Quack of the Week: John McCain

To be fair, both Democratic presidential hopefuls seem to be unaware of the science surrounding the issue^* but John McCain is being singled out for the following statement which he made last week:

"It's indisputable that (autism) is on the rise amongst children, the question is what's causing it. And we go back and forth and there's strong evidence that indicates that it's got to do with a preservative in vaccines."

Probably the same kind of 'strong evidence' that there were WMDs in Iraq. But this isn't really about McCain per se - there is no shortage of people who share this stance - it's about the idea that mercury-based preservatives (thimerosal) in vaccines are responsible for a rise in autism.

First of all, is it even true that autism and autism spectrum disorders (ASD) are on the rise? It's possible that what we're seeing is an increase in diagnosis of ASD rather than an increase in the disorders themselves - either due to improved diagnostic tests or changes in the way these disorders are classified. This paper from the journal Pediatrics puts some numbers to that idea and shows that autism diagnoses took an upswing at the same time that diagnoses of mental retardation and learning disability declined suggesting that changes in diagnostics may explain the apparant autism epidemic. Orac at Respectful Insolence blogs about this paper in greater depth.

Regardless of whether autism is actually on the rise or not, it's still important to find out the underlying cause. Are mercury-containing vaccines the culprit? The science says no. A link between thimerosal-containing vaccines is not supported by science. Several studies have been done that show no causal link between mercury in vaccines and autism.

"But still," some will say, "we avoid eating certain fish because of mercury levels, so having it in vaccines makes no sense." First of all, the mercury build-up in fish (methyl-mercury) is different from the form in thimerosal (ethyl-mercury). Ethyl-mercury has a much shorter half life and does not build up in the body the same way. And that's beside the point: in the US thimerosal hasn't been used as a preservative in recommended childhood vaccines since 2001. Yet in the US and other countries that no longer use thimerosal, autism diagnoses continue to rise. This tells us that it's not the mercury, but more likely - as mentioned above - changes in the way autism spectrum disorders are identified.

For those who refuse getting vaccinated "to be on the safe side": Don't. Vaccines are responsible for the eradication of smallpox, near-eradication of polio and a host of other diseases (presumably, this is why anti-vaxers feel they can get away with it - because there's little fear of smallpox, polio and the like). Not receiving these routine childhood vaccinations has real public health implications. It's irresponsible to not have your children vaccinated.

As for John McCain, it's irresponsible for HIM to make such strong claims about a connection between mercury in vaccines and an autism epidemic. He needs to surround himself with better science advisors than that.



^*Clinton says: "I am committed to make investments to find the causes of autism, including possible environmental causes like vaccines. [...] I will ensure that all vaccines are as safe as possible for our children by working to ensure that Thimerosal and mercury are removed from vaccines."

Obama says: "An Obama administration will go where the science and the facts lead us, whether it is about climate change or toxic heavy metals in our environment. [...] I support the removal of thimerosal from all vaccines and work to ensure that Americans have access to vaccines that are mercury free."

(source: Age of Autism)

Some people argue that these viewpoints are worse than the stronger McCain statement.

16p11.2 linked to autism

We've discussed in the past how copy number variation in the genome may be the smoking gun behind autism. Well a new paper in the NEJM attributes 1% of all autism cases to a CNV on chromosome 16 at 16p11.2. Hopefully this discovery will shed some light on physiological basis of autism. There are some indications that autism may be reversible if only we knew what was wrong. In fact it seems that even a fever can temporarily improve behavior of autistic patients... The paper fails to really get into the details, but find bellow a list of genes in that region, and what little information I could find on them. Notice that there is an enrichment in brain transcripts, implying a multigenic basis of the disease. Place your bets now on which gene will the next big breakthrough:

BOLA2: cell proliferation & cell cycle
GIYD1/2: sulfotransferase family
SULT1A3/4: sulfotransferase family
spn: important for lymphocyte funstion, is also defective in WAS, an x-linked mental retardation
QPRT: Elevation of qprt in the brain may be involved in the pathogenesis of neurodegenerative disorders
c16orf54: hypothetical protein
kif22: kinesin-like protein family
maz: MYC-associated zinc finger protein
prrt2: glucan 1,4-alpha-glucosidase activity
c16orf53: hypothetical protein
mvp: resistance to lung infection
cdipt: phosphatidylinositols biosynthetic pathway
sez6l2: homologous to a seizure related genes
asphd1: A transcript abundant in the brain, catalyses oxidative reactions in a range of metabolic processes
KCTD13: binding partner to PCNA at replication foci
loc124446: hypotethical protein
toak2: ?
hirp3: ?
ccdc95:coiled-coil domain containing 95, function unknown
doc2a: implicated in neurotransmitter release
fam57b: transmembrane protein of the cerebellum
aldoa: aldolase
pp4c: Serine/threonine-protein phosphatase
tbx6: required for choice between mesodermal and a neuronal differentiation pathway during gastrulation
ypel3: cell division,mitotic spindle
gdpd3: glycerol metabolism
mapk3: erk signalling and growth. erk1 knockout mice have behavioral problems


ref: January 9, 2008 (DOI: 10.1056/NEJMoa075974), in print February 14, 2008

why trial by jury sucks

Who doesn't like democracy right? Well sometimes, a group of people is worse at taking decisions than multiple independent individuals. At least that's Dawkin's very convincing opinion which starts like this: "Trial by jury must be one of the most conspicuously bad good ideas anyone ever had. Its devisers can hardly be blamed. They lived before the principles of statistical sampling and experimental design had been worked out. They weren’t scientists. Let me explain using an analogy. And if, at the end, somebody objects to my argument on the grounds that humans aren’t herring gulls, I’ll have failed to get my point across.".

All of this to introduce this story, about a federal hearing of a court case against the US government presented on behalf of a family of a girl afflicted with severe autism. You may have heard of it, there has been a long suspicion that childhood vaccines such as the Measles-Mumps-Rubella (MMR), the vaccine additive thimerosal, or the mercury preservative in vaccines may be the cause of autism. Yet time and time again studies have shown the absence of causation, and that the correlation between the age of onset of autism and childhood vaccines is just that, a correlation. Even the correlated rise of autism cases and vaccination in the past decades can be explained simply by better records and diagnosis of autism, there probably hasn't been a major increase in actual cases. So lets hope that this particular jury is not a loose canon and can appreciate the subtleties of causation and correlation, and that there is a good expert witness at that trial, since there are 5000 other similar cases for which it might set a costly precedent.

Genomic diversity and cancer

First we sequenced the human genome, then we started working on SNPs to generate the Hapmap, and finally we started mapping methylation sites to generate the epigenome. Together these polymorphisms generate the bulk of the human phenotypic diversity. Recently it has come to my attention that there exists an additional source of diversity between humans: copy number variation (CNV). This, I am ashamed to admit, was complete news to me. Apparently you and I differ not only in the alleles we carry but also in the number of copies of the allele we carry. Traditionally we think of allelic variation has having 0, 1 or 2 alleles of a particular gene. However it is also possible to have the locus of the gene deleted, or have that locus duplicated or in multiple copies whitout having any sort of disease. Imagine if you had 4 copies of the gene for brown eyes versus someone who was just a regular homozygous brown+/+. Because of gene dosage your eyes would be darker. A recent study estimates that CNVs account for up to 17% of genetic variation in gene expression. And this isn't even a rare event, most people will have copy variants within their genome, and there are already almost 1500 variable regions covering 12% of the genome. They do tend to be outside of coding areas, probably because of negative selection, but are prevalent in regions containing genes important for immunity, environmental stress, etc... in other words regions where faster evolution is more likely to be beneficial. They may also account for complex "spectrum" diseases like autism. Cancer is another obvious disease which has been known to use this evolutionary mechanism for a long time. It is an evolutionary disease that uses all of the variation producing tricks evolution has to offer to mutate it's way to full blown malignancy. It uses normal genetic events like gene conversion, duplication, deletion, methylation, single base mutation but at an accelerated rate. If it manages to stay alive while generating enough diversity to evade the immune system and come-up with solutions to all of the built-in roadblocks we have it will have reached an "escape velocity". At this point it has reached a perfect balance of mutation rate / functionality in the genome, and has deviated so much from a normal cell it is barely recognizable. The corrolary is that the mutation rate is a fine balance, too low and you can't escape the body's counter measures, too high and you risk reaching catastrophy where the cell is no longer able to duplicate sucessfully. This is why radiation therapy is like having too much of a good thing for cancer. Remarkably the same thing is observed in viruses, as I recently learned at a seminar. The mutation rate is a very fine balance between evading the immune system but not compromising virulence. The lesson of the story is, if we can figure out how to increase mutation rates in cancer cells selectively, we might have a way to fight fire with fire...