|
|
||||||||
The Johns Hopkins University School of Medicine, Departments of Medicine,
* Biological Chemistry, and
* Pediatrics, Baltimore, Maryland 21205, USA
2Correspondence: Department of Biological Chemistry, Physiology 615, The Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205, USA. E-mail: gsack{at}jhmi.edu
| ABSTRACT |
|---|
|
|
|---|
(TNF-
), interleukin 1ß (IL-1ß), and IL-6. However, brain
SAA1 expression fell after injecting either virus or
control saline and remained low despite increases in TNF-
and IL-6,
which are known to induce its expression in hepatocytes. Our data thus
show that expression of the murine SAA1 gene has different,
unprecedented control in mouse brain, suggesting that the protein
itself may have a different physiological role there.Tucker, P. C., Sack, G. H., Jr. Expression of serum amyloid A (SAA) genes in
mouse brain: unprecedented response to inflammatory mediators.
Key Words: inflammation neuroimmunology acute-phase reactants gene regulation SAA
| INTRODUCTION |
|---|
|
|
|---|
) and
interleukins 1 and 6 (IL-1, IL-6) (5
Despite their prominent induction during the acute-phase response and
several proposed roles in vivo and in vitro, the primary physiological
function(s) for SAA proteins remain unclear. SAA proteins are small
(104 amino acids in humans, 103 in mice) and well conserved in
evolution (9)
. They are lipophilic (10
, 11)
and poorly soluble in aqueous solution. SAA binds as a
lipoprotein to the high density lipoprotein
(HDL)3 fraction of serum, altering delivery of
cholesterol-laden HDL to hepatocytes, macrophages, and other cells
(12
13
14)
. Cholesterol binding is mediated by amino acids
118 of acute-phase SAA (15)
. At least one member of the
SAA protein family is an autocrine stimulator of collagenase in
inflamed synovial cells (16
17)
.
Because SAA proteins bind cholesterol, show lipid solubility, and are
prominent in inflammation, we have sought evidence for their synthesis
in the brains of mice where many details of inflammatory responses to
experimental infections have been studied. We used RT-PCR to evaluate
SAA mRNA production in brains of normal mice and in response to a viral
infection. We have found constitutive SAA gene expression in normal
murine brain as well as a remarkable decrease in its expression during
viral infection despite the presence of inflammatory mediators
(TNF-
, IL-6, and IL-1ß).
| MATERIALS AND METHODS |
|---|
|
|
|---|
Detection of SAA and cytokines by RT-PCR
mRNA was detected by RT-PCR as described (19)
.
Cellular RNA was extracted from frozen homogenized brain or liver with
RNAZOL (Tel-Test, Inc., Friendswood, TX) and primed with oligo(dT).
Reverse transcription was carried out with avian myeloblastosis virus
reverse transcriptase (Boehringer Mannheim, Indianapolis, IN).
The PCR used primers whose sequences are conserved between expressed
murine SAA genes (20)
to amplify a 312 nucleotide product
that crossed splice sites to avoid confusion between amplification from
cDNA and genomic DNA (see Table 1
). Different numbers of PCR cycles were performed to ensure that
amplification was occurring in the linear range. PCR using negative
controls (water used instead of RNA for synthesis) and a positive
control (cDNA for glyceraldehyde-3-phosphate dehydrogenase) were
included in every run. Serial dilutions of the positive control were
amplified at 25, 30, and 35 cycles, generating a standard curve to
ensure a fixed relationship between initial RNA input and densitometric
readout.
|
A portion of the PCR reaction was electrophoresed through a 1.2%
agarose gel and transferred to nylon (Amersham, Arlington Heights, IL).
An oligonucleotide probe (Table 1)
internal to the PCR primers was
labeled with 32P-ATP using T4 polynucleotide
kinase and then used for Southern blot analysis. Radioactivity of the
bands on the autoradiograms was estimated by laser scanning
densitometry (Molecular Dynamics, Sunnyvale, CA). The relative
intensity of each mRNA band was divided by the intensity of the
autoradiogram band for the internal positive control to determine the
relative amount of mRNA. All studies were performed in duplicate.
Sequencing of RT-PCR products
The product of 30 amplification cycles was run on a 2% Nu-sieve
agarose gel (FMC Bioproducts, Rockland, ME). The gel was stained with
ethidium bromide and the prominent band was cut from the gel. DNA was
eluted from the gel slice using a Qiagen column. Sequencing of both
strands was performed using the original PCR primers and an Applied
Biosystems (Foster City, CA) sequencer.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
|
The RT-PCR product corresponds to murine
SAA1
Figure 2
shows the sequence of the RT-PCR product. Only a single sequence was
found, corresponding precisely to the transcript expected from the
murine SAA1 gene (20)
. Splice
junctions are present, consistent with this being an authentic
transcript containing sequence information from exons 24 of the
chromosomal gene (20)
as delimited by the primer
sequences.
|
These data show that the murine SAA1 gene is
constitutively transcribed in brains of normal BALB/cJ mice. Although
this observation was unexpected (because the SAA proteins have been
described as synthesized largely in liver but also in spleen and
leukocytes), the lipophilicity of SAA proteins (10
, 11)
at
least suggested that they might have legitimate roles in brain
physiology.
Murine brain SAA1 transcription pattern after virus
infection
Figure 3
shows levels of SAA1 expression compared with the
patterns of IL-6, TNF-
, and IL-lß before and at intervals after
Sindbis virus infection. Several features are apparent:
|
1) SAA mRNA levels fall immediately after infection and remain low for up to 6 wk. Control mice inoculated with HBSS showed a similar decrease in SAA mRNA levels, suggesting that inoculation trauma or HBSS itself may affect SAA expression;
2) IL-6 and TNF-
mRNA levels reach peaks at 6 h and
7 days postinfection, respectively. Finding low levels on day 4 is
consistent with the earlier report by Wesselingh et al.
(19)
, who found the highest inter-animal variations on
days 34; they attributed these variations to differences in rates of
mononuclear cell influx into the brain. Thus, the trends to a
consistent rise in TNF-
mRNA levels to day 7 and a consistent fall
in IL-6 mRNA levels from the outset are more likely the case and
similar to those found in the earlier study where measurements were not
reported for day 4 (19)
. IL-1ß mRNA levels rise later,
with a peak around day 9;
3) Brain SAA mRNA levels do not rise in response to an
increased expression of mRNAs for TNF-
, IL-6, and IL-1ß (which
presumably reflects the synthesis of the respective proteins) of nearly
two orders of magnitude.
Our finding low levels of brain SAA mRNA in the presence of an increase
of nearly two orders of magnitude in TNF-
and IL-6 mRNA levels is a
striking contrast to the effects of these same cytokines on
hepatocytes, where they induce a large increase in SAA mRNA expression
in the first 1824 h after exposure (4
5
6
7
8
, 21)
. In our
mice, liver SAA mRNA levels rose modestly only 12 wk after brain
infection (see Fig. 4
). This rise was seen at the same time as modest doubling of the liver
TNF-
mRNA level and is compatible with liver infection by newly
replicated AR339. The relatively low level of stimulation of liver
TNF-
mRNA expression presumably reflects the low titer of AR339
released from the infected brain and its low efficiency for liver
infection. Liver IL-6 mRNA was not induced (data not shown).
|
This remarkable contrast between SAA mRNA response patterns in brain and liver is compatible with several possibilities:
1) TNF-
and IL-6 in brain do not reach the cells
responsible for SAA mRNA expression;
2) Signal transduction pathways responding to TNF-
and
IL-6 in brain differ from those in liver. This is consistent with our
observation that levels of brain SAA mRNA begin to fall early after
infection, even before the peaks of TNF-
and IL-6 mRNA;
3) The trauma of intracerebral inoculation may reduce SAA mRNA transcription in brain without affecting the cytokine responses;
4) Neurons may be the site of constitutive brain SAA mRNA
expression in mice. In human brains, in situ hybridization has shown
SAA mRNA expression in cortical pyramidal neurons and cerebellar
Purkinje cells (22)
. Because neurons are primary targets
of Sindbis virus infection (23)
and Sindbis virus shuts
off host cell protein synthesis early in infection, synthesis of SAA
mRNA itself or some mediator of SAA mRNA expression may be reduced.
(TNF-
and IL-6 are synthesized in other cell types.) However, this
does not explain the decreased SAA mRNA levels after inoculation of
HBSS alone.
As shown in Fig. 3
, there is no apparent delay or interruption of
transcription of cytokines IL-1ß, IL-6, and TNF-
in response to
AR339 infection; levels of these are similar to those reported before
(18
, 19)
. Thus, the persistent low levels of SAA mRNA
after 8 days (when the virus infection has been cleared; 18
, 19
, 23
) are unexpected, and we consider at least two explanations.
1) The neurons responsible for synthesizing SAA mRNA or its
mediators may have been killed or damaged by the infection. However,
mice recover from AR339 infection uneventfully, making it unlikely that
enough neurons would be lost to inhibit SAA mRNA transcription to this
extent unless a certain population of neurons, susceptible to AR339
infection, is also responsible for SAA mRNA transcription.
2) There may have been long-term postinoculation change(s)
in molecule(s) responsible for inducing SAA mRNA expression.
If the second explanation is correct, however, this response pattern
appears limited to the brain because there is a modest but significant
increase in SAA (and TNF-
) mRNA induced in the liver at late times
after brain infection (see Fig. 4
).
Hardardóttir et al. did not detect SAA mRNA in brains of
unstimulated Syrian hamsters, although it did appear after
intraperitoneal (i.p.) injection of LPS. i.p. injection of TNF-
and
IL-1ß did not lead to SAA mRNA production in these brains
(24)
. SAA was not detected by immunohistochemistry in the
brains of normal rhesus monkeys (25)
. In contrast to the
findings in Syrian hamsters and monkeys, our data show that SAA mRNA
transcription is constitutive in mouse brain. Low levels were detected
by immunohistochemistry in normal human brains (26)
.
SAA gene expression has been studied in liver or hepatocyte-derived
systems (4
, 6
7
8
, 20)
. Mice have low constitutive levels
of hepatic SAA mRNA that rise rapidly after acute-phase stimuli
(4
, 8
, 27)
. SAA genes differ in 5' control elements. For
example, murine SAA1 and
SAA2 have NF-
B sites (7
, 28)
whereas human SAA4 lacks an NF-
B site, but has a prominent AP-1 site
(29)
. Additional 5' sequence information has been
identified for SAA genes in humans and mice, but these regions have not
been characterized fully. TNF-
activates NF-
B but, as our data
show, SAA mRNA transcription in the mouse brain is not stimulated by
local synthesis of TNF-
.
Our observations are significant in relation to both normal brain
function and central nervous system responses to infection and
inflammation. For example, the pathological changes of Alzheimers
disease involve inflammatory cells and proteins. Activated
microglia/macrophages with MHC class II antigens are conspicuous
(30
31
32
33)
. Senile plaques contain cytokines (IL-1, IL-6,
TGF, TNF-
), complement factors, pentraxins, and other inflammatory
proteins. Local astrocytes, microglia, and neurons of Alzheimers
disease brains contain mRNAs for most of these proteins, supporting the
notion that these cytokines arise endogenously (34
35
36
37)
. A
causal relationship between the chronic activation of brain glia that
occurs in aging and the deposition of inflammatory proteins and
ß-amyloid in brain plaques has been proposed (38)
. Thus,
although these cytokines (known inducers of hepatic acute-phase protein
synthesis) are present, our data indicate that they do not induce SAA
mRNA expression in mouse brain after Sindbis virus (AR339) infection.
Liang et al. found SAA1 proteins in extracts of
brains from humans with Alzheimers disease as well as low levels of
SAA mRNA, but the stimulus(i) for SAA gene expression was(were) not
identified (39)
. Subsequently, immunohistochemistry of
brains from individuals with Alzheimers disease showed SAA deposition
in myelin sheaths and axonal membranes but not in plaques and tangles
(26)
. Similar changes were found in brains from
individuals with multiple sclerosis (26)
.
Immunohistochemistry also localized SAA to capillaries and
microinfarcts of hypertensive (but not normotensive) rhesus monkeys
(25)
. Thus, vascular damage or primary inflammation of
myelin (as in multiple sclerosis) can be associated with localized SAA
accumulation in the brain. These pathological associations are
consistent with the known interactions of SAA with lipids (10
, 11)
and cholesterol (15
, 40)
.
T cell-dependent factors are unlikely to explain the early fall in SAA
mRNA expression that we have observed, because the T cell response to
Sindbis virus infection develops only
3 days after infection. By
contrast, synthesis of IL-1, IL-4, IL-6, IL-lß, TNF-
, and TGF-ß
in cells intrinsic to the brain begins immediately after infection
(23
, 41
, 42)
; our data have reproduced several of these
patterns.
It will be important to determine the site(s) of SAA mRNA synthesis in
the mouse brain, particularly to distinguish between neuronal and
non-neuronal cells. Using immunohistochemistry and in situ
hybridization in normal human tissues, Urieli-Shoval et al. showed that
pyramidal cells in the cortex and Purkinje cells in the cerebellum
express SAA mRNA and protein (22)
. Circulating blood
leukocytes are unlikely to be responsible for the brain SAA mRNA
expression we have observed. Relatively few leukocytes are present in
our specimens; if present, they would be expected to show stimulated
SAA mRNA expression in response to the cytokines. It also will be
important to determine the concentration of SAA proteins present.
The basis for the striking lipophilicity of native SAA proteins is
unknown (10
, 11
, 15
, 40)
. Primary sequences of SAA
proteins include
40% charged residues, and so the lipophilicity
must at least reflect folding. Clearly, the amino-terminal residues of
SAA proteins are lipophilic (1
, 2
, 15)
, but our
unpublished observations have shown that removing the 6 amino-terminal
residues does not increase solubility and leads to denaturation.
Unfortunately, because of their intrinsic insolubility SAA proteins
have resisted many structural studies. Our study of an SAA fusion
protein has shown that both helical and ß-sheet domains are present
(43)
. Further characterization of the structure of folded
SAA should help explain the function and distribution of SAA in the
normal brain as well as in response to infection, inflammation, aging,
and other degenerative processes.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Received for publication April 4, 2001.
Revision received June 19, 2001.
| REFERENCES |
|---|
|
|
|---|
B and NF-IL6 transactivating factors in the synergistic activation of human serum amyloid A gene expression by Interleukin-1 and Interleukin-6. J. Biol. Chem. 268,25624-25631
B-like transcription factor. Mol. Cell. Biol. 9,1908-1916This article has been cited by other articles:
![]() |
A. Grolleau, D. E. Misek, R. Kuick, S. Hanash, and J. J. Mule Inducible Expression of Macrophage Receptor Marco by Dendritic Cells Following Phagocytic Uptake of Dead Cells Uncovered by Oligonucleotide Arrays J. Immunol., September 15, 2003; 171(6): 2879 - 2888. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. Mori, F. Goshima, T. Koshizuka, N. Koide, T. Sugiyama, T. Yoshida, T. Yokochi, Y. Nishiyama, and Y. Kimura Differential activation of the c-Jun N-terminal kinase/stress-activated protein kinase and p38 mitogen-activated protein kinase signal transduction pathways in the mouse brain upon infection with neurovirulent influenza A virus J. Gen. Virol., September 1, 2003; 84(9): 2401 - 2408. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |