News Synthesis | 2 Apr 2012

Whodunit?

Risk I: Overview of risk factors in multiple sclerosis

Amy Maxmen

If the hunt for causes of multiple sclerosis were a murder-mystery board game like Clue, investigators would be in the midst of one epic match. In that game, players race to figure out which suspect committed murder, where the crime took place, and with what weapon. In MS, investigators similarly scramble to learn which genes, viruses, and other factors trigger the harmful inflammation and debilitating nerve damage that characterize the disease, where in the body these agents act, and when during a person’s life they deliver their blows.

But uncovering the truth is a far bigger challenge in the case of the disease than it is in the game. Solving Clue’s mystery depends only on ruling out innocent characters and incorrect locations and weapons; MS does not lend itself to such a systematic exercise. In the disease, for instance, the suspects don’t act alone. If they did, the sleuths who have tracked MS over the past century almost certainly would have solved the puzzle already. Instead, investigators have learned that MS is not like cholera, which is contracted only after infection with the bacterium Vibrio cholerae, nor like Tay-Sachs disease, which develops only when a child inherits matching defects in a particular gene.

Rather, the agents that can underlie MS don’t appear to be present in 100% of the cases, nor does their presence always trigger the disease. For example, not all MS patients share the same genetic variants linked to the malady, and many people who possess the variants never get the disease. These troublemaking genes could present a problem only in the context of particular risk factors—environmental as well as genetic ones—and pinpointing relevant interactions is tricky. As a result, identifying the causes of MS will be as difficult as winning a game of Clue in which multiple Professor Plums, Miss Scarlets, and Colonel Mustards inhabit various rooms where crimes may or may not have been perpetrated, with wrenches, ropes, or revolvers.

“Different genes can load the gun, and the environment pulls the trigger,” says Pierre-Antoine Gourraud, a neurologist at the University of California, San Francisco (and an MSDF scientific adviser). “But there may be very different ways to pull the trigger.”

Despite the complications, investigators march onward in their risk-factor hunt because rooting out the causes of MS could help banish the disease. Even if they can’t prove a suspect guilty, strong implication might provide reason enough to intervene, says Kassandra Munger, an epidemiologist at Harvard School of Public Health in Boston. "We may not be able to prove causation without a doubt,” she says, “but we can still take action if the evidence is consistent with causality."

When the root of a disease isn’t straightforward, epidemiologists often turn to a classic list of nine items to help them assess whether a candidate likely contributes. These criteria, published nearly 40 years ago by epidemiologist and statistician Austin Bradford Hill, help investigators home in on certain suspects and rule out unlikely ones (Hill, 1965). First, the more often a factor coincides with the outcome of interest and not with the population at large, the more likely the association is causal. Second, this association should be found in multiple populations. Other Bradford Hill criteria favored by MS sleuths include the requirement that the cause occurs before the effect, that it could lead to the effect through a reasonable mechanism, and that greater exposure to the cause leads to a greater risk of the effect. A lack of data on these principles—temporality, mechanism, and dose dependence—doesn’t exclude a candidate, but evidence of their existence strengthens a case.

Investigators have eased up on many suspects over the years when they failed to meet Bradford Hill’s criteria of association and consistency. For example, the bacterium that causes a form of pneumonia, Chlamydia pneumoniae, once landed near the top of investigators’ list when bits of the microbes were detected in the spinal fluid of 97% of MS patients and only 18% of controls (Sriram et al., 1999). But support for the case against this pathogen diminished when other groups could not reproduce these findings (for instance, see “C. pneumoniae and MS: fact or fiction” in Giovannoni et al., 2006). This example illustrates why Bradford Hill’s consistency criterion is vital to MS research; tests for pathogens, nutrients, and even MS diagnosis can be misleading, so initial reports need to be confirmed by other methods and research groups.

A handful of environmental and genetic factors have stood up to many of Bradford Hill’s criteria. They’ve emerged from features that are more common among patients than in the community at large. For example, women are more likely than men to have MS, so gender-related factors such as hormones are suspected to be involved in the disease (see “Discriminatory Disease”). The common Epstein-Barr virus (EBV) is also a suspect because more people with MS have been infected with it than have people in the general population (see “Viral Villain”). Other fishy factors include low levels of vitamin D (see “The Sunshine Suspect”), smoking and stress, and certain genes (see “Genetic Associations”), as well as crossroads where genes and other factors intersect.

Investigators are interrogating these suspects, but the candidate troublemakers are not offering up easy evidence. Some factors, for instance, are difficult to measure. A person’s experience with stress, say, is harder to calculate than assessing whether he or she has been infected by EBV. The lag between the time when putative perpetrators first exert their influence and MS onset occurs complicates researchers’ ability to pin down when and how the suspects might stir up harm. Hundreds of other events take place in the intervening years, which hampers strategies that rely on the process of elimination to identify sources of the malady. Furthermore, the imperfect nature of memory can muddle exposure data that must be recalled from long ago—such as whether one took vitamins during childhood or how often one’s parents fought (a gauge of stress). And some factors might exert different effects, depending on the age at which an individual encounters them. Figuring out whether and how any of the possible culprits team up poses an additional challenge. It is already clear that some combinations add up to more—or, in some cases, less—than the sum of their individual contributions. Scientists are just beginning to get a handle on such situations. Some investigators are performing wet-lab experiments to unravel the details of the interactions, and others are building new statistical computational tools to probe how environmental agents combine with genetic variants to alter biochemical pathways in the body (see, for instance, Moore et al., 2010, for a discussion of new bioinformatics strategies for studying genome-wide association studies).

The ultimate test for a causal relationship is currently out of reach. Just as jailing a serial killer ought to end his brand of murders, eliminating a guilty agent should make MS disappear. This experiment isn’t feasible, however, because methods to restrain most of the risk factors are not available.

However, even if a single risk factor’s detrimental activities could be foiled, MS frequency might drop. Therefore, while seeking signs of causality, MS investigators also consider how they might keep the suspects in check. “If scientists know an intervention is safe and relatively simple, health officials may want to take policy action even without understanding the mechanism,” says Gary Cutter, a biostatistician at the University of Alabama, Birmingham. “People used aspirin to treat pain for years before we knew how it worked.” For this reason, although vitamin D insufficiency is no stronger a risk factor than gender or certain genes, it’s the best candidate for preventing MS because supplements raise blood levels and have few, if any, side effects.

Thus, for vitamin D and other suspects, investigators are simultaneously seeking signs of causation and evidence that an intervention might slow the rate of new MS cases. Despite playful similarities to Clue, this is no game. The health of future generations relies, in part, on investigators’ next moves.

Next article in series: "The Sunshine Suspect"