Tuesday, 17 May 2011

Immunophenotypes and autism

I knew it. Once I started blogging about the immune findings potentially connected to autism I would have difficulty stopping and indeed it has happened as predicted. It was the same when I got my very first Sinclair ZX Spectrum computer all those years ago: '10 PRINT "HELLO", 20 GOTO 10' - once bitten by the computer bug, forever smitten.

I do however promise that this blog with still continue to look at as wide a variety of research as possible so please read on confident that if this is not your cup of tea, soon there will (hopefully) be something.
The paper in question for this post is this one from that research behemoth, the UC Davis MIND Institute.

The title: In search of cellular immunophenotypes in the blood of children with autism published in PLoS One. I was drawn to this post because: (i) it is an interesting paper and (ii) some of the authors also lead on this paper including a technique very close to my heart (TOF LC-MS).

The title is perhaps more complicated than the study implies: look in blood samples taken from children with autism and controls for various immune cell populations and see if you can spot a 'signature' profile which might define one or other group and differentiate them. Pretty much the same concept as that wonderful 'metabolomics' field which I discussed previously.

OK the study did get pretty complicated but together we will go through the main findings.

The basics first: initially 2 groups: autism (n=70) and age-matched controls (4-6 years old) not presenting with autistic features (n=70). The autism group were subsequently divided into 'low' and 'high' functioning on the basis of IQ measures. There was a lot of diagnostic work-up on the participants to ensure that things like co-morbidities linked to immune dysfunction were minimised as much as possible.

5ml of blood were collected from each participant and subject to laser scanning cytometry, basically a very sophisticated counting machine (the molecular biologists might not like my simplification) coupled with some analysis of the intensity of antigens.

Alot of counting of various immune cell populations was done and then the results: depending on the significance level, anywhere between 21-151 markers were different between autism and controls. A smaller number of markers showed differences between the 'low' and 'high' functioning groups (between 1-33). Absolute numbers of B cells and NK (natural killer) cells were reported significantly higher in autism vs. controls. A few other types of cell populations were almost significantly different, including our old friends the eosinophils (p=0.066). There were a few other interesting results, not least the intensity of cell surface markers expressed on immune cells such as CD26/dipeptidyl peptidase-IV on CD8 T cells in the autism group over controls. Some might have heard about CD26/dipeptidyl peptidase-IV from some suggestion a while back on a possible relation to autism and peptide chemistry.

There is quite a bit of overlap from these results and the research literature so far although not all of it is in the same direction. NK cell activity in this paper for example, was lower in autism. Even the late Reed Warren and his early results suggested a similar pattern. The current paper explains such differences in terms of the younger age group included and methodological/analytical differences. The fact that the current participant group were so tightly controlled in terms of co-morbidity might also be a factor.

Given that the aim of this blog is to question answers, I have to say that I am struggling with this one. I could perhaps argue that an additional control group made up of children diagnosed with non-specific learning difficulties (without autism) might have given an additional dimension to the study. One could also suggest that a similar analysis based on either the same participant group 3-5 years down the line when diagnoses potentially become a little less unstable might also be a good idea or failing that an independent adult participant group. Finally I can't seem to find anything in the paper about participant use of medication, diets, etc (whether this would have an effect, I don't know). These are but observations, and I don't want to unduly criticise just for the sake of it.

I probably have not done this new paper justice in terms of my summarising such an extensive piece of work into a short blog description. I do however tip my hat and say to rack up another one for the MIND Institute (who I have heard have just received quite a large pot of money to start looking at gastrointestinal co-morbiditiy in autism).