Neuroimaging the Neuropathogenesis of HIV

Abstract

Purpose of Review

This review highlights neuroimaging studies of HIV conducted over the last 2 years and discusses how relevant findings further our knowledge of the neuropathology of HIV. Three major avenues of neuroimaging research are covered with a particular emphasis on inflammation, aging, and substance use in persons living with HIV (PLWH).

Recent Findings

Neuroimaging has been a critical tool for understanding the neuropathological underpinnings observed in HIV. Recent studies comparing levels of neuroinflammation in PLWH and HIV-negative controls show inconsistent results but report an association between elevated neuroinflammation and poorer cognition in PLWH. Other recent neuroimaging studies suggest that older PLWH are at increased risk for brain and cognitive compromise compared to their younger counterparts. Finally, recent findings also suggest that the effects of HIV may be exacerbated by alcohol and drug abuse.

Summary

These neuroimaging studies provide insight into the structural, functional, and molecular changes occurring in the brain due to HIV. HIV triggers a strong neuroimmune response and may lead to a cascade of events including increased chronic inflammation and cognitive decline. These outcomes are further exacerbated by age and age-related comorbidities, as well as lifestyle factors such as drug use/abuse.

Introduction

Since HIV/AIDS was first described in 1981, incredible progress has been made in the care of persons living with HIV (PLWH). The UNAIDS reported that in 2019, 81% of PLWH knew their status, 82% of those who knew their status had access to treatment, and 88% of those being treated were virally suppressed (https://www.unaids.org/en/resources/fact-sheet). In the past, HIV was considered a fatal disease, but with the development of combined antiretroviral therapy (cART), the life expectancy of PLWH is now approaching general population norms [1]. Despite this progress, PLWH continue to experience health complications related to ongoing HIV infection, such as HIV-associated neurocognitive disorder (HAND) [2].

An estimated 30–50% of PLWH develop some form of HAND [3, 4]. PLWH most commonly exhibit impairment in the learning/memory and executive functioning domains [5, 6]. In the post-cART era, most PLWH with HAND have milder forms, either asymptomatic neurocognitive impairment (ANI) or mild neurocognitive disorder (MND) [7]. Furthermore, since the introduction of cART, less than 5% of PLWH diagnosed with HAND have HIV-associated dementia (HAD) [3, 5]. Despite these improvements, cognitive impairment continues to affect quality of life and interfere with activities of day-to-day living of PLWH [7].

HIV enters the central nervous system (CNS) soon after initial infection [8]. Even after initiating cART and achieving viral suppression, some PLWH can have viral escape with the continued presence of HIV RNA that is detectable in the cerebrospinal fluid (CSF) [9]. The source of this persistent HIV infection is still debated. There is evidence that HIV viral load in the CNS is maintained by ongoing infection by peripheral cells, such as monocytes [10, 11]. Alternatively, studies have demonstrated that CD4 T cells and/or microglia are the site of this HIV reservoir in the CNS [12]. Ongoing research has focused on understanding how persistent HIV infection and neuroinflammation affects CNS function.

Neuroimaging has been a vital tool for understanding the neuropathology seen with HIV. A diverse set of techniques enable researchers to evaluate structural, functional, and molecular changes in the brain due to HIV (Table 1). Furthermore, neuroimaging provides a noninvasive technique to repeatedly measure brain changes, making longitudinal assessment easier. Given that HIV is a chronic disease, this is especially important for analyzing the long-term effects of the virus. This review highlights studies conducted over the last 2 years regarding HIV and neuroimaging and discusses how relevant findings further our understanding of the neuropathology of HIV. Three major avenues of neuroimaging research are covered with a particular emphasis on inflammation, aging, and substance use in PLWH.

Table 1.

Neuroimaging modalities in persons living with HIV

Structural magnetic resonance imaging (MRI)	Cortical and subcortical brain volumes
Magnetic resonance spectroscopy (MRS)	Metabolites associated with inflammation, neuronal integrity, and neuronal transmission
TSPO PET imaging	Activated microglia imaging indicative of neuroinflammation
Functional MRI	Neural network activity and functional connectivity
Diffusion tensor imaging	White matter integrity

TSPO, 18-kDa translocator protein; PET, positron-emission tomography; MRI, magnetic resonance spectroscopy

Inflammation

HIV enters the CNS and triggers an immune response. Many studies have shown a significant decrease in immune activation in PLWH after initiating cART [13-19]. Despite adhering to cART, low levels of viremia persist in the CNS and may lead to chronic immune activation in the brain [11]. While acute inflammation is an important immune response, chronic inflammation can cause significant damage to the brain and result in the cognitive deficits seen in PLWH. Inflammatory markers in the plasma or CSF are frequently used, but these measures are often indirect measures of neuronal processes that may be occurring in the brain. In the case of lumbar puncture for obtaining CSF, this method can be somewhat invasive. In contrast, neuroimaging is noninvasive and can spatially localize inflammation to specific brain regions. Several neuroimaging methods can evaluate neuroinflammation including injection of tagged radioisotopes using positron-emission tomography (PET) and MRI methods, specifically magnetic resonance spectroscopy (MRS).

PET imaging has been used to assess neuroinflammation due to chronic HIV infection. Using the radiotracer [11C]-PBR28, which targets the 18-kDa translocator protein (TSPO) of activated microglia, levels of neuroinflammation have been compared between virologically well-controlled PLWH and HIV-uninfected controls. Previous studies using TSPO PET imaging have shown inconsistent results when comparing PLWH on cART to HIV-uninfected controls [20-24]. One study observed no differences between PLWH and HIV-uninfected controls, while others measured an increase in neuroinflammation in PLWH but in various non-overlapping brain regions. More recently, a larger study conducted by Boerwinkle and colleagues (2020) demonstrated no significant differences in neuroinflammation were observed between virally controlled PLWH on cART and HIV-uninfected controls [25•]. Within the PLWH, an inverse correlation was observed with greater neuroinflammation in the frontal, parietal, occipital cortices, and thalamus associated with worse executive function. Levels of neuroinflammation were not associated with duration of infection or CNS penetration-effectiveness (CPE) score of cART, consistent with previous TSPO PET studies. Additionally, neuroinflammation as measured by TSPO PET was not significantly associated with plasma measures of inflammatory biomarkers.

Magnetic resonance spectroscopy (MRS), a technique that can measure regional immune activation, has also shown mixed results. A recent study by Chaganti et al. (2019) observed no differences in myo-inositol and glutamate/glutamine levels but did observe an elevation of choline in the basal ganglia and frontal white matter of cART-suppressed PLWH compared to HIV-uninfected individuals [26]. However, previous studies have observed an increase in myo-inositol but not choline in PLWH compared to HIV-uninfected individuals [27, 28]. A recent follow-up study by Boban and colleagues (2019) assessed changes in MRS measures of inflammation over 5 years in virologically well-controlled PLWH [29•]. PLWH had a significant increase of myo-inositol within multiple regions (ventral and dorsal anterior cingulate gyrus, posterior cingulate gyrus, and intraparietal sulcus). In addition, choline was increased within the dorsal anterior cingulate gyrus of PLWH. Finally, a recent MRS study by Alakkas et al. (2019) assessed the relationship between neuroinflammation and cognitive impairment in PLWH. Results showed that choline was significantly elevated in the frontal cortex of virologically suppressed PLWH who had cognitive impairment compared to unimpaired PLWH [30].

Given the inconsistencies among the results of these and previous neuroimaging studies, additional work is necessary to help elucidate the source of these differences. For example, longitudinal neuroimaging studies would be beneficial to assess the impact of events such as viral blips or CSF HIV escape on levels of chronic neuroimmune activation in PLWH. Factors such as gender, chronic stress, and age have been suggested to influence immune activation in PLWH, and future studies may provide insight into these relationships [31, 32]. Overall, additional neuroimaging studies are needed to understand the relationship between neuroinflammation and HIV.

Aging

Since the introduction of cART, the proportion of PLWH who reach geriatric age has dramatically increased with nearly 63% of all PLWH in the USA aged 50 years or older. Accordingly, there has been increased effort to determine the effects of aging and aging-related comorbidities within older PLWH. It has been hypothesized that the prevalence of HAND may be over-inflated due to pre-existing comorbidities. Estimates suggest that minimal-to-moderate comorbidities exist in 20–50% of all individuals with HAND and these may exert deleterious effects on older PLWH [4, 33, 34]. Determining the neuroimaging correlates of comorbidities is therefore critical. For example, cardiovascular disorders (CVD), such as hypertension and diabetes mellitus, increase with age and contribute to central nervous system dysfunction in PLWH as well as the general population. The negative effects of CVD are particularly apparent in relation to white matter integrity. Recent research has demonstrated that white matter microstructural alterations are greatest in virally suppressed older PLWH who have CVD risk factors compared to those without CVD [35•]. A synergistic interaction between aging and related comorbidities (e.g., CVD) may occur independent of HIV.

In addition to white matter, gray matter integrity is also negatively affected by comorbidities in older PLWH. For example, when investigating associations between comorbidity burden and brain integrity, as well as the moderating effect of age, Saloner and colleagues (2019) observed specific neuroimaging signatures that were associated with confounding conditions. Comorbidity burden (neuromedical and neuropsychiatric) was determined using the Frascati criteria for HAND, and individuals were classified as having either incidental (i.e., mild), contributing (i.e., moderate), or confounding (i.e., severe-exclusionary) conditions with regards to a HAND diagnosis. Multiple regression models revealed that although the groups did not differ by HIV disease characteristics, structural brain differences persisted. Specifically, PLWH who had confounding conditions had smaller cortical gray matter volumes, larger ventricles due to atrophy, and greater white matter burden compared to PLWH who had only incidental conditions. Confounded individuals additionally had greater neuroinflammation (e.g., elevated choline and myo-inositol on MRS) and reduced neural integrity (e.g., N-acetyl aspartate on MRS). Age differentially affected the various comorbidity groups. In the incidental and contributing comorbidity groups, older age was associated with reductions in cortical gray matter. The confounded group demonstrated greater white matter abnormalities, reduced total white matter, and smaller subcortical gray matter volume with increasing age compared to the other groups. Notably, the deleterious effects of age on white matter burden were the greatest in individuals who also had hypertension and hepatitis C coinfection [36•].

Due to the extreme nuance of factors affecting and interacting with HIV, recent research has focused on defining individual trajectories in order to better predict outcomes and inform treatment for personalized medicine. For example, Popov and colleagues (2020) used a data-driven mixed membership trajectory model (MMTM) to assess the relationship between regional brain volumes and individual trajectories of impairment in adults 50 years and older [37]. A total of 302 men (167 PLWH) from the Multicenter AIDS Cohort Study (MACS) underwent MRI and were assigned to one of three canonical profiles: “normal aging” (i.e., probability of normal cognition was high), “premature aging” (i.e., probability of mild impairment occurring between 45–50 years old), or “unhealthy” (i.e., probability of normal cognition nearly zero regardless of age). Their analysis revealed that the “unhealthy” trajectory was associated with volume loss within the posterior cingulate/precuneus, putamen, insula, inferior frontal cortex, and caudate nucleus gray matter, while the “premature aging” profile was associated with a reduction in the precuneus volume. It was concluded that these cognitive impairment trajectories were the result of atrophy in regions involved in normal and pathological aging [38].

Findings such as those from Popov et al. (2020) and Calon et al. (2020) highlight the need to determine whether an interaction exists between HIV and age, and if so, to what extent. As mentioned previously, white matter burden due to CVD is particularly vulnerable to the effects of aging with a potential interaction occurring with HIV status, especially in relation to white matter microstructure [39]. Specifically, using diffusion tensor imaging (DTI), it has been demonstrated that in PLWH, increasing age is associated with lower fractional anisotropy (FA), and higher mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) along with multiple white matter tracts relative to either younger PLWH or HIV− individuals. Interactions between HIV and aging were strongest within fibers projecting to the temporal and frontal lobes, consistent with previous literature implicating frontal-subcortical involvement due to HIV [39-43]. Critically, these effects are also present when controlling for confounding factors such as sex, ethnicity, education level, nadir CD4 count, and highest lifetime HIV viral load [39]. Joint effects of HIV and aging are similarly observed in alterations in neural activity within the visuo-motor and perceptual systems, which also demonstrate disruptions in cerebral blood flow, volume, and functional connectivity using functional magnetic resonance imaging (fMRI) in individuals with HIV and frailty [44, 45]. Collectively, these results suggest that older PLWH are at increased risk for brain and cognitive compromise compared to their younger counterparts.

Substance Use

Substance use is highly prevalent and is associated with global disease burden, high economic costs, and morbidity and mortality [46, 47]. Substance use is more common among PLWH compared to HIV− controls [48, 49]. Abuse can lead to alterations in the blood-brain barrier which can contribute to the continued presence of brain viral reservoirs and chronic neuroinflammation [50]. Substance use may also contribute to alterations in the structure and functional organization of the brain, including brain atrophy, white matter damage, and functional network strength, and can increase the risk of developing HIV-associated neurological disorders and/or contribute to lack of adherence to antiretroviral therapy [51-56]. Compounding these effects, polysubstance use, or the use of multiple drugs simultaneously, has become more common [57, 58]. Numerous neuroimaging studies have shown alterations in the structure and function of the brain as a result of use and/or abuse of tobacco, alcohol, cannabis, stimulants, and opioids [40, 54, 56, 59]. An in-depth understanding of the characteristic phenotypes, mechanisms, synergistic effects, and high-order interactions of substance use in PLWH is needed to provide optimal patient care. However, limited published studies, a lack of consistency in published results, or the general exclusion of individuals with substance use disorders in studies have led to a lack of understanding of these effects in PLWH.

The prevalence of tobacco use in PLWH is 2–3 times higher than the general US population [54, 60]. Tobacco use may affect ART drug metabolism, and certain cART regimens may affect the metabolism of chemicals found in tobacco [61, 62]. Moreover, the presence of virotoxins and chemicals found in tobacco may break down the blood-brain barrier, facilitating the entry of proinflammatory cytokines into the brain [63]. Neuroimaging studies have shown that the tobacco use in PLWH can have a deleterious effect on brain structure and function, possibly due to the chemicals in tobacco exacerbating neuroinflammation and immune dysregulation [54, 61, 64, 65]. For example, Chang et al. [54] evaluated 21 PLWH smokers, 25 PLWH nonsmokers, 25 HIV− smokers, and 23 HIV− nonsmokers with DTI. Eight cerebral fiber tracts and five subcortical regions were evaluated, along with cognitive performance assessments in seven domains. They found additive and synergistic effects of HIV and smoking in multiple white matter tracks, with PLWH who were smokers showing the highest MD and lowest FA. Furthermore, higher diffusivity in some regions was predictive of poorer cognitive performance. They concluded that the proinflammatory and demyelination effects of tobacco use combined with HIV may lead to an increased risk of white matter abnormalities and HAND. However, a retrospective study conducted by Tsima et al. [66] on 3033 PLWH of which 49% were self-reported smokers found no differences in cognitive impairment using the Mental Alternation Test. They noted that after controlling for confounding factors, there was a 12% higher risk of cognitive impairment in PLWH who smoked, but this result was not statistically significant. While the prevalence of smoking in PLWH remains a serious health concern, further studies are needed to ascertain its impact on the brain structure and function.

Alcohol is the most commonly abused drug in the world [47], and alcohol abuse is prevalent among PLWH [49]. Studies have shown that alcohol use is associated with changes in neuroimmune and neuroinflammatory processes, including alterations in blood-brain barrier permeability and an increase in proinflammatory cytokines [67]. Alcohol use may also lead to lack of adherence to cART or interfere with the mechanisms of pharmacological treatment [53]. Neuroimaging research on the combined effects of alcohol use and HIV is limited, as previous studies have generally excluded individuals with excessive alcohol use [68]. Gullett et al. [40•] evaluated DTI-based metrics of white matter integrity in four white matter tracks in 37 PLWH, 25 of which had alcohol use disorder. Participants with alcohol use disorder were found to have increased axial diffusivity in the anterior thalamic radiation. When accounting for nadir CD4 and age-adjusted length of infection, a diagnosis of alcohol use disorder explained over 36% of the variance in this white matter track. As a result, the authors suggest a possible link between reduced frontal white matter integrity and alcohol use disorder in PLWH.

Cannabis use is common among PLWH, with studies reporting rates of use ranging from 20 to 60% [51, 69, 70]. Many PLWH report the use of cannabis as a means of ameliorating HIV symptoms and/or side effects associated with treatment (e.g., pain, nausea, mood problems, and appetite) [71-74]. Studies have shown that both HIV and cannabis use can lead to disruptions in the intrinsic brain functional connectivity [75-77]. Meade et al. [56] examined the independent and additive effects of HIV and cannabis use. This study included 20 PLWH who were cannabis users, 19 HIV− cannabis users, 29 PLWH who did not use cannabis, and 25 HIV− who did not use cannabis. All participants performed a counting Stroop task during a functional magnetic resonance imaging (fMRI) scan. In mixed-effects analysis, an interaction was observed in the left fronto-insular cortex of PLWH who used cannabis. This group had the largest increase in activation compared to the other groups. Conversely, Hall et al. [78] applied graph theoretical measures to fMRI collected from 20 PLWH who used cannabis, 17 HIV− individuals who used cannabis, 20 PLWH who did not use cannabis, and 22 HIV− controls who did not use cannabis. Their results showed that PLWH and HIV− individuals who used cannabis had no significant differences in brain network organization, but differences were observed when compared to HIV− individuals who did not use cannabis. The authors conclude that a synergistic effect exist between cannabis use and HIV that could lead to alterations in brain networks. When evaluating structural MRI, Thames et al. [79] evaluated 24 PLWH who used cannabis, 24 PLWH who did not use cannabis, 13 HIV− individuals who used cannabis, and 16 HIV− individuals who did not use cannabis. They observed that cannabis use, regardless of HIV status, was associated with smaller volumes in the entorhinal cortex and fusiform gyrus. An interaction was seen between HIV and cannabis for global cognition, as measured by cognitive testing that encompassed multiple domains. This effect was based on the amount of cannabis consumed during a typical week. Interestingly, those who consumed less cannabis (<1.43 g per week) were more likely to show impairment compared to those who consumed greater amounts (≥1.43 g per week). The authors conclude that the combination of HIV and cannabis use does not lead to a greater risk for adverse brain or cognitive outcomes compared to HIV− cannabis users.

PLWH have higher rates of cocaine use disorder (CUD) than the general population [55]. The use of cocaine leads to an increase in dopamine, which is believed to result in an inflammatory response and increased permeability of the blood-brain barrier [80]. Studies of PLWH with CUD have primarily focused on reward-based decision-making and risk-taking behavior. Bell et al. [55] use fMRI to evaluated the interactive effects of HIV and cocaine use disorder within 19 PLWH who use cocaine, 18 HIV− individuals who use cocaine, 24 PLWH who did not use cocaine, and 18 HIV− who did not use cocaine. A Wheel of Fortune task that assessed neural activation in response to monetary risk was performed during the scan. In response to increasing risk, the PLWH who used cocaine exhibited higher neural activation across multiple brain regions, including the left precuneus, posterior cingulate cortex, hippocampus, and right postcentral gyrus, lateral occipital cortex, cerebellum, and posterior parietal cortex. The authors conclude that the combination of HIV infection and cocaine use may lead to reduced efficiency when processing risk involved in decision-making. Similar studies have observed altered neural activity in several regions, including the frontal pole, prefrontal cortex, precentral gyrus, and cerebellum [81]. Conversely, Cordero et al. [82] found no evidence of an independent effect of cocaine on white matter integrity in PLWH. Their study involved 26 PLWH who use cocaine, 37 PLWHwho do not use cocaine, 37 HIV− individuals who use cocaine, and 35 HIV− individuals who do not use cocaine. These individuals were evaluated with DTI measures of FA and MD. Their analysis revealed a significant HIV effect on FA and MD, but no evidence that cocaine exacerbated the effects of HIV.

Methamphetamine (METH) use is common among PLWH [83, 84]. Studies have shown that the use of METH can cause impairment in multiple cognitive domains, is associated with a higher likelihood of frailty, and can lead to brain structural and functional alterations [85-88]. In PLWH, METH use can lead to impaired behavioral inhibition, which in turn can lead to worse medication adherence, greater risk taking behavior, and a greater likelihood ofHIV transmission [89]. MacDuffie et al. [90] utilized structural MRI to evaluate the effects of METH and potential interactions with HIV when evaluating cortical thickness. T1-weighted structural MRI was obtained on 31 PLWH who use METH, 38 PLWH who do not use METH, 28 HIV− individuals who use METH, and 40 HIV− individuals who do not use METH. Using an ROI-based analysis, no interaction was observed between HIV and METH for cortical thickness measures, although trends were observed in the caudal anterior cingulate, caudal middle frontal, cuneus, isthmus cingulate, lateral orbital frontal, and precentral gyrus. However, these trends were not statistically significant after correction for multiple comparisons. The authors concluded that they saw no evidence that METH use exacerbates adverse effects of HIV on cortical thickness with any changes observed likely being additive. These results are similar to past research regarding the effects of METH and HIV [91, 92].

Conclusions

Due to increase in the life expectancy of PLWH and the prevalence of HAND, it has become increasingly imperative to identify the neuropathological changes due to HIV. As highlighted in this review, HIV triggers a strong immune response, leading to a cascade of events including increased chronic inflammation and cognitive decline. These outcomes are further exacerbated by age and age-related comorbidities, as well as lifestyle factors such as drug use/abuse. Advances in neuroimaging have greatly aided in the pursuit of understanding HIV pathology and have provided insight into the structural, functional, and molecular changes occurring in the brain due to HIV. However, despite these advances, more research is needed to fully understand the nuances of how the virus alters brain structure and functioning. Delineating such effects will ultimately enable increased individualized care and treatment for PLWH.