Journal of Neuroimmunology 157 (2004) 111 – 119

www.elsevier.com/locate/jneuroim

Acquired neuronal channelopathies in HIV-associated dementia

Benjamin B. Gelman a,b,*, Vicki M. Soukup c, Kimberly W. Schuenke a, Michael J. Keherly a,
Charles Holzer III d, Frances J. Richey a, Christopher J. Lahart e
a
Texas NeuroAIDS Research Center Department of Pathology, Rt 0785, University of Texas Medical Branch, Galveston, TX 77555-0785, United States
b
Departments of Anatomy and Neuroscience, University of Texas Medical Branch, Galveston, TX, United States
c
Department of Neurology, University of Texas Medical Branch, Galveston, TX, United States
d
Department of Psychiatry, University of Texas Medical Branch, Galveston, TX, United States
e
Department of Internal Medicine, Baylor College of Medicine, Houston, TX, United States

Accepted 30 August 2004

Abstract

A gene expression profile of the human brain cortex was performed in people with HIV-1-associated dementia (HAD) using Affymetrix
HG-U133 chips. Messenger RNA transcripts in middle frontal gyrus from subjects with HAD or milder neurocognitive dysfunction were
compared to HIV-negative people. The analysis focused on ionic conductance carriers that control membrane excitation. Overexpressed
ionic channel genes in brain cortex of subjects with dementia included (1) a calcium-driven K+ channel that prolongs afterhyperpolarization
(AHP) current, (2) a leak type of K+ channel that prolongs the AHP, (3) an adenosine receptor that modulates cationic current via G proteins,
(4) a G protein-coupled serotonin receptor that modulates cyclic AMP-linked current transduction, (5) a G protein-coupled dopamine
receptor, (6) a GABA receptor subunit that conducts chloride current. Underexpressed current generators in the demented subjects included
(1) two voltage-gated K+ channels that influence refractory periods and the onset of AHP, (2) a Na+ channel subunit that modifies current
inactivation and the onset of the AHP, (3) a neuronal type of voltage-sensitive Ca2+ channel that controls postsynaptic membrane
excitability, (4) a metabotropic glutamate receptor that regulates cationic gating via G protein coupling, (5) A specific Ga protein that
transduces metabotropic cationic current, (6) an NMDA receptor subunit, (7) a glycine receptor subunit that modulates chloride current.
These gene expression shifts probably occurred in neurons because they were not present in gyral white matter. Acquired neuronal
channelopathies were not associated with a generalized shift of neuronal or glial cell markers, which suggest that they were not an artifact
produced by neurodegeneration and/or glial cell proliferation. Channelopathies were not correlated with a generalized increase of
inflammatory cell transcripts and were present in demented people without, and with HIV encephalitis (HIVE). We surveyed experimentally
induced perturbations of these channels to determine the implications for brain function. Eleven experimental channelopathies produced
decreased neuronal firing frequencies and pacemaker rates in model neurons; seven channelopathies increase neuronal firing rates
experimentally. The implied disruption of neuronal excitability is consistent with some features of HAD, including its potential reversibility
after HIV-1 replication is suppressed, the abnormal electroencephalographic recordings, the lack of clear-cut correlation with
neurodegeneration and the lack of strict correlation with brain inflammation. The channelopathy concept may have wide relevance to
the subcortical dementias.
D 2004 Elsevier B.V. All rights reserved.

Keywords: AIDS; Autopsy; Afterhyperpolarization; Calmodulin; Channelopathy; Chloride; Calcium; Dementia; G protein; HIV encephalitis; Ionotropic;
Membrane potential; Metabotropic; Potassium; Sodium

1. Introduction
* Corresponding author. Texas NeuroAIDS Research Center, Depart- Cellular pathology is not an end if one cannot see any
ment of Pathology, Rt 0785, University of Texas Medical Branch,
Galveston, TX 77555-0785, United States. Tel.: +1 409 772 5316; fax:
alteration in the cell. Chemistry brings the clarification of
+1 409 772 5220. living processes nearer than does anatomy. Each anatom-
E-mail address: bgelman@utmb.edu (B.B. Gelman). ical change must have been preceded by a chemical one.
0165-5728/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jneuroim.2004.08.044
112 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119

–attributed to Rudolph Virchow (Coper and Herken,
1963)
Infection with the human immunodeficiency virus type 1
(HIV-1) can produce dementia (HAD) and mild cognitive
and motor disorder (MCMD). The prevalence of HAD
during the first decade of the epidemic of acquired
immunodeficiency syndrome (AIDS) was at least 20%.
The prevalence of brain dysfunction overall, including
MCMD was greater (Heaton et al., 1995). The incidence
of HAD decreased by at least half after highly active
antiretroviral therapy (HAART) was introduced in the USA.
Effective suppression of HIV-1 replication with HAART
often seems to reverse the progressively worsening course
of HAD, and can produce clinical improvement in many
patients (Dougherty et al., 2002; Robertson et al., 2004). It
is suggested that HAD may be a neurodegenerative disease,
akin to the senile dementias such as Alzheimer’s disease
(AD) (Esiri et al., 1998; Everall et al., 1994; Masliah et al.,
1992). Classical neurodegeneration does, however, leave
certain aspects of HAD unexplained. HAD is potentially
reversible and probably stems from a subcortical disturb-
ance, which differs from the irreversible cortical pathology
of AD. Cumulative nerve cell death, the hallmark of
neurodegenerative diseases such as AD (Roth, 2001; Terry
et al., 1991), is not an established requirement for HAD
(Budka et al., 1991; Everall et al., 1994; Ketzler et al.,
1990). The senile dementias and most neurodegenerative
diseases generally produce bsignatureQ neuropathological
lesions at autopsy, whereas a hallmark neuropathological
lesion was not found in some cases of HAD (Bell, 1998;
Glass et al., 1995). Brain inflammation is hypothesized to be
critical in HAD (Wiley and Achim, 1994), yet some people
with HAD did not have HIV encephalitis (HIVE) or other
specific inflammatory change at autopsy (Glass et al., 1995).
As well, some people with a postmortem diagnosis of HIVE
did not have strong historical evidence of HAD.
Gene expression profiling (GEP) offers a potentially
revealing way to screen for biochemical evidence of
neuronal dysfunction of the human brain. GEP can detect
abnormal expression of brain proteins that propagate ionic
currents and action potentials in neurons, which in turn, can
provide insights into abnormal neurophysiology. GEP of
autopsy brain specimens can be clinically relevant provided
that neurocognitive dysfunction is measured shortly before
the autopsy, as was done by the National NeuroAIDS
Tissue Consortium (NNTC) (Morgello et al., 2001). To
elucidate the neurochemical basis of neurocognitive dys-
function, we used GEP technology and NNTC brain tissue
to determine whether HAD is associated with transcrip-
tional changes in proteins that drive ionic currents in
neuronal membranes.
2. Methods
2.1. Subject selection
The study subjects were recruited prospectively accord-
ing to protocols employed by the NNTC (Morgello et al.,
2001). Briefly, informed consent was obtained from
individuals who were infected with HIV-1 and who had
a high risk of mortality due to AIDS. The NNTC specimen
repository in Galveston, TX was screened for persons with
a clinical diagnosis of HAD or MCMD or NPI-O (see
Table 1). Nine decedents were selected that had global
clinical ratings of neurocognitive impairment, an overall
measure of their test performance, which ranged from 5
through 9 (see below for methods). A score higher than 4
indicated significant clinical impairment and 7 or higher
qualified for a diagnosis of dementia. All of the impaired
subjects had a degree of dementia that was greater than 1.0
using the Memorial Sloan Kettering (MSK) dementia scale

Table 1
Study subjects for GeneChip analysis of HIV-associated dementia
Subject Age (years) Gender (M/F) HIV/AIDS PMI (h) Impaired Comorbid condition HIVE *CIR
1 Control 27 F No 3 No No No (b4)
2 Control 34 M No 24 No No No (b4)
3 Control 39 M No 8 No No No (b4)
4 Control 64 M No 24 No No No (b4)
5 Control 58 F No 7 No No No (b4)
6 NPI-O 53 M Yes 5 Probably Yes No 8
7 NPI-O 49 M Yes 15 Probably Yes No 8
8 MCMD 35 M Yes 9 Yes No No 6
9 HAD 39 M Yes 3 Yes No No 8
10 HAD 30 M Yes 6 Yes No No 9
11 HAD 38 M Yes 6 Yes No Yes 9
12 HAD 42 M Yes 6 Yes No Yes 6
13 HAD 49 M Yes N24 Yes No Yes 8
14 MCMD 32 M Yes 14 Yes No Yes 5
Acquired immunodeficiency syndrome (AIDS), HIV-associated dementia (HAD), human immunodeficiency virus (HIV), HIV encephalitis (HIVE), Minor
Cognitive and Motor Disorder (MCMD), neuropsychologically impaired but not definitely caused by HIV infection due to comorbid factor (NPI-O), Clinical
Impairment Rating (*CIR). An CIR of b4 is normal or asymptomatic, and N4 is clinically abnormal. The listed order of subjects in the table matches the order
shown in Fig. 2.
 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119
(Price and Brew, 1988); two out of nine had impairment
that was diagnosed as MCMD. Two neuropsychologically
impaired subjects were deliberately included who had a
comorbid condition in addition to HIV-1 infection (cocaine
abuse and potential mental retardation); these two subjects
are diagnosed bNPI-OQ. For the statistical comparison, all
nine HIV-1-infected people with neurocognitive impair-
ment were compared to the five HIV negative control
specimens. The abnormal brain mRNAs presented herein
were significantly different using all nine impaired people;
if significance depended on including the NPI-O subjects
they were rejected from this analysis. Clinical ratings in
use by the NNTC permitted us to examine a wide range of
functional impairment (Woods et al., 2004). Four subjects
had global clinical rating scores of 5–7; five people had
more severe neurocognitive impairment with scores of 8 or
9. To determine whether neurodegeneration or inflamma-
tion was required to produce HAD-associated transcrip-
tional changes, the presence and absence of neuropatho-
logical change also was built into the study design. Four of
the nine subjects with neurocognitive dysfunction had a
neuropathological diagnosis of HIVE; five of them did not
have HIVE and were neuropathologically normal. Since
the majority of demented people did not have HIVE in this
series, significant differences did not depend upon brain
inflammation per se. Subjects were selected in part
according to whether the postmortem time interval was
short, so as to permit maximal preservation of brain
mRNAs.
2.2. Neurocognitive and neurological evaluation
Subjects had regular neuropsychological testing and
structured neurological examinations repeated at 6-month
intervals. Data from the final testing session before death
were used. Briefly, the evaluation was comprised of
standardized measures of attention/working memory, speed
of information processing, verbal fluency, learning, mem-
ory, executive function, and motor speed, which are
compromised in MCMD and HAD (Heaton et al., 1981;
Heaton et al., 1995; Woods et al., 2004). Raw scores were
transformed into age, education, or ethnicity-corrected
standardized T scores. The transformed scores yielded
deficit scores (ranging from 0 to 5), and clinical ratings
of impairment (ranging from 1 to 9) for individual testing
domains. Subjects were assigned a global deficit score and
a global clinical rating of impairment (Woods et al., 2004).
A screening interview was conducted to obtain data
regarding potential confounds in diagnosis. Substance
abuse and psychiatric history were surveyed systematically
using an established instrument (Hasin et al., 1996). A
neurologist performed structured neurological examinations
according to criteria established by the NNTC. The neuro-
cognitive diagnosis was assigned in conference using
neurobehavioral test scores, neurological data and neuro-
psychiatric data.
113
2.3. Postmortem neuropathology and brain banking
Subjects received a complete autopsy with neuropatho-
logical examination. One hemisphere of the brain was sliced
coronally in the fresh state, was chilled on copper plates to a
temperature of 83 8C and maintained in the Texas
Repository of the NNTC in Galveston, TX. The neuro-
pathological diagnosis was established in the contralateral
unfrozen hemisphere, which was immersed in 10% neutral
buffered formalin at 5 8C for 10 days before being dissected.
We selected frontal lobe for study (Brodmann area 8)
because it resists postmortem degradation (Schuenke and
Gelman, 2003), is involved in subcortical dementias and is
functionally active during working memory delay (Filley,
2001; Passingham and Rowe, 2002). Using a circular
porcelain cutting tool (Model 395 Type 5, Dremel, Racine,
WI), about 750 mg of tissue was excised in the frozen state
and immediately placed in chilled TRI ReagentR (T9424,
Sigma-Aldrich, St. Louis, MO) for mRNA extraction. We
dissected a sample of cortical gray matter and adjacent gyral
white matter both. The two samples were obtained between
5 and 10 mm distant from each other, which would include
interconnecting axonal projections to maximize the chance
that the pair was functionally interrelated (Fig. 1).
2.4. Extraction of brain mRNA and gene expression
profiling
Total RNA was isolated from the dissected gray and
white matter samples according to manufacturer’s recom-
mended protocol (Chomczynski and Sacchi, 1987). Gray
and white matter samples were analyzed separately on
Affymetrix HG-U133 chips according to the manufacturer’s
instructions in a dedicated core facility maintained by the
University of Texas Medical Branch. All in all, 14 samples
of gray matter and 14 samples of white matter were obtained
for gene expression profiling from decedents.
3. Results
3.1. Spotfire screening of dysregulated transcripts
To deal with type b1Q error, all 44,780 raw signals (probe
sets) on the Affymetrix HG-U133 A and B GeneChips were
filtered using a one way analysis of variance to exclude all
nondetected transcripts, i.e., those transcripts not detected in
any of the human brain tissue specimens (Absent filtered
genes). The raw signal scores of the remaining filtered probe
sets were normalized by Z scores to reduce the variability in
message signal between the probe sets (Scaling process)
using Spotfire DecisionSite 7.1 software. A profile search
was used to identify those probe sets that were different
between the five control samples and the nine samples from
HIV+ subjects who had neurocognitive impairment (Two-
Group Model). From the profile search we generated a list
114 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119

Fig. 1. This drawing depicts the brain tissue sampling strategy. Two samples were obtained from middle frontal gyrus in Brodmann area 8 as shown at left. One
was a full-thickness sample of cortex that contained predominantly triangular-shaped neuron cell bodies and few small round oligodendrocytes (top). The
adjacent sample of gyral white matter contained predominantly oligodendrocytes (bottom) and no neuron cell bodies (bottom). The two samples were obtained
5–10 mm apart to obtain clear-cut separation and still preserve afferent and efferent axonal connections, so as to impart some functional interrelationship
between them.

of 1000 transcripts that best fit this model (top 500 where for a possible function in ion channel conductance. The
gene expression was higher in HIV+ versus control samples database was specifically searched for mRNA transcripts
and top 500 where the gene expression was lower in HIV+ that are associated with ionic current conduction directly as
versus control samples) and then individually screened them with channel pore proteins, or indirectly as with proteins

Fig. 2. Hierarchical representation of normalized signal intensities (bheat mapsQ) of 18 ion channel mRNAs in people with HAD/MCMD (bdementedQ) versus
nondemented subjects without HIV (bcontrolsQ). Red=high signal intensity, Green=low intensity, gray/black=intermediate intensity. Ion channel mRNAs were
altered in brain cortex (left), but not in adjacent white matter (right). Blank horizontal spaces in the white matter heat map at right denote probe signal intensities
too weak to be considered present. Gene symbols in the midline are identified fully in the text and Table 1. See Table 2 for statistical analysis. NP1 and NP 2 are
the two subjects diagnosed NPI-O because they had a comorbid factor other than HIV infection that can produce neuropsychological impairment. HIVE=HIV
encephalitis. HAD=HIV-associated dementia. C1–C5=subjects not infected with HIV.
 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119 115

that modify channel pore behavior. We included proteins
that do not form ion channel pores but play obligatory roles
in regulating the gating properties of channels. As well, we
scrutinized ionotropic and metabotropic receptors, and key
polypeptides required for transduction of channel modifi-
cation. Voltage-sensitive and -insensitive ion channels both
were included if the current they produce involves
membrane excitation. The selected 18 transcripts for ion
channels or channel pore modifiers were analyzed by the
Hierarchical Clustering Method of Unweighted Pair-Group
Method with arithmetic mean based on Similarity measure-
ments (Euclidean distance) (Mirkin, 1996) to generate a
bHeat MapQ (Fig. 2). These 18 transcripts were found to be
statistically different between the two groups (Control and
HIV+), as analyzed by two-tailed Student t test ( pb0.017)
(Table 2). Six of these genes were overexpressed and 12
were underexpressed as shown in Table 2 and Fig. 2.
3.2. Abnormal ion channel shifts that would decrease
repetitive firing
We surveyed the literature of experimentally induced
perturbations of the abnormally expressed ion channels, to
determine implications for neuronal physiology and brain

Table 2
Affymetrix signal intensities of ion channel related messages in human brain cortex
Affymetrix Gene name Gene Control group HIV Group Student Experimental Reference
probe set symbol (averageF (averageF t test effect on
S.E.M., N=5) S.E.M., N=9) p value neuron
firing rate
244680_at glycine receptor, beta GLRB 545F41 328F30 0.0010 z Rees et al., 2002;
Muller et al., 2003
224861_at guanine nucleotide binding protein GNAQ 2710F326 1513F175 0.0038 A Lesage et al., 2000;
(G protein), q polypeptide Chemin et al., 2003
238077_at hypothetical protein MGC27385 MGC27385 1168F178 513F73 0.0017 A Pongs, 1999
(KCTD6)
213108_at calcium/calmodulin-dependent CAMK2A 10688F1964 2449F689 0.0004 z Wang et al., 2002;
protein kinase (CaM kinase) Churn et al., 2000;
II alpha Silva et al., 1992
225019_at calcium/calmodulin-dependent CAMK2D 2403F263 1456F116 0.0024 z Wang et al., 2002;
protein kinase (CaM kinase) II delta Churn et al., 2000
235781_at calcium channel, voltage-dependent, CACNA1B 1040F75 659F54 0.0013 z Hallworth et al.,
L type, alpha 1B subunit 2003
212243_at glutamate receptor, ionotropic, GRINL1A 974F58 753F36 0.0050 A Liu et al., 2004
N-methyl d-aspartate-like 1A
220294_at potassium channel, subfamily V, KCNV1 290F42 107F19 0.0007 A Pongs, 1999
member 1
207299_s_at glutamate receptor, metabotropic 1 GRM1 495F44 261F35 0.0016 A Chemin et al., 2003;
Kohler et al., 1996
204723_at sodium channel, voltage-gated, SCN3B 3679F503 1540F331 0.0031 A Vijayaragavan et al.,
type III, beta 2001
232401_at potassium voltage-gated channel, KCNS2 236F26 67F20 0.0003 A Pongs, 1999
delayed-rectifier, subfamily S,
member 2
212757_s_at calcium/calmodulin-dependent CAMK2G 3560F198 1706F351 0.0030 z Wang et al., 2002;
protein kinase (CaM kinase) Churn et al., 2000
II gamma
216220_s_at adenosine A1 receptor ADORA1 422F63 718F68 0.0142 A Lin and Phillis,
1991; Marks et al.,
1993
206525_at gamma-aminobutyric acid (GABA) GABRR1 99F15 195F24 0.0163 A Cheng et al., 2001;
receptor, rho 1 Zheng et al., 2003
207577_at 5-hydroxytryptamine (serotonin) HTR4 164F23 416F65 0.0169 z Blondel et al., 1997
receptor 4
205902_at potassium intermediate/small KCNN3 109F51 909F158 0.0033 A Wolfart et al., 2001;
conductance calcium-activated Stackman et al.,
channel, subfamily N, member 3 2002; Hallworth
et al., 2003
220727_at potassium channel, subfamily K, KCNK10 255F56 532F30 0.0004 A Kohler et al., 1996;
member 10 Pongs, 1999;
Lesage et al., 2000;
Chemin et al., 2003
216924_s_at dopamine receptor D2 DRD2 216F32 444F28 0.0003 z Wang et al., 2004
116 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119
function. We focused on repetitive firing rates and pace-
maker firing to derived predictions concerning how the 18
abnormal ion channels and modifiers influence the excita-
tion of cortical neurons. Eleven of the altered channels
produced decreased firing rates in model neurons when
channel function was shifted in the same direction as
observed in HAD brain cortex. Four K+ channel superfamily
genes were abnormally expressed. TREK2 (KCNK10) is a
K+ bleakQ channel that controls action potential firing rates
by modulating the duration of the afterhyperpolarization
(AHP). An increase in the number of channels in HAD
predicts that the resting membrane potential would become
more negative; in model neurons this change decreases
repetitive firing rates by prolonging the AHP (Pongs, 1999).
The increase in KCNN3, a Ca2+ driven K+ channel (SK3, or
bGardos channelQ) has a similar influence. Increasing
KCNN3 channel opening slows repetitive firing by prolong-
ing the AHP (Hallworth et al., 2003; Stackman et al., 2002;
Wolfart et al., 2001). A voltage-driven K+ channel (KCNV1,
KCNB3, bHNKAQ, Kv8.1) and its coexpressed subunit
(KCNS2, Kv9.2) were decreased in tandem in HAD and
would decrease the firing rate. Here intraspike repolarization
is slow, which prolongs the refractory period and delays the
onset of the AHP (Pongs, 1999). The decrease of
MGC27385 (KCTD6), a hypothetical voltage-driven K+
channel, is likely to behave in a similar fashion. Out of at
least 12 voltage-gated Na+ channel subunits screened on the
chip, one was abnormal. The decrease of SCN3B slows
sodium channel pore closing (inactivation) during the action
potential, which prolongs the refractory period, delays the
onset of the AHP and decreases firing rates (Vijayaragavan
et al., 2001). One out of four members of the adenosine
receptor family, ADORA1, was increased selectively. This
change decreases neuronal firing rates, and is of interest in
HAD because the receptor may be stimulated by ATP that is
released during inflammation (Lin and Phillis, 1991; Marks
et al., 1993). The increase in GABAc rho 1 receptor
(GABRR1) decreases firing rates experimentally; as its
expression in the brain is limited, its role is not well
characterized (Cheng et al., 2001; Zheng et al., 2003). One
member out of eight in the family of metabotropic glutamate
receptors, mGluR1a (GRM1), was decreased. In tandem,
expression of the protein required for transduction of ligand
binding, the G alpha q subunit (GNAQ) also was decreased
and would exacerbate the decrease of mGluR1a. The
physiological effect of decreased mGluR1a is pleiotropic
because it modulates gating of numerous ion channel pores,
including two that were mentioned above (KCNK10 and
KCNN3). On balance decreased firing is likely due to
activation of KCNK10 (TREK2) (Chemin et al., 2003;
Kohler et al., 1996; Lesage et al., 2000). GRNL1A is a
newly recognized NMDA receptor subunit and is likely to
behave generally as other subunits; downregulation would
decrease excitation (Liu et al., 2004). All told, 11 out of 18
transcriptional shifts in HAD produced decreased firing
rates in model neurons.
3.3. Abnormal ion channel shifts that increase repetitive
firing rates
Seven of the altered channel-related genes in HAD
produce excitation in model neurons. Decreased ionotropic
glycine receptor beta subunit (GLRB), decreased voltage-
dependent Ca2+ channel N type alpha 1B (CACNA1B,
Cav2.2) and decreased Cam kinase II subunits (CAMK2)
increase firing rates experimentally (Churn et al., 2000;
Hallworth et al., 2003; Muller et al., 2003; Rees et al., 2002;
Silva et al., 1992). CAMK2 subunits drive membrane
physiology by phosphorylating subunit domains on many
channel proteins that control pore opening (Wang et al.,
2002). The decreased CAMK2 expression in HAD is highly
relevant to dementia because it produces problems in
memory and learning experimentally (Silva et al., 1992).
Dopamine 2 (DRD2) and serotonin 4 (5HT4) receptors are
members of multigene families; both were selectively
increased in HAD and produce increased firing rates
experimentally (Blondel et al., 1997; Lesage et al., 2000;
Wang et al., 2004). It is noteworthy that DRD2 mRNA is
localized primarily in cortical lamina V, which contains
excitatory pyramidal neurons that generate cortical output
and produce currents that drive abnormal brain wave
activity as measured with electroencephalography (EEG)
using scalp electrodes (Lidow et al., 1998; Martin, 1991).
4. Discussion
This is the first suggestion that acquired neuronal
channelopathies are involved in human neurocognitive
dysfunction. The concept of ion channel dysfunction in
human disease has gained momentum. Grouped under the
rubric of bchannelopathy,Q potential new examples are being
discovered steadily (Waxman, 2001a,b; Craner et al., 2004).
Inherited mutations of ion channels usually produce over-
excitation including brain seizures, muscle spasticity or
cardiac dysrhythmia (Kullmann, 2002). The diseases asso-
ciated with acquired changes in ion channel function are an
emerging class of disorders (Waxman, 2001a,b). Transcrip-
tional channelopathies recognized thus far include sodium
and calcium channelopathies that occur in multiple sclerosis
(MS), and sodium, potassium, and calcium channelopathies
in models of painful peripheral neuropathy (Waxman,
2001b). Experimental data shows clearly that disruption of
ion channel synthesis or channel pore gating causes very
substantial brain dysfunction, including alteration of long-
term potentiation and synaptic strength which are highly
relevant to clinical dementia (Giese et al., 2001). Here we
used GEP and clinically characterized human brain speci-
mens from the NNTC to survey HAD brains for evidence of
acquired neuronal channelopathies. The results and correla-
tion with the experimental physiology literature show that
altered transcriptional control of ion channel genes could be
important in HAD. The altered mRNAs are most likely to be
 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119
synthesized by neurons because they were not changed (or
not detected) in adjacent subcortical white matter, which
contains little neuronal mRNA (Fig. 2).
Altered ion channel gene expression could explain some
heretofore vexing features of HAD. Channelopathy theory
helps to explain why HAART can improve neurocognitive
dysfunction in HAD, especially the speed of psychomotor
processing. It also helps to explain why HAD autopsies do
not always reveal the classical pathological appearance of
neurodegeneration or neuroinflammation (Glass et al.,
1995). The potential reversibility of HAD differs from most
classical neurodegenerative diseases, including those that
produce neurocognitive dysfunction. For example, senile
dementia in AD results from the cumulative destruction of
the complexity of the neuronal network, including death of
whole neurons and neuropil components such as synapses
(Roth, 2001; Terry et al., 1991). The death of adult brain
neurons (neurodegeneration) is considered a permanent loss
and recovery of function after cell death is severely
restricted. In contrast to AD, functional restoration and the
lack of neuronal death are quite consistent with the principle
of a channelopathy. The channelopathy hypothesis of HAD
explains functional improvement after HAART as being a
return of viable neurons to normal electrophysiological
function, with restoration of optimal efficiency of cortical
electrical output. Channelopathy theory and neurodegenera-
tion are potentially contradictory, as channelopathy is
reversible while neuron loss is not. However, the two
processes might not be mutually exclusive, and could work
in tandem in HAD. Some evidence has suggested that
aberrantly expressed sodium channel pores are localized to
sites of axonal injury in people with MS (Craner et al.,
2004). Axonal injury has been observed in some HIV
infected people (An et al., 1997) and could be associated
with an axonal channelopathy similar to MS. Since
irreversible and reversible changes both are likely to occur
in HAD, we hypothesize that channelopathy and neuro-
degeneration both are important, and could represent
interactive processes. Whether or not transcriptional chan-
nelopathies participate in the process of neuronal injury and
death in HAD remains to be determined. One plausible
scenario is that neuronal channelopathies arise due to an
HIV-1-induced disturbance that generates inflammatory or
viral toxins (Nath and Geiger, 1998). Toxin-mediated
theories of HAD emphasize neuronal death and neuro-
degeneration because apoptosis is observed experimentally.
Toxin-induced stresses in the HIV-infected brain are dose-
dependent and potentially sublethal; these could have
produced the altered ion channel gene expression in HAD.
Prolonged and more concentrated exposure to these putative
neurotoxins may have the potential to produce a more
permanent nerve injury.
Channelopathy theory also may elucidate how abnormal
electrical current output from brain cortex is produced in
HAD, as documented using electroencephalography (EEG).
EEG reflects postsynaptic potentials of brain cortex that are
117
recorded using scalp electrodes; studies have suggested that
the amplitude of alpha and theta waves is increased in HAD
(Baldeweg and Gruzelier, 1997; Jeong et al., 2001; Newton
et al., 1994; Polich et al., 2000; Tinuper et al., 1990). EEG
discharges from brain cortex are believed to result from
feedback loops between firing of neocortical pyramidal
cells, and subcortical inputs primarily from thalamic relays
and the basal ganglia (Llinás et al., 1999; Martin, 1991).
Abnormal brain wave oscillation in HAD is consonant with
the underlying theory of the subcortical dementias; that is,
subcortical inputs to the cortex are disturbed that, in turn,
interfere with the coordination and timing of cortical firing
patterns, that in turn produces the characteristic decrease in
the speed of information processing (Filley, 2001). The
channelopathy theory of HAD predicts that dysrhythmic
cortical output as recorded on EEG may be the physio-
logical drive behind the decreased speed of processing. The
scenario of a subcortical disturbance driving cortical neuro-
nal channelopathies agrees with neuropathological reports
that emphasized the vulnerability of subcortical structures
(Bell, 1998). Widely cited pre-HAART era data showed that
a correlation between dementia and increased brain macro-
phages (inflammation) was present in subcortical brain
tissue, but not in a sample of frontal cortex (Glass et al.,
1995).
A limitation of GEP is that it is difficult to determine
if altered mRNA concentration in a homogenized brain
specimen represents a transcriptional change in neurons,
or instead, reflects abnormal proportions of cells due to
inflammatory cell infiltration, glial cell proliferation and
neuronal dropout. Many aspects of this study strongly
favor altered transcriptional regulation within neurons: (1)
There were substantial increases of certain neuronal
mRNA transcripts, and neurons do not proliferate in the
mature brain. (2) A high proportion of transcripts that
encoded characteristic structural elements of neurons were
not changed in either direction (data not shown), which
does not support a generalized dropout of neurons. (3)
There was no systematic shift in characteristic glial cell
markers mRNAs in brain cortex, which does not support
a substantial dilution of neuronal mRNAs due to
proliferation of glial cells. (4) Bidirectional changes of
neuron-specific messages were present; they cannot be
explained by cell population shifting in one direction as
would occur in neurodegeneration. It remains possible
that there was dropout of specific subpopulations of
neurons that were enriched with the apparently abnormal
mRNAs; we plan to evaluate that possibility in future
investigation. The acquired neuronal channelopathy con-
cept could apply to the subcortical dementias generally,
including HAD, Parkinson’s Disease, Huntington’s Dis-
ease, Binswanger’s Disease and others (Berger and
Arendt, 2000; Draper, 1991; Filley, 2001; Llinás et al.,
1999). More study is needed on ion channel expression
and the posttranscriptional control of ion channel gating
in demented people.
118 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119
Acknowledgements
We thank Steve Schuenke for the drawing in Fig. 1 and
Dr. Harvey Fishman for physiological consultations. Carlos
Huerta Jr. and Joshua G. Lisinicchia provided technical
assistance. This research was supported by the National
Institutes of Health (NS45491, MH59656, MH69200,
DC04970).
References
An, S.F., Giometto, B., Groves, M., Miller, R.F., Beckett, A.A., Gray, F.,
Tavolato, B., Scaravilli, F., 1997. Axonal damage revealed by
accumulation of beta-APP in HIV-positive individuals without AIDS.
J. Neuropathol. Exp. Neurol. 56, 1262 – 1268.
Baldeweg, T., Gruzelier, J.H., 1997. Alpha EEG activity and subcortical
pathology in HIV infection. Int. J. Psychophysiol. 26, 431 – 442.
Bell, J.E., 1998. The neuropathology of adult HIV infection. Rev. Neurol.
(Paris) 154, 816 – 829.
Berger, J.R., Arendt, G., 2000. HIV dementia: the role of the basal ganglia
and dopaminergic systems. J. Psychopharmacol. 14, 214 – 221.
Blondel, O., Vandecasteele, G., Gastineau, M., Leclerc, S., Dahmoune, Y.,
Langlois, M., Fischmeister, R., 1997. Molecular and functional
characterization of a 5-HT4 receptor cloned from human atrium. FEBS
Lett. 412, 465 – 474.
Budka, H., Wiley, C.A., Kleihues, P., Artigas, J., Asbury, A.K., Cho, E.S.,
Cornblath, D.R., Dal Canto, M.C., DeGirolami, U., Dickson, D., 1991.
HIV-associated disease of the nervous system: review of nomenclature
and proposal for neuropathology-based terminology. Brain Pathol. 1,
143 – 152.
Chemin, J., Girard, C., Duprat, F., Lesage, F., Romey, G., Lazdunski, M.,
2003. Mechanisms underlying excitatory effects of group I metabo-
tropic glutamate receptors via inhibition of 2P domain K+ channels.
EMBO J. 22, 5403 – 5411.
Cheng, Q., Kulli, J.C., Yang, J., 2001. Suppression of neuronal hyper-
excitability and associated delayed neuronal death by adenoviral
expression of GABA(C) receptors. J. Neurosci. 21, 3419 – 3428.
Chomczynski, P., Sacchi, N., 1987. Single-step method of RNA isolation
by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal.
Biochem. 162, 156 – 159.
Churn, S.B., Sombati, S., Jakoi, E.R., Severt, L., DeLorenzo, R.J., 2000.
Inhibition of calcium/calmodulin kinase II alpha subunit expression
results in epileptiform activity in cultured hippocampal neurons. Proc.
Natl. Acad. Sci. U. S. A. 97, 5604 – 5609.
Coper, H., Herken, H., 1963. Damage to the central nervous system by
antimetabolites of nicotinamide. A contribution to molecular
pathology of pyridine nucleotides. Dtsch. Med. Wochenschr. 88,
2025 – 2036.
Craner, M.J., Newcombe, J., Black, J.A., Hartle, C., Cuzner, M.L.,
Waxman, S.G., 2004. Molecular changes in neurons in multiple
sclerosis; altered axonal expression of Nav1.2, and Nav1.6 sodium
channels and Na+/Ca2+ exchanger. Proc. Natl. Acad. Sci. U. S. A. 101,
8168 – 8173.
Dougherty, R.H., Skolasky Jr., R.L., McArthur, J.C., 2002. Progression
of HIV-associated dementia treated with HAART. AIDS Read. 12,
69 – 74.
Draper, B., 1991. Potentially reversible dementia: a review. Aust. N. Z. J.
Psychiatry 25, 506 – 518.
Esiri, M.M., Biddolph, S.C., Morris, C.S., 1998. Prevalence of Alzheimer
plaques in AIDS. J. Neurol. Neurosurg. Psychiatry 65, 29 – 33.
Everall, I.P., Glass, J.D., McArthur, J., Spargo, E., Lantos, P., 1994.
Neuronal density in the superior frontal and temporal gyri does not
correlate with the degree of human immunodeficiency virus-associated
dementia. Acta Neuropathol. (Berl.) 88, 538 – 544.
Filley, C.M., 2001. The Behavioral Neurology of White Matter. Oxford
University Press, New York, pp. 247 – 267.
Giese, K.P., Peters, M., Vernon, J., 2001. Modulation of excitability as a
learning and memory mechanism: a molecular genetic perspective.
Physiol. Behav. 73, 803 – 810.
Glass, J.D., Fedor, H., Wesselingh, S.L., McArthur, J.C., 1995. Immuno-
cytochemical quantitation of human immunodeficiency virus in the
brain: correlations with dementia. Ann. Neurol. 38, 755 – 762.
Hallworth, N.E., Wilson, C.J., Bevan, M.D., 2003. Apamin-sensitive small
conductance calcium-activated potassium channels, through their
selective coupling to voltage-gated calcium channels, are critical
determinants of the precision, pace, and pattern of action potential
generation in rat subthalamic nucleus neurons in vitro. J. Neurosci. 23,
7525 – 7542.
Hasin, D.S., Trautman, K.D., Miele, G.M., Samet, S., Smith, M., Endicott,
J., 1996. Psychiatric Research Interview for Substance and Mental
Disorders (PRISM): reliability for substance abusers. Am. J. Psychiatry
153, 1195 – 1201.
Heaton, R.K., Grant, I., Anthony, W.Z., Lehman, R.A., 1981. A comparison
of clinical and automated interpretation of the Halstead–Reitan Battery.
J. Clin. Neuropsychol. 3, 121 – 141.
Heaton, R.K., Grant, I., Butters, N., White, D.A., Kirson, D., Atkinson,
J.H., McCutchan, J.A., Taylor, M.J., Kelly, M.D., Ellis, R.J., et al.,
1995. The HNRC 500-neuropsychology of HIV infection at different
disease stages. HIV Neurobehavioral Research Center. J. Int. Neuro-
psychol. Soc. 1, 231 – 251.
Jeong, J., Chae, J.H., Kim, S.Y., Han, S.H., 2001. Nonlinear dynamic
analysis of the EEG in patients with Alzheimer’s disease and vascular
dementia. J. Clin. Neurophysiol. 18, 58 – 67.
Ketzler, S., Weis, S., Haug, H., Budka, H., 1990. Loss of neurons
in the frontal cortex in AIDS brains. Acta Neuropathol. (Berl.) 80,
92 – 94.
Kohler, M., Hirschberg, B., Bond, C.T., Kinzie, J.M., Marrion, N.V.,
Maylie, J., Adelman, J.P., 1996. Small-conductance, calcium-activated
potassium channels from mammalian brain. Science 273, 1709 – 1714.
Kullmann, D.M., 2002. The neuronal channelopathies. Brain 125,
1177 – 1195.
Lesage, F., Terrenoire, C., Romey, G., Lazdunski, M., 2000. Human
TREK2, a 2P domain mechano-sensitive K+ channel with multiple
regulations by polyunsaturated fatty acids, lysophospholipids, and
Gs, Gi, and Gq protein-coupled receptors. J. Biol. Chem. 275,
28398 – 28405.
Lidow, M.S., Wang, F., Cao, Y., Goldman-Rakic, P.S., 1998. Layer V
neurons bear the majority of mRNAs encoding the five distinct
dopamine receptor subtypes in the primate prefrontal cortex. Synapse
28, 10 – 20.
Lin, Y., Phillis, J.W., 1991. Characterization of the depression of rat
cerebral cortical neurons by selective adenosine agonists. Brain Res.
540, 307 – 310.
Liu, L., Wong, T.P., Pozza, M.F., Lingenhoehl, K., Wang, Y., Sheng, M.,
Auberson, Y.P., Wang, Y.T., 2004. Role of NMDA receptor subtypes in
governing the direction of hippocampal synaptic plasticity. Science 304,
1021 – 1024.
Llinás, R.R., Ribary, U., Jeanmonod, D., Kronberg, E., Mitra, P.P., 1999.
Thalamocortical dysrhythmia: a neurological and neuropsychiatric
syndrome characterized by magnetoencephalography. Proc. Natl. Acad.
Sci. U. S. A. 96, 15222 – 15227.
Marks, J.D., Donnelly, D.F., Haddad, G.G., 1993. Adenosine-induced
inhibition of vagal motoneuron excitability: receptor subtype and
mechanisms. Am. J. Physiol. 264, L124 – L132.
Martin, J.H., 1991. The collective electrical behavior of cortical neurons:
the electroencephalogram and the mechanisms of epilepsy, Chapter 50.
In: Kandel, E.R., Schwartz, J.H., Jessel, T.M. (Eds.), Principle of Neural
Science. Appleton and Lange, Norwalk, CT, pp. 779 – 785.
Masliah, E., Ge, N., Achim, C.L., Hansen, L.A., Wiley, C.A., 1992.
Selective neuronal vulnerability in HIV encephalitis. J. Neuropathol.
Exp. Neurol. 51, 585 – 593.
 B.B. Gelman et al. / Journal of Neuroimmunology 157 (2004) 111–119
Mirkin, B., 1996. Mathematical Classification and Clustering, Nonconvex
Optimization Its Applications. Kluwer Academic Publishers, The
Netherlands.
Morgello, S., Gelman, B.B., Kozlowski, P.B., Vinters, H.V., Masliah, E.,
Cornford, M., Cavert, W., Marra, C., Grant, I., Singer, E.J., 2001. The
National NeuroAIDS Tissue Consortium: a new paradigm in brain
banking with an emphasis on infectious disease. Neuropathol. Appl.
Neurobiol. 27, 326 – 335.
Muller, F., Heinke, B., Sandkuhler, J., 2003. Reduction of glycine
receptor-mediated miniature inhibitory postsynaptic currents in rat
spinal lamina I neurons after peripheral inflammation. Neuroscience
122, 799 – 805.
Nath, A., Geiger, J., 1998. Neurobiological aspects of human immunode-
ficiency virus infection: neurotoxic mechanisms. Prog. Neurobiol. 54,
19 – 33.
Newton, T.F., Leuchter, A.F., Miller, E.N., Weiner, H., 1994. Quantitative
EEG in patients with AIDS and asymptomatic HIV infection. Clin.
Electroencephalogr. 25, 18 – 25.
Passingham, R.B., Rowe, J.B., 2002. Dorsal prefontal cortex: maintenance
in memory or attention selection? In: Stuss, D.T., Knight, H. (Eds.),
Principles of Frontal Lobe Function. Oxford University Press, New
York, N.Y, pp. 225.
Polich, J., Ilan, A., Poceta, J.S., Mitler, M.M., Darko, D.F., 2000.
Neuroelectric assessment of HIV: EEG, ERP, and viral load. Int. J.
Psychophysiol. 38, 97 – 108.
Pongs, O., 1999. Voltage-gated potassium channels: from hyperexcitability
to excitement. FEBS Lett. 452, 31 – 35.
Price, R.W., Brew, B.J., 1988. The AIDS dementia complex. J. Infect. Dis.
158, 1079 – 1083.
Rees, M.I., Lewis, T.M., Kwok, J.B., Mortier, G.R., Govaert, P., Snell,
R.G., Schofield, P.R., Owen, M.J., 2002. Hyperekplexia associated with
compound heterozygote mutations in the beta-subunit of the human
inhibitory glycine receptor (GLRB). Hum. Mol. Genet. 11, 853 – 860.
Robertson, K.R., Robertson, W.T., Ford, S., Watson, D., Fiscus, S.,
Harp, A.G., Hall, C.D., 2004. Highly active antiretroviral therapy
improves neurocognitive functioning. J. Acquir. Immune Defic. Syndr.
36, 562 – 566.
Roth, K.A., 2001. Caspases, apoptosis, and Alzheimer disease: causation,
correlation, and confusion. J. Neuropathol. Exp. Neurol. 60, 829 – 838.
Schuenke, K., Gelman, B.B., 2003. Human microglial cell isolation from
adult autopsy brain: brain pH, regional variation, and infection with
human immunodeficiency virus type 1. J. Neurovirology 9, 346 – 357.
119
Silva, A.J., Stevens, C.F., Tonegawa, S., Wang, Y., 1992. Deficient
hippocampal long-term potentiation in alpha-calcium-calmodulin kin-
ase II mutant mice. Science 257, 201 – 206.
Stackman, R.W., Hammond, R.S., Linardatos, E., Gerlach, A., Maylie, J.,
Adelman, J.P., Tzounopoulos, T., 2002. Small conductance Ca2+-
activated K+ channels modulate synaptic plasticity and memory
encoding. J. Neurosci. 22, 10163 – 10171.
Terry, R.D., Masliah, E., Salmon, D.P., Butters, N., DeTeresa, R., Hill, R.,
Hansen, L.A., Katzman, R., 1991. Physical basis of cognitive
alterations in Alzheimer’s disease: synapse loss is the major correlate
of cognitive impairment. Ann. Neurol. 30, 572 – 580.
Tinuper, P., de Carolis, P., Galeotti, M., Baldrati, A., Gritti, F.M., Sacquegna,
T., 1990. Electroencephalogram and HIV infection: a prospective study
in 100 patients. Clin. Electroencephalogr. 21, 145 – 150.
Vijayaragavan, K., O’Leary, M.E., Chahine, M., 2001. Gating properties of
Na(v)1.7 and Na(v)1.8 peripheral nerve sodium channels. J. Neurosci.
21, 7909 – 7918.
Wang, Z., Wilson, G.F., Griffith, L.C., 2002. Calcium/calmodulin-depend-
ent protein kinase II phosphorylates and regulates the Drosophila eag
potassium channel. J. Biol. Chem. 277, 24022 – 24029.
Wang, M., Vijayraghavan, S., Goldman-Rakic, P.S., 2004. Selective D2
receptor actions on the functional circuitry of working memory. Science
303, 853 – 856.
Waxman, S.G., 2001a. Acquired channelopathies in nerve injury and MS.
Neurology 56, 1621 – 1627.
Waxman, S.G., 2001b. Transcriptional channelopathies: an emerging class
of disorders. Nat. Rev., Neurosci. 2, 652 – 659.
Wiley, C.A., Achim, C., 1994. Human immunodeficiency virus encephalitis
is the pathological correlate of dementia in acquired immunodeficiency
syndrome. Ann. Neurol. 36, 673 – 676.
Wolfart, J., Neuhoff, H., Franz, O., Roeper, J., 2001. Differential expression
of the small-conductance, calcium-activated potassium channel SK3 is
critical for pacemaker control in dopaminergic midbrain neurons. J.
Neurosci. 21, 3443 – 3456.
Woods, S.P., Rippeth, J.D., Frol, A.B., Hinkin, C.H., Levy, J.L., Ryan, E.,
Soukup, V.M., Lazzaretto, D., Cherner, M., Marcotte, T.D., Gelman,
B.B., Morgello, S., Singer, E.J., Grant, I., Henton, R.K., 2004. Interrater
reliability of clinical ratings and neurocognitive diagnosis in HIV
dementia. J. Clin. Exp. Neuropsychol. 26, 759 – 778.
Zheng, W., Xie, W., Zhang, J., Strong, J.A., Wang, L., Yu, L., Xu, M., Lu,
L., 2003. Function of gamma-aminobutyric acid receptor/channel rho 1
subunits in spinal cord. J. Biol. Chem. 278, 48321 – 48329.