This is the fifth in a series of blog posts about my genome, which I recently had sequenced through Illumina's Understand Your Genome program.
Last time, we manually examined some of my read mappings and called one A/A homozygous variant. I didn't choose this example at random, of course, but rather because it's an interesting variant with a fairly life-altering phenotype.
As indicated in the annotations above the reference DNA sequence, the variant lies within a protein-coding exon of ALDH2, a gene encoding a key enzyme in the pathway for metabolizing alcohol. The G$\rightarrow$A variant furthermore falls in the first position of a codon, giving rise to a GAA$\rightarrow$AAA codon substitution and a corresponding change in the encoded amino acid, from glutamic acid to lysine. The protein's primary sequence consists of more than 500 amino acids, but this single missense variant largely disables its function in the alcohol metabolism pathway.
The typical consequence of carrying this disabled ALDH2 allele, known in the literature as ALDH2*2, is probably familiar to anyone who went to a college with a significant Asian population: the alcohol flush reaction, perhaps better known as "Asian glow".
Alcohol flush reaction. This is an image not of your blogger, but rather an anonymous individual who consented to the ongoing use of his image in Brooks et al. (2009). Kudos to that dude!
Without adequate ALDH2 activity, ingesting alcohol leads to a rapid build-up of acetylaldehyde, an intermediate product of the alcohol metabolism pathway and the substrate of ALDH2 (acetylaldehyde dehydrogenase). Acetylaldehyde is the toxin mainly responsible for the flush reaction.
As a matter of both prior probability and ascertainment bias, people you've seen experience the alcohol flush reaction are probably ALDH2*2 heterozygotes, meaning they do have one good copy of the gene, with which they can metabolically limp along after drinking alcohol. By "limp along," I mean heterozygotes are typically capable of far less than 100% or even 50% of the normal ALDH2 activity. That's because the enzyme actually acts as a homotetramer, and a bad subunit has a disproportionate effect on the whole. (Impressively, the single amino acid substitution manages to wreck both the protein-substrate and protein-protein binding interfaces.) In classical genetic terms, ALDH2*2 is partially dominant.
I, on the other hand, am homozygous for ALDH2*2, and consequently I'm essentially incapable of acetylaldehyde metabolism via ALDH2. This renders me, and other homozygotes, unable to consume alcoholic beverages to any useful or enjoyable end - the flush reaction's effects are overwhelming. (The acetylaldehyde that arises upon a futile attempt is eventually disposed by a different, inefficient pathway.) Interestingly, a large portion of the scientific research on this topic is in the area of alcoholism, and in that context, ALDH2*2 is considered a protective allele for that disease. Indeed, the proportion of alcoholics with my genotype is indistinguishable from zero.
Not so with heterozygotes - and here we come to a deadly serious note. ALDH2*2 heterozygotes who build tolerance to the side effects and become heavy drinkers place themselves at dramatically increased risk of esophageal cancer, and others. That's because acetylaldehyde, in addition to provoking the immediate flush reaction, is a potent DNA-damaging agent - and heavy drinking by such individuals stews their tissues in it. Brooks et al. (2009) estimated that, among Japanese men, the total occurrence of esophageal squamous cell carcinomas - which have a very poor prognosis - would be halved if all ALDH2*2 carriers consumed alcohol lightly or not at all.
Recently there's been some promising research into a pharmacological chaperone to improve the mutant ALDH2*2 function. Now this admittedly may not seem like the most urgent therapeutic target, but there are reasons to pursue it besides to enable my recreational drinking - namely, it's becoming apparent that ALDH2 plays additional roles in modulating tissue response to ischemia, with relevance to coronary artery disease and recovery from heart attacks and open-heart surgery. In any case, a small molecule called Alda-1, illustrated below, appears to be quite effective in restoring ALDH2 function, at least in vitro and in rodent models. It will be interesting to follow ongoing research into Alda-1, with the awareness that plenty of promising agents end up proving unsuitable for use in humans.
Population genetics of ALDH2*2According to data from the 1000 Genomes project, the ALDH2*2 allele is rare or undetectable in non-Asian populations across the world; it's not called Asian glow for no reason! Among East Asians, 1000 Genomes measured an overall allele frequency of 22%, with 34% of individuals genotyped as heterozygous carriers and just 4.5% as homozygotes like myself.
Previously, Li et al. (2009) collected ALDH2*2 allele frequency data at many locations throughout the region, providing a finer-scale "map" of its prevalence. I've reproduced a version of their Figure 1 using their supplementary data, which shows an allele frequency as high as 40% in some areas of eastern China. Although they did not report the individual genotypes, that suggests most individuals in those areas carry at least one copy of ALDH2*2 ($2q-q^2$ using Hardy-Weinberg).
Minor allele frequency (MAF) for ALDH2*2 (rs671) at points throughout East Asia, as reported by Li et al. (2009). Before reusing this figure please see a note of caution in the source code.Naturally, we ought to wonder why ALDH2*2 is common in East Asian populations and no others. Did it spread there mainly through historical happenstance, the result of founder effect or genetic drift? Or was it also propelled to high frequency by natural selection, perhaps having itself conferred some survival or reproductive advantage? There seems to be little evidence for the latter hypothesis; despite some earlier signs and speculation, the locus so far does not stand out in genome-wide screens for evidence of selective sweeps. Like a great many other outcomes of evolution, there may turn out to be no deeper answer to "why?" than merely historical reasons. (ALDH2 lies within a region with relatively low genetic recombination rates, making selection particularly difficult to detect.)
Aside: most Asians also carry a variant of ADH1B that accelerates the initial metabolism of ethanol into acetylaldehyde - potentially compounding the effect of ALDH2*2's inability to deal with the latter. I'm homozygous for this variant, as well! There's much stronger evidence for positive selection associated with that one.
Moving onHad I known my ALDH2 genotype and its implications from a young age, it certainly could have saved me a few embarrassing moments over the years. As it happened, I'd pretty quickly learned to avoid alcohol through simple Pavlovian conditioning, and so it is for many less-than-life-and-death genetic predispositions: they're often perfectly well discoverable without a precise molecular test. On the other hand, it's also easy to see this as a relatively benign example of an adverse drug reaction predictable through pharmacogenetics, a category which includes some very serious cases. It's also notable that my homozygosity implies that any and all biological children I have will carry ALDH2*2 and the attendant risks mentioned above.
Determining my ALDH2 status - one single-nucleotide variant - from whole-genome sequencing was a bit like killing a fly with a bazooka. That's actually not an uncommon thing to see in this field nowadays, as we do our best to integrate our newfound high-throughput technologies with the preceding century's worth of accumulated knowledge about human genetics. We'll be back to big-data bioinformatics shortly, but I'll be on the lookout for other instructive, low-throughput excursions along the way.