Progress on MS, and a somewhat typical genetics story

Multiple sclerosis (MS) is yet another chronic often devastating disease for which cause is not known.  Some epidemiological studies suggest an infectious cause, since cases often clusters geographically, but the fact that prevalence tends to increase with distance from the equator in childhood or adolescence has suggested to other researchers a link with sunlight exposure, and thus perhaps to vitamin D deficiency, because sunlight is important to vitamin D synthesis.  Because causation isn't clear, many also suggest a gene-by-environment interaction, in which some genetic variants react differently to a given environmental factor than other variants do.  And now a new study, reported all over the web, including here, and published in the Annals of Neurology (here, with subscription), reports a genetic link to vitamin D and risk of MS.

MS is an autoimmune disease characterized by degradation of the myelin sheath that surrounds and protects nerve cells.  The sheath is like insulation in electric wires, and protects the nerve signal from dissipating as it travels from the brain to where its action is triggered.  Nerve damage occurs due to the inflammation that happens when the body's immune cells attack the nervous system, interrupting the signals sent along the affected nerves.  Sex and age are related to risk, and among sufferers the symptoms vary widely, depending on which nerves are damaged.  Prognosis also varies widely, and is difficult to predict.  Many studies, including but long before GWAS, have found HLA class II genes, and others with modest effect, but the effect varies by whether cases are sporadic or familial.  This all suggests that the disease is heterogeneous, with varied causation -- not at all unusual for complex chronic diseases (nor even for 'simple' genetic disease).   

The aim of the study just reported, carried out by Oxford researchers in collaboration with colleagues at the University of Ottawa in Canada, was to look for rare genetic variants that might explain the excess of cases in some of the 30,000 families participating in a Canadian study of MS, although they do point out that shared family environment could be important.  They began by sequencing the exomes -- all the gene coding regions -- of a randomly selected affected individual from 43 of the 82 families with four or more members with MS (a fairly rare occurrence given that there seems to be a genetic component to risk).  Using families with multiple affected members is well-known to increase the likelihood of finding relative simple and tractable genetic causation.  They didn't in fact find any shared rare variants in this way, so they then focused on MS genes identified in previous GWA studies they had done, that is, comparing cases with unrelated controls.

This eventually led them to zero in on a single change in the CYP27B1 gene that was shared by at least one unaffected parent in the multiply affected families, and by all the affected members.  This is a candidate gene previously tentatively suggested by a GWA study.  They used data from 3046 additional parent-affected child trios, 422 parent-affected sibling pairs and 1873 healthy individuals (in total 12,579 people) from their study to replicate their results.  

As reported in the story in MedicalExpress, the lead author, George Ebers at Oxford, concluded:

"The odds are very much less probable than being hit by lightning", he said. "Is this gene variant causative in multiple sclerosis? Pretty much! The cornerstone for causation has always been the strength of association."

Further, this mutation is known to be associated with low levels of vitamin D, which implies that MS is somehow linked with vitamin D levels.  In the rare instances when people have 2 copies of this variant, they also have rickets, which is also associated with vitamin D levels.   From the conclusion to the paper:

Most biological effects of vitamin D are mediated by calcitriol acting via the vitamin D receptor (VDR). We have previously shown that vitamin D regulates more than 80% of MS associated genes 25, and thus it is likely that lower levels of calcitriol as a result of CYP27B1 mutations leads to a disruption to critical gene-environment interactions important for the developing immune or nervous system which then predisposes to MS.

Indeed, because there are unaffected family members with the mutation -- it was present in 33% of genotyped unaffected family members -- clearly any genetic cause here is more complex than one-to-one.  Or, in rather imprecise genetic jargon, it's "incompletely penetrant".  And indeed, it's a rare variant, present in only a subset of people with MS.  And, clearly, if there is a vitamin D component to causation, it's also not one-to-one, because while prevalence of MS is higher in higher than lower latitudes, it's still a fairly rare disease, even in people with low levels of vitamin D.  The idea of the latitude effect has long been thought to involve sunlight exposure amounts, but other things are correlated with latitude, including immune exposures.  So, as with most complex chronic diseases, causation doesn't come down to simply one mutation or one environmental factor.

The study does appear to implicate vitamin D biology, though in a curious way since the penetrance even of the clearest alleles is far from complete. There are either many pathways to MS, or there is a specific as yet unspecified subset of the diseases with this causation.  It's not really clear what fraction of risk this new gene explains, but the point is not that the gene 'for' MS has been found, but that this particular mutation is associated with low vitamin D levels, as well as some risk of MS.  The roughly 4-1 female-male sex ratio for MS in general, and its usually adult onset, and highly variable pathology all point to very stochastic (probabilistic) effects and/or cofactors that mean there is still much to learn.   But this may be a case where GWAS have made a somewhat greater than usual contribution to risk, not by identifying genes with strong effects but rather by suggesting a train of thought that may perhaps lead to more systematic understanding of a devastating disease.

Hope, not hype -- stimulating remyelination may be a possible route to multiple sclerosis (MS) therapy

We don't do this very often, but here we point out an excellent use of genetics with a potentially very important therapeutic outcome.  Authors writing in Nature Neuroscience on Dec 5 report that the ability of myelin sheaths in the CNS to regenerate following acute demyelination is limited in dymyelinating diseases such as multiple sclerosis (MS), but that with pharmacological and genetic manipulation methods, they were able to enhance remyelination in rats.  This has the potential to be a significant advance for treating demyelinating diseases. 

The molecular basis of remyelination has not been well-characterized, but it has been known, as stated in the paper, that following demyelination, "adult oligodendrocyte precursor cells (OPCs) can migrate to the area of injury, differentiate into oligodendrocytes and restore myelin sheaths".  In individuals with MS, in part the problem is that following the demyelination episodes that characterize the disease, OPCs don't differentiate into myelinating oligodendrocytes.  (OPCs are the CNS equivalent to Schwann cells, which insulate axons.)

To identify genes of interest in the remyelinating process, Robin Franklin's group demyelinated nerve sheaths in rats with a toxin, and generated a list of all the genes expressed in the lesions during the remyelination process, a complete "transcriptome".  They found thousands of genes differentially expressed over time in the lesions as they regenerated myelin, including, early on, genes involved in the immune response, and later, a number of genes already known to be involved with myelination, cell metabolism and proliferation and differentiation.

These results show that the overall molecular signature of CNS remyelination involves distinct and temporally regulated signaling pathways that are characterized by active inflammation at 5 dpl [days post lesion] and by the initiation of remyelination at 14 dpl.

But one gene was of special interest because it was one of the most significantly upregulated genes at the time when remyelination was occurring, 14 dpl, and clustered with many genes involved in myelination.  This is retinoid X receptor gamma, RXR-gamma, found in the tissue of individuals with multiple sclerosis as well, and involved in the regulation of cell proliferation, differentiation and apoptosis, which they verified experimentally.  The paper shows this to have been a very careful characterization of this gene and its role in remyelination. 

To determine whether they could effectively activate RXR and promote CNS remyelination, they tested the effects of 9-cis-retinoic acid (9cRA) on RXR activity.  9cRA is an RXR ligand which is known to activate transcription of MBP, or myelin basic protein, which is expressed in differentiated oligodendrocytes.  If RXR signaling is involved in differentiation in these cells, it would be via MBP.  It turned out that OPC cultures treated with 9cRA showed indication of OPC differentiation, and in vivo tests of the effects of RXR signaling via this pathway were positive. They then determined that, in culture, in OPCs in which RXR activity was blocked, oligodendrocyte differentiation was inhibited.

So, this series of experiments, from gene discovery to characterization of function, suggests a possible pathway to treatment for demyelination diseases -- stimulate RXR activity.  A short interview with the senior author in the story on the BBC website shows him to be cautiously hopeful, but measured in his promises for a cure. When pressed on when this work might lead to treatment, he said that it was very difficult to say, but perhaps within 15 years. 

The BBC quotes him:

Professor Robin Franklin, director of the MS Society's Centre for Myelin Repair at the University of Cambridge, said: "Therapies that repair damage are the missing link in treating MS.

"In this study we have identified a means by which the brain's own stem cells can be encouraged to undertake this repair, opening up the possibility of a new regenerative medicine for this devastating disease."

The study takes advantage of the idea of genes for processes rather than things per se.  Myelin is, in a sense, an insulating compound: it's not alive, but is secreted by cells. Rather than try to apply some sort of coating to individual nerve cells by therapy the idea is to induce the cells to do it, when they're defective in that process.  That means tinker with signaling!  Thus, the problem here is with a signal reception and the consequent disruption of subsequent signal cascades leadiing to cellular production of myelin. By activating the signal efficiency, the downstream event is triggered -- and no meddling with the many intermediate steps is needed, as they are intact in the individual's genome.  In that sense, or rather in the way we've tried to describe in our book The Mermaid's Tale, it is communication and cooperation that have failed, and here, fortunately, appear to be able to be restored.

This is an excellent example of the use of genetic knowledge to explore a disease and potential treatment pathway.  A lot of work still must be done to figure out, among other things, how to stimulate remyelination in vivo, so caution is warranted.  Still, we hope along with the researchers that their work continues to be fruitful.

First 2010 Misrepresentation of the Year Award Goes to Nature

The first 2010 Misrepresentation of the Year Award goes to Nature magazine for last week's cover story. The caption on the cover is "The MS Genome" and is accompanied by silhouettes of one person standing and her twin in a wheelchair. But is the story about 'the multiple sclerosis genome' (whatever that would be)? No.

Compare the authors' actual title of their report to the cover's:
"Genome, epigenome and RNA sequences of monozygotic twins discordant for multiple sclerosis."

This is a representative, scientifically responsible title for the paper. There is a difference between genomes from MS patients and 'the MS genome.' What the  authors report is the complete sequencing of the genomes of one discordant pair of identical twins (where one twin has MS and the other doesn't), and gene expression differences between 3 sets of discordant identical twins. They discuss issues of data accuracy and replicability, and they find no genetic signature that can explain the discordance.

• Is the story well-done work? Yes, we're not experts but it seems to be
• Is the science valid? Yes, it seems to be
• Have the authors explained MS? No, but they don't claim to have done so
• Have they found the MS genome? No and they don't claim that either
• Have they identified which genome is 'the' MS genome? No
• Have they shown that there is such a thing as 'the' MS genome? No
• Is the story an important contribution to the understanding of MS? That's debatable

Genetic differences between the two twins in each pair were found, presumably the result of somatic mutations that occurred during or after embryogenesis, but none were associated with MS. Because MS has a curious relationship to climate and has sometimes been attributed to viral infection, sequences that weren't identifiable as human from each set of twins were aligned with viral genomes. That was a clever thing to try, and might have identified an MS 'infectome', but unfortunately no differences were found between the viral load of affected vs. unaffected twins.

The twins did not show replicable discordance at the immune system (HLA) genes, or other genes, that previous mapping and other studies had shown are likely contributors to MS risk. The twins varied a lot in terms of DNA sequence (presumably, somatic mutation) and gene expression levels--but in a way that could be attributable to their MS discordance.

This is another installment in the Life is Complex Department. We said above that this may not be an important contribution, because there are so many possible reasons why the authors didn't find what they hoped to find. That would not be their fault, and indeed, their work shows some of the many reasons why these kinds of data are problematic (sequencing errors, difficulty replicating expression levels, and so on). Perhaps they should have known this would have been the case from other similar kinds of studies (such as comparing genomes of cancer and normal tissue from the same person).

They confined their gene expression comparisons to a very confined set of sites -- necessarily, given current sequencing methods -- but that means that in no sense was this study exhaustive. It could be said that this was a premature use of new technology. But that's not our issue here!

The idea was a clever one, but perhaps built too much on hope against what we know about disease complexity. In general, one would be surprised if a clear result had been found, because these twins did not even have the suggested risk genotypes at those candidate genes, and because a genetic signature for MS has been so elusive before now.

Perhaps identical twins are not as good a comparison as they seem on the surface, since we know that stochastic and somatic changes can be responsible for many differences during life (that's what the cancer studies show).

Perhaps twins that did carry the suspected HLA or other risk genotypes, but who are nonetheless discordant for MS, would provide a more cogent comparison. Discordance in them would suggest that something else--perhaps environmental pathogen exposure or somatic mutation in other genes in one of the twins--was responsible.

The etiology of MS remains mysterious and this was one way to take a shot at a discovery.

But multiple sclerosis is no joking matter. The Nature cover is a kind of cruel disservice to those who suffer from MS, or their loved ones, by suggesting that the genetic cause had been found, with the obvious innuendo that a cure is just round the corner. Why else the cover story in Nature??

Their kind of misrepresentation is what diverts resources to inefficient or even lost causes. In fact, in no way has 'the' MS genome been found, nor is it defensible to suggest that there is such a singular thing. And only half the studied people were affected. Maybe the affected twin experienced somatic genetic changes, while the unaffected twin reflected the inherited 'resistant' or 'normal' genome of the pair?

It's not the authors but the journal that is responsible for the misrepresentation. It was a shot in the dark, perhaps, that had no business as a cover story. We hope that research will not be discouraged as a result--that is, that the authors' lack of positive findings does not lead to their work being thought of as 'negative' (if the bait-and-switch of the journal cover leaves hopeful readers deflated when they see the real story).

Nature has been quite successful at its attempt to match the quality standard of People magazine, and we assume that will continue, as they're pros and they know how to reach their market. Their naked savage and 'ancient' Khoisan genome cover of a couple of months ago ran this one a close second, and we wrote about that at the time. Though we're sure they would deserve it, we want to keep these Misrepresentation Awards one per customer per year, so we hope Nature (but not People) will wait for its next one til 2011.