You are on page 1of 9

CROI 2015 Highlights: Neurologic Complications Volume 23 Issue 1 March/April 2015

Neurologic Complications of HIV Infection
Serena S. Spudich, MD; Beau M. Ances, MD, PhD

More than 30 years into the HIV epidemic, research efforts are focusing on HIV rebounded first in blood and then
better understanding how the central nervous system (CNS) is adversely in CSF. Rebounding HIV was occa-
affected by HIV and on improving the quality of life of HIV-infected sionally equilibrated, meaning that
individuals. At the 2015 Conference on Retroviruses and Opportunistic the HIV found in CSF was identical to
Infections, neurologic presentations concentrated on characterization that found in blood. However, in many
of potential CNS reservoirs of HIV, the pathogenesis of HIV-associated participants (10/14), HIV rebound-
neurocognitive disorders (HAND), diagnosis of cognitive dysfunction caused ing in CSF was distinct from that in
by HIV, neuroimaging biomarkers of HAND, and treatment of modifiable blood in at least 1 gene, most typically
risk factors of HAND. Studies presented also highlighted research on CNS env. Although this was a retrospective
disorders in international, resource-limited settings, setting the stage for study with variable timing of sampling
a growing collection of collaborative studies that will directly impact the after treatment interruption, the 4 par-
largest concentrations of people living with HIV worldwide. ticipants who had samples taken in
the first weeks after treatment inter-
Keywords: CROI 2015, HIV, central nervous system, HIV-associated neurocognitive ruption all had compartmentalization
disorder, neuroimaging, neuropathogenesis, cerebrospinal fluid of rebounding HIV. In 11 participants,
serial sampling after treatment inter-
A potentially deleterious impact of Central Nervous System HIV ruption provided a window to study
HIV on the nervous system in other- Persistence and Latency: if compartmentalization in the CSF
wise stable-appearing individuals has Evidence, Measurements, and was sustained; in a majority of these,
garnered growing concern. HIV may Mechanisms the compartmentalization persisted.
establish infection and immune acti- These studies suggest a source of HIV
vation in the nervous system early Several studies at CROI 2015 focused in CNS independent from blood after
during infection, and recognition of on characterizing and interrogating treatment interruption. Methods such
these early events has implications HIV in CNS tissues in order to shed as these, including sampling after
for potential persistent sites of HIV light on the possibility and nature of a treatment interruption and the use of
infection and for early and long-term CNS reservoir. Gianella and colleagues next-generation sequencing, are likely
neuropathogenesis.1 Although the inci- (Abstract 58) examined genetic attri- to be powerful tools for dissecting
dence of more advanced forms of HIV- butes of paired cerebrospinal fluid compartmentalized sources of HIV and
associated neurocognitive disorders (CSF) and plasma samples from 14 their characteristics in future studies.
(HAND) has declined with the use of HIV-infected men after interruption A key challenge to assessing HIV
potent antiretroviral therapy, milder of antiretroviral therapy. This study persistence in the CNS compartment is
forms of HAND remain relatively com- investigated the patterns of genetic the need to quantify HIV in the CNS
mon.2,3 Further, HIV infection and compartmentalization between CSF in HIV-infected participants, using
associated inflammation may persist and blood that might reveal the source current tissue sampling strategies.
in the central nervous system (CNS) of HIV rebound after interruption of To date, most studies have relied on
in some individuals during antiretro- therapy. measuring cell-free HIV RNA from CSF
viral treatment. Studies presented at Study participants had been on supernatant to provide an assessment
the 2015 Conference on Retroviruses antiretroviral therapy for more than of HIV burden in the CNS. Hellmuth
and Opportunistic Infections (CROI), 4 years, although whether they were and colleagues (Abstract 438) com-
held from February 22 to 26, aimed to all virally suppressed during this pared levels of HIV RNA in CSF and
unravel the biologic underpinnings of period was unknown. Next-generation blood, using standard viral load assays
HIV effects in the CNS and to develop sequencing using the 454 platform was (lower limit of detection of 50 copies/
therapeutic strategies to address ongo- employed to examine viral sequences mL in blood and 100 copies/mL in
ing abnormalities related to these in (including env, gag, and pol) in CSF and CSF) in HIV-infected individuals dur-
HIV-infected individuals. blood after treatment interruption. ing acute and chronic infection, with a

Dr Spudich is Associate Professor of Neurology at Yale University School of Medicine in New Haven, Connecticut. Dr Ances is Associate Profes-
sor of Neurology at Washington University in St Louis, Missouri. Send correspondence to Serena S. Spudich, MD, Yale University Department
of Neurology, PO Box 208018, New Haven, CT 06520. Received on March 25, 2015; accepted on March 30, 2015.

IASUSA Topics in Antiviral Medicine

higher ratio of plasma to CSF HIV RNA testing before and after antiretroviral pellet samples from the entire group,
observed during early than in chronic therapy (regimens were determined and in 50% of CSF samples from indi-
infection. Further, 10 of 42 individuals separately from the study). A lumbar viduals with undetectable HIV RNA
assessed during acute HIV infection puncture and neuropsychologic test- in both compartments. Although a
had levels of HIV RNA in CSF that were ing (11 tests, normed and averaged higher proportion of PBMC samples
below the limit of detection and were to yield a total z score) were obtained had detectable HIV DNA (100% of the
associated with lower levels of biomark- prior to treatment, with repeat test- overall group), median levels of HIV
ers of immune activation in plasma ing at intervals following antiretroviral DNA were similar or higher in CSF
and CSF. This study raises the question therapy initiation: lumbar puncture at (3.4 log10 copies/million cells) than in
of whether initiation of antiretroviral 2 weeks to 4 weeks and neuropsycho- PBMCs (2.2 log10 copies/million cells)
therapy during acute HIV infection may logic testing at 24 weeks. The median in individuals taking suppressive anti-
reduce the burden of persistent CNS CSF HIV RNA level of 3.14 log10 cop- retroviral therapy.
infection or prevent dissemination to ies/mL at baseline was reduced to In the group overall, higher levels
the CNS, given the lack of detectable 1.60 log10 copies/mL after 2 weeks of of HIV DNA in CSF correlated with
HIV RNA in the CNS in 24% of indi- antiretroviral therapy, and the base- higher levels measured in PBMCs and
viduals during this period. line summarized total z score of -0.91 with higher levels of HIV RNA in CSF.
Two studies that also focused on had improved to -0.71 at the time of More than 4.0 log10 copies/million CSF
HIV RNA recovered from the CSF sug- follow-up. cells were detected in some individu-
gested a relationship between the Analysis demonstrated relation- als with undetectable HIV RNA levels,
extent of local HIV replication within ships between the degree and rapidity indicating persistence of HIV in CNS
the CNS compartment and neuro- of HIV RNA reduction in the first 2 cells during suppressive antiretroviral
pathogenesis. Joseph and colleagues weeks of antiretroviral treatment and therapy. These data provide rationale
(Abstract 440) examined the genetic the extent of improvement in neuro- for future studies focused on charac-
attributes and phylogenetic relation- cognitive functioning during the first terization of HIV detected in CNS cells
ships of HIV in paired CSF and blood 6 months after initiating therapy, sug- during suppressive antiretroviral ther-
samples from 40 individuals with neu- gesting that early attenuation of HIV apy, to potentially yield further insight
ropsychologic testingbased diagnoses RNA replication in the CNS may lead into the sites and mechanisms of HIV
of distinct categories of HAND. In to enhanced neurocognitive responses persistence.
some individuals, partial env genes to treatment. Alternatively, these data In further studies examining HIV
were amplified using deep sequenc- could suggest that individuals with a persistence and latency in the CNS,
ing with a primer ID, and in others, slower or less pronounced reduction in Gelman and colleagues (Abstract 61)
full-length env was amplified by single HIV RNA initially have a more severe studied brain specimens from 40 indi-
genome amplification. The investi- local CNS HIV infection, which is asso- viduals who died with HIV infection
gators found increasing proportions ciated with poorer long-term reversal and from 20 HIV-uninfected individu-
of CNS compartmentalization with in response to standard antiretroviral als, using autopsy materials obtained
increasing severity of HAND: 29% in treatment. through the National NeuroAIDS Tis-
neurologically normal participants, CSF HIV RNA is typically reduced to sue Consortium. In order to determine
40% in participants with asympto- below levels of standard detection in tissue biomarkers and mechanisms
matic neurocognitive impairment individuals taking antiretroviral treat- associated with HIV DNA in the brain,
(ANI) or mild neurocognitive disorder ment. Attention has recently turned to the investigators examined the rela-
(MND), and 70% in participants with whether HIV nucleic acid detected in tionship between macrophage markers
HIV-associated dementia (HAD). Using cells recovered from the CSF may be associated with and levels of HIV DNA
an Affinofile assay to determine puta- a useful measure of HIV persistence relative to HIV RNA measured in the
tive macrophage infectivity of HIV in the CNS. de Oliviera and colleagues dorsal prefrontal cortex, an area often
based on entry into cells with vary- (Abstract 435) measured pol using a affected by HIV. Although a number of
ing surface levels of CD4 receptors, the droplet digital polymerase chain reac- the markers considered standard indi-
investigators determined that 71% of tion (ddPCR) assay, to quantitate HIV cators of macrophages and microglia
compartmentalized viruses derived DNA in peripheral blood mononuclear (including CD16, CD14, and CD163)
from CSF appeared to be adapted to cells (PBMCs) and CSF cell pellets did not differentiate between those
replication within macrophages. obtained from 29 HIV-infected indi- with proportionally higher HIV DNA,
Robertson and colleagues (Abstract viduals. Twenty of these individuals other markers (including interferon
439) also used CSF HIV RNA as a mea- were taking antiretroviral therapy for regulatory factor 4 [IRF-4]; C-type lectin
sure of CNS HIV infection, using the a median of 2.3 years and were viro- domain family 4, member A [CLEC4A,
HIV RNA level as a proxy for extent logically suppressed, with HIV RNA also termed DCIR]; and interleukin 10
of CNS infection in a group of approxi- levels in blood and CSF of less than [IL-10]) were higher in this group.
mately 30 individuals evaluated with 50 copies/mL. ddPCR yielded detect- These results are of interest because
CSF collection and neuropsychologic able HIV DNA in 66% of CSF cell IRF-4 is a transcription factor that could

CROI 2015 Highlights: Neurologic Complications Volume 23 Issue 1 March/April 2015

potentially mediate the relationship the brain tissue compartment. Further combination of agents, with a 10-fold
between integrated HIV DNA and studies focused on measuring not higher SIV RNA level in CSF than
expressed HIV RNA. IRF-4 is also only whole drug concentrations but in plasma and development of SIV
regulated by a polycomb repressive also concentrations of phosphorylated encephalitis despite continued treat-
complex such that the methyltrans- nucleotide analogue reverse transcrip- ment with antiretroviral therapy. This
ferase enhancer of zeste homolog 2 tase inhibitor anabolites may further animal also had increases in levels
(EZH2) suppresses the expression of enhance precision of antiviral activity of CSF neopterin (a marker of mac-
IRF-4; thus, if important as a regulator, estimates in the CNS. rophage activation), chemokine (CC
EZH2 might lower HIV DNA relative motif) ligand 2 (CCL2; a marker of
to HIV RNA. Dual staining of macro- New approaches for assessing HIV per- monocyte chemoattraction), and neu-
phages in the leptomeninges of the sistence or replication in the CNS include rofilament light chain (NFL; a marker
brains of these individuals showed that of neuronal injury) after intervention.
comparison of HIV variants in compart-
the presence of EZH2 was associated Moreover, SIV RNA was elevated in
ments after treatment interruption,
with increased expression of HIV RNA, the occipital cortex in this animal at
ddPCR measurement of CSF fluid cell-
suggesting that this pathway might be sacrifice, and a quantitative outgrowth
associated DNA, and next-generation
related to control of viral expression assay to determine absolute number
versus maintenance of HIV DNA in a sequencing to genetically characterize of infected resting CD4+ T cells har-
latent state. Although these analyses HIV in the CNS compartment. vested from PBMCs revealed that in
were performed in autopsy studies both treated animals, there was a
of HIV-infected individuals who were Gama and colleagues (Abstract decline in infected resting CD4+ T
not necessarily taking suppressive 416) examined issues of latency and cells after use of latency-reactivating
antiretroviral therapy at the time of explored strategies for reducing per- agents. Although a small study in a
death, these results reveal potentially sistent lentiviral integration in the simian model developed for study of
important mechanisms of endogenous CNS by employing latency-reactivating encephalitis and thus of accelerated
regulation of HIV DNA transcription in agents in neurologic studies of rhe- disease with uncertain generalizabil-
the human brain. sus macaques. Using an accelerated ity to human CNS HIV infection, this
Another study presented at CROI macaque model of neuroAIDS created is the first study to demonstrate the
2015 that was focused on brain by dual infection with 2 neurovirulent activity of latency-reactivating agents
autopsy tissue explored an issue of rel- strains of simian immunodeficiency in the CNS and the potential deleteri-
evance to control of HIV replication in virus (SIV; SIVDeltaB670 and SIV/17- ous effects of this strategy in the CNS
the setting of antiretroviral treatment. Fr), these investigators treated 3 ani- compartment.
In order to assess how antiretroviral mals with antiretroviral therapy such
drugs may access brain tissue and that they were virally suppressed to Mechanisms of Neuropatho-
potentially impact levels of HIV infec- less than 100 copies/mL of SIV RNA
genesis in HIV: Immune Activa-
tion in this compartment, Bumpus and in plasma for 500 days. They then
tion, Mitochondrial Dysfunction,
colleagues (Abstract 436) examined exposed 2 of the animals to the protein
and Toxicity
samples from 3 brain regions (glo- kinase C activator ingenol-3-hexanoate
bus pallidus, cortical grey matter, and (Ing-B, a putative latency-reactivating That activation of the host immune
white matter) obtained from autopsy agent) for 40 days, allowed a 2-week system resulting in inflammatory-
studies from the California NeuroAIDS washout period, and then exposed mediated damage in the CNS is a key
Tissue Consortium. Concentrations of these animals to Ing-B plus vorinostat substrate of HIV-related neurologic
atazanavir, efavirenz, emtricitabine, (total 6 mg/kg) for 15 days. One control injuries is well established. However,
and lamivudine in the brain were simi- animal received antiretroviral therapy the details of the processes and
lar to those reported in CSF. However, alone without Ing-B or vorinostat. pathways involved in this immune
tenofovir, which is considered to have CSF assessments were obtained response are still not well understood,
potentially poor CNS efficacy owing throughout the study, and the animals and must be determined in order to
to low measured concentrations in were euthanized for examination of develop effective therapies for injuries
the CSF, had concentrations that were brain tissue after the interventions, to the nervous system in individuals
notably higher in all brain regions than with measurement of CSF and plasma infected with HIV. As a parallel pro-
in the CSF. Lopinavir concentrations SIV RNA by quantitative PCR (qPCR) cess to cellular infiltration and soluble
were also higher in the brain in the and ddPCR, measurement of SIV inflammatory mediator activation in
frontal white matter than in the CSF. DNA in brain by qPCR, and measure- the CNS, perturbation of blood-CNS
This study reveals that although con- ments of SIV RNA in brain by in situ barriers may serve an important role
centrations of certain drugs reaching hybridization. One animal treated in neuropathogenesis of HIV, by allow-
the brain can be extrapolated from lev- with latency-reactivating agents had ing increased influx of infected and
els found in the CSF, in some cases CSF a marked rise in plasma and CSF activated cells and of toxic soluble
may underestimate concentrations in SIV RNA after treatment with the products into the CNS.

IASUSA Topics in Antiviral Medicine

Two oral presentations at CROI Albumin ratio measured at baseline baseline and follow-up CSF and neu-
2015 examined the status of the (a median 3 months postinfection) ropsychologic testing through the
blood-CNS barrier in HIV-infected indi- in 108 individuals with primary HIV CHARTER (CNS HIV Anti-Retroviral
viduals, using a measure of albumin infection was elevated compared Therapy Effects Research) study or the
concentration in CSF compared with with an age-matched group of HIV- HIV Neurobehavioral Research Center
blood (CSF-toplasma albumin ratio). uninfected individuals. Over a median (HNRC). Participants were classified
Anesten and colleagues (Abstract 59) 1 and one-half years of longitudinal as neurologically normal or as having
measured CSF-toplasma albumin ratio follow-up prior to initiation of anti- neurocognitive impairment (ANI or
in 657 HIV-infected individuals cat- retroviral therapy, CSF-toplasma albu- MND) and were assessed for changes
egorized by HIV treatment status, and min ratios did not change, suggesting in neuropsychologic testing perfor-
by clinically based diagnosis of HAD a lack of resolution of blood-CNS bar- mance between 2 time points. CSF
versus lack of neurologic symptoms rier disruption during this early pe- NFL levels were not elevated in the
(neuroasymptomatic) in individuals riod. Moreover, in a smaller cohort group of individuals with neurocogni-
not taking antiretroviral therapy. (n=57), the CSF-toplasma albumin tive impairment at baseline (n=70)
Neuroasymptomatic individuals were ratio did not statistically change 1 year or in those who had neurocognitive
further divided based on CD4+ cell after initiating antiretroviral therapy. decline over the course of follow-up
count range. When albumin ratio in The CSF-toplasma albumin ratio (n=32). However, CSF neopterin was
each HIV-infected group was com- correlated with CSF NFL level in indi- elevated in those with impairment at
pared with that in a group of 53 HIV- viduals with primary HIV infection at baseline compared with neurocogni-
uninfected controls who had CSF sam- baseline and in longitudinal follow- tively normal individuals, and was also
ples collected for research purposes, up, and inversely correlated with elevated in the group that experienced
statistically significant elevations were N-acetylaspartateto-creatine ratio, neurocognitive decline. These data
only found in the group with HAD. a neuroimaging measure of neuronal indicate that heightened intrathecal
No differences were noted across the integrity in the parietal grey matter. macrophage activation is associated
CD4+ cell count spectrum between Differing results between this study with the presence and progression
HIV-uninfected controls and HIV- and those noted in Abstract 59 may of impairment in participants on sys-
infected individuals whether they were relate to a lack of age matching in HIV- temically successful treatment. These
or were not taking suppressive anti- uninfected comparison participants are the first data to tie a mechanism of
retroviral therapy. in the larger study, or to a distinct neurologic injury to progressive clini-
When compared with age- and dynamic pattern of blood-CNS cal signs in well-treated individuals
determined published cutoffs of upper perturbation that is specific to early with HIV infection, and have important
limit of normal albumin ratio, abnor- HIV infection. In this same cohort of implications for strategies to reverse or
mally elevated levels were detected in primary HIV infection, Wright and col- ameliorate HAND.
16% of participants in the neuroas- leagues (Abstract 60) demonstrated Peluso and colleagues (Abstract 473)
ymptomatic group that was not taking that a higher CSF-toplasma albumin examined a novel immune activation
antiretroviral therapy and in 68% of ratio correlated with reduced putam- marker assessed longitudinally in CSF
individuals with HAD. Despite albumin inal volume. These data underscore before and after initiation of antiret-
ratios within the normal range in most the potential pathogenic significance roviral therapy in study participants
individuals in the study, NFL (as noted of blood-CNS barrier perturbation in with acute or chronic HIV infection in
above, a marker of active neurologic CNS and emphasize that processes Thailand. CSF YKL-40, a systemic bio-
injury) level correlated with albumin in HIV infection associated with neu- marker of inflammation and cancer
ratio in the HIV-infected groups, and rologic damage and neurocognitive that localizes to activated microglial
in the antiretroviral treated, virally sup- impairment are initiated during the cells and reactive astrocytes in the
pressed group. In a multivariate model, early stages of infection. CNS, was measured in 33 individuals
albumin ratio was a predictor of NFL Edn and colleagues (Abstract 474) with acute HIV infection (median 18
level independent of age. These results focused on the concept that progres- estimated days of infection) compared
indicate that blood-CNS barrier disrup- sive neurologic injury in the CNS with chronic infection (n=34), owing
tion is associated with neurologic injury in individuals taking suppressive to its predictive value in development
in HIV infection, and suggest that this antiretroviral therapy may relate to of neurodegeneration (including SIV
process is a late-stage complication of underlying mechanisms of persistent encephalitis). In individuals with acute
HIV that is specifically related to severe immune activation. They explored HIV infection prior to intitiation of
encephalitis and dementia. measures of intrathecal macrophage antiretroviral therapy, CSF YKL-40
In a longitudinal study of primary activation (CSF neopterin) and neuro- was lower than in participants with
HIV infectiondefined as within the nal injury (CSF NFL) in 100 individuals chronic HIV infection and was not
first year of HIV acquisitionRahimy taking antiretroviral therapy who had different than in HIV-uninfected Thai
and colleagues (Abstract 62) exam- successful plasma viral suppression volunteers (n=18). After initation of
ined the CSF-toplasma albumin ratio. (HIV RNA level < 50 copies/mL) and antiretroviral therapy during acute HIV

CROI 2015 Highlights: Neurologic Complications Volume 23 Issue 1 March/April 2015

infection (6-month treatment dura- target for the treatment of depression Several studies focused on heme
tion), the median level of CSF YKL-40 regardless of HIV infection. oxygenase-1 (HO-1) as a potentially
remained stable and was statistically Alteration in mitochondrial func- neuroprotective factor in HIV-related
significantly lower than in those start- tion is postulated to potentially result brain injury. HO-1 has been previously
ing treatment during chronic HIV from HIV infection itself and the toxic shown to be reduced in the frontal
infection (12-month treatment dura- effects of antiretroviral medications. cortex and striatum of HIV-infected
tion). These findings suggest that Funes and colleagues presented data individuals with HAD, and inversely
microglial and perhaps astrocyte from a study that investigated nitric related to HIV RNA level and neuro-
activation may be prevented or ame- oxide as a potentially crucial molecule inflammation.4 Gill and colleagues
liorated by very early antiretroviral in the mechanism of efavirenz-induced (Abstract 501) focused on the relation-
therapy. neuronal toxicity (Abstract 500). Lev- ship between HIV strain and degree
Numerous studies presented at els of nitric oxide synthase and nitric of macrophage HO-1 down regula-
CROI 2015 examined how altered oxide (a free radical associated with tion by studying an in vitro model
cellular bioenergetics or metabolism mitochondrial dysfunction and inflam- of HIV-infected monocyte-derived
may contribute to neuropathogen- mation) were measured in human macrophages (MDMs) isolated from
esis of HAND. Two studies employed brain tumor cell lines and cultured rat noninfected donors and infected the
metabolomic profiling to assess pro- cortical neurons after brief exposure cells with 15 HIV strains. In this study,
cesses associated with neurocognitive to efavirenz. In this system, efavirenz replication of HIV strains consistently
dysfunction. Haughey and colleagues provoked inducible nitric oxide syn- reduced HO-1 in MDMs. Levels of HO-1
(Abstract 497) explored how the thase in glial cells, and this increase were inversely associated with levels of
composition of energy metabolites in nitric oxide impaired mitochondrial viral replication and extracellular gluta-
in the CSF, as measured by hydrogen function. These effects were not seen mate measured in supernatant in this
nuclear magnetic resonance (1H-NMR) in neurons, suggesting that bioener- MDM model, suggesting that enhanc-
spectroscopy, may relate to neurocog- getic toxicity of efavirenz may occur ing production of HO-1 in MDMs may
nitive status over time in HIV-infected through glial dysfunction rather than benefit the CNS in HIV.
individuals. Using partial least squares neuronal injury. Therapies aimed at
regression, which allowed them to reducing nitric oxide or its precur- HIV neuropathogenesis appears to re-
identify predictors associated with sors may reduce the potential adverse late to processes such as blood-CNS
outcomes from a very large number effects associated with efavirenz.
barrier disruption and intrathecal im-
of potentially contributing variables, Another molecule that has been
mune activation, which may persist de-
these investigators found that the implicated in the etiology of HAND
spite initiation of antiretroviral therapy.
metabolites in CSF involved with aero- through induction of cell death and
Alteration of intracellular energy me-
bic glycolysis along with other clinical CNS inflammation is HIV transacti-
factors were predictive of neurocogni- vator of transcription (Tat). Brew and tabolism may be associated with neuro-
tive decline, whereas those involved colleagues (Abstract 505) explored logic or psychiatric morbidity and may
with anaerobic glycolysis were predic- whether HIV Tat levels in CSF might be worsened by neurotoxic medications.
tive of neurocognitive improvement. remain persistently abnormal in
This pattern suggests potential novel individuals taking suppressive antiret- Duncan and colleagues (Abstract
approaches for predicting response to roviral therapy, as Tat can be secreted 502) presented a study that identified
therapy in HIV-infected individuals. In by infected cells even during therapy. atorvastatin as a medication with poten-
a related study, Cassol and colleagues HIV Tat remained detectable in the CSF tial benefit for the HO-1 deficiency
(Abstract 498) analyzed cellular metab- in 5 (13.5%) HIV-infected individuals noted in HIV infection. Atorvastatin
olites in blood plasma using liquid or taking antiretroviral therapy who had belongs to a class of 3-hydroxy-3-
gas chromatography followed by mass HIV RNA levels below 50 copies/mL methylglutaryl-coenzyme A (HMG-CoA)
spectrometry in 68 HIV-infected and in blood or CSF (n=37). Detectable reductase inhibitors that has numerous
36 HIV-uninfected study participants. Tat was not associated with current putative immunomodulatory effects,
Participants were further classified HAND status or B-cell CLL/lymphoma including reduction of monocyte or
by depression status based on a self- 11B (zinc finger protein) (BCL11B), macrophage activation. Thus, the
administered depression inventory. a putative measure of viral latency. investigators used an in vitro model
Depressed individuals had lower However, further studies are needed to of primary MDMs infected with a
levels of metabolites of phenylalanine- investigate whether levels of CSF Tat macrophage-tropic HIV strain to assess
tyrosine catabolism and acylcarnitine detected in individuals on successful the impact of atorvastatin treatment on
than individuals without depression. systemic treatment may be associated HO-1. They found that although HIV
Interestingly, these results were found with other highly sensitive markers of infection of MDMs reduced HO-1,
both in the HIV-infected and HIV- neuronal injury, such as CSF NFL or atorvastatin added to MDMs at 8 days
uninfected groups, suggesting that imaging markers of inflammation or postinfection increased HO-1 pro-
this pathway might be an important neuronal injury. tein expression. This study not only

IASUSA Topics in Antiviral Medicine

provides rationale for a targeted treat- associated with greater cognitive HIV-infected individuals with ANI or
ment study based on these preliminary decline (as assessed by neuropsycho- normal cognition. Observed areas of
findings, it also suggests a potential logic performance testing) at sub- increased amyloid deposition were dif-
beneficial mechanism of broad statin sequent follow-up. Collectively, these ferent than those typically seen with
use in ameliorating the neurologic pro- results suggest that screening of higher- Alzheimers disease. However, amyloid
cesses associated with HAND. risk HIV-infected participants based on depositions were not different for HIV-
cardio- and cerebrovascular abnormali- infected individuals compared with
ties (including smoking) may assist in HIV-uninfected individuals.
Diagnosis of HAND
the early diagnosis of HAND. Treating Using another PET ligand, Vera and
Key questions remain concerning how these modifiable risk factors could lead colleagues (Abstract 477) examined
to identify HIV-infected individuals at to a reduction in HAND. the relationship between microbial
increased risk for HAND. Rourke and translocation (measured by 16S ribo-
colleagues (Abstract 465) assessed 575 Identification and treatment of modi- somal [r] DNA) and brain inflamma-
HIV-infected adults taking antiretrovi- fiable risk factors (eg, smoking, obe- tion ([11C]peripheral benzodiazepine
ral therapy in the OCS (Ontario HIV sity, or depression) could reduce HIV- receptor [PBR] 28) and structure (by
Treatment Network Cohort Study). At diffusion tensor imaging [DTI]) in
associated neurocognitive disorders.
baseline, participants had neuropsy- HIV-infected individuals (n=12). An
chologic performance testing and were association existed between plasma
categorized as normal (n=299) or as 16S rDNA and brain biomarkers (11C-
having ANI (n=276). At 52 weeks of PBR28 and DTI metrics). In addition,
follow-up a greater percentage of indi- A variety of neuroimaging techniques Smith and colleagues (Abstract 485)
viduals with ANI progressed to more were used to assess the effects of HIV observed increased inflammation
severe forms of HAND (MND or HAD) in the CNS. Becker and colleagues using a 3D postcontrast T2-weighted
compared with the cognitively normal (Abstract 494) used a trajectory model fluid-attenuated inversion recovery
HIV-infected individuals. Factors asso- based on volumetric data from 3892 (FLAIR) imaging technique. A greater
ciated with faster progression included HIV-infected individuals followed by proportion of HIV-infected participants
depression and history of smoking. the MACS (Multicenter AIDS Cohort had focal leptomeningeal contrast
Similar results were observed by Study) to identify 3 possible trajecto- enhancement than did HIV-uninfected
Brouillette and colleagues (Abstract ries of disease progression: 1) normal participants. Observed enhancement
469) in a subset of the CHARTER aging, a profile with relatively low was not correlated with CD4+ cell
cohort (n=191). Factors associated probability of even mild impairment count, CD4+ cell nadir, duration of
with cognitive decline included ath- until middle age; 2) premature aging, HIV infection, or history of neurologic
erosclerotic vascular disease, duration a profile with the probability of mild disorders. Finally, Cohen and col-
of HIV infection, and education level. impairment occurring at age 45 years leagues (Abstract 935) demonstrated
The investigators noted that a large to 50 years; and 3) unhealthy, a profile that perinatally HIV-infected children
proportion of the CHARTER cohort with a high probability of impairment (n=35) had statistically significant
(approximately 80%) had modifiable at a young age. Changes in the poste- structural imaging (DTI and volumetric)
risk factors, including smoking and rior cingulateprecuneus cortex, the and neuropsychologic performance
a body mass index of 25 or higher. hippocampus, and the inferior fron- differences compared with healthy,
These risk factors could be targeted by tal cortex were associated with the HIV-uninfected children (n=37). These
primary care physicians. unhealthy profile, and changes in the neuroimaging studies hint at the possi-
The effects of potentially modifiable cingulate gyrus, the insula, and the bility of including imaging biomarkers
risk factors were further confirmed by basal ganglia were associated with the in diagnostic criteria for HAND.
several groups, using the VACS (Veter- premature aging profile.
ans Aging Cohort Study) Index score Observed neuroimaging differences Novel neuroimaging markers may
(risk points assigned for age, CD4+ may assist in differentiating between noninvasively assess structural and
cell count, plasma HIV RNA level, the effects of HIV and those with aging. inflammatory changes seen in asymp-
hemoglobin value, fibrosis stage, renal Sacktor and colleagues (Abstract 482) tomatic HIV-infected individuals, and
glomerular filtration rate, and pres- obtained 18F-AV-45 positron emission thus may provide preclinical markers of
ence of hepatitis C virus infection). tomography (PET) scanning to assess
and insight into pathogenesis of HIV-
Calcagno and colleagues (Abstract amyloid deposition in HIV-infected par-
associated neurocognitive disorders.
487) demonstrated that HIV-infected ticipants (n=25) and HIV-uninfected
participants with HAND (n=441) were controls (n=6). HIV-infected individ-
at higher cardiovascular risk based on uals with symptomatic HAND (MND Treatment of HAND
the VACS Index. Rourke and colleagues and HAD) had mildly increased amy-
(Abstract 467) showed that a higher loid depositions in the hippocampus A number of different treatment
VACS Index score at baseline was and basal ganglia compared with options are now available and should

CROI 2015 Highlights: Neurologic Complications Volume 23 Issue 1 March/April 2015

be considered for HIV-infected patients, regimen. However, those receiving the maraviroc-containing arm. Evering and
especially those with HAND. Bow- nevirapine-containing regimen had colleagues (Abstract 446) studied indi-
man and colleagues (Abstract 445) more adverse events than those receiv- viduals taking prolonged (median, 5.7
assessed HIV-infected, antiretroviral ing treatment containing efavirenz. For years) antiretroviral therapy that had
treatmentnaive participants (n=40) each of these studies, the investigators been initiated at a median 1.6 months
before and 2 weeks after initiating hypothesized that observed differences after infection, who did not have
therapy. Viral load in blood decayed may reflect differences in drug distri- comorbidities such as depression or
faster than in CSF after initiation of bution into the CNS or neurotoxicity. substance abuse. Cognitive impair-
antiretroviral therapy. In general, HIV With regard to other therapies cur- ment was observed in only 4% (1/26)
protease inhibitors were associated rently being used to treat HIV-infected of this group, suggesting the potential
with faster CSF viral suppression, and individuals with HAND, Caramatti and benefit of early treatment. Although
integrase strand transfer inhibitors were colleagues (Abstract 442) observed no many of these preliminary studies are
associated with slower suppression. differences in neuropsychologic per- promising, larger studies with longer
There is increasing concern regard- formance in participants (n=37) who follow-up are needed for HIV-infected
ing the potential adverse effects of received monotherapy with ritonavir- participants with varying degrees of
efavirenz on cognition. In a large Can- boosted (/r) atazanavir compared with cognitive impairment.
adian cohort (n=831), Rourke and those who received triple therapy that
colleagues (Abstract 448) observed no included atazanvir/r. For both groups, a Antiretroviral therapy enhancement
differences among HIV-infected par- substantial decrease in HAND was with maraviroc may lead to reduced in-
ticipants who were currently taking seen at 96 weeks. These results com- flammation and better neurocognition.
efavirenz, those who had previously plement work by Ferretti and col-
However, larger longitudinal studies
received efavirenz, or those who had leagues (Abstract 443) in a smaller
are needed.
never taken efavirenz. Although these nested cohort (n=23), which showed
results complement previous pub- that CSF viral escape was similar
lished studies nicely,5,6 they conflict between HIV-infected participants Adjunctive measures (eg, exercise
with other reports, including studies receiving long-term, successful mono- and engagement in mental exercises)
presented at CROI 2015 that suggest therapy with atazanvir/r and those were also considered for HIV-infected
the potential neurotoxicity of efavirenz. receiving triple therapy that included participants at risk for cognitive disor-
Ma and colleagues (Abstract 444) atazanvir/r. Baker and colleagues ders. Basco and colleagues (Abstract
assessed the incidence of neurocog- (Abstract 447) demonstrated that the 488) studied neuropsychologic perfor-
nitive impairment in HIV-infected CNS penetration effectiveness (CPE) of mance, neuroimaging, and self-reported
participants (n=23) in China assigned antiretroviral therapy did not affect aerobic exercise in HIV-infected indi-
an antiretroviral regimen of tenofovir, neuropsychologic performance and viduals categorized as physically active
lamivudine, and efavirenz compared brain volumetrics in a cohort of HIV- (n=22) or sedentary (n=48). Physi-
with a regimen of zidovudine, lami- infected participants (n=64). No dif- cally active participants performed
vudine, and nevirapine. A higher ferences in neuropsychologic perform- statistically significantly better than
incidence of cognitive impairment ance testing results or brain volumet- sedentary HIV-infected participants on
was seen in HIV-infected participants rics existed between groups with low neuropsychologic performance tests of
receiving the regimen that included versus high CPE scores. executive function but not of motor
tenofovir than in those receiving the Gates and colleagues (Abstract 441) function. Monroe and colleagues (Ab-
efavirenz-containing regimen. As conducted a small randomized control- stract 489) found similar results in the
part of the same randomized study, led trial among virologically sup- large MACS cohort of HIV-infected men
Letendre and colleagues (Abstract 56) pressed HIV-infected participants with (n=622). High physical activity was
presented results of these regimens HAND (n=19) who were assigned to associated with better neuropsycho-
used in antiretroviral treatmentnaive, receive antiretroviral therapy or anti- logic performance on executive and
HIV-infected adults in China. Par- retroviral therapy enhanced with psychomotor tests than was low phys-
ticipants were randomly assigned to maraviroc. Maraviroc was chosen ical activity. Whether exercise leads to
receive an open-label antiretroviral reg- because of its high level of CNS pene- cognitive improvement or whether
imen of zidovudine, lamivudine, and tration and dual antiretroviral and cognitive improvement leads to an
nevirapine or tenofovir, lamivudine, and anti-inflammatory activity. At 52 increased ability to participate in exer-
efavirenz, and were assessed longitudi- weeks, neuropsychologic performance cise remains in question.
nally by neuropsychologic performance had improved more in the maraviroc- Pinnetti and colleagues (Abstract
testing at 48 weeks and 96 weeks. The containing arm than in the control 63) demonstrated that in a single-site
group taking the efavirenz-containing arm, and neuroimaging measures of cohort (n=569), better virologic con-
regimen had a greater risk of incident glutamate concentrations (a possible trol (higher current CD4+ cell count
neurocognitive impairment than the measure of excitotoxicity) were high- and lower viral load) and higher educa-
group taking the nevirapine-containing er in the control arm than in the tion level were associated with reduced

IASUSA Topics in Antiviral Medicine

risk of developing HAND in the current examined by neurologic testing at moderate impairment, 3% with severe
antiretroviral therapy era. Results from baseline, week 48, and week 96: the impairment, and 54% falling in the
Milanini and colleagues (Abstract 495) Color Trails Test 1 (measuring attention normal range. Effective antiretroviral
showed that in a cohort of 50 HIV- and concentration), the Color Trails therapy reduced neurocognitive impair-
infected participants, higher cognitive Test 2 (measuring cognitive flexibility), ment substantially over time, from
reserve (assessed by IQ) was associ- and the Grooved Pegboard test (mea- 46% at baseline to 28% at week 168
ated with lower risk of ANI. Overall, suring psychomotor speed/fine motor of treatment.
these studies suggest that engagement skills). Test scores were standardized Sacktor and colleagues (Abstract
in physical, intellectual, and social to a z score based on demographi- 452) conducted a large study of HIV-
activities may independently protect cally adjusted, US-derived norms and infected (n=299) and HIV-uninfected
against cognitive impairment in HIV- were then averaged into an overall indiviuals (n=210) living in rural Rakai,
infected individuals, but larger studies score (NPZ-3 score). Mean composite Uganda, and found a very high rate of
with more detailed measurements of z score at baseline was -2.96, suggest- HAD in antiretroviral treatmentnaive
physical function and cognitive reserve ing that individuals in this study had individuals compared with normative
before and after an intervention are performance levels approximately 3 data from HIV-uninfected individuals
needed. standard deviations below the norm at in Kampala, Uganda. Twenty-seven
baseline. In multivariable analyses, z percent of individuals with HIV infec-
scores were independently lower with a tion met criteria for HAD compared
Adjunctive measures (including engage-
number of factors, including older age, with 7% of HIV-uninfected individuals.
ment in physical and mental exercises) lower body weight, higher viral load, Most HIV-infected individuals in this
may help reduce HIV-associated neu- lower hemoglobin value, and fewer region harbor HIV subtype (or clade)
rocognitive disorders. However, larger years of education. Scores improved A or D, allowing for important future
longitudinal studies are needed. substantially after starting antiretrovi- analyses regarding possible associa-
ral treatment, with equal improvement tions between HIV subtype and risk for
between regimen arms. The dramati- HAND. Finally, Valcour and colleagues
cally low z scores of individuals in the (Abstract 459) studied more than 900
New Frontiers for Under-
EARNEST study at baseline may reflect participants from East Africa in order to
standing HAND and HIV
true substantial cognitive impairment investigate determinants of neuropsy-
Neuropathogenesis: Resource-
or may indicate that norms derived chologic performance in the AFRICOS
Limited Settings
from the United States may not yield (The African Cohort Study) study. Cog-
At CROI 2015, neurologic studies con- accurate results when assessing indi- nitive impairment, as measured by all
ducted in resource-limited settings, viduals in resource-limited settings. testing methods, was associated with
led by or in collaboration with local A number of other studies detailed HIV infection status, age, and level
investigators were emphasized. Kam- the prevalence or incidence of HAND, of education. Nadir CD4+ cell count
bugu and colleagues (Abstract 57) or response to treatment in resource- was also associated with performance
presented data from the EARNEST limited settings. Robertson and col- on 2 specific tests. However, cognitive
(Europe-Africa Research Network for leagues (Abstract 451) presented the performance was not associated with
Evaluation of Second-Line Therapy) results of the AIDS Clinical Trials Group number of infectious and noninfectious
study examining the magnitude of and (ACTG) 5199 (International Neurological comorbidities, a finding distinct from
factors associated with neurocognitive Study) and the ACTG 5271 (Interna- studies in resource-endowed settings.
function at the time of failure of initial tional Neurocognitive Normative Study) All of these studies highlight the need
antiretroviral therapy, and assessing studies. These large studies accrued for appropriate norms to study corre-
any changes in neurocognitive func- more than 3200 participants over 12 lates to neurocognitive disorder that are
tion that may occur after switching years in various countries (Malawi, relevant in resource-limited settings,
antiretroviral treatment. The EARNEST Thailand, Brazil, India, Peru, Zimbabwe, and the need to collaborate with and
trial enrolled 1277 participants whose and South Africa), specifically included foster investigations by local experts in
initial antiretroviral therapy had failed site-specific HIV-uninfected individuals order to implement highly rigorous and
according to World Health Organiza- appropriately matched to HIV-infected relevant research in these settings.
tion clinical and immunologic criteria articipants, and performed extensive
at the time of the study design. Indi- coordinated oversight of the sites in- Financial affiliations in the past 12 months:
Drs Spudich and Ances have no relevant
viduals were randomly assigned to volved in the study for quality control.
financial affiliations to disclose.
receive lopinavir/r plus 2 or 3 nucleo- These studies demonstrated that the
tide analogue reverse transcriptase prevalence of neurocognitive impair-
inhibitors; a protease inhibitor plus ment compared with well-matched All cited abstracts appear in the
raltegravir; or protease inhibitor mono HIV-uninfected controls is similar CROI 2015 Program and Abstracts
therapy (with a 12-week raltegravir to what is seen in the United States: eBook, available online at www.
induction period). Three domains were 25% with mild impairment, 17% with
CROI 2015 Highlights: Neurologic Complications Volume 23 Issue 1 March/April 2015

3. Cysique LA, Brew BJ. Prevalence of non- 6. Leutscher PD, Stecher C, Storgaard M,
Additional References confounded HIV-associated neurocogni- Larsen CS. Discontinuation of efavirenz
tive impairment in the context of plasma therapy in HIV patients due to neuropsy-
1. Spudich S, Gisslen M, Hagberg L, et al. HIV RNA suppression. J Neurovirol. 2011; chiatric adverse effects. Scand J Infect Dis.
Central nervous system immune acti- 17(2):176-183. 2013;45(8):645-651.
vation characterizes primary human 4. Gill AJ, Kovacsics CE, Cross SA, et al.
immunodeficiency virus 1 infection even Heme oxygenase-1 deficiency accompa-
in participants with minimal cerebro- nies neuropathogenesis of HIV-associated
spinal fluid viral burden. J Infect Dis. neurocognitive disorders. J Clin Invest.
2011;204(5):753-760. 2014;124(10):4459-4472.
2. Heaton RK, Clifford DB, Franklin DRJ, et 5. Ciccarelli N, Fabbiani M, Di Giambene-
al. HIV-associated neurocognitive disor- detto S, et al. Efavirenz associated with
ders persist in the era of potent antiretro- cognitive disorders in otherwise asymp-
viral therapy: CHARTER Study. Neurology. tomatic HIV-infected patients. Neurology. Top Antivir Med. 2015;23(1):47-55.
2010;75(23):2087-2096. 2011;76(16):1403-1409. 2015, IASUSA. All rights reserved