91av

IT’S DARK. As you drift off to sleep, a noise outside yanks you back into the
wakeful world. Instantly alert, your heart pounds against your ribs. A moment
later you hear a snarling bark and a crash as the intruders beat a hasty
retreat. Panic over—good old Fido has seen them off.

But imagine that instead of settling back to sleep, your loyal watchdog comes
bounding into your bedroom, teeth bared, eyes flashing, and starts attacking
you. How do you feel as your flesh is ripped apart? Now you can begin to
understand the shock felt by someone who has been told that their own immune
system—their body’s molecular watchdog—has turned traitor.

Autoimmune diseases, including multiple sclerosis, lupus and rheumatoid
arthritis, claim millions of new victims each year. Their methods vary, but all
involve an overzealous immune system attacking the very body it was designed to
defend. Cell by cell, they slowly disable and ultimately kill many of their
victims. Despite decades of research, few theories have emerged to explain why
natural selection would tolerate such a critical design flaw. Now a few
biologists are starting to point the finger at alien invaders that have been
trapped inside our cells for millions of years. And if they are right, practices
such as gene therapy and xenotransplantation may be riskier than anyone
thought.

Researchers agree that the key to the puzzle of autoimmunity must lie in the
major histocompatibility complex (MHC) —our immune system genes. The MHC
is an unusually diverse region of the genome, which probably reflects our
intense and ongoing co-evolutionary arms race with countless disease organisms.
Each MHC gene in a population can have dozens of versions—or
alleles—although only a couple of these will be present in any given
individual. Over 200 MHC genes are packed tightly together in our genome and
their job is to produce proteins that detect and destroy invaders.

But some MHC alleles or haplotypes—combinations of alleles—have a
major drawback: people who carry them are more likely to get an autoimmune
disease. For example, nearly every victim of Hirata’s disease—in which the
body attacks its insulin-producing cells—carries the MHC allele known as
DR4w13. Two-thirds of those who suffer from rheumatoid arthritis
possess another risky allele called DQB1. The haplotypes dubbed
DQ and DR also confer an increased risk of rheumatoid arthritis
and type I diabetes. And researchers have identified dozens more of these
associations. But strangely, not everybody who carries a risky allele or
haplotype becomes ill. So the simplistic notion that these risky alleles lead
inexorably to autoimmune disease has been abandoned in favour of more
sophisticated explanations.

The widely accepted view is that a risky allele must be switched on before it
turns traitor. The trigger is thought to be bits of foreign proteins, from food
or infectious invaders, that resemble the body’s own proteins. MHC alleles that
respond to such “molecular mimics” will then be primed to attack the body’s own
proteins as well. Such alleles could persist through evolutionary time if they
help fight off infectious diseases in early life and do not trigger autoimmune
reactions until middle or late life. So people who carry them are more likely to
survive long enough to reproduce and pass along their double-edged immunological
inheritance.

Although this theory doesn’t explain susceptibility to diseases that kick in
during childhood, such as type I diabetes, many find it appealing. Not so Graham
Boyd, emeritus professor of medicine at the University of Tasmania in Hobart. He
doesn’t attribute autoimmune diseases to molecular mimics at all. Instead, Boyd
and a growing number of like-minded theorists point to what at first glance may
seem an unlikely culprit: ancient viruses stuck in the human genome, known as
endogenous retroviruses or ERVs.

Outlandish as it sounds, we are the genetic descendants of viruses as well as
primates. The viral ancestors of ERVs invaded the cells of our forebears during
infections millions of years ago and liked it so much they decided to stay.
Happily integrated into their new home, ERVs have become part of our own genome,
passed down through the generations. In fact, virologists have spotted ERVs in
the genome of every mammal they have checked. Repeated invasions over more than
30 million years have left a surprisingly large viral legacy. “Up to 1 per cent
of the human genome is represented by human ERVs and their fragments,” says
Eugene Sverdlov, a geneticist at the Russian Academy of Sciences, Moscow.

ERVs are relatively simple creatures, genetically speaking. Like wild
retroviruses—which include HIV—they have a few genes coding for
enzymes and structural proteins. These are sandwiched between long terminal
repeat sequences (LTRs), which act like on-off switches regulating the
production of viral genes. They are called retroviruses because their genes are
encoded in RNA rather than DNA and they infiltrate the host genome by creating
DNA copies of themselves. Infected cells may then be tricked into duplicating
the viral genes as though they were merely instructions for one of the body’s
own cellular proteins (see Diagram).

Sverdlov calls ERVs “the perpetually mobile footprints of ancient
infections”. Many of the resident aliens’ genes have been broken up by
mutations, but at least a few are still intact and able to make viral proteins.
ERVs also have a nasty habit of hopping around the genome, duplicating as they
go. And either behaviour—jumping or producing viral proteins—could
explain why certain MHC alleles are linked to autoimmune disease.

One way ERVs might trigger auto-immune disease is by causing regulatory
problems in the MHC genes, suggests Klaus Badenhoop from the University of
Frankfurt, Germany. Although nobody knows exactly why, ERVs appear to have a
particular affinity for the MHC region. By some estimates, they are 10 times as
common in the MHC as elsewhere in our genome. When an ERV’s regulatory
instructions (its LTR) land near a host regulatory sequence, the viral on-off
switch can be mistaken for the host’s own genetic gadgetry, with disastrous
results. The ERV can enhance or modify the expression of adjacent genes, says
Badenhoop.

Together with Ralf Tvnjes from the Paul Ehrlich Institute in Langen, Germany,
and others, Badenhoop is investigating whether risky MHC alleles might wreak
their havoc because they mistake the ERV instructions for their own, and so are
accidentally switched on. If this is so, you would expect to find more LTRs near
risky alleles than normal MHC alleles. And this is exactly what Badenhoop and
Christian Seidl of the University of Frankfurt discovered when they looked at
the region around the haplotypes DQ and DR that confer a high
risk of type I diabetes and rheumatoid arthritis.

Badenhoop sees promise in the idea that ERVs contribute to autoimmune disease
by causing confusion within the immune system, but is reluctant to draw firm
conclusions yet. “There is sufficient evidence to regard ERV long terminal
repeats in the MHC as genetic markers for autoimmune disease,” he says. But
“their function—how and where they contribute to pathogenesis—still
needs to be elucidated”.

Boyd also believes that ERVs could play a role in autoimmune disease. But
instead of seeing people with autoimmune diseases as hapless victims, he prefers
a more co-evolutionary explanation, which he has dubbed “balanced dynamic
polymorphism”. Boyd sees viruses and the hosts they live in as opposing teams in
a dynamic co-evolutionary arms race.

Like exotic species settling in new ecosystems—rats on an island, for
example—ERVs can be disruptive when they first arrive. But like the rats’
descendants, the viral lineages tend eventually to become better adapted to
their surroundings. Once an ERV is integrated into another genome, its survival
is hitched to that of its host. So the longer the association, the more likely
it is that evolution will have quelled an ERV’s more unneighbourly instincts.
Boyd likens it to long-running tribal warfare. “As the years went by,” he says,
“there would be a sort of truce whereby the survivors from both sides would
generally agree that all aggression should be curbed.”

This truce even goes so far as allowing ERVs to play a major role in our
evolution, by doing away with the need to lay eggs
(91av, 12 June 1999, p 26).
And recent studies reveal cases where viral genes have been
co-opted by hosts to serve useful functions—ironically, often helping to
fight disease. One such, called P-5, is involved in producing immune
lymph cells. It has “a possible role in immunity to retrovirus infection”, says
Jerzy Kulski from the University of Western Australia in Nedlands, who made the
discovery with his colleague Roger Dawkins.

Such domestication events may explain why Badenhoop’s team found that while
many risky genes are associated with viral LTRs, at least one such haplotype
lacks them. In this case, the viral on-off switches are found near normal
versions of a gene, suggesting that they may provide some protection against
autoimmune disease. ERVs can also help defend hosts against wild viruses in
other ways, according to Roswitha Löwer, a geneticist from the Paul Ehrlich
Institute in Langen, Germany. Chickens and mice are protected from infection by
endogenous proteins that stop viruses sticking to host cells, Löwer says.
And in mice, ERVs interfere with the replication of wild viruses inside host
cells.

Normally, then, ERVs pose little threat to their hosts. But Boyd believes the
truce is an uneasy one. “There would always be renegade rogues on both sides,”
he says. Although most ERVs do not normally produce the viral proteins that
provoke attack from the host immune system, many retain the genetic code for
such particles. This unwillingness to disarm, Boyd believes, may set the stage
for autoimmune disease. ERVs that retain their protein-producing potential are
more like tenuously tamed wolves than loyal puppy dogs. Every once in a while
these ERVs awake from their civilised slumber and begin pumping out viral
proteins.

So what makes domesticated ERVs turn feral? Löwer speculates that UV
light and bacteria are possible alarm clocks that wake ERVs. Boyd believes that
the key is repeated damage to cells, either from infection by wild viruses or
from severe psychological stress. The effects of such stress can affect the
sympathetic nerves controlling arterial blood supply. If the nerves are shut
down, the temporary loss of blood supply can cause cellular damage, which might
contribute to ERV activation, according to Boyd.

Either way, once ERVs start producing molecules that look like antigens from
wild viruses, MHC genes may kick in to fight off the perceived invasion. The
result is an immune attack against your own cells. The question then is why
natural selection hasn’t eliminated those MHC alleles prone to mistake ERV
products for infections. Boyd argues that hosts are in an evolutionary bind, a
Darwinian catch-22. So long as wild relatives of ERVs exist in nature and pose a
threat, there will be a survival advantage in possessing MHC alleles that can
fight them off—even though individuals carrying such alleles are
susceptible to autoimmune disease.

His ideas fit in with today’s knowledge of autoimmune susceptibility alleles
and ERVs. He believes that the conflicting selection pressures for maintaining
and eliminating risky MHC alleles—those that mistake ERV proteins for
foreign invaders—result in some, but not all, individuals possessing
dangerous alleles. If he is right, geneticists will find geographic patterns in
the distribution of risky alleles that reflect varying disease risks.

For the moment, the jury is still out on whether ERVs are an accessory to
autoimmune disease. If they are, the implications for medicine are wide-ranging.
Löwer is particularly concerned about gene therapy and cross-species organ
transplantation. The retroviruses used as vehicles in gene therapy might
activate a patient’s ERVs. And the promise of plentiful supplies of organs from
pigs and baboons for use in xenotransplantation may be dangerous for similar
reasons. The ERVs in pigs and primates, Löwer points out, are close
relatives of ours. “There is a risk of uncontrolled human ERV amplifications,”
she says.

Löwer’s concerns seem well founded. Earlier this year, a team of her
colleagues led by Frank Czauderna reported that some pig ERVs do indeed code for
viable viruses. Worse still, these viruses can replicate in human tissues, at
least under lab conditions. Czauderna worries that pig and human ERVs could
hybridise, yielding infectious viruses with new and unfamiliar properties. But
he is hopeful that it will be possible to use genetic engineering techniques to
knock out pig ERVs and create a cloned lineage of donor animals for
xenotransplantations.

And there is a more positive side to all this: it could lead to ways of
fighting off diseases that now seem to strike at random. If ERVs are activated
by repeated stress, as Boyd suspects, then identifying and avoiding such stress
could one day eliminate the immunological betrayals that lead to disease.
“Autoimmune disease is not inevitable,” says Boyd.