| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Commentary |
From the Cerebrovascular Disease Group, Department of Psychiatry and the Institute for Health of the Elderly, University of Newcastle-upon-Tyne, United Kingdom
Induction of cell surface molecules on cerebral endothelial cells, microglia, oligodendrocytes, choroidal epithelial cells, and (to a lesser degree) astrocytes has been reported in a few chronic inflammatory conditions and infectious diseases of the central nervous system (CNS). The elucidation of autoimmune mechanisms in multiple sclerosis and its experimental model, experimental allergic encephalomyelitis (EAE), has provided knowledge of the pleiotropic actions of cytokines and specific cellular interactions between the circulation-derived immune cells and brain elements. Although stereotypical immune interactions are mimicked in the CNS, they do not appear to measure up to what is encountered within systemic organs. This is due not only to the specialized cellular consistency of the CNS but also to its isolation behind the blood-brain barrier. It may not, therefore, be surprising that the targets, namely endothelial cells or macrophages (microglia) of the brain, do not fully respond to injury or infection in the same manner as or with the expected resilience of those in the periphery. However, inflammatory responses resulting from infections or injury of the CNS activate the brain endothelium and other nonneural cells of the brain to various degrees depending on the type, titer, or strength and duration of exposure to the agent or insult. The activation of these cells may be modulated by the action of one or more cytokines and relies on expression of respective cell surface receptors, though cytokines are known to work via receptor-independent systems. In an effort to understand the immunopathogenesis and find rational treatments several rodent and nonhuman primate models of infectious disease of the CNS have been devised. Such efforts have further been rewarded by the use of several mouse knockout strains or models deficient in molecules of interest.
Infections of the CNS by viruses, bacteria, and protozoa reveal the
presence of inflammatory nodules and perivascular cuffing around blood
vessels by inflammatory cells including infiltrated lymphocytes,
monocytes, and microglia. Astrocytes and oligodendrocytes also
invariably become hypertrophic and exhibit hyperplasia.
However, the consequences of these agents
on neurones remain largely unknown. Protozoal infections of the CNS
such as cerebral malaria (Plasmodium falciparum),
toxoplasmosis (Toxoplasma gondii), and
trypanosomiasis (Trypanosoma brucei and Chagas'
disease) lead to unique immunopathological features, although they also
have features in common with those apparent in chronic inflammatory
conditions such as multiple sclerosis. The molecular mechanisms
involved in these conditions and how these might impinge on
understanding brain inflammatory mechanisms are largely unexplored
topics. Toxoplasmosis, an obligate intracellular protozoan, has
recently attracted considerable attention because it is the most common
CNS infection producing a mass lesion in acquired immune deficiency
syndrome (AIDS). Ten to 25 percent of AIDS patients have toxoplasmosis
of the CNS and Toxoplasma abscesses are a late complication
of HIV infection associated with a low CD4+ cell count. It is well
known that during chronic infection in AIDS and mouse models of
toxoplasmosis, parasite-specific T lymphocytes release high levels of
the cytokine interferon-
(IFN-), which is required to prevent
cyst reactivation and likely initiates several cascades of inflammatory
and immune responses.
Specific Actions of Cytokines on Cerebral Endothelium and Microglia
In vitro and in vivo studies demonstrate
that cytokines such as IFN- and tumor necrosis factor-
(TNF-)
can readily induce a variety of critical cell adhesion molecules as
well as the major histocompatibility complex (MHC) antigens in
nonneural cells of the CNS. IFN- and TNF- seem equally important
stimulators of endothelial cells, microglia and oligodendrocytes but
not astrocytes or neurones. Both class I and class II MHC molecules may
be readily induced in brains of rats and mice depending on the dose and
duration of the presence of IFN-.1,2
It is now well
recognized that the induction of these molecules or adhesion proteins,
such as intracellular cell adhesion molecule (ICAM-1), in endothelial
cells as well as microglia depends also on the strain and species of
animal. For example, IFN- or TNF- can elicit more robust
responses in mice susceptible to EAE or cerebral malaria than in
disease-resistant mouse strains.3
IFN- induced
expression of MHC class II was significantly higher in mice susceptible
to cerebral malaria than in resistant mice. Similarly, site-specific
responses have been noted with respect to these cytokines. Inasmuch as
there appear to be certain differences in the induction of MHC antigens
in endothelial cells between cerebral and peripheral vessels, it is
known that circulation-derived macrophages display a more robust
antigen-presenting capacity function than brain-resident
microglia.4
The activation stage of cells is another
variable important to achieving the desired outcome. For example, the
antigen-presenting capacity of activated (but not unactivated)
microglia to naïve T cells could be increased by treatment with
IFN-.5
The differential actions and the modulation of responses of the
cytokines may be achieved via different receptors. Deckert-Schluter and
colleagues6
previously showed that expression of inducible
nitric oxide synthase (iNOS) was necessary by signaling through the
55-kd TNF receptor-1 (TNFR1) rather than the 75-kd TNFR2. In mice
subjected to encephalitis by Toxoplasma gondii infection the
expression of iNOS protein and mRNA in microglia were reduced, whereas
that of IFN-, TNF-, or interleukin-1ß (IL-1ß), as well
as the recruitment of immune cells in brain, were unaffected in
TNFR1-deficient mice. Whether such preferential signaling occurs in the
cerebral endothelium is unknown. In accord with the differential
effects of TNFR, however, tissue factor synthesis was regulated through
TNFR1 but not TNFR2. This study also showed that endothelial secretion
of tissue factor occurs via the synergistic action of TNF and vascular
endothelial growth factor.7
In this issue of The American Journal of Pathology,
Deckert-Schluter et al8
report on the role of IFN-R-
and TNFR-mediated immune reactions in the activation of cerebral
endothelial cells in the same mouse model of Toxoplasma
encephalitis. Although a strong up-regulation of the cell adhesion
molecules vascular cell adhesion molecule (VCAM) and ICAM-1 and the MHC
antigens was observed in wild-type and TNFR-deficient mice with
Toxoplasma encephalitis, it was not observed in the
IFN-R-deficient mice. Similar lack of up-regulation of ICAM-1 and
its ligand LFA-1 and of the MHC class I and II molecules was also noted
in microglia of the IFN-R-deficient mice compared to the
TNFR-deficient or wild-type mice. These observations suggest that the
IFN-R rather than the TNFR signaling pathway is involved in
activation of the cerebral endothelium and microglia in mice with
Toxoplasma encephalitis. Whether such signaling is specific
to the condition remains to be seen but it is not unlikely that there
will be differences, albeit small ones, between various inflammatory
conditions and infectious diseases, including cerebral malaria.
However, the study also showed that although maximal induction of VCAM
and ICAM was IFN--dependent, the IFN-R-deficient mice showed
significant up-regulation of VCAM upon infection. This observation
suggests that, in addition to IFN-, other factors such as IL-1ß
(also produced in IFN-deficient mice) may be involved in the
up-regulation of these molecules. That IFN--independent factors are
important in other CNS infections is evident in another mouse
encephalitis caused by the lymphocytic choriomeningitis virus, where
normal induction of ICAM-1 and VCAM was observed in IFN--deficient
mice.8
These findings suggest that there may be
considerable downstream cross-talk between different signaling pathways
in response to particular pathogenetic mechanisms.
Deckert-Schluter and colleagues further demonstrated that the
recruitment of immune cells (CD4+, CD8+, and macrophages) was not
impaired in IFN--deficient mice and was similar to that in wild-type
mice.8
They suggest that the low level of induction of
cell adhesion molecules was sufficient in IFN--deficient mice to
ensure entry of leukocytes into CNS. However, recruitment of
macrophages but not leukocytes was impaired in the TNFR-deficient mice.
This is comparable to the EAE mouse model where TNF- deficient mice
also had impaired recruitment of macrophages but not of T cells to the
brain.9
These observations imply that some cytokines may
play divergent roles in guiding the movement of inflammatory cells into
brain parenchyma that may be explained by different pathogenic
mechanisms of such diseases.
Synergistic Actions of Cytokines
Recent studies with brain endothelial cultures have shown that
IFN- and TNF- also act in synergy with each other or with IL-1
but it is unclear yet how they might interact and differentially
regulate downstream signaling pathways. Tanaka and
McCarron10
had previously shown that TNF- inhibits
induction of Ia antigens by IFN- in cultures of cerebral
endothelial cells and that IL-1 may act synergistically with TNF- to
down-regulate or alter immune responses in the endothelial cells. Using
primary cultures of human brain endothelial cells, Dorovini-Zis and
colleagues11
have clearly shown that unlike TNF-,
IFN- has little or no effect in inducing cell adhesion molecules
including ICAM-1, VCAM-1, and E selectin. However, IFN- and
TNF- in combination maximally induced ICAM-1. In contrast, IFN-
alone readily induced MHC class II antigens in the endothelial cells
and likely increased permeability of the endothelium via
receptor-independent mechanisms. Interestingly, even in bovine brain
endothelial cells12
endothelin-1 (and probably
other vasoactive substances) was induced by the combined actions of
TNF- and IFN-. Therefore, direct actions of cytokines that may
result in a change in tone or permeability of the cerebrovascular
endothelium should not be overlooked.
Activation of iNOS and production of nitric oxide (NO) further
illustrate the synergistic actions of cytokines. The combination of
IFN- and TNF- (with or without IL-1ß) was required to
induce iNOS and its mRNA in brain endothelial cells and in microglia
from rats and mice. Such induction occurs via protein tyrosine kinase
and the transcription factor NF-
B. IFN-, in concert with TNF-,
also markedly enhances the ability of IL-1ß-primed endothelial cells
to release reactive nitrogen intermediates, which may affect vascular
tone and permeability. The synergistic actions of cytokines such as IFN
and TNF are undoubtedly important in magnifying their biological
effects during brain injury and inflammation. For example, IFN- can
induce TNFR1 and TNFR1 mRNA in cerebral endothelial cells from
EAE-susceptible mice.13
In support of this,
Deckert-Schluter et al8
also demonstrated that TNF mRNA
was not induced in the IFN--deficient mice with
Toxoplasma encephalitis. A recent study suggesting that
induction of TNFR1 by IFN- facilitates the actions of
TNF- at the transcription level shows the importance of this
finding in oligodendrocytes,14
although it is not
unlikely that brain endothelial cells may respond in a
similar manner. Agresti et al14
demonstrated that neither
IFN- nor TNF- was capable of inducing MHC class I regulatory
element (MHC-CRE) binding activity when administered alone. However,
following exposure of rat oligo-dendrocytes to IFN-, TNFR1
expression was transcriptionally induced by the binding of signal
transducer and activator of transcription-1 (STAT-1) homodimers to the
IFN--activated site (GAS) present in the gene promoter. The
up-regulation of TNFR1 allowed TNF- to induce binding of NF-B to
the MHC-CRE site. Thus, IFN- and TNF- synergistically stimulated
interferon regulatory factor (IRF-1) gene expression. IFN- directly
induced the binding of STAT-1 homodimers to the GAS element, whereas
NF-B binding to B sequence was activated by TNF- only after
IFN- treatment.
IFN-R and TNFR- Signaling
Current studies show convincingly that the actions of TNF in
endothelial cells are mediated via protein tyrosine kinase and protein
kinase C (PKC), rather than protein kinase A (PKA), signaling
pathways.15,16
Human microvascular endothelial
cells exposed to ultraviolet light B irradiation induced ICAM-1
up-regulation and lymphocyte-endothelium interaction.17
These authors also observed that PKC inhibitors rather than PKA
inhibitors were effective in attenuating this induction and that
PKC- was translocated from the cytosol to the membrane, indicating
enzyme activation. However, IFN- but not TNF- antibodies blocked
the irradiation-induced ICAM-1 up-regulation. In another approach, Grau
et al3
described TNF--induced up-regulation of
both TNFR1 and TNFR2 receptors in brain endothelial cells from cerebral
malaria-susceptible but not from resistant mice. PKC inhibitors blocked
the response to TNF- in both mouse strains but an inhibitor of PKA
selectively abolished the response to TNF- in the cerebral
malaria-resistant mice. Although these observations support the
importance of TNR-mediated PKC-dependent signaling in certain
inflammatory conditions, they collectively imply divergent signaling
pathways for cytokine-induced adhesion molecule expression.
IFN- and several other cytokines have been found to activate the
Janus kinases (Jaks) and the STAT proteins.18
Although
similar signaling ought to be present in all cells expected to
respond to cytokines, the components of the signaling pathways have not
yet been characterized in CNS cells with IFN-R or other cytokine
receptors. Jaks are a unique class of tyrosine kinases that associate
with cytokine receptors. On ligand binding, they activate (see
above) members of the STAT family through phosphorylation on a single
tyrosine. Activated STATs form dimers, translocate to the nucleus, bind
to specific response elements (eg, GAS) in promoters of target gene,
and transcriptionally activate these genes. Both positive and
negative regulations of the Jak-STAT pathway have been identified. It
has also been shown that STAT-1 knockout mice have impaired IFN
signaling.18
Convergence of the actions of TNF- and IFN- is thought to occur
through the interferon regulatory factor (IRF-1). The induction is a
primary transcriptional response that occurs rapidly without further
requirement for protein synthesis. Synergism is mediated by a novel
composite element in the IRF-1 promoter that includes GAS, overlapped
by a nonconsensus site for NF-B. Synergistic induction of IRF-1 is
likely to be an important early step in regulatory networks critical to
the synergism of TNF- and IFN-. The GAS/NF-B element
may mediate synergistic transcriptional induction of IFR-1 by other
pairs of ligands that together activate NF-B and STAT family
members. There is also some evidence to suggest that transmodulation
between the STAT and SMAD (homologues of
Drosophilia mao proteins) signal transduction pathways
occurs, providing a further means of cross-talk between IFN- and
TGF-ß signaling pathways.
Cerebral endothelial cells and microglia are activated by
cytokines to induce a variety of cell surface molecules during
inflammation and infectious disease. The degree of induction may depend
not only on the nature of the cytokine but also on the pathogenetic
features of the condition.19
Although IFN- and TNF-
act independently, it is apparent that their synergy in the induction
of characteristic cell adhesion and MHC molecules is important in
modulation of specific responses in certain CNS infections. However,
the specific downstream signaling components after cytokine receptor
stimulation remain to be elucidated for a full understanding of immune
responses in models of infectious disease of the CNS such as
Toxoplasma encephalitis and cerebral malaria. Although
more peculiarities of cytokine actions remain to be elucidated in the
pathogenesis of the these models or other CNS diseases, it is important
to bear in mind that variable results often arise due to
different experimental designs, time point(s) of the course of
infection being studied, and species and strains of animals
used.
Footnotes
Address reprint requests to Prof. R. N. Kalaria, CBV Path Group, MRC Unit, Newcastle General Hospital, Westgate Road, Newcastle-upon-Tyne NE4 6BE, UK. E-mail: r.n.kalaria{at}ncl.ac.uk
Accepted for publication March 19, 1999.
References
induced major histocompatibility complex expression in the central nervous system: astrocytes fail to express major histocompatibility complex class I and II molecules. Lab Invest 1999, 79:235-242[Medline]
in vivo. Glia 1997, 19:181-189[Medline]
receptor-, but not tumour necrosis factor receptor type 1- or type 2-mediated signalling is crucial for the activation of cerebral blood vessel endothelial cells, and microglia in murine Toxoplasma encephalitis. Am J Pathol 1999, 154:1549-1561 on endothelial cells from murine central nervous system microvessels. J Neuroimmunol 1990, 27:209-215[Medline]
on primary cultures of human brain microvessel endothelial cells. Am J Pathol 1993, 142:1265-1278[Abstract]
on the production of vasoactive substances by cerebral endothelial and smooth muscle cells inn culture. J Cereb Blood Flow Metab 1995, 15:920-928[Medline]
and TNF- on cultured cells. J Neuroimmunol 1995, 62:161-167[Medline]
and TNF- in oligodendrocytes. Eur J Neurosci 1998, 10:2975-2983[Medline]
in brain microvascular endothelial cells is mediated by a protein tyrosine kinase and protein kinase C-dependent pathway. J Neuroimmunol 1996, 70:199-206[Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
