HIV infection is associated with structural deficits of the basal ganglia (BG). Volume deficits may be related to changes in cerebral metabolites measured with MRS. Altered relationships between metabolite levels and BG volumes may provide a clearer picture of the effects of HIV on the brain than one measure alone. However, relationships between these measures in healthy children are not known. Previous studies in HIV-infected adults have found lower BG N-acetyl-aspartate (NAA) to be associated with reduced temporal lobe⁽¹⁾, putamen, and thalamus volumes⁽²⁾. Lower Glutamate/Glutamine (GluGln) has been associated with reduced cortical gray matter (GM), caudate and thalamus volumes⁽¹⁾, but also with increased putamen and thalamus volumes⁽²⁾. Our aims were to investigate the relationship between NAA/GluGln and BG volumes in healthy children and to determine whether these relationships are altered in HIV-infected children on antiretroviral therapy (ART).

Participants were 104 isiXhosa children from a neuroimaging follow-on study of the Children with HIV Early Antiretroviral Therapy (CHER) trial. All HIV-infected children had started ART before 18 months of age and were stable on ART by the time of scanning at age 7.

Scanning was performed on a 3T Allegra (Siemens, Erlangen, Germany). The protocol included a high-resolution T1-weighted 3D EPI-navigated⁽³⁾ multiecho magnetisation prepared rapid gradient echo (MEMPRAGE)⁽⁴⁾ (FOV 224 x 224 mm², TR 2530 ms, TI 1160 ms, TE’s = 1.53/3.19/4.86/6.53 ms, bandwidth 650Hz/px, 144 sagittal slices, 1.3 x 1.0 x 1.0 mm³). Single voxel ¹H spectroscopy (SVS) in the BG was obtained using an EPI volumetric navigated point-resolved spectroscopy (PRESS) sequence with real-time first-order shim and motion correction⁽⁵⁾ (TR = 2000 ms, TE = 30 ms, 64 measurements, vector size 512, spectral bandwidth 1000 Hz). Water unsuppressed ¹H MRS measurements were acquired for water scaling and eddy current correction. Scans were performed according to protocols approved by the Human Research Ethics Committees of the Universities of Cape Town and Stellenbosch.

FreeSurfer version 5.1.0 was used to perform automated segmentation of subcortical structures to generate regional volume measures. LCModel⁽⁶⁾ was used to calculate metabolite concentrations from frequency and phase-corrected water-suppressed spectra using water scaling. The SVS voxel was segmented using SPM12 to determine tissue type composition. SVS data were not acquired for 9 children, and spectra from a further 32 were excluded due to SNRs <7, leaving 63 children with both volume and SVS data (38 HIV-infected, 25 controls, age 7.2 ± 0.16 years, 26 boys).

Statistical analysis was performed using R. Subcortical GM, right and left caudate, putamen and nucleus accumbens volumes were regressed against NAA and GluGln concentrations, testing for an interaction of metabolite level with HIV infection status and accounting for age, gender and intracranial volume. Group differences in metabolites and volumes between HIV-infected children and controls were tested using ANCOVA controlling for the confounders.

In controls, NAA in BG was negatively associated with volumes of the right (slope=-180, p=0.008, figure 1) and left (slope=-143, p=0.03, figure 2) nucleus accumbens and left putamen (slope= -727, p=0.04). There was no significant interaction of HIV status. GluGln related positively to subcortical GM volume in controls only (slope=380, p=0.02), with a significant interaction effect of HIV status (p=0.04, figure 3). ANCOVAs controlling for age, gender and ICV indicated smaller right (p=0.01) and left (p=0.0008) nucleus accumbens, lef t putamen (p=0.009) and subcortical GM (p=0.01) in HIV-infected children. Higher mean NAA and GluGln in the HIV-infected children than controls was not significant (p=0.9, p=0.4 respectively).

Increased NAA – a marker of neuronal health – is associated with smaller volumes of nucleus accumbens and putamen in both HIV-infected children and controls. This suggests that smaller volumes reflect denser neuronal organization in both groups, consistent with previous studies indicating that NAA gradually increases during childhood⁽⁷⁾, while absolute volumes of subcortical structures including nucleus accumbens decrease during the same period⁽⁸⁾. However, it does not explain why in our sample the controls have larger nucleus accumbens and putamen volumes than HIV-infected children, unless this is due to accelerated maturation in the HIV-infected group that is only evident in the narrow age range of this sample.

The positive relationship of GluGln with subcortical GM volume seen in controls is not seen in HIV-infected children on ART, suggesting a possible effect of HIV/ART.

Measurement of metabolites as well as volumes may provide a more complete – and more complex – description of the effects of HIV/ART on the brain. Normal relationships between metabolites and BG volumes should be established to determine when the breakdown of these relationships may be an indicator of disease.