Introduction

The introduction of effective combination antiretroviral (cART) therapy has led to dramatic reductions in the incidence and prevalence of HIV-associated central nervous system (CNS) opportunistic infections and HIV encephalopathy [1, 2]. Despite the significant impact of cART on HIV-associated CNS disease, clinically significant cognitive deficits affecting people living with HIV have become increasingly apparent in recent years [3]. Prevalence rates of HIV-associated neurocognitive disorders, also known as HAND of between 15 to 50% have been observed across clinical and geographical settings, with effectively treated HIV-positive individuals having a higher prevalence of cognitive impairment compared to the general population [4–9]. HAND diagnosis is based on neuropsychometric testing and includes asymptomatic neurocognitive impairment (ANI), mild neurocognitive impairment (MCI), and the most severe form, HIV-associated dementia (HAD) [10]. In countries where cART is widely available, HIV-associated dementia has dramatically declined, however milder forms of cognitive impairment persist despite the availability of potent cART. HAND in people living with HIV has several possible pathogenic mechanisms including poor drug concentration of antiretrovirals in CNS, legacy effect of CNS damage sustained during early stages of HIV infection of the brain, antiretroviral neurotoxicity, persistent brain immune activation, and comorbidities such as cerebrovascular disease, syphilis, and hepatitis C co-infection [11–18].

Neuroimaging techniques that evaluate brain structure, metabolism, neurochemistry, and brain function have been used widely in clinical settings to aid in the diagnosis, and description of neurological disorders associated with HIV. Recent technical advances in these techniques offer new perspectives for studying the neuropathogenesis of HIV-associated brain disease in the era of cART. Positron emission tomography (PET) is now well established for clinical use in neurodegenerative disorders, especially in the diagnosis of dementia. The development of PET brain radiotracers that allow imaging of neuroinflammation, functional activity, neurotransmitter systems, and amyloid plaque deposition might have the potential to advance our knowledge of the pathophysiology of HAND. This review will highlight PET radiotracers used to study HIV-associated brain disease, focusing on PET neuroimaging studies that have contributed to our understanding of HAND in the era of combination antiretroviral therapy (Table 1).

Table 1 Summary of PET studies in HAND in HIV-positive individuals on cART
Full size table

PET-FDG brain imaging in HAND

Fluorodeoxyglucose [18F]-FDG PET brain imaging has been widely used in clinical and research settings for the evaluation of several neurological diseases, including HIV-associated brain diseases [30]. [18F]-FDG PET allows the in vivo quantification of cerebral glucose metabolism, which reflects neuronal and synaptic activity. [18F]-FDG PET has the ability to quantify regional differences in cerebral glucose metabolism, which can provide information about the distribution and activation of brain patterns. [18F]-FDG PET has proven to be useful as a diagnostic tool in the clinical evaluation of CNS lesions in patients with HIV, particularly in those with CNS opportunistic infections where [18F]-FDG PET can be used to distinguish between CNS malignancy and infection [31, 32]. [18F]-FDG PET has also been used to explore regional cerebral glucose metabolism patterns in HIV-positive patients with and without cognitive impairment. Early pre-cART studies using [18F]-FDG PET have shown consistently the presence of subcortical hypermetabolism in the basal ganglia, striatum, and thalamus of HIV-positive patients with early stages of HIV-associated dementia, as well as asymptomatic HIV-positive individuals, suggesting that increased glucose metabolism of subcortical structures, like the basal ganglia, are characteristic of HAND in patients without cART [33, 34]. A limited number of studies have investigated the impact of cART on cerebral glucose metabolism using [18F]-FDG PET. They have demonstrated changes in [18F]-FDG PET uptake despite effective virological suppression with cART. Changes include reductions in [18F]-FDG PET uptake in frontal regions in HIV-positive patients without cognitive impairment [19], and subtle basal ganglia hypermetabolism despite virological control with cART [20]. In summary, [18F]-FDG PET is helpful in the clinical evaluation of HIV positive patients with space-occupying lesions demonstrated on MRI or CT. However, the role of [18F]-FDG PET in the clinical assessment of HAND is less clear, as [18F]-FDG PET brain uptake correlates poorly with clinical and cognitive parameters in patients with different forms of HAND. Further research is needed to determine the role of [18F]-FDG PET in this context.

PET imaging of neurotransmitters in HAND

The presence of acute-onset parkinsonism in HIV-positive patients with HIV-associated dementia suggested a vulnerability of the dopaminergic system to the effects of HIV infection of the brain [35, 36]. This clinical observation was followed by reports of hypersensitivity to dopamine blockers, and the reduced concentration of CSF dopamine levels in HIV-positive patients with and without cognitive deficits [37]. Histopathological and [18F]-FDG PET imaging support the propensity of HIV-1 to affect basal ganglia structures [38, 39], which are known to have the highest density of dopaminergic terminals. Dopamine function can be assessed using PET imaging. [11C]-cocaine allows the assessment of presynaptic dopaminergic transporters (DAT) availability, while [11C]-raclopride measures D2 dopaminergic receptor availability, which mostly reflects postsynaptic sites [40]. Using these radioligands, Wang et al. found that although there were no differences in D2 receptors between HIV-positive patients and controls in any brain region, a significant decrease in DAT availability was observed in the putamen and ventral striatum of HIV-positive patients with HIV-associated dementia compared with HIV-negative controls. This small study (n = 15, HIV-positive individuals) was the first to suggest that decreased DAT might contribute towards the pathogenesis of HIV-associated dementia (Fig. 1) [21]. These findings were validated in a larger cohort of HIV-positive patients (n = 35) with cognitive impairment, but no HIV-associated dementia. In this study, compared to HIV-negative controls, HIV-positive individuals with cognitive impairment had reduced dopaminergic function (lower DAT), and this was associated with poorer cognitive performance [22]. Interestingly, in both studies, despite the significant decreases in dopaminergic function in the basal ganglia structures, there were only mild motor symptoms such as bradykinesia and rigidity, including intermittent choreiform movements in only three HIV-positive individuals with HIV-associated dementia. Further studies are needed to determine the effect of cART on DAT and D2 availability in HIV-positive individuals with and without cognitive impairment and its relationship with cognitive function.

Fig. 1
figure 1

Distribution volume ratio images of PET with 11C-cocaine (DA transporter) at the level of the basal ganglia. A significant decrease in DAT availability was observed in the putamen and ventral striatum of HIV-positive patients with HIV-associated dementia compared with HIV-negative controls. The images are scaled with respect to the maximum value obtained in the control subject and presented using the rainbow scale from [21] Wang et al. with permission

Full size image

PET imaging of neuroinflammation in HAND

Chronic activation of brain microglia has been suggested to be a major contributor towards HIV-associated brain disease [41]. In HIV, activation of microglia has been associated with several factors, including persistent low-level HIV RNA replication, reduced concentration of cART in the CSF, cART neurotoxicity, co-infections (hepatitis C, syphilis) and lifestyle factors (smoking alcohol and recreational drug use). All of these factors might lead to activation of microglia despite effective control of HIV RNA with cART [42]. PET imaging allows in vivo quantification of neuroinflammation by measuring the density of the translocator protein 18 kDa (TSPO). TSPO is highly expressed in the mitochondria of microglia and astrocytes [43]. Following activation through host responses to cellular injury, microglia and astrocytes increase expression of TSPO [44]. Increased binding of a TSPO radioligand therefore provides a proxy measure of brain microglial activation that can be assessed in vivo with PET. Few studies have investigated microglial activation using TSPO PET in people living with HIV, with the majority of studies employing [11C]-PK11195, a first-generation TSPO radioligand. The findings have been contradictory with some studies demonstrating differences in [11C]-PK11195 binding between HIV-positive individuals with and without cognitive impairment [27, 45] and others showing no significant differences between groups [26]. Possible explanations for these discrepancies could be related to the difficulties in making accurate measures of brain binding with this “first-generation” ligand because of a lower proportion of the signal that arises from specifically bound [11C]-PK11195 [46].

Second-generation TSPO radioligands such as [11C]-DPA-713 and [11C]-PBR28 have higher affinity for TSPO and significantly better sensitivity compared with [11C] PK11195 [47]. One of the limitations of second-generation radioligands is that affinity for the target protein is determined by the rs6971 single-nucleotide polymorphism (SNP) in the TSPO gene [48, 49] that leads to an amino acid substitution (Ala147Thr), which is associated with in vitro affinity of TSPO in platelets [50, 51]. Three patterns of TSPO binding phenotypes have been identified in humans: high affinity binders (HABs) who are those subjects without the polymorphism (HH), low affinity binders (LABs) who are homozygotes (LL) and mixed affinity binders (MABs) who are heterozygotes (HL) expressing both low- and high-affinity binding to TSPO. HABs and LABs express a single binding site for TSPO with either high or low affinity, whereas MABs express equal amounts of HAB and LAB binding sites. A genotype analysis in plasma is available to characterize expected TSPO affinity that allows the incorporation of genotypic data enabling more accurate quantitative interpretation of TSPO PET data [48]. Another limitation relates to the TSPO itself. TSPO is regulated by a number of physiological factors including stress, steroid, and cholesterol metabolism, which can affect the analysis of TSPO imaging studies, and might contribute to poor tests reproducibility [29]. To control for some of these factors, analysis of second-generation TSPO data using a brain region that serves as a region of reference has been suggested. Brain regions of reference are selected because they are believed to be brain regions were binding activity is nonspecific in the disease process studied, however when postmortem and receptor imaging studies are performed, a significant percentage of specific binding is encountered, complicating the interpretation of the data [52]. Studies using this model of analysis have employed the gray matter because of the uniform pattern of binding displayed in healthy subjects, but not HIV-positive patients using the first-generation TSPO radioligand [11C] PK11195.

Two studies using “second-generation” TSPO radioligands investigated microglial activation in people living with HIV on cART using the gray white matter (GM) as a region of reference. Coughlin et al. used [11C]-DPA-713 to evaluate regional brain distribution of TSPO in HIV-positive patients compared with seronegative controls [29]. The study compared 23 HIV-positive individuals with and without HAND with 12 HIV-negative controls. HIV-positive individuals demonstrated significantly higher TSPO normalized volume of distribution (VTGM) in the white matter, cingulate cortex, and the supramarginal gyrus. An increase in TSPO VTGM within the frontal cortex was specifically linked to individuals with HIV-associated dementia. Using [11C]-PBR28 PET, Vera et al. found evidence of increases in brain TSPO radioligand uptake, with the greatest uptake of [11C]PBR28 in subcortical brain grey matter, particularly in the basal ganglia (globus pallidus, caudate and striatum) in a group of cognitively healthy HIV-positive individuals without HAND and effective cART [18] (Fig. 2). In this study a significant relationship between increased TSPO distribution volume ratios VTDVR, in the basal ganglia and poorer cognitive performance in tasks assessing verbal and visual memory was observed, as well as white matter microstructural abnormalities on diffusion tensor imaging MRI [18]. These findings evidence that suggest that effective cART, HIV-positive patients microglial and possibly, astroglial activation is present and that microglial activation could potentially lead to the development of HIV-associated brain disease and HAND. More research is required to determine the feasibility of using TSPO PET in the clinical assessment of patients with HAND.

Fig. 2
figure 2

Volume of distribution parametric maps (0–90 min) of [11C]PBR28 TSPO PET in a 55-year-old HIV-positive man without cognitive impairment on the left and a 53-year-old control subject on the right. Both individuals were high affinity binders. Images are transverse, coronal, and sagittal, from top to bottom

Full size image

PET imaging of neuropathology in HAND

Neurodegeneration has been suggested as one the pathogenic mechanisms associated with HAND. However, until recently, it has not been possible to examine in vivo biomarkers of neurodegeneration in HIV-positive patients effectively treated with cART. The deposition of Aβ42−amyloid is considered a key marker of neuronal degeneration in Alzheimer’s disease (AD). Several reports have described a significant increase in brain Aβ42-amyloid deposition in patients with acquired immunodeficiency syndrome (AIDS) in brain pathology following post-mortem examinations [53, 54], as well as reductions in Aβ42-amyloid in CSF of treated HIV-positive patients with HAND [55–57], indicating increased deposition of Aβ42-amyloid in brain parenchyma. With the improvements in life expectancy of people living with HIV, increasing concerns have arisen as to whether HIV-positive patients are at increased risk of developing AD, and the challenges that clinicians will face differentiating between HAND and other neurodegenerative disorders in aging HIV-infected populations. The development of radiotracers such as the amyloid binding agents [11C]-labeled Pittsburgh Compound B [11C]-PIB, [18F]-flobetapir, [18F]-florbetaben, and [18F]-flutametamol offer the possibility to measure in vivo, the degree and distribution of Aβ42-amyloid deposition in the brains of HIV-positive individuals with cognitive impairment, and to distinguished between HAND and AD. In HIV-negative individuals, [11C]-PiB has demonstrated amyloid deposition in preclinical AD [58]. Two studies have used [11C]-PiB in HIV-positive individuals [23, 24] with unexpected results, as both studies were unable to demonstrate the presence of amyloid accumulation in patients with HAND. There are several possible explanations for these results. First, [11C]-PiB has significantly better affinity for fibrillar plaques rather than diffuse amyloid plaques, which have been traditionally associated with HAND. Second, there might be differences in amyloid metabolism between AD and HAND that [11C]-PiB is unable to detect. Finally, relatively younger individuals were included in these studies (the oldest was 67 years old) [59 ]. Indeed, Turner et al. reported for the first time the presence of in vivo amyloid deposition in a 71-year-old HIV-positive individual with HAND using [18F]-florbetaben, which has better affinity for diffuse amyloid plaques. [18F]-florbetaben PET CT showed abnormal scan appearances with pronounced cortical radiotracer deposition suggesting for the first time the presence of in vivo Aβ42-amyloid deposition in a person living with HIV [25]. Further research is required to determine the role of amyloid imaging in the clinical assessment of people living with HIV with cognitive impairment.

Conclusions and future directions

The introduction of cART has dramatically changed the pattern of brain disease associated with HIV infection. In settings where cART is readily available, CNS opportunistic infections and HIV-associated dementia are now rare, while patients on effective treatment are presenting with cognitive symptoms associated with HAND. This change in the natural history of HIV-associated CNS brain disease along with an aging HIV population at risk of neurodegenerative diseases generates new challenges to clinicians and researchers alike, as defining the cause of cognitive deficits in older HIV-positive patients can be difficult, due to the multifactorial nature of HAND. PET brain neuroimaging using novel radioligands might provide insight into different pathophysiological changes in the brain, which then combined with other neuroimaging techniques such as MRI could increase our knowledge of the relationship between structural, chemical, and functional changes in the brains of people living with HIV. In this respect, radioligands targeting microglial and/or astrocyte activation may be of particular importance, as neuroinflammation in HIV appears to be a major contributing factor for the development of HAND [60, 61]. PET brain tracers targeting inflammation, dopaminergic function, and possibly amyloid deposition have the potential to both serve as biomarkers supporting the diagnosis and management of HAND, and to facilitate the evaluation of therapeutic interventions. Considerably more research is needed to evaluate and establish the role of PET brain imaging techniques in the diagnosis and management of HAND in people living with HIV.