think about it, we know that

the Y chromosome has to come from the grandfather, to the father and to these

grandsons. This is the only place where these Y

chromosomes can come from. And it's possible that there some

epigenetic change to the Y chromosome that's being transgenerationally

inherited. But it's also possible that there's been

some genetic change on the Y chromosome that could be being passed through in

the very normal inheritance type of fashion in this case.

When we think about the exposure of this grandfather, if he's being exposed and

this was in his slow growth period from 9 to 12 years of age.

At this stage when he's being exposed he's clearly not pregnant, he's a

male, but he is developing germ cells. So, that will effect this next generation

but does not explain how the subsequent generation could be effected in this

case. So, grandpaternal effects have different

considerations to grandmaternal effects. There's been one other study, actually by

the same group as the people who have the Överkalix data that I'll

mention. There have been some, some others that

are coming out now that looked at paternal effects of the environment.

In this case, it's to do with prepubertal smoking.

So, this is where the onset of paternal smoking begins in that same slow growth

period. So 9 to 12 years in boys, as opposed

to if they started smoking later, in adolescence or later in adulthood.

And what they found was that if these males started smoking in this

prepubertal period, then they had an increased body mass index in their sons.

So, as compared with if they smoked, started smoking later.

So, in other words, their sons were more likely to be overweight or obese or

morbidly obese if their father started smoking in this period.

So again, there's no molecular data yet to go with this finding.

Again it's a very interesting finding and something that we'd really like to know

about. But because it's only from a father to a

son again it's possible that yes, that it's the Y chromosome, which is

epigenetically changed or it's the Y chromosome which is genetically changed.

Because we know smoking is actually also mutagenic.

So, we know, its very well known that smoking has a carcinogenic effect

because of this mutagenic effect.

It's really likely in this case, or it's certainly possible that there may be some

sort of genetic mediator here rather than an epigenetic mediator.

And until we have molecular data it's hard to work out which of these two

might be the case. So, if you want to summarise what I've

spoken about for this lecture and the last lecture, can we really think

about transgenerational epigenetic inheritance through the gametes in

humans? Does this really happen?

Well, I think the point what I want to make is that, it may well happen and if

it does it's extremely important first to know and very interesting to study.

But at the moment it's extremely difficult to study in humans.

So, it's very difficult to really have the right controls and feel that we know

we are studying transgenerational epigenetic inheritance through gametes

rather than some genetic alterations that we haven't yet found.

So, as I said they’re can be genetic differences that could account

for different epigenetic state and we'll talk about an example of this later in

the week. We know that in humans it's very

difficult or next to impossible to have formal proof of epigenetic inheritance

via the gametes. Because the sorts of controls we need to

do here are simply not possible. So, we can't transfer gametes from one

parent into another to control for you know, that particular parent's

environment, if you like, and even if we could do those sorts of studies,

so, usually we can't access gametes, so while we probably could access

sperm, we're certainly not going to be able to access very many eggs.

Even if we could do those we often can't study the cell type that's relevant.

So, if for example we have a phenotype that's associated with the brain, we

can't go and take a little piece of brain just to be able to work out whether this

is transgenerational epigenetic inheritance.

And so, while this concept is very interesting at the moment, most of the

studies that have looked at the potential of transgenerational epigenetic

inheritance through the gametes in humans have been epidemiological in nature

because this is what's feasible to do in humans.

So, this is why, over the next several lectures, what we're going to do is,

we're going to turn to thinking about mouse and rat model systems where we can

actually answer these questions because if we can understand mechanisms that are

occurring in these organisms then we might be better placed to understand how

it's happening in humans. But of course if transgenerational epigenetic inheritance

through the gametes is true, if in other words we inherit more than just

our genetics from our parents. But also some epigenetic marks that they

potentially acquire during their lifetime,

then this has a huge influence on how we interpret our inheritance of phenotypic

traits. And so, we're really interested to see how this happens.

5.5 Human epidemiological studies on the Overkalix cohort, grandparental effects and possibility of transgenerational epigenetic inheritance in humans

Epigenetic Control of Gene Expression

与讲师见面