¿ìè¶ÌÊÓƵ

Rheaumatoid arthritis is the end of a cascade of self-destruction, triggered
by confusion in the immune system. Biologists have known for more than 50
years that the immune system can attack the body’s own tissues. But how
this happens remains a puzzle. More recently, researchers have begun to
suspect that the mechanism is linked with the occasional failure of the
immune system to distinguish between proteins on bacteria and proteins the
body produces under stress. If they are right, they could be on the trail
of a cure for a crippling disease.

This type of arthritis afflicts as many as two in every hundred people.
People who have it suffer from stiffness and painful swelling in the joints,
which becomes steadily worse. The sequence of events that leads to inflammation
in the joint is already in place by the time these symptoms appear, although
it is still poorly understood. Briefly, white blood cells, initially attracted
by regulating chemicals called cytokines, cross the walls of the blood vessels
in the surrounding soft synovial tissue and invade the joint. The cells
of the synovial membrane divide rapidly and membrane tissue grows into the
joint cavity. The white cells release enzymes such as hyaluronidase, collagenase
and elastase and stimulate connective tissue cells to secrete these and
other chemicals such as prostaglandins, which break down cartilage and bone
(see Figure 1).

There is no cure for the disease. In severe cases treatment may involve
immunosuppressive drugs with unwanted side effects. Doctors may ease the
pain and pressure in joints by drawing off the excess synovial fluid that
accumulates during inflammation, but sometimes they have to replace joints
such as the hip and knee if these have suffered irreparable damage. If researchers
could find out exactly what triggers the invasion of joints by white cells,
they could treat the cause of the disease instead of its symptoms.

Nearly 100 years ago, at the 1897 Congres Francais de Chirurgie in Paris,
a French doctor called Poncet described how some of his patients with acute
tuberculosis had also developed inflammatory arthritis. But despite isolated
claims, no one has consistently found bacteria in the joints of people with
rheumatoid arthritis. The notion that bacteria may be involved in autoimmunity
has been the focus of much controversy ever since.

In 1986, a group at the Weizmann Institute in Israel led by immunologists
Joseph Holoshitz and Irun Cohen showed that T cells-a subset of white blood
cells-isolated from rheumatoid synovial fluid divided rapidly when crude
mixtures of proteins from Mycobacterium tuberculosis, the bacterium that
causes tuberculosis, were added to them. These findings lent support to
the increasingly popular theory that the same human cells that react to
mycobacteria cause rheumatoid arthritis.

Two years later, with the help of recombinant DNA technology, a Dutch
group led by Pieter Res identified the bacterial antigens that the T cells
had recognised. One protein in particular, with a molecular size of 65 kilodaltons
stimulated T cells derived from the synovial fluid. Even more strikingly,
the T cells reacted most strongly to this protein during the early stages
of the disease.

By 1989, Holoshitz had moved to Stanford University in California where
he succeeded in isolating and cloning the T cells. Now he had a way of studying
the link between mycobacterial immunity and rheumatoid arthritis. He could
analyse each target antigen in detail because each T cell clone recognises
just a small region, or epitope, of a single protein. Another thing that
interested Holoshitz and his team was the type of T cells they found. These
had a rare set of surface proteins called gamma-delta, whereas most T cells
have alpha-beta. It is through these surface proteins that each T cell recognises
its specific antigen. (In a normal immune response this recognition is the
trigger for a chain of events involving other parts of the immune system,
culminating in the destruction of the foreign organism.

Mystery antigen

A year earlier, immunologist Fionula Brennan and her colleagues at the
Charing Cross Sunley Research Centre in London had found more gamma-delta
T cells in synovial fluids than in the circulating blood of people with
rheumatoid arthritis. Perhaps these T cells had recognised their antigen
in the joint-even though no one had identified the antigen-and had then
multiplied as they would in a typical immune response. But there are no
bacteria in the joint. So what else could be the target for the self-attacking
immune system?

At London’s Hammersmith Hospital in 1988, biochemist Doug Young and
his colleagues discovered a group of five proteins that were the body’s
targets in its fight against tuberculosis and leprosy-both infectious diseases
caused by mycobacteria. Of this group, one protein had a molecular size
of 65 kilodaltons. The protein-p65-also bore a striking resemblance to so-called
stress proteins.

Stress proteins (or heat shock proteins) are essential for survival
and are detectable in all cells, including bacterial and mammalian, after
injury. Normally they are present at low levels but when cells are subjected
to ultraviolet light, heat, chemicals, change in pH, or even alcohol, cells
produce large amounts of them for protection (see ‘Heat shock proteins to
the rescue’, ¿ìè¶ÌÊÓƵ, 1 April 1989). In conversation with Young, Holoshitz
realised that the mycobacterial target for his T cells was identical to
Young’s p65 stress protein.

Other researchers now believe that inflammation and destruction in the
rheumatoid joint triggers the production of heat shock proteins in the cells
of the joint-both connective tissue cells and invading white cells. Two
years ago a Swedish group led by Lars Klareskog at the University of Uppsala
tagged a dye to a monoclonal antibody they knew would bind to human p65.
Looking through a light microscope they saw that the antibody stained inflamed
tissue from a rheumatoid joint. In particular, the dye-and therefore the
p65-was present not only in white cells, including ‘scavenger’ macrophage
cells, but also in cartilage-forming cells called chondrocytes. Tissue from
healthy joints did not stain.

Ravinder Maini and his colleagues at the Kennedy Institute of Rheumatology
in London did similar experiments. Maini is confident that antigens, in
the form of heat shock proteins, are present in rheumatoid tissue. ‘There’s
no doubt that anti-p65 antibodies have stained some inflamed joints,’ he
says. It is possible that heat shock proteins in joints themselves stimulate
T cells to continue the cycle of destruction. But this does not explain
why Holoshitz’s cloned T cells reacted to bacterial p65. And it still leaves
the question of how mycobacteria fit into this autoimmune disease.

Biologists now attempting to answer these questions have adopted the
theory of ‘antigen mimicry’. Among others, molecular virologist Michael
Oldstone at the Scripps Institute in San Diego, California, proposed that
the reason why the human immune system may not be able to distinguish between
a bacterial or viral protein and one of the body’s own proteins is that
they have similar amino acid sequences. This is possible because the epitope
to which the T cell binds may only be a short stretch of 8 to 12 amino acids.

Research by several groups seems to support this idea. Tetsuya Koga
and Stefan Kaufmann, immunologists at the University of Ulm in Germany,
took macrophages from the blood of mice and stressed them in the laboratory
to produce cell proteins, which they found to be 65 kilodaltons in size.
Then they tagged proteins (which were either whole or fragments) on the
cell surface with fluorescent antibodies. To stress the macrophages they
used interferon-gamma-a chemical secreted by activated white cells which
may also be present in the rheumatoid joint. The mouse p65 proteins made
the macrophages ‘visible’ to the T cells, which Koga and Kaufmann had deliberately
selected for their immunity to mycobacterial p65. It seems that the T cells
responded to regions of mouse p65 that closely resembled the mycobacterial
protein.

Some researchers might argue that the macrophages may have been harbouring
bacteria. But Kaufmann says they extracted macrophages from bone marrow,
in a synthetically prepared culture medium that has no possible contamination
from infected blood products. They have yet to confirm that the protein
is mouse p65, and he warns that it may instead be cross-reactive, a reaction
of the T cells to a different protein of a similar shape. Nevertheless,
the T cells responded only marginally to untreated macrophages, just as
you would expect for an immune system that is working normally where T cells
are tolerant to other cells-including macrophages-which come from the same
organism.

The second line of evidence for antigen mimicry in T cells comes from
experiments with human T cells. In 1988 immunologist Tom Ottenhoff, then
at the Armauer-Hansen Research Institute in Addis Ababa, Ethiopia and Jan
Van Embden of the National Institute of Public Health and Hygiene in Bilthoven,
the Netherlands, studied human T cells immunised against mycobacterial proteins.
They discovered that these T cells were crossreactive: they reacted not
only to human macrophages from the same individual, when these were pretreated
with the mycobacterial proteins, but also with untreated macrophages. It
seems that the T cells killed the macrophages because they appeared similar
to mycobacterial protein. What may have made them similar, the researchers
suggested, was the human stress proteins they produced (see Figure 2).

Immunologist Jon Lamb and his colleagues at the Hammersmith Hospital
took this one step further. They showed that clones of T cells from the
blood of someone with tuberculosis recognised both mycobacterial p65 and
human p65 isolated from heat-shocked white blood cells.

So it is beginning to look as if the chronic inflammation of rheumatoid
arthritis may be continually refuelled by human stress proteins. In theory,
this could happen in rheumatoid joints in the absence of bacteria. But if
bacteria were also present, they could help to stimulate crossreactive T
cells.

However, immunity to bacterial stress proteins alone will not necessarily
cause autoimmunity. A person’s genetic background also has a strong influence
in determining his or her susceptibility to autoimmune disease. In addition,
unknown environmental changes during a person’s life may alter the potential
range or ‘repertoire’ of T cells available to respond. The circumstances
leading to autoimmunity are complex and not fully understood. For example,
both people with tuberculosis and healthy people can be immune to mycobacteria.
But very few of either develop rheumatoid arthritis, and there has been
no epidemiological study to establish whether or not a link exists between
mycobacterial infection and the disease.

Antigen mimicry explains how the same T cells react to both bacterial
and human stress proteins. It also allows us to speculate that bacteria
may trigger autoimmune disease. When we are fighting a bacterial infection,
some of the immune response may become misdirected towards our body’s own
stress proteins. The subsequent inflammation could create a vicious cycle
of cell damage, stress and further T cell stimulation, mediated by cytokines.

So far, we only have evidence from animal models that this may be true.
In 1986 Irun Cohen and his team at the Weizmann Institute showed that rat
T cells immunised in the laboratory to mycobacterial antigens caused arthritis
when injected back into healthy rats. When they looked at the cells in greater
detail, the researchers found that the cells appeared to be attacking a
protein in rat cartilage. Later, Cohen, with Jan Van Eden in the Netherlands,
showed that the mycobacterial target of the cell was p65. But so far, no
one has identified the cross-reactive protein in cartilage.

David Latchman, a molecular biologist who studies stress proteins at
the University College and Middlesex School of Medicine believes that the
trigger for the disease is the increase in production or expression of human
proteins within the cell: ‘You can argue that infection with bacteria or
pathogen can be the trigger, but maybe you need an endogenous trigger also,
to raise heat shock proteins through to a level where autoimmunity is to
´Ç³¦³¦³Ü°ù.’

Other researchers at the School of Medicine suspect that cross-reactive
antibodies may also be important. ‘Whether mycobacteria per se are the cause
is open to question,’ says immunologist Peter Lydyard, but he also speculates
that infection with mycobacteria might lead to elevated levels of antibodies
to both mycobacterial and human p65 stress proteins in the sera of people
with arthritis. Even more intriguing, given our frequent exposures to mycobacteria,
is the suggestion that exposure to some bacteria may even be protective
against developing autoimmunity.

No one yet knows where the boundary lies between a controlled immune
response and uncontrolled inflammation. At the Sunley Research Centre, we
are studying synthetic peptides of human and mycobacterial p65 to find out
precisely which regions gamma-delta and alpha-beta T cells recognise. We
are also interested in finding out what first attracts T cells to the joints,
since p65 appears to be a ubiquitous protein. Because it is so common, we
believe that other autoantigens are involved. Marco Londrei, Zoe Leech and
their colleagues from the Sunley have isolated T cells which recognise collagen
type II, the cartilage-specific type of the protein collagen from the joint
or joints of one patient.

Rheumatologist Hill Gaston and his colleagues at the University of Birmingham
also see a need to identify T cell targets. They have studied T cell clones
induced by bacterial stress proteins in another disease, reactive arthritis,
which is not chronic and is triggered by bacterial infections such as gastroenteritis.
Gaston says that despite immunity to whole mycobacterial proteins there
is no proof that, in rheumatoid arthritis, the mycobacterial p65 epitopes
are shared with human p65.

This work is crucial to understanding rheumatoid arthritis and may lead
the way to new treatments for this debilitating disease. Ideally, researchers
would like to target those cells, such as T cells and other white cells,
which cause damage to tissue, so that the rest of the immune system could
continue to function without side effects. Already the Cytel Corporation
in San Diego, California is making peptides based on this principle. What
researchers there have done is to build into the design of the peptides
alterations to trick the T cells at the site of disease into non-responsiveness.
These include substitution of certain key amino acids in the peptides so
that they bind tightly to the antigen-presenting cells that activate T cells.
In this way they displace the body’s own peptides and/or make it impossible
for them to bind. Thus, the ‘bad’ autoimmune T cells fail to find their
target peptide-possibly p65-and so do not trigger their usual reactions.
This approach is called ‘peptide blocking therapy’.

In rheumatoid arthritis, it seems that the usual mechanisms for bringing
the immune system back under control have failed. Without using drugs, the
cycle of inflammation continues unabated. Researchers aim to design effective
drugs that will act before damage is done. If they could achieve this, the
millions of sufferers from rheumatoid arthritis the world over would be
only too pleased.

Julie Clayton is a post-doctoral researcher in rheumatoid arthritis
at the Charing Cross Sunley Research Centre in London.