Download Lifestyle in the sperm

Published online: November 8, 2014
Science & Society
Lifestyle in the sperm
There is growing evidence that epigenetic marks can be inherited. But what is the nature of the
information they store and over how many generations do they prevail?
Katrin Weigmann
M
ore than 200 years ago, JeanBaptiste Lamarck (1744–1829)
proposed a theory for the major
question in biology: How do organisms
change over time to adapt to their environment and create new species? Lamarck
reasoned that if an organism changes to
better adapt to its environment, these
changes would be passed on to its offspring,
and this would explain changes over generations. He was eventually refuted, however,
by Charles Darwin’s theory of evolution by
natural selection, and Gregor Mendel’s
genetics, which explained how random mutations, not adaptive changes, generate diversity. The discovery of the structure of DNA,
which mechanistically explained inheritance
and mutations, finally moved Lamarck’s
theory to the huge pile of disproven scientific
hypotheses.
Recent research has re-opened a debate
about Lamarck, however, as so-called
epigenetic inheritance—based on reports
that stress, nutrition, toxins, and infections
not only affect an individual’s behavior,
body weight, and health and defense
system, but also that of its offspring—goes
some way to rehabilitating his theory. “A
Revolution in Evolution: A Return to
Lamarck?” wrote Gregg Henriques, Professor of Psychology at James Madison
University, in December last year on the
blog of Psychology Today (http://www.psychologytoday.com/), while Science magazine asked whether the discovery that
epigenetic changes can be inherited is
“evolutionary heresy” [1]. In reality, epigenetics does not really challenge our basic
model of evolutionary inheritance, founded
on the work of Darwin and Mendel, but it
does add some important features to the
model, and research into epigenetics and
whether such epigenetic marks are heritable is expanding rapidly.
G
enerally, epigenetics is the study of
heritable changes in gene regulation
that are not a result of changes to
the underlying DNA sequence. Mechanisms
that produce such changes include chemical
modifications to the DNA, such as DNA
methylation, or to proteins associated with
the DNA (such as histone modification), but
also other factors, most notably non-coding
RNAs. Epigenetic processes play a central
role in cell development and explain how
the same genome can give rise to many
different cell types; how these cells maintain their identity throughout life; and how
this information is passed on to daughter
cells.
Two recent developments have contributed to the growing interest in epigenetic
inheritance. First, it has become clear that
environmental factors—such as toxins,
nutrition, or stressful experiences—can
leave their marks on the epigenome, permanently altering gene expression and impacting on health, disease, and behavior many
years later. Second, and even more fascinating, epigenetic changes can be inherited
across generations, despite the fact that
nearly all epigenetic marks are thought to be
erased during gametogenesis and in early
embryonic development in a process called
“epigenetic reprogramming”. But to prove
Lamarck right, epigenetic inheritance would
have to meet a number of additional criteria.
Do epigenetic changes induced by the environment reflect beneficial adaptations to the
environment? And if so, are these changes
inherited over many generations in an
epigenetic fashion or are they eventually
fixed into the genome itself to play a role in
evolution?
The best evidence for the inheritance of
environmentally induced traits in humans
comes from epidemiological studies of the
so-called Dutch hunger winter and the
Överkalix cohort. During the winter of
1944–1945, the German-occupied part of
the Netherlands suffered a devastating
famine. Women who were pregnant during
that winter gave birth to children who
developed a number of clinical disorders,
and their grandchildren also suffered from
poor health. Överkalix is a remote community in northern Sweden for which detailed
accounts of harvests, births, and deaths are
available. Tracing back family trees,
researchers
have
found
correlations
between food availability during the prepuberty years of grandparents and the
mortality risk of subsequent grandchildren.
These studies suggest that starvation during
critical periods of development—fetal
development and early adolescence—
induces epigenetic changes in the germline
that persist to the children and possibly
grandchildren.
O
f course, these studies are merely
observational and retrospective
and carry the risk of possible bias
and confounding factors. However, several
recent rodent studies further support the
idea that acquired traits can be transmitted
across generations. Most of these studies
have focused on the inheritance from
fathers rather than mothers to rule out
maternal effects that are not germline
specific, such as in utero embryonic development or maternal care after birth. For
Freelance journalist in Oldenburg, Germany. E-mail: [email protected]
DOI 10.15252/embr.201439759 | Published online 8 November 2014
ª 2014 The Author
EMBO reports Vol 15 | No 12 | 2014
1233
Published online: November 8, 2014
EMBO reports
example, a high-fat diet in male rats
induced adiposity, impaired glucose tolerance, and insulin sensitivity in their
female offspring, along with an altered
expression of pancreatic islet genes. The
sperm of pre-diabetic male mice had
altered methylation patterns, and their
offspring showed increased glucose intolerance. In addition, the progeny of male
1234
EMBO reports Vol 15 | No 12 | 2014
Lifestyle in the sperm
mice that had been fed a low-protein diet
exhibited altered expression of genes
involved in lipid and cholesterol metabolism [2].
Oliver Rando, a molecular biologist at the
University of Massachusetts Medical School
in Worcester, USA, and lead author of the
low-protein study, is looking for the mechanisms behind this effect. Together with his
Katrin Weigmann
colleagues, he has initiated an epigenomewide mapping of histone modification,
cytosine methylation, and small RNAs.
Although the researchers see at least some
changes in all three types of marks, some
seem to be more promising than others. “As
a guess, I am most optimistic about the
small RNAs. But the cytosine methylation
changes we find are also appealing to us
ª 2014 The Author
Published online: November 8, 2014
Katrin Weigmann
because they occur at regions that escape
demethylation at fertilization”, Rando said.
In addition to nutrition, scientists are
also investigating other examples of epigenetic inheritance. Isabelle Mansuy, a
neuroscientist at the University Zürich and
the Swiss Federal Institute of Technology
in Zürich, Switzerland, has demonstrated
that early postnatal trauma in mice can
affect behavior across generations. In her
experiments, she separates newborn mice
from their mothers at unpredictable times
every day for 2 weeks. “The effect is quite
dramatic”, Mansuy said, “it severely alters
the animals’ behavior”. When the pups
become adult, they suffer from depression,
loss of behavioral control, impaired cognitive functions, and social skills. These
symptoms, along with the altered expression of several genes, including stressrelated genes, are transmitted to the progeny for several generations.
......................................................
“The extent to which useful
adaptations can be inherited
epigenetically may depend on
what Rando calls the
“bandwidth” of epigenetic
information in sperm”
......................................................
In a recent article, Mansuy and her
colleagues identified small non-coding
RNAs as potential mediators of inheritable
epigenetic information [3]. They injected
RNA extract from the sperm of traumatized
males into fertilized eggs of wild-type
females. The resulting offspring expressed
similar behavioral alterations as their
fathers, showing that sperm RNA is likely
a carrier of a transmissible epigenetic
signal. However, additional mechanisms
may also be involved. “Non-coding RNAs
are not altered in sperm of the progeny”,
Mansuy explained. “But since the phenotype does persist in the third generation,
we believe the change in RNA in sperm is
transferred to another likely more stable
form of epigenetic mark”.
E
nvironmental information can leave
epigenetic marks that are relayed to
the next and even further generations.
However, the resulting phenotypes are not
necessarily useful adaptations to the environment, as one would expect in a truly
ª 2014 The Author
EMBO reports
Lifestyle in the sperm
Lamarckian evolution. “Stress can, in some
conditions, improve behavioral flexibility”,
Mansuy commented. But this is rather the
exception than the rule. In general terms,
depression and impaired memory can hardly
be called a useful adaptation. Similarly, the
metabolic changes in Rando’s feeding paradigm are not unambiguously adaptive.
“Whether or not the offspring of starved
mice are better adapted to starvation has not
been analyzed”, he said.
The extent to which useful adaptations
can be inherited epigenetically may depend
on what Rando calls the “bandwidth” of
epigenetic information in sperm. “You can
imagine two extremes”, Rando explained.
“One extreme is that sperm carry an infinite amount of environmental information.
Every methylation site could carry one bit
of information and, in terms of RNA
profile, the high-dimensional complexity of
the transcriptome could be used. The alternative is that sperm transmit some blunt
measure of quality of life: Life was terrible,
life was ok, life was great”. And there is,
according to Rando, a good reason to lean
toward the latter interpretation: Different
paradigms produce an overlapping set of
phenotypes in the progeny. “You might
find metabolic phenotypes in a parental
stress paradigm or behavioral phenotypes
in nutritional paradigms”, Rando said.
Whatever causes these epigenetic effects
might be more related to sperm quality
than to adaptive traits.
......................................................
“Disease risk may therefore
not only depend on our own
genes and environmental exposure, but also on environmental exposure of our parents”
......................................................
However, a recent discovery seems to
contradict this view. Brian Dias and Kerry
Ressler at Emory University in Atlanta,
Georgia, USA, reported that mice can
inherit olfactory memories from their
fathers [4]. They trained mice to fear
specific odors by associating them with
mild foot shocks and showed that the
offspring of these mice—children and
grandchildren—were also more sensitive to
that particular smell. “It will certainly be
important to see other examples in the
same domain. But if this proves to be the
first of many papers, that would demonstrate high-dimensional information transfer”, Rando commented. To date, however,
data are sparse and there is a good portion
of skepticism. In addition, the concept of
inheritance of memory raises questions. “If
things learned by a grandfather would be
inherited in a substantial manner, this
would have been noticed by now”, Mansuy noted.
Inherited memories or not, it is becoming
increasingly accepted among the scientific
community that at least some environmental
information can be passed on to offspring.
This is an important insight for clinical
research as human and rodent studies have
shown that disease risk may not only
depend on an individual’s genetic setup and
environment, but also on the environment
of his or her ancestors. Many complex
diseases have a strong heritable component.
Although genomewide association studies
have successfully identified hundreds of
variants that are associated with disease
risks, these variants only explain a small
portion of the observed heritability, which
has been referred to as “missing heritability”. According to Mansuy, epigenetic inheritance could provide an explanation for this
puzzle. “Today, many people agree that
epigenetic inheritance largely contributes to
diseases like depression, personality disorder, even Alzheimer’s disease”, she said.
I
n searching for potential mechanisms
of multigenerational epigenetic inheritance, an obvious place to look is in
C. elegans. In 1998, Andrew Fire, then working at the Carnegie Institution of Washington
in Baltimore, Maryland, USA, and Craig
Mello, who was at the University of Massachusetts Cancer Center in Worcester, USA,
published a seminal paper on a process
called RNA interference (RNAi), for which
they were later awarded the Nobel Prize.
Short pieces of double-stranded RNA, they
found, can silence specific target genes. And
these interference effects, as Fire and Mello
mention in their original publication, “can
persist well into the next generation, even
though many endogenous RNA transcripts
are rapidly degraded in the early embryo”
[5].
In 2012—more than 15 years after the
Nobel
Prize-winning
publication—the
research groups of Craig Mello and Eric
Miska, a molecular geneticist at the University of Cambridge, UK, published a set of
EMBO reports Vol 15 | No 12 | 2014
1235
Published online: November 8, 2014
EMBO reports
papers that shed further light on the underlying mechanisms. “We found a nuclear
RNAi pathway that seems to make this
multi-generational inheritance possible”,
Miska said. The researchers unraveled a
silencing machinery that is guided to
nascent RNA transcripts at specific loci by
small RNAs, where it modulates chromatin
structure to shut down genes [6]. In addition, when the scientists induced silencing
using piRNAs—a germline-specific set of
small RNAs—instead of external RNAs, they
generated a very strong silencing effect. “We
call this paramutation”, Miska said, named
after a stable inheritance mechanism of
epigenetic marks in plants. “This seems to
be the first example in animals”, he added.
......................................................
“Although it seems conceiv-
able that the RNAi and piRNA
pathways could transmit
environmental adaptions
across several generations, it
is still not clear if this happens
in nature”
......................................................
Similar to mice and humans, most chromatin marks are erased during germline
development in C. elegans. Accordingly,
Miska thinks that RNA is the most likely
candidate to carry information from one
generation to the next. “Our model is that
small RNA is inherited and then these molecules re-establish a certain chromatin mark”,
he said. This idea is in line with a previous
study by Andrew Fire and colleagues who
showed that the genomic locus is not
required to pass RNAi-induced silencing
from parents to offspring; all you need is
some diffusible substance in the sperm.
If small RNA can transmit information
across generations, what is the nature of
this information? Does it include environmental cues? Feeding worms RNA to silence
genes in laboratory experiments can hardly
be called an environmental cue, as Miska
conceded. “That would be cheating a little
bit, although I’d love to see it as an
environmental cue”, he said. Starvation, a
temperature change, infections, or behavioral experiences would be more typical
environmental changes encountered by a
worm in real life.
Several laboratories have developed
paradigms in C. elegans to investigate the
1236
EMBO reports Vol 15 | No 12 | 2014
Lifestyle in the sperm
transgenerational inheritance of such environmentally induced traits. Some reports
show that an antiviral RNAi response can be
inherited to the next generation. In addition,
the inheritance of an acquired olfactory
behavior has been described and starvation
has been shown to trigger the expression of
small RNAs that are inherited to the next
generation. But in most of these cases, there
are still open questions: Is the nuclear RNAi
machinery discovered by the Miska and
Mello laboratories really involved? Do
inherited RNA expression profiles reflect
useful physiological adaptations? Is the
phenotype stably transmitted over several
generations? None of the studies can answer
all of the questions. Although it seems
conceivable that the RNAi and piRNA pathways could transmit environmental adaptions across several generations, it is still not
clear whether this happens in nature. Just
like in mammals, the question of “bandwidth” arises. How specific is the information that is inherited across generations?
Even though the system could potentially
carry a lot of information, this does not
imply that all of it is used. Again, epigenetic
inheritance could be refined to some gross
evaluation of life’s quality: “environment is
bad, better warn the kids”, as Miska
described it.
W
hile there is massive epigenetic
reprogramming in embryonic
and germline development in
mammals, this is not the case in plants
where, accordingly, epigenetic inheritance is
much more common. To investigate the
impact of heritable epigenetic variation,
Frank Johannes, a specialist in population
epigenetics and epigenomics at the University of Groningen, the Netherlands, in
collaboration with researchers from Paris,
created so-called “epigenetic recombinant
inbred lines” of Arabidopsis. These lines are
genetically equal, but show different methylation patterns. “In this experimental system,
certain methylation changes can be stably
inherited for many generations”, Johannes
said. Moreover, plant lines with different
methylation patterns also show different
phenotypes. They vary in flowering time,
plant height, fruit size, and other traits. In a
recent study [7], Johannes and his colleagues determined specifically which of the
differentially methylated regions in the plant
genome cause variation in flowering time
and root length. “This has really been the
Katrin Weigmann
first systematic demonstration that DNA
methylation across the genome can affect
complex traits in a heritable way”, he said.
Whereas phenotypic diversity within populations is commonly attributed to genetic
variation, these results show that epigenetic
changes also contribute.
......................................................
“. . . even if epimutations are
not environmentally induced,
epigenetics may play an
important role in evolution”
......................................................
Plants have a wide scope of options for
epigenetic regulation, including a unique
epigenetic pathway, RNA-directed DNA
methylation, where small RNAs guide
changes to the chromatin. But are such
mechanisms used to pass environmentally
induced traits to the next generation? “This
is a bit of a controversial area of research”,
Johannes said. “Some studies have investigated the effect of environmental treatments
on the methylome and whether changes that
are induced would persist over multiple
generations. It turns out that evidence for
this is quite limited”, he said. “As of now it
seems the environment is not ‘strong’
enough to induce lasting methylation
changes. But there is certainly more room
for investigation”.
N
onetheless, even if epimutations are
not environmentally induced, epigenetics may play an important role in
evolution. To understand how a population
adapts to environmental changes, scientists
have long used genetic models of evolution.
One parameter that often enters into such
models is the mutation rate. But, according
to Johannes, the epimutation rate should
also be considered. “We already know that
the epimutation rate, at least at individual
cytosines, is actually much higher than the
genetic mutation rate. There is a much
higher probability to come up with a potentially favorable epiallele, which may facilitate adaptation. This makes the system
much less constrained”, he said. As plants
are sessile organisms that cannot move to a
more favorable environment, one can easily
imagine how a faster adaptation rate can be
particularly profitable for them.
But epigenetic changes are also more
likely to undergo reversion. Epigenetics may
play an important role in the initial steps
ª 2014 The Author
Published online: November 8, 2014
Katrin Weigmann
toward adaptation, but advantageous phenotypic traits ultimately need to be fixed by
changes to the genome. “Theoretical modeling shows that a population can be rapidly
brought to an adaptive peak through a series
of epimutations, and that subsequent genetic
mutations can produce favorable alleles
that stabilize the population around the
peak. Of course, these processes are largely
hypothetical and what the epigenetics
community needs is further experimental
data to support such theoretical insights”,
Johannes explained.
Epigenetic inheritance is an intriguing
mechanism with important implications, as
epigenetic changes inherited from our
ancestors may impact on metabolic and
psychiatric disorders. In addition, the inheritance of epigenetic marks in plants can
affect complex traits, and epimutation rates
could be an important concept for evolutionary models. As yet, however, there is
little evidence that epigenetic inheritance
could provoke directed evolution, as
proposed by Lamarck. It is questionable,
using Rando’s words, whether the
ª 2014 The Author
EMBO reports
Lifestyle in the sperm
“bandwidth” of epigenetic information is
broad enough to afford the stable inheritance of detailed environmental information
through the epigenetic code over many
generations. There are a handful of reports
about inherited environmental adaptions
that do point to significant bandwidth, but
the experiments reported have not yet been
repeated independently and the data have
not yet convinced the majority of the
scientific community. “It’s like seeing a
unicorn in the garden”, Miska said. “If you
see one you better have a lot of evidence
that it is true”. To date, this evidence is
still lacking.
References
1.
2.
3.
Gapp K, Jawaid A, Sarkies P, Bohacek J,
Pelczar P, Prados J, Farinelli L, Miska E,
Mansuy IM (2014) Implication of sperm RNAs
in transgenerational inheritance of the effects
of early trauma in mice. Nat Neurosci 17:
667 – 669
4.
Dias BG, Ressler KJ (2014) Parental olfactory
experience influences behavior and neural
structure in subsequent generations. Nat
Neurosci 17: 89 – 96
5.
Fire A, Xu S, Montgomery MK, Kostas SA,
Driver SE, Mello CC (1998) Potent and specific
genetic interference by double-stranded RNA
in Caenorhabditis elegans. Nature 391:
806 – 811
6.
Ashe A, Sapetschnig A, Weick EM, Mitchell J,
Bagijn MP, Cording AC, Doebley AL, Goldstein
Pennisi E (2013) Plant biology. Evolution
LD, Lehrbach NJ, Le Pen J et al (2012) piRNAs
heresy? Epigenetics underlies heritable plant
can trigger a multigenerational epigenetic
traits. Science 341: 1055
memory in the germline of C. elegans. Cell 150:
Carone BR, Fauquier L, Habib N, Shea JM, Hart
88 – 99
CE, Li R, Bock C, Li C, Gu H, Zamore PD et al
(2010) Paternally induced transgenerational
7.
Cortijo S, Wardenaar R, Colomé-Tatché M,
Gilly A, Etcheverry M, Labadie K, Caillieux E,
environmental reprogramming of metabolic
Hospital F, Aury JM, Wincker P et al (2014)
gene expression in mammals. Cell 143:
Mapping the epigenetic basis of complex
1084 – 1096
traits. Science 343: 1145 – 1148
EMBO reports Vol 15 | No 12 | 2014
1237