Honors Thesis - USC Dornsife - University of Southern California

Download PDF
0 downloads 0 Views 6MB Size Report
Comment
Apr 11, 2016 - for regions of the corona radiata were positively associated with ... (12-14) If treatment is not administered, cortical atrophy with ..... cingulate gyrus and rate of RD for the left superior corona radiata, ..... MRI of stroke recovery.
University of Southern California

Honors Thesis

Consequences of HIV Infection and Combination Antiretroviral Therapy on White Matter Microstructure during Childhood Brain Development

By Arvin Saremi

USC Imaging Genetics Center Dr. Neda Jahanshad Dr. Paul Thompson April 11, 2016



Saremi 2

Table of Contents Abstract ……………………………………....………………………………….. 3 Introduction …………………………………….………………………… …….. 4 Methods…………………………………………………………………………… 7 Participant Selection……………………………………....………………….

7

Clinical Assessment……………………………………....…………………..

8

Brain image acquisition……………………………………....……………….

9

Quality Control……………………………………....………………………

10

Image Processing……………………………………....…………………….

10

Statistical Analysis……………………………………....…………………...

11

Results…………………………………………...………………………………… 13 HIV Infection……………………………………....………………………...

13

Exposure……………………………………....…………………………….

13

Combination Antiretroviral Therapy………....………………………………..

13

IQ/index……………………………………....……………………………..

15

Discussion ………………………………………………………………………… 15 Acknowledgement………………………………………………………………… 19 Funding …………………………………………...………………………………. 19 References …………………………………………...……………………………. 20 Tables…………………………………………...…………………………………. 24 Figures …………………………………………...……………………………….. 29



Saremi 3

Abstract Vertical transmission of the human immunodeficiency virus (HIV) from mother to child is still a major public health problem in many countries. As reported in 2013, 3.2 million children under the age of 15 still live with HIV. In the 21 “global priority” countries, 66% of the HIV-infected children receive combination antiretroviral therapy (cART). However, the precise progression pattern of HIV affecting neural development has not yet been established, and little is known about how cART affects neurodevelopment in young children. In this study, we set out to examine whether brain abnormalities are detectable using neuroimaging - between HIVinfected and HIV-uninfected children - and if there are any changes associated with cART status through measuring and comparing diffusion in dominant white matter fibers, cross-sectionally and longitudinally. Our investigation is mainly exploratory in terms of the affected neurocircuitry, and we evaluate white matter integrity and maturation throughout the brain rather than focus on any particular systems, to ensure better understanding of neurodevelopmental and neuropsychiatric differences. We hypothesize that higher neural integrity should suggest more mature neural development and be positively correlated with better neuropsychological performance. T1-weighted and diffusion weighted images (DWI) of 92 children consisting of 39 HIV-infected (30 on cART before initial scan) and 53 HIV-uninfected children were selected from the Pediatric Randomized Early versus Deferred Initiation in Cambodia and Thailand (PREDICT) cohort study. Fractional anisotropy and diffusion maps were extracted from preprocessed DWIs using FSL tract-based spatial statistics (TBSS). Significantly higher baseline axial diffusivity (AD) and mean diffusivity (MD) were found in HIV-infected children in multiple brain regions, such as the superior longitudinal fasciculus, with no detectable differences in rate of change in the diffusion tensor measures. Significantly higher baseline AD was found in HIV-infected children on cART compared to ones without treatment in inferior



Saremi 4

fronto-occipital fasciculus and superior fronto-occipital fasciculus with a significantly lower rate of change in FA and higher rate of change in radial diffusivity (RD) for multiple regions, such as the corpus callosum and superior longitudinal fasciculus. Additional regions were significantly different for rate of change in FA and RD between on-cART and without-cART children when testing for associations with lifetime on cART (time on cART over age). Rate of change in FA for regions of the corona radiata were positively associated with neuropsychiatric scores. Results suggest slowed neurodevelopment in children on antiretroviral medication, which could be the foundation for future neurodevelopmental deficits. Continued longitudinal monitoring will help to determine if these findings are maintained through later stages of neurodevelopment.

Introduction Vertical transmission of the human immunodeficiency virus (HIV) from mother to child is a major public health problem in many countries. The number of new cases of HIV infections amongst children has declined 48% since 2009, going from 330,000 in 2009 to 170,000 in 2014.(1) However, as reported in 2013, 3.2 million children under the age of 15 still live with HIV.(2) When infected, HIV attacks the host’s immune system, making the body vulnerable to other infections and diseases. Secondary infections in the central nervous system (CNS) include progressive leukoencephalopathy, toxoplasmosis, cryptococcosis, and meningoencephalitis.(3) However, the most common neurological disease in the CNS of patients infected with HIV is from direct involvement of the virus itself, resulting in cognitive impairments and motor symptoms.(4,5) Soon after contraction, HIV can be detected in the central nervous system, followed by inflammation of the brain.(6) The inflammatory responses also produce neurotoxins that can cause neurodegeneration and, eventually, lead to encephalopathy.(7) Numerous structural deficits such as ventricular and corpus callosum abnormalities, cerebral atrophy, and loss of white matter volume have been reported in people infected with HIV.(8-10)



Saremi 5 Mother-to-child transmission of the virus is associated with the greatest risk of permanent

neurological damage. In the developing brain, these inflammatory responses and neurotoxins exhibit a more damaging effect due to higher sensitivity of neurons during synaptogenesis and pruning.(7) Studies have reported a greater frequency and structural damage in infected children than in adults, with the greatest deficits in vertically infected infants.(11) Cognitive, motor, and language impairments are consistently reported in developing children infected with HIV, increasing with disease severity.(12-14) If treatment is not administered, cortical atrophy with dilation of lateral ventricles, white matter lesions, and basal ganglia calcification can follow.(15,16) However, a precise progression pattern of HIV affecting neural development has not yet been established. In the 21 “global priority” countries, 66% of HIV-infected children receive antiretroviral medication (ART).(1) The medications have been very successful in suppressing viral replication and restoring the body’s immune system. Best results have been seen with combination antiretroviral therapies (cART) where three or more classes of medication are combined.(17,18) In both adults and children, cART can improve neurocognitive performance and suppress its decline.(13,19,20) However, these medications can also penetrate the blood-brain barrier, leading to side-effects such as memory loss, insomnia, hallucinations, depression, and in rare cases, mitochondriopathy, raising concerns about its long-term use.(21,22) Yet, little is known about how ART affects the neurodevelopment of young children. Diffusion tensor imaging (DTI) is a magnetic resonance imaging modality that tracks the Brownian motion of water molecules as they diffuse through the brain’s white matter tissue, providing insight into white matter microstructure and organization.(23) Using computational methods, values summarizing diffusion at each voxel can be extracted from the scan. Common measures include fractional anisotropy (FA) and mean diffusivity (MD). DTI parameters can



Saremi 6

change in response to alteration in white matter structure, such as ischemia, edema, demyelination, axonal damage, and inflammation.(24,25) FA is a normalized value ranging from 0 to 1 representing the non-uniform (anisotropic) diffusion of water, measuring the relative amount of diffusion in the direction of white matter tracts. FA is highly sensitive to microstructural alterations, but it shows less specificity to type of change.(24) MD is the average magnitude of diffusion in all directions, increasing in response to less constrained diffusion. Damage to white matter microstructure (e.g. demyelination and neurodegeneration) are typically observed as a decrease in FA and increase in MD.(25) Axial diffusivity (AD) and radial diffusivity (RD) are also evaluated to provide further insight on white matter organization. Diffusivity parallel (axial) and perpendicular (radial) to the principal direction of the local dominant fiber orientation can be used for more accurate interpretation of FA and MD values. For example, abnormal diffusion along local dominant fibers may represent axonal injury while increase in perpendicular diffusion can represent loss of myelination.(26,27) Diffusion imaging can provide insight into normal white matter architecture in healthy populations, and serve as a method of comparison to make inferences about how disease affects the brain.(28,29) Prior studies of DTI differences between HIV-infected and uninfected children found lower FA and higher MD in regions such as the corpus callosum and superior longitudinal fasciculus,

(30)

with more extensive white matter deficits in adults.(31-33) However, studies on

cART’s effects on white matter integrity have demonstrated inconsistent findings. One study reported no differences associated with cART in HIV-infected children,(34) while another found significant changes in regions of the corpus callosum and corona radiata in cART treated children.(35) Therefore, further evaluation into the neurodevelopmental effects of HIV and antiretroviral medication in children is required.



Saremi 7 In this study, we examined whether abnormalities are detectable between HIV-infected

and HIV-uninfected children and if there are any changes associated with cART status through measuring and comparing diffusion in dominant white matter fibers, cross-sectionally and longitudinally. Detection of anatomical differences could be an indicator of future neurodevelopmental deficits, providing practicing clinicians guidance on whether to administer antiretroviral treatments to young children. We further sought to demonstrate correlations between DTI measures and neuropsychological scores among these children in efforts to reveal any decline in cognitive abilities that might be due to reduction in white matter integrity. Our investigation is mainly exploratory in terms of the affected neurocircuitry, and we evaluate white matter integrity and maturation throughout the brain rather than focus on any particular systems, to ensure better understanding of neurodevelopmental and neuropsychiatric differences. We hypothesize that higher neural integrity should suggest more mature neural development and be positively correlated with better neuropsychological performance.(36,37)

Methods Participant Selection Participants have been selected from the Pediatric Randomized Early versus Deferred Initiation in Cambodia and Thailand (PREDICT) Cohort study,(38) an ongoing investigation of HIV-infected children and comparing their developmental trajectories to each other as well as to HIV-uninfected children. HIV-uninfected subjects are composed of two groups: HIV-unexposed and uninfected (HUU) and HIV-exposed but uninfected (HEU) children. HIV-infected subjects are also divided into two groups: HIV-infected on medication (HOM) and HIV-infected not on medication (HNM) children. The HEU and HUU children were recruited from (1) children born to HIV-infected mothers at the PREDICT study sites and (2) well-child clinics in hospitals associated with the respective study sites. All children satisfied the following inclusion criteria:



Saremi 8

(1) age 18 years. Children were excluded from enrollment for the following reasons: (1) previous or current brain infection; (2) known neurological or psychiatric disorder; (3) any congenital abnormality; or (4) head injury with a loss of consciousness. Images were collected at Chiang Mai University Hospital in Chiang Mai and Chulalongkorn University Hospital in Bangkok, Thailand. Scans at three different time points, approximately one year apart, were provided for each subject by the PREDICT group for a crosssectional and longitudinal analysis. To better age-match the HIV-infected group to the HIVuninfected, only two time points were used for this study. For HIV-infected children, the second and third time points served as the baseline and follow-up scans (39 subjects; age at baseline 10.81±2.17; time between assessments 1.21±0.36). For the HIV-uninfected children, the first and second time points served as the baseline and follow-up scans (53 subjects; age at baseline 11.01±2.80; time between assessments 0.88±0.11). The Institutional Review Boards of each study site, the Thai Ministry of Public Health, University of California, San Francisco, and the University of Southern California, in Los Angeles each approved the study. Clinical Assessment Medical history for the children was provided by their caregivers. Physical and psychological examinations were performed by a pediatrician on dates corresponding with the MRI scan dates. cART is defined as a combination of three or more antiretroviral medication. Age of cART initiation was listed as the date the first medication was prescribed. Children were listed as on-medication (cART +) even if they were exposed to any ART 24-hours before their initial scan to control for any inflammation that might be caused by the drugs (30 out of 39 HIV-



Saremi 9

infected subjects; age of cART initiation 7.91±2.17; time on cART before assessment 2.81±0.72). All HIV-infected children were eventually placed on cART in the months following the initial scan but before reassessment. Children were listed as exposed if they were born from an HIV-infected mother but did not contract the virus (25 out of 53 HIV-uninfected subjects). The exposure of HEU children to ART is unknown as the majority of their caregivers were unable to provide that information. However, 12 out of 25 of the HEU children were born before the use of ART for the prevention of mother-to-child transmission (PMTCT) was implemented in Thailand in the year 2000 and, therefore, are less likely to have been exposed to it. Caregiver’s income was unavailable for 9 children (7 HOM and 2 HNM). The Wechsler Preschool and Primary Scale of Intelligence-III (ages 4–7 years) and the Wechsler Intelligence Scale for Children-III (age 7–16 years) were used to assess cognitive skills at each MRI time point.(39,40) The tests include a subset of core, supplemental, and optional tasks that are used in the computation of the core verbal, performance, and full scale IQ and related indexes. To create consistency in behavioral results, only neuropsychological measures for children who completed the Wechsler Intelligence Scale for Children-III were taken into consideration. 15 children did not complete or have the appropriate neuropsychological assessment completed at baseline (1 HUU, 3 HEU, 5 HOM, and 6 HNM); 9 children had partially reported baseline neuropsychological scores (1 HUU, 7 HOM, and 1 HNM); and, 11 children had no follow-up neuropsychological evaluation (4 HUU, 6 HEU, and 1 HOM). Brain image acquisition Whole brain structural T1-weighted MRI and diffusion-weighted imaging (DWI) were performed on GE 1.5 tesla scanners at both study sites. The structural imaging used the following protocol: axial plane, 3D SPGR images with a minimum TE at full echo, TR = 11.2 ms, slice thickness =1.0mm; 256x256 imaging matrix. DWI was collected in duplicate using a



Saremi 10

single-shot echo-planar imaging protocol; 25 diffusion-sensitized directions were collected with a b-value of 1000 s/mm2 with voxels sized 3mm3. Quality Control One of the challenges of imaging young children is the higher frequency of motion exhibited by them inside the scanner. Young children tend to move more frequently compared to their older counterparts. This movement can affect both the quality of the T1-weighted and the diffusion-weighted images (Figure 1). HIV-infected subjects were younger on average and displayed the most artifacts in their images. Age-matching the HIV-infected group to the HIVuninfected allowed the use of scans that were generated at an older age and thus higher in quality. This reduced the need to withdraw any participants completely from the analysis. All children had T1-weighted images with acceptable quality. However, for 3 of the subjects, only one DWI was provided by the PREDICT group (1 HIV-uninfected at baseline and 2 HIVinfected at reassessment) either due to the low quality of the image or subject’s lack of cooperation. Furthermore, the vibration of the scanner also introduces motion artifacts in children with lighter heads, observed in the posterior-inferior regions such as the cerebellum and brainstem. For this reason, analysis of the cerebellum and brainstem was excluded from this study. Image Processing Extracerebral tissue was removed for the DWI scans using the brain extraction tool from the Functional MRI of the Brain Software Library (FSL; http://www.fmrib.ox.ac.uk/fsl).(41,42) DWI were corrected for head motion and eddy current distortions using the FSL eddy_correct tool. T1-weighted images were intensity normalized using the FSL automated segmentation tool.(43) White matter boundaries were extracted using FreeSurfer, a brain parcellation software developed at the Martinos Center for Biomedical Imaging by the Laboratory for Computational



Saremi 11

Neuroimaging (http://surfer.nmr.mgh.harvard.edu/fswiki),(44) to implement a boundary-based coregistration of the T1-weighted and DWI scans to an asymmetric pediatric template (age 4.518.5) developed by the NIH MRI Study of Normal Brain Development Pediatric database (NIHPD; http://www.bic.mni.mcgill.ca/ServicesAtlases).(45) Registration was performed using the FSL linear image registration tool.(46-48) Once in the same space, to remove echo-planar imaging (EPI) distortion from the DWI, the non-diffusion weighted b0 image was elastically registered to the T1-weighted scan using registration tools provided by the Advanced Normalization Tools (ANTs; http://stnava.github.io/ANTs/).(49-51) The transformation was then applied to the rest of the DWI volumes. Fractional anisotropy and diffusion maps were extracted from the preprocessed DWIs by fitting a diffusion tensor model at each voxel using FSL dtifit tool and then registered to the ENIGMA-DTI atlas using FSL tract-based spatial statistics for extraction of FA/diffusivity values at specified regions of interest.(52) Detail on the preprocessing steps and guidelines on how to perform them can be found on the ENIGMA-DTI Protocol website (http://enigma.ini.usc.edu/protocols/dti-protocols/).(53) Statistical Analysis A linear random-effects regression was used to test for effects of HIV infection, HIV exposure, and cART. Tests were adjusted for imaging site as the random variable and covaried for sex and age. Exposure and cART status were also added as covariates when testing the effects of HIV infection. The extracted diffusion tensor measures from baseline scans ― fractional anisotropy, mean diffusivity, axial diffusivity, and radial diffusivity ― were used as indicators for cross-sectional differences, while their relative rate of change (the relative change in each diffusion tensor measures between baseline and reassessment scans divided by the exhibited time between them) were used as an indicator for longitudinal differences between each group.



Saremi 12 To investigate correlations with time spent on cART, the ratio of time on cART over age

(lifetime on cART) was used to control for the logarithmic pattern observed in the change of diffusion measures,(54,55) and to put more weight on the effects of cART on younger children. To increase statistical power, the analysis was performed using all HIV-infected children and setting the value for lifetime on cART to 0% for subjects that were not on treatment at the time of baseline evaluation. Regions that were found to be significantly different between the groups were further analyzed for correlations within the on-treatment group. Tests were adjusted for imaging site and covaried with sex and age. A regression analysis was performed on each diffusion tensor measure and its relative rate of change to evaluate the effects of infection, exposure, cART, and lifetime on cART in bilaterally averaged and medial regions, followed by a separate analysis of lateral regions to examine if any of the effects exhibit bilateral asymmetry. Statistical testing was limited to regions where FA was >0.2 to exclude gray matter and cerebrospinal fluid.(56,57) As mentioned in previous sections, the cerebellum and brainstem were not evaluated because of a higher level of motion artifacts (Figure 1). The false discovery rate (FDR) method was used to control for the false positive rate of each map at q = 0.05.(58) Furthermore, supplemental regression analyses were performed to find associations in diffusion tensor measures with baseline IQ/indexes and relative rate of change of IQ/indexes (Computed similarly as previously described) in the entire group (covarying with sex, age, and HIV status) and in the HIV-infected group (covarying with sex, age, and cART status). No FDR was set for these tests as they were performed on regions that were previously found to be significant.



Saremi 13

Results HIV Infection Multiple structural differences were found between HIV-infected and HIV-uninfected children (Figure 14): cross-sectional analysis of diffusion tensor measures found MD for the anterior limb of the internal capsule (ALIC) and the superior fronto-occipital fasciculus (SFO) and AD for the internal capsule (IC), posterior limb of the internal capsule (PLIC), corona radiata (CR), anterior corona radiata (ACR), inferior fronto-occipital fasciculus (IFO), and superior fronto-occipital fasciculus to be significantly higher on average in HIV-infected children (Figure 2). Examining bilateral regions individually showed a symmetric pattern for AD and a laterally variant pattern for MD with a preference towards the left hemisphere in regions passing FDR (Table 2). Furthermore, AD for the left superior corona radiata (SCR), even though its average was not found to be significant, survived multiple comparison correction. Longitudinal analysis was able to find regions where relative rate of change in diffusion tensor measures between the two groups passed the p=0.05 threshold, but none survived after correcting for multiple comparison error (Figure 3). Exposure No structural differences between HEU and HUU children were found. Cross-sectional and longitudinal analyses failed to find regions that passed FDR (Figure 4 & 5). Combination Antiretroviral Therapy Multiple structural differences were found between HIV-infected children on cART and without cART (Figure 14): cross-sectional analysis of diffusion tensor measures found AD for the inferior fronto-occipital fasciculus and superior fronto-occipital fasciculus to be significantly lower on average in HIV-infected children on cART (Figure 6). Examining bilateral regions individually showed a laterally variant pattern for AD with no hemispheric preference in regions



Saremi 14

passing FDR (Table 3). Longitudinal analysis of diffusion tensor measures found rate of FA for the corona radiata, anterior corona radiata, superior corona radiata, superior longitudinal fasciculus (SLF), corpus callosum (CC), and body of the corpus callosum (BCC) to be significantly lower on average and rate of RD for the superior longitudinal fasciculus, corpus callosum, and body of the corpus callosum to be significantly higher on average in HIV-infected children on cART (Figure 7, 8, & 9). Examining bilateral regions individually showed a laterally variant pattern for rate of FA with no hemispheric preference in regions passing FDR; no subregions survived multiple comparison correction for rate of RD (Table 3). Furthermore, rate of FA for the left cingulum of the cingulate gyrus (CGC), even though its average was not found to be significant, survived multiple comparison correction. When examining diffusion tensor measures for correlations with lifetime on cART, new regions emerged (Figure 14): Correlations with baseline measures failed to show any regions passing FDR (Figure 10). However, longitudinal analysis of diffusion tensor measures, in addition to the previously named regions, found rate of FA for the internal capsule, anterior limb of the internal capsule, superior fronto-occipital fasciculus, splenium of the corpus callosum (SCC), genu of the corpus callosum (GCC), sagittal stratum (SS), fornix (FX), and the average FA among all regions to be negatively correlated with lifetime on cART, and found rate of RD for anterior limb of the internal, cingulum of the cingulate gyrus, sagittal stratum, splenium of the corpus callosum, corona radiata, posterior corona radiata (PCR), posterior thalamic radiation (PTR), uncinate fasciculus (UNC), and the average RD among all regions to be positively correlated with lifetime on cART (Figure 11). Examining bilateral regions individually showed both rate of FA and rate of RD exhibit a symmetric pattern in some regions while laterally variant with no hemispheric preference in others, with little consensus between the patterns of the two rates (Table 3). Furthermore, rate of FA for the left cingulum of the



Saremi 15

cingulate gyrus and rate of RD for the left superior corona radiata, even though their averages were not found to be significant, survived multiple comparison correction. Examining regions that were found to be both significantly different between the groups and correlated with lifetime on cART showed only rate of FA for superior longitudinal fasciculus, corpus callosum, and body of the corpus callosum to be significantly correlated with lifetime on cART within the HOM group (Figure 12; Table 4). IQ/indexes Further examination of regions that were found significant in the cross-sectional and longitudinal analyses of HIV infection and cART revealed a few diffusion measures to be correlated with few of the neuropsychological indexes (Table 5). Examination of cross-sectional results in the whole group showed MD for the right superior longitudinal fasciculus to be associated with processing speed. Examination of longitudinal results in the HIV-infected group showed rate of FA for the average/right/left corona radiata, average/right/left superior corona radiata, and average/right anterior corona radiata to be associated with freedom from distractibility (Figure 13).

Discussion This work has followed the development of the white matter trajectory in children born to mothers infected with HIV. This randomized trial is the first of its kind to follow the children longitudinally and monitor whether children undergoing treatment immediately at the start of the trial develop differently than those whose treatment is deferred until they reach WHO (World Health Organization) mandated criteria. We made several important discoveries in this work. First, we were able to use diffusion tensor imaging to detect differences between HIVinfected and HIV-uninfected children. There was substantial increase in the AD of regions found in HIV-infected children which corresponded with regions that also had an increase in MD



Saremi 16

(Figure 2). A previous study performed by Hoare et al. (2012) was also able to find differences between these two groups, yet their findings were different than ours.(30) They were able to find higher MD in regions such as the superior longitudinal fasciculus and increase in RD for the same region and the corpus callosum, with no differences in AD. This contradiction in findings could be explained by the difference in subject selection. Our HIV-infected subject pool is consisted of mostly children on treatment, while their HIV-infected subjects were not exposed to medication. However, the results show the same pattern. An increase in AD and MD, followed by a positive trend in increasing RD (even though insignificant), in HIV-infected children can be interpreted as a progression towards homogeneity rather than anisotropy of white matter diffusivity. We were not able find any correlation between baseline diffusion measures and IQ/indexes (Table 5); however, Hoare et al. (2012) was able to find some associations between the measures. The inspections of previous studies and our own findings have made it tempting to conclude that white matter integrity has been compromised in HIV-infected children. However, the increasing trend found in rate of FA and AD in our longitudinal analysis (even though insignificant) hints on a quicker and more mature neurodevelopment in HIV-infected children (Figure 3). It is important to note most children that are vertically infected with HIV do not make it past one year of age. The children investigated in studies such at this are survivors and show strong resistance to the virus. Conversely, the increase in rate can also be interpreted as the brain compensating for existing damage in white matter microstructure, an example of developmental plasticity.(59) Therefore, continued longitudinal monitoring of the HIV-infected children will be necessary to make any final conclusion. We were also not able to find any differences associated with HIV-exposure in the HIVuninfected group, cross-sectionally or longitudinally (Figure 4 & 5). Our results correlate with



Saremi 17

findings of Jahanshad et al. (2015) which looked at baseline differences in brain diffusion between HEU and HUU children and was not able to find significant differences associated with HIV exposure.(59) This gives us the confidence to conclude that HIV exposure does not lead to any measurable differences in white matter microstructure and organization in this sample population. Continued longitudinal monitoring will be performed to determine whether these findings hold through in later stages of neurodevelopment. Studies investigating the effects of cART on white matter integrity have had inconsistent findings, with some showing differences associated with treatment while others do not.(34,35) Our cross-sectional regression analysis of baseline diffusion measures was able to find some differences between HOM and HNM children (Figure 6): AD was shown to be higher in HNM children in regions that were found significantly different between HIV-infected and HIVuninfected groups, suggesting a higher difference between HIV-uninfected children and HNM. These results further complicate the debate on the effect of cART on white matter integrity; even though difference were found, the extend of the effect is not wide-spreading enough to make solid conclusions. However, a longitudinal analysis of the measures has shown there is more to the effect of cART than what is shown at baseline (Figure 7). Regions such as the superior longitudinal fasciculus, the corpus callosum, and the frontal portions of the corona radiata, which are still in the process of development,(61,62) show a lower rate of FA (Figure 8), and a higher rate of RD in the superior longitudinal fasciculus and corpus callosum is seen as well (Figure 9), a pattern associated with damage to white matter microstructure and reduction of integrity. Furthermore, when examining associations with lifetime on cART, more region’s rate of FA and RD show up to be under the effect of cART (Figure 11), with the superior longitudinal fasciculus and corpus callosum FA and RD rates experiencing extensive association both when inspecting the entire HIV-infected group and HOM children (Figure 12; Table 4).



Saremi 18

This is an important finding as it suggests cART has measurable effects on the neurodevelopment of young children. Behavioral testing also showed a higher rate of FA is positively correlated with the rate of change of freedom from distractibility in the corona radiata and its subregion (Figure 13; Table 5). As there’s a significant amount of missing longitudinal neuropsychological data (18 subjects), we are skeptical of the accuracy of the behavioral correlations, but do speculate more regions would be found significantly associated with neuropsychological scores if power is increased. Further longitudinal neuropsychological evaluations must be performed to examine this hypothesis. The primary focus of the current work has been on white matter integrity and development. Now that we have established patterns of effects, in future works, we aim to investigate this further by looking into the structural connectome and how the pattern and network of cortical connections may be altered in the course of development with HIV infection, and any alterations which may arise from deferring treatment. These works will be accomplished through cortical parcellation, which we have done using FreeSurfer and rigorously quality controlled, and utilizing diffusion based tractography to map streamlines which are thought to correspond to the white matter tracts. These cortical regions are the core of the human brain’s functional network, and we may yield insight into the particular systems affected. The PREDICT study has completed in its image collection course, and a new Resilience study has been initiated to more specifically target the neuropsychological and behavioral profile of at-risk infected youth and their abilities to adjust in society, resilient to the social and neurological consequences of HIV-infection. This study follows many of the same children, yet they are being imaged on new scanners placed throughout the sites in Thailand. A new image processing endeavour I am interested in will aim to bridge the connection between the two scanners and profile the development of the children across the two studies. This will be made



Saremi 19

possible through the use of adult human phantoms being scanned with both scanners within a short time interval. The scope of this research project and the work I have done over the past two years alongside the faculty at the USC Imaging Genetics Center has been a unique learning experience. The expedition has broadened my intellectual horizon by introducing me to the current practice and research principles of neuroimaging, a discipline comprised of innovative technology, complex mathematical analyses, and fascinating computer modeling. Specifically, it has shown me how rudimentary the understanding of an undergraduate student actually is compared to the vast knowledge that is managed by the research community. Current research is far more intricate than what is presently taught in a neuroscience course. However, the experience has served as a reinforcement, inspiring me to further increase my understanding of the field to excel both in academics and in research.

Acknowledgment I would to thank my mentor, Dr. Neda Jahanshad for her inspiration and guidance through every step of this project, my coworkers at the USC Imaging Genetics Center for their help and support, as well as Dr. Paul Thompson, co-PI of the PREDICT study and Director of the IGC. I would also like to thank Dr. Victor Valcour at UCSF, who is also co-PI of the study, along with his team, and the entire Thailand team of clinicians and researchers at PREDICT and the Thai Red Cross. This study could not have been possible without the continued participation of all children enrolled.

Funding This work was supported by grants R01MH089722 and R01MH102151, funding provided by the USC Provost’s Undergrad Research Fellowships, and in part, by the NIH 'Big Data to Knowledge' (BD2K) Center of Excellence grant U54 EB020403.



Saremi 20

References 1. United Nations Programme on HIV/AIDS (UNAIDS). (2015). 2015 progress report on the global plan towards the elimination of new HIV infections among children and keeping their mothers alive. UNAIDS. 2. United Nations Programme on HIV/AIDS (UNAIDS). (2015). The GAP report 2014. Geneva: UNAIDS. 3. Smith, A. B., Smirniotopoulos, J. G., & Rushing, E. J. (2008). Central Nervous System Infections Associated with Human Immunodeficiency Virus Infection: RadiologicPathologic Correlation 1. Radiographics, 28(7), 2033-2058. 4. Sibtain, N. A., & Chinn, R. J. S. (2002). Imaging of the central nervous system in HIV infection. Imaging, 14(1), 48-59. 5. Navia, B. A., Jordan, B. D., & Price, R. W. (1986). The AIDS dementia complex: I. Clinical features. Annals of neurology, 19(6), 517-524. 6. Valcour, V., Chalermchai, T., Sailasuta, N., Marovich, M., Lerdlum, S., Suttichom, D., ... & van Griensven, F. (2012). Central nervous system viral invasion and inflammation during acute HIV infection. Journal of Infectious Diseases, 206(2), 275-282. 7. Epstein, L. G., & Gelbard, H. A. (1999). HIV-1-induced neuronal injury in the developing brain. Journal of leukocyte biology, 65(4), 453-457. 8. Thompson, P. M., Dutton, R. A., Hayashi, K. M., Toga, A. W., Lopez, O. L., Aizenstein, H. J., & Becker, J. T. (2005). Thinning of the cerebral cortex visualized in HIV/AIDS reflects CD4+ T lymphocyte decline. Proceedings of the National Academy of Sciences, 102(43), 15647-15652. 9. Thompson, P. M., Dutton, R. A., Hayashi, K. M., Lu, A., Lee, S. E., Lee, J. Y., ... & Becker, J. T. (2006). 3D mapping of ventricular and corpus callosum abnormalities in HIV/AIDS. Neuroimage, 31(1), 12-23. 10. Chiang, M. C., Dutton, R. A., Hayashi, K. M., Lopez, O. L., Aizenstein, H. J., Toga, A. W., ... & Thompson, P. M. (2007). 3D pattern of brain atrophy in HIV/AIDS visualized using tensor-based morphometry. Neuroimage, 34(1), 44-60. 11. Tardieu, M., Le Chenadec, J., Persoz, A., Meyer, L., Blanche, S., & Mayaux, M. J. (2000). HIV-1–related encephalopathy in infants compared with children and adults. Neurology, 54(5), 1089-1095. 12. Chase, C., Ware, J., Hittelman, J., Blasini, I., Smith, R., Llorente, A., ... & Kaligh, L. I. (2000). Early cognitive and motor development among infants born to women infected with human immunodeficiency virus. Pediatrics, 106(2), e25-e25. 13. Le Doaré, K., Bland, R., & Newell, M. L. (2012). Neurodevelopment in children born to HIV-infected mothers by infection and treatment status. Pediatrics, peds-2012. 14. Smith, R., Malee, K., Leighty, R., Brouwers, P., Mellins, C., Hittelman, J., ... & Blasini, I. (2006). Effects of perinatal HIV infection and associated risk factors on cognitive development among young children. Pediatrics, 117(3), 851-862. 15. Safriel, Y. I., Haller, J. O., Lefton, D. R., & Obedian, R. (2000). Imaging of the brain in the HIV-positive child. Pediatric radiology, 30(11), 725-732. 16. George, R., Andronikou, S., Du Plessis, J., du Plessis, A. M., Van Toorn, R., & Maydell, A. (2009). Central nervous system manifestations of HIV infection in children. Pediatric radiology, 39(6), 575-585.



Saremi 21 17. Mathis, S., Khanlari, B., Pulido, F., Schechter, M., Negredo, E., Nelson, M., ... & Bucher, H. C. (2011). Effectiveness of protease inhibitor monotherapy versus combination antiretroviral maintenance therapy: a meta-analysis. PloS one, 6(7), e22003. 18. Taiwo, B., Murphy, R. L., & Katlama, C. (2010). Novel Antiretroviral Combinations in Treatment-Experienced Patients with HIV Infection. Drugs, 70(13), 1629-1642. 19. Ellis, R. (2010). HIV and antiretroviral therapy: impact on the central nervous system. Progress in neurobiology, 91(2), 185-187. 20. Whitehead, N., Potterton, J., & Coovadia, A. (2014). The neurodevelopment of HIVinfected infants on HAART compared to HIV-exposed but uninfected infants. AIDS care, 26(4), 497-504. 21. Ene, L., Duiculescu, D., & Ruta, S. M. (2011). How much do antiretroviral drugs penetrate into the central nervous system?. Journal of medicine and life, 4(4), 432. 22. Abers, M. S., Shandera, W. X., & Kass, J. S. (2014). Neurological and psychiatric adverse effects of antiretroviral drugs. CNS drugs, 28(2), 131-145. 23. Basser, P. J., & Pierpaoli, C. (2011). Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. Journal of magnetic resonance, 213(2), 560-570. 24. Thomason, M. E., & Thompson, P. M. (2011). Diffusion imaging, white matter, and psychopathology. Clinical Psychology, 7(1), 63. 25. Alexander, A. L., Lee, J. E., Lazar, M., & Field, A. S. (2007). Diffusion tensor imaging of the brain. Neurotherapeutics, 4(3), 316-329. 26. Le Bihan, D., Mangin, J. F., Poupon, C., Clark, C. A., Pappata, S., Molko, N., & Chabriat, H. (2001). Diffusion tensor imaging: concepts and applications. Journal of magnetic resonance imaging, 13(4), 534-546. 27. Song, S. K., Sun, S. W., Ramsbottom, M. J., Chang, C., Russell, J., & Cross, A. H. (2002). Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage, 17(3), 1429-1436. 28. Hermoye, L., Saint-Martin, C., Cosnard, G., Lee, S. K., Kim, J., Nassogne, M. C., ... & Wakana, S. (2006). Pediatric diffusion tensor imaging: normal database and observation of the white matter maturation in early childhood. Neuroimage, 29(2), 493-504. 29. Dennis, E. L., & Thompson, P. M. (2013). Typical and atypical brain development: a review of neuroimaging studies. Dialogues Clin Neurosci, 15(3), 359-384. 30. Hoare, J., Fouche, J. P., Spottiswoode, B., Donald, K., Philipps, N., Bezuidenhout, H., ... & Paul, R. (2012). A diffusion tensor imaging and neurocognitive study of HIV-positive children who are HAART-naïve “slow progressors”. Journal of neurovirology, 18(3), 205212. 31. Nir, T. M., Jahanshad, N., Busovaca, E., Wendelken, L., Nicolas, K., Thompson, P. M., & Valcour, V. G. (2014). Mapping white matter integrity in elderly people with HIV. Human brain mapping, 35(3), 975-992. 32. Gongvatana, A., Cohen, R. A., Correia, S., Devlin, K. N., Miles, J., Kang, H., ... & Tashima, K. T. (2011). Clinical contributors to cerebral white matter integrity in HIVinfected individuals. Journal of neurovirology, 17(5), 477-486. 33. Pfefferbaum, A., Rosenbloom, M. J., Rohlfing, T., Kemper, C. A., Deresinski, S., & Sullivan, E. V. (2009). Frontostriatal fiber bundle compromise in HIV infection without dementia. AIDS (London, England), 23(15), 1977.



Saremi 22 34. Li, R., Tang, Z., Mi, H., Zhao, J., Yuan, D., & Li, H. (2015). Diffusion tensor imaging study of early white matter integrity in HIV-infected patients: A tract-based spatial statistics analysis. Radiology of Infectious Diseases, 2(4), 183-191. 35. Hoare, J., Fouche, J. P., Phillips, N., Joska, J. A., Paul, R., Donald, K. A., ... & Stein, D. J. (2015). White matter micro-structural changes in ART-naive and ART-treated children and adolescents infected with HIV in South Africa. AIDS, 29(14), 1793-1801. 36. Chaddock-Heyman, L., Erickson, K. I., Voss, M. W., Powers, J. P., Knecht, A. M., Pontifex, M. B., ... & Hillman, C. H. (2013). White matter microstructure is associated with cognitive control in children. Biological psychology, 94(1), 109-115. 37. Tamnes, C. K., Østby, Y., Walhovd, K. B., Westlye, L. T., Due-Tønnessen, P., & Fjell, A. M. (2010). Intellectual abilities and white matter microstructure in development: a diffusion tensor imaging study. Human brain mapping, 31(10), 1609-1625. 38. Puthanakit, T., Saphonn, V., Ananworanich, J., Kosalaraksa, P., Hansudewechakul, R., Vibol, U., ... & Luesomboon, W. (2012). Early versus deferred antiretroviral therapy for children older than 1 year infected with HIV (PREDICT): a multicentre, randomised, openlabel trial. The Lancet infectious diseases, 12(12), 933-941. 39. Wechsler, D. (1989). Wechsler preschool and primary scale of intelligence-revised. Psychological Corporation. 40. Wechsler, D. (1991). WISC-III Wechsler Intelligence Scale for Children–third edition. San Antonio, TX: The Psychological Corporation. 41. Smith, S. M. (2002). Fast robust automated brain extraction. Human brain mapping, 17(3), 143-155. 42. Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., & Smith, S. M. (2012). Fsl. Neuroimage, 62(2), 782-790. 43. Zhang, Y., Brady, M., & Smith, S. (2001). Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. Medical Imaging, IEEE Transactions on, 20(1), 45-57. 44. Fischl, B., van der Kouwe, A., Destrieux, C., Halgren, E., Ségonne, F., Salat, D. H., ... & Caviness, V. (2004). Automatically parcellating the human cerebral cortex. Cerebral cortex, 14(1), 11-22. 45. Fonov, V., Evans, A. C., Botteron, K., Almli, C. R., McKinstry, R. C., Collins, D. L., & Brain Development Cooperative Group. (2011). Unbiased average age-appropriate atlases for pediatric studies. NeuroImage, 54(1), 313-327. 46. Jenkinson, M., Bannister, P., Brady, M., & Smith, S. (2002). Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage, 17(2), 825-841. 47. Jenkinson, M., & Smith, S. (2001). A global optimisation method for robust affine registration of brain images. Medical image analysis, 5(2), 143-156. 48. Greve, D. N., & Fischl, B. (2009). Accurate and robust brain image alignment using boundary-based registration. Neuroimage, 48(1), 63-72. 49. Leow, A. D., Yanovsky, I., Chiang, M. C., Lee, A. D., Klunder, A. D., Lu, A., ... & Thompson, P. M. (2007). Statistical properties of Jacobian maps and the realization of unbiased large-deformation nonlinear image registration. Medical Imaging, IEEE Transactions on, 26(6), 822-832. 50. Huang, H., Ceritoglu, C., Li, X., Qiu, A., Miller, M. I., van Zijl, P. C., & Mori, S. (2008). Correction of B0 susceptibility induced distortion in diffusion-weighted images using



Saremi 23 large-deformation diffeomorphic metric mapping. Magnetic resonance imaging, 26(9), 1294-1302. 51. Avants, B. B., Tustison, N. J., Song, G., Cook, P. A., Klein, A., & Gee, J. C. (2011). A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage, 54(3), 2033-2044. 52. Smith, S. M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T. E., Mackay, C. E., ... & Behrens, T. E. (2006). Tract-based spatial statistics: voxelwise analysis of multisubject diffusion data. Neuroimage, 31(4), 1487-1505. 53. Jahanshad, N., Kochunov, P. V., Sprooten, E., Mandl, R. C., Nichols, T. E., Almasy, L., ... & Duggirala, R. (2013). Multi-site genetic analysis of diffusion images and voxelwise heritability analysis: A pilot project of the ENIGMA–DTI working group. Neuroimage, 81, 455-469. 54. Wakana, S., Nagae-Poetscher, L. M., Jiang, H., van Zijl, P., Golay, X., & Mori, S. (2005). Macroscopic orientation component analysis of brain white matter and thalamus based on diffusion tensor imaging. Magnetic resonance in medicine, 53(3), 649-657. 55. Lebel, C., Walker, L., Leemans, A., Phillips, L., & Beaulieu, C. (2008). Microstructural maturation of the human brain from childhood to adulthood. Neuroimage, 40(3), 10441055. 56. Lebel, C., Gee, M., Camicioli, R., Wieler, M., Martin, W., & Beaulieu, C. (2012). Diffusion tensor imaging of white matter tract evolution over the lifespan. Neuroimage, 60(1), 340-352. 57. Smith, S. M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T. E., Mackay, C. E., ... & Behrens, T. E. (2006). Tract-based spatial statistics: voxelwise analysis of multisubject diffusion data. Neuroimage, 31(4), 1487-1505. 58. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 289-300. 59. Jiang, Q., Zhang, Z. G., & Chopp, M. (2010). MRI of stroke recovery. Stroke, 41(2), 410414. 60. Jahanshad, N., Couture, M. C., Prasitsuebsai, W., Nir, T. M., Aurpibul, L., Thompson, P. M., ... & Desai, A. (2015). Brain imaging and neurodevelopment in HIV-uninfected Thai children born to HIV-infected mothers. The Pediatric infectious disease journal, 34(9), e211-e216. 61. Kochunov, P., Glahn, D. C., Lancaster, J., Thompson, P. M., Kochunov, V., Rogers, B., ... & Williamson, D. E. (2011). Fractional anisotropy of cerebral white matter and thickness of cortical gray matter across the lifespan. Neuroimage, 58(1), 41-49. 62. Asato, M. R., Terwilliger, R., Woo, J., & Luna, B. (2010). White matter development in adolescence: a DTI study. Cerebral Cortex, 20(9), 2122-2131.





Saremi 24

Tables



HIV-infected

HIV-uninfected

P-value

39

53

0.09

10.81 (2.17)

11.01 (2.80)

0.65

Sex, % Male

%56

%42

0.24

School Attendance, % Attended

%100

%100

1.0

7 22 1



0.07

16 10



1.21 (0.36)

0.88 (0.11)

) have a Tvalue that goes beyond what is displayed on the graph.





Figure 7. Longitudinal results of diffusion measures between HOM and HNM children. Tracks are displayed in order of significance. Darker color signifies passing the p=0.05 threshold. Regions marked with a (*) passed FDR at q=0.05. Regions marked with (>) have a T-value that goes beyond what is displayed on the graph.



Saremi 32



Figure 8. A longitudinal comparison of the rate of FA between HIV-uninfected children (HIV -), HOM children (cART +), and HNM children (cART -) for regions passing FDR in the longitudinal analysis of cART. Each line represents one subject’s change in FA between their initial and reassessment scan. Subject color and trendline pattern are indicated on the right of each graph.





Figure 9. A longitudinal comparison of the rate of RD between HIV-uninfected children (HIV -), HOM children (cART +), and HNM children (cART -) for regions passing FDR in the longitudinal analysis of cART. Each line represents one subject’s change in FA between their initial and reassessment scan. Subject color and trendline pattern are indicated on the right of each graph.



Saremi 33



Figure 10. Cross-sectional correlation between diffusion measures and lifetime on cART for HIV-infected children. Tracks are displayed in order of significance. Darker color signifies passing the p=0.05 threshold. Regions marked with a (*) passed FDR at q=0.05.





Figure 11. Longitudinal correlation between diffusion measures and lifetime on cART for HIV-infected children. Tracks are displayed in order of significance. Darker color signifies passing the p=0.05 threshold. Regions marked with a (*) passed FDR at q=0.05. Regions marked with (>) have a T-value that goes beyond what is displayed on the graph.



Saremi 34



Figure 12. A comparison of the rate of FA between HOM (cART +) and HNM children (cART -) for regions that were both significantly different between the groups and were correlated with lifetime on cART. Subject color and trendline pattern are indicated on the right of each graph.









Figure 13. Visualization of regions which rate of FA were associated with freedom of distractibility for HIV-infected children. Only limited number of subjects (21 out of the 39 HIV-infected subjects) had the necessary behavioral measures performed (both at baseline and reassessment) for a longitudinal analysis. Subject color and trendline pattern are indicated on the right of each graph.



Saremi 35





Figure 14. Plot displays diffusion tensor measure/rate of medial and bilateral subregions that were found to be significantly different/correlated with HIV/cART/lifetime on cART. Colors indicate “-log” of the p-value of the performed test for that region; a change in color indicates a change in the region's significance by a factor of 10. Only subregions passing FDR are shown; displayed bilateral regions were required to pass multiple comparison correction both when inspecting their average in initial testing and separately in the follow-up bilateral analysis.