вход по аккаунту


Diffuse microgliosis associated with cerebral atrophy in the acquired immunodeficiency syndrome.

код для вставкиСкачать
Ddhse Microghosis Associated
with Cerebral Atrophy in the Acquired
Immunodeficiency Syndrome
Benjamin B. Gelman, MD, PhD
The cause of cerebral atrophy in patients with acquired immunodeficiency syndrome (AIDS) is obscure because human
immunodeficiency virus type 1 (HIV-1)-related histopathological changes hardly correlate with cerebral atrophy. In
this study, brain ventricular expansion was compared to the frontal lobe density of mononuclear and astroglial cells
at autopsy. Twenty-eight male patients with AIDS displaying varying degrees of atrophy were compared to 17
age-matched male control subjects without AIDS or atrophy. An index of ventricular expansion was measured in
uniformly sliced, formalin-fixed brain specimens, and immunochemically marked cells in coronal sections of the left
superior frontal gyrus (Brodmann area 8) were quantified by field counting and planimetry. In the cortex, diffuse
ferritin-stained microglia and glial fibrillary acidic protein-positive astrocytes were about twice as numerous in the
patients with AIDS. Sixty-five percent (18/28) of the patients with AIDS had a microglial cell density greater than 2
standard deviations above the control mean. Microglial cell density was correlated positively with the severity of
ventricular expansion ( r = 0.71,p < 0.0001),while hypertrophied astroglial cells were very weakly related. In white
matter, Ham-56-positive macrophages and glial fibrillary acidic protein-positive astrocytes were not meaningfully
correlated with the index of ventricular expansion. Brain ventricular expansion and diffuse cortical microgliosis are
highly prevalent anomalies in patients with AIDS, and their interrelationship may be more important than previously
Gelman BB. Diffuse microgliosis associated wtih cerebral atrophy in the
acquired immunodeficiency syndrome. Ann Neurol 1993;34:65-70
lar expansion and increased numbers of stained microglial cells in the cerebral cortex.
Cerebral atrophy is the most common abnormality observed in brain images of patients with the acquired
immunodeficiency syndrome (AIDS). Atrophy may be
less evident at autopsy, and the neuropathological diagnosis tends to be rendered more conservatively [ 1, 21.
The clinical significance of the lesion in brain images
is unclear because the mechanism has not been elucidated. Defining the cellular basis of brain atrophy will
lead to a more lucid interpretation of brain images, and
will validate pathologically their use in the research
of AIDS-associated neurodegeneration. Previously, we
showed that each of the main microscopic abnormalities in the central nervous system (CNS) of patients
with AIDS correlate weakly or not at all with cerebral
atrophy [l}.The objective of this study was to quantify
more subtle cellular abnormalities to determine if they
are related to brain ventricular expansion. Ventricular
expansion was estimated planimetrically in fixed brain
specimens at autopsy, and cellular anomalies detected
using immunostaining were quantified in uniformly
sampled frontal lobe sections using field counts. The
results reveal a prevalent connection between ventricu-
Twenty-eight male patients who died from complications of
AIDS and 17 age-matched control men who underwent autopsy were studied. All patients died at the University of
Texas Medical Branch Hospitals in Galveston, Texas, as detailed previously [l, 31. The postmortem incidences of opportunistic infections and human immunodeficiency virus
type 1 (HIV-1)-related changes within and outside of the
CNS were similar to those reported in large reviews [4, 51
and were comparable to data in other reports from East Texas
161. Patients with AIDS who had a CNS opportunistic infection, CNS tumor, gross cerebral edema, terminal hypoxia or
ischemia, or stroke were excluded. Patients with AIDS had
cerebral atrophy ranging from nil to severe; most had at least
one HIV-1-associated abnormality (see [ l , 3-51). Six patients had vague nonfocal neurological symptoms prior to
their demise, and 1 had dementia. The control men had no
neuropathological abnormality and their causes of death were
hepatic failure (4 subjects), malignant neoplasm (4subjects),
From rhe Departmenr of Parhology, Universirv of Texas Medical
Branch, Galveston, TX.
Address correspondence to Dr Gelman, Department of Parhology
G - 8 5 , Universiry of Texas Medical Branch, Galveston, TX 7 7 5 5 0 .
Patients and Methods
Received Nov 25, 1992, and in revised form Feb 23, 1993. Accepted
for publication Mar 3, 1993.
Copyright 0 1993 by the American Neurological Association
thermal injury (3 subjects), acute thoracic trauma (3 s u b
jects), pulmonary embolus (1 subject), acute myocardial infarction (1 subject), and renal failure (1 subject) The mean
( 2 1 standard deviation) ages of the control and AIDS
groups, respecuvely, were 33 I 10 years and 35 2 8 years
The mean time intervals between death and bran removal
in the two groups were 11 + 8 hours and 13
10 hours.
Gross Neuropathology and Measurement of
Ventricular I n A x
After removal, the brains were suspended in 209f buffered
formalin for 10 to 14 days prior to dissection. Fixed brains
were sliced uniformly in the coronal plane at well-defined
basilar surface landmarks; dimensions of the ventricular
spaces were measured to derive the ventricular index (VI),
using a protocol that was described in detail previously El).
For histological analysis, the superior frontal gyrus in Brodmann area 8 was consistently sampled i7).
Staining and Counting o f Frontal
Cortex M icrogliuL Cells
Tissue blocks embedded in paraffin were sectioned at a thickness of 8 and mounted o n glass slides coated with poly-~lysine. Microglial cells in the cerebral cortex were visualized
by immunostaining ferritin using polyclonal rabbit antihuman ferritin (Boerhinger-Mannheim, Indianapolis, IN)
and avidin-biotin complex ( A X ) as: described previously 13,
81. All specimens in every staining procedure were run in a
single batch under identical conditions. Ferritin-stained microglia in 30 random 40 x fields (area per field = 0.19 mm?)
were counted without knowledge of the patient groupings:
10 fields in laminae I and 11, 10 in laminae 111 and IV, and
10 in laminae V and VJ. In some specimens, up to 70 field
counts were obtained. All microglial cells tallied were “diffuse’’ (fields containing microglial nodules were excluded).
Only stained microglial cells with a nucleus in the plane of
section were tallied; isolated fragments of ramified or ameboid microglial cell processes lacking a nucleus were not
counted. A section of a subacute infarct containing macrophages and microglia was used for positive control [3). Results are given as the average number of stained microglial
cells per unit field area.
Staining und Counting of Frontal Lobe Glial
Fibrillary Acidic Protein-Positive Astrocyte.tes
Polyclonal anti-cow glial fibrillary acidic protein (GFAP) was
used at a concentration of 1:500 as the primary antiserum
(Uakopatts, Carpinteria, CA) followed by ABC [ 3 , 91. In
cortex, GFAP-positive cells were counted in the middle laminae (11, 111, IV, and V), without knowledge of the patient
groupings; the glia limitans and gray-white junction were
avoided. Stained cells with a nucleus in the plane of section
were counted in at least 20 fields at a magnification of 40 X ;
rarely up to 70 fields were counted. Glial processes lacking
a nucleus in the plane of section were not tallied. In white
matter, 10 fields were counted (area per field .- 0.793 mm2).
Staining and Counting
White Matter Macrophages
Macrophages were stained for Ham-56 using monoclonal antibody (Enzo Diagnostics, Syosset, NY), pronase digestion,
66 Annals of Neurology Vol 34 No 1 July 1003
and the ARC protocol 13, 8, 101. Stained macrophages in
the white matter were counted in the entire section using
successive sweeps at a magnification of 20 x , as described
[11. Human spleen containing abundant positive macrophages was used as a positive control. The total area of white
matter in each section was measured by tracing the borders
of the white matter in the Lux01 fast blue (LFB)-stained
section, and then measuring the inclusive area of the image
using computer-assisted planimetry (JAVA,Jandel Scientific,
Corte Madera, CA). White matter microglia were not quantified because ferritin staining of myelin causes interference
c3, 81.
Staining of Dendrites
Rabbit polyclonal antiserum (Sigma, St. Louis, MO) against
microtubule-associated protein 2 (IMAP-2) was used as described and characterized elsewhere E1 1, 121, except that
bound antibody was visualized with ABC as above. Human
dorsal root ganglia served as negatively stained control
Linear regression analysis with Pearson’s correlation coefficient, two-tailed Student’s t test, and one-way analysis of variance (ANOVA) 1131 were used, as indicated in the figure
legends and tables.
Microglia, Astroglia, and Dendrites i n Frontal Cortex
The VI was significantly higher in the patients with
AIDS (Table), as previously documented 111. Because
increased cerebrospinal fluid mass compensates for lost
parenchyma, fresh brain weight was not decreased 1I].
T h e postmortem fixation interval did not correlate significantly with any measured result. The density offerritin-stained microglial cells in the cerebral cortex was
twofold higher in the patients with A I D S (Table, Fig
1).Sixty-five percent (lSi28) of the patients had a microglial cell density greater than 2 standard deviations
above the control mcan. Regression analysis between
microglial cell density and VI (Fig 2A) revealed a
strong positive correlation that was statistically significant ( r = 0.706,p < 0.0001). When the AIDS patients
were considered separately, t h e relationship was not as
strong bur still was highly significant (see Table). Six
of 7 patients with neurological symptoms had an increased VI and microgliosis; 1 had an increased VI
without microgliosis. The density of GFAP-positive
astrocytes was significantly higher in the patients with
AIDS, but was correlated weakly with VI (Table, Fig
28). Large apical dendrites were well visualized with
MAP-2 immunostaining; t h e r e was no evidence of
dendritic anomalies in t h e patients with AIDS, using
light microscopy (see Fig 1 ).
Macrophages and htroglia in Frontal White Matter
Ham-56-stained macrophages in the white matter
were distributed around blood vessels 13, 51. T h e i r
Sammary Statir tics
Correlation with Ventricular Index
All Patients
Parameter Measured
Control Subjects”
Venticular index
Cortical rnicroglial cells!mm”
Cortical GFAP-positive astrocytasirnm’
White matter Ham-%-positive rnacrophages/mm2
White matter GFAP-positive astrocyresi’rnm’
* 3.4
27 i 11
5.0 2 6.4
1.5 5 0.6
* 24
10 2
2.4 i 1.3‘
AIDS Patients
1 standard deviation.
‘Different from control mean, p < 0.0000, two-cailed Srudenc’s t test.
CProbabilir)iof type I error with a value < 0.05 considered significant scatiatically.
“Different from control group, p < 0.005 1. using one-way analysis of variance: median value is shown Lx~causcof :wnnormal 2istrihutiori f k f t
“Diffcrent from ronrrol mean, p < O.OjX9, two-tailed Student’s t test.
number of patients; AIDS
acquired immunocleilciency syndrome; GFAP
density was significantly higher ( 6 0 q ) in the patients
with AIDS, but there was no quantitative relationship
to ventricular expansion (Table, Fig 3A). ’The number
of GFAP-positive astrocytes in the white matter was
minimally increased in the AIDS group; no relationship to VI was evident (Table, Fig 38).
The cause of dementia in patients with AIDS is not
clear. The key finding of this study establishes a significant positive correlation between the amount of
bran ventricular expansion and the increased numbers
of stained microglial cells seen in the cerebral cortex
of AIDS patients (referred to hereafter as “diffuse microgliosis”). Detection of microglial cells was enhanced
by using ferritin staining, which gives a high degree of
sensitivity using paraffin-embedded tissue Fpecimens
13,81.Sixty-five percent of the patients with AIDS had
diffuse microgliosis (defined as more than 2 standard
deviations above the mean of control values), compared to only 30% with microglial nodules and 2574
with neurological symptoms. A clear relationship between diffuse microgliosis and ventricular expansion
has not been emphasized previously, perhaps because
both are nonspecific changes that may be too subtle
to measure reliably at autopsy. The connection was
relatively specific because the other brain cellular disturbances measured were not clearly correlated. The
severity of microgliosis accounted for 110 more than
half of the variation in the index of ventricular expansion (r2 = 0.4‘)8), and the data do not establish causation between microgliosis and cerebral atrophy. Their
interrelationship is, nevertheless, potentially important
because an active role of microglial cells h e . , as effector cells) in the wasting and neurodegeneration of
AIDS is often suggested [14-171. Microglia are immune cells belonging to the family of mononuclear
glial hbrillary acidic protein
phagocytes produced in bone marrow [lt3},and they
are considered to be the major source of productive
HIV-1 infection in the CNS {lj]. In vitro evidence
that has been presented demonstrates neurodegeneration in response to secretory products of human mononuclear phagocytes that were infected with HIV- 1
f17). In vivo demonstration of HIV-1-mediated neurotoxicity remains elusive, in part becausc productive
HIV-1 infection in the CNS is sometimes difficult to
document at autopsy [l6, 191. Furthermore, a neurotoxic cytokine has not been isolated from degenerated
brain tissue as yet, and cerebrospinal Huid cytokine
concentrations do not seem to correlate with HIV- 1associated neuropathology [ 161. Therefore, the cause
of the diffuse microgliosis in AIDS is not clear, and it
cciuld be a nonspecific response to CNS degeneration
{IS, 20:. Future studies attempting to establish an effector role of HIV-1-infected microglial cells in neurodegeneration must address the exact nature of the
putative neurotoxin, its disease specificity, its cellular
source, and its concentration in the diseased brain
Patients with AIDS who have advanced neurodegeneration display a nonspecific cortical dendritic anomaly
in a subpopulation of Golgi-impregnated neurons [ 11,
12, 2 11. Dendritic disturbances are potentially related
to brain wasting by “stripping” and/or phagocytosis of
dendrites mediated by activated microglial cells il51.
Visualization of cortical dendrites with MAP-2 immunostaining did not sugqest a link with diffuse microgliosis or ventricular expansion, perhaps because conventional optical focusing was not sufficiently sensitive.
Abnormalities in the staining of myelin in cerebral
white matter also are considered to be relatively prevalent in AIDS patients [l, 3-51. There were no meaningful relationships demonstrated between ventricular
exyansion and GFAP-positive astrocytes or Ham-56Cerebral Atrophy and Microgliosis in AIDS Patients 67
3 M)z
Fzg 1. Immunoperoxidase rtaznrng of the cerebral cortex for m croglzal cdh zn a control Jubyect (A)und a patient uzth AIDS
(B) Cortical mzcroglzul cells contain demr depocztr of ferrztin
and are much more numerozcc in the patient uvth AIDS
( X 135 ) Normal-appearzng a'endrztes of cortical neurons i n
a swerely atrophzed drum stained for mtcrotithule-arrorlate~
protezn 2 (C) ( x 340 )
68 Annals of Neurology Vol 14 No 1 July 1993
Fig 2. Area density ?ffirritin-.rtained microglial celh (A).
glial fibrillay acidic protein (GFAP)-stained astroglial celh
(B) in the cerebral cortex of putients with AIDS (squarcs) and
age-matched control subjects (circles) as a function of zerztriculur index. The ai'erage nzicroglial cell density toas po.rztiz.'e(~~
correlated uith the uentricular index (p = 0.706, p < 0.0001; the
least squares regyeaion line equation is y = 11.4 2 1.88 x i .
The GFAP-positive atroglial cell densitj was signtfirantb
higher in the AIDS patients (p < 0.0051. one-u:ay anahis of
tiariani-e)and waJ correlated vey u v a k ~uith
~ the zvntricular
index (r = 0.ij80. p < 0.007i.
O !
0 0
Fig 3. Area density or Ham-j &stained macrophages (A)and
glialjbri/laq acidic protein (GFAP)-stained astrocytes ( B ) in
the cerebral white matter of patients uith AIDS (squares) and
age-matched control subjects (circles), plotted as a /Ilnction of
c'entrzcularindex with least squares regressiu~zline. Patients
uiith AIDS had significantly more macrophages ip < 0.0389),
with no szgnzjkant correlation with the zmtricular index. The
nvmher oJ GljAP-puiti?v astroqtes ivi the AIDS patierits waj
not signzjkant~~
diflwevtfrom that in control .subjects and had
no significant relationship t o the rwitricular index (see Table).
positive macrophages in the white matter. This finding
is consistent with previously reported nonsignificant
correlations between cerebral atrophy and abnormal
myelin staining {l] or white matter siderotic macrophages [ 3 ] . One case report emphasized specifically a
relative lack of a white matter disturbance in a severely
atrophied brain with AIDS-associated encephalopathy
[22}. Recent findings emphasizing the importance of a
defect in the blood-brain barrier in white matter may
explain this discordance {23]. An increased microglial
cell density in white matter could not be measured in
this study because ferritin staining of myelin produces
interference {3, 81.
In conclusion, these data demonstrate a positive relationship between brain ventricular expansion and diffuse cortical microgliosis in AIDS patients. Both of
these highly prevalent anomalies are subject to substantial variation, but their linkage is probably more
important than previously recognized. Underlying
mechanisms and potential relationships to brain dysfunction need to be emphasized in future investigations.
1. Gelman BB, Guinto FC Jr. Morphometry, histopathology, and
tomography of cerebral atrophy in the acquired immunodeficiency syndrome. Ann Neurol 1992;31:32-40
2. Raininko R, Elovaara I, Virta A, et al. Radiologic study of the
brain at various stages of human immunodeficiency virus infection: carly development of brain atrophy. Neuroradiology 1992;
3. Gelman BB, Rodriguez-Wolf M, Wen J, et al. Sideroric cerebral
macrophages in the acquired immunodeficiency syndrome. Arch
Pathol Lab Med 1992;116:509-516
4. Price RW, Brew 8 , Sidtis j,et al. The brain in AIDS: Central
nervous system HIV-1 infection and AIDS dementia complex.
Science 1988;239:386-592
3. Sharer LR. Pathology of HIV-1 of the central nervous system.
J Neuropathol Exp Neurol 1992;52:3-11
6. Burns DK. Risscr RC, White CL. The neuropathology of human
immunodeficiency virus infection. The Dallas, Texas experience. Arch Pathol Lab Med 1491;115:1112-1124
7. Carpenter MB. Core text of neuroanatomy. 4th ed. Baltimore:
Williams & Wilkins, 1991:399
8. Kaneko Y, Kitamoto T, Tareishi J, et a]. Ferritin immunohistochemistry as a marker for microglia. Acta Neuropathol (Bed)
9. Gelman BB, Goodrum JF, Bouldin TW. Macrophage apolipoprotein synthesis and endoneurial distribution as a response to
segmental demyelination. j Neuropathol Exp Neurol 1491;50:
10. Hulette CM, Downey BT, Burger PC. Macrophage markers in
di,agnostic neuropathology. Am J Surg Parhol 1992;16:493-499
11. Wiley CA, iMasliah E, Morey M, et al. Neocortical damage during HIV infection. Ann Neurol 1901;29:651-657
12. Masliah E, Ge N, Morey M, et al. Cortical dendritic pathology
in human immunodeficiency virus encephalitis. Lab Invest
1992;66:285-29 1
13. Remington RD, Schork MA. Statistics with applications to the
biological and health sciences. Englcwood Cliffs, NJ: PrenticeHall, 1970:193-323
Gelman: Cerebral Atrophy and
Microgliosis in AIDS Patients 69
14 Lipton SA HIV-related neurotoxicity Brain Pathol 1991 , l
15 Dickson DE, Mattiace LA, Kure K, et a1 Microglia in human
disease, with an emphasis on acquired immune deficiency syndrome Lab Invest 1991,64 135-156
16 Tyor WR, Glass JD, Griffin JW, et al Cytokine expression in
the bran during the acquired immunodeficiency syndrome Ann
Neurol 1992.31 349-360
17 Guilian D, Vaca K, Noonan CA Secretion of neurotoxins by
mononuclear phagocytes infected wirh HIV-1 Science 1990,
250 1593-1596
18 Hickey WF, Kmura H Penvascular microglial cells of the CNS
are bone marrow derived and present antigen in vivo Science
1988,239 270-292
19 Budka H Human immunodeficiency virus (HIV) envelope and
70 Annals of Neurology Vol 34 No 1 July 1093
core proteins in CNS tissues of patients with the acquired immune deficiency syndrome (AIDS) Acta Neuropathol (Berl)
Dolman CL Microgha In Davis RL., Robertson DM, eds Textbook of neuropathology 2nd ed Baltimore Wilhams & Wilkins, 1991 141-163
Kaufmann WE Cerebrocortical changes in AIDS Lab Invest
1992,66 261-264
Gray F, Haug H, Chimelli L, er al Prominent Lortical dtrophv
with neuronal loss as correlate of human immunodeficiency virus encephalopathy ActaNeuropdthol (Bed) 1991,82 229-233
Petito CK, Cash KS Blood-brain barrier abnormalities In the
acqmred munodeficiency syndrome immunohistochemical
locdization of serum proteins in postmorrem brain Ann Neurol
1992,12 658-666
Без категории
Размер файла
529 Кб
atrophy, diffuse, associates, immunodeficiency, microgliosis, acquire, syndrome, cerebral
Пожаловаться на содержимое документа