Alterations in Lipid Metabolism and Anti-Oxidant Status as Specific Biomarkers of Autism Plasma

Background: Our initial findings using a Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTMS) based comprehensive non-targeted metabolomic platform revealed alterations in the levels of very long chain fatty acid (VLCFA) containing phosphatidylethanolamines (PtdEtn) and in DHA containing plasmalogens (PlsEtn) in autism plasma (unpublished results).

Objectives: To validate and expand these findings, we investigated more thoroughly fatty acid elongation, DHA synthesis, and plasmalogen synthesis using serum from autistic subjects. We measured the plasma levels of ethanolamine phospholipids containing fatty acids ranging from 16 to 40 carbon units and from zero to six double bonds. We also investigated markers of glutathione pathway in order to investigate a possible correlation between lipid metabolism and oxidative stress in autism.

Methods: Three plasma samples from 15 autism subjects and 12 non-autism controls were collected over a 12 month period (six month interval between samplings).   Phospholipid and thiol levels were determined by tandem LC-MS/MS analyses using a linear ion trap mass spectrometer (4000 Q TRAP, Applied Biosystems) coupled with an Agilent 1100 LC system. 

Results:

Plasma levels of 136 phosphatidyl ethanolamines and 15 ethanolamine plasmalogens were measured and compared to levels observed in non-autistic subjects. The results of these analyses revealed that autistic subjects had increased levels of fatty acid synthesis, elongation, and desaturation products relative to controls. Every non carnitine-supplemented autistic subject exhibited elevated plasma levels (p<0.05) of either docosahexaenoic acid containing ethanolamine plasmalogens (DHA-PlsEtn) or very long chain fatty acid containing phosphatidylethanolamines (VLCFA-PtdEtn), whereas all autistic subjects taking carnitine supplementation had normal levels of DHA metabolites and 2/4 had slightly elevated levels of VLCFA metabolites.  Since over half of the ethanolamine phospholipids in the membrane of neurons are plasmalogens and DHA plays an essential role in the membrane biophysical properties, the abnormalities we observed in the non carnitine supplemented autistic subjects are expected to alter neuron plasticity and therefore neurotransmission. An overall increase in plasmalogens (increase in some specific plasmalogens without a decrease in others) can be hypothesized to explain the hypersensitivity frequently observed in autism.

In a second step, we found that all autistic subjects had decreased levels of reduced glutathione (GSH) and metabolic precursors of GSH, regardless of their supplementation status.  Basal levels of carnitine and acetyl carnitine were normal in non-supplemented subjects and elevated in carnitine supplemented subjects. The decreased anti-oxidant capacities we observed in all autistic subjects confirm the results reported by other groups and delineate the limitations of carnitine supplementation in autism.

Conclusions:

The serum of autistic subjects displayed markers of increased oxidative stress, in association with a specific phospholipid pattern for the non carnitine supplemented subjects. It is our hypothesis that the phospholipid pattern arises from a chronic exposure to glutamate as we discuss in our second abstract (in Model Systems). These results raise the promising possibility of diagnosing autism by a simple plasma draw.