Organic acids as biomarkers of autism?

Whilst I am always a little cautious about the use of the word 'biomarker' when applied to a heterogeneous condition like autism, even the autisms, I am nevertheless always intrigued at any reasonable prospect reported in the scientific literature. So it was when I read the paper by Joanna Kałużna-Czaplińska and colleagues [1] and their assertion that "there is a significant metabolic difference between autistic and non-autistic children" and onwards that "21 metabolites were identified as potential biomarkers".

Let me expand on this a little...

• This was a small study looking at potential biomarker identification on the basis of the analysis of urine samples via gas chromatography-mass spectrometry (GC-MS). If you want some further background on this technique applied to autism research, have a look at a previous post (see here) where it has been utilised. Overnight urine samples from 14 children (aged 4-10 years) diagnosed with an autism spectrum disorder (ASD) undergoing "rehabilition" (whatever that means) were analysed in comparison to samples from 10 asymptomatic controls.
• Quite a bit of information is included about sample treatment and the analytical method. Each sample result was represented as a TIC (total ion count) and, as is often the case with such methods, data processing was an important part of the analysis. Most compounds were identified by cross-referencing with the NIST mass spectra library and via fragmentation patterns. Principal component analysis (PCA) was "applied to check the dataset structure and assess the variability of the profiles belonging to groups of autistic vs. non-autistic children". 
• Results: as indicated, 21 metabolites were deemed as "potential marker metabolites" some detected in higher quantities in the autism samples, and some lower. Fourteen of these compounds were described as organic acids. Without hopefully breaking any copyright, I've attached a copy of the table included in the paper with all the compounds differing between autism and control samples. The eagle-eyed will also note the big 'H' - homocysteine - to be a part of that list, and as expected, elevations in urinary homocysteine for the autism group as per other work in this area (see here).
• Given the title of this post I'll point out a few organic acids which seemed to be important differentiators between autism and control samples: (i) levels of beta hydroxybutyric acid were elevated in autism sample. This compounds has been talked about previously on this blog with regards to inborn errors of metabolism and autism (see here). (ii) Hydroxybenzoic acid was again elevated and perhaps ties into other findings from this group [2] potentially indicative of intestinal dysbiosis. (iii) Succinic acid levels were also generally elevated, and as the authors point out: "is considered a potential marker for deficiency of CoQ10 and riboflavin in children with autism". Co-enzyme Q10 y'say? I could go on, but won't.
• Various statistical models (PCA) were applied to the datasets which led authors to find: "The group of samples from non-autistic control children [were] more homogeneous than the group from autistic children". Further: "There is a clear distinction between those two groups of samples". ROC analysis looking at the performance of the PCA models was also applied leading authors to conclude that there may be something in their results from a diagnostic point of view.

Obviously the Kałużna-Czaplińska results are preliminary and in need of further independent replication. I note that quite a bit of the other literature in this area of biomarkers tend to use both training and test sets, where training samples provide your initial compounds of interest and test sets do just that, test your biomarker assumptions (see here). This wasn't the case in the current study but still leaves the door open to independent verification. That also the word 'comorbidity' does not seem to be mentioned as part and parcel of the autism group means the questions of how widespread comorbidity was in the autism participant group and whether this might have exerted an effect on the results obtained are unanswered. I might also quibble about the way that peaks in the TIC were assigned a compound name: "Peaks with the similarity index more than 80% were assigned compound names..." but now I'm just nit-picking.

That all being said, I do see some promise in the results obtained by Kałużna-Czaplińska et al. I note in another paper by some of the authors [3] they talk about how probiotic therapy might impact on both some of the behavioural measures of autism and also levels of one of the compounds picked up in their latest analysis, D-arabinitol. Again, I'd like to see more research done on this, alongside their other suggestion on the use of B vitamins (and magnesium) potentially affecting organic acids in cases of paediatric autism [4] talked about in a previous post (see here). The focus on the inner working of the gut, and particularly the trillions of gut bacteria which call us home, potentially being connected to some of these biomarkers, ties in well with an emerging autism research area (see here).

Music to close, and yet again my brood provide the inspiration as Bob Marley is fast becoming a YouTube favourite in our home with the classic One Love. You know you're getting old when your kids start listening to cooler music than you do...

----------

[1] Kałużna-Czaplińska J. et al. Identification of organic acids as potential biomarkers in the urine of autistic children using gas chromatography/mass spectrometry. Journal of Chromatography B. 2014. Feb 2.

[2] Kałużna-Czaplińska J. Noninvasive urinary organic acids test to assess biochemical and nutritional individuality in autistic children. Clin Biochem. 2011 Jun;44(8-9):686-91.

[3] Kałużna-Czaplińska J. & Błaszczyk S. The level of arabinitol in autistic children after probiotic therapy. Nutrition. 2012 Feb;28(2):124-6.

[4] Kałużna-Czaplińska J. et al. B vitamin supplementation reduces excretion of urinary dicarboxylic acids in autistic children. Nutr Res. 2011 Jul;31(7):497-502.

----------

Kałużna-Czaplińska J, Zurawicz E, Struck W, & Markuszewski M (2014). Identification of organic acids as potential biomarkers in the urine of autistic children using gas chromatography/mass spectrometry. Journal of chromatography. B, Analytical technologies in the biomedical and life sciences PMID: 24565890