вход по аккаунту


Cerebral white matter changes in acquired immunodeficiency syndrome dementia Alterations of the blood-brain barrier.

код для вставкиСкачать
Cerebral Wlxte Matter Changes in Acquired
Immunodeficiency Syndrome Dementia:
Alterations of the Blood-Brain Barrier
Christopher Power, MD,* Pei-Ann Kong, BSc," Thomas 0. Crawford, MD,"?
Steven Wesselingh, BMBS, PhD," Jonathan D. Glass, MD,"f Justin C. McArthur, MBBS, MPH,"
and Bruce D. Trapp, PhD"
The cause of acquired immunodeficiency syndrome (AIDS) dementia, which is a frequent late manifestation of human
immunodeficiency virus (HIV) infection, is unknown but radiological and pathological studies have implicated alterations in subcortical white matter. To investigate the pathological basis of these white matter abnormalities, we
performed an immunocytochemical and histological analysis of subcortical white matter from AIDS patients with and
without dementia, from pre-AIDS patients (asymptomatic HIV-seropositive patients), and from HIV-seronegative
control subjects. Reduced intensity of Luxol fast blue staining, designated "diffuse myelin pallor," was detected in 8
of 15 AlDS dementia patients, 3 of 13 AIDS nondemented patients, and none of the pre-AIDS patients (n = 2) or
control subjects (n = 9). In contrast to Lwrol fast blue staining, sections stained immunocytochemically for myelin
proteins did not show decreased staining intensities in regions of diffuse myelin pallor. In addition, neither demyelinated axons nor active demyelination were detected in light and electron micrographs of subcortical white matter from
brains of patients with AIDS dementia. An increase in the number of perivascular macrophages and hypertrophy of
astrocytes and microglia occurred in brain sections from HIV-infected patients. These changes were not specific to
dementia or regions of diffuse myelin pallor and they occurred in both gray and white matter. In contrast to the lack
of myelin pathology in AIDS dementia brains, significant accumulations of serum proteins in white matter glia were
detected in the brains of 12 of 12 patients with AIDS dementia and 6 of 12 AIDS patients without dementia. Serum
protein-immunopositive cortical neurons were detected in the frontal cortex of 11 of 12 patients with AIDS dementia
and 3 of 12 nondemented AIDS patients. Seronegative control subjects showed minimal serum protein immunoreactivity in both cortex and white matter. We conclude therefore that alterations in the blood-brain barrier and not demyelination contribute to the development of AlDS dementia.
Power C, Kong P-A, Crawford TO, Wesselingh S, Glass JD, McArthur JC, Trapp BD.
Cerebral white matter changes in acquired immunodeficiency syndrome dementia:
alterations of the blood-brain barrier. Ann Neurol 1993;34.139-350
Acquired immunodeficiency syndrome (AIDS) dementia is a subcortical dementia characterized by psychomotor slowing, memory impairment, behavioral
dysfunction { 1, 21, and subcortical white matter abnormalities [2-41. Cranial magnetic resonance imaging
(MRI) scans of patients with AIDS dementia show
hyperintensities of subcortical white matter without
contrast enhancement [2). Luxol fast blue (LFB) staining of subcortical white matter from the majority of
AIDS dementia brains shows regions of diffuse myelin
pallor (DMP) with preserved staining of subcortical U
fibers 131. Other subcortical white matter changes associated with human immunodeficiency virus (HIV) infection include astrogliosis, mdtinucleated giant cells,
microglial nodules, increased numbers of perivascular
macrophages [?I, and a possible loss of oligodendro-
cytes 161. Damaged dendrites, reduced numbers of
cortical neurons, and a loss of synapses have also been
documented in AIDS brains [7, 81. Neuronal loss and
altered neuronal morphology have been attributed to
the neurotoxic effects of glycoprotein 120 (gp120), cytokines, and quinolinic acid 19-11).
DMP is a common neuropathological change that is
associated with the clinical expression of AIDS dementia [3]. The DMP associated with AIDS dementia has
been assumed to be secondary to demyelination or
some other myelin pathology [12]; however, extensive
demyelination in brains of patients with AIDS dementia has not been reported. Focal regions of demyelination in the brains of patients with AIDS have been
described, although these findings have not been associated with AIDS dementia [ S , 121.
From the Departments of "Neurology, tPediauics, and $Pathology,
The Johns Hopkins University School of Medicine, Baltimore, MD.
Address correspondence to Dr Trapp, Johns Hopkins University
School of Medicine, Pathology 627C, 600 N. Wolfe Street, Baltimore' IMD 21287-6965'
Received Jan 11, 1993, and in revised form Apr 19. Accepted for
publication Apr 21, 1993.
Copyright Q 1993 by the American Neurological Association
Table 1 . Summary of Clinical, Radiological, and Histopatholoxical Features
Age (yr)
Control (n = 7)
Pre-AIDS (n = 2)
Nondemented AIDS (n = 13)
AIDS dementia (n = 15)
O I N D (n = 7 )
46 -+ 17
31 ? 5
40 2 12
41 r 12
37 ? 10
Mean T 4 Count
Time (hr)
767 2 796
108 5 160
37 % 52
15 ?
18 k
13 %
11 -+
14 2 7
of DMP
"Four patients had had MRl scans.
bFourteen parients had had MR1 scans.
diffuse myelin pallor; ND = not done; pre-AIDS
asymptomatic HIV-seropositive cases; OIND = ocher inflarnrnacory neurological
An alternative explanation for DMP and MRI abnormalities is an alteration of the blood-brain barrier
(BBB). Conditions associated with BBB perturbation,
such as metastatic tumors, can show increased signal
in white matter on MRI scans and myelin pallor on
histological sections [ 13, 14}. Previous reports described BBB alterations in HIV-infected individuals.
Increased cerebrospinal fluid (CSF)-serum albumin ratios, an indication of BBB damage in neurological diseases, occur during HIV infection and become abnormally high in patients with AIDS dementia [15, 163.
Additional evidence of BBB damage in HlV infection
comes from immunocytochemical studies that have detected serum proteins in the brains of AIDS patients.
Rhodes [17] reported that most AIDS patients with
central nervous system (CNS) lesions had serum protein immunoreactivity in some neurons and glia. Petito
and Cash 1181 examined the extravasation of serum
proteins into the brains of HIV-infected individuals
with or without HIV encephalitis and into the brains
of HIV-seronegative control subjects. Compared to
control brains, brains from HIV-infected individuals
had a significantly higher incidence of serum protein
immunoreactivity; however, no differences were found
between those individuals with and those without HIV
encephalitis. Fifty percent of all AIDS brains had a
diffuse BBB leakage that occurred in the absence of
other recognized pathologies, including HIV encephalitis. Based on these observations, the possibility was
raised that circulating cytokines induce alterations in
the BBB and that alterations in the BBB contribute to
viral entry into the brain, to the development of DMP,
and to the gliosis that is common to all patients with
AIDS. Additional evidence for alterations in the BBB
comes from a histological analysis of three AIDS dementia brains {19}that had hyperdensities of the centrum semiovale on computed tomography (CT) scans
and DMP of subcortical white matter. Striking changes
in the microcirculation in subcortical white matter were
found, and included mural thickening, increased cellu-
340 Annals of Neurology
Vol 34
larity, enlargement and pleomorphism of endothelial
cells, and increased perivascular macrophages that often contained hemosiderin pigment. The relationship
between alterations in the BBB and the clinical status
of HIV-infected patients was not investigated in these
studies {17--19].
The objective of this study was to determine
whether white matter abnormalities observed in patients with AIDS dementia were related to demyelination or to an alteration in the RBB. Analysis of tissue
from the frontal lobes by histological, irnmunocytochemical, and ultrastructural techniques did not detect
abnormalities of myelin in regions of DMP and MRI
hyperintensity. In contrast, marked accumulations of
serum proteins were found in the brains of patients
with AIDS dementia. This study raises the possibility,
therefore, that the dementia associated with HIV infection is caused by alterations in the BBB.
Materials and Methods
HlV-infected brains were selected from the AIDS Brain
Bank at Johns Hopkins Hospital, based on clinical status and
pathological findings. The AIDS Brain Bank contains frozen
and fixed brains from over 300 autopsied patients, many of
whom were followed clinically by the AIDS Neurology
Group. HIV-infected brains were chosen for this study based
on the premortem clinical status. HIV-infected patients were
defined as pre-AIDS (asymptomatic HIV-seropositive), nondemented AIDS, or demented AIDS. Diagnostic criteria for
AIDS dementia were (1) HIV-1 seropositivity, (2) history of
progressive cognitive and behavioral decline, (3) neurological
and/or neuropsychological hndings consistent with a decline
from the premorbid baseiine, and ( 4 )opportunistic processes
in the CNS excluded by C T or MRI and CSF analysis. Tissue
from control subjects and from patients with other inflammatory neurological diseases was obtained from the Johns
Hopkins Hospital Autopsy Service and the Maryland Medical Examiner's Office in Baltimore.
Thirty HIV-infected patients and 16 uninfected patients
were studied (Table 1).Twenty-eight HIV-infected patients
had AIDS and 2 were designated as pre-AIDS. Fifteen of
No 3 September 1993
the AIDS patients had AIDS dementia based o n the above
criteria. Control brains (n = 9) were obtained from patients
who died from cardiac failure (n = 2), leukemia (n = l),
liver failure (n = l ) , Hodgkin’s disease (n = l), trauma (n
= l), gunshot wound (n = 2), or myocardial infarction in =
1). Brains from patients with other inflammatory neurological
diseases were also studied. These included herpes simplex
encephalitis (n = 3), multiple sclerosis (n = l), Neisserzu
meningitis (n = l),encephalopathy associated with sepsis (n
= 11, and lupus erythematosus (n = 1). Mean ages at the
time of death for each group were similar and are listed in
Table 1. Fourteen of the AIDS dementia patients and 4 of
the nondemented AIDS patients had undergone cranial
MRI. Increased signal in the subcortical white matter in both
hemispheres on T2-weighted images was found in 9 of 14
AIDS dementia patients and in none of the 4 nondemented
AIDS patients. Nondemented AIDS patients can have increased T2-weighted signals in subcortical white matter. The
lack of hyperintensities in the present study is due to the
small number of nondemented patients scanned. All but 2
of the nondemented AIDS patients had been evaluated neurologically with cognitive status testing within 30 days of
death. The other 2 patients had been assessed clinically
within the same time frame and were not found to have any
neurological signs or symptoms.
Tissue Samples
Brain tissue was sampled from the right frontal lobe. Five
adjacent segments (I-V) were taken: Segments I, 111, IV,
and V were 3 x 3 x 0.5 cm; segment I1 was 3 x 3 x 0.2
cm. Segment I was fixed in 4% paraformaldehyde, embedded in paraffin, and used for histological and immunocytochemical staining. Segment I1 was fixed in 470 glutaraldehyde
and embedded in epoxy resin (Epon). Segment 111 was fixed
in 4F paraformaldehyde, cryoprotected, and sectioned into
40-pm-thick, free-floating sections and used for immunocytochemisrry. Segments IV and V were frozen and stored at
- 70°C. The remaining brain tissue was fixed in 10% formalin.
Paraffin-embedded sections ( 20-pm thick) were stained immunocytochemically by the avidin-biotin complex (ABC)
procedure with 0.5 M Tris buffer (pH 8.0). Sections were
depardffinized and incubated in the following solutions: 3%
normal goat serum containing 0.5:;f powdered skim milk
for 1 hour at 22T,primary antibody for 18 hours at 4”C,
appropriate biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA) diluted 1:100 for 30 minutes
at 22”C, enzyme (HRPblinked biotin (1: 100) and avidin
(1: 100) in Tris buffer for 30 minutes, diaminobenzidineihydrogen peroxide for 8 minutes, and 0.4% osmium tetroxide
for 2 minutes.
In selected cases, tissue from segment 111 was cryoprotected in 20%) glycerol and 0.08 M Sorensen’s phosphate
buffer. These samples were embedded in 30%; sucrose on a
dry-ice platform, sectioned on a sliding microtome (40-pm
thick), and immunostained by the ABC technique. Prior to
immunostaining, sections were permeabilized with 10% Triton X-100 (Rohm and Haas) for 30 minutes.
The primary antibodies used are well characterized and
include rabbit anti-glial fibrillary acidic protein (GFAP;
Dako, Carpenteria, CA), rabbit anti-myelin basic protein
(MBP; Dako), rabbit anti-myelin-associated glycoprotein
(MAG {20}), mouse anti-HLA-DR (HLA-DR; Dako),
mouse anti-CD45 (leukocyte common antigen {LCAJ;
Dako), mouse antimacrophage (EBM-11; Dako), goat antihuman albumin (Antibodies, Inc, Davis, CA), and rabbit
anti-human haptoglobin (Binding Site, San Diego, CA). To
confirm the specificity of binding of the haptoglobin and albumin antibodies, antisera were incubated with the corresponding peptide for 2 hours at 22°C before they were applied to tissue sections.
Electron Microscopy
Segment I1 was fixed in 4% glutaraldehyde, postfixed in 4%
osmium tetroxide, and embedded in epoxy resin by standard
procedures. One-micron-thick sections were stained with toluidine blue and examined with a Zeiss Axiophot light microscope. Ultrathin sections were stained with uranyl acetate
and lead citrate and examined with a Hitachi H-600 electron
Paraffin-embedded sections from each brain were immunostained with haptoglobin antibodies and analyzed by a
blinded operator with computerized gray-scale image analysis
software (Bioquant, R & M Biometrics, Memphis, TN). Section thickness and DAB incubation times were identical for
all specimens. The number of haptoglobin-positive cells in
subcortical white matter was quantitated in ten nonadjacent
fields from regions distinguished by relative uniformity of
staining and the absence of arrifacts. Optimal threshold values
were selected for each image based o n the relative intensity
of cell and background staining. Cells with a cross-sectional
areagreater than 40 ym’ were counted per square millimeter
of tissue, by using a 20 x objective.
A paraffin-embedded section from segment I for each
brain was stained with hematoxylin-eosin and analyzed
histologically. Multinucleated giant cells were found in
1 of 15 of the A I D S dementia, 1 of 13 of the nondemented A I D S , and none of the pre-AIDS and seronegative control sections from segment I. Microglial nodules were identified i n 4 of 15 AIDS dementia, 3 of
13 nondemented AIDS, and none of t h e pre-AIDS
and HIV-seronegative control sections. Opportunistic
infections were not detected in the tissue segments
used in this study; they were, however, found i n other
brain regions (Table 2) and occurred with similar frequency in AIDS dementia (40%) and nondemented
AIDS brains (31%,) but were not seen in pre-AIDS
brains. Based o n repeated clinical and radiological evaluations, the opportunistic infections in the AIDS patients with dementia did not contribute to t h e symptoms and signs of AIDS dementia.
Power et al: BBB in AIDS Dementia
Table 2. Summar7,of Opportuizistic InfectiunJ and Their
Location in AIDS Patients
Patient No.
Cingulate gyrus, globus
pallidus, hippocampus
Toxoplasmosis Occipital lobe, thalamus
Spinal cord
Parietal lobe, caudate
Cerebellum, thalamus
Toxoplasmosis Thalamus
Parietal and occipital
Internal capsule
= cytomegalovirus; PML = progressive multifocal leukoencephalopathy.
In paraffin-embedded sections from control brains,
LFB stained all regions of subcortical white matter with
equal intensity (Fig 1A). In contrast, cortical gray matter showed little detectable LFB staining. Sections cut
adjacent to the LFB-stained control sections were immunostained with antibodies directed against the myelin proteins MBP and MAG. Similar to LFB-stained
sections, myelin protein antibodies (Fig 1B) stained all
regions of subcortical white matter with similar intensity.
In sections from 3 of 13 nondemented AIDS brains
and from 8 of 15 AIDS dementia brains, the intensity
of LFB staining was reduced in subcortical white matter
(Fig 1C). Figure 1C illustrates an example of more severe DMP. DMP was confluent; it was not enhanced
in the vicinity of blood vessels but it often spared subcortical U fibers. Small focal areas of decreased LFB
staining that are common in diseases like multiple sclerosis and progressive multifocal leukoencephalopathy
were not prominent in HIV-infected brains. Sections
cut adjacent to those revealing DMP with IPB staining
were immunostained with MBP and MAG antisera.
Unlike the variable LFB staining pattern, myelin protein antibodies stained all regions of subcortical white
matter with similar intensity (Fig 1D). White matter
stained uniformly when MBP and MAG antibodies
were applied at dilutions ranging from 1:100 to
1:5,000. Enhancement in myelin protein immunoreactivity, as observed in acute demyelination {21), was
not seen in areas of DMP, nor was myelin protein
immunoreactivity detected within parenchymal or perivascular macrophages. Silver staining of paraffinembedded sections from AIDS patients did not demonstrate significant axonal injury or altered axonal
morphology in regions of DMP.
Epoxy resin-embedded sections from segment I1 of
control (n = 4), nondemented AIDS (n = 4), and
AIDS dementia (n = 4 ) brains were analyzed by light
and electron microscopy. Demyelinated axons were
not seen in tissue from control or HIV-infected brains
(data not shown). Inflammatory infiltrates and macrophages containing myelin debris were rare in both control and HIV-infected sections. Although analysis of
subtle ultrastructural changes within the myelin sheath
was not possible because of the limits of fixation, analysis of epoxy resin-embedded sections provided no evidence of active demyelination in HIV-infected brains.
GFAP immunostaining of subcortical white matter revealed increased astrogliosis in tissue from demented
and nondemented AIDS brains (Fig 2C) compared to
control (Fig 2A) and pre-AIDS brains (Fig 2B). GFAPimmunoreactive nucleated cells were counted in five
fields of subcortical white matter (magnification x
250) by an examiner unaware of the case identities.
There was a statistically significant increase in the mean
number of GFAP-positive cells detected in tissue from
the two AIDS groups (28.1 -+ 4.98) compared to the
control group (20.1 k 7.5) ( p = 0.04, Mann-Whitney
two-tailed test). The increased numbers of immunodetectable astrocytes were accompanied by hypertrophy
of the astrocyte cell body and elongation of cell processes in the AIDS groups compared to the control
group. There were no statistically significant differences in the numbers of astrocytes in brains stratified
for the presence of DMP or dementia.
Mucrophuges and Microglia
If active demyelination or wallerian degeneration were
occurring in HIV-infected brains, phagocytic macrophages should be present. To test this hypothesis,
40-~m-thick,free-floating sections from control, preAIDS, nondemented AIDS, and AIDS dementia
brains were immunostained with antibodies directed
against the microglial and/or macrophage proteins,
LCA, EBM-11, and HLA-DR. Microglial hypertrophy
and increased expression of LCA, EBM-I 1, and HLADR were observed in HIV-infected brains. The intensity and patterns of immunoreactivity, however, did
not differ between regions with and without DMP
or between the nondemented and demented AIDS
LCA antibody is directed against CD45, a cell-surface antigen found on
cells of lymphocytic and monocytic lineage. In sections
from control brains, LCA stained microglia and perivascular macrophages throughout the gray and white
matter (Fig 3A). Small LCA-positive microglial cell
bodies extended numerous LCA-positive processes
342 Annals of Neurology Vol 34 Nu 3 September 1993
Fig 1. Comparison of Laxol fast blue (LFB)staining (A, Cj
with myelin basic protein (MBP) (B, 0)immunoreactivity in
10-pm-thick, parafin-embedded sections of frontal cortex-from
control (A, B ) and AIDS dementia (C, Di brains. LFB and
M B P show similar staining patterns in control brains. A
marked decrease in LPB staining of subcortical white matter in
AIDS dementia bruins (C, is not rejected by a loss of MBP immunoreactivity (0).
Scale bar = 2 mm.
Power et al: BBB in AIDS Dementia
Fig 2. Glialjbrillavy acidic protein inmunostaining of 40pm-thick, free-floating sections of subcortical white matter from
control (A),pre-AIDS (Bj. and AIDS (C) brains. Astrocytic bypertrophy increases during HIV infection but is not spec& for
AIDS dementia o r regions of dqfase myelin pallor. Scale bar =
160 p m .
that rarely overlapped with processes of adjacent microglia. The LCA-positive microglia in the demented
and nondemented AIDS brains were larger than those
in control brains and extended thicker processes that
often overlapped (Fig 3C). Sections from the 2 preAIDS brains showed LCA-positive microglia (Fig 3B)
with larger perikarya and thicker processes than control microglia, but these changes were less apparent
than in AIDS brains.
EBM-11. The EBM-11 antibody reacts with a lysosomal protein (CD68) found in cells of monocytic lineage, especially phagocytic macrophages. In sections
from control brains, EBM-11 produced weak immunostaining of perivascular cells (Fig 3D) and a weak particulate staining throughout the neuropil that had a distribution similar to that of the LCA-positive microglial
perikarya. This staining represents a low basal level
of lysosomal activity in perivascular macrophages and
microglia in normal brain. The EBM-11 antibody
stained perivascular macrophages and microglia with
greater intensity in sections from the AIDS brains (Fig
3F) when compared to sections from control brains
(Fig 3D). EBM-1 1-positive perivascular macrophages
extended thick processes along the abluminal aspect of
blood vessels. The pre-AIDS sections contained increased EBM-11 immunoreactivity (Fig 3E) compared
to control sections, but again these changes were less
prominent than those observed in the AIDS groups.
344 Annals of Neurology Vol 34
HLA-DR. The HLA-DR antibody reacts with major
histocompatibility complex (MHC) class I1 molecules.
Expression of MHC class I1 can be induced in a variety
of cells and can play a role in antigen presentation. The
HLA-DR antibody produced little detectable immunoreactivity in sections from control brains (Fig 3G). In
sections from the AIDS brains, perivascular macrophages (Fig 31, arrows) and parenchymal microglial
cells (Fig 31, arrowheads) were HLA-DR-positive.
Sections from one of the AIDS dementia brains
showed round, hypertrophied, HLA-DR-positive microglia with extremely short, thickened processes (data
not shown); none of the pre-AIDS or nondemented
AIDS brains, however, displayed this finding. HLADR-stained sections from the pre-AIDS brains (Fig
3H) had changes in microglia and perivascular macrophages that were similar to but less striking than those
of the AIDS groups.
In summary, the intensity or pattern of microglial
and macrophage immunoreactivity was similar in AIDS
dementia and nondemented AIDS groups and did not
differ in regions of DMP. The AIDS groups, however,
did show a “generalized activation” of microglial and
perivascular macrophages within white and gray matter
when compared to the control group. The large, round,
debris-filled macrophages that are prominent in regions
of active demyelination {2 1) were not observed in subcortical white matter from HIV-infected brains.
Blood-Brain Bawier
To assess BBB permeability, sections from control,
nondemented AIDS, and AIDS dementia brains were
stained with antisera directed against haptoglobin and
albumin (Fig 4), two serum proteins rarely found
in normal brain parenchyma [22]. The albumin and
3 September 1993
F i g 3. Forty-mz~ron-thick,free-flouting sectionj of subcortical
uthite matter from control (A, D , G), pre-AIDS iB. E , H I . and
AIDS (C F , I ) bruins Leukocyte common dntzgen (LCA) zmmunostaining (A-C) shous hypertrophied resident m i rogltu in
the AIDS b r a m (C) compared to the control (A)and pre-AIDS
(B) brutnr. EBM-1 I ID-FI rtuznzng produced prominent perzzascukzr macrophuge rtuinmg (arrows) and leu intense particu-
late ~tuinzngof resident microgliu (arrowheads) In the AIDS
(I;) und pre-AIDS (E) bruins, but produced lzttle Jtuining in
the control (0)brarnr HLA-DR (G-I) stuzned perzva~cular
mcrophugeJ (arrow) and microglta (arrowheads) 2n the AIDS
( I ) and pre-AIDS (H) bruins, wzth mznzmul staining t n the
control iC) brains Scale bur = 160 p n z
Power et al: BBB in AIDS Dementia
Fig 4. Haptoglobin immunostaining of 40-pm-thick, free-$oating sections of subcortical white matter (A-C) and cortical gray
matter (0-1) from control (A, DI, AIDS dementia (B. C.
F-I), and nondemented AIDS (EI brains. Control brains shw
predominantly intravusrular staining (A, arrowhead). AIDS
cGementaa brains with dqfuse myelin pallor show prominent
glial (B, arrowheads) and dqfuse neuropil staining. Prior immunoadsorption of the haptoglobin antibody eliminates immuno-
staining o/mtions from AIDS dementia brains (GI. Serum protein immunoreactivity in graji mutter was not prominent in
control brains (Di while neumpil (E-l), glial ( E and F , arrowheads), and neuronal @-I) staining wasfound in AIDS
brains. See text for frequencies. Neurons contained cytoplusmic
(G-I) and in some instances nuclear (I, arrowhead) staining.
A-C: scale bar = 160 pm; D-P: scale bar = 80 pm: G I :
scale bar = 20 pm.
346 Anrials of Neurology Vol 34 No 3 September 1993
haptoglobin immunostaining patterns were similar, and
therefore are described and discussed as one pattern.
In brain sections from the AIDS groups, two striking
patterns of serum protein immunostaining were observed. First, all brains with DMP and 12 of 12 brains
from patients with AIDS dementia showed extensive
serum protein immunoreactivity in subcortical white
matter; and second, 11 of 12 brains from patients with
AIDS dementia showed serum protein-immunopositive neuronal perikarya.
In sections from control brains, serum
protein immunoreactivity was usually restricted to intravascular and perivascular regions (Fig 4A). Perivascular leakage of serum proteins is a recognized postmortem phenomenon in control brains {23]. Twelve
of 12 AIDS dementia and 6 of 12 nondemented AIDS
brains contained serum protein-positive glial cells in
subcortical white matter (Fig 4B). In addition to cellular staining, diffuse extracellular staining was also present and was uniform throughout the sections without
the perivascular enhancement seen in control sections.
All haptoglobin staining in sections from AIDS dementia brains was eliminated by prior absorption of the
haptoglobin antibody with haptoglobin peptide (Fig
4C). Sections from nondemented AIDS brains lacking
serum protein-positive immunostaining of white matter contained intravascular and some perivascular staining as seen in control sections.
Morphometric analysis revealed that the mean numbers of serum protein-immunoreactive glia for each
group were as follows: controls, 3.3 0.86 ceUs/mm’;
nondemented AIDS, 46 5 7.5 cells/mm2; AIDS dementia, 86 2 12.7 cellsimm’; and other inflammatory
neurological disease, 156 t 33.9 cells/mm2 (ANOVA,
p = 0.0001). Post hoc comparisons with Bonferroni
corrections showed statistical differences between the
control and AIDS dementia ( p = 0.01) groups, and
the control group and the group with other inflammatory neurological diseases ( p = 0.004). The coincidence of DMP with MRI white matter hyperintensities
was more frequent in sections with glial staining (x’ =
13.1, p = 0.005). The frequency of DMP, MRI white
matter hyperintensities, and increased GFAP immunoreactivity was significantly associated with the clinical
expression of dementia (x’ = 18.4, p = 0.002).
In 7 of 8 control brains, serum protein immunoreactivity was restricted to vascular and perivascular regions (see Fig 3D). The other
control brain contained some parenchymal serum protein staining. Four of 12 nondemented AIDS brains
showed diffuse protein staining of neuropil and 3 of
these contained serum protein-positive cells. Most
of the serum protein-positive cells in nondemented
AIDS sections were glia (see Fig 4E).
N = 12
N = 12
N = 7
Fig 5 . Fregzlencies of dt#use myelin pallor, sevum protein-positiiie subcortical gliu, and .serum protein-po.titive cortical ne14rom in control, notidemented AIDS (HIVND),A1DS dementia
(HIVD), and other it$ammatovy neurological diseuse (OIND)
Eleven of 12 AIDS dementia brains contained serum
protein immunoreactivity in the frontal cortex. Serum
proteins were distributed diffusely throughout the neurophil and located in neuronal and glial perikarya (see
Fig 4F). Serum protein-positive neurons in AIDS dementia sections were randomly distributed in all layers
of the cortex and ranged in size from large pyramidal
neurons to small interneurons. Serum protein-positive
neurons were often located next to serum-negative
neurons (see Figs 4F, 41), indicating a selective neuronal vulnerability to serum protein uptake. The distribution of serum proteins in neurons displayed two general patterns; the most common was diffuse perinuclear
cytoplasmic staining that extended into processes (see
Fig 4G).A less common pattern consisted of both cytoplasmic and nuclear serum protein staining (see Figs
4 H , 41). Neurons with serum protein-positive nuclei
generally showed fewer and shorter cellular processes
(see Fig 41).
Neurons containing serum protein were observed
in 5 of 7 brains with other inflammatory neurological
diseases and intense neuronal nuclear staining was seen
in 4. The frequency of serum protein-stained neurons
was significantly higher in the AIDS dementia group
than in the combined nondemented AIDS and control
groups (Fisher’s exact test, p = 0.0001) but the frequencies of stained neurons did not differ significantly
between the AIDS dementia and other inflammatory
neurological disease groups.
Figure 5 summarizes the significant immunocytochemical findings described above. Compared to control and nondemented AIDS brains, the frequency
(based on Fisher’s exact test) of DMP ( p = 0.013),
serum protein-positive glia ( p = 0.0001), and serum
protein-positive neurons ( p = 0.0001) was greater in
AIDS dementia brains. The frequency of DMP and
Power et al: BBB in AIDS Dementia
serum protein immunoreactivity in glia and neurons
did not differ between the other inflammatory neurological disease group and the AIDS dementia group.
AIDS dementia has been described phenomenologically as a subcortical dementia 111. Radiological and
pathological abnormalities in subcortical white matter
have been documented in the brains from AIDS patients with dementia. Demyelination 1123 and altered
BBB or altered microvasculature 117- 191 have been
proposed as possible mechanisms leading to the white
matter abnormalities in patients with AIDS dementia.
We have identified changes in subcortical white matter, indicating that the BBB in AIDS dementia is perturbed without evidence of widespread demyelination.
The brains of patients with AIDS dementia analyzed
in this study showed increased BBB permeability in
subcortical white matter and in cerebral cortex from
the frontal lobe. Serum protein immunoreactivity in
some of the brains of nondemented AIDS patients suggests that the abnormalities of the BBB in HIV infection are chronic and precede the onset of AIDS dementia. There was a statistically significant association
between DMP, MRI white matter hyperintensities, serum protein immunoreactivity in white matter glia, and
the clinical expression of AIDS dementia.
Increased CSE;/serurn albumin ratios have also been
used as an indication of alterations in the integrity of
the BBB [241. In support of the hypothesis that HIVinfected individuals have a chronic, slowly progressing
breach of the BBB, the CSFiserum albumin ratio increases with the durauon of HIV infection a i d becomes abnormally high in patients with AIDS dementia [l5, 16’1. Cryoinjury of rat cerebral cortex also leads
to increased permeability of the BBB and to serum
protein immunoreactivity in neurons 1251. Serum protein accumulation in neuronal nuclei after cryoinjury
closely correlates with neuronal loss 1251. A recent
immunocytochemical study of serum protein distribution following brain contusions concluded that the uptake of serum proteins by neurons was an indication of
neuronal injury that occurs before reliable histological
changes have developed 1261. Serum protein-positive
neurons with short or distorted dendritic arborizations,
serum protein-positive nuclei, or both, were common
in AIDS dementia brains and are likely to be dysfunctional or dying.
Several mechanisms could be responsible for altered
BBB function in HIV-infected individuals. HIV could
infect cerebral endothelial cells and alter thex barrier
properties; however, HIV has not been detected in
vivo in cerebral endothelial cells. We consider it most
likely that systemic abnormalities such as generalized
immune activation and elevated serum cytokine levels
12.71 are
responsible for HIV-related BBB dysfunction. In experimental brain infection, interleukin-1
and tumor necrosis factor (TNF)-u appear to augment
BBB permeability [28-30} by increasing cerebral endothelial pinocytosis and by altering endothelial tight
junctions [28}. These changes occur as albumin and
leukocytes enter the CNS 128-301. Other serum factors, such as free radicals, could also contribute to BBB
damage by inducing lipid peroxidation of endothelial
plasma membranes 1311. Perivascular macrophages
may also alter BBB function; their numbers are increased in HIV-infected brains, they are often HIVpositive, and if activated, they can secrete cytokines or
other factors that damage endothelial cells.
Unlike the acute and rapid BBB breakdown associated with cerebrovascular accidents or multiple sclerosis, BBB changes in HIV-infected brains appear to be
chronic and slowly progressive. Acute breakdown of
the BBB does not induce dementia {32l, and to our
knowledge, there are no experimental animal models
that mimic the chronic BBB changes found in HIVinfected brain. We believe that increased BBB permeability is an initial or “priming” event that contributes
to CNS injury associated with AIDS dementia. Altered
BBB function could contribute to the pathogenesis of
AIDS dementia in several ways. A permeable BBB
could facilitate entry of HIV or H1V-infected cells and
increase the CNS viral burden. Once the BBB is compromised, serum-derived agents such as HIV gp120
[C,} or TNF-u [lo, 333 could be toxic to neurons and
glia. Serum factors could also activate microglia, astrocytes, or CNS macrophages and increase their production of cytokines [lo], quinolinic acid 1111, arachidonic
acid metabolites 134, 351, and other toxic factors 136,
371. Previous studies by our group [lo, 331 revealed
increased TNF-u immunoreactivity within microglia
and increased levels of TNF-a messenger RNA in
AIDS dementia brains that were analyzed in the present study. A more permeable BBB could also develop
a chronic CNS edema leading to gliosis { 131, an altered
ionic milieu, and altered function of excitable mernbranes within the CNS.
The frontal lobe was chosen for study because the
clinical features of AIDS dementia suggest that the
frontal lobe is affected and because marked neuronal
dropout in the frontal cortex of HIV-infected brains
has been reported [7,83. The intensity of diffuse extracellular serum protein immunoreactivity, especially apparent in AIDS dementia brains, was more prominent
on the free-floating than on the paraffin-embedded sections. Diffuse extracellular staining of serum proteins
is reduced on paraffin-embedded sections because the
solvents used during processing extract extracellular soluble proteins 1381. Brain sections from patients
with other neurological inflammatory diseases that are
known to induce alterations in the BBB showed neu-
348 Annals of Neurology Vol 34 No 3 September 1993
ronal serum protein-staining frequencies similar to
those found in the AIDS dementia group (see Fig 5 ) ,
thus supporting the postulate of increased cerebrovascular permeability in AIDS dementia. Age- and autopsy time-matched groups were used in this study to
avoid the pitfalls that may occur when serum protein
immunoreactivity is used as a measure of BBB permeability [23]. The incidence of systemic illnesses such
as multiorgan failure and sepsis was similar in AIDS
demented, AIDS nondemented, and HIV-seronegative groups. It is unlikely, therefore, that premortem
conditions induced BBB breakdown.
The astrogliosis and hypertrophy of microglial and
perivascular macrophages that have been reported previously in HIV-infected brains [39] were prominent
features of the HIV-infected brains analyzed in this
study, N~ obvious differences between the AIDS
groups based on immunocytochemical staining of astrocytes and microglia were found, indicating that the
of these cells alone does not correlate with
the hypertrophy and increased numbers of perivascular macrophages in AIDS
dementia sections as identified by EBM-11, a lysosomal marker, raise the possibility that perivascular
macrophages play some role in perturbing the BBB. In
support of this hypothesis, sections from HIV-infected
patients without hypertrophied perivascular macrophages contained less parenchymal serum protein immunoreactivit y.
Immunocytochemical studies did not detect myelin
damage in regions of subcortical DMP. The uniformity
of the myelin protein immunoreactivity abrogates the
idea that demyelination causes DMP, since myelin
protein immunoreactivity is enhanced during myelin
damage [2lJ Regions of DMP did not contain foamy
macrophages or demyelinated axons, two hallmarks of
chronic demyelinating lesions. Quantitative Western
blot analysis of tissue from the same patients described
in this report detected similar levels of MBP and proteolipid protein in the subcortical white matter of
AIDS demented, AIDS nondemented, and control
groups [40). Our failure to detect widespread demyelination in DMP does not imply that myelin pathology
does not occur in HIV infection, merely that frank
demyelination is not the cause of DMP. Although the
cause of decreased LFB staining intensity in HIVinfected brains is not known, we believe it is related
to alterations in the BBB. Factors contributing to DMP
may occur before or after death. The DMP associated
with metastatic brain tumors presumably arises from
sequestration of exudative fluid in myelinated tracts,
which provide the least resistance to fluid shifts [41).
This interpretation is supported by experimental studies of cerebral vasogenic edema showing augmented
extracellular spaces, separation of myelin larnellae, and
distention of periaxond spaces {42-44}.
The results of this study establish a strong correlation between alterations of the BBB and the presence
of AIDS dementia. Whether the perturbation of BBB
is related to circulating serum factors or the numerous
“activated” perivascular macrophages seen in HIVinfected brains remains to be determined. Identification of local or systemic factors, or both, that contribute to the sequestration of macrophages around
cerebral vessels, including the presence of endothelial
adhesion molecules and HIV viral proteins, will address this issue.
This work was supported by grants NS26643 (to B. D. T.), NS01577
(to J, D. G.), A1 72634 (to J. C. M.), and RR00722 from the National Institutes of Health. Dr Power is a recipient of a Medical
Research Council (MRC) fellowship (Canada). Dr Wesselingh is a
recipient of an N H and MRC fellowship (Australia). Dr Glass is a
recipient of a Lilly Clinician-ScientistAward from the Johns Hopkins
University School of Medicine,
The authors thank Mr R. Graham for manuscript preparation; Ms
B. Nsah and Mr P. Hauer for technical assistance; and Drs R. T.
Johnson, P. Talalay,J. W. Griffin, C. Pardo, and G. Kidd for helpful
1. Navia BA, Jordan BD, Price RW. The AIDS dementia complex:
1. Clinical features. Ann Neurol 1986;19:5 17-524
2. McArthur JC. Neurologic manifestations of AIDS. Medicine
1987;66:407-43 7
3. Navia BA, Cho E-S, Petito CK, Price RW. The AIDS dementia
complex. 11. Neuropathology. Ann Neurol 1986;19:525-535
4. Budka H. Human immunodeficiency virus (H1V)-induced disease of the central nervous system: pathology and implications
for pathogenesis. Acta Neuropathol (Berl) 1989;77:225-236
5. Budka H. Neuropathology of human immunodeficiency virus
infection. Brain Pathol 1991;1:163- 175
6. Esiri MM, Morris CS, Millard PR. Fate of oligodendrocytes in
HIV-1 infection. AIDS 1991;5:1081-1088
7. Everall IP, Luthbert PJ, Lantos PL. Neuronal loss in the frontal
cortex in HIV infection. Lancet 1991;337:1119-1121
8. Wiley CA, Masliah E, Morey M, et al. Neocortical damage during HIV infection. Ann Neurol 1991;29:651-657
9. Kaiser PK, Offermann JT, Lipton SA. Neuronal injury due to
HIV-1 envelope protein is blocked by anti-gpl20 antibodies
but not by anti-CD4 antibodies. Neurology 1990;40:17571761
10. Tyor WR, Glass JD, Becker PS, et al. Cytokine expression in
the brain during AIDS. Ann Neurol 1992;31:349-360
11. Heyes MP, Brew BJ, Martin A, et al. Quinolinic acid in cerebrospinal fluid and serum in HIV-1 infection: relationship to clinical
and neurological status. Ann Neurol 1991;29:202-209
12. Kleihues P, Lang W, Burger PC, et al. Progressive diffuse leukoencephalopathy in patients with acquired immune deficiency
syndrome (AIDS). Acta Neuropathol (Berl) 1985;68:333-339
13. Okazaki 0.Fundamentals of neuropathology: morphologic basis
of neurologic disorders. 2nd ed. New York: Igaku-Shoin, 1989
14. Atlas SW. Intracranial brain tumors. In: Atlas SW, ed. Magnetic
resonance imaging of the brain and spine. New York: Raven,
15. McArthur JC, Nance-Sproson TE, Griffin DE, et al. The diagnostic utility of elevation in cerebrospinal fluid beta,-microglobd i n in HIV-1 dementia. Neurology 1992;42:1707-1712
Power et al: BBB in AIDS Dementia
16. Marshall DW, Brey RL, Butzin CA, et al. CSF changes in a
longitudinal study of 124 neurologically normal HIV-1-infected
United States Air Force personnel. J Acquir Immune Defic
Syndr 1991;4:77 7-78 1
17. Rhodes RH. Evidence of serum-protein leakage across the
blood-brain barrier in the acquired immunodeficiency syndrome. J Neuropathol Exp Neurol 1991;50:171-183
18. Petito CK, Cash KS. Blood-brain barrier abnormalities in the
acquired immunodeficiency syndrome: immunohistochemical
localization of serum proteins in postmortem brain. Ann Neurol
19. Smith TW,DeGirolami U, Henin D, et al. Human immunodeficiency virus (HIV) leukoencephalopathy and the microcirculation. J Neuropathol Exp Neurol 1990;49:357-370
20. Quarks RH, Johnson D, Brady RO, Sternberger NH. Preparation and characterization of antisera to the myelin-associated
glycoprotein. Neurochem Res 1981;6:1109-1121
21. PrineasJW, KwonEE, Sternberger NH. Lennon VA. The distribution of myelin-associated glycoprotein and myelin basic protein in actively demyelinating multiple sclerosis lesions. J Neurnimmunol 1984;6:25 1-264
22. L u HM, Atack JR, Rapoport S1. lmmunohistochemical localization of intracellular plasma proteins in the human central nervous system. Acta Neuropathol (Berl) 1989;78:16-2 1
23. Mori S, Sternberger N H , Herman MM, Sternberger LA. Leakage and neuronal uptake of serum protein in aged and Alzheimer brains: a postmortem phenomenon with antemortem etiology. Lab Invest 1991;64:345-35 1
24. Adams RD, Victor M. Principles of neurology. 3rd ed. New
York: McGraw-Hill, 1985:14
25. Loberg EM, Torvik A. Uptake of plasma proteins into damaged
neurons: an experimental study on cryogenic lesions in rats.
Acta Neuropathol (Berl) 1991;81:479-485
26. Loberg EM, Torvik A. Neuronal uptake of plasma proteins in
brain contusions: an immunohistochemical study. Acta Neuropathol (Berl) 1992;84:234-2 37
27. Rosenberg ZF, Fauci AS. Immunopathogenir mechanisms of
HIV infection: cytokine induction of HIV expression. Immunol
Today 1990;ll:176-180
28. Quagliarello VJ, Wispelway B, Long WJ Jr, Scheld WM. Recombinant human interleukin-1 induces meningitis and blood-brain
barrier injury in the rat. J Clin Invest 1991;87:1360-1366
29. Saukkonen K, Sande S, Cioffe C, et al. The role of cytokines
in the generation of inflammation and tissue damage in experimental gram-positive meningitis. J Exp Med 1990;17 1:439-448
30. Ramilo 0,
Saez-Llorens X, Mertsola J, et al. Tumor necrosis
factor-alphaicachectin and interleukin 1-beta iuitiate meningeal
inflammation. J Exp Med 1990;172:497-507
31. Povlktock JT, Kontos HA. The pathophysiology of pial and
intraparenchymal vasculature dysfunction. In: Grossman G,
Gildenberg PL, eds. Head injury: basic and clinical apects. New
York: Raven, 1980:15-29
32. Rapoport SI, Thompson HK. Osmotic opening of the bloodbrain barrier in the monkey without associated neurological
deficits. Science 1973;180:971
33. Wesselingh SL, Power C , Glass JD, et al. Intracerebral cytolune
mRNA expression in AIDS. Ann Neurol (in press)
34. Genis P, Jett M, Bernton EW, et al. Cytokines and arachidonic
metabolites produced during human immunodeficiency virus
(H1V)-infected macrophage-astroglia interactions: implications
for the neuropathogenesis of HIV disease. J Exp Med 1992;
176:1703-17 18
35. Wahl LM, Corcoran ML, Pyle SW, et al. Human immunodeficiency virus glycoprotein (gp120) induction of monocyte arachidonic acid metabolites and interleukin 1. Immunology 1989;
36. G i d a n D, Vaca K, Noonan CA. Secretion of neurotoxins by
mononuclear phagocptes infected with HIV-1. Science 1990;
250: 1593-1 596
37. Giulian D, Vaca K, Corpuz M. Brain glia release f'actors with
opposing actions upon neuronal survival. J Neurosci 1993;
13129-3 7
38. Itoyama Y, Sternberger N H , Kies M, et al. lmmunocytochemical merhod to identify myelin basic protein in oligodendroglia
and myelin sheaths of the human nervous system. Ann Neurol
39. Ciardi A, Sinclair E, Scaravilli F, et al. The involvement of the
cerebral cortex in human immunodeficiency virus encephalopathy: a morphological and immunohistochemicd study. Acta
Neuropathol (Berl) 1990;81:51-59
40. Moeller JR, Durr PG, Quarks RH, et al. Analysis of myelin
proteins in postmortem AIDS brain. Soc Neurosci Abst
1992;18:1591 (Abstract)
41. Marmarou A, Takazi H, Shulman K. Biomechanics of brain
edema and effects on local cerebral flow. Adv Neurol 1980;
42. Powell HC, Myers RR, Lampert PW. Edema in neurotoxic injury. In: Spencer PS, Schaumburg HH, eds. Experimental and
clinical neurotoxicology. Baltimore: Williams & Wilkins, 1980:
43. Brucher JM. Leukoencephalopathies in anoxic-ischemic processes. In: Vinken PJ, Bruyn GW, Klawans HL, Koetsier JC,
eds. Demyelinating diseases. Handbook of clinical neurology,
vol 47.Amsterdam: Elsevier, 1985:525-549
44. Hirano A. Fine structure of edematous encephalopachy. Adv
Neural 1980;28:83-97
350 Annals of Neurology Vol 34 N o 3 September 1993
Без категории
Размер файла
1 320 Кб
immunodeficiency, matter, change, acquire, syndrome, dementia, white, brain, barriers, alteration, cerebral, blood
Пожаловаться на содержимое документа