I’m currently at the Autism brain research meeting in San Diego, having finally, finally made it through immigrational control and customs in Los Angeles. Two people to check a jumbo jet’s worth of passengers. Possibly not enough. That’s all I’m saying.
So far, the meeting has been pretty interesting. Day 1 was focusing mostly on genetics, animal models of autism, and neuropathology with a little bit on brain imaging. Not really stuff that I’m over familiar with but a fascinating learning experience nonetheless. Here's a brief(ish) summary:
 |
| I think I'm currently in one of those tall buildings. Not sure which. |
So far, the meeting has been pretty interesting. Day 1 was focusing mostly on genetics, animal models of autism, and neuropathology with a little bit on brain imaging. Not really stuff that I’m over familiar with but a fascinating learning experience nonetheless. Here's a brief(ish) summary:
Genetics
The main theme of the early sessions was genetics. Recent research seems to suggest that we need to think about genetic causes of autism as a combination of (a) rare genetic variations that have large effects (i.e., if you’ve got the allele or mutation then you’re at high risk of autism) and (b) more common variations that each carry a low risk.
Consistent with this, Bernie Devlin pointed out that the search for common genetic factors has been largely unsuccessful. He highlighted three recent high-profile studies that have each identified genetic loci associated with autism in their sample. However, there was no overlap between any of the studies, so none of the loci have been validated. The problem, he suggested, is that these studies only had the power to detect genetic variation that was both common and carried a high risk of autism. And so much larger studies would be needed to identify genes that carried a lower risk.
Indeed, in the previous talk, Steve Scherer noted that evolution would be expected to select against common high risk genetic factors that, by causing autism, would reduce an individual’s likelihood of reproducing. So perhaps it’s not too surprising that none have been found. However, Scherer suggested that various rare genetic events might all have common effects by acting upon the same neurodevelopmental pathways and that looking at these functional groups of genes is the way forward.
Joseph Buxbaum added a further caveat – that different genetic variations in the same loci could have positive, negative, and neutral effects. The solution, he argued, was to look at the distribution of outcomes rather than just the overall relative risk.
The highlight of the session, for me, was a presentation by Catalina Betancur. She reported a study of 180 families in France with at least one autistic child. Via a combination of clinical evaluations, metabolic tests, and genetic screening, her team were able to identify a specific genetic disorder in 32 (18%) of the families. These included Fragile X, Angelmann, Di George, Timothy, Smith-Magenis, and Jacobsen syndromes. Many of these disorders are linked to intellectual disability and epilepsy, but none of them necessarily lead to autism. She concluded that autism is not a single disorder but the behavioural manifestation of many different genetic disorders – only some of which we are currently able to easily identify. However, as Dr Betancur pointed out, in those cases where a genetic etiology can be determined, this may have implications for the specific treatments and therapies offered.
Animal models:
Session 2 focused on animal models of autism. First up was Jacqueline Crawley, who talked about her group’s research investigating behavioural analogues of autistic impairments in genetically engineered mice. Her focus was more on the methods rather than specific genes. However, in each case, she provided evidence that mice with mutations linked to autism showed evidence of abnormal behaviour. Social impairments were measured by seeing whether a mouse would choose to approach a cage containing another mouse or an empty cage, or by recording social behaviour such as sniffing and climbing over and under another mouse. Communication was indexed by scent marking behaviour and by recording ultrasonic vocalizations in different social contexts (e.g., when an intruder mouse was added to the cage). Although, as one questioner pointed out, some of these associations may reflect the influence of genetic mutations on sexual behaviour rather than social behaviour. Most convincing, perhaps, were the measures of sterotyped behaviour, particularly the ability to learn and then re-learn a simple maze, which seems to correspond more directly to the cognitive flexibility impairments in autistic humans.
Presentations then followed from Tom Jongens (models of Fragile X syndrome in fruit flies) and Joseph Buxbaum (mouse models of the SHANK3 deletion, linked to autism). Finally, Alcino Silva talked about genes related to tuberous sclerosis, neurofibromatosis, and schizophrenia in mice. The common theme across the three presentations was that knowledge of the neurodevelopmental pathways affected by these genetic manipulations led to biological treatments that could ameliorate at least some of their behavioural consequences. These studies offer the prospect of pharmacological treatments for the corresponding disorders in humans, although they are still some way off. And, as the genetic studies presented earlier showed (and Dr Silva acknowledged), each treatment is only likely to be relevant to at best a small proportion of the autistic population.
I must admit to having had some skepticism about the relevance of animal models. Does it make sense to think of autistic mice? Or even fruit flies? Some of the speakers stressed that we mustn’t anthropomorphise animal behaviour. But then they kind of all did anyway. The problem for me is trying to define what is an equivalent behaviour in humans and non-humans. In fact, the most interesting animal research I heard about was a poster presented later on, by Mike Gandal and colleagues. Their approach is to develop biomarkers for autism that can be used with little alteration in both humans and animal models. The poster focused on brain responses to auditory stimuli, particularly the synchronization of gamma oscillations. They showed reduced gamma synchronization, both in autistic children (measured using MEG) and in mice (measured via intra-cortical electrodes) that were administered valproic acid prenatally. The connection here is that mothers who are forced to take valproic acid during pregnancy to counteract their epilepsy are at much higher risk of having a child with autism.
Neuropathology:
The afternoon session began with Eric Courchesne, who has previously reported cross-sectional data showing an increase in brain size early in post-natal development in autism. This increase is particularly noticeable in the dorsolateral prefrontal cortex (DLPFC). Here, he presented results from detailed studies of neural architecture in the DLPFC in the brains of seven autistic children who had died between the ages of two and fifteen years. Compared with control brains (N=6), the autistic brains had an estimated 68% increase in the number of neurons in DLPFC, as well as abnormal clustering, irregular spacing, and disorganization of the cortical layers. There was also much more patchy expression of various genes in the different cortical layers of the autistic brains. Inevitably (and thankfully), the sample size was small. The age range was also large and it wasn’t clear what had caused these tragic deaths or whether that might have been related to the neuropathology rather than it being autism-related. So we have to be cautious about these (nevertheless intriguing) results.
Later in the day, Cyndi Schumann presented a further series of neuropathological studies, this time focusing on the amygdala. Again, there is cross-sectional data suggesting initial amygdala overgrowth in young children with autism. Her sample was somewhat older (10-44 years) and she took care to exclude people with a history of seizure disorder that might have confounded results. Although there were no differences between autism and control brains in terms of overall volume, there were significant reductions in the number of neurons in various nuclei of the amygdala. She and her colleagues hypothesized that there may be degenerative processes at work, leading to loss of cells. Consistent with this, they found on average an increased number and size of microglia, indicating some form of neuroinflammation. However, there was considerable individual variability and the predicted inverse correlation between microglial number and neuron number wasn’t found.
Following on from that, Patrick Hof presented work on von Economo neurons – these are large neurons that have so far only been found in large-brained social mammals, including humpback whales, elephants, gorillas, and hippos (as well as humans). In most of these species, they are only found in the anterior cingulate and frontal insular cortex. Von Economo neurons are depleted in fronto-temporal dementia (a degenerative disorder related to inappropriate social behaviour), and have an abnormal distribution in Riley-Day syndrome (a disorder of the autonomic system). Somewhat surprisingly, the neuropathological study conducted by Dr Hof showed an increase in number in autism and a broader distribution. What this means wasn’t entirely clear to me.
Brain imaging:
Surprisingly, there were only two talks on brain imaging (and none tomorrow). Karen Pierce talked about fMRI studies of children as young as one year old, identified by paediatricians as possible cases of autism, language disorder, or developmental delay. The only way this kind of study is possible is to test the children while they are sleeping. Remarkably, playing them speech still leads to sensible responses in language-related regions. In the first study of its kind, Redcay and Courchesne found that children with autism had reduced activation in left temporal lobe but increased right hemisphere activation. And in fact the autistic children with greatest right hemisphere activity had the best language skills, suggesting perhaps that their right hemisphere is recruited to compensate for left hemisphere problems. A more recent study by Pierce and colleagues found that in non-autistic children, playing speech samples including the child’s name increased activation in the superior temporal sulcus. However, this was not true in autistic children. As Dr Pierce noted, failure to respond to their own name is a common sign of autism in young children. While still in its early days, this research has the potential to tell us a lot about early brain development in autism. My only slight concern is that, given the prevalence of sleep abnormalities in autism (e.g., reduction in REM sleep), it may be important to somehow monitor what stage of sleep the kids are in when the brain scan is being conducted.
The second brain imaging talk was presented by Declan Murphy, who gave a whistlestop tour of a number of recent and ongoing MRI studies with autistic adults. Using traditional structural MRI, his research group have shown differences in temporal, occipital, and frontal lobe grey matter volume. Using magnetic resonance spectroscopy, they’ve shown reductions in glutamate concentrations in the basal ganglia. Using diffusion MRI, they’ve found differences in the arcuate fasciculus - the white matter pathway linking language regions of the left hemisphere. There was more, but I wasn’t writing fast enough to get it all down. Murphy’s aim is to be able to use some of these measures to assist in diagnosis. I’ve previously blogged about a study this group conducted using MRI to diagnose autism and, as various other people did at the time, suggested that there are many hurdles to overcome before that becomes a reality. Prof Murphy was careful to emphasise that they are still very much at the proof of concept stage, although he was confident that objective brain-based diagnoses are ultimately achievable.
There were a couple more studies at the end of the day by Drs Blatt and Pardo, but sadly I can't make much sense of my notes. Will try and do a better job tomorrow!
PS: If there's anything here that you notice is wrong (and there almost certainly will be), please comment and I'll make the corrections. As I say, this is all pretty much new to me!
Update (18/03/11):
Many (if not all) of the presenters at the meeting contributed to a special issue of the journal, Brain Research dedicated to The Emerging Neuroscience of Autism.
In case it wasn't obvious, I blogged Day 2 here.
Update (18/03/11):
Many (if not all) of the presenters at the meeting contributed to a special issue of the journal, Brain Research dedicated to The Emerging Neuroscience of Autism.
In case it wasn't obvious, I blogged Day 2 here.