Hidden Inheritance

Gail Vines

This article first appeared in New Scientist (28 November 1998, No. 2162, pp. 27-30) and is republished here with permission. It is followed by a response from Jim Cummins which was published in New Scientist 'Letters' section (19/26 December 1998, p. 103).

Imagine aliens from Outer Space announcing that they had engineered lasting alterations into the human race. The changes are going to make our children, and our children's children, smaller, weaker and easier to control. But grassroots resistance is fierce, and soon technicians are working round the clock to screen millions upon millions of human genomes in an effort to weed out anyone whose genes show signs of alien tinkering.

A mystery swiftly unfolds. The human genes appear to be untouched, yet downsized babies are born in ever- increasing numbers. Then, just when it looks like our number's up, the aliens take pity and decide to reveal their biotechnical knowhow. "There's more to heredity than DNA," an alien boffin begins ...

Back in the real world, molecular biologists now sequencing DNA as part of the multimillion-dollar human genome project will finish the job in a few years. Yet masters of the genome we won't be. A spate of mysterious observations made by Earthling scientists suggest that those alien boffins are right -- that there is a lot more to heredity than DNA.

Just as cells inherit genes, they also inherit a set of instructions that tell the genes when to become active, in which tissue and to what extent. This much is uncontroversial. Without this "epigenetic" instruction manual, multicellular organisms would be impossible. Every cell, whether it's a liver cell or a skin cell, inherits exactly the same set of genes, and it is the manual, which has different instructions for different cell types, that allows the cell to develop its distinctive identity.

Established theory has it that the instruction manual is wiped clean during the formation of sperm and egg cells, ensuring that all genes are equally available, until the embryo starts to develop specific tissues. But outlandish evidence now suggests that changes in the epigenetic instruction manual can sometimes be passed from parent to offspring. These findings have even inspired some biologists to suggest that changes in the manual passed down through the generations could provide a way for populations of animals to quickly adapt to their environment, creating a fast-track supplement to the more sedate Darwinian selection.

Speculation aside, one thing is certain. "Bizarre things are going on that we are just beginning to get a handle on," says Marcus Pembrey, a clinical geneticist at the Institute of Child Health in London. Consider the pregnant Dutch women who starved during the famine of the Second World War. Not unexpectedly, they had small babies. Far more surprisingly, those babies went on to have small babies, even though the postwar generation was well fed and no genes had been tinkered with.

Then there are the perplexing findings in mice and rats. Give just one generation of male rats a drug called alloxan, which decreases the body's sensitivity to the hormone insulin, and their offspring and their offspring's offspring become progressively more prone to diabetes. Expose mice to high doses of morphine and the damage to the nervous system persists in their descendants. And one injection of the thyroid hormone thyroxine into a newborn rodent permanently depresses levels of both that hormone and thyroid stimulating hormone -- and levels remain low in the next generation, too.

Many of these observations are decades old and have long been relegated to the scrap heap of unexplained and inconvenient findings. They trouble geneticists, because they seem to fly in the face of classical genetics, even smacking of Lamarckian inheritance, the discredited notion that animals actively acquire characteristics and pass them on to their offspring-by Lamarck's reckoning, body builders would beget muscle-bound babies.

In fact, the way mammals are built should stop a parent's environment having any direct impact on its offspring's genes. The sperm and eggs are packed away in ovaries and testes from very early in development. While other cells become specialised, turning genes on and off to create the different tissues of the body, these "germ" cells remain quietly sequestered, shielded from the environment, until called upon to pass their still pristine genes on to the next generation.

So it's not surprising that scientists have tried to explain away the disturbing aftermath of the Dutch famine and the results of the mice and rat experiments with more conventional reasoning. Did the first generation of small babies suffer some strange hormonal imbalance which, when they reached adulthood, affected the growth of their infants in the womb? Or in the case of the rodent experiments -- which passed down the male line, too -- were the experiments just plain suspect? No firm conclusions were ever reached, but doubts lingered.

"It has become difficult for people to think of heredity as involving non-genetic material," says Steven Rose, a biologist at Britain's Open University in Milton Keynes. The research has continued, he says, but epigenetic research "remains semi-underground. You're not supposed to talk about it". That, however, could be about to change. Last year, Wolf Reik, a molecular biologist at the Babraham Institute outside Cambridge, and his colleagues at the Free University in Berlin, stumbled upon the best evidence yet that epigenetic changes can pass from one generation of mammals to the next.

Reik's main interest is in an epigenetic phenomenon called "imprinting". Genes exist in pairs, one from the mother, one from the father. And whereas most genes in animals such as mice and humans behave in exactly the same way regardless of which parent they come from, imprinted genes are different. In some cases, an imprinted gene is activated only if it is inherited from the father; in other cases, only if it comes from the mother. No one knows quite how this process works, but clearly some sort of "mark" must persist through the generations to tell the offspring's cells which genes to re-imprint.

While much about imprinted genes remains a mystery, initial studies suggest that they often help to regulate the growth of the fetus, and that they are marked for shutdown by small, molecular clusters called methyl groups ("Where did you get your brains?" New Scientist, 3 May 1997, p 34). The methyl groups both block transcription-the first step in gene activation-and, by binding certain proteins, help to fold the DNA into tight, inaccessible coils. Other control mechanisms, still poorly understood, are also at work. But however they work, the existence of imprinted genes demonstrates that, each generation, not all genes are wiped totally clean of their epigenetic marks.

Last year, Reik and his colleagues found clues to the identity of genes that potentially, at least, carry epigenetic information with them as they move from parents to offspring. First, the researchers discovered that some genes become methylated if you move the nucleus from a just-fertilised mouse embryo into the egg of a mouse of a different strain that had had its nucleus removed, and then put the newly manufactured embryo into the womb of another mouse and let it develop normally. The resulting mouse pups were also noticeably smaller.

By measuring the amount of protein in the livers, brains and hearts of these mice, Reik was able to show that two genes had been shut down: a gene for a liver protein called major urinary protein (MUP) and a gene for a protein made in the cells lining the nose, called olfactory marker protein (OMP). Although the DNA sequence of each gene remained unchanged, they had been methylated.

But the real bombshell was yet to come. MUP proteins are usually secreted in mouse urine and, along with pheromones, are signalling chemicals vital to normal sexual behaviour in mice. OMP, on the other hand, is part of the olfactory system that allows mice to recognise pheromones. Not surprisingly, when the smaller mice grew up they were slow to mate. When they did mate eventually, Reik and his colleagues Irmgard Roemer, Wendy Dean and Joachim Klose were amazed to discover that not only were the offspring smaller than usual, but that the MUP and OMP genes were again methylated and switched off. The epigenetic changes had passed down from one generation to the next.

Once you accept that epigenetic inheritance occurs, it's far easier to envisage how drugs, hormones and starvation could have created the bizarre transgenerational effects in rodents and perhaps even in humans, says Reik: the chemicals and the diet may have triggered the heritable methylation of certain genes. At first, however, "we tried very hard to disbelieve our results", says Reik. But as they checked and double-checked their data, and studied the literature, things just fell into place.

It turned out that there had been a smattering of earlier reports of mice inheriting epigenetic changes. Ten years ago, Christine Pourcel at the Pasteur Institute in Paris discovered that when a gene from a virus was inserted into mice it became methylated and silenced, and that the modification was passed on to the offspring. And in 1990, Azim Surani and his team at the Wellcome Trust and Cancer Research Campaign Institute of Cancer and Developmental Biology in Cambridge found other cases of epigenetic inheritance when genes were shifted from viruses into mice. Those earlier transgenic experiments were generally deemed too artificial to be of any consequence in the natural world.

Not so Reik's mice, it seems. "It's lovely work," says Lawrence Hurst, an evolutionary geneticist at the University of Bath. Transferring a nucleus from one mouse egg to another is undoubtedly an unnatural thing to do, but as Reik points out, the procedure could mimic changes that happen naturally. In of development, the activity of genes is in tremendous flux, being turned up and down as methyl groups and proteins are added and removed. Similarly, as the nucleus is moved from one egg to another in Reik's experiment, it experiences differences in temperature and concentrations of various chemicals, all of which could permanently change the methylation of certain genes.

Curiously, cloned lambs and calves created by nuclear transfer -- a technique similar to the one used to create Reik's undersized mice -- may be up to twice as large as normal. No one knows what causes the phenomenon, whether genes are "inappropriately" methylated or whether the oversized offspring, if bred, would pass the trait on. "But our observations raise the question of whether or not such manipulations could actually have a long-term impact by being transmitted to future generations," says Reik.

And if physical manipulations of embryos is all it takes to trigger inappropriate methylation of some genes, then that may be a good reason to worry about what happens to human sperm, eggs and embryos during high-tech fertility treatments. All three are routinely squirted through pipettes, swirled around in lab dishes, or frozen during procedures such as in vitro fertilisation or genetic testing of embryos. What's more, there have been some reports -- albeit controversial -- that babies born following IVF are smaller than normal (see "Shots in the dark for infertility", New Scientist, 27 November 1993, p 13).

Reik's mice also highlight another potentially worrying issue. Hurst, and developmental biologists such as Martin Johnson of the University of Cambridge, argue that in an effort to sell the genome sequencing projects to the public and the funding agencies, molecular biologists have created the misleading impression that genes alone run the show. The constant emphasis on the power of genes, he says, has created "a 20th-century form of fatalistic predestination", in which people believe they are the product of their genes, nothing more, nothing less. Even geneticists, he says, have lost sight of the huge range of environmental factors that can change a gene's activity, ranging from an adult's diet to certain high-tech fertility treatments. For those reasons, some geneticists are calling for a new definition of the gene, based on not only its DNA sequence, but also its epigenetic instruction manual -- the degree of methylation, for example.

But can epigenetic alterations, heritable or otherwise, really be worth the fuss? Yes, according to Eva Jablonka, an evolutionary biologist at Tel-Aviv University. In her book with Marion Lamb, Epigenetic Inheritance and Evolution, The Lamarckian Dimension, she points out that the idea that the effect of the environment on one generation's epigenetic instruction manual can be passed to the next is old hat to students of simpler organisms like bacteria, yeast, plants, and even fruit flies. For example, in yeast, the epigenetic silencing of one of two genes produces changes in sex that are inherited. And just a few months ago, Renato Pare of the Centre for Molecular Biology in Heidelberg, Germany, reported a striking example of epigenetic inheritance in laboratory fruit flies (Cell, vol 93, p 505). The activity -- but not the sequence -- of a key gene was changed in embryos that went through a brief heat shock, activating another gene that caused the flies to have red eyes, a trait they passed on to their offspring.

Jablonka theorises that epigenetic inheritance in lower organisms at the very least play a key role in evolution by providing an additional source of variation on which selective pressures can act. Although epigenetic changes may be as random as mutations in the DNA sequence, they could also be adaptive, triggered by environmental changes to enable simple organisms to respond quickly to a fluctuating environment. For example, if one source of bacterial food is in short supply, heritable epigenetic modifications could help populations of bacteria to switch to another food source. Jablonka also points out that epigenetic inheritance is not at odds with classic inheritance via the genes. Instead, it would be a complementary inheritance system, with Darwin's natural selection acting on both the modified gene and on the genes that control epigenetic modifications.

Meanwhile, Pembrey, provocatively calling himself a "neo-Lamarckian", is prepared to stick his neck out even further, and suggest an adaptive role for epigenetic inheritance in higher organisms such as humans. He speculates that the inheritance of epigenetic factors which control a few select genes may have enabled human populations to regulate the growth of individuals according to food availability. Food shortages could generate physiological responses in adults, say, a change in hormone levels, that influence the activity of key growth genes. This could then be passed on to their offspring by varying the genes' methylation.

In the short term, such an adaptive mechanism could, for example, ensure that the baby's head is not too big for the mother's birth canal. In the longer term, if the offspring also passed those epigenetic changes on to their offspring, it would result in generations of progressively smaller people, until a period of plenty created the epigenetic changes that reversed the trend. The two generations of small babies that followed the Dutch famine could be explained by just such epigenetic adaptation, says Pembrey Perhaps, he says, the giants of Patagonia (literally "the place of big feet") reported by Ferdinand Magellan in the 16th century and countless later European travellers, really did exist.

"What we can see now is the tip of the iceberg," says Marilyn Monk, a molecular embryologist and geneticist and a colleague of Pembrey's at the Institute of Child Health in London. She predicts that many more examples of epigenetic inheritance in mammals will come to light once geneticists develop ways to monitor methylation across the entire genome during an embryo's development. What's more, she says, the much- cherished notion that sperm and egg genes are totally sheltered in the ovaries and testes starts to look shaky when you examine it more closely: in humans, the primordial cells that generate eggs and sperm are busy dividing up until the 15th week of development.

Not everyone is prepared to take such radical positions as those of Lamb and Pembrey. John Maynard Smith, an evolutionary biologist at the University of Sussex, remains sceptical. He points out that even if epigenetic modifications occur naturally in mammals and are passed down the generations, there is still no reason to suspect that they are any more "adaptive" than random gene mutations that are passed on to offspring. Reik, too, cautions against overinterpreting his results. "Whether any such epimutations have any adaptive significance remains to be established," he says. No one has yet shown that inherited epigenetic changes occur naturally in mammals, and even if they did they may still be rare, random and inconsequential events -- even downright dangerous.

Whatever the final verdict on the significance of epigenetic changes, one thing is already clear, says Hurst: "Epigenetics matters." As the human genome project rushes to completion, the really interesting insights are going to come not from the sequences, he predicts, but "from working out how genes are controlled".

Further reading

Epigenetic Inheritance and Evolution, The Lamarckian Dimension by Eva Jablonka and Marion Lamb (Oxford University Press, 1995)
Genomic Imprinting, edited by Wolf Reik and Azim Surani (Oxford University Press, 1997)
"Epigenetic programming of differential gene expression in development and evolution" by Marilyn Monk, Developmental Genetics, vol 17, p 188 (1995)
"Epigenetic inheritance in the mouse" by Irmgard Roemer and others, Current Biology, vol 7, p 277 (1997)
"Imprinting and transgenerational modulation of gene expression: human growth as a model" by Marcus Pembrey, Acta Genet Med Gemmellol, vol 45, p111 (1996)
"Transgenerational effects of drugs and hormone treatment in mammals: a review of observations and ideas" by J. Campbell and P Perkins, Progress in Brain Research, vol 73, p 535 (1988)

From the Letters page, New Scientist, 19/26 December 1998, p. 103

Not by genes alone

Gail Vines's nice article on epigenetic effects ("Hidden inheritance", 28 November, p 26) rightly points to imprinted genes as the main players -- although it would be interesting to look at marsupials, where imprinting supposedly doesn't occur.

However, there may be other non-genetic factors. For example, a sperm contributes a structure known as the centriole to the fertilised egg, and this is now known to be the template for the first cleavage spindle in most mammals (mice are an exception). In addition, calcium oscillations responsible for activating the oocyte are triggered by an extranuclear factor or factors from a sperm's perinuclear region.

Sperm also carry mitochondria that are normally destroyed but they may occasionally evade this process -- possibly by fusing with an egg's mitochondria. There are a number of other components such as a unique alpha-tubulin protein and elements of the tail that could contribute to axis formation and factors such as nucleo-cytoplasmic ratios in the early embryo.

As human-assisted reproductive technologies are becoming increasingly intrusive -- for example, attempts to rescue bad eggs by cytoplasmic transfer -- it is imperative that we come to grips with the biology of what's going on. Small babies may foreshadow much nastier anomalies, and we now have boys being born with the Y chromosome deletions that caused infertility in their fathers.

We urgently need research in appropriate animal models, particularly primates. Mice are not especially relevant to human embryogenesis.

Jim Cummins
Anatomy, Division of Veterinary and Biomedical Sciences,
Murdoch University,
Western Australia 6150
Email: cummins@central.murdoch.edu.au

The material on this page is the copyright of Reed Business Information, publishers of

For more science news and views, check out New Scientist Planet Science

To Ifgene home page