M. Andreina Mendez, Jamie Horder, Nicola Gillan, Suzanne Coghlan, Declan Murphy

Abnormalities In Subcortical Glutamate/Glutamine, But Not GABA, In Adults With An ASD

A [1H]MRS Study

M. Andreina Mendez, Jamie Horder, Nicola Gillan, Suzanne Coghlan, Declan Murphy

Background - General

● GABA / glutamate balance implicated in ASD?

Background - General

● GABA / glutamate balance implicated in ASD?• Rodent models

• e.g. reduced GABA-A receptors, increased glutamate mGluR receptors (subtype-specific) (Selby et al 2007)

Background - General

● GABA / glutamate balance implicated in ASD?• Rodent models

• e.g. reduced GABA-A receptors, increased glutamate mGluR receptors (subtype-specific) (Selby et al 2007)

• Human Postmortem Studies (Blatt & Fatemi 2011)

Background - General

● GABA / glutamate balance implicated in ASD?• Rodent models

• e.g. reduced GABA-A receptors, increased glutamate mGluR receptors (subtype-specific) (Selby et al 2007)

• Human Postmortem Studies (Blatt & Fatemi 2011)

• Genetics – ASD CNVs include GABA-A receptors e.g. 15q11-13 (Urraca et al 2013)

Background - General

● GABA / glutamate balance implicated in ASD?• Rodent models

• e.g. reduced GABA-A receptors, increased glutamate mGluR receptors (subtype-specific) (Selby et al 2007)

• Human Postmortem Studies (Blatt & Fatemi 2011)

• Genetics – ASD CNVs include GABA-A receptors e.g. 15q11-13 (Urraca et al 2013)

• Comorbidities – epilepsy common in ASD (20% vs 2% in non-ASD) (Kohane et al 2012)

Background - PET

• Positron Emission Tomography (PET) can measure GABA-A receptor density

• We showed preliminary evidence of reduced GABA-A alpha5 (Mendez et al 2013)

• But PET can’t measure neurotransmitters…

[11C]Ro154513 VT parametric images, averages of 3 healthy adult males and 3 ASD males. This view illustrates lower average binding in the nucleus accumbens, and other regions

0 4 8 12 16VT

HEALTHY ASD

Background - MRS

● Proton magnetic resonance spectroscopy (MRS) can measure GABA, glutamate and glutamine (Glx)

● Previous studies have found evidence of glutamate/glutamine (Glx) abnormalities….

But how about GABA?

Methods - GABA/Glx MRS

Left striatum35 x 30 x 25 mmFunction: action and movement control

Bilateral midline dorsal ACC

25 x 40 x 30 mm

Function: cognitive control, error monitoring, executive

Participants

● n = 22 adult males with a diagnosis of ASD and IQ>70 and 27 control males matched on verbal, performance and full-scale IQ (t-test, all p>0.4, mean FSIQ 119 vs. 116.

● Diagnosis by expert clinical judgement using ICD-10-R criteria based on ADOS Module IV + ADI-R (where possible)

● Exclusion criteria included:

• Epilepsy• Genetic disorder• Psychosis• Substance dependence.• Any psychoactive medication at the time of scanning, within previous six

weeks minimum.

Results #1 – MRS

We found a significant reduction in Glx in the left striatum (p = 0.033 two-tailed).

However, no differences were seen in Glx in the DMPFC, or in GABA in either voxel.

Therefore, replicating previous findings using PRESS MRS (Horder et al 2013), we showed reduced Glx in the basal ganglia region in adults with ASD.

(Previous Findings

Discussion #1: MRS – fMRI

● Might a reduction in striatal Glx have functional implications?● We used resting state fMRI to examine the correlation between

[1H]MRS Glx signal and the functional connectivity to/from the basal ganglia region (putamen) as a seed (seed-voxel correlations):

Results #2: MRS-fMRI association in all participants (n=42)

A significant (p < 0.05, whole-brain cluster corrected) effect of striatal Glx on the pattern of connectivity with the left putamen was found. Areas of significant differences included the anterior cingulate cortex.

Inspection of the scatterplot revealed that this effect was not driven by outliers.

Results #3: Glx vs. Connectivity: HV vs. ASD

The correlation between left basal ganglia Glx and left putamen connectivity was stronger (p<0.05 whole brain corrected) for healthy controls than for ASD patients in two regions:

• Medial prefrontal cortex

• Lateral temporal lobe including STG and MTG.

Conclusions

● Subcortical (striatum) glutamate/glutamine is reduced in adults with ASD

● However, we found no differences in GABA in either the cortex or the subcortex

● Contrasts with prior evidence and theory?● Or understandable in terms of multiple ‘pools’ of brain

GABA, only some of which may be abnormal?

Detailed Methods

fMRI

● Instructions were to remain awake with eyes open and fixate on a static cross on a dark background

● 256 volumes, TR = 2000 ms. Analysed in FSL v 4.18. Preprocessing: MCFLIRT motion correction; registration and normalization to MNI 2mm; 5 mm smoothing.

● EVs of no interest: 6 motion parameters + pulse and breathing (recorded in scan) + mean signal over all grey, white, and CSF voxels.

MRS

● J-edited MEGAPRESS [1H]MRS at 3T was used, TE = 68 ms.

● Unsuppressed water signal was acquired and used for absolute quantitation

● Voxel grey, white and CSF composition was calculated from T1 structural MRI. Water-scaled metabolite concentrations were corrected for CSF.

● Data analysed in LCModel software. Unedited spectra were used to quantify Glx while J-edited ON/OFF subtraction spectra were analysed to measure “GABA+” (GABA + macromolecules).

Acknowledgements

Prof Gareth BarkerChristine EckerRichard EddenGrainne McAlonanRuth O’GormanProf Steve WilliamsEileen DalyDave LythgoeEllie WilsonJessica Faulkner

EU-AIMS Consortium

Medical Research Council (MRC)

National Institute for Health Research (NIHR)

Wellcome Trust

For more information, please contact: maria.mendez@kcl.ac.uk