Last month, Ars Technica reviewed Epidemic. The write-up mostly works like a synopsis. But in the final section, the review veers into a kind of reality distortion field.
Immune dysregulation is not implicated in cancer, cardiovascular disease or obesity, the reviewer proclaims.
Yet inflammatory involvement in all three is fairly well established, at this point. It’s verging on old news, in fact.
Some recent reviews on inflammation in colon cancer; adipose tissue, inflammation and cardiovascular disease; and obesity, inflammation and Atherosclerosis.
That’s not what I wanted to talk about, though. I wanted to talk about genetics.
The review finishes up with a common trope: The causes of autism are mostly genetic, it proclaims, not environmental.
Au contraire, I must declare. First, there’s the epidemiology over time. If a disorder has increased by an order of magnitude in just a few decades — well, that’s unlikely to be genetic. “Bad” genes don’t spread throughout a population that quickly.
The actual increase in autism is probably significantly less than ten-fold. This careful study in California, for example, found it to be less than half the 7- to 8-fold measured increase in that state. Still, no scientists I’ve encountered argue that the entire increase is an artifact. And a tripling — which is plenty alarming — is still far more than “bad” genes can account for.
Then there’s the link between autism and autoimmune, allergic and metabolic disorders in the mother. (See the source list for refs.) All three of these risk factors have increased in prevalence in recent decades. Between 5 and 8 percent of Americans have an autoimmune disorder. Eight percent of adults have asthma. Obesity, a rough predictor of metabolic syndrome, has more than doubled among adults since 1970. It afflicts one-third of Americans.
About the old mantra — correlation does not imply causation — consider the evidence: elevated markers of inflammation in amniotic fluid from children who later develop autism, coupled with animal models of prenatal immune activation producing behavioral problems in offspring. (As an aside, experiments on sheep demonstrate that obesity increases proinflammatory factors at the placenta. No behavioral outcomes were measured in this study, though.) Together, this is somewhat compelling evidence for a causal relationship between prenatal inflammation and behavioral disorders in at least a subset of ASDs.
The faith in genetic explanation for autism, meanwhile, rests on oldish twin studies. The first, published in 1977, absolved parents — and mothers in particular — from accusations of “cold parenting.” It found 90 percent heritability of autism in identical twins, and none in fraternal twins.
A 1995 study found a concordance of 92 percent for autism spectrum disorders among identical twins, and just 10 percent for nonidentical twins — a seeming confirmation of the earlier study. Half the subjects in this study in fact came from that earlier study.
But then came a larger, better designed twin study including 192 twin pairs — nearly tenfold the sample size of the 1977 study, which was 21, and roughly quadruple the 1995 study. It also required a doctor diagnosis of autism, which previous studies had not.
This study, published in 2011, found that concordance for strict autism between identical twins was just 58 percent. For non-identical twins, however, it was 21 percent.
In other words, concordance was lower among people with the same genome than previously reported. And it was higher among people with non-identical genomes. All of which strongly implicates environmental factors, especially prenatal ones.
The scientists put a number to it, actually: 55 percent of cases were attributable to environment, while just 37 percent could be pegged to genetics. Environmental influences were dominant, in other words.
As Peter Szatmari wrote in the Archives of General Psychiatry, “The autism field can now join the chorus and ask, ‘Where did all the heritability go?’ We now appear to have an answer, at least in part: those original estimates were inflated.”
What of the geneticists valiantly seeking the autism genes? Different groups have identified hundreds of risk alleles at this point, most of which increase odds by miniscule amounts. In some cases, they’ve failed to replicate each other’s findings. And even they say that the constellation of genes they’ve identified so far only explain 10 – 20 percent of autism cases.
Add to that the burgeoning research on the prenatal origins of disease generally — an argument based on environmental influences acting via epigenetics, not genetics — and, as this nice (and free) review points out, “The need to move forward with more extensive investigation of environmental risk factors in autism is now widely accepted.”
NIH echoes the call, emphasizing the necessity for more research into environmental causes of autism.
But wait! What about those de novo mutations, spontaneous alterations to genes that aren’t found in parents, but which, according to some studies — most recently of old dads — are enriched in autistic children?
As Thomas Insel, director of the National Institute of Mental Health, points out, “It is important to understand that de novo mutations may represent environmental effects. In other words, environmental factors can cause changes in our DNA that can raise the risk for autism and other disorders.”
Furthermore, when it comes to gene variants — not the same as de novo mutations — a given variant does not always produce the same outcome. Depending on environment, in fact, it may produce opposite results. Some of the most intriguing research on this comes from Finnish and Russian Karelia.
Both study populations are ethnically Finnish (i.e. Karelian), but they’ve been separated since WWII. They live at the same latitude, have the same prevalence of genes associated with allergic and autoimmune disease, but live in very difference socioeconomic circumstances. Remarkably, this study found that the very alleles associated with asthma in Finland PROTECTED against the disease in poorer Russia.
Likewise, HLA alleles associated with celiac are equally prevalent on both sides of the border. Yet the prevalence of the disease itself drops by a factor of five on the Russian side compared to the Finnish (rounded out: 1 in 100 in Finland versus 1 in 500 for celiac in Russia). Environment PLUS genes produces disease, it seems, not genes alone — and yes, the Russians consume plenty of gluten, more, in fact, than the Finns.
Some are beginning to plumb this kind of gene-environment interaction in autism. This German group, for example, finds that mutations associated with tuberous sclerosis — and also autism — only produced behavioral problems in mice that were ALSO exposed to prenatal inflammation.
What’s the takeaway? “Genes are not Stalinist dictators,” David Barker, a pioneer in the Fetal Origins field, once said. “They live in a democracy, and what they do is conditioned by what else is going on around them.”
The central argument of my book is that “what’s going on around them” has changed so dramatically in the past 150 years that once beneficial genes are now proving deleterious.
So back to Ars Technica’s — and others’ — quasi-religious faith in genetic explanations. What accounts for the fingers-in-both-ears insistence on genes in the face of so much evidence to the contrary? Beyond pointing out that it has little to do with science at this point, hard to say.
But I will note the following. I met the Ars Technica reviewer, Diana Gitig, for coffee. She was pleasant, engaged, interested. And she related an anecdote that I think nicely illuminates this question. We were discussing Intelligent Design — the idea that an omniscient being created life on Earth. Engineers, she said, often secretly ascribed to intelligent design. Why? As professionals who worked from meticulously laid out plans, they found the chaotic and unplanned process of evolution unsettling. The notion of central planning was far more appealing. Thus the preference for an omniscient central planner.
She meant this in a joking way, I’m sure, as do I. But sometimes I suspect that this type of bias — an esthetic preference, really — is widespread. We prefer clean, simple answers that involve a unidirectional chain of causality. That’s the genetic explanation. Too bad reality doesn’t always comply.
