5
Supplementary Methods
MRI Acquisition. Scanning was performed at the Northwestern University Center for Advanced Magnetic Resonance Imaging using a 3-T Siemens whole body scanner (Siemens Trio Tim B15 software, Elangen, Germany) with a 12-channel head coil. The imaging protocol included structural MR and DTI. Structural MRI: High-resolution 3D T1-weighted MRI scans were acquired using a magnetization-prepared rapid gradient echo sequence (MPRAGE) (TR= 2,300 ms, TE =2.91 ms, flip angle = 9°, field of view 256*256 mm, voxel dimension= 1 mm isotropic, 176 slices and acquisition time = 10 min). DTI: 2D spin-echo EPI acquisition (SE-EPI-DTI) utilizing a pair of diffusion weighted gradients (Basser et al., 1994). A b=0 reference image and 64 diffusion-weighted images with a b-value of 1000 sec/mm² were acquired at each slice location (FOV: 130 mm, matrix size: 1.97×1.97×2.0 mm , TR= 9600ms), bandwidth ±1325 Hz. The standard Siemens 12-channel coil was used. Echo planar readout, accelerated with sensitivity encoding, was performed, and all images were reconstructed to a 256x256 matrix after zero filling. The duration of the SE-EPI-DTI scan was 10 minutes and 54 seconds. Eddy-current-induced distortions were corrected in the SE-EPI-DTI images by registering all DW images to the B0 image for each slice.
DTI Parametric Maps. This investigation acquired measurements for FA and for MD. Pixel-by-pixel FA and MD parametric maps were derived using custom software on a Linux workstation based on standard equations.
DTI parameters: From the eigenvalues of D, λ1, λ2, λ3, the trace of the diffusion tensor can be estimated, which is proportional to the mean diffusivity: Trace = λ1 + λ2 + λ3 and the MD, or Mean Diffusivity=Trace/3. Fractional anisotropy (FA), which quantifies the directional diffusion, can be estimated from the eigenvalues of D (as shown). Fractional Anisotropy and Mean Diffusivity can be estimated in each voxel and then summarized for specific regions, volumes of interest or calculated for the entire brain.
Automated Atlas Based Analysis. Images were transferred to a Linux workstation for post-processing. Brain volumetric measurements were derived using image processing tools requiring minimal operator input. The image analysis tool, SIENAX (Oxford University, Oxford, England) was used to calculate normalized volumetric measures of gray matter, white matter, and ventricular CSF within the intracranial cavity.(Smith et al, 2002)
In order to produce region of interest (ROI) measurements without manual placement, an automated segmentation algorithm was incorporated using Freesurfer.(Fischl et al, 2002) We derive comprehensive measurements of over 50 discrete structures and neuroanatomical landmarks. The volume for each structure was divided by the volume of the individual intracranial cavity in order to adjust for individual differences in brain size. Segmentation algorithms, which are indispensable for collapsing brain tissue into broad constituent tissue classes of gray matter, white matter or CSF, are susceptible to wide spectrum signal intensity variation. This potentially introduces considerable overlap between tissue classes. The fully automated Freesurfer algorithm (Martinos Center for Biomedical Imaging, Charleston, MA) was used to derive comprehensive measurements of specific neuroanatomical landmarks, structures and brain regions including critical subcortical structures that may be difficult to evaluate otherwise.(Fischl et al, 2002) This obviates the necessity for labor-intensive, manual outlining across multiple brain images by expert operators. This algorithm uses an affine rigid linear transformation, combining voxel intensity information relative to a probability distribution for tissue classes and information concerning the spatial relationship of the voxel to location of neighboring structures from an integrated a priori anatomic atlas.
FA and MD maps were coregistered to segmented brain structural masks.(Wu et al, 2012) For DTI, the following regions were included in this study: corpus callosum, whole brain white matter, left and right caudate, cerebral cortex, cerebral white matter, hippocampus, putamen, and thalamus proper. Mean FA and mean MD were calculated as the average value of all voxels within a segmented anatomical structure in three dimensions, in which consecutive slices cover the observed region of interest.
Neuropsychological Assessments. The neuropsychological test battery is based on similar batteries in HIV clinical trials and neurological outcome studies.(Butters et al, 1990; Sevigny et al, 2004) Specific cognitive tests included the Rey Auditory Verbal Learning Test (verbal memory), ReyOsterrieth Complex Figure Recall Test (visual memory) and ReyOsterrieth Complex Copy Test (constructional skills),(Rey, 1941) as well as the Letter-Number Sequencing Test (verbal memory).(Wechsler, 1997a; Wechsler, 1997b) Frontal systems were assessed with Verbal Fluency,(Benton and Hamsher, 1976) OddManOut,(Flowers and Robertson, 1985) and Trail Making(Partington and Leiter, 1949) tests. Psychomotor skills were assessed using the Digit Symbol Test(Wechsler, 1981) and motor skills were assessed using Grooved Pegboard(Klove, 1963) and Timed Gait(Robertson et al, 2006) tests. Reaction times were measured using the California Computerized Assessment package (CALCAP).(Miller et al, 1991) Neuropsychological measures were summed across trials and variations of a test.
Supplementary References
Benton AL, Hamsher KD (1976). Multilingual Aphasia Examination. University of Iowa: Iowa City.
Butters N, Grant I, Haxby J, Judd LL, Martin A, McClelland J, Pequegnat W, Schacter D, Stover E (1990). Assessment of Aids-Related Cognitive Changes: Recommendations of the NIMH Workshop on Neuropsychological Assessment Approaches. Journal of Clinical & Experimental Neuropsychology: Official Journal of the International Neuropsychological Society 12: 963-978.
Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, van der Kouwe A, Killiany R, Kennedy D, Klaveness S, Montillo A, Makris N, Rosen B, Dale AM (2002). Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 33: 341-55.
Flowers KA, Robertson C (1985). The effect of Parkinson's disease on the ability to maintain a mental set. J Neurol Neurosurg Psychiatry 48: 517-29.
Klove H (1963). The Medical Clinics of North America. The Medical Clinics of North America: New York.
Miller EN, Satz P, Visscher B (1991). Computerized and conventional neuropsychological assessment of HIV-1-infected homosexual men. Neurology 41: 1608-16.
Partington JE, Leiter RG (1949). Partington's Pathway Test. The Psychological Service Center Bulletin 1: 9-20.
Rey A (1941). L'examen psychologique dans les cas d'encephalopathie traumatique. Arch Psychologie: 28286-340.
Robertson KR, Parsons TD, Sidtis JJ, Hanlon Inman T, Robertson WT, Hall CD, Price RW (2006). Timed Gait test: normative data for the assessment of the AIDS dementia complex. Journal of Clinical & Experimental Neuropsychology: Official Journal of the International Neuropsychological Society 28: 1053-64.
Sevigny JJ, Albert SM, McDermott MP, McArthur JC, Sacktor N, Conant K, Schifitto G, Selnes OA, Stern Y, McClernon DR, Palumbo D, Kieburtz K, Riggs G, Cohen B, Epstein LG, Marder K (2004). Evaluation of HIV RNA and markers of immune activation as predictors of HIV-associated dementia. Neurology 63: 2084-90.
Smith SM, Zhang Y, Jenkinson M, Chen J, Matthews PM, Federico A, De Stefano N (2002). Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 17: 479-89.
Wechsler D (1981). Wechsler Adult Intelligence Scale Revised. The Psychological Corporation: New York.
Wechsler D (1997a). Wechsler Adult Intelligence Scale-III. The Psychological Corporation: San Antonio.
Wechsler D (1997b). Wechsler Memory Scale. Third edition manual. The Psychological Corporation: San Antonio.
Wu Y, Du H, Storey P, Glielmi C, Malone F, Sidharthan S, Ragin A, Tofts PS, Edelman RR (2012). Comprehensive brain analysis with automated high-resolution magnetization transfer measurements. Journal of Magnetic Resonance Imaging 35: 309-17.
Subgroups / Group 1 / Group 2 / ANOVA / Planned t-tests
Viral Load / Aviremic / Viremic / 0/1 / 0/2 / 1/2
MMP-1 / 1,303 ± 966 / 1,147 ± 949 / 0.659 / 0.311 / 0.567 / 0.638
MMP-2 / 38,208 ± 10,212 / 42,890 ± 14,081 / <0.001** / <0.001** / <0.001** / 0.314
MMP-7 / 5,770 ± 2,484 / 4,775 ± 2,330 / 0.253 / 0.081† / 0.542 / 0.230
MMP-9a / 33,255 ± 17,756 / 44,611 ± 44,071 / 0.998 / 0.890 / 0.886 / 0.887
MMP-10 / 655 ± 484 / 509 ± 291 / 0.237 / 0.242 / 0.497 / 0.229
MMP measurements are in units of pg/mL.
aKruskal-Wallis and Mann-Whitney tests used for MMP-9.
**Significant at the 0.01 level. *Significant at the 0.05 level. †Nearly significant (p < 0.10).
Table S2: MMP Correlations with Brain Tissue Classes
Plasma / CSF
SIENAX / MMP-1 / MMP-2 / MMP-7 / MMP-9 / MMP-10 / MMP-2 / MMP-9
Cortical Gray Matter / 0.178 / 0.136 / -0.116 / -0.054 / -0.062 / -0.417 / 0.317
Gray Matter / 0.214 / 0.136 / -0.163 / -0.002 / 0.043 / -0.417 / 0.433
White Matter / -0.118 / -0.080 / 0.129 / 0.043 / -0.139 / 0.150 / -0.167
Ventricular CSF / 0.021 / -0.067 / 0.119 / 0.135 / -0.141 / 0.200 / -0.200
Total Brain Volume / 0.065 / 0.055 / -0.040 / 0.058 / -0.058 / 0.083 / 0.333
Pearson correlation coefficients (Spearman used for MMP-9 and CSF).
**Significant at the 0.01 level. *Significant at the 0.05 level.