Académique Documents
Professionnel Documents
Culture Documents
www.elsevier.com/locate/jneuroim
Abstract
The etiology of HIV-associated dementia (HAD) is still unknown although direct viral effects have not been supported. Although
evidence supports a role for products of activated macrophages, other evidence suggested the possibility of associated autoimmune
phenomena at least as a marker. In a blinded analysis, non-HIV-infected whole brain material was immunoblotted with samples of serum, and
in certain cases cerebrospinal fluid (CSF), from HAD patients and controls. Distinct antibrain antibodies were detected in 11/12 of HIV+
HAD patients, 7/19 of HIV+ patients without HAD, and 0/11 HIV seronegative controls who were either healthy or had other neurologic
diseases. Reactivity against control tissue was negative. Though the etiopathogenetic relation of these antibrain antibodies remains to be
delineated, the data suggest that they may be a marker of HAD.
D 2003 Published by Elsevier Science B.V.
finding of increased intrathecal IgG synthesis in AIDS the ability to shift sets, and memory for verbal and figural
patients of which only 0.2 –2.8% is directed against HIV material (Levin, 1994).
(Resnick et al., 1985). HIV-1 serological status was determined in each by
Early studies with different methods produced different enzyme-linked immunoabsorbent assay (ELISA) and con-
results with respect to antibrain Abs. Although they did not firmed by Western blot. Laboratory studies also included
examine CSF, Kumar et al. (1989) demonstrated antibrain blood for B12 levels, VDRL, and FTA-ABS. Magnetic
serum Ab against a 45-kDa protein of human hippocampal resonance image (MR) of the brain was obtained on all
tissue. They found that 14/18 AIDS-dementia sera samples subjects. As part of the study design, lumbar puncture was
were positive for the brain reactive Ab as were 4/12 samples performed on all HIV-1 positive subjects. Routine CSF
from AIDS patients without the dementia. Mathiesen et al. analysis included cell count and differential, protein, glu-
(1989) demonstrated CSF autoAb complexed to MBP in cose, VDRL, FTA-ABS, immunoelectrophoresis, and viral
HAD patients. In contrast to this, Lenhardt and Wiley (1989), cultures on human foreskin fibroblasts, human embryonic
investigating the possibility of antibrain immunoglobulin in lung fibroblasts, and primary monkey kidney cells. Subjects
AIDS patients, were not able to find a significant association. with features of encephalopathy also had detailed micro-
Trujillo et al. (1994) did not find any distinct associations. biological studies of their CSF including bacterial stain and
We decided to investigate the existence of an antibrain cultures, fungal cultures, India ink preparations, and cryp-
antibody in the setting of HAD more thoroughly as a marker tococcal antigens. All subjects provided informed consent.
for the existence of the disorder because it would have great This study was approved by the institutional review board of
clinical utility. Furthermore, if autoAbs contributed to the the University of Miami School of Medicine.
pathogenesis of the disorder, their effect might be poten- AIDS dementia complex (ADC now termed HAD) was
tially reversible. This reversibility is illustrated by the CNS defined as in accordance with the Working Group of the
lupus psychosis syndrome; the classic humoral autoimmune American Academy of Neurology AIDS Task Force (1991).
disease, myasthenia gravis, caused by autoAbs directed Each experienced a ‘‘loss of intellectual abilities of suffi-
against the nicotinic acetyl choline receptor; Guillain –Barré cient severity to interview with social or occupational
Syndrome responsive to plasmapheresis; and the Lambert – functioning.’’ The criteria also included: (1) an acquired
Eaton myasthenic syndrome with autoAbs directed against abnormality in at least two cognitive and abilities present for
the presynaptic active zone sites. 1 or more months; acquired abnormality in order to function
We investigated the hypothesis that humoral immune or performance end/or declining in motivation, or emotional
mechanisms could be occurring in HAD by several methods control or change in social behavior; (3) no clouding of
probing for serum and CSF antibrain Abs. consciousness; and (4) no other identifiable etiology. As
indicated above, all subjects were systemicatically screened
for confounding factors that might contribute to their
2. Materials and methods dementing illness. HIV-infected subjects with dementia that
may have resulted from a neoplastic, nutritional, metabolic,
2.1. Sample collection, patient population, clinical evalua- or infectious (other than HIV) etiology were excluded from
tion this study. Also excluded were those with a severe premor-
bid psychiatric disorder, chronic seizure disorder, or head
Serum and CSF samples were obtained from 31 ran- trauma with residual dysfunction.
domly selected HIV-1 positive subjects from among 185
subjects participating in a longitudinal, prospective study of 2.1.1. Control samples
the neurological complications of HIV infection performed Control samples came from our bank of paired CSF and
at the University of Miami, Miami, FL. Subjects had a sera on normals and patients with other neurological dis-
complete clinical evaluation which included a detailed eases including neurologic Lyme disease with cognitive
history, general physical and neurological examination, dysfunction, viral and bacterial meningitis, amyotrophic
and extensive neuropsychological battery. The neurological lateral sclerosis, Guillain – Barré syndrome, herpes encepha-
examination included assessments of muscle tone, extremity litis, Creutzfeldt – Jakob disease, multiple sclerosis, syphilis,
dexterity, gait, balance and postural stability, and frontal sciatica, transverse myelitis, lupus, Alzheimer, and back
release signs (snout, glabellar, involuntary grasp). A modi- pain. The CSF normals derive from patients who had
fied version of the mini-mental status examination (Folstein symptoms such as back pain, had negative workups, and
et al., 1983) was administered. This battery was comprised eventually recovered and remained symptom-free. Many
of tests of orientation, simple calculations, ability to follow specimens were obtained before the AIDS epidemic.
commands, spelling forward and backward, naming and
recent memory and was graded on a 40-point scale. The 2.2. Analysis of autoAbs against brain
detailed neuropsychological battery included tests designed
to evaluate language function, judgment in reasoning, visuo- We initially used a traditional alkaline phosphatase West-
spatial constructive abilities, auditory and visual attention, ern blot technique and then modified the procedure to a
158 S.E. Schutzer et al. / Journal of Neuroimmunology 138 (2003) 156–161
biotin – avidin blot to analyze autoAb in the CSF and serum lanes, pre-stained SDS-PAGE standards) and diluted brain
samples by blotting normal brain. Control Ags included were all boiled 3– 5 min, cooled, and applied to wells of the
HIV negative human liver and neural Ags. Indirect immu- gel and 125 constant volts applied until samples lined up at
nofluorescence of brain was used initially and primarily to bottom of stacking gel, and then power supply switched to
screen large numbers of samples. 200 constant volts until dye front marker ran to bottom of
gel.
2.2.1. Alkaline phosphatase Western blots
Target Ags were optimized for protein, run on 10% 2.3.3. Electrotransfer
SDS-PAGE gels, then transferred to nitrocellulose mem- Gels were removed from plates and equilibrated for 15
branes. The resultant blots were probed with the appro- min in transfer buffer (25 mM Tris, 190 mM Glycine, 20%
priate class-specific alkaline phosphatase-conjugated Methanol), put in a sandwich containing wetted nitrocellu-
antihuman immunoglobulin antisera, then, stained. Molec- lose (MSI Nitro-Bind 0.2 Am) and transferred for 1 h at 100
ular weight standards and HIV lysates or the major core V (approximately 0.25 Amps) using Hoeffer apparatus (TE
and envelop proteins were run in tandem to these 22). Membrane blots were measured for dye front migra-
samples to provide markers and distinguish them from tion, appropriately marked if pre-stained markers were not
HIV Ags. used, rinsed, dried, and could be stored in the cold or stained
Target Ags included: for further use.
1. Normal human brain. This was obtained from a 2.3.4. Immunoblot stain
subject within hours of death, and was snap-frozen Nitrocellulose blots were stained in a 0.2% Ponceau S
in liquid nitrogen and stored at 70 jC. Samples solution (Sigma) which served two purposes. Firstly, the
were prepared as described by Hall and Choppin stained bands along with the Methyl Green dye front
(1981). marker helped to emboss the blot in proper alignment
2. Structural brain components Ags included galacto- with the 28 mini-well immunoblotter (Immunetics). Sec-
cerebroside (Sigma C-8752, Type I), and ganglioside ondly, the light red bands confirm that approximately 1
(Sigma G-2250, Type II) as positive controls in Ag of protein/well, the lower limit of Ponceau S, was
combination with the corresponding monoclonal Ab obtained during blotting, corresponding to the 1:64 dilu-
probe. tion, 10-Al sample of brain initially applied to the original
3. HIV lysates and Ags were used to distinguish an Ab gel well predetermined by a limiting dilution experiment
directed against brain from that against the HIV itself. of brain using secondary reagents in the biotin –avidin
assay.
2.3. Biotin –avidin Western blots The membrane was blocked in Tris-buffered saline
(TBS, 50 mM Tris – HCL, 0.2 M NaCl, 3 mM KCl, pH
2.3.1. Brain preparation 7.5, 0.02% Azide) containing 1% Non-Fat Dry Milk
A piece of brain which had been flash-frozen at autopsy (Carnation) and 1% normal goat serum (NGS) for 30 min
and kept at 70 jC was sliced with scalpel, and 0.226 g in a tray (on a slowly rocking platform (Reliable Scien-
was placed in a (P35) petri dish on ice with 1 ml of Laemmli tific). After rinsing in TBS, the membrane was placed in
gel sample-reducing buffer consisting of 20% SDS, 10% 60 the-28 lane miniblotter apparatus (MN28, Immunetics), and
mM Tris – HCL (pH 6.8), 0.1 M DTT, 0.005% bromphenol lanes aspirated dry with suction. Patient’s serum samples
blue, 1 mM PMSF. After mincing with a scalpel, the were diluted 1:100 in Tris-buffered saline –tween 20 (0.1%
solution was forcibly passaged up and down 10 times each v/v) (TBS-T) containing 1% NGS, patient CSF was used
through an 18-G followed by a 21-G needle. A portion was neat or at 1:10 dilution, and positive control mouse mono-
further diluted in sample buffer, and undiluted aliquots were clonal anti-neurofilament 68, NR-4 (Sigma) was diluted
frozen at 70 jC for later use. Before electrophoresis, the 1:500.
diluted sample was boiled for 3 – 5 min, cooled, and applied Sixty microliters was pipetted into each lane (diluent
to gel. alone in secondary reagent controls, and TBS in biotiny-
lated standard wells) and miniblotter was slowly rocked
2.3.2. Gel electrophoresis (setting of 2.8 on rocking platform) for 2 h for the
Twelve percent SDS-polyacrylamide (7 8 cm) mini- primary incubation. Wells were aspirated and filled with
gels, 0.75 mm thick with stack were run using Hoeffer TBS-T and miniblotter rocked rapidly for 5 min. This
apparatus (Mighty small, SE 250). Five microliters of a was repeated with TBS-T three more times and served as
0.1% Methyl Green dye solution was added to each well to the standard wash procedure between reagent incubations.
serve as a blotted dye front marker (for chamber orienta- Secondary Abs consisted of goat biotinylated antihuman
tion). Samples including SDS-PAGE biotinylated low-range IgG (gamma chain specific), biotinylated antihuman IgM
molecular weight standards (Bio-Rad) diluted in sample (mu chain specific) which were diluted 1:250, and bio-
buffer containing 5% beta-mercaptoethanol (and in some tinylated horse anti-mouse IgG (H + L) which was diluted
S.E. Schutzer et al. / Journal of Neuroimmunology 138 (2003) 156–161 159
1:500 (Vector Laboratories) in TBS-T plus 1% NGS. 3.2. Western blot demonstration of specific antibrain Abs in
After washing, secondary Abs were added and incubated serum and CSF
with slow rocking for 30 min. During this time, the
avidin (reagent A) and biotinylated alkaline phosphatase Western blots against non-HIV-infected brain tissue
(reagent B) complex was formed by adding 10 Al of demonstrated specific IgG antibrain autoAbs in cerebrospi-
reagent A and 10 Al of reagent B to 1 ml of TBS-T and nal fluid (CSF) and serum from HAD patients (n = 6).
vortexing. After standing at room temperature (RT) for 30 Reactivity was seen against two antigens between 41 and
min, it was further diluted by adding 2 ml of TBS-T. All 48.5 kDa; and a third between 48.5 and 58 kDa. The stained
wells of the miniblotter, including the biotinylated MW bands were distinct from those of HIV lysates probed in
standards were washed four times, and the diluted avi- parallel by these patients’ sera. As a control, multiple
din – biotinylated alkaline phosphatase complex (ABC) sclerosis serum and CSF did not stain any of these bands.
solution was added for a further 30-min incubation. After This suggested that there are at least some auto-IgG Abs
standard wash, the membrane was removed from the reactive to brain components in AIDS patients with demen-
miniblotter, rinsed once with TBS and developed in a tia and that they appear to be distinct from anti-HIV Abs and
tray using Vector Kit II Alkaline phosphatase substrate in are not found in controls. We considered these findings as
100 mM Tris –HCl pH 9.5 buffer, washed twice with preliminary but provocative to perform a blinded study with
water and dried. HAD patients and controls using a more sensitive biotin –
avidin blot.
3. Indirect immunofluorescence 3.2.1. Detection of antibrain Abs in serum and CSF from
HAD patients and controls by biotin – avidin Western blots
This assay was used as preliminary screen of the In a blinded analysis, the brain material was probed
concept of antibrain Abs using a different set of patient with samples of serum and cerebrospinal fluid from
samples. Undiluted CSF and diluted serum (1:5, 1:10, clinically well-defined groups from the University of
1:20) were overlaid on sections from normal human brain Miami. Distinct antibrain brain Abs were detected in 11/
which had been obtained within 2 h after death and snap- 12 of HIV+ HAD patients, 7/19 of HIV+ patients without
frozen in liquid nitrogen and stored at 70 jC until use. HAD, and 0/11 HIV seronegative controls who were
Brain sections were cut to 10-Am thickness by a micro- healthy or had other neurological disorders including
tome cryostat, placed on microscope slides, desiccated cognitive dysfunction. In one HAD patient, the reactive
overnight at 4 jC, placed in acetone for 1 min, washed in antibody was detected only in the CSF. Both IgM and IgG
phosphate-buffered saline (PBS), and then overlaid with antibrain Abs were detected. Statistical analysis by Fisher
the CSF or serum as above for 1 h at room temperature. exact test showed these findings to be significant (see
After washing again in PBS, a 1:40 dilution of fluores- Table 1). The weight (kDa) of these Ags were: 26, 31, 36,
cein-labeled goat antihuman IgG (Sigma) was applied for 40, 45, 55, 60, 80 –82. A representative blot is shown in
30 min. The slides were rewashed in PBS and mounted Fig. 1.
in 1% glycerol/PBS. They were read under epifluores- The data suggest that differences between HAD+ and
cence with a Zeiss photomicroscope III (Garvey et al., HAD HIV+ is in the percentage having some antibrain
1977). Abs rather than the total number of bands on the blot.
Liver and spleen served as tissue controls. Stated another way, it may be that HAD HIV patients
who have positive bands are on their way to developing
3.1. Positive indirect immunofluorescence of HIV+ sera on clinically detectable HAD but are not yet at that thresh-
normal mouse brain: demonstration of antibrain Abs old.
staining pattern. These findings suggested that IgG directed HAD , HIV + 7 12 2.5 10 2
against brain tissue was present in the serum of AIDS Non-HIV controls 0 11 reference
patients with dementia and perhaps those with subclinical Suggests possible marker role for ABA in HAD group and possible
disease. predictive marker of which HIV+ patients might progress to HAD.
160 S.E. Schutzer et al. / Journal of Neuroimmunology 138 (2003) 156–161
Fig. 1. Probe for reactive serum IgG to normal (non-HIV infected) brain. Lanes 1, 9, 10, and 12 show reactivity from HIV+ but not encephalopathy patients
(representing 7 of 19 patients). Lanes 2 – 8, 11 show reactivity from HIV+ HAD patients (representing 11 of 12 patients). Note that the major difference is in the
percentage of each subset reacting to these antigens.
Though the etiopathogenetic relation of these autoAbs al. (1987), in a study of encephalomyelitis and polyneuritis
remains to be delineated, the data suggest that they may be a following rabies vaccination, identified myelin basic protein
marker of HAD. (MBP) as an encephalitogen.
Although not of proven pathogenetic significance, in
multiple sclerosis (MS), autoAbs directed against brain
4. Discussion components have been described (Ryberg, 1982) and
anti-MBP in the CSF (even when hidden in the immune
The results of this set of experiments suggest that anti- complex form) has been demonstrated by several authors
brain Abs may be associated with the HAD. The data, while (Coyle and Procyk-Dougherty, 1984; Coyle, 1985; Warren
supporting a humoral autoimmune process in this syndrome, and Catz, 1986). Toh et al. (1985) have demonstrated
do not distinguish between the role of the antibody as a autoAbs against neurofilamental proteins from patients
marker for the disease or a pathogenetically important with the subacute spongiform encephalopathies of kuru
molecule. and Creutzfeldt – Jakob disease. Also noted was similar
Our studies began in an open exploratory fashion. staining by some Alzheimer patients and some normal
Indirect immunofluorescence as performed on freshly individuals (Bahmanyar et al., 1983). Gajdusek (1985)
obtained mouse brain was an exploratory and screening hypothesized that autoAbs interfering with axonal transport
study. This preceded the use of human brain tissue. Com- could be a common pathogenetic mechanism in certain
parison to controls including normal individuals when CNS diseases.
employing the Western blot is especially important because Whether these antibrain antibodies play any role in the
even if autoAbs are established as etiologically important in pathogenesis of HAD remains to be determined. Conceiv-
the brain their origin would still be unknown. Questions that ably, HIV could cause brain dysfunction either by cross
arise are: (1) does the antibrain autoAb arise de novo after reactivity with either normal or altered brain Ags or in fact a
the disease; is it present in undetectable amounts even in de novo response against an altered Ag. Conversely, the
normals, but produced by activated B cells after HIV has profound immunosuppression of AIDS, could unleash a
perturbed the immune system; and (2) is it a cross reactive normally suppressed autoAb.
Ab directed against epitopes of HIV, altered brain Ag, or In summary, we believe these results indicate that anti-
other altered tissue Ags such as those on lymphocytes? brain Abs can be demonstrated in patients with HAD and
These are unknown, but the preliminary evidence demon- may possibly be demonstrated in those who might have
strating Ab reactivity distinct from reactivity against HIV subclinical disease. This information may lead to new
Ags suggests this is not a simple cross reaction. directions in the investigation of the pathogenesis of
Precedent exists for an infectious agent or its components HAD. Certainly, the etiopathologic role could only be
triggering an autoimmune process that ultimately results in determined with future studies. The findings do indicate
neurologic symptoms. Johnson et al. (1984) have provided that further investigation in these area is warranted because
evidence that a cell-mediated immunologic reaction trig- of the potential to identify a marker of the syndrome at a
gered by measles is responsible for post-infectious measles point when some reversibility of the manifestations, and
encephalomyelitis. Neuroparalytic accidents following consequently, improvement of quality of life may be
administration of vaccines based on neural tissue have also achieved with therapeutic intervention (Cherner et al.,
been attributed to immune hyperreactivity. Hemachudha et 2002; Ernst et al., 2002).
S.E. Schutzer et al. / Journal of Neuroimmunology 138 (2003) 156–161 161