HIV-associated neurocognitive disorder — pathogenesis and prospects for treatment

Abstract

In the past two decades, several advancements have improved the care of HIV-infected individuals. Most importantly, the development and deployment of combination antiretroviral therapy (CART) has resulted in a dramatic decline in the rate of deaths from AIDS, so that people living with HIV today have nearly normal life expectancies if treated with CART. The term HIV-associated neurocognitive disorder (HAND) has been used to describe the spectrum of neurocognitive dysfunction associated with HIV infection. HIV can enter the CNS during early stages of infection, and persistent CNS HIV infection and inflammation probably contribute to the development of HAND. The brain can subsequently serve as a sanctuary for ongoing HIV replication, even when systemic viral suppression has been achieved. HAND can remain in patients treated with CART, and its effects on survival, quality of life and everyday functioning make it an important unresolved issue. In this Review, we describe the epidemiology of HAND, the evolving concepts of its neuropathogenesis, novel insights from animal models, and new approaches to treatment. We also discuss how inflammation is sustained in chronic HIV infection. Moreover, we suggest that adjunctive therapies — treatments targeting CNS inflammation and other metabolic processes, including glutamate homeostasis, lipid and energy metabolism — are needed to reverse or improve HAND-related neurological dysfunction.

Several advances have dramatically improved the care and prognosis of HIV-positive (HIV+) individuals in the past 20 years, changing the course of HIV from a life-limiting infection that was frequently complicated by fatal opportunistic infections and malignancies to a chronic infection that can be managed, is associated with a near-normal lifespan and in which opportunistic infections are rare¹.

The first major advancement was an understanding of the direct relationship between HIV replication and subsequent immunological and clinical progression. This finding emphasized the need to completely suppress HIV replication in order to control disease progression.

The second major advancement was the development and deployment of combination antiretroviral therapy (CART), which can provide effective systemic suppression of HIV replication. The introduction of CART in the mid-1990s resulted in a 50% decline in the rate of death from AIDS, substantial decreases in rates of maternal–infant transmission, reduced incidence of opportunistic infections, and a 40–50% decrease in the incidence of HIV-associated dementia (HAD), which was previously common and is the most severe form of cognitive impairment associated with the infection².

The third major change in the care of HIV+ patients was the ability to monitor the efficacy of CART through the reliable and widespread measurement of CD4⁺ helper T cells, plasma HIV RNA levels and antiretroviral resistance profiles, all of which are now fully integrated into routine clinical care in the developed world and used to optimize treatment for individual patients. Plasma viral load is now routinely monitored in HIV+ patients to ensure complete systemic viral suppression, and resistance profiles are used to monitor for the development of resistance to antiretroviral therapy and to choose CART regimens with optimal effectiveness in a given patient³.

The latest major advancement in the care of HIV+ individuals is the recommendation to begin CART as soon as a patient is willing to commit to this lifelong therapy, regardless of CD4⁺ T cell count³. This recommendation is bolstered by results from the recent Strategic Timing of Antiretroviral Therapy (START) trial, which proved the benefits and safety of earlier CART initiation on overall outcomes in HIV infection⁴. In this international study of >4,600 CART-naive HIV+ patients with CD4⁺ T cell counts >500 cells/μL, immediate initiation of CART resulted in fewer AIDS-related and non-AIDS related events than did deferring initiation of CART until CD4⁺ T cell counts fell below 350 cells/μL.

Key points.

• Despite entering the era of combination antiretroviral therapy (CART), HIV-associated neurocognitive disorder (HAND) remains prevalent; however, less severe forms of HAND now predominate, and the most severe form, HIV-associated dementia, is rare

• In individuals treated with CART, the risk of HAND increases with age and in the presence of cardiovascular disease risk factors

• Latent HIV can persist in the brain even when systemic virological control is achieved with CART, thereby hampering efforts to eradicate HIV

• Animal models of CNS HIV infection — such as macaques infected with simian immunodeficiency virus — develop severe HAND, viral encephalitis and neuronal apoptosis, and are central to understanding the immunopathogenesis of HIV-induced CNS damage

• A growing body of work indicates that mild HAND can be modelled in immunocompetent mice infected with chimeric HIV (a model known as EcoHIV), and in chronically HIV-infected immunodeficient mice reconstituted with human immune systems

• To date, clinical trials of HAND therapies have been unsuccessful, but further trials for the treatment of HAND are forthcoming, including a trial of intranasal insuli

For almost a decade, the term HIV-associated neurocognitive disorder (HAND) has been used to describe the range of neurocognitive dysfunction associated with HIV infection⁵. Just as the course of HIV or AIDS has changed significantly over the past two decades, so has the course of HAND (FIG. 1). Nevertheless, despite our increasing knowledge and understanding of HAND, there is still no definitive marker or specific treatment: CART is the only option to prevent or delay the progression of HAND, but it is effective only in a subset of patients. The development of HAND remains an important issue for HIV+ patients, as it affects not only survival and quality of life, but also everyday functioning⁶. Worldwide, HAND remains a common cause of cognitive impairment and has persisted even in individuals who have received CART^7,8. As CART becomes more widely distributed in resource-limited settings and improves survival, the long-term global impact of HAND will become even more significant. In addition, early HIV infection of the CNS is believed to contribute to the development of HAND, and evidence suggests that the CNS can subsequently serve as a reservoir for ongoing HIV replication, thereby limiting the opportunity for a sterilizing cure or eradication⁹.

Figure 1. Timeline of advances in neuro-AIDS research.

Since the discovery of AIDS in 1981 and HIV in 1983, important advances have been made in research into and the prevention and treatment of HIV-associated neurocognitive disorder (HAND). AZT, azidothymidine; CART, combination antiretroviral therapy; HNRC, HIV Neurobehavioral Research Center; MACS, Multicenter AIDS Cohort Study; UCSD, University of California San Diego.

This Review will focus on HAND, describing its changing epidemiology and its neuropathogenesis, including recent insights from animal models. We will review known risk factors for HAND and consider projections of the epidemiology in resource-limited countries and among the ageing population with HIV infection. We will also detail the evidence for early brain infection and the brain as a sanctuary for HIV, as well as considering how and why inflammation is sustained in chronic HIV infection, even when systemic virological control is achieved. Finally, we will discuss new approaches to the treatment of HAND and their implications in an era when HIV eradication might be feasible. In particular, adjunctive therapies targeting CNS inflammation and other metabolic processes, including glutamate homeostasis, lipid and energy metabolism, could be necessary to prevent or improve HAND-related neurological dysfunction¹⁰.

Clinical features of HAND

Epidemiology and risk factors

HAND refers to a spectrum of neurocognitive impairment that includes asymptomatic neurocognitive impairment (ANI), mild neurocognitive disorder (MND), and HIV-associated dementia (HAD), and is diagnosed using neuropsychological testing and functional status assessments⁵ (TABLE 1). In the pre-CART era, HAD was the most common form of HAND and was almost invariably fatal^11,12. However, the prevalence of HAD has substantially declined with the widespread implementation of CART. Before 1991, 20% of participants enrolled in the Multicenter AIDS Cohort Study (MACS) met the criteria for HAD, but only 5% met the criteria in 2001–2003 (REFS 13,14). As a result, HAD — the most severe form of HAND — is rare in the developed world today.

Table 1.

Classification of HIV-associated neurocognitive disorders

HIV-associated neurocognitive dysfunction (HAND) type*	Prevalence in CART-treated HIV+ individuals	Diagnostic criteria5
Asymptomatic neurocognitive impairment (ANI)	30%	Impairment in ≥2 neurocognitive domains (≥1 SD) Does not interfere with daily functioning
Mild neurocognitive disorder (MND)	20–30%	Impairment in ≥2 neurocognitive domains (≥1 SD) Mild to moderate interference in daily functioning
HIV-associated dementia (HAD)	2–8%	Marked (≥2 SD) impairment in ≥2 neurocognitive domains Marked interference in daily functioning

SD, standard deviation.

^*With no evidence of other cause. Adapted from Antinori, A. et al. Neurology 69, 1789–1799 (2007).

Changes in HAND severity in the CART era

Despite improved life expectancies and a dramatic decline in the rates of CNS opportunistic infections in HIV+ people as a result of CART, HAND remains a major cause of morbidity: 15–55% of HIV+ individuals are estimated to have HAND — a proportion that remains similar to that in the pre-CART era^7,14 (FIG. 2). It should be noted, though, that these HAND cases primarily represent the more mild forms of the condition^7,14. In both the USA and Sub-Saharan Africa, patients receiving CART have much better neuropsychological function than CART-naive patients or individuals treated with zidovudine monotherapy^13,15. As a result, the prevalence of milder forms of HAND has increased so that ANI now accounts for approximately 70% of all forms of HAND⁷.

Figure 2. More-effective therapies have reduced the severity of HIV-associated the severity of HIV-associated neurocognitive disorders.

Since the introduction of combination antiretroviral therapies (CARTs) in 1996, the proportion of HIV+ individuals with neurocognitive symptoms has remained unchanged, but the proportion of people with severe symptoms has declined so that HIV-associated dementia (HAD) is much less common and asymptomatic neurocognitive impairment (ANI) now accounts for the majority of cases. MND, mild neurocognitive disorder. Adapted from McArthur, J. C. et al. Ann. Neurol. 67, 699–714 (2010).

Conversion from asymptomatic to symptomatic HAND

Despite being asymptomatic, ANI is clinically relevant because individuals with ANI can transition to one of the more severe forms of HAND: for example, participants of the CNS HIV Antiretroviral Therapy Effects Research (CHARTER) study who had ANI at baseline were two to six times more likely to develop symptomatic HAND during several years of follow-up than those who were neurocognitively normal at baseline¹⁶. The increased risk of conversion to symptomatic dementia for individuals with ANI could reflect the finding that some individuals have very early involvement of the brain after HIV infection. For example, structural brain changes can sometimes be identified by neuroimaging within 100 days of primary infection, even in the absence of symptomatic involvement^17,18. However, the term ANI should be reserved for research studies, as its use in clinical settings remains controversial.

HAND and immunosuppression

Besides the reduced severity of HAND, other epidemiological features of the condition have also changed in the CART era. For example, in the pre-CART era, HAD was primarily seen in advanced HIV disease¹⁹. Although HAD is much less prevalent in patients receiving CART, when it does occur, it now often does so in patients with less severe immunosuppression²⁰. Moreover, in the pre-CART era, low CD4⁺ T cell counts²¹ and high plasma and cerebrospinal fluid (CSF) viral loads^20,22 were associated with HAD, but these biomarkers of viral infection are not consistently associated with cognitive impairment in CART-treated patients²⁰, and new predictive biomarkers are being sought. On the other hand, CD4⁺ T cell count nadir remains strongly associated with HAND, even in virologically suppressed patients on CART, and a history of clinically-defined AIDS is associated with onset of cognitive impairment at a younger age (<50 years)^7,20,23,24.

HAND progression

Perhaps surprisingly, the true impact of CART on HAND remains ill-defined. HAND generally remains stable during CART, but rarely resolves completely. A 4-year study of 197 CART-treated individuals demonstrated that 77% remained neurocognitively stable, with only 13% deteriorating to a more severe form of HAND, and 10% improving²⁵. Thus, HAND is typically not progressive in the majority of aviraemic HIV+ individuals on CART. The fact that lower CD4⁺ T cell nadir is a risk factor for HAND⁷ suggests that earlier HIV treatment to prevent severe immunosuppression could reduce the severity of HAND (that is, shift the phenotype from HAD to MND or ANI). However, the recent START trial failed to confirm a major effect of early CART on HAND^4,26. Although immediate initiation of CART resulted in fewer AIDS-related events than did deferred initiation of CART, performance on neuropsychological testing did not differ between these groups after a mean of 3.3 years of follow-up, indicating that earlier treatment might not markedly affect the development of HAND²⁷.

Clinical characteristics of and risk factors for HAND have also changed with CART. Typically, HAND presents with executive dysfunction and memory impairment with prominent disruption of attention, multitasking, impulse control, judgement and memory encoding and retrieval. HAND can also be associated with motor dysfunction, including bradykinesia, loss of coordination and gait imbalance. Whereas deficits in motor skills and psychomotor speed were the most common manifestations of HAND before CART, deficits in learning and/or memory and executive function are more common symptoms today²⁰.

Risk factors for HAND

Cardiovascular risk factors

Cardiovascular risk factors were linked to poorer cognitive performance in the MACS and other cohort studies, and central obesity and diabetes were important risk factors for HAND in the CHARTER cohort^28,29. In a study of 245 HIV+ individuals in Italy, diabetes, carotid intima-media thickness and cardiovascular risk factors — including hyperlipidaemia and tobacco use — were strongly associated with lower cognitive performance³⁰.

Age

Older age is associated with an increased risk of HAND. Older HIV+ adults (>50 years) in the Hawaii Ageing Cohort were twice as likely as their younger HIV+ counterparts to have HAD after adjusting for other known dementia risk factors^31,32. Older age has also been associated with increased risk of HAND in several South African cohorts^33,34; however, the effect of age on HAND risk is likely confounded by the increased prevalence of cerebrovascular risk factors at older ages^26,35. In the CHARTER cohort, older age, elevated systolic blood pressure, high BMI, high serum cholesterol, and a diagnosis of AIDS were associated with worse global neuropsychological performance, suggesting that small or large vessel atherosclerotic disease could be contributing to cognitive impairment in older HIV+ individuals³⁶.

Hepatitis C virus infection

In one cohort, hepatitis C co-infection was found to nearly double the risk of cognitive impairment in HIV+ individuals compared with those without hepatitis C³⁷; however, a later study found hepatitis C infection to have no effect on the cognitive performance of HIV+ individuals³⁸. With the prospect of a true cure for hepatitis C virus with new and potent antiviral regimens, this controversial point needs further research.

Additional risk factors

Substances of abuse, particularly methamphetamine, have deleterious effects on cognition, which are more pronounced when combined with HIV infection³⁹. Cognitive reserve could also be important: in the MACS, cognitive impairment was observed in 38% of HIV+ participants with less than a high-school education, but only 17% of HIV+ participants with at least a high-school education¹³. Clinically, several comorbidities can contribute to cognitive impairment in HIV+ individuals (BOX 1). If these comorbidities are present, it is difficult to ascertain with certainty whether cognitive impairment is caused by direct effects of HIV, direct effects of the comorbidities, or a combination of both, thereby making a definitive diagnosis of HAND more difficult. Additional risk factors for HAND in the CART era are summarized in BOX 2.

Box 1. Comorbidities to be considered in HAND.

• Age-related cognitive deficits

• Alcohol and substance abuse

• Viral co-infection: HCV, HIV-2 and HTLV-I

• Nutritional and vitamin deficiencies

• Accelerated atherosclerosis

• Traumatic brain injury

• Obstructive sleep apnoea, disturbed sleep

• Psychiatric illnesses: anxiety disorders, major depression, bipolar disorder

HAND, HIV-associated neurocognitive syndrome; HCV, hepatitis C virus; HIV-2, human immunodeficiency virus type 2; HTLV, human T-lymphotropic virus I.

Box 2. Risk factors for HAND in the CART era.

• Advanced age

• Low CD4⁺ T cells nadir

• Use of illicit drugs

• Hepatitis C co-infection

• Cerebrovascular disease risk factors: diabetes, hypertension, hypercholesterolaemia, obesity

• Sleep disorders: insomnia, obstructive sleep apnoea, sleep fragmentation

• Psychiatric comorbidity: major depression, anxiety disorders, bipolar disease

CART, combination antiretroviral therapy; HAND, HIV-associated neurocognitive disorder.

HAND in the global setting

More than 70% of the world’s HIV+ population lives in Sub-Saharan Africa, and HIV+ individuals throughout this region tend to have worse neuropsychological function than HIV-uninfected (HIV−) counterparts^33,40–45. In a study of HIV+ outpatients in Uganda, nearly one-third met the criteria for HAD, with advanced age and low CD4⁺ T cell count conferring increased risk⁴⁶. In a South African study, 25% of CART-naive HIV+ individuals met the criteria for HAD, and an additional 42% met the criteria for MND³³. Large longitudinal studies from resource-limited settings are lacking, but if these proportions are accurate, HAND would be the most common form of young-age neurocognitive impairment worldwide. In addition, owing to an increased risk of HAND with advancing age, the burden of HAND in the developing world might continue to increase as life expectancy rises with increased availability of CART. However, this change could be offset by the wider deployment of CART: the AIDS Clinical Trials Group (ACTG) A5199 trial tested the effects of different CART regimens in resource-limited settings and showed improved neuropsychological performance, regardless of the specific regimen⁴⁷. CART was also shown to improve cognitive performance and everyday function in a clinical setting in Uganda¹⁵.

Biomarkers in HAND

Although a variety of potential biomarkers for HAND have been identified (TABLE 2), the majority of these are actually markers associated with HAD rather than ANI and MND, which currently represent much more common forms of cognitive impairment. Validated biomarkers for these more common forms are desperately needed to more accurately diagnose and delineate the early stages of HAND (given that it is often very challenging to distinguish ANI and MND from other comorbidities) or, more importantly, to predict the trajectory of cognitive function in HIV+ patients. Biomarkers that could identify a preclinical stage of HAND or predict cognitive worsening (expected in 23% of patients with ANI⁴⁸) would open the possibility of treatment at the earliest stage of neurological decline, when interventions are likely to have the greatest impact. Equally important is the identification of biomarkers that are associated with cognitive improvement, which would enable more accurate assessment of the interventions in phase I/II clinical trials that are too short for reliable assessment of effectiveness with neuropsychological tests. Finally, understanding the underlying molecular mechanisms and how biomarkers differ across the spectrum of HAND will ultimately facilitate the identification and development of precision therapeutics.

Table 2.

Surrogate biomarkers for cognitive status in HIV

Pathophysiological mechanism	Primary infection	ANI/MND	HIV-associated dementia
Cell stress	Not determined	CSF sphingomyelin* (REFS 72,75,76) CSF ceramide* (REFS 72,75,76) CSF esterified cholesterols* (REFS 75,76); sphingomyelinase activity76 CSF hydroxynonenals72	Brain and CSF sphingomyelin72 Brain and CSF ceramide72
Neuronal injury/protection	CSF NFL95 CSF p-tau95 CSF β-amyloid95 Brain NAA224	Not determined	CSF NFL82,91,92,225 CSF t-tau91; CSF sAPPβ91 Brain NAA70,79 CSF quinolinic acid226
Oxidative stress	Not determined	CSF haem oxygenase-1* (REF. 142) CSF hydroxynonenals72,227 CSF protein carbonyls228	CSF haem oxygenase-1 (REF. 142) Brain SOD-1 (REF. 229) Brain iNOS229 CSF 3-nitrotyrosine230 CSF protein carbonyls228
Energy metabolism	Brain choline57,224	CSF Krebs cycle substrates*‡§ (REF. 73) CSF triglycerides‡ (REF. 75) CSF fatty acids* (REF. 76)	Brain choline79
Immune activation	Plasma sCD163 (REF. 56) Brain myo-inositol224 Plasma IL-1α, IL 12, TNF, lymphotoxin, IL-10, IP-10, MCP-3, eotaxin, IFN-α17 Plasma MMP 2 (REF. 55)	Brain HLA-DR72 Panel of cytokines from plasma and CSF*‡§ (REF. 96) Plasma65 and CSF67 sCD14 Plasma sCD163 (REF. 66)	CSF MCP-1 (REFS 231,232) Plasma TNF231 Brain HLA-DR72 Brain myo-inositol70,79 Plasma sCD14 (REFS 65,233) CSF neopterin66,234 CSF osteopontin235 CSF fractalkine236 Brain IL-1β229 Brain IL-10 (REF. 229) Brain STAT-1 (REF. 237) CSF S100β238
Glutamate regulation	Brain glutamate224	Brain glutamate* (REF. 77) CSF glutamine*‡ (REF. 73)	Brain Glx70 Brain glutaminase C237

ANI, asymptomatic neurocognitive impairment; sCD, soluble cluster of differentiation; CSF, cerebrospinal fluid; Glx, glutamate/glutamine complex; HLA DR, human leukocyte antigen–antigen D related; IFN, interferon; iNOS, inducible nitrous oxide synthase; IP, inducible protein; MCP, monocyte specific chemokine; MMP, matrix metalloprotease; MND, mild neurocognitive disorder; NAA, N-acetyl aspartate; NFL, neurofilament light chain; p-tau, phosphorylated tau; sAPPβ, soluble amyloid precursor protein beta; STAT, signal transducers and activators of transcription; SOD 1, superoxide dismutase 1; TNF, tumour necrosis factor; t-tau, total tau.

^*Changes in these markers indicate cognitive decline in HIV+ patients.

^‡Changes in these markers indicate cognitive improvement in HIV+ patients.

^§Some of the tested molecules were associated with cognitive decline, some with cognitive improvement.

Biomarkers for HAND can be broadly classified into four groups: soluble markers of immune activation, markers of metabolic or cellular stress, neuronal injury markers, and neuroimaging markers. As several recent reviews have summarized the current knowledge of neuroimaging markers at various stages of HAND^49–52, we will only briefly discuss selected imaging markers that are related to soluble indicators. Our focus here is to briefly discuss biomarkers in the context of HAND staging, and the progress made in informatics approaches that attempt to incorporate biological markers with clinical and demographic features so as to identify groups of prognostic indicators for change in cognitive status.

Innovative approaches to foster community networking between populations that are at risk of HIV infection have provided unique opportunities to study biological and neuropsychological changes that occur at very early time points following HIV infection. These studies have revealed that neurological involvement occurs very rapidly following infection. Early changes in brain structure, including increased permeability of the blood–brain barrier, reductions in brain volume^17,53 and decreases in diffusion measures of white matter⁵⁴, can appear within the first few months following infection. These structural modifications are accompanied by increases in circulating levels of inflammatory cytokines, immune activation^17,55,56, evidence of acute metabolic disturbances⁵⁷, and measurable deficits in cognitive and psychomotor functions⁵⁸. Many of these structural, cognitive and inflammatory modifications do not improve to a clinically relevant degree following the initiation of CART^57–59 and can persist or even worsen over time in some individuals, despite systemic viral suppression^60–64. These pathological modifications have considerable individual variability, and it is currently not known how these very early events affect the development and/or trajectory of cognitive impairments later in life. Continuing to follow these patients from very early infection through long-term CART will provide valuable information on how individual responses to infection affect long-term outcomes.

Biomarkers implicated in HAND

In individuals who develop ANI and MND, markers of immune^65–67 and cytokine activation^68,69 are more pronounced than in HIV+ individuals without cognitive impairment. Likewise, changes in bioenergetics, as measured with neuroimaging^70–72 and metabolomic appraches^73,74, are readily apparent in individuals with ANI or MND compared with cognitively normal HIV+ individuals, as are accumulations of bioactive lipids, such as ceramide, and sterol markers of cell stress^75,76. Brain structural changes⁸⁴ and progressive impairments in energy and lipid metabolism^75,77–80, immune regulation^67,81,82 and metabolism^74,83 worsen with age and duration of infection^50,85–89. Whether markers of neuronal and axonal injury are elevated during acute infection is not entirely clear. Initial studies reported that markers of neuronal injury, such as neurofilament light chain protein, tau and amyloid precursor proteins, were not elevated during early stages of ANI, or at early time points following the initiation of CART^90,91. Rather, these markers typically appear in the mid to late stages of the disease process^91–93 and correlate with low CD4⁺ T cell counts⁹⁴ and MRI markers of neuronal damage⁹⁵. However, ultrasensitive measures of neurofilament light chain levels identified that 44% of people with primary HIV infection did show elevated markers of neuronal damage, suggesting that a subset of newly infected individuals show signs of neurological injury. These findings demonstrate that neurological involvement occurs within months of initial HIV infection⁹⁵, but the precise relationship between this early involvement and the onset of cognitive impairment is unknown. The interval between early involvement and clinical HAND symptoms suggests the existence of a lengthy therapeutic window, during which targeted interventions could preserve cognitive function.

The need for composite biomarkers

Our current lack of clinically validated biomarkers for ANI and MND suggests that any single biomarker might be insufficient to identify early stages of HAND, and that alternative methods that can identify combinations of markers might facilitate efforts to reliably identify the earliest stages of HAND in clinical practice. Advanced statistical approaches, such as machine learning and multivariate statistical modelling, have been increasingly used to interrogate complex biological and clinical data. For neurodegenerative conditions, these approaches are based on the notion that complex diseases such as HAND can be better understood by incorporating multiple biological and clinical variables using nonparametric approaches. Studies using these approaches have begun to identify clinical and demographic factors, including time-dependent treatment effects, historical or current comorbid conditions and metabolic pathways associated with lipid metabolism, bioenergetics and inflammation, that are progressively perturbed during the onset and worsening of ANI and MND^73,75,96. Validation of these models and of the underlying mechanistic pathways identified will be critical in assessing their utility for clinical practice and the identification of molecular targets for precision therapeutics.

Pathogenesis in HAND

The neuropathology of HAND has changed considerably since the introduction of CART: the frequency with which HIV encephalitis is observed at autopsy has reduced from 54% before CART to 15% in the CART era⁹⁷. Encephalitis and outright neuronal loss were historically thought to have central roles in HAND, but these factors are no longer sufficient to explain neurological dysfunction in the CART era, as these pathologies are no longer typical⁹⁸. The paucity of overt neuropathology specific to HIV infection in CART-treated patients suggests that the underlying pathophysiology of HAND is more likely to be associated with functional alterations in neurons (FIG. 3). This paradigm shift necessitates new therapeutic strategies tailored to preserve brain function in CART-treated patients.

Figure 3. Neuropathogenic mechanisms that contribute to HIV-associated neurocognitive disorders.

HIV-infected macrophages and microglial cells release neurotoxic viral proteins that trigger astrocyte activation, which results in increased glutamate release and reduced glutamate uptake. Elevated extracellular glutamate levels cause neuronal bioenergetic disturbances that lead to aberrant synaptodendritic pruning and neuronal injury. Moreover, systemic inflammation and microbial translocation products lead to microglial activation and increased production of chemokines and cytokines that contribute to neuronal injury. Adapted from Williams, D. W. et al. Curr. HIV Res. 12, 85–96 (2014).

Inflammation and HIV neurotoxicity in HAND

Compelling evidence suggests that inflammation has an important role in triggering events that lead to neurodegeneration in HIV infection. However, robust inflammation is not always seen in HAND, especially early in the disease process.

Navia et al.⁹⁹ provided one of the first comprehensive neuropathological studies of HAD in 1986, describing their findings as follows: “Most commonly noted was diffuse pallor in the white matter, which in the pathologically milder cases was accompanied by scanty perivascular infiltrates of lymphocytes and brown-pigmented macrophages, and in the most advanced cases by clusters of foamy macrophages and multinucleated cells associated with multifocal rarefaction of the white matter. However, in nearly one third of the demented cases the histopathological findings were remarkably bland in relation to the severity of clinical dysfunction. In addition, similar mild changes were noted in over one half of the non-demented patients, consistent with subclinical involvement”. Their observations remain relevant today. Contemporary pathological studies in CART-treated individuals do not usually report white matter pallor, but more-subtle changes in white matter integrity are apparent on diffusion tensor imaging; these changes seem to worsen with increasing age and duration of infection^85,100–105. Activated circulating monocytes continue to have a critical role, both for the introduction of HIV into the brain via transmigration across the blood–brain barrier in response to chemotactic signals expressed within the parenchyma, and for the subsequent establishment of infection within CNS perivascular macrophages^106–108, microglia¹⁰⁹ and astrocytes¹¹⁰. Although astrocytes do not seem capable of producing intact virions under normal conditions, they can produce and export non-structural proteins such as tat (an HIV transcription factor), Rev, and Nef, all of which promote inflammation and neuronal damage^111–113.

One question that remains unanswered is why CNS inflammation is sustained even when the initial stimulus — viral replication — is suppressed by CART^95,114. According to one hypothesis, the inflammatory responses initiated by HIV direct the proteasome to become an ‘immunoproteasome’ that impedes turnover of folded proteins in brain cells and affects cellular homeostasis and response to stress¹¹⁵, resulting in perturbed neuronal and synaptic protein dynamics, possibly contributing to HAND. Another postulated mechanism for the sustained CNS inflammation is microglial priming from circulating microbial translocation products derived from gut bacteria and a disturbed microbiome¹¹⁶. It has also been suggested that the CNS inflammation in CART-treated individuals could be an attenuated form of immune reconstitution inflammatory syndrome¹. The role of genetic control of inflammatory responses, specifically, polymorphisms in genes encoding CCL3L1 and CCR5, has also been invoked to explain the individual variability in the course of HAND as well as the occurrence of HAND in only a subset of HIV+ individuals¹¹⁷.

The brain as a reservoir for HIV persistence

The importance of the brain as a potential reservoir for persistent HIV infection has gained renewed importance, with the intense focus of efforts being on eradication strategies. Clearly, a true sterilizing cure cannot be achieved if the brain harbours latent HIV that can be reactivated and can then reseed systemic infection. Simian immunodeficiency virus (SIV) models (see below) have produced robust evidence for the persistence of SIV DNA even after prolonged suppression of viral replication with CART¹¹⁸. Numerous studies of HIV RNA levels in the CSF have also suggested the presence of latent infection in the brain. The phenomenon of CSF viral escape supports the concept of a CNS reservoir⁹. This phenomenon can occur in as many as 5–10% of CART recipients and is associated with immune activation¹¹⁹ and major depressive disorder¹²⁰.

Clearance of both latent and productive HIV from the brain must underpin successful eradication. Macrophages and microglia, the cells within the brain that harbour HIV and produce infectious virions, are very long-lived, with turn-over rates of months or years¹²¹. The importance of the relative penetrance of different antiretrovirals into the brain remains debated¹²², but it is certainly plausible that lower concentrations of antiretrovirals within the CNS might lead to sub-optimal virological suppression. The cellular pharmacology of CART is relatively understudied with regard to tissue macrophages. A much higher maximal effective concentration (EC₅₀) in macrophages than in lymphocytes might reduce CART efficacy in this cell type¹²³. Furthermore, up to 20% of astrocytes isolated directly from autopsy brain tissues of HIV+ individuals contain integrated HIV¹²⁴. The ability of HIV-1 to integrate into terminally differentiated astrocytes suggests a permanent reservoir of provirus in the brain that influences the development and likely success of strategies aimed at eradicating HIV-1.

Early and progressive disturbances of bioenergetics

The sparing of neurons in CART-treated individuals with HAND suggests that functional changes underlie cognitive impairment in HIV+ individuals. Several lines of evidence support the notion that a loss of bioenergetic homeostasis could be an early event that primes the CNS for functional deficits. Brain gene expression profiling studies have identified that HIV+ patients exhibit widespread alterations in the expression of genes that regulate brain energy metabolism, and dysregulation progresses over time¹²⁵. PET studies have revealed varying degrees of reduced glucose uptake in the mesial frontal gyrus¹²⁶, as well as evidence for small but consistent age-related reductions of glucose uptake in the anterior cingulate cortex⁷¹ in individuals with undetectable plasma viral loads. Moreover, magnetic resonance spectroscopy imaging studies have identified progressive abnormalities in levels of choline, N-acetylaspartate, glutamate and glutamine-containing compounds in multiple brain regions of HIV-infected individuals on CART^77,78; these abnormalities correlated with deficits in motor and psychomotor speed, attention and working memory^70,77,78. High field-strength nuclear magnetic resonance (NMR) spectroscopy analysis of energy metabolites in the CSF has revealed accumulation of specific tricarboxylic acid cycle and glycolytic intermediates that were associated with changes in cognitive status in CART-treated HIV+ patients⁷³. The pattern of change in these metabolic features suggested that worsening cognitive function is associated with increased aerobic glycolysis, and improvements in cognitive function are associated with a shift in metabolism to promote anaerobic glycolysis. These disruptions in cellular energetics could explain abnormal accumulation of sphingolipids and proteins, such as amyloid-β, in dysfunctional endolysosomal compartments^{72,75,76,80,127–129}: reduced or modified cellular energy production would impair the function of the proton pumps that are dependent on adenosine triphosphate (ATP) — these pumps are necessary to maintain an acidic luminal pH, which is required for efficient functioning of >50 hydrolytic enzymes in lysosomes that degrade cellular products^128,130,131. The discussed studies have identified several possible targets for therapeutic intervention that include modulators of glucose metabolism (intranasal insulin), ceramide metabolism^72,132,133 and endolysosomal function¹³¹. Small-molecule therapeutics designed to affect these targets are in early to mid-stage development^131,134, and a clinical trial is planned to determine the effectiveness of intranasal insulin to treat ANI and MND in HIV+ patients.

Evidence for abnormal glutamate homeostasis

Both viral and host factors are thought to perturb brain glutamate metabolism and neurotransmission, and thereby have an important role in the development of HAND^135,136. In support of this notion, CSF levels of glutamate are fivefold greater in HIV+ individuals than in healthy controls¹³⁷, and recent studies of HIV+ patients who received CART revealed selective increases in CSF levels of glutamate in patients with HAND compared with patients without neurocognitive impairment⁷⁴. Tat and envelope glycoprotein gp120 have been shown to decrease glial and synaptic glutamate uptake^138,139, stimulate glutamate release from nerve endings^138,140, and phosphorylate glutamate receptors, thus potentiating the toxicity of the neurotransmitter¹⁴¹. Although glutamate levels in the CSF are increased in HIV+ patients with cognitive impairment, glutamate levels are selectively lower in the parietal grey matter, basal ganglia and cortex^77,78. These findings are consistent with incomplete recycling of glutamate by the glutamate–glutamine shuttle, leading to increased output of glutamate into the extracellular space and a reduction in the total amount of intraneuronal glutamate. HIV-infected macrophages have been shown to release ATP, which triggers production of neurotoxic levels of glutamate¹⁴² and decreases expression of the cytoprotective enzyme haem-oxygenase-1.

Several strategies for modulating glutamate-mediated neuronal toxicity have been evaluated. Early work focused on N-methyl-D-aspartate (NMDA) receptor antagonism, in particular the use of 1-amino-3,5-dimethyladamantane (memantine), a noncompetitive, low-affinity antagonist approved for treatment of Alzheimer disease. Unfortunately, despite preclinical studies suggesting a promising efficacy^143–145, initial clinical trials in HAND showed no effect¹⁴⁶. One alternative to direct receptor blockade is the modulation of enzymes that are responsible for production of glutamate, such as glutaminase and glutamate carboxypeptidase II. Glutaminase mRNA and protein are strongly upregulated following HIV infection, and inhibition of this upregulation blocks glutamate release and provides neuroprotection in pre-clinical models of HAND^147,148. Similarly, small-molecule inhibitors of glutamate carboxypeptidase II, which blocks conversion of the abundant neuropeptide N-acetyl-aspartylglutamate into glutamate, have been shown to protect against gp120-induced toxicity¹⁴⁹. Another suggested approach is to regulate the transporters responsible for modulation of extracellular glutamate, such as the cystine–glutamate transporter, which is profoundly upregulated in microglia that are activated by tat¹⁵⁰. Unfortunately, no clinically available brain-penetrating inhibitors exist to test this hypothesis in patients, but preclinical development is underway^151,152.

Animal models of neuro-HIV

Primate models

SIV-infected macaque models are of great value for studying the pathogenesis of HAND, including attempts to discover neuroprotective host genes and predictive plasma and CSF biomarkers¹⁵³. SIV-induced neuropathology closely resembles HIV-induced alterations, including multifocal perivascular aggregates in the brain that are composed of macrophages and multi-nucleate giant cells that contain replicating virus¹⁵⁴. Macaque models are particularly useful for studying the neuropathogenesis of HIV because plasma, CSF, and CNS samples can be obtained at multiple time points throughout infection, from acute through asymptomatic to terminal stages. In addition, SIV-infected macaques can be treated with suppressive antiretroviral therapy to study HAND in the context of treatment¹⁵⁵.

Various SIV models have been established to study HIV-induced CNS damage. One such model uses intravenous inoculation of pigtailed macaques with both a neurovirulent clone, SIVmac/17E-Fr, and an immunosuppressive swarm, SIV/DeltaB670. With this inoculation combination, approximately two-thirds of macaques develop SIV encephalitis within 84 days^156,157. Of interest is the fact that, though most studies of SIV pathogenesis use rhesus macaques, pigtailed macaques develop CNS disease more often than do rhesus macaques that receive the same SIV inoculum¹⁵⁸. Key features of the pigtailed macaque SIV model include development of CNS inflammation that correlates with high viral load in the brain, cognitive and motor deficits typical of HIV, and the classic lesions of HIV encephalitis^{156,159–163}. Furthermore, the pigtailed macaque SIV model enables the impact of antiretroviral therapy on the CNS to be studied, as neuroinflammation persists despite suppression of plasma and CSF viral load¹⁵⁵. In another model, SIV encephalitis is induced in most SIVmac251-infected rhesus macaques through depletion of CD8⁺ cells^164–166. Although this model illustrates key role of CD8⁺ cells in the neuropathogenesis and has been informative with respect to CNS macrophage biology, the elimination of both CD8⁺ T cells and natural killer cells constrains the study of cell-mediated immunity in the CNS.

The physiological relevance of primate models to mild HAND in CART-treated individuals has not yet been established. Starting CART for infected animals as early as 4 days after infection had significant benefits for acute brain disease, including suppression of SIV expression in the brain and improvements in inflammation and immunological aspects of the disease^155,167. Initiation of CART treatment after acute infection reduced SIV burden in the brain and prevented neurophysiological and locomotional alterations¹⁶⁸. Longer-term CART regimens and neurocognitive assessments of SIV-infected macaques treated with CART might, therefore, be necessary to demonstrate mild HAND in this model.

Rodent models

Rodents lack the receptors and some cellular factors to support high-level productive HIV infection¹⁶⁹. The first mouse model made to study the impact of HIV on the brain included selective expression of envelope (env)¹⁷⁰, a pathogenic HIV protein, in astrocytes under the control of the glial fibrillary acidic promoter (GFAP)¹⁷¹. These transgenic animals showed evidence of neurotoxicity and behavioural deficits^171–174, as well as defective neurogenesis^173,175. A different mouse model, based on inducible doxycycline-dependent expression of tat under the control of GFAP^{175–177,175}, resulted in synaptic pathology, learning and memory deficits, and anxiety^178–181.

Limitations of rodent models

Although a great deal about the pathophysiological effects of HIV env and tat proteins has been learned from rodent models^182,183 and these models continue to be relevant for research into some pathologies associated with HAND, they also have limitations. For example, such transgenic models cannot mimic complex aspects of HIV infection in the human host, such as invasion of the brain by activated HIV-infected and uninfected monocytes and macrophages and by free virus, which are key events in the development of HAND¹⁸⁴.

Towards better rodent models of neuro-HIV

Alternative approaches to address the limitations of rodent models include the expression of human receptors and co-receptors, tat-cofactors, and/or the viral genome in rodents^185–187. Mice and rats that carry a gag-pol deleted HIV transgene show a broad range of pathologies, including nephropathy, pulmonary disease and brain abnormalities^188–190. The transgenic HIV rat model, in particular, has been used extensively by several research groups as a model for HIV-associated brain disease; the phenotypic alterations of these rats include HAND-like gene expression profiles, changes in energy metabolism in the brain, synaptodendritic damage and behavioural deficits^191–193. These pathologies are probably caused by HIV long terminal repeat (LTR)-driven expression of viral RNA and multiple viral proteins in host cells¹⁹⁴, so might at least partially reflect the physiological complexity of HIV–host interactions in human disease. However, there are also important limitations of the model, including the presence of an incomplete HIV genome, absence of natural infectious processes, host immune tolerance to HIV transgene products, and the fact that the HIV genome is present in all cells and expressed in a variety of cell types and tissues. In a better representation of the natural HIV infection process, mice and rats carrying human CD4 and CCR5 transgenes exhibited low-level HIV infection and expression in vivo^185,195. As in humans, HIV infection in these animals can be blocked by antiretroviral drugs^185,195. There are indications, however, that HIV infection of CD4 or CCR5 transgenic rodents is limited and does not propagate¹⁹⁶, and there have been no reports of HIV brain entry and neuropathogenesis in this model to date.

In another approach, a mouse model of HIV infection was created by causing HIV tropism in mice by introducing HIV with gp120 replaced by the ecotropic murine leukaemia virus gp80 envelope gene¹⁹⁷. This chimeric virus, called EcoHIV, gains entry into murine cells through the cationic amino acid transporter-1 (mCAT)¹⁹⁸. Conventional immunocompetent mice that are exposed to EcoHIV acquire efficient HIV infection that can be prevented by treatment with antiretroviral drugs that are in clinical use¹⁹⁹. Despite widespread expression of mCAT in mouse tissues¹⁹⁸, persistent HIV infection is preferentially detected in lymphoid tissues and the brain, specifically in CD4⁺ T cells, macrophages and microglia, but not in the liver or lung^197,200,201. Infected mice seroconvert and develop CD8⁺ T cell-mediated responses, which limit systemic virus expression. Similar to processes observed in patients receiving CART, these animals do not show progression to immunodeficiency and AIDS^{197,199,201,202}. Despite immune control, HIV spreads to lymphoid tissues and the brain, and residual virus remains infective^197,199. Virus burden in the brain was low, and no gross brain pathology was observed, but gene expression tests in brain tissue revealed low-level inflammatory and type I interferon responses^199,201. Efficient HIV expression and microglia and astrocyte activation were observed after stereotactic EcoHIV inoculation into mouse basal ganglia, but these changes were limited by type I interferon responses²⁰¹.

Thus, the EcoHIV model might capture many of the features observed in individuals with HAND who are receiving CART, including maintenance of functional immunity, viral persistence at low levels in the periphery and brain, and minimal brain pathology despite the presence of molecular changes that are associated with neuroinflammation and cognitive dysfunction^199,201,202. Some distinctions from human disease are clear. Brain abnormalities in EcoHIV-infected mice occur in the presence of a functional host immune system^155,160,202, so some proposed determinants of mild HAND in humans (for example, CD4⁺ T cell nadir (REF. 24)), might not be required in this model. Absence of gp120 from EcoHIV prevents analysis of the contributions that env makes to neuropathogenesis during viral infection of mice.

Research in rodent models has been stimulated by improvements in the efficiency and stability of human haematopoietic stem cell grafts into immunodeficient mice^203,204. NOD/Scid IL-2R-gamma null (NSG) mice that are engrafted with human CD34⁺ stem cells (NSG-hCD34⁺), which differentiate into mature human T lymphocytes, monocytes and macrophages, can be efficiently infected with HIV in a sustained manner^203–206. Chronic HIV infection in this model is characterized by high HIV plasma burdens, CD4⁺ T cell depletion, and low-level HIV neuroinvasion^205,206. The latter is probably linked to transmigration of HIV-infected monocytes and macrophages into the mouse CNS. These human cells localize predominantly in the meninges and perivascular spaces and, to a lesser extent, in brain parenchyma^205,207. Despite low viral burdens, chronically HIV-infected NSG-hCD34⁺ mice show several important characteristics of HAND, including activation of resident microglia and astrocytes in some brain regions, limited brain pathology in some mice, elevated markers of neuroinflammation, and evidence of neuronal injury and neurodegeneration, assessed with brain metabolite analysis and immunofluorescence staining^205–207. Some of these changes were reversed by nanoparticle-based CART²⁰⁸. This model has also reproduced some aspects of the cognitive deficits of HAND; for example, animals exhibit increased anxiety in an open field exploratory behaviour test²⁰⁷. Physical fragility of NSG-hCD34⁺ mice precludes testing of cognitive impairment in infected mice with more conventional tests of learning and memory, such as the Morris water maze²⁰⁵. Despite the potential limitations of this model, which also include variability in the efficiency of human cell engraftment and the low rates of graft-versus-host disease, it holds promise for studies of prognostic and diagnostic translatable neuroimaging and biomarkers, and for providing a model in which to test novel therapeutic approaches to on-going cognitive impairment or its prevention^175,209.

Each of the mouse models discussed above should be aided by the recent development of sophisticated and reproducible behavioural tests for executive dysfunction and attention deficits in mice. Impaired performance in these tests can serve as markers of cognitive impairment in these models^210–212.

Therapeutic advances for HAND

The widespread implementation of CART means that it is more important than ever to consider HAND therapy in the context of ageing HIV+ patients who have received CART for years or even decades, but have persistent systemic and CNS inflammation. The development of validated biomarkers or clinical neurocognitive tests that can accurately stratify the risk of developing HAND will be important steps in improving therapy.

Eradication strategies

As discussed above, advancements in HIV-eradication strategies have drawn attention to the CNS as a potentially important reservoir for HIV. While CD4⁺ memory T cells are clearly the major viral reservoir, other sites — including gut-associated lymphoid tissue, peripheral blood, bone marrow and the brain — could also be important reservoirs. The SIV encephalitis model provides convincing evidence that viral DNA persists even after complete suppression of SIV in the blood and CSF¹¹⁸. It is critical not to overlook the CNS as a potential reservoir site when eradication strategies are deployed⁹. A central premise of HIV eradication is that latent viral reservoirs might need to be activated so as to be targeted for elimination. Novel latency-reversing agents (LRAs) might be used to activate latent viruses and purge persistent reservoirs in resting memory CD4⁺ T cells and throughout the body, but the degree of reservoir reduction that is necessary for a true ‘cure’ is unknown²¹³. In addition, LRAs might pose a challenge in the CNS, as activating latent viral reservoirs in the brain in immunocompetent patients might result in an overabundant inflammatory response that leads to brain oedema and profound neurological complications.

CNS escape and CNS CART penetration

In some people with chronic HIV infection, HIV-1 RNA can be found at higher concentrations in the CSF than in the blood, possibly as a result of poor delivery of antiretroviral drugs into the CNS. Published reports have identified that low-level HIV is present in the CSF in up to 28% of adults receiving CART^214,215. Nevertheless, the clinical relevance of this CSF viral escape is not well understood, because the phenomenon has no consistent correlation with CNS penetration of CART or with the development of HAND.

The extent to which antiretroviral drug distribution and toxicity in the CNS affect clinical outcomes is also debated. CART regimens with high CNS penetration–effectiveness (CPE) have been associated with a reduced proportion of patients with detectable CSF viral loads²¹⁶. By contrast, one large cohort of 51,938 HIV+ individuals who were CART-naive at enrolment found a 74% increased risk of HAD in those receiving CART with high CPE. A trial that focused on optimizing the CNS penetration of CART regimens failed to show an effect of this strategy on neurocognitive performance²¹⁷. Given the potent CART agents available today and this conflicting evidence, in clinical practice we use the simplest, most potent and least toxic regimens in HIV+ patients with or without HAND, and do not consider the theoretical CPE of a given CART regimen.

Neurotoxicity of CART

The suggestion that neurocognitive function is worse with high CPE CART regimens has prompted concerns that antiretrovirals themselves might be neurotoxic, thus contributing to the persistence of HAND in the CART era²¹⁸. Some in vitro investigations have supported these concerns. For example, in one study, MAP-2 staining, dendritic arborization complexity, and neural responses to exogenous calcium were used as markers for neuronal damage and revealed neuronal toxicity of 15 different antiretroviral drugs from different drug classes²¹⁹. Our research group has also shown that metabolites of efavirenz, a commonly used antiretroviral, may induce neuronal injury in vitro²²⁰, and clinical observations have suggested negative neurocognitive effects of this drug²²¹. Whether these observations are clinically relevant for other CART regimens is uncertain. The use of CART is undoubtedly life-saving, so CART should not be interrupted or deferred on the theoretical grounds of neurotoxicity.

Drug discovery in HAND

One of the challenges that the HAND research community has faced is the lack of interest from the pharmaceutical industry in the development of therapeutics for HAND, principally because the condition has not been perceived as a viable target. This reluctance of the commercial pharmaceutical sector has imposed an increasing burden on the academic sector to develop new therapeutics for HAND. One example of the efforts led by academic researchers is the development of intranasal insulin as a possible therapeutic agent for HAND, which we are studying in preclinical and human studies. A number of studies have successfully used intranasal insulin to improve cognitive function in healthy individuals, and in individuals with impaired cognitive performance as a result of ageing or Alzheimer disease²²². The mechanistic explanation for these protective effects is not well understood, but insulin has a variety of metabolic and trophic effects and might directly protect neurons and dampen inflammatory cytokine expression²²³. These multi-target effects of insulin, coupled with intranasal delivery to selectively target the CNS, make intranasal insulin an attractive candidate for a neuroprotective therapy in HAND.

Conclusions

Although significant progress has been made in understanding the clinical features, epidemiology, and neuropathogenesis of HAND, a number of critical and unanswered questions remain (BOX 3). We hope that in the next decade, substantial progress will be made in the development of validated biomarkers and an effective adjunctive therapy that can be added to CART regimens to prevent and/or ameliorate the neurocognitive deficits of HAND.

Box 3. Critical unanswered questions regarding HAND.

• In the setting of complete, durable systemic virological suppression with CART, do individuals with HIV infection continue to develop HAND?

• Despite improved understanding of the pathogenic mechanisms that underlie HAND, why are there no definitive adjunctive treatments?

• Can in vitro or in vivo models be used to more effectively develop and translate novel therapeutics agents for clinical trials?

• Can validated surrogate markers be used to improve the efficiency of clinical trials for HAND?

• Can screening tests and methods to assess for HAND be optimized to identify those at risk of developing HAND and those at risk of progression of HAND?

• How can knowledge of optimal screening tests and methods of assessment be more widely and effectively disseminated to HIV care providers around the world?

CART, combination antiretroviral therapy; HAND, HIV-associated neurocognitive syndrome.

Glossary

HIV-associated dementia (HAD)

Marked cognitive impairment involving at least two cognitive domains that substantially interferes with daily functioning

Sterilizing cure

Elimination of all HIV-infected cells from the individual

Asymptomatic neurocognitive impairment (ANI)

Cognitive impairment involving at least two cognitive domains that does not interfere with everyday functioning

Mild neurocognitive disorder (MND)

Cognitive impairment involving at least two cognitive domains that produces at least mild interference in daily functioning

CSF viral escape

Presence of detectable HIV in the cerebrospinal fluid (CSF) despite undetectable HIV RNA levels in the plasma