Archive for the ‘Genetics’ Category
A Brief Overview Of Glial Priming, How It (Probably) Applies To (Some Cases Of) Autism, And Worrisome Speculation On A Model Of A Low Penetrant Effect
Posted on: October 26, 2012
Hello friends –
The concept of glial priming (and implicit double multi hits) is the nexus of developmental programming, low penetrant effects, and an altered microglial responsiveness, a blueprint for a change in function in the tightly entangled neuroimmune environment; sort of an all time greats theory mashup for this blog. The basic idea is that microglia can become sensitized to insults and subsequently respond to similar insults with greater robustness and/or for increased timespans later in life. Here is a snippet from Microglia in the developing brain: A potential target with lifetime effects on the primed glial phenotype:
There is a significant amount of evidence regarding what is often termed ‘‘priming’’ and ‘‘preconditioning’’ events that serve to either exacerbate or provide neuroprotection from a secondary insult, respectively. In these states, the constitutive level of proinflammatory mediators would not be altered; however, upon subsequent challenge, an exaggerated response would be induced. The phenomena of priming represent a phenotypic shift of the cells toward a more sensitized state. Thus, primed microglia will respond to a secondary ‘‘triggering’’ stimulus more rapidly and to a greater degree than would be expected if non-primed.
Glial priming may be the fulcrum on which much of the underlying early immune activation research balances, the machinery that drives environmental influences during development leading to irregular neuroimmune functionality through the lifespan. Even though this type of finding is not really unexpected when considered within the prism of programming effects in other systems and the perturbed immune milieu in many (all?) neurological disorders, it is still pretty cool.
The first paper that I read that specifically mentioned glial priming was Glial activation links early-life seizures and long-term neurologic dysfunction: evidence using a small molecule inhibitor of proinflammatory cytokine upregulation, (Somera-Molina KC , 2007) which totally kicked ass. They brought a lot of heat at design time of the study; (very powerful) seizures were induced /saline given in animals at postnatal day 15 and 45; at day 55 animals were analyzed and showed distinct increases in microglial activation, neurologic injury, and future susceptibility to seizures in the ‘two hit’ group (i.e., animals that got seizure inducing kainic acid instead of saline on both day 15 and 45). Even better, it was shown that a CNS available inhibitor of inflammatory cytokine production rescued the effect of the seizure. In other words, it didn’t matter if the animals had a seizure, what mattered was the presence or absence of an unmitigated inflammatory response associated with the seizure.
Treatment with Minozac, a small molecule inhibitor of proinflammatory cytokine upregulation, following early-life seizures prevented both the long-term increase in activated glia and the associated behavioral impairment.
That is an important step in understanding the participation of inflammation in seizure pathology. There were also observable effects (worse) in animals that got seizures just once, if they got induced on day 15 versus 45, and even worse symptoms for the “double hit” animals. That was pretty fancy stuff in 2007. The similarity in terms of seizure susceptibility really reminded me of another paper, Postnatal Inflammation Increases Seizure Susceptibility in Adult Rats, which also showed altered susceptibility to seizures in animals subjected to seizures in early life, with the effect mediated through inflammation related cytokines. Here, however, the same effect observed, but with the addition of clinical evidence of chronically perturbed microglia phenotype in the treatment group. Nice!
The same group followed up with Enhanced microglial activation and proinflammatory cytokine upregulation are linked to increased susceptibility to seizures and neurologic injury in a ‘two-hit’ seizure model (full version), with more of the same. Here is part of the Discussion:
First, in response to a second KA ‘hit’ in adulthood, there is an enhancement of both the upregulation of proinflammatory cytokines, microglial activation, and expression of the chemokine CCL2 in adult animals who had previously experienced early-life seizures. Consistent with the exaggerated proinflammatory cytokine and microglial activation responses after the second hit, these animals also show greater susceptibility to seizures and greater neuronal injury. Second, administration of Mzc to suppress of the upregulation of proinflammatory cytokines produced by early-life seizures prevents the exaggerated cytokine and microglial responses to the second KA hit in adulthood. Importantly, regulating the cytokine response to early-life seizures also prevents the enhanced neuronal injury, behavioral impairment, and increased susceptibility to seizures associated with the second KA insult. These results implicate microglial activation in the mechanisms by which early-life seizures lead to increased susceptibility to seizures and enhanced neurologic injury with a second hit in adulthood.
Not only that, but the authors speculated on the possibility of a rescue effect through neuroimmune modulation!
Our data support a role for activated glia responses in the mechanisms by which early-life seizures produce greater susceptibility to a second neurologic insult. The improved outcomes with Mzc administration in multiple acute or chronic injury models where proinflammatory cytokine upregulation contributes to neurologic injury (Hu et al., 2007; Somera-Molina et al., 2007; Karpus et al., 2008; Lloyd et al., 2008) suggest that disease-specific interventions may be more effective if combined with therapies that modulate glial responses. These results are additional evidence that glial activation may be a common pathophysiologic mechanism and therapeutic target in diverse forms of neurologic injury (Akiyama et al., 2000; Craft et al., 2005; Emsley et al., 2005; Hu et al., 2005; Perry et al., 2007). Therapies, which selectively target glial activation following acute brain injury in childhood, may serve to prevent neurologic disorders in adulthood. These findings raise the possibility that interventions after early-life seizures with therapies that modulate the acute microglial activation and proinflammatory cytokine response may reduce the long-term neurologic sequelae and increased vulnerability to seizures in adulthood.
(Please note, the agent used in the above studies, kainic acid, is powerful stuff, and the seizures induced were status epileptcus, a big deal and a lot different than febrile seizures. That doesn’t mean that febrile seizures are without effect, I don’t think we are nearly clever enough to understand that question with the level of detail that is needed, but they are qualitatively different and not to be confused.)
The idea of modulating glial function as a preventative measure seems especially salient to the autism community alongside the recent (totally great) bone marrow studies observing benefits to a Rett model and an early life immune activation model of neurodevelopment.
A lot of kids with autism go on to develop epilepsy in adolescence, with some studies finding prevalence in the range of 30%, which terrifies the shit out of me. Is a primed microglial phenotype, a sensitization and increased susceptibility to seizures one of the mechanisms that drive this finding?
After Somera-Molina, I started noticing a growing mention of glial priming as a possible explanation for altered neuroimmune mechanics in a lot of places. Much of the early life immune literature has sections on glial priming, Early-Life Programming of Later-Life Brain and Behavior: A Critical Role for the Immune System (full / highly recommended / Staci Bilbo!) is a nice review of 2010 data that includes this:
However, there is increasing support for the concept of “glial priming”, in which cells can become sensitized by an insult, challenge, or injury, such that subsequent responses to a challenge are exaggerated (Perry et al., 2003). For instance, a systemic inflammatory challenge in an animal with a chronic neurodegenerative disease leads to exaggerated brain inflammation compared to a control animal (Combrinck et al., 2002). The morphology of primed glial cells is similar to that of “activated” cells (e.g. amoeboid, phagocytic), but primed glial cells do not chronically produce cytokines and other pro-inflammatory mediators typical of cells in an activated state. Upon challenge, however, such as infection or injury in the periphery, these primed cells will over-produce cytokines within the brain compared to cells that were not previously primed or sensitized (Perry et al., 2002). This overproduction may then lead to cognitive and/or other impairments (Cunningham et al., 2005; Frank et al., 2006; Godbout et al., 2005).
Other studies included increased effects of pesticide exposure following immune challenge, Inflammatory priming of the substantia nigra influences the impact of later paraquat exposure: Neuroimmune sensitization of neurodegeneration, which includes, “These data suggest that inflammatory priming may influence DA neuronal sensitivity to subsequent environmental toxins by modulating the state of glial and immune factors, and these findings may be important for neurodegenerative conditions, such as Parkinson’s disease (PD).” Stress was also found to serve as a priming agent in Glucocorticoids mediate stress-induced priming of microglial pro-inflammatory responses, which studied the effect of stress mediated chemicals on inflammatory challenges; the authors get bonus points for using glucocorticoid receptor agonists and surgical procedures to eliminate glucocorticoid creation to observe a priming effect of stress on neuroimmune response.
Here is a terrifying but increasingly unsurprising study on how neonatal experience modifies the physical experience of pain in adulthood, recently published in Brain, Priming of adult pain responses by neonatal pain experience: maintenance by central neuroimmune activity
Adult brain connectivity is shaped by the balance of sensory inputs in early life. In the case of pain pathways, it is less clear whether nociceptive inputs in infancy can have a lasting influence upon central pain processing and adult pain sensitivity. Here, we show that adult pain responses in the rat are ‘primed’ by tissue injury in the neonatal period. Rats that experience hind-paw incision injury at 3 days of age, display an increased magnitude and duration of hyperalgesia following incision in adulthood when compared with those with no early life pain experience. This priming of spinal reflex sensitivity was measured by both reductions in behavioural withdrawal thresholds and increased flexor muscle electromyographic responses to graded suprathreshold hind-paw stimuli in the 4 weeks following adult incision. Prior neonatal injury also ‘primed’ the spinal microglial response to adult injury, resulting in an increased intensity, spatial distribution and duration of ionized calcium-binding adaptor molecule-1-positive microglial reactivity in the dorsal horn. Intrathecal minocycline at the time of adult injury selectively prevented both the hyperalgesia and early microglial reactivity associated with prior neonatal injury. The enhanced neuroimmune response seen in neonatally primed animals could also be demonstrated in the absence of peripheral tissue injury by direct electrical stimulation of tibial nerve fibres, confirming that centrally mediated mechanisms contribute to these long-term effects. These data suggest that early life injury may predispose individuals to enhanced sensitivity to painful events.
One of the primal drivers of behavior in any animal, pain, can be persistently modified at a molecular level! Have you ever known someone that seemed to have a higher pain tolerance than you? Maybe they did, and the training of their microglia (or yours) in early life might be why. The most basic physiologic responses can be organized through the crucible of early life events sensitizing microglia to the future environment. Multi hit wow!
The effect that befalls us all, getting older, has a ton of studies on the effect of aging on glial priming, with greatest, err, ‘hits’ including Immune and behavioral consequences of microglial reactivity in the aged brain, Aging, microglial cell priming, and the discordant central inflammatory response to signals from the peripheral immune system (full),Immune and behavioral consequences of microglial reactivity in the aged brain (full), and the autism implication heavy Microglia of the Aged Brain: Primed to be Activated and Resistant to Regulation, and others. Broadly, these studies spoke of the same pattern, a primed neuroimmune response, except in this instance, the “hits” that predisposed towards altered microglial reactivity weren’t a vigorous insult during development, but just the hum drum activity of growing older. It wasn’t a hit so much, more like a then gentle force of a relentless tide, but the functional effect on microglia response was largely similar, responses to stimuli were changed and programming was observed. I do not believe that the underlying instrument of change in age related priming is understood, but the thought occurs to me that it could simply be an exhaustion effect; a lifetime of exposure to inflammatory cytokines gradually changes the microglial phenotype.
So what about autism?
First and foremost, it provides us a line of insight into the likelyhood of a causal relationship between an altered neuroimmune milieu and autism (or nearly any other neurological disorder); that is, the question of whether or not our continued and repeated findings of altered neuroimmune parameters in the autism population represent a participating force in autism, as opposed to an artifact, a function of something else, which is also causing autism, or perhaps a result of having autism. While these are still possible explanations, the findings of glial priming provide additional detail on available mechanisms to affect brain activity and behavior through neuroimmune modifications alone.
If nothing else, we now know that we need not rely on models with no underlying substrate except the lamentations of ‘correlation does not equal causation’ and the brash faith of another, as of yet undefined, explanation. These models tell us that immune mediated pathologies can be created (and removed!) in very well established animal models of behavioral disturbances with corollaries to autism findings.
For more direct links to autism, we can look at the autism immune biomarker data set and find evidence of primed peripheral (i.e., outside the CNS) programming, literal examples where the autism population responds with a different pattern than the control group including an increased response to some pathogen type agonists, increased immune response following exposure to pollutants, of even dietary proteins.
The pattern we see of an altered microglial phenotype in the autism population, a state of chronic activity, is certainly consistent with disturbed developmental programming; it does not seem unlikely to me that a priming effect is also present, the initial prime seems to be responsible for the programming. As far as I know, there are no studies that have directly attempted to evaluate for a primed phenotype in the microglia of the autism population; I’d be happy to be corrected on this point.
Thinking about the possibility of increased microglial responsiveness and possible cognitive effects of a sustained neuroimmune toggling got me wondering if this is one of the mechanisms of a change in behavior following sickness? Or, alternatively, for some of us, “Is This Why My Child Goes Goddamn Insane And Stims Like Crazy For A Week After He Gets Sick?”
If we look to a lot of the studies that have shown a priming effect, they share a common causative pathway as some cases of autism, an early life immune insult. For some examples, the interested reader could check out Neonatal programming of the rat neuroimmune response: stimulus specifc changes elicited by bacterial and viral mimetics (full paper), Modulation of immune cell function by an early life experience, or the often mentioned Postnatal Inflammation Increases Seizure Susceptibility in Adult Rats (full paper). If there are some cases of autism that have an early life immune insult as a participating input, it is very likely a primed microglia phenotype is also present.
The studies on aging are bothering me, not only am I getting older, but the findings suggest that a priming need not necessarily mandate a distinct ‘hit’, it can be more like a persistent nudge. Our fetuses and infants develop in an environment with an unprecedented number of different nudges in the past few decades as we have replaced infection with inflammation. Acknowledging this reality, however, raises the troubling thought that our embrace of lifestyles associated with increased inflammation has reached a tipping point that we are literally training the microglia of our children to act and react differently; we aren’t waiting a lifetime to expose our fetuses and infants to environments of increased inflammation, we are getting started from the get go.
Even with all of that, however, there is a genuinely microscopic Google footprint if you search for “autism ‘glial priming’”. So, either I’m seeing phantoms (very possible), or the rest of the autism research community hasn’t caught on yet, at least in such a way that Google is notified.
Even if I am chasing phantoms, there is evidence of a widespread lack of understanding of the depth of the neuroimmune/behavioral crosstalk literature, even by the people who should be paying the most attention. This was brought to my attention by a post at Paul Patterson’s blog, where Tom Insel was quoted as finding the recent Patterson and Derecki findings ‘unexpected’.
A bone marrow transplant, which replaces the immune system, corrected both the immune response and the behavior. This finding, which was unexpected, is surprisingly similar to another recent paper reporting disappearance of the symptoms of Rett syndrome in mice following a bone marrow transplant.
Keep in mind, this is from the guy who is the head of the IACC! I can tell you one thing; while the studies were impressive, I don’t think that the findings were especially unexpected. The researchers took the time to give mice bone marrow transplants, and in Wild-type microglia arrest pathology in a mouse model of Rett syndrome, the authors utilized a variety of knockout mice and even partial body irradiation to illuminate the question of neuroimmune participation in disorder. This work was not initiated in a vacuum, they did not throw a dart at a barn door sized diagram of study methodologies and land on ‘bone marrow transplants with subsequent analysis of microglia population properties and behaviors, accounting for different exposure timeframes, radiation techniques, and genotypes’. These were efforts that had a lot of supporting literature in place to justify the expense and researcher time. [I really want to find time to blog both of those papers in detail, but for the record, I did feel the rescue effects are particularly nice touches.]
So given that the head of the IAAC was surprised to find that immune system replacement having an effect on behavior was ‘surprising’, I’m not all together shocked at the relative lack of links on ‘glial priming’ and autism, but I don’t think it will stay that way for too much longer. As more experiments demonstrating a primed phenotype start stacking up, we are going to have to find a way to understand if generation autism exhibits a primed glial phenotype. I don’t think we are going to like the answer to that question very much, and the questions that come afterwards are going to get very, very inconvenient.
Spelling it out a bit more, with bonus speculation, we should remember our recent findings of the critical role microglia are playing in shaping the neural network; our microglia are supposed to be helping form the physical contours of the brain, a once in a lifetime optimization of synaptic structures that has heavy investment from fetushood to toddlerhood. Unfortunately, it appears that microglia perform this maintenance while in a resting state, i.e., not when they have been alerted of an immune response and taken on a morphology consistent with an ‘activated state’. An altered microglia morphology can be instigated during infection, or perceived infection and consequent immune response. For examples of peripheral immune challenges changing microglial morphology, the neuroimmune environment and behavior some examples include: Peripheral innate immune challenge exaggerated microglia activation, increased the number of inflammatory CNS macrophages, and prolonged social withdrawal in socially defeated mice, Exaggerated neuroinflammation and sickness behavior in aged mice following activation of the peripheral innate immune system, or Long-term changes of spine dynamics and microglia after transient peripheral immune response triggered by LPS in vivo.
But what if we have a susceptible population, a population sensitized such that the effects of an immune challenge would result in an exaggerated and extended microglial response, effectively increasing the length of time the microglia would be ‘not resting’. What might be the changes in this population in response to a series of ‘hits’?
It does not seem to be a large logical leap to assume that if some of the altered brain physiology in autism is due to abnormal microglia function during the period of robust synaptic pruning, triggering the microglia to leave their resting state for an extended period in response could be a reasonable participant. Think of it as an exaggerated loss of opportunity effect, essentially a longer timeframe during which the microglia are not performing synaptic upkeep when compared to the microglia in an individual that is not sensitized. While our brains do show a lot of ability to ‘heal’, that does not mean that all things or times are created equally; there are some very distinct examples of time and spatially dependent neurochemical environments during early synapse development, environments that change as time goes on; i.e., Dynamic gene and protein expression patterns of the autism-associated met receptor tyrosine kinase in the developing mouse forebrain (full paper), or A new synaptic player leading to autism risk: Met receptor tyrosine kinase. In other words, recovering from a delay in microglial participation in synaptic pruning during development may not be as simple as ‘catching up’; if the right chemical environment isn’t available when the microglia get done responding, you might not be able to restart like a game of solitaire. The Met levels might be different, the neurexin levels might be different, a thousand other chemical rally points could be set that much of a nudge differently; in a system dependent on so many moving variables being just so, an opportunity missed is an opportunity lost. For good.
While the effects of a series of challenges and consequent obstructions of synaptic maintenance might not be acutely clear, I am becoming less and less convinced of the ‘safety’ of an observed lack of immediately obvious effects. I think that an intellectually honest evaluation of our recent ‘discoveries’ in many areas of early life disturbances (i.e., antibiotics and IDB risk, C-section and obesity risk, birth weight and cardiovascular risk) tell us that subtle changes are still changes, and many rise to the level of a low penetrant, environmentally induced effect once we get clever enough to ask the right question. And boy are we a bunch of dummies.
Taking all of this into consideration, all I can think is thank goodness we haven’t been artificially triggering the immune system of our infants for the past two decades while we were blissfully unaware of the realities of microglial maintenance of the brain and glial priming! What a relief that we did not rely on an assumption of lack of effect as a primary reason not to study the effect of an immune challenge. If we had done those things, we might start kicking ourselves when we realized out that our actions could be affecting susceptible subsets of children who were predisposed to reacting in difficult to measure but real ways that could literally affect the physical structure of their brains.
Oops.
– pD
Piling Up Small Changes, The Selective Skepticism Of Replacing One Fallacy With Another, And The Seductive Lure of Hubris Versus The Dispassionate Rules Of Nature
Posted on: September 28, 2012
Hello friends –
We keep on finding things that seem to very gently alter developmental trajectory towards (or away from) an eventual diagnosis of autism; a genetic variant here or there, an environmental exposure, or one of our very many experiments in cultural engineering. When these nudges are founded on genetic variances, they are often referred to as “low penetrance” risk factors; here is a snipet from the wiki definition for “Penetrance”
An allele with low penetrance will only sometimes produce the symptom or trait with which it has been associated at a detectable level.
I would argue, and have previously on this blog, that there isn’t a good reason that the descriptive of low penetrance should be relegated solely to genetic inputs. The ‘non-genetic’ factors we seem to have associated with autism risk, or protection, seem to inherit the same quality of a low grade impact; the risk of an autism diagnosis isn’t altered by too much, but instead, just a little.
There are a great number of examples of environmental impacts that seem to follow a low penetrance model of effect; maternal obesity, paternal age, cesarean section, maternal asthma, maternal folate ingestion [protective!], maternal use of anti-depressants (or being depressed?), low birth weight, and some perhaps some drugs given during pregnancy.
[Please, please note: I’m not “blaming the mother” here, but we do not have the luxury of invoking Bettleheim as a mechanism for avoiding evident truths. A dispassionate analysis of the data mandates we accept that the prenatal environment is critical.
If you think that some percentage of the autism ‘epidemic’ is real, you should realize that this issue is too important to be scuttled by emotional hotspots. You cannot blame yourself for things that were unknown to you during your pregnancy. If, instead, you don’t think autism rates have changed, none of the above impacts can be meaningful. Finally, if you believe that autism is more gift than disorder, then you aren’t getting blamed for anything anyways.]
Unfortunately, a mixture of subtle changes makes for a messy situation for our researchers for a few reasons; environmental studies contain a difficult to contend with set of confounders; knowing what to measure, when to measure it, and the often times necessary evil of usage of self reporting, computer models, or other proxies for exposure measurements. Making things even worse, it is biologically plausible, indeed, mandatory, that low penetrant effects operate with each other. What we will eventually need to be working on, for example, is determining the specific genetic dispositions that act in concert with a low birth weight and with gestational anti-depressant exposure to perturb neurodevelopment toward autism. That’s a tough thing to do.
Throwing this kind of disparate data into a blender at study time looks to be largely beyond our current capacities; researchers are struggling to identify single gene-environment interactions, for example, MET-C/pollutants, or the terrifying notion of RORA demythlation/endocrine disruptors interacting together. Looking at more, or a handful, as is likely necessary, is a long ways off.
I’ve been thinking about the intersection of these two things lately; our relative inability to evaluate for several, subtle, interacting forces, with the growing evidence that a great many mysterious conditions, including autism, seem to be governed by lots of small things occurring differently. I am left with the idea that are woefully unready to understand the participating factors in any particular case of autism, with similar reservations regarding our ability to know how much, if any, of the autism ‘epidemic’ is real.
A few weeks ago, there was an Op-Ed in the New York Times that speculated on the link between an in-utero environment characterized by increased inflammation and an eventual diagnosis of autism. I was largely in agreement with Moises Velasquez-Manoff on a the basic premise of his argument; especially regarding the state of the science on the immune findings in the autism realm, the use of helminths, not so much. A very widely read response by Emily Willingham accused the author of the piece of invoking a naturalist theory of the past:
Whether he means to or not, Velasquez-Manoff then echoes one of the favorite refrains of the anti-vaccine movement, that back when the world was a beautiful place of dirty, worm-infested children, clean water, 100% breastfeeding, and no television, it was a place where the immune system could do its work peacefully and with presumably Zen-like calm, weeding out the weak among us and leaving behind the strong.
I don’t think that the NYT article did anything of the sort, the author merely stated that there seem to be fewer signs of immune dysregulation and autoimmune conditions in some types of living conditions.
Then, a few weeks later, a widely publicized metadata study on organic eating came out. Again, the skeptics were ready to pummel the bruised body of the naturalistic fallacy, in this case, Stephen Novella at SBM:
Environmental claims for organic farming are complex and controversial – I will just say that such claims largely fall prey to the naturalistic and false dichotomy fallacies.
Stephen Novella’s version here is terse, but I think it is fair to say that in this context, the idea is that that if something is labeled as ‘natural’, that it then must be somehow superior to a ‘non-natural’ alternative, is a fair characterization of a naturalistic fallacy.
[The masochist could read through a few comments on that thread to see my take on the organic/non organic study; but the TL;DR version is, the study could have just as easily been titled, “Evaluations of Organic Eating Insufficiently Powered Or Designed To Know More Than The Most Primitive Endpoints, At Best”. Here is an NPR transcript where the presenter is a little more up front in that the state of the science is that health benefits have not been evaluated for.
But what I should point out here is that the studies of people were very limited. They were short-term and, like, narrowly focused. So they would look at pregnant women, for instance, and say, are pregnant women eating organic, are their children – did their children have left eczema or allergic conditions? So these are sort of narrowly focused studies. They were short-term, and there weren’t very many of them.
One of the few human studies in this metadata analysis involved a dietary intervention of one apple. What we have is a lack of evaluation, as opposed to a lack of findings, a familiar situation.]
Even so, it must be stated: The naturalistic fallacy(ies), as presented by the skeptics, and as believed by some fraction of grape-nut-eating-tarot-card-flipping people out there, is bogus. Things weren’t better way back then. Just because something is ‘natural’ doesn’t mean it is better, or without unknown consequences. Washing your hands is good, but antibiotics are also good, and work better when necessary. Breastfeeding is good, but it doesn’t keep your infant from getting cholera. Vaccines work. Modern agriculture is feeding a lot more of us than we used to be able to feed, and the hard truth be told, it is policies and habits that are leaving lots of people hungry. I don’t know if eating a organic diet is better for you or not, but I do know that I do like supermarkets.
But.
Our history is littered with the discarded arguments of people just as smart as us using rudimentary tools to understand complicated systems, declaring a lack of effect and throwing a contemptuous look over their shoulder at the rubes who long for the hilariously outdated solutions of yesteryear. We shouldn’t be concerned with the fact that the naturalistic fallacy is intellectually bankrupt; we should be concerned with the fact that our incredibly stupid species is changing our environment with reckless abandon on the assumption that we are smart enough to understand what we are doing. If the naturalistic fallacy is bad, the perfection-of-progess fallacy is almost as bad, with bonus negative points of being invoked by people who should know better.
How many examples do we need of our previous hubris until we realize that we are just barely less dumb now than we were then?
First we thought lead was safe as a pesticide, in paint, and as a gasoline additive. Then, we figured out it was only safe for paint and gasoline; then just in gasoline. Now, we know that any amount of measurable levels of lead are associated with cognitive effects. Any individual reader of this column was very likely an adult in 2002, and at that time, the state of our knowledge didn’t tell us that any amount of lead was less safe than no amount of lead. Ten goddamn years ago, the FDA thought there was a level of lead that in the bloodstream that did not affect cognitive function in children.
We have been performing increasingly optional cesarean sections for decades before starting to figure out that they are associated with adverse health effects for the lifespan. Only within the past few years have we discovered that this procedure is associated with altered microbiomes, obesity, and asthma.
We have been so successful at distributing products with based on plastic that over 90% of every human on the planet has detectable levels of component chemicals in their bloodstream. Only now that we have insured that nearly every human has been touched, we consistently find associations with metabolic and reproductive changes.
After near thirty years, the recommendations over administering Tylenol to infants was changed. In the 1980s we saw Reyes syndrome, made the association with aspirin, failed to observe any acute differences in infants given Tylenol, and pulled the trigger on global recommendation to replace aspirin with acetaminophen. It took decades before we were clever enough realize that eliminating Reyes might not have been the only thing we did, because we were too stupid to realize that effects do not have to be immediately obvious in order to have profound outcomes.
Human bodies were forged through the crucible of evolution, thousands of generations of adaptation, to be ready to start reproducing by the teens, and we have decided to start putting that process of for a decade, or two.
All of these examples are founded of the specificity of our analytical abilities, or rather a relative lack of specificity. We weren’t clever enough to understand to look for associations, so they remained invisible to us. A question never asked is never answered. Even worse, some of these are discrete events, disturbances orders of magnitude more simplistic to analyze compared to ‘eating organic’.
A lot of the skeptical sites will utilize the idea that humans are ‘pattern seekers’, especially when it comes to people reporting temporal associations with development of autistic behaviors and vaccination. I kind of like the idea of the pattern seeking human in general; the biggest pattern we seem to be seeing is the one that tells us that our current state of knowledge gives us enough information to understand what we are doing, a type of uber-pattern.
The idea that we have a decent understanding the effect of ingesting increased pesticide residue, a finding included in the organic metadata study, is a joke. The idea that we have the faintest clue of the outcomes of replacing infection with inflammation, a practice we have embraced with great enthusiasm, is a total fucking joke. We have barely bothered to look. Do not believe anyone who tells you otherwise.
This is what bothers me so much about a casual wielding of the naturalistic fallacy; it is so frequently a feint from critical questions. The discordance with reality of the naturalist fallacy has been established. It is great how much less suffering there is now, compared to then, but let’s not rest on our laurels. Am I the only one worried about how wrong we are here, now?
I don’t know if eating less pesticide is better than eating more pesticide, and I also can’t be sure that a lifestyle characterized by increased inflammation is a risk factor for developmental differences. I do know that the rules implemented by the natural world have no care for our hubris. Those same rules have violated our once pristine knowledge so dispassionately and with such regularity that I can find no pleasure in hurling the accusation of the naturalistic fallacy at anyone. Instead, the idea fills me with a sense of honorable mention at best; we are more capable than last century, last generation, last year, but we remain at the mercy of machinations which hold no regard for such incremental progress in knowledge in the face of unprecedented changes to our environment.
– pD
The Fairytale of a Static Rate of Autism Part 4: Troubling Realities Acknowledged, The Incredible Shrinking Gods of the Gaps, and Otherwise Rational People Using ‘Small’ As An Empirical Measure To Answer A Critical Question
Posted on: August 19, 2011
Hello friends –
These have been rough times for the people who are heavily invested in the kissing cousin theories of autism as a predominantly genetic disorder and the static, or near static rate of autism. The California twin study that is old news by the time I get this finished showed much different rates of genetic participation than previously believed. These findings exposed the underlying frailty of gene-based causation theories, namely that some of the most widely referenced studies in the autism literature, studies used repeatedly as a basis for the notion that autism was ‘the most highly heritable neurodevelopmental disorder’, were, in fact, relatively underpowered, and suffered from serious temporal and methodological shortcomings.
By contrast, the California study looked at two hundred twin pairs, a lot more twins than any previous study and actually performed autism diagnostics on all of the participating children, whereas other studies relied on medical records. Performing dedicated ADOS diagnosis prospectively on the children allowed the researchers to discern between autism and PDD-NOS, something that not all previous studies were not able to perform, if for no other reason than the DSM-IV wasn’t even released when several of the most often cited studies were published. This is from the Comment section of the California twin study:
The results suggest that environmental factors common to twins explain about 55% of the liability to autism. Although genetic factors also play an important role, they are of substantially lower magnitude than estimates from prior twin studies of autism. Nearly identical estimates emerged for ASD, suggesting that ASD presents the same liability spectrum as strict autism.
This is on top of the fact that there is a quiet, but growing acknowledgement of the fact that literally decades of genetic studies have failed to be able to explain more than a fraction of autism cases despite sequencing of tens of thousands of genomes. This is a very similar situation to a great number of other disorders which we thought we would cure once the human genome was decoded. [Note: That isn’t to say that we haven’t learned a lot from sequencing the genome, just that we didn’t quite get what we thought we were going to get.]
This ‘double hit’, so to speak, has reached a critical mass such that health officials are making politically shrewd, but refreshingly realistic statements, and dare I say, a sliver of common sense may be about to infiltrate the discussion about autism prevalence. For example, as pointed out by Sullivan, Tom Insel, head of the National Institute of Mental Health keeps a blog where he recently blogged ‘Autism Spring’, which included this nugget within the context of continued failure of genetic studies to explain any substantial part of autism, “It is quite possible that these heritability estimates were too high. . .” Ouch. (I would recommend the entire blog posting by Mr. Insel.)
The high heritability estimates, and implicit genetically-mediated cause of autism, are foundational pillars of the argument that autism rates have not changed over time. Though overused, or used wrongly in many instances, there is a kernel of dispassionate reality behind the statement, ‘there is no such thing as a genetic epidemic’. Without the crutch of exceedingly high heritability to rely on, the notion of a stable rate of autism loses the only hard science (read: replicable, biologically-plausible), i.e.,genetics, it ever had, and must place complete reliance on the softer sciences (read: unquantifiable, ‘greater awareness’), i.e.,sociology. This is great news if you love impossible to verify estimates of prevalence and anecdotes about crazy uncle George who would have been diagnosed with autism forty years ago. However, if you think we should be relying less on psychologists and cultural anthropologists to answer critical questions, and rely more on hard science, this means that the old narrative on autism prevalence holds even less allure than it did in the past, for those of you who thought this was possible.
Before Kid Autism came around, I would occasionally read discussion boards on the creationism versus evolution ‘debate’. One thing that I noticed was that the creationists would often employ a ‘God of the Gaps’-style argument: anything that couldn’t be explained by science (yet), or anything necessary to support whatever fanciful construct had been erected to protect biblical creation fables, was ascribed to the work of God. That’s one thing you have to give to God, he (or she!) can handle it all; it didn’t matter what primitive logical test biblical creation was failing to pass, the golden parachute clause was always that God could have just made things that way. It was a nifty out on the part of the creationists, kind of like a get out of jail free card. The autism prevalence discussion has been working just like this, and the funny part is that the people that are always claiming to have the intellectual high ground, the supposed skeptics, are playing the part of the creationists! Zing!
Here is how it works:
Concerned Parent: It sure does seem like there is more autism than there used to be, what with there being X in a thousand kids with it! That’s much, much more than even ten years ago! My brothers, sisters and I all knew kids with mental retardation and Down’s syndrome, but we just don’t remember kids like we see today.
Supposed Skeptic: It is diagnostic substitution and ‘greater awareness’; autism incidence has been stable. The DSM was changed which resulted in more children being labeled.
Concerned Parent: It sure does seem like there’s more autism than there used to be. Now there are Y kids in a thousand having autism! Why does my son’s preschool teacher keep insisting something is changing?
Supposed Skeptic: It is diagnostic substitution and ‘greater awareness’; autism incidence has been stable. The DSM was changed which resulted in more children being labeled.
Concerned Parent: What the hell? Now there are Z kids in a thousand having autism! When are those genetic studies going to figure autism out, anyway?
Supposed Skeptic: It is diagnostic substitution and ‘greater awareness’; autism incidence has been stable. When does the new DSM come out again?
(Replace X/Y/Z with any progressively larger numbers.)
It doesn’t matter what prevalence number is thrown about–even the astronomical one in thirty-eight figure bandied about for South Korean children didn’t cause so much as a raised eyebrow; the autism equivalent of God of the Gaps, greater awareness and loosening of diagnostic criteria can handle any amount of increase gracefully. It is the equivalent of an uber-absorbent autism paper towel, capable of soaking up any number of new children with a diagnosis; there is, literally, no amount of an increase that the God of the Gaps can’t handle.
If, instead the question was posed like this, ‘How much of the apparent increase in autism is real?’, the answer was always, ‘Zero’, regardless of what the current rates of autism were when you asked the question
Then a funny thing happened, a series of studies from several researchers showed a consistent trend of older parents giving rise to more children with autism than younger parents. There were differences between the studies on just how much of an effect an older parent had, but the overall direction of association was clear. In this instance, there was also the luxury of a plausible biological mechanism that involved the mediator in favor, genetics. The idea is that advancing age in the parent meant more years for gametes to get knocked by a random cosmic zap or other environmental nastygram and this disturbance created genetic problems down the line for the offspring, a theory I think is probably pretty good. Once a couple of these studies started to pile up, there was a small shift in the narrative regarding autism prevalence; after all, nobody could bother to try to deny that parents were getting older compared to past generations. Here is how it looked:
Concerned Parent: What the hell? Now there are X kids in a thousand having autism!
Supposed Skeptic: Greater awareness and diagnostic substitution are primarily responsible for our observations of increased autism, although, ‘a real, small increase’ cannot be ruled out.
And with that, there was a little less autism prevalence for the God of the Gaps to handle. It never seemed to bother anyone that implicit in this argument is an impossible to quantify concept ‘small increase’. If you were to ask someone what rate of autism ‘a small increase’ amounted to with more precision, the answer is whatever amount rises to the level of autism minus the difficult to quantify effect of older parents. That is some lazy stuff.
Here are some examples of prominent online skeptics discussing the possibility of a true rise in autism. See if you can detect a pattern.
Here is Stephen Novella pushing The Fairytale in 2009:
While a real small increase cannot be ruled out by the data, the observed increase in diagnostic rates can be explained based upon increased surveillance and a broadening of the definition – in fact autism is now referred to as autism spectrum disorder.
[Here we see the notion that everything can be explained by the God of the Gaps.]
Here is an example of Orac toying around with this filibuster just the other day, in August of 2011:
True, the studies aren’t so bulletproof that they don’t completely rule out a small real increase in autism/ASD prevalence, but they do pretty authoritatively close the door on their being an autism “epidemic.”
These aren’t the only examples, far from it. Check it out:
It should be noted that the data cannot rule out a small true increase in autism prevalence. (Stephen Novella in 2008)
If the true prevalence rate of autism and ASDs has increased, it has not increased by very much. (David Gorski, 2010)
We should have the curiosity to wonder, what, exactly, does small mean in these contexts? What percentage size increase should we consider small enough to hide within the data? Five percent? Ten percent? What does ‘small’ mean, numerically, within a range? Is a ten to twenty percent rise in autism rates reason for us to take comfort in the fact that the effect of greater awareness is real? At what level does the percentage of ‘real’ autism increase mandate more than superficial lip service, more than a paragraph about ‘gene-environment interactions’ at the end of a two-thousand word blog post that takes pride in the intellectual chops of outthinking Jenny McCarthy? You won’t get anyone to answer this question; they can’t, because they don’t really know what they mean when they say, ‘small’, other than, ‘it can’t be vaccination’.
How do we know the amount of this increase must, in fact, even be ‘small’? This becomes especially problematic when we consider the smackdown that the canard of autism as ‘among the most heritable neurological conditions’ has taken as of late. If the high heritability estimates of autism are incorrect, yet so often repeated as gospel, why should we also assign confidence to the idea that the increase is trivial? Isn’t one argument the foundation of the other? Did either really have quality data behind them?
This is a terrible, awful, horrible, completely fucking idiotic way to address a question as important as whether or not a generation of children is fundamentally different. We cannot afford the ramifications of being wrong on this, but we seem to find ourselves in an epidemic of otherwise intelligent people willing to accept the pontifications of cultural anthropologists and the feebleness of social scientists on this critical question. I am not arguing against the realities of diagnostic switching and greater awareness affecting autism diagnosis rates. But we can understand that while they are a factor, we must also admit that we have little more than a rudimentary understanding of these impacts, and when we consider the implications of being incorrect, the potential disaster of a very real, not ‘small’ increase in the number of children with autism, we shouldn’t be overselling our knowledge for the sake of expedient arrival at a comforting conclusion. We should be doing the opposite.
If we can’t have the robustly defendable values on autism rates right now, that’s fine, because that is the reality, but we should at least have the courage to acknowledge this truth. This is the nature of still learning about something, which we are obviously doing in terms of autism, but in that situation, we don’t have the currency of scientific debate, decent data, to be saying with authority that any true increase in autism is small.
Unfortunately for the purveyors of The Fairytale, things are going to get a lot worse. The problem is that we are starting to identify extremely common, in some cases, recently more common, environmental influences that subtly increase the risk of autism. These are further problems for a genetic dominant model and effectively mandate that the ‘small increase’ is going to have to start getting bigger as a measurement, with a correlated decrease in the amount of autism that cultural shuffling can be held responsible for. Will anyone notice?
By way of example, we now have several studies that link the seasons of gestation with neurodevelopmental disorders including autism and schizophrenia; i.e., Season of birth in Danish children with language disorder born in the 1958-1976 period, Month of conception and risk of autism, or Variation in season of birth in singleton and multiple births concordant for autism spectrum disorders, which includes in the abstract, “The presence of seasonal trends in ASD singletons and concordant multiple births suggests a role for non-heritable factors operating during the pre- or perinatal period, even among cases with a genetic susceptibility.” Right! As I looked up some of these titles, I found that the evidence for this type of relationship has been well known for a long time; schizophrenia, in particular has a lot of studies in this regard, i.e., Seasonality of births in schizophrenia and bipolar disorder: a review of the literature, which is a review of over 250 studies that show an effect, and I also found Birth seasonality in developmentally disabled children, which includes children with autism and was published in 1989, which is like 1889 in autism research years.
Our seasons have remained constant (but probably won’t stay too constant for much longer. . . ), but this still throws a whole barrel of monkey wrenches into the meme of a disorder primarily mediated through genetics.
More damning for the Fairytale are some studies presented at this year’s IMFAR, and some others just published, that tell us that abnormal immune profiles during pregnancy appear to provide slightly increased risk for autism, roughly doubling the chance of a child receiving a diagnosis. The groovy part is that the studies utilized both direct and indirect measurements of an activated immune system to draw similar conclusions, a sort of biomarker / phenotype crossfire.
From the direct measurement end, we have Cytokine Levels In Amniotic Fluid : a Marker of Maternal Immune Activation In Autism?, which reports that mothers with the highest decile of tnf-alpha levels in the amniotic fluid had about a one and a half times increased risk for autism in their children. This makes a lot of sense considering the robustness of animal models of an acute inflammatory response during pregnancy and its impact on behavior.
Another study, this one from the MIND Institute in California (which I love), is Increased mid-gestational IFN-gamma, IL-4, and IL-5 in women giving birth to a child with autism: a case-control study (full paper). They found that in pregnant mothers, increased levels of IFN-gamma led to a roughly 50% increased risk of an autism diagnosis. Here is a snipet:
The profile of elevated serum IFN-γ, IL-4 and IL-5 was more common in women who gave birth to a child subsequently diagnosed with ASD. An alternative profile of increased IL-2, IL-4 and IL-6 was more common for women who gave birth to a child subsequently diagnosed with DD without autism.
This study took a lot of measurements, and goes to great lengths to explicitly call for additional analysis into the phenomena. IFN-gamma is typically considered pro-inflammatory, while IL-4 and IL-5 are considered regulatory cytokines. In order to determine if these findings were chance or not, the researchers determined if there was a correlation between the levels of IFN-gamma, IL-4, and IL-5, which they reported with very robust results. Less clear is what might be causing these profiles, or how, precisely, they might give rise to an increased risk of autism. The interconnectedness of the brain and the immune systemwould be a good place to start looking for an answer to the last question though.
What about indirect measurements? It just so happens, another paper was published at IMFAR this year that observed the flip side of the coin, conditions associated with altered cytokine profiles in the mother and this study also found an increased risk of autism. The Role of Maternal Diabetes and Related Conditions In Autism and Other Developmental Delays, studied a thousand children and the presence of diabetes, hypertension, and obesity in their mothers in regards to the risk of a childhood autism diagnosis. The findings indicate that having a mother with one or more of those conditions roughly doubles the chances of autism in the offspring. Obesity, in particular, has an intriguing animal model Enduring consequences of maternal obesity for brain inflammation and behavior of offspring, a crazy study that I blogged about when it was published. A variety of auto immune disorders in the parents have been associated with an autism diagnosis in several studies.
The obesity data is particularly troublesome for the idea of a ‘small’ increase in autism, just like parents have been getting older, parents have also been getting fatter, waaaay fatter, (and more likely to have diabetes) the last few decades. There isn’t any squirming out of these facts. If, indeed, being obese or carrying associated metabolic profiles is associated with an increased risk of autism, ‘small’ is getting ready to absorb a big chunk of real increase. But is there any clinical data to support this possible relationship, do we have any way to link obesity data with this autism data from the perspective of harder figures?
It further turns out, there are some very simple to navigate logical jumps between the above studies. Remembering that our clinical measurements indicated that increased INF-gamma, IL-4, and IL-5 from the plasma of the mothers was associated with increased risk, we can see very similar patterns in Increased levels of both Th1 and Th2 cytokines in subjects with metabolic syndrome (CURES-103). Here is part of the abstract, with my emphasis.
Metabolic syndrome (MS) is a cluster of metabolic abnormalities associated with obesity, insulin resistance (IR), dyslipidemia, and hypertension in which inflammation plays an important role. Few studies have addressed the role played by T cell-derived cytokines in MS. The aim of the tudy was to look at the T-helper (Th) 1 (interleukin [IL]-12, IL-2, and interferon-gamma [IFN-gamma]) and Th2 (IL-4, IL-5, and IL-13) cytokines in MS in the high-risk Asian Indian population.
Both Th1 and Th2 cytokines showed up-regulation in MS. IL-12 (5.40 pg/mL in MS vs. 3.24 pg/mL in non-MS; P < 0.01), IFN-gamma (6.8 pg/mL in MS vs. 4.7 pg/mL in non-MS; P < 0.05), IL-4 (0.61 pg/mL in MS vs. 0.34 pg/mL in non-MS; P < 0.001), IL-5 (4.39 pg/mL in MS vs. 2.36 pg/mL in non-MS; P < 0.001), and IL-13 (3.42 pg in MS vs. 2.72 pg/mL in non-MS; P < 0.01) were significantly increased in subjects with MS compared with those without. Both Th1 and Th2 cytokines showed a significant association with fasting plasma glucose level even after adjusting for age and gender. The Th1 and Th2 cytokines also showed a negative association with adiponectin and a positive association with the homeostasis model of assessment of IR and high-sensitivity C-reactive protein.
Check that shit out! Seriously, check that out; increased IFN-gamma, IL-4, and IL-5 in the ‘metabolic syndrome’ group, comprised of people with, among other things, obesity, insulin resistance, and hypertension; the same increased cytokines and risk factors found to increase the risk of autism.
If we look to studies that have measured for TNF-alpha in the amniotic fluid during pregnancy, we quickly find, Second-trimester amniotic fluid proinflammatory cytokine levels in normal and overweight women
There were significant differences in amniotic fluid CRP and TNF-alpha levels among the studied groups: CRP, 0.018 (+/-0.010), 0.019 (+/-0.013), and 0.035 (+/-0.028) mg/dL (P=.007); and TNF-alpha, 3.98 (+/-1.63), 3.53 (+/-1.38), and 5.46 (+/-1.69) pg/mL (P=.003), for lean, overweight, and obese women, respectively. Both proinflammatory mediators increased in women with obesity compared with both overweight and normal women (P=.01 and P=.008 for CRP; P=.003 and P=.01 for TNF-alpha, respectively). There were significant correlations between maternal BMI and amniotic fluid CRP (r=0.396; P=.001), TNF-alpha (r=0.357; P=.003) and resistin (r=0.353; P=.003).
Nice.
What we are really looking at are five studies the findings of which speak directly to one another; a link to metabolic syndrome during pregnancy and increased IFN-gamma, IL-4, and IL-5, a link to obesity and hypertension in pregnant mothers and autism risk, and an increased risk of autism in mothers wherein IFN-gamma, IL-4, and IL-5 were found to be increased outside of placenta. Further, we have a link between amniotic fluid levels of TNF-alpha and metabolic syndrome, metabolic syndrome in mothers and autism risk, and increased risk from increased tnf-alpha in the amniotic fluid.
As I have said previously, one thing that I have learned during this journey is that when we look at a problem in different ways and see the same thing, it speaks well towards validity of the observations. What we see above is a tough set of data to overcome; we need several types of studies looking at the relationship between metabolic syndrome, immune profiles during pregnancy, and autism from different angles to have reached the same wrong conclusion, something that is increasingly unlikely. We are in an epidemic of obesity and the associated endocrine mish mash of metabolic syndrome, there simply isn’t any diagnostic fuzziness on this. It is happening all around us. Even though the total increase in risk is relatively small, the sheer quantity of people experiencing this condition of risk mandates that the numbers game looks favorable towards a real increase in autism. If we acknowledge this, how can we continue to have faith in the concept that any true increase in the autism rates must be ‘small’?
Is the next argument going to be that besides increased parental age, and heavier or more diabetic mothers, the rest of the autism increase is the result of diagnostic three card monte? (Just how much is the rest, anyways?)
And even though these studies, and likely more in the future, expose the crystal delicate backbone of the ‘small true increase’ argument, I have great pessimism that the people so enamored with invoking this phrase will ever acknowledge its shifting size, much less the implications of being wrong on such a grand scale.
– pD
Elegant Observations, Feedback Loops Within Entangled Systems, and Possible Clues On Why So Many People With Autism Are Male – “Sex Hormones in Autism: Androgens and Estrogens Differentially and Reciprocally Regulate RORA, a Novel Candidate Gene for Autism”
Posted on: March 8, 2011
- In: Autism | Beautiful Complexity | Epigenetics | Epigenome | Feedback Loops | Genetics | Hu | RORA
- 2 Comments
Hello friends –
Whatever your take on the predominant cause of autism(s), your thoughts on the appropriateness, or inappropriateness of research allocations in the autism realm, one thing can’t be denied; collectively, a lot of researcher time and dollars have been poured into autism research. We’ve learned some important things, some cool things, some confusing things, some obscure things, and some useless things. But even with all of the resources we have applied towards understanding autism, one of the most curious unknowns is also one obvious to the most rudimentary observations, the persistently skewed male to female ratio, with a finding of three or four males to every female with autism. It is one of the most vexing questions, almost taunting us with seeming obviousness, but consistently elusive. It isn’t just autism, lots of other neurological conditions are similarly tilted, and in a bazillion animal models it seems an unfortunate fact that being born male simple predisposes you (or rat you) to a variety of things you’d rather not have.
The ideas I have seen floated most frequently to explain this observation involve the effect of prenatal testosterone and the associated ‘extreme male brain’ theory, a loss of genetic backups to compensate for mixups, and synergistic effects of testosterone on chemicals, notably, mercury; explanations which I generally like a little, a little less, and almost not at all, respectively. An idea I’ve floated a couple of times, but that seemed even less accepted (or, more likely, completely unnoticed), was that estrogen might be acting in a protective manner; as it is known to exhibit attenuate the effects of neuroinflammation and oxidative stress. Indeed, it is starting to look like estrogen receptors are expressed in a large variety of situations salient to CNS processes.
A few weeks ago, the same group that has published some immensely cool studies on epigenetics and brain proteins, genetic expression differentials within twin siblings with autism, and circadian rhythm alterations comes a paper which may give us insight into this question, so that is pretty exciting. Even cooler, it invokes a negative feedback loop in a complicated system and is built upon the foundation of several earlier studies on a protein implicated in lots of things we know are awry in autism. That study is Sex Hormones in Autism: Androgens and Estrogens Differentially and Reciprocally Regulate RORA, a Novel Candidate Gene for Autism (full paper). Here is the abstract:
Autism, a pervasive neurodevelopmental disorder manifested by deficits in social behavior and interpersonal communication, and by stereotyped, repetitive behaviors, is inexplicably biased towards males by a ratio of ~4:1, with no clear understanding of whether or how the sex hormones may play a role in autism susceptibility. Here, we show that male and female hormones differentially regulate the expression of a novel autism candidate gene, retinoic acid-related orphan receptor-alpha (RORA) in a neuronal cell line, SH-SY5Y. In addition, we demonstrate that RORA transcriptionally regulates aromatase, an enzyme that converts testosterone to estrogen. We further show that aromatase protein is significantly reduced in the frontal cortex of autistic subjects relative to sex- and age-matched controls, and is strongly correlated with RORA protein levels in the brain. These results indicate that RORA has the potential to be under both negative and positive feedback regulation by male and female hormones, respectively, through one of its transcriptional targets, aromatase, and further suggest a mechanism for introducing sex bias in autism.
The press release and google news cycle for this paper seemed to have been well ahead of the pubmed robot; Kev had a post on this study a few weeks before it hit pubmed with a postdate. I generally skip out on the interest story/vaccine fairytale story/vaccine nightmare story/lost child nightmare story that is the google news autism feed, but in this case, it harbored a story on a paper that I was actually interested in. In any situation, the paper landed in pubmed this morning, and is available in full via PLOS, so great stuff is available to us all.
The paper starts with some of the backstory, the ‘inexplicable’ male predominance in autism, some of the theories on why this might be the case, and most importantly, details on previous findings by this set of researchers on reduced levels of RORA in the CNS of people with autism, a protein with a great number of functions of interest to the autism community.
Together, these results link molecular changes in RORA in peripheral cells to molecular pathology in the brain of autistic individuals. These findings are particularly relevant to ASD as RORA is involved in several key processes negatively impacted in autism, including Purkinje cell differentiation, cerebellar development, protection of neurons against oxidative stress, suppression of inflammation, and regulation of circadian rhythm. Behavioral studies on the RORA-deficient staggerer (RORA+/sg) mouse, primarily used as a model to study ataxia and dystonia[13], further show that RORA is also associated with restricted behaviors reminiscent of ASD, such as perseverative tendencies, limited maze patrolling, anomalous object exploration as well as deficits in spatial learning.
It’s tough to find a protein with a greater key word hitlist for our population of interest than Purkinje cell differentiation, cerebellar development, protection from oxidative stress, suppression of inflammation, and regulation of the circadian rhythm. In fact, I’d be shocked to find a protein touching so many fracture points that wasn’t found altered in the autism population; it makes too much sense within the framework of an entangled system and what we already know about the physiology of autism. Remember that the previous paper found decreased RORA in the brain of people with autism; i.e., less of a protein that protects from oxidative stress, supports Purkinje cell development, and suppresses the inflammatory response. A relative lack of RORA makes a depressingly good amount of sense.
That being said, what makes the current paper so interesting is that they found the RORA is differentially, and inversely modulated by female and male hormones (i.e., testosterone and estrogen). But even more insidiously, one of the downstream products regulated by RORA, aromatase, participates in the cleavage of testosterone to estrogen; the authors essentially describe a negative feedback loop. It turns out, not only is RORA decreased in the CNS of autism, but so too is aromatase.
We also show that one of the transcriptional targets of RORA is aromatase, which is a crucial enzyme in the biosynthesis of estrogen from testosterone. It is noteworthy that both RORA and aromatase proteins are decreased in the frontal cortex of autistic subjects, and that the level of aromatase protein is strongly correlated with the level of RORA protein in the brain tissues. We therefore propose that the reduction of RORA observed in autism is exacerbated by a negative feedback mechanism involving decreased aromatase level, which further causes accumulation of its substrate, testosterone, and reduction of its product, estradiol. Testosterone and estradiol respectively exhibit negative and positive feedback regulation of RORA expression as illustrated in Fig. 5, which summarizes the principal findings of this study. Thus, a deficiency in RORA in autistic brain is expected to be further aggravated by increased levels of testosterone due to suppression of aromatase, a transcriptional target of RORA.
This is pretty neat; it shows how simply being male can lead to the downregulation of a system with tendrils attached to a great number downstream processes we know to be disturbed in autism.
I particularly liked that this paper established a chain of learning more, something I think we can all agree is a great idea. Some of the people on this study have been plugging away with some interesting ideas for a while, all of which, I believe, are ancestors of these findings. They had two really neat papers on genetic expression in autism twins with differential degrees of autism severity, both of which used genomic bioinformatic tools to understand which the genetic pathways were affected. This is actually rather brilliant; they essentially leveraged the genetic uniqueness of the twins to gain more insight into which processes were being affected in autism by seeing which genes were differentially expressed in identical twins that manifested differently, using genetics to learn about what is happening a layer above the genome. Next, the original RORA paper began to probe the mechanism by which the previously observed expression was achieved, they found that a particular protein, RORA, was overmethylated and consequently at depressed levels. Another bioinformatic approach told them that RORA was a particularly attractive candidate for further evaluation based on its descendant interactions, and the association between RORA, aromatase, and sexual hormones appeared. Beautiful.
All that coolness not withstanding, some of the articles I saw on this lacked the caution and nuance we ought to see with these kinds of findings; the paper was pretty clear that previous CNS studies hadn’t shown decreased RORA in all of their samples, just most of them. This doesn’t answer all of our questions about the male dominance of autism, but we do know more than when this study was published, and that is pretty cool. Hooray for knowledge.
– pD
Hello friends –
I have decidedly mixed feelings on the genetic side of autism research; clearly genetics plays a part, but it does appear that autism has largely mirrored other complicated conditions in that what we thought we were getting when we cracked the genetic code has, for all practical purposes, failed to materialize. To what extent our genetic makeup really plays a part in autism more than any other condition that is currently mystifying us, I don’t think we can say with much certainty; unless you want to count some.
To my mind, one particularly bright spot in the gene realm is the associations of the MET-C allele and an increased risk of an autism diagnosis. At first glance, MET doesn’t seem like a big deal; lots of people have the MET-C mutation, in fact, nearly half of everyone has it. But people with autism have it just a little more frequently, an observation that has been replicated many times. But what is exciting is not only that the MET-C findings are robust, but they can also affect a lot of implicated systems in autism in biologically relevant ways. From an ideological standpoint, the fissure in the autism community about research priorities regarding genetics versus environment, the MET-C studies are a superb example of just how much useful knowledge there is by starting at the genome and working upwards, and finding once we get there that the reality involves lots more than just genes. There is something for everyone!
Getting to the big picture where we can appreciate the beautiful complexity takes a little bit of digging, but it’s worth the effort.
Every now and again you’ll see a period piece about the forties, fifties or sixties, and you’ll get a glimpse of the female operator, someone who would take a call and literally connect two parties together; the gatekeeper. The operator’s actions were binary; either she connected the lines and the call went through, or she didn’t, and nothing happened. Of course, one operator couldn’t connect you to any other phone, but participated in groupings of phones with some logical or functional structure. Ultimately, the operators were the enabler of communication, physically putting two entities into contact to perform whatever business they had with each other.
Within our bodies, tyrosine kinases are enzymes responsible transferring phosphate to proteins; a chemical exchange critical towards a great number of cellular functions, and in a sense, the tyrosine kinases act as cellular operators, helping implement a physical swap of chemicals that ultimately set in motion a great number of processes. Some very rudimentary cellular functions are initiated by the tyrosine kinases; for example, cell division, which is why mutated kinases can lead to the generation of tumors; i.e., the signaling for cell division gets turned on, and never gets turned off. Inhibiting tyrosine kinases is the mechanism of action for some drugs that target cancer.
The MET gene is responsible for creating the MET receptor tyrosine kinase. This particular receptor is involved in lots of processes that are of great interest to autism; the MET receptor is expressed heavily during embryogenesis in the brain, has immune modulating capacities, and is associated with wound healing, and is particularly implicated in repair of the gastro-intestinal track.
Kinases don’t just fire away, shuttling phosphates around any old time, they must be activated by a triggering molecule, or a ligand. There is only one known ligand for the MET receptor; hepatocyte growth factor, or HGF (also sometimes referred to as HGF/SF, or hepatocyte growth factor/scatter factor). We’ll get to why we bother worrying about HGF a little later on, but it is important to keep in mind that without HGF, the functions affected by the MET-C receptor, early brain development, immune modulating, and wound repair cannot be achieved.
So what about autism, and why is it a beautiful illustration of complexity? Walking our way through the MET findings in autism is a rewarding task; it is one of the few instances I’ve seen where the glimpses of relevance gleaned from straight genetic studies have been incrementally built upon to achieve a much grander understanding of autism. This is the kind of thing that I think a lot of people who dismiss the utility of genetic studies are missing; genetics are only the first piece of the puzzle, it doesn’t only implicate genes, it tells us about the processes and the proteins disturbed in autism; and with that knowledge, we can perform targeted analysis for environmental participants.
The first clues about MET involvement with autism came in 2006, when A genetic variant that disrupts MET transcription is associated with autism (full paper) was published. The abstract is longish, but here is a snipet:
MET signaling participates in neocortical and cerebellar growth and maturation, immune function, and gastrointestinal repair, consistent with reported medical complications in some children with autism. Here, we show genetic association (P = 0.0005) of a common C allele in the promoter region of the MET gene in 204 autism families. The allelic association at this MET variant was confirmed in a replication sample of 539 autism families (P = 0.001) and in the combined sample (P = 0.000005). Multiplex families, in which more than one child has autism, exhibited the strongest allelic association (P = 0.000007).
I appreciate the pleiotropic nature of what we are seeing here, a gene that is involved with brain growth and maturation, immune function, and GI repair. The association in ‘multiplex’ (i.e., families with more than one child with autism) was very, very strong. Even still, this was a pretty short paper, and it was all genetics. Coolness factor: 3.
Neater studies were on the horizon shortly thereafter, a year later, some of the same group looked for expression of MET in post mortem brain tissue and found significantly decreased levels of MET protein in Disruption of cerebral cortex MET signaling in autism spectrum disorder.
MET protein levels were significantly decreased in ASD cases compared with control subjects. This was accompanied in ASD brains by increased messenger RNA expression for proteins involved in regulating MET signaling activity. Analyses of coexpression of MET and HGF demonstrated a positive correlation in control subjects that was disrupted in ASD cases.
This is a nice follow up; lots of times a genetic study might suggest a hit, but we really don’t even know how such a genetic change might manifest physiologically, like having a jigsaw puzzle of solid black and finding two pieces that fit together. In those instances, we can’t really go looking for different levels of the protein, so there you are. In this case, the authors found an allele worth investigating, and then went looking to see if relevant proteins were altered in the population, and in the CNS no less! Not only that, but they also looked at the initiating end of the process, the ligand, HGF, and found abnormalities. Good stuff. Unfortunately, I haven’t found myself a copy of this paper yet, but the fact that other proteins in the pathway were altered is another line of evidence that something is amiss. I’ve begun to appreciate the fact that I have spent a long time under appreciating the interconnectedness of biological systems; you aren’t going to have a disturbance in one system without altering the way upstream, and downstream processes are working; so the fact that we see other proteins, those related to MET functions, modified, makes beautiful sense. Coolness factor: 5.
Likely because of the mixed findings of skewed proteins in the MET pathway (?), the next study in line is, Genetic Evidence Implicating Multiple Genes in the MET Receptor Tyrosine Kinase Pathway in Autism Spectrum Disorder (full paper available). Here’s the abstract:
A functional promoter variant of the gene encoding the MET receptor tyrosine kinase alters SP1 and SUB1 transcription factor binding, and is associated with autism spectrum disorder (ASD). Recent analyses of postmortem cerebral cortex from ASD patients revealed altered expression of MET protein and three transcripts encoding proteins that regulate MET signaling, hepatocyte growth factor (HGF), urokinase plasminogen activator receptor (PLAUR) and plasminogen activator inhibitor-1 (SERPINE1). To address potential risk conferred by multiple genes in the MET signaling pathway, we screened all exons and 5′ promoter regions for variants in the five genes encoding proteins that regulate MET expression and activity. Identified variants were genotyped in 664 families (2,712 individuals including 1,228 with ASD) and 312 unrelated controls. Replicating our initial findings, family-based association test (FBAT) analyses demonstrated that the MET promoter variant rs1858830 C allele was associated with ASD in 101 new families (P=0.033). Two other genes in the MET signaling pathway also may confer risk. A haplotype of the SERPINE1 gene exhibited significant association. In addition, the PLAUR promoter variant rs344781 T allele was associated with ASD by both FBAT (P=0.006) and case-control analyses (P=0.007). The PLAUR promoter rs344781 relative risk was 1.93 (95% Confidence Interval [CI]: 1.12−3.31) for genotype TT and 2.42 (95% CI: 1.38−4.25) for genotype CT compared to genotype CC. Gene-gene interaction analyses suggested a significant interaction between MET and PLAUR. These data further support our hypothesis that genetic susceptibility impacting multiple components of the MET signaling pathway contributes to ASD risk.
We’ve got two new genes added to the mix, PLAUR and SERPINE. The juicy part here is that the authors didn’t look for these variants at random, but performed a targeted search; they knew that the proteins encoded by these genes interact with either MET receptor function or HGF, and they also had found altered expression of these genes in the CNS study. From the Introduction:
The hepatocyte growth factor (HGF) gene encodes the activating ligand for the MET receptor. HGF is translated as an inactive precursor protein that requires cleavage for efficient binding to the MET receptor [Lokker et al 1992]. The activating cleavage of HGF is achieved most efficiently by the enzyme plasminogen activator (urokinase-type; uPA; gene symbol: PLAU) under conditions in which uPA binds to its receptor, the urokinase plasminogen activator receptor (uPAR; gene symbol: PLAUR). Activating cleavage of HGF can be suppressed by the plasminogen activator inhibitor-1 (PAI-1; gene symbol: SERPINE1). Together, these proteins regulate the activity of MET receptor tyrosine kinase signaling, and our recent microarray analyses of postmortem temporal lobe of individuals with ASD indicate that disrupted MET signaling may be common to ASD pathophysiology [Campbell et al 2007]. For example, we found that there is increased expression of the HGF, PLAUR and SERPINE1 transcripts in ASD in postmortem cerebral cortex. The observation of disrupted expression suggests a general dysfunction of MET signaling in the cerebral cortex of individuals with ASD.
The proteins encoded by PLAUR and SERPINE were also found increased in the expression study; a finding further supported by the genetic study here. The really grand slice here is that the SERPINE protein suppresses cleavage of HGF; essentially another way MET function can be affected, from a disturbance upstream of HGF binding. In other words, more SERPINE (possibly as a result of a ‘promoter allele’) would result in less MET receptor activation because the SERPINE interferes with the cleavage of HGF, and thus, another pathway to reduced MET activation. In a finding that seems 20/20 with hindsight, a functional promoter of the SERPINE gene was found to increase autism risk; i.e., if you have more SERPINE, you get less functional HGF, and therefore less triggering of the MET receptor. This is cool and begins a portrait of the complexity; it shows how the effect of reduced MET functionality can come from multiple drivers; the reduced MET allele, or, the promoter SERPINE allele, and what’s more, having both is an even bigger risk; the authors are describing a synergy of low penetrance genes.
From the discussion section of the paper:
Beyond genetic susceptibility, the functional integrity of the MET signaling system also is sensitive to environmental factors. This concept is supported by bioinformatics analyses that identified PLAUR, SERPINE1 and HGF as genes active in immune response regulation, sensitive to environmental exposures, and within chromosomal regions previously implicated in ASD linkage studies [Herbert et al 2006]. Moreover, a recent cell biological study shows that chemically diverse toxicants reduce the expression of MET in oligodendrocyte progenitor cells, a result that is interpreted as the convergence of toxicant effects on oxidative status and the MET-regulating Fyn/c-Cbl pathway
Here are links to the Hebert paper, Autism and environmental genomics, and the Li paper, Chemically Diverse Toxicants Converge on Fyn and c-Cbl to Disrupt Precursor Cell Function. What is neat here is that we are starting to be able to see a pathway of genes, and resultant proteins, that can effect disparate systems. I believe that there is a subset of acupuncture, acupressure that relies on more knuckles than needles, and while the science on accu* based therapies isn’t very good, it does occur to me that in a sense, our lattice work of HGF-PLAUR-SERPINE proteins that participate in the MET-C process are pressure points in a delicate system, push a little bit and things will bend down the line accordingly. It also exemplifies why I am offended by highly negative attitudes on genetic studies held by people who believe in a non trivial, environmentally mediated increase in the rates of autism; we are approaching a nearly impossible to overturn reality that genes we know to be associated with autism are particularly sensitive to interference from environmental agents, and participate in immune function. That is important information. Coolness factor 8. First glimpse of beauty factor: 10.
Next up we have Dynamic gene and protein expression patterns of the autism-associated Met receptor tyrosine kinase in the developing mouse forebrain (full paper).
The establishment of appropriate neural circuitry depends upon the coordination of multiple developmental events across space and time. These events include proliferation, migration, differentiation, and survival – all of which can be mediated by hepatocyte growth factor (HGF) signaling through the Met receptor tyrosine kinase. We previously found a functional promoter variant of the MET gene to be associated with autism spectrum disorder, suggesting that forebrain circuits governing social and emotional function may be especially vulnerable to developmental disruptions in HGF/Met signaling. However, little is known about the spatiotemporal distribution of Met expression in the forebrain during the development of such circuits. To advance our understanding of the neurodevelopmental influences of Met activation, we employed complementary Western blotting, in situ hybridization and immunohistochemistry to comprehensively map Met transcript and protein expression throughout perinatal and postnatal development of the mouse forebrain. Our studies reveal complex and dynamic spatiotemporal patterns of expression during this period. Spatially, Met transcript is localized primarily to specific populations of projection neurons within the neocortex and in structures of the limbic system, including the amygdala, hippocampus and septum. Met protein appears to be principally located in axon tracts. Temporally, peak expression of transcript and protein occurs during the second postnatal week. This period is characterized by extensive neurite outgrowth and synaptogenesis, supporting a role for the receptor in these processes. Collectively, these data suggest that Met signaling may be necessary for the appropriate wiring of forebrain circuits with particular relevance to social and emotional dimensions of behavior.
Coooooool. Here we touch on the complexity of brain formation, all the little things that need to go exactly right, and how MET might play a role in that incredibly complicated dance. Even better, a mouse model is used to gain an understanding of where and when peak expression of MET proteins occur, a period of significant changes to neural structures and the formation of synapses, the physical structures that enable thought. This is a dense paper, too dense to get deeply into blockquoting for this posting, but there are some parts that deserve notice, namely, documentation of spatially localized MET expression in brain areas associated with social behaviors and some fine grained information on the specific parts of synapse formation that utilize MET. Coolness factor: 8. Complexity Factor: 12.
Here is a paper that a lot of people that play skeptics on the Internet ought to hate, Distinct genetic risk based on association of MET in families with co-occurring autism and gastrointestinal conditions. (full paper)
In the entire 214-family sample, the MET rs1858830 C allele was associated with both autism spectrum disorder and gastrointestinal conditions. Stratification by the presence of gastrointestinal conditions revealed that the MET C allele was associated with both autism spectrum disorder and gastrointestinal conditions in 118 families containing at least 1 child with co-occurring autism spectrum disorder and gastrointestinal conditions. In contrast, there was no association of the MET polymorphism with autism spectrum disorder in the 96 families lacking a child with co-occurring autism spectrum disorder and gastrointestinal conditions. chi(2) analyses of MET rs1858830 genotypes indicated over-representation of the C allele in individuals with co-occurring autism spectrum disorder and gastrointestinal conditions compared with non-autism spectrum disorder siblings, parents, and unrelated controls.
There is a lot of caution in this paper, but the nice part is that there are biologically plausible mechanisms by which a reduction in MET could snowball into problems in the gastro-intestinal track.
In the gastrointestinal system, MET signaling modulates intestinal epithelial cell proliferation, and thus acts as a critical factor in intestinal wound healing. For example, activation of MET signaling via application of exogenous hepatocyte growth factor has been shown to reduce the effects of experimentally induced colitis, inflammatory bowel disease, and diarrhea.
Pushing on the other end of the balloon, increasing MET signaling, has been shown to help GI problems; no less than evidence that a genetic change associated with autism has biologically plausible mechanisms by which GI problems would be more prevalent. In fact, unless our findings of MET alleles are in error, or our clinical findings of the effects of HGF are spurious, it is absolutely expected. There is also a section with the startlingly simple, and simultaneously great idea of why findings like these might be useful markers for phenotypic categorization in studies in the future; i.e., to discern the prevalence of GI problems in autism, it might, for example, make sense to design that study to take presence or absence of MET alleles into consideration. Nice. Coolness Factor: 7. Insidiousness factor: 9.
Here’s another one that found associations with MET and social behavior, and GI disturbances again. Association of MET with social and communication phenotypes in individuals with autism spectrum disorder
Autism is a complex neurodevelopmental disorder diagnosed by impairments in social interaction, communication, and behavioral flexibility. Autism is highly heritable, but it is not known whether a genetic risk factor contributes to all three core domains of the disorder or autism results from the confluence of multiple genetic risk factors for each domain. We and others reported previously association of variants in the gene encoding the MET receptor tyrosine kinase in five independent samples. We further described enriched association of the MET promoter variant rs1858830 C allele in families with co-occurring autism and gastrointestinal conditions. To test the contribution of this functional MET promoter variant to the domains of autism, we analyzed its association with quantitative scores derived from three instruments used to diagnose and describe autism phenotypes: the Autism Diagnostic Interview-Revised (ADI-R), the Autism Diagnostic Observation Schedule (ADOS), and both the parent and the teacher report forms of the Social Responsiveness Scale (SRS). In 748 individuals from 367 families, the transmission of the MET C allele from parent to child was consistently associated with both social and communication phenotypes of autism. Stratification by gastrointestinal conditions revealed a similar pattern of association with both social and communication phenotypes in 242 individuals with autism from 118 families with co-occurring gastrointestinal conditions, but a lack of association with any domain in 181 individuals from 96 families with ASD and no co-occurring gastrointestinal condition. These data indicate that the MET C allele influences at least two of the three domains of the autism triad.
Really sort of plain, but very nice to see the GI component validated in another data set. Coolness factor 5.
Then a few months ago, Prenatal polycyclic aromatic hydrocarbon exposure leads to behavioral deficits and downregulation of receptor tyrosine kinase, MET was released, an uber cool showcase of the autism bigfoot, the often regaled, only very rarely documented, gene/environment interaction.
Gene by environment interactions (G × E) are thought to underlie neurodevelopmental disorder, etiology, neurodegenerative disorders, including the multiple forms of autism spectrum disorder. However, there is limited biological information, indicating an interaction between specific genes and environmental components. The present study focuses on a major component of airborne pollutants, polycyclic aromatic hydrocarbons (PAHs), such as benzo(a)pyrene [B(a)P], which negatively impacts cognitive development in children who have been exposed in utero. In our study, prenatal exposure of Cpr(lox/lox) timed-pregnant dams to B(a)P (0, 150, 300, and 600 μg/kg body weight via oral gavage) on embryonic day (E14-E17) consistent with our susceptibility-exposure paradigm was combined with the analysis of a replicated autism risk gene, the receptor tyrosine kinase, Met. The results demonstrate a dose-dependent increase in B(a)P metabolite generation in B(a)P-exposed Cpr(lox/lox) offspring. Additionally, a sustained persistence of hydroxy metabolites during the onset of synapse formation was noted, corresponding to the peak of Met expression. Prenatal B(a)P exposure also downregulated Met RNA and protein levels and dysregulated normal temporal patterns of expression during synaptogenesis. Consistent with these data, transcriptional cell-based assays demonstrated that B(a)P exposure directly reduces human MET promoter activity. Furthermore, a functional readout of in utero B(a)P exposure showed a robust reduction in novel object discrimination in B(a)P-exposed Cpr(lox/lox) offspring. These results confirm the notion that common pollutants, such as the PAH B(a)P, can have a direct negative impact on the regulated developmental expression of an autism risk gene with associated negative behavioral learning and memory outcomes.
Oh snap. A common pollutant (well, common in the last few decades anyways), is shown to interact with MET in a dose dependent fashion to reduce protein expression in the brain during embryonic development and cause ‘a robust reduction in novel object discrimination’. (Ouch) This is an example of just what we mentioned above, referenced Herbert, concerning the possibility of MET as a gene sensitive to ‘environmental exposures’. Indeed. From the discussions section:
The results from the present study demonstrate that the transcription and developmental expression patterns of a replicated ASD risk gene, MET, are highly sensitive to a common PAH pollutant. In utero exposure to B(a)P produces an oxidative milieu of B(a)P metabolites in offspring during a key postnatal period of synapse development, providing evidence that environmental exposure creates a sustained cerebral cortical burden that likely contributes to an increased oxidative load. Oxidative stressors in the form of metabolites would be expected to negatively impact gene expression (Kerzee and Ramos 2000) and, more specifically, receptor tyrosine kinase function, including Met (Li et al. 2007). These data suggest that B(a)P-induced exposure would impact the expression of key neurodevelopmental genes, including Met. Additionally, the predominance of the 3-OH and 9-OH metabolites places a sustained burden in the brain because of the potential for further oxidization to form B(a)P quinones (McCallister et al. 2008, Hood et al. 2000, Brown et al. 2007) which undergo redox cycling to generate reactive oxygen species (Kerzee and Ramos 2000, Bolton et al., 2000).
And
In conclusion, specific developmental events such as glutamatergic excitatory synapse formation and maturation may be particularly vulnerable to G x E effects that impact regulatory and signaling proteins involved in this process. While we do not suggest that the current study reflects specific defects related to a complex clinical condition such as the ASDs, current molecular, behavioral and functional imaging data are converging on the concept that the ASDs are a manifestation of altered local and long-distance cortical connectivity (Geschwind et al. 2007, Bill and Geschwind 2009, Geschwind and Levitt, 2007, Levitt and Campbell 2009). Also, Met and other related signaling components of this receptor tyrosine kinase pathway have been implicated in both syndromic and idiopathic disorders where the ASDs are diagnosed at a high rate. In combination with risk alleles in key genes, the in utero exposure to PAHs such as B(a)P, which results in both a reduction in absolute levels and the mistiming of peak Met expression, could drive the system toward a pathophysiological threshold that neither genetic risk nor environmental factors could produce individually. The present study focused on the neocortex, but given the highly restricted spatial and temporal expression of Met in mouse limbic circuits associated with social-emotional development and cognition (Judson et al. 2009), it is likely that perturbations occur throughout these key circuits, including in the hippocampus.
Really cool stuff; particularly the finding that developmental, in utero exposure was capable of driving abnormal protein expression well after birth. This is the best of both sides of the genetics versus environment conundrum; the kind of finding that sheds light on how environmental pollutants could be participating in increasing the number of children with autism by interacting with genetically susceptible children. But what I love about this is that it is the death knell of the fairytale of a static rate, or near static rate of autism, just having the genes or the exposure isn’t enough; instead, the interaction of alleles and timed exposure ‘could drive the system toward a pathophysiolical threshold that neither genetic risk nor environmental factors could produce individually’. I think there are some more findings coming from this group soon that might be exciting, or terrifying, depending on how you see it. (or both). Coolness factor: 99.
So what have we learned and just how cool is it?
1) The MET receptor enables some types of cellular signaling that have relevance to the autism community including synapse formation, immune modulation, and gastro intestinal function. The ligand, or trigger of the MET receptor is HGF.
2) Certain alleles of the MET gene that result in decreased expression are more common in children with autism than people without autism.
3) Consistent with findings of increased prevalence of MET alleles, MET protein expression was found to be decreased in brain tissue from people with autism. Other, related proteins, HGF, PLAUR, and SERPINE were also found to be disturbed.
4) Following up on the differential findings of SERPINE and PLAUR, genetic studies found gene to gene interactions between the MET allele and alleles involved with production of SERPINE and PLAUR. Some of the proteins in question are known to be particularly vulnerable to environmental interference.
5) Animal models tell us that MET is heavily expressed in many areas of the mammalian brain during prenatal and postnatal development, and we gain insight into the spatial and temporal expression of MET during the intricate dance of brain formation.
6) Two studies add evidence that the one function of decreased MET expression, GI disturbances, are indeed found with greater consistency within children with autism and the MET allele. This should be a relatively unsurprising finding considering what we know about MET and children with autism.
7) Finally, a portrait of genetic / environmental interactions capable of disturbing physiology and behavior in ways consistent with findings in autism is rendered using an agent that is the product of the automobile age and already associated with decreased cognitive skills for groups with the highest gestational exposure.
It should be noted that this is just a slice of the MET papers out there in the autism realm; they all shared one or more authors, I picked them because they seem to show a nice progression of knowledge, and incremental approach towards learning more. There is a lof more to learn, in particular, I think that the immune modulating effects of reduced expression would be an interesting subject, but one that will have to wait for another posting.
– pD
Low Penetrance Environmental Impacts, Gene Environmental Interactions, and the Depressingly Bad Jokes that Infiltrate Autism Discussions
Posted on: January 8, 2011
Hello friends –
There’s been something at the back of my mind for a while now regarding the potential for environmental influences to participate in autism, and indeed, a true rise in the number of children that have developmental problems that I’ve been struggling with articulating elegantly. The right course came to me while reading threads where the recent autism risk as proximity to highways paper was discussed. I’m actually not too big on the paper, it is very preliminary, uses some terms that are kind of confusing, and at very best, should be used as a guide for more targeted studies. For anyone who didn’t see it when it came out, essentially it reported a small increase of risk of having a baby with autism as the pregnant mother lived closer to some types of highways.
What I liked about this study is that at the core, there was a twinge of a biologically plausible mechanism, specifically, exposure to pollutants during development and consequent interference with neural development. Examples given in the text including possible endocrine disrupting effects of some types of automotive exhaust, and studies showing altered glutamate expression and associated plasticity defects resulting from pollutants.
What I didn’t like about the study is that it didn’t include any biomarkers and seemed relatively soft on the definitional terms. It was essentially a GIS placement and association lookup; lots of data and easy to find phantoms. A methodologically similar study by Bearman was released a few months previously; purporting to assign a very specific percentage of autism increase (16%) to the spatial proximity of other parents with children with autism, with the idea being that those chatty parents convinced their close neighbors to get their child diagnosed, while those people who more than 500 meters from a child with autism, and therefore don’t talk to as many people, failed to get their child diagnosed. I came down pretty hard on Bearman and don’t see much difference to apply less skepticism here. I will note, however, with no small amount of amusement, that when Bearman was discussed, no one seemed too concerned about the lack of control for urbanicity in the ‘skeptical’ realm. Big surprise.
The skeptics took the freeway paper apart, or in some instances, took apart a reporter or blogger who was spinning the findings as stronger than they were. I was more or less in agreement with the skeptics ideas on this one; this paper certainly was not sufficiently strong to make any conclusive statements and as usual, some headlines got it way wrong.
On the other hand, according to my underlying principles of subtle change still being meaningful, the humbling complexity of poking around with systems like embryonic development, and the difficult to overstate gulf between what we know and what we think we know about the effects of our reckless introduction of a galaxy of sythentic chemicals into the environment our infants are born into, this study fit in pretty nicely; at the very least as a reason to perform bioinformatic analysis of pregnant women to test for biomarkers of exhaust exposure and cognitive outcomes a few years down the ‘road’.
It didn’t take long before the gross over simplifications started rolling in though; i.e., ‘If this study is valid, we should have seen the rise in autism when the Interstate program was initiated in the 1950’s!’ [cue laugh track], or ‘I guess I have genes that made me live near an Interstate’. [cue whoot whoot track] It occurred to me that the Interstate jokes are a good illustration of what is largely wrong with nearly every single discussion on environmental participation you stumble into on the Internet. On one hand, the notion that unless an environmental study has sufficient power to prove a causal relationship for autism, or indeed, can be shown to be unable to account for all autism cases, it is safe to be mocked, or for the more academically minded, accused of being the result of data dredging. Similarly, anything showing a glimmer of plausibility that isn’t a genetic finding can lends itself towards showing how worthless the genetic angle is. These are useful cards to play if your goal is to bash environmental causation theories (and thereby, vaccination causation theories), or if your goal is to bash genetic theories; but ultimately are wastes of time if we want to understand a condition with the murky history and multifaceted manifestations of autism. The crux of what really bothered me about both sides of the Internet joke is that they each ignore meaningful information that can be offered from the other side. It is worse than dumb, it is wasteful.
Stepping away from the environmental end for a moment, I think it is safe to say that everyone is beginning to realize that the hunt for high impact genetic changes that can explain more than a tiny fraction of our autism cases is an abject failure. While there are some genetic changes, like Fragile X, that confer extremely high risk of autism, the absolute number of people with such changes is relatively simple to determine, and they comprise a vanishingly small subset of the children with autism. What we do seem to be finding is that there are lots of genetic changes that confer a small risk of having autism, the so called, low penetrance genetic changes. The idea here is that if you have many, (maybe as many as a dozen or more) low penetrance genes, the cumulative effects build up until a physiological end point is reached wherein autistic behaviors manifest. I actually like the idea behind low penetrance genes a lot; it makes a lot of our finings of genetics make sense, and I absolutely believe in a strong genetic participation in autism.
Remember, at the end of the day, genes are nothing more than blueprints for building proteins. Most genetic alterations don’t involve complete additions, or removals, of proteins, but rather, creation of a little less, or a little more of a protein, or perhaps, creation of proteins that are just a tiny bit different than ‘normal’, sort of like autism itself. While the environment these proteins enter, or are regulated into entering, starts influencing the eventual biological outcome in the most immediate sense imaginable, the end points of genetics, these proteins and their precise structures are indisputably important in what is happening in everything our bodies do; including, in some instances, have autism.
Consider the tightly orchestrated formation of the microscopic chasms between neurons, the process of synaptogenesis. Dozens (or hundreds) of chemicals dance together in order to form the structures in our brains that exchange chemical messengers, neurotransmitters, that literally form the foundation of neuron to neuron communication, and thus, cognition; the physical constructs of thought. It is a biological cauldron that we are just beginning to comprehend, the mind formingly intricate, time dependent interplay of a chemical deck of cards orders of magnitude more complicated than sequencing the genome.
The evidence for altered synapses, and modified synaptic function in autism, and most (all?) other developmental disorders is impossible for an intellectually honest observer to deny. Some of the most commonly found genetic alterations in people with autism involve genes known to participate in the formation, maintenance, or functioning of synapses. For example, neurexin , shank, and neuroligin, are some well known, or at least, well reported reported genes intimately attached to synapse function also found associated with autism, and our list should also include calcium expression and adhesion genes (and many, many others). Each of these genes or processes contribute to the synapse in subtle, but different ways, at different times, and yet we can see that interferences anywhere down the functional class of chemicals is associated with autism. Yet, very few people, (I’ve read of none), have been found to have a neurixin allele, a shank allele, and a neruoligin mutation. And there are some people who have the same mutations, but do not exhibit autistic behavior. There are also a great many people that have no known mutations in any of these genes, and still, receive an autism diagnosis. What does this tell us?
It should tell us that while there are lots of genetic ways that synapse function can be altered in such a way that autistic behaviors bubble up to the diagnosis endpoint, but more importantly, the critical question need not necessarily revolve around what genes you have, but rather, is synapse function manipulated? Furthermore, we should be able to conclude that simply having a single modifier (i.e., one shank mutation) go wrong isn’t a guarantee of an autism diagnosis, and thus; the participation of individual mutations is real, but small. [I would also argue that it is likely that there are a great number of as of yet, undetected genetic misprints that contribute in the same real, but subtle ways.]
Another more accessible example of a low penetrance gene is the MET gene, which produces a protein known to interact with a lot of important processes involved in autism, including brain formation, immune system functioning, and intestinal repair. There have been a lot of high quality studies on the MET mutations in the past few years including those that report higher incidences of MET mutations in children with autism and gastrointestinal problems, higher findings of MET alleles in autism, association to communication phenotypes and MET expression, replication of above studies, evidence of interaction with other genes known to be associated with autism, decreased expression in post mortem brain tissue, and animal studies showing differential, time dependent expression of MET. (and many, many others).
The kicker towards this discussion, howeever, is that the changes to the MET gene are really, very, very common. Nearly one half of everyone has the low MET production gene, but even still, many more people with autism have it. So, while it is clearly implicated, other changes are obviously necessary for that particular genetic change to result in autism. What we are learning about the systems affected by MET, or lots of the genes implicated in autism, is that very subtle changes towards critical processes are sufficient to modify the course of development. Somewhat counter intuitively, I would argue that the implication of this is compelling evidence (or terrifying news) for those of us with worries about the possibility of an environmentally driven increase in the number of people with an autism diagnosis; indeed, it argues that just like genetics, we must admit the reality that if genes can be low penetrance, so too, then, can environmental impacts.
For example, back to brain formation. We know that the neurexin proteins participate in forming our synapses. But we have evidence that hypo
thyroidism can lead to structural changes during development, and we also know that there is increasing evidence that endocrine disruptors can interferre with thyroid metabolism, or for that matter, a wide range of findings on endocrine disruptors and cognitive function. Or if we look to pesticides, we have evidence that developmental exposure to diazonon can modify neurotransmitter function, with similar findings are available for other classes of pesticides. Similarly with heavy metals.
The skeptics would claim with some legitimacy that there are significant dose dependency problems to be addressed before we should start pointing to every experimental slice of evidence of potential harm and claiming that the sky is falling. But. What if, in fact, we need only perturb the process of brain development a little bit, and with a little help from other, low penetrance genes or other exposures, developmental trajectories begin to alter? This would seem to be precisely what we are learning from the genetic angle; it isn’t one big thing incorrectly designed, it is lots of small things. And while our genetic code has, for the most part, remained stable; our environment today is vastly overpopulated with chemicals capable of minor, but real, effects when compared to yesteryears past.
The search for a single environmental impact with the ability to explain a significant portion of autism diagnosis is as futile as the hunt based on genetics. This makes for a far messier landscape, but also one that fits my terrifying, over arching principle of the Fairytale of a static (or near static) rate of autism, that our uncontrolled experiment of introducing synthetic chemicals into our environment coupled with widespread social changes with real physiological impacts, a set of experiments absolutely unprecedented in the history of living things on planet Earth, that changes to our offspring are unavoidable. To suggest otherwise, strikes me as either the height of arrogance, or the depths of ignorance.
Going back to the freeway study for a minute, I ran into a paper while writing this piece that involves pollutants, interaction with the MET gene, gene x environment interactions, and low penetrance impacts that I think has salience towards this discussion.
Here is the abstract:
Gene by environment interactions (G × E) are thought to underlie neurodevelopmental disorder, etiology, neurodegenerative disorders, including the multiple forms of autism spectrum disorder. However, there is limited biological information, indicating an interaction between specific genes and environmental components. The present study focuses on a major component of airborne pollutants, polycyclic aromatic hydrocarbons (PAHs), such as benzo(a)pyrene [B(a)P], which negatively impacts cognitive development in children who have been exposed in utero. In our study, prenatal exposure of Cpr(lox/lox) timed-pregnant dams to B(a)P (0, 150, 300, and 600 μg/kg body weight via oral gavage) on embryonic day (E14-E17) consistent with our susceptibility-exposure paradigm was combined with the analysis of a replicated autism risk gene, the receptor tyrosine kinase, Met. The results demonstrate a dose-dependent increase in B(a)P metabolite generation in B(a)P-exposed Cpr(lox/lox) offspring. Additionally, a sustained persistence of hydroxy metabolites during the onset of synapse formation was noted, corresponding to the peak of Met expression. Prenatal B(a)P exposure also downregulated Met RNA and protein levels and dysregulated normal temporal patterns of expression during synaptogenesis (!). Consistent with these data, transcriptional cell-based assays demonstrated that B(a)P exposure directly reduces human MET promoter activity. Furthermore, a functional readout of in utero B(a)P exposure showed a robust reduction in novel object discrimination in B(a)P-exposed Cpr(lox/lox) offspring. These results confirm the notion that common pollutants, such as the PAH B(a)P, can have a direct negative impact on the regulated developmental expression of an autism risk gene with associated negative behavioral learning and memory outcomes.
(my emphasis)
I have to say, finding this paper was a bit of tragic humor for me; it was published in December 2010, with zero fanfare from the press, as opposed to the confounder heavy, Residential Proximity to Freeways and Autism in the CHARGE study, study, which had a thousand similar articles in Google News. But here we find a superb example of what gets bandied around a lot when in quick passing but rarely with any meat behind the discussion; a real life, experimentally sound version of a gene environment interaction that integrates biologically plausible mechanisms that is able to describe what is observed physiologically in autism with dose responses. Beautiful. But, it gets even better. It just so happens, the classifications of agents in use in this study, polycyclic aromatic hydrocarbons, are generated, in some instances, by car exhaust. In fact, in Detection of polycyclic aromatic hydrocarbon exposure from automobile exhaust fumes using urinary 1-hydroxypyrene level as an index, the authors conclude in part that “Automobile exhaust fume exposed subjects have a higher risk to be exposed to PAHs than the non-exposed subjects”. Go figure.
Whatever the problems with the freeway CHARGE study, they pale in comparison to the problems that the notion that because we didn’t observe increases in autism when the Interstate system was constructed, the findings must be spurious. Similarly, genetic predisposition is an indisputable fact; and knowing which genes are implicated in autism can help us intelligently target environmental factors that might be changing our infants.
– pD
- In: Autism | BDNF | Biological Plausibility | BPA | Brain | Epigenetics | Epigenome | Genetics | Impending Doom | Intriguing | PBDE | Pesticides | Prevalence | Purkinje | Reelin | Some Jerk On The Internet | Synthetics | The Fairytale | Uncategorized
- 33 Comments
Hello friends –
I ran into a few abstracts, read a few papers, and tried to get my way through one really dense paper in the past few weeks that got me thinking about this post. It’s all shook up, like pasta primavera in my head, but hopefully something cogent will come out the other end. (?)
Of the metabolic conditions known to be associated with having a child with autism, hypothyroidism is one that I keep on running into by way of the pubmed alert grapevine. By way of example, we have two studies that looked for autoimmune conditions in family members which found hypothyroidism to be one of many autoimmune diseases as a risk factor for autism, including, Familial clustering of autoimmune disorders and evaluation of medical risk factors in autism, and Increased prevalence of familial autoimmunity in probands with pervasive developmental disorders. This shouldn’t be too surprising, we know that, for example, perinatal hypothyroidism is a leading cause of mental retardation, with similar findings for the condition during pregnancy. It turns out, it appears that rates of hypothyroidism are slightly increasing, though at this time, the increases are of relatively small proportions, and as such, may be artifacts unrelated to an actual increase in classically recognized hypothyroidism. In any case, I think it is safe to say that interference with thyroid metabolism is something to be avoided at all costs when possible.
So after having read about that, this paper showed up in my inbox a while ago:
Thyroid hormones have long been known to play important roles in the development and functions of the central nervous system, however, the precise molecular mechanisms that regulate thyroid hormone-responsive gene expression are not well understood. The present study investigated the role of DNA methylaion and histone acetylation in the effects of perinatal hypothyroidism on regulation of reelin and brain-derived neurotrophic factor (BDNF) gene expression in rat hippocampus. The findings indicated that the activities of DNA methyltransferase (DNMT), methylated reelin and BDNF genes were up-regulated, whereas, the activities of histone acetylases (HAT), the levels of global acetylated histone 3 (H3) and global acetylated histone 4 (H4), and acetylated H3, acetylated H4 at reelin promoter and at BDNF gene promoter for exon II were down-regulated in the hippocampus at the developmental stage of the hypothyroid animals. These results suggest that epigenetic modification of chromatin might underlie the mechanisms of hypothyroidism-induced down-regulation of reelin and BDNF gene expression in developmental rat hippocampus
This gets interesting for autism because reelin, and bdnf levels have been found to be decreased in several studies in the autism population, with direct measurements, genetic expression, mouse knockout based models of autism , and genomic alterations all being implicated. There have been some negative genetic studies, but considering that it isn’t always the genes you have, but the genes you use, our other available evidence certainly points to BDNF and reelin involvement with some percentage of children with autism, and the association is such that a reduction in reelin or BDNF is a risk factor for developing autism. It would seem that the paper above might give some insight into the lower level details of the effects of hypothyroidism and subsequent developmental trajectories; modifications of reelin expression; through epigentic mechanisms, no less!. That’s pretty cool!
Then, I got my hands on a review paper that tries to go into detail as to the functional mechanism by which reelin deficiency could contribute to ASD, Neuroendocrine pathways altered in autism. Special role of reelin. It is a review that touches on a variety of ways that reelin contributes to neurodevelopment that have findings in the autism realm, including neuronal targeting and migration during brain formation, interactions with the serotonin and GABA systems, testosterone, and oxytocin. In short, there are plenty of ways that decreased reelin expression can impact development in ways that mirror our some of our observations in autism.
Of the many things that convince me that we are doomed, the proliferation of chemical compounds whose interactions within our bodies we scarcely understand is among them. In my readings on endocrine disruptors, one thing I found that seemed to be worrying lots of researchers was that some classes of these chemicals are capable of interfering with thyroid metabolism, and in some cases interfering with development of cells known to be associated with autism. Terrifyingly enough, since I read those papers, several others have come out, including Polybrominated Diphenylether (PBDE) Flame Retardants and Thyroid Hormone during Pregnancy and Mini-review: polybrominated diphenyl ether (PBDE) flame retardants as potential autism risk factors. At this point, it is important to point out that, as far as I know, there have not been any studies showing that non occupational exposure to PDBEs or other environmental pollutants can lead to classically defined hypothyroidism, at least none that I know of. (?) Be that as it may, I think it is realistic to assume any interference in thyroid metabolism is a bad thing, and while finding people in the outlier regions of hypo (or hyper) thyroidism gives us information on extreme environments, it would take someone with a lot of misplaced faith to assume that we can safely disturb thyroid metabolism just a little bit, and everything will come out in the wash.
I’ve had the argument made to me in the past that environmental pollutant driven increases in autism lacked biological plausible mechanisms; this argument is almost always made within a context of trying to defend the concept of a static rate of autism. While the papers I’ve linked to above do not provide conclusive proof that our changing environment is causing more children to be born with autism, they do provide increasing evidence of a pathway from pollutants to ASD, and indeed, the lack of biological plausibility becomes an increasingly flacid foundation on which to assume that our observations of an increased rate of autism are illusory. Unfortunately, in my opinion, the focus on vaccines has contributed to the mindset that a static rate of autism (or nowadays, maybe a tiny increase), must be protected at all costs, including some ideas on the application of a precautionary principle that seem outright insane to me (or at least, the exact opposite of what I would consider to be a precautionary path).
One thing is for certain, the number of child bearing women in developing countries with measurable concentrations of chemicals known to interferre with thyroid metabolism nears 100% in the industrialized nation as we eat , drink, breathe and bathe in the microscopic remnants of packaging materials, deteriorating carpet fibers, and baby clothes that are made to be fire resistant. This is an environment unambiguously different than that encountered by any other generation of infants in the history of mankind. To believe that we can modify our environment so drastically without having an impact seems incredibly naive to me, or on some days, just plain old stupid.
– pD
Confusing and Fascinating Findings – Gestational Age at Delivery and Special Educational Need
Posted on: June 11, 2010
Hello friends –
I ran across this one on accident the other day (why wasn’t it in one of my pubmed alerts?):
Background
Previous studies have demonstrated an association between preterm delivery and increased risk of special educational need (SEN). The aim of our study was to examine the risk of SEN across the full range of gestation.
Methods and Findings
We conducted a population-based, retrospective study by linking school census data on the 407,503 eligible school-aged children resident in 19 Scottish Local Authority areas (total population 3.8 million) to their routine birth data. SEN was recorded in 17,784 (4.9%) children; 1,565 (8.4%) of those born preterm and 16,219 (4.7%) of those born at term. The risk of SEN increased across the whole range of gestation from 40 to 24 wk: 37–39 wk adjusted odds ratio (OR) 1.16, 95% confidence interval (CI) 1.12–1.20; 33–36 wk adjusted OR 1.53, 95% CI 1.43–1.63; 28–32 wk adjusted OR 2.66, 95% CI 2.38–2.97; 24–27 wk adjusted OR 6.92, 95% CI 5.58–8.58. There was no interaction between elective versus spontaneous delivery. Overall, gestation at delivery accounted for 10% of the adjusted population attributable fraction of SEN. Because of their high frequency, early term deliveries (37–39 wk) accounted for 5.5% of cases of SEN compared with preterm deliveries (<37 wk), which accounted for only 3.6% of cases.
Conclusions
Gestation at delivery had a strong, dose-dependent relationship with SEN that was apparent across the whole range of gestation. Because early term delivery is more common than preterm delivery, the former accounts for a higher percentage of SEN cases. Our findings have important implications for clinical practice in relation to the timing of elective delivery
[Full paper from link. Emphasis is mine]
Essentially the authors evaluated gestational lengths with a fine tooth comb to discern if ‘early’, though not technically ‘pre-term’ delivery was associated with a ‘special education need’ (SEN), which in this case embodies a range of developmental problems including dyslexia, autism, or even physical problems like deafness or vision problems.
What the authors found was that there were subtle, but real effects in the likelyhood of having a special education need for non full term births that was dose dependent, but even included children that would not necessarily be considered early by existing standards.
Our study demonstrated a strong trend of decreasing risk of SEN with advancing gestational age at birth. The key finding of the present analysis is that this trend continued across gestational ages classified as term. Although the risk of SEN was highest among infants who were delivered preterm (<37 wk gestation), these accounted for only 5.1% of deliveries. Therefore, only a relatively small proportion of SEN (3.5%) could be attributed to preterm delivery. By contrast, 39.6% of infants were delivered between 37 and 39 wk gestation. Therefore, whilst these early term infants had only a moderately increased risk, 5.3% of SEN cases could be attributed to early term delivery.
The authors claim that the finding of effects at early, but not pre-term gestational lengths is one that is largely missing from existing studies, which have not taken these date ranges into consideration, or the ones that did, were not studying for cognitive problems, and indeed, excluded children with these criteria. Curiously, they also report an increase in SEN in children who had extra gestational periods, i.e., > 41 weeks in some studies.
The authors make absolutely no speculation as to what might be driving increased special education needs as the result of premature or early birth.
Looking at their results, one of the most striking things is that the impact did not alter if elective (i.e. C-Section) versus non-elective births were used as a variable. But this has deep ramifications for the autism storyline, which holds that if there are environmental factors that can contribute to autism, they are prenatal, and indeed, are often thought to involve insults very early in the prenatal period. In this case, we know that a genetic or environmental force isn’t contributing to the early birth, because it didn’t matter if the birth was spontaneous or not. The only area for an effect is postnatal. That is a big, big difference in the narrative.
Is this a matter of some just in time epigenetic programming happening in the womb that doesn’t get a chance to finish up in early births? Alternatively it could be that early birth allows for environmental exposures that the infant is not quite prepared to deal with. Or it could be both, or neither, or an illusory finding, but if these findings can be replicated, it raises a lot of questions about the sacred line between prenatal and postnatal environmental influences.
Unfortunately, the raw data for this project doesn’t seem to be available online; it might be really nice to see if there were patterns to be observed had particular salience to our population of interest.
– pD
Fascinating Study – Immune transcriptome alterations in the temporal cortex of subjects with autism
Posted on: June 7, 2010
Hello friends –
I’ve been referencing this paper in some discussions online for a while; I’ve read it, and in fact, while working on another project, got the opportunity to speak with one of the authors of the paper. It’s a very cool paper with a lot of information in it, some of which, could be considered inconvenient findings. Here is the abstract:
Immune transcriptome alterations in the temporal cortex of subjects with autism
Autism is a severe disorder that involves both genetic and environmental factors. Expression profiling of the superior temporal gyrus of six autistic subjects and matched controls revealed increased transcript levels of many immune system related genes. We also noticed changes in transcripts related to cell communication, differentiation, cell cycle regulation and chaperone systems. Critical expression changes were confirmed by qPCR (BCL6, CHI3L1, CYR61, IFI16, IFITM3, MAP2K3, PTDSR, RFX4, SPP1, RELN, NOTCH2, RIT1, SFN, GADD45B, HSPA6, HSPB8and SERPINH1). Overall, these expression patterns appear to be more associated with the late recovery phase of autoimmune brain disorders, than with the innate immune response characteristic of neurodegenerative diseases. Moreover, a variance-based analysis revealed much greater transcript variability in brains from autistic subjects compared to the control group, suggesting that these genes may represent autism susceptibility genes and should be assessed in follow-up genetic studies.
(emphasis is mine) [Full paper freely available from that link]
I am particularly intrigued by the second bolded sentence regarding the “these expression patterns appear to be more associated with the later recovery phase of autoimmune brain disorders, than the innate immune response characteristic of neurodegenerative diseases”. I’ve had it put to me previously that we should not necessarily implicate neuroinflammation in autism, the argument being that even though we had evidence of chronically activated microglia, what we do not seem to have evidence for is actual damage to the brain, and ergo, the neuroinflammation may actually be a byproduct of having autism, as opposed to playing a causative role, or that in fact, the neuroinflammation might even be beneficial. There have been some other places where the claim has been made that because our profile of neuroinflammation doesn’t match more classically recognized neurodegenerative disorders (i.e., MS/Alzheimer’s/Parkinson’s), that therefore, certain environmental agents need not be fully investigated as a potential contributor to autism. This is the first time that I am aware that someone has attempted to classify the neuroinflammatory pattern observed in autism not only as distinctly different from classical neurodegenerative diseases, but to also go so far as to provide a more refined example.
From the Introduction:
In order to better understand the molecular changes associated with ASD, we assessed the transcriptome of the temporal cortex of postmortem brains from autistic subjects and compared it to matched healthy controls. This assessment was performed using oligonucleotide DNA microarrays on six autistic-control pairs. While the sample size is limited by the availability of high-quality RNA from postmortem subjects with ASD, this sample size is sufficient to uncover robust and relatively uniform changes that may be characteristic of the majority of subjects. Our study revealed a dramatic increase in the expression of immune system-related genes. Furthermore, transcripts of genes involved in cell communication, differentiation, cell cycle regulation and cell death were also profoundly affected. Many of the genes altered in the temporal cortex of autistic subjects are part of the cytokine signaling/regulatory pathway, suggesting that a dysreactive immune process is a critical driver of the observed ASD-related transcriptome profile.
I was initially very skeptical about this, with a sample set so small, wasn’t it difficult to ascertain if their findings were by chance or not? It turns out, the answer depends on the type of datapoint you are evaluating against. A powerful tool in use by the researchers is a recent addition to the genetic analysis research suite, not only the ability to scan for thousands of gene activity levels simultaneously, but the use of known gene networks to identify if among those thousands of results, related genes are being expressed differentially. This is important for some amazingly robust findings presented later in the paper, so lets sidetrack a little bit. Here is a nice overview of the process being used:
Although genomewide RNA expression analysis has become a routine tool in biomedical research, extracting biological insight from such information remains a major challenge. Here, we describe a powerful analytical method called Gene Set Enrichment Analysis (GSEA) for interpreting gene expression data. The method derives its power by focusing on gene sets, that is, groups of genes that share common biological function, chromosomal location, or regulation. A common approach involves focusing on a handful of genes at the top and bottom of L (i.e., those showing the largest difference) to discern telltale biological clues. This approach has a few major limitations.
(i) After correcting for multiple hypotheses testing, no individual gene may meet the threshold for statistical significance, because the relevant biological differences are modest relative to the noise inherent to the microarray technology.
(ii) Alternatively, one may be left with a long list of statistically significant genes without any unifying biological theme. Interpretation can be daunting and ad hoc, being dependent on a biologist’s area of expertise.
(iii) Single-gene analysis may miss important effects on pathways. Cellular processes often affect sets of genes acting in concert. An increase of 20% in all genes encoding members of a metabolic pathway may dramatically alter the flux through the pathway and may be more important than a 20-fold increase in a single gene.
(iv) When different groups study the same biological system, the list of statistically significant genes from the two studies may show distressingly little overlap (3).
So, back to Garbett, not only did the authors find a great number of genes overexpressed in the autism group (and a smaller number, underexpressed), when they threw their thousands of results of individual genes into the GSEA, what came back was that several genetic pathways were very significantly altered, many of them immune mediated. This is a big step in understanding in my opinion. I believe we have likely come full circle on our understanding of very high penetrance genes that might be driving towards a developmental trajectory of autism; i.e., Rhett, Fragile-X. But using this technique we can determine if entire biological pathways are altered by measuring the output of genes. Specifically the point made in bullet (iii) stands out to me; having a twenty fold increase in a single gene might not be too big a deal if the other participants in the proteins function cannot be altered by twenty fold as a result due to other rate limiting constraints; but if we can see related sets of genes with similar expression profiles, we can get a much better picture of the biological results of different expression.
The methods get dense pretty quickly, but are worth a shot to show how thorough the researchers were to insure that their findings were likely to be signficant. Essentially they performed three different statistical tests against their results of differentially expressed genes and broke their results down into genes that passed all three tests, two of three test, or one of three tests. Furthermore, a selected twenty genes were targeted with qPCR validation, which in all cases showed the expected directionality; i.e., if the expression was increased in the transcriptome analysis, qPCR analysis confirmed the increased expression.To provide another benchmark, they tested for other genes known to be associated with autism, REELN, and GFAP and found results consistent with other papers.
Having determined a large number of differentially expressed genes, the authors then went to try to analyze the known function of these genes.
These classifications were performed on a selected gene set that is differentially expressed between AUT and CONT subjects; based on the success of our qPCR validation, we decided to perform this analysis using transcripts that both reported an |ALR>1| and that reached p<0.05 in at least 2/3 statistical significance comparisons. Of 221 such transcripts, 186 had increased expression in AUT compared to CONT, while only 35 genes showed reduced expression in the AUT samples. We subjected these transcripts to an extensive literature search and observed that 72 out of 193 (37.3%) annotated and differentially expressed transcripts were either immune system related or cytokine responsive transcripts (Supplemental Material 2). Following this first classification, we were able to more precisely sub-classify these 72 annotated genes into three major functional subcategories, which overlap to a different degree; 1) cell communication and motility, 2) cell fate and differentiation, and 3) chaperones (Figure 3). The deregulation of these gene pathways might indicate that the profound molecular differences observed in the temporal cortex of autistic subjects possibly originate from an inability to attenuate a cytokine activation signal.
That last sentence packs a lot of punch for a couple of reasons. It would seem to be consistent with their statements regarding a “late recovery phase” of an autoimmune disorder; i.e., an immune response was initiated at some point in the past, but has yet to be completely silenced. This also isn’t the first time that the idea of problems regulating an immune response (i.e., the inability to attenuate a cytokine activation signal) has been suggested from clinical findings, for example, in Decreased transforming Growth Factor Beta1 in Autism: A Potential Link Between Immune Dysregulation and Impairment in Clinical Behavioral Outcomes, the authors found an inverse correlation between TGF-Beta1 and autism behavioral severity:
Given that a major role of TGFβ1 is to control inflammation, the negative correlations observed for TGFβ1 and behaviors may suggest that there is increased inflammation and/or ongoing inflammatory processes in subjects that exhibit higher (worse) behavioral scores.
As such, TGFβ has often been considered as one of the crucial regulators within the immune system and a key mediator in the development of autoimmune and systemic inflammation.
In summary, this study demonstrates that there is a significant reduction in TGFβ1 levels in the plasma of young children who have ASD compared with typically developing children and with non-ASD developmentally delayed controls who were frequency-matched on age. Such immune dysregulation may predispose to the development of autoimmunity and/or adverse neuroimmune interactions that could occur during critical windows in development.
[full paper from the link]
The theme of a critical window of development and enduring consequences of insults during that window is one that is getting more and more attention recently; this is an area that is going to get more and more attention as time goes by, and eventually, as the clinical data continues to pile up, meaningless taglines aren’t going to be enough to keep us from dispassionately evaluating our actions.
The Discussion section is particularly nice, I’ll try not to just quote the entire thing. Here are the really juicy parts.
The results of our study suggest that 1) in autism, transcript induction events greatly outnumbers transcript repression processes; 2) the neocortical transcriptome of autistic individuals is characterized by a strong immune response; 3) the transcription of genes related to cell communication, differentiation and cell cycle regulation is altered, putatively in an immune system-dependent manner, and 4) transcriptome variability is increased among autistic subjects, as compared to matched controls. Furthermore, our study also provides additional support for previously reported involvement of MET, GAD1, GFAP, RELN and other genes in the pathophysiology of autism. While the findings were obtained on a limited sample size, the statistical power, together with the previously reported postmortem data by other investigators suggest that the observed gene expression changes are likely to be critically related to the pathophysiology seen in the brain of the majority of ASD patients.
There is some description of studies using gene expression testing in the autism realm where the authors ultimately conclude that technical and methodological differences between the studies make them difficult to tie together coherently. There is another small section re-iterating the findings that were similar to single gene studies; i.e., REELN, MET, and GAD genes.
The most prominent expression changes in our dataset are clearly related to neuroimmune disturbances in the cortical tissue of autistic subjects. The idea of brain inflammatory changes in autism is not novel; epidemiological, (DeLong et al., 1981; Yamashita et al., 2003; Libbey et al., 2005) serological studies (Vargas et al., 2005; Ashwood et al., 2006) and postmortem studies (Pardo et al., 2005; Vargas et al., 2005; Korkmaz et al., 2006) over the last 10 years have provided compelling evidence that immune system response is an essential contributor to the pathophysiology of this disorder (Ashwood et al. 2006). Finally, converging post-mortem assessments and measurements of cytokines in the CSF of autistic children (Vargas et al., 2005), may indicating an ongoing immunological process involving multiple brain regions
Nothing really new here to anyone that is paying attention, but good information for the extremely common, gross oversimplification that ‘immune abnormalities’ have been found in autism, but we don’t have any good reason to think they may be part of the problem. Of course, this is an argument you’ll see a lot of the time regarding everyone’s favorite environmental agent.
Altered immune system genes are often observed across various brain disorders, albeit there are notable differences between the observed transcriptome patterns. The majority of neuroimmune genes found activated in the autistic brains overlap with mouse genes that are activated during the late recovery or “repair” phase in experimental autoimmune encephalomyelitis (Baranzini et al., 2005). This suggests a presence of an innate immune response in autism. However, the altered IL2RB, TH1TH2, and FAS pathways suggest a simultaneously occurring, T cell-mediated acquired immune response. Based on these combined findings we propose that the expression pattern in the autistic brains resembles a late stage autoimmune event rather than an acute autoimmune response or a non-specific immune activation seen in neurodegenerative diseases. Furthermore, the presence of an acquired immune component could conceivably point toward a potential viral trigger for an early-onset chronic autoimmune process leading to altered neurodevelopment and to persistent immune activation in the brain. Interestingly, recently obtained gene expression signatures of subjects with schizophrenia (Arion et al., 2007) show a partial, but important overlap with the altered neuroimmune genes found here in autism. These commonly observed immune changes may represent a long-lasting consequence of a shared, early life immune challenge, perhaps occurring at different developmental stages and thus affecting different brain regions, or yielding distinct clinical phenotypes due to different underlying premorbid genetic backgrounds.
The last sentence, regarding ‘long-lasting’ consequences of early life immune challenges is one that has a large, and growing body of evidence in the literature that report physiological and behavioral similarities to autism. We also have recent evidence that hospitilization for viral or bacterial infection during childhood is associated with an autism diagnosis. There is, of course, a liberal sprinkling of ‘mays’, ‘propose’, and ‘conceivably’ caveats in place here.
Earlier I mentioned that the authors studied gene networks in addition to single gene expressions., and that some of those findings were very significant. The results of this are found in Table 2. In one discussion, I had it pointed out to me by ScienceMom that it appeared that some of the networks were not found to be statistically significant (and ergo we should not necessarily assume that immune dysfunction was a participant in autism). [If you look at Table 2, some networks like a p value of 0000]. I decided to use the data in this paper for another project that isn’t ready yet, but in that process I was able to speak directly with one of the authors of this paper. I asked him about this, and he told me that this was a function of space limitations; all of the gene networks described were found to be statistically signficant, but in some instances there wasn’t enough space to typeset the p value. In fact, some networks were found to be differentially expressed with a p-value of .000000000000001. (!!!!!!!!) That isn’t a value that you see very often.
I recently got a copy of Mitochondrial dysfunction in Autism Spectrum Disorders: cause or effect, which shares an author with this paper, Persico. In that paper, they reference Immune Transcriptome Alterations In the Temporal Cortex of Subjects With Autism, invoking a potential cascade effect of prenatal immune challenge, inherited calcium transport deficiencies, and resultant mitochondrial dysfunction that could lead to autism. I’ve generally stayed away from the mitochondria stuff in the discussion realm; even though I think it is probably somewhat important to some children, and critically important to a select few children, I’ve mostly found that the discussion of mitochondrial issues is comprised of two sets of people talking past one another so as to prove something, or disprove something about everyones favorite environmental agent; but this is a neat paper that I’d like to get to eventually.
– pD
passionless Droning about autism 