| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Biotechnology Laboratory and
Department of Microbiology and Immunology, University of British Columbia, Vancouver, British Columbia, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
,
indicating that host cell responses depend on the activation state of
the cell. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the in vivo mouse model of human typhoid fever, IFN- is released
by NK and T cells 2–3 days following S. typhimurium
infection. IFN- is a potent stimulator of macrophage gene expression
and is necessary for clearance of S. typhimurium and other
intracellular bacteria (6, 7, 8). A variety of studies,
including the use of gene arrays, have supplied a wealth of data
regarding differential gene expression in response to IFN-
stimulation (9, 10). These pleiotropic effects on gene
expression translate into alterations of receptor expression, Ag
presentation, phagocytosis, cell proliferation, metabolism, and the
antimicrobial oxidative and NO burst (11, 12). While
IFN- is thought to prime the macrophage to respond more rapidly and
effectively against invading pathogens, the spectrum of genes whose
expression is altered during bacterial infection in unprimed vs
IFN--primed cells has not been extensively analyzed. Investigating
how IFN- activation alters the ability of S. typhimurium
to affect macrophage gene expression may lead to the identification of
genes that contribute to IFN-’s critical role during S.
typhimurium infection.
Macrophages have evolved the ability to recognize bacterial products and initiate an immune response to clear the microbe. An innate pattern of macrophage response is triggered by conserved bacterial products such as LPS, porins and other outer membrane proteins, fimbrial proteins, flagella, lipoproteins, glycoproteins, and peptidoglycan (13). These bacterial components, termed modulins, signal through CD14 or other pattern recognition receptors to modulate overlapping as well as unique host cell gene expression. These signals help to initiate the innate and specific immune responses to clear the bacterial infection (14, 15). The bacterial surface component LPS is a potent immunostimulatory molecule that initiates both rapid changes in macrophage signaling pathways and adaptive changes in macrophage gene expression. LPS alters the expression of a variety of genes including transcription factors, cytokines, chemokines, receptors, and cationic antimicrobial peptides (16, 17, 18, 19). Other structural components of Salmonella such as porins and flagella induce cytokine gene expression independently of LPS (20, 21, 22). To promote their survival, bacterial pathogens such as S. typhimurium secrete specialized protein effectors that induce alterations in host cells responses (23). These effectors specifically affect host cell functions such as cytoskeletal architecture, vesicle trafficking, cell signaling, and apoptosis to create a more hospitable intracellular niche (24, 25, 26, 27, 28). Most studies to date have shown how bacterial effectors modify existing host proteins rather than examining how host gene transcription is affected.
One way to analyze both the complex interactions between host and
pathogen as well as the priming effects of IFN- is with a general
approach such as gene arrays. Gene array technology has recently been
used for a more global view of differential gene expression in such
fields as inflammatory diseases (29), tumor biology
(30), human cytomegalovirus infection (31),
superantigen stimulation of T cells (32), S.
cerevisiae metabolism (33), and genetic variability
of Mycobacterium tuberculosis (34, 35). One
proven strength of this experimental approach has been the ability to
study the expression of hundreds of genes simultaneously without
biasing conclusions drawn from a subset of genes presumed to be
involved in a particular process. We capitalized on gene array
technology to obtain, for the first time, a more comprehensive picture
of how host gene expression is altered during infection by a pathogenic
bacterium. Differential host cell gene expression was examined in an in
vitro model of S. typhimurium infection using the RAW 264.7
murine macrophage cell line, a common model for the intracellular
growth of S. typhimurium. Gene arrays were used to test two
hypotheses: 1) that most of the gene expression changes in macrophages
infected by S. typhimurium can be induced by LPS, the major
constituent of S. typhimurium outer membranes, and 2) that
the priming of macrophages by IFN- alters the spectrum of genes
induced by S. typhimurium infection. We found that S.
typhimurium infection altered the expression of a large number of
macrophage genes and that an individual virulence factor, LPS, could
itself cause many of the same changes in host gene expression. The
macrophage gene expression profile following infection was altered by
priming with IFN-, revealing how host cell activation state alters
macrophage responses to bacterial infection at the molecular level.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The S. typhimurium strain SL1344 was obtained from
the American Type Culture Collection (ATCC, Manassas, VA) and grown in
Luria-Bertani broth. For macrophage infections, highly invasive
bacterial cultures were prepared by diluting an overnight culture 1:34
in Luria-Bertani broth and subculturing aerobically with shaking for
3 h at 37°C. The murine macrophage cell line RAW 264.7 (ATCC)
was maintained in DMEM (Life Technologies, Burlington, ON) supplemented
with 10% FBS (Life Technologies) without antibiotics at 37°C in 5%
CO2. Where indicated, the cells were cultured
with 200 U/ml IFN- (Genzyme, Cambridge, MA) for 24 h before
infection.
Infection conditions
For immunofluorescence studies and bacterial colony counts,
24-well plates were seeded with 2.5 x 105
RAW 264.7 cells per well. Bacteria were diluted in culture medium to
give a nominal multiplicity of infection
(MOI)3 of 
20.
Invasion was allowed to proceed for 10 min in a 37°C,
CO2 incubator. Cells were washed two times with
PBS to remove extracellular bacteria and then incubated in DMEM plus
10% FBS containing 50 µg/ml gentamicin (Sigma, St. Louis, MO) to
kill any remaining extracellular bacteria and prevent reinfection.
After 2 h, the gentamicin concentration was lowered to 5 µg/ml.
Colony counts and immunofluorescence were subsequently performed in
parallel to compare the variability in the actual number of
intracellular bacteria per cell with the average number per cell for
the population, as determined by colony counts. To determine invasion
efficiency, samples of cells were washed twice with PBS to remove
gentamicin and lysed with 1% Triton X-100/0.1% SDS in PBS at 2 h
postinfection. Numbers of intracellular bacteria were calculated by
colony counts. At various times postinfection, immunofluorescence was
performed as previously described (36) using a rabbit
polyclonal anti-LPS Ab diluted 1:200 (S. typhimurium O
Ag group B factors 1, 4, 5, and 12; Difco, Detroit, MI) and Alexa
488-conjugated mouse anti-rabbit secondary Ab diluted 1:400
(Molecular Probes, Eugene OR). Cells were counted within randomly
selected fields. Consistently, macrophages were infected by an average
of one to three bacteria per cell as assessed by standard plate counts
and immunofluorescence studies.
RNA isolation
RAW 264.7 macrophage cells were seeded at 5.6 x 106 cells in 20 ml media per 150 mm diameter tissue culture dishes and cultured overnight. RAW 264.7 macrophages were infected with S. typhimurium at an MOI of 20 or stimulated with 100 ng/ml S. typhimurium LPS (Sigma) for 4 h. After stimulation, the culture medium was removed for measurement of cytokine production. The cells were washed once with diethyl pyrocarbonate-treated PBS and scraped to detach the cells from the dish. RNA was then isolated using Trizol according to the manufacturer’s directions (Life Technologies). The RNA pellet was resuspended in RNase-free water containing RNase inhibitor (Ambion, Austin, TX). Contaminating genomic DNA was removed using DNaseI (Clontech, Palo Alto, CA) in the presence of 50 U RNase inhibitor for 1 h at 37°C. The reaction was stopped by adding 1/10 volume 10x termination mix (0.1 M EDTA, pH 8.0, 1 mg/ml glycogen) and extracted twice with phenol:chloroform:isoamyl alcohol (25:24:1) and once with chloroform. The RNA was then precipitated with 2.5 volumes 100% ethanol and 1/10 volume sodium acetate, pH 5.2, resuspended in RNase-free water with RNase inhibitor, and stored at -70°C in aliquots to minimize freeze-thaw cycles. Thirty micrograms of total RNA, as determined by OD260 reading, was routinely isolated from one 150-mm dish of cells. The quality of the RNA was assessed by gel electrophoresis and ethidium bromide staining. The absence of genomic contamination was confirmed by using the isolated RNA as a template for PCR amplification using ß-actin-specific primers (5'-GTCCCTGTATGCCTCTGGTC-3' and 5'-GATGTCACGCACGATTTCC-3') in the absence of reverse transcriptase. The absence of an amplicon after 35 cycles was checked by agarose gel electrophoresis and ethidium bromide staining.
Mouse cDNA expression arrays
Atlas mouse cDNA expression arrays I (no. 7741-1; Clontech) consist of a matched set of positively charged membranes containing duplicate spots of 588 mouse partial cDNAs. Information on the genes represented on these arrays and hybridization protocols can be found on the manufacturer’s website: www.clontech.com. Briefly, 32P-radiolabeled first-strand cDNA probes were prepared from 2–5 µg of total RNA from each cell population using Moloney murine leukemia virus reverse transcriptase and pooled primers specific for the 588 genes. 32P-labeled cDNA probe was separated from unincorporated nucleotides using the provided ChromaSpin columns, and probe activity was measured using a scintillation counter. The arrays were prehybridized for 1 h with ExpressHyb containing 100 µg/ml heat-denatured herring sperm DNA (Sigma) to block nonspecific hybridization. The filters were then incubated with 1-5 x 106 cpm of denatured cDNA probes in 5 ml of hybridization solution in hybridization bottles. Hybridization was performed overnight at 71°C in a hybridization oven, and bottles were rotated at 5 rpm. The filters were then extensively washed at low- and high-stringency conditions in hybridization bottles at a rotation speed of 15 rpm, exposed to a phosphoimager screen (Molecular Dynamics, Sunnyvale, CA) for 3–5 days at 4°C, and the resulting hybridization signals measured using a PSI Phosphoimager (Molecular Dynamics).
Image analysis
Atlas Image 1.0 (Clontech) and Excel 5.0 (Microsoft, Redmond, WA) software were used to quantify and compare the hybridization signals. The intensities for each spot were corrected for background levels and normalized for differences in probe labeling using the average values for genes observed to vary little between our stimulation conditions: ß-actin, ubiquitin, GAPDH, calcium binding protein CAB45, and ribosomal protein S29 (37). Spots with an intensity <300 under all conditions, as calculated by Atlas Image, exhibited higher variability and a low signal-to-noise ratio and were therefore not included in the analysis. Genes included in all tables were selected by the following criteria: the mean hybridization intensity values for macrophage genes were altered by >2-fold upon S. typhimurium infection; the averaged data was representative of the individual data sets; duplicate spots on the array gave similar hybridization signals; and the specific hybridization signal was not confounded by background hybridization. Intensity values of zero were replaced by the value of 20 to permit ratio calculation.
Northern blots
cDNA was prepared from total RNA purified from RAW 264.7 cells using oligo(dT) and SuperScript II reverse transcriptase (Life Technologies). The following primer pairs were designed to amplify portions of the indicated macrophage cDNAs to produce templates for probe synthesis: DRFT polypeptide-1 (DP-1), 5'-TCCAATGGGTCTCAGTACAGC-3', 5'-TGTCATTGTCACTGGTGTGG-3'; IL-1ß, 5'-TCCAGGATGAGGACATGAGC-3', 5'-CTTGTGCTCTGCTTGTGAGG-3'; cyclin D1, 5'-CAGCTTAATGTGCCCTCTCC-3', 5'-GGTAATGCCATCATGGTTCC-3'; tristetraprolin, 5'-GGACTCGTCATTGCTGTGG-3', 5'-CAATGGCTTTGGCTATTTGC-3'; CD14, 5'-CTGATCTCAGCCCTCTGTCC-3', 5'-CAGGAGGATGCAAATGTTCC-3'; GAPDH, 5'-AGAACATCATCCCTGCATCC-3', 5'-CTGGGATGGAAATTGTGAGG-3'. Antisense cDNA probes were prepared by PCR using 50 ng of the appropriate PCR product template, the 3' oligo, and modified nucleotides to facilitate repeated stripping of blots (Strip-EZ PCR; Ambion). These single-stranded PCR products were column purified (Qiagen, Mississauga, ON) and labeled with biotin using psoralen-biotin (Ambion) and cross-linking with 365 nm UV light. Northern blots were performed with the NorthernMax-Gly kit (Ambion) which uses glyoxal/DMSO to denature the RNA as an alternative to formaldehyde. RNA was transferred to a positively charged membrane (Ambion) and cross-linked with long-wave UV light and baked at 80°C for 30 min. Labeled probe (3 ng in 10 ml UltraHyb or ZipHyb; Ambion) was used for hybridization at 45°C. The BrightStar nonisotopic detection kit (Ambion) was used for probe detection according to the manufacturer’s protocols. Northern blots were analyzed using an AlphaImager system (Alpha Innotech, San Leandro, CA).
Cytokine assays
The concentration of TNF-
, IL-1ß, and macrophage
inflammatory protein (MIP)-1 in culture supernatantnts from RAW
264.7 cells was determined by ELISA (R&D Systems, Minneapolis,
MN).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Gene array technology was used to examine differential gene
expression in the RAW 264.7 murine macrophage cell line following
S. typhimurium infection. The arrays chosen for this study
contained 588 murine cDNAs encoding proteins with a wide range of
functions and included several gene families whose role during
macrophage responses to infection have not been characterized.
Macrophages were infected with S. typhimurium SL1344 for 10
min, after which cells were washed and treated with gentamicin to kill
any remaining extracellular bacteria and prevent reinfection. The short
invasion time permitted a synchronous wave of bacterial invasion to
induce a coordinated change in gene expression that could be measured
4 h postinfection. Total RNA was isolated from RAW 264.7 cells
that were unstimulated or stimulated with virulent S.
typhimurium or 100 ng/ml purified S. typhimurium LPS.
Fig. 1
shows images of identical arrays
hybridized with 32P-labeled cDNA probes prepared
from RAW 264.7 macrophages that were either left unstimulated, infected
with S. typhimurium, or stimulated with purified S.
typhimurium LPS.
|
compares the
hybridization intensities of these five genes and shows that their
expression levels under different experimental conditions deviated by
not more than 0.7- to 1.4-fold, indicating valid data normalization. To
determine the reproducibility of the gene arrays, we compared the
hybridization intensities of two identical array membranes hybridized
with probes synthesized from two separate RNA preparations of
un- stimulated RAW 264.7 cells. Less than 5% of genes
expressed by unstimulated cells varied by >2-fold between the two
hybridization experiments (data not shown).
|
Our application of gene array technology provided a cross-section
of the diversity of genes whose expression is altered at a given time
point after S. typhimurium infection. Due to the extensive
amount of data accumulated from the gene array experiments, we
have made the data sets for all 588 genes available on our web pages
(http://www.cmdr.ubc.ca/salmonellaarray). At 4 h
postinfection with S. typhimurium, the expression
levels for 40 of the 588 genes represented on the array were altered in
RAW 264.7 macrophages by 4-fold or greater from their uninfected level
(Fig. 2). When a cut-off of 2-fold
induction or inhibition was applied to the data, 77 genes showed
changes in expression. Fig. 2 shows the mean hybridization intensity
values for macrophage genes, encoding a broad spectrum of proteins,
that were induced by >4-fold upon S. typhimurium infection.
Many of these up-regulated genes encode effectors with
well-characterized proinflammatory or direct antimicrobial properties.
For example, inducible NO synthase (iNOS), which encodes the enzyme
responsible for producing the potent antibacterial molecule NO, was
strongly induced upon S. typhimurium infection
(38). Highly elevated expression levels were also observed
for the chemokines MIP-1, MIP-1ß, (39), and MIP-2
(40), which selectively recruit other effector cells to
infection sites (41). The expression of IL-1ß, which
contributes to the proinflammatory and acute-phase responses, was also
up-regulated (42). S. typhimurium infection
also elevated the expression of receptors that allow macrophages to
communicate with other cells of the immune system. Expression of the
gene encoding the receptor for the proinflammatory cytokine TNF- was
up-regulated, as was CD40. CD40 binds to a ligand on T lymphocytes, and
this interaction induces the production of many inflammatory
mediators, primes T cells (43), and augments survival
of mice infected with Salmonella dublin (44). A
subset of the induced genes shown in Fig. 2 may serve to control or
inhibit the inflammatory response. Tristetraprolin was highly induced
upon S. typhimurium infection and can decrease TNF-
synthesis by decreasing mRNA stability (45). Elevated
transcription of the inhibitory 
B (I-B) and ß inhibitory
subunits of NF-B was also observed, and these proteins are known to
down-regulate the transcriptional program initiated by the
translocation of NF-B to the nucleus (46). Elevated
expression of the antiinflammatory cytokines TGF-ß1 and -ß2 was
also observed. TGF-ß can have potent effects on macrophage
activities, and administration of recombinant TGF-ß has been shown to
protect mice from a lethal dose of S. typhimurium
(47). Elevated mRNA levels for signaling molecules that
are involved in cell death or the response to IFN- were also
observed. These include the apoptosis-associated genes ICE protease
(caspase 1), TNF receptor 1, Fas, TDAG51, and TRAIL, and the
IFN--induced IFN regulatory factor 1 (IRF-1). Some of the S.
typhimurium-up-regulated genes also encode proteins involved in
macrophage migration. For example, ICAM-1 is required for vascular
extravasation during migration to sites of infection, and urokinase
plasminogen activator receptor participates in extracellular matrix
remodeling (48). Dystroglycan 1 promotes extracellular
matrix formation, and its transcriptional down-regulation (Fig. 3) may cooperate with the up-regulated
genes encoding various proteases to remodel the extracellular matrix
and promote tissue infiltration by macrophages (49, 50).
|
|
). For example, leukemia inhibitory
factor (LIF) was up-regulated. LIF is secreted by macrophages in
response to LPS and promotes myeloid differentiation (51).
A number of transcription factors were also regulated by S.
typhimurium infection. Expression of Egr-1, NF-E2, IRF-1, and
c-rel was up-regulated, while expression of Ski, B-myb,
Fli-1, and c-Fes was down-regulated by >2-fold (Fig. 3). Egr-1
controls both monocyte development and appears necessary for
maintenance of macrophage differentiation, as the expression of many
cytokines and receptors important during infection are regulated by
Egr-1 activity (52). B-myb is a negative regulator of
macrophage terminal differentiation, and its down-regulation by
bacterial products promotes macrophage development. These transcription
factors all regulate macrophage differentiation, and their coordinated
expression in response to bacterial products may serve to promote
development of the macrophage’s antibacterial abilities (53, 54). Expression of these transcription factors during macrophage
maturation is usually coupled with an inhibition of cell proliferation.
The expression level of many genes controlling cell cycle
G1 to S phase transition were down-regulated.
Modest decreases in the mRNA levels of cyclin D1 and its partner
cyclin-dependent kinase (cdk) 4 were measured. This kinase complex
phosphorylates the retinoblastoma gene product, causing it to
dissociate from the DP-1:E2F heterodimer, which then translocates to
the nucleus and initiates cell cycle progression. The array results
also revealed down-regulation of DP-1 and cyclin E, as well as
up-regulated expression of various retinoblastoma-related genes (data
not shown), all known to block entry into S phase
(55).
Effect of IFN- activation on gene expression by infected
macrophages
IFN- primes macrophages for enhanced microbicidal responses to
bacterial infection. The established importance of IFN- production
during S. typhimurium infection invites a molecular
examination of how the macrophage’s gene expression profile following
S. typhimurium infection is affected by prior IFN-
activation. To this end, gene arrays were hybridized with cDNA probes
prepared from uninfected and S. typhimurium-infected RAW
264.7 macrophages, with or without prior IFN- activation.
Table II presents genes that were differentially
expressed by IFN--activated and unactivated macrophages 4 h
after S. typhimurium infection.
|
treatment altered the expression of a number of
genes and, importantly, that it modulated the ability of S.
typhimurium to alter macrophage gene expression. IFN- often
up-regulated gene expression in uninfected cells, such as BST-1, MIG
monokine, and MIP-1. For some genes, this expression level was
further enhanced by S. typhimurium infection. Examples
include iNOS, I-Bß, NF-B p65, JunB, JunD, TDAG51,
tristetraprolin, and TNF-. For other genes, such as MIG, IFN-
up-regulated their expression but bacterial products did not
significantly increase expression levels above the IFN--stimulated
level. Prior IFN- stimulation resulted in gene expression upon
infection, such as the transcription factors Cdx2 and
Brn3.2, which was not observed at the same time point in infected cells
not primed by IFN-. For other genes, IFN- treatment up-regulated
mRNA levels in uninfected cells, which was repressed following S.
typhimurium infection. The IFN-inducible protein 1 is an example
of this pattern of gene expression that may provide negative
feedback.
The most striking trend was an increase in the steady-state mRNA levels
encoding transcription factors such as tristetraprolin, three members
of the jun family, Fos B, I-B and I-Bß, C-EBP, Stat 5a, and
elk-1. Of note was an induction in the expression of the homeobox (Hox)
family of transcription factors. Expression of the Hox-4.2, caudal-type
homeobox 2, and Brn 3.2 POU transcription factors was up-regulated by
S. typhimurium infection of IFN--treated macrophages to a
much greater extent than in unprimed macrophages. Homeobox genes play
critical roles during development, and the homeobox genes Hox-B3,
Hox-B4, and Hox-B7 have been implicated in orchestrating various stages
of myeloid differentiation (53, 56). This is the first
data, to our knowledge, suggesting that other homeobox genes may play a
role in macrophage responses stimulated by bacterial products.
Contribution of LPS signaling to S. typhimurium-induced changes in gene expression
LPS is a potent inducer of macrophage inflammatory functions
(13, 17). Because S. typhimurium is a
Gram-negative bacteria with an outer membrane rich in LPS, our
hypothesis was that many of the effects of S. typhimurium on
macrophage gene expression are due to its LPS. Gene arrays were used to
identify the relative contribution of the bacterial component, LPS, to
the overall pattern of macrophage gene expression observed during
S. typhimurium infection. This analysis revealed that the
gene expression profiles overlapped considerably (Figs. 2 and 3). In
most cases, 100 ng/ml LPS caused equivalent or greater increases in
steady-state mRNA levels than S. typhimurium infection.
The 100 ng/ml dose of purified LPS used was probably
greater than the amount of LPS that the macrophages encountered during
a 10-min invasion by S. typhimurium. Therefore, of special
interest are genes, such as tristetraprolin, that this semiquantitative
technique suggests are preferentially induced or repressed by
Salmonella invasion in comparison to LPS stimulation.
Confirmation of array data using Northern blots and ELISAs
Despite its reproducibility, gene array analysis is only semiquantitative. Therefore, Northern blots were used to confirm and more accurately measure the regulation of genes identified in our gene array analysis to be regulated by S. typhimurium infection or LPS stimulation. mRNA levels for both CD14, a receptor for LPS, and IL-1ß, a proinflammatory cytokine, were up-regulated, while cyclin D1 levels were decreased in macrophages by S. typhimurium and purified LPS from Northern blot analysis, confirming previously published data (data not shown). Northern blots were also used to confirm the induction or repression of candidate genes identified using array technology where there was little precedence in the literature. We analyzed mRNA levels of DP-1 and tristetraprolin relative to GAPDH in RAW 264.7 macrophages at 1, 4, and 6 h following S. typhimurium infection or LPS stimulation. DP-1 binds to members of the E2F gene family to form a heterodimeric transcription factor that can regulate cell cycle progression (57, 58). Expression of DP-1 is necessary for progression from G1 to S phase, as shown by studies with dominant negative mutants (59). To date, two DP genes and five E2F genes have been identified, and heterodimer subunit composition determines specificity for different E2F DNA binding sites (60). Therefore, regulated expression of DP-1 may coordinate expression of a subset of genes involved in entry into S phase. According to the two array hybridization results, both S. typhimurium and LPS stimulation decreased DP-1 expression by 40% in unprimed macrophages. We confirmed this data by Northern blot analysis, in that DP-1 expression decreased at 6 h following infection or LPS stimulation (Fig. 4A). To our knowledge, this is the first report of repressed DP-1 mRNA levels in macrophages during bacterial infection. An important finding from this Northern blot analyses is that a decrease in macrophage gene expression as small as 40% can be detected by array hybridization and confirmed and quantified by Northern blot analysis.
The expression of tristetraprolin was greatly up-regulated by both
Salmonella infection and LPS, according to the array data
sets. Tristetraprolin, encoded by the gene zfp-36, has been
hypothesized to be a transcription factor due to its zinc finger motif
and its ability to translocate to the nucleus (61).
Tristetraprolin regulates mRNA stability as studies with knockout mice
show that tristetraprolin lowers TNF- protein levels by binding to
the AU-rich elements in TNF- mRNA and destabilizing it
(62). Tristetraprolin is encoded by an early response gene
that is rapidly induced by mitogens (63) and LPS
(45). In Northern blot experiments, we found that
expression of tristetraprolin was increased as early as 1 h
poststimulation by virulent S. typhimurium or by LPS (data
not shown) and then decreased to a lower level at 4 and 6 h (Fig. 4B). The apparent increase in tristetraprolin mRNA levels
was smaller when quantified by Northern blot analysis compared with
the array data, suggesting that the array technique accurately detects
trends in altered gene expression but can overestimate ratios. This
could be explained by the inability of the semiquantitative array
technique to accurately quantify low levels of gene expression, for
example in unstimulated cells. Quantitation of the Northern blotting
results revealed that macrophages infected by S. typhimurium
exhibited a higher level of tristetraprolin mRNA compared with
macrophages stimulated by 100 ng/ml LPS. This confirmed the array data,
which suggested that infection by one to three bacteria per macrophage
induced a 30% higher level of tristetraprolin mRNA than following
stimulation by LPS.
|
(both
conditions resulted in 6.3–8.8 ng/ml) and TNF- (infection, 1.3–2
ng/ml; LPS, 2.8–3.4 ng/ml) were elevated at 4 h, while levels of
IL-1ß were elevated at 24 h (0.3–0.5 ng/ml) when compared with
unstimulated cells. For each ELISA, proinflammatory cytokine
concentrations in culture supernatants of cells stimulated by S.
typhimurium or LPS were similar, supporting our array data. By
array analysis, iNOS expression was induced by S.
typhimurium infection and LPS stimulation. Elevated levels of
nitrate in the culture supernatants were detected at 24 h (data
not shown), indicating increased iNOS activity and confirming that
elevated iNOS expression translated into increased NO production. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Gene array technology is a powerful tool that can be used to expand our
current understanding of this relationship for a number of reasons.
First, this technique permits one to study simultaneous changes in
expression of a large number of genes under uniform experimental
conditions, including infectious dose and cell passage number. While
the selection of genes for inclusion on the array introduces some bias,
the wide range of gene families allows rapid identification of genes
previously not known to be involved in the host response to pathogens.
In this study, we identified genes that have never been directly
implicated in macrophage responses to S. typhimurium
infection and identified novel gene targets of LPS signaling. These
include dystroglycan, which is involved in extracellular matrix
formation, and DP-1, which regulates cell cycle progression. Second,
gene arrays permit comparison of expression profiles obtained from
multiple stages of infection, from stimulation with purified microbial
products, or from infection with bacterial virulence factor mutants.
Our comparison of macrophage gene expression altered by bacterial
infection to stimulation with purified LPS suggests that LPS serves a
principal role in altering host gene expression during S.
typhimurium infection. Third, gene arrays measure changes in
individual genes in the context of how the expression of other members
of the gene family, their receptors, ligands, or transcriptional
activators are altered. This allows a more comprehensive understanding
of host responses to bacterial infection by identifying patterns of
gene expression that would not be evident from studying each gene in
isolation. Indeed, this approach enabled us to detect the induction of
families of transcription factors in IFN--activated macrophages
following S. typhimurium infection.
We were able to identify novel macrophage gene targets of IFN-
activation or S. typhimurium infection by looking at <600
genes. The genes presented in this study likely underestimate the total
number of affected genes due to limitations of accurately quantifying
very low levels of gene expression. This suggests that gene array
filters, used in this study, can complement the use of gene chip
technology, which can analyze the expression of thousands of genes,
because different cross-sections of the genome can be studied in each
case. The use of commercially available filter-based gene arrays is an
accessible approach to generate testable hypotheses of how hosts
respond to pathogens. These arrays have the advantage of containing
characterized genes for which reagents such as Abs, mutant cell lines,
and knockout mice may be available for hypothesis testing. An even more
comprehensive view of host response could be obtained by extending this
approach to using gene microarrays incorporating thousands of genes.
For this to be successful, improved bioinformatics resources are needed
as well as a conceptual shift in the way we analyze and publish large
amounts of data. The findings of many studies similarly rest on our
assumption that changes in steady-state mRNA levels often correlate
with meaningful changes in protein levels. While increased protein
levels have been measured for many of the genes found to be
differentially expressed in this study, others are bound to be
regulated at the level of protein synthesis, posttranslational
modification, or intracellular localization. This also highlights the
need for high-throughput strategies to confirm changes in genes of
interest at the level of transcription, translation, and protein
localization to pursue the biological relevance of array data.
Our gene array results suggest that the macrophage’s transcriptional program undergoes a massive overhaul during bacterial infection and highlight the myriad of ways in which macrophages attempt to control and clear Salmonella infection. The majority of differentially expressed genes were up-regulated upon S. typhimurium infection, and several of these are known to play well-characterized roles during bacterial infection. In general, we observed a strong proinflammatory response that may be tempered by up-regulated expression of TGF-ß, IL-10, and tristetraprolin, all of which have demonstrated antiinflammatory properties. This suggests that there may be a balance between proinflammatory responses and negative feedback regulation during S. typhimurium infection (42). Stimulation by LPS enhances the macrophage’s ability to interact with other cells through the coordinated expression of various receptors, such as CD40 and ICAM-1 (44). Extracellular matrix remodeling, through alterations in the expression of various proteases, protease inhibitors, and dystroglycan may promote macrophage entry into infected tissues (48). Differentially expressed genes identified using the arrays were not limited to genes with characterized proinflammatory or antibacterial properties, because S. typhimurium had numerous effects on the cell cycle regulator and transcription factor gene families within macrophages. With myeloid cells, LPS has anti-mitotic effects by down-regulating the expression of cyclins and cyclin-dependent kinases and by influencing levels of positive and negative transcriptional activators (55). Northern blots for cyclin D1 and the transcription factor DP-1 revealed that the expression of both are decreased to an equivalent extent by LPS and S. typhimurium. This suggests that Salmonella infection may affect the cell cycle via LPS signaling. Our data supports a reprioritizing of host gene expression away from normal physiology toward establishing an antibacterial state.
While S. typhimurium initially invade naive unactivated
murine macrophages in vivo, macrophages are more likely to be
stimulated by IFN- during later stages of S. typhimurium
infection (7). IFN--activated macrophages display
enhanced microbicidal activities upon bacterial infection, due to
changes in the expression of genes such as iNOS and MIP chemokines
(64). However, the spectrum of host responses affected by
IFN- priming is not fully understood at the molecular level. We
analyzed the expression patterns of hundreds of genes to gain a more
comprehensive understanding of how priming by IFN- alters macrophage
gene expression, and hence responses, to S. typhimurium
infection. We identified a variety of gene expression patterns in
IFN--primed RAW 264.7 macrophages, which included up-regulated gene
expression in uninfected cells, synergistic effects between IFN- and
S. typhimurium infection, and elevated expression of genes
following infection of IFN--primed cells that was not seen following
infection of unprimed cells. IFN- signaling has been shown to
increase the amount of NF-B in the macrophage cytoplasm that, upon
LPS stimulation, translocates to the nucleus more rapidly and
effectively than without prior priming by IFN- (64, 65). This model of priming by IFN- may explain the
differential response to S. typhimurium mediated by IFN-,
by altering the kinetics of gene activation, so that genes are elevated
at our 4 h window. Alternatively, IFN- may supply a necessary
first signal so that a second stimulus provided by the bacteria
triggers gene expression, which is not possible in unactivated cells.
Either mechanism could make IFN--primed macrophages more sensitive
to stimulation by bacterial products and permit a more rapid and
effective antimicrobial response against invading S.
typhimurium.
To our knowledge, this is the first report of the application of gene
arrays to the study of macrophage biology by profiling how RAW 264.7
macrophages respond to various stimuli, such as IFN- and LPS.
Maturation of myeloid cells into terminally differentiated macrophages
involves an arrest in proliferation and the differential expression of
many transcription factors (54), some of which were
identified using the arrays. Both LPS and IFN- exert
anti-mitotic effects while promoting development of the
antimicrobial properties of myeloid cells. Many of the cell cycle
regulatory and transcription factor genes expressed by RAW 264.7 cells
in response to LPS stimulation have previously been reported using
primary macrophages (18, 45, 46, 53, 66, 67, 68, 69, 70). This
suggests that RAW 264.7 cells may provide an adequate model for
identifying genes involved in macrophage responses to infection, which
can then be further characterized using primary macrophages. The most
striking class of gene induction in IFN--activated cells 4 h
after S. typhimurium infection was a group of >15
transcription factors. In infected unactivated cells, many of these
transcriptional activators, namely of the homeodomain class, were not
induced above our detection level. Hox transcription factors play
crucial roles during developmental patterning (71). A
previous report has connected the processes of developmental patterning
and macrophage differentiation by implicating the expression of the Hox
transcription factor Hox-2.4 (Hox-B8) in the terminal differentiation
of a hemopoietic cell line along the macrophage lineage
(56). This differentiation required expression of Egr-1,
which was up-regulated upon infection of IFN--activated RAW 264.7
macrophages in this study. Because IFN- activation of macrophages
results in differentiation of monocytes into macrophages, it is
possible that expression of Hox transcription factors upon infection of
RAW 264.7 macrophages, identified in this study, may promote further
maturation of the cell’s antibacterial phenotype. Alternatively, these
transcription factors may serve an as yet uncharacterized role during
macrophage response to S. typhimurium infection
We hypothesized that LPS, a structural component of all Gram-negative
bacteria and the most well-characterized modulin, should play a
principal role in stimulating the early innate response of macrophages
to bacterial infection. To test this hypothesis, we compared changes in
host gene expression caused by virulent S. typhimurium and
purified S. typhimurium LPS to investigate the relative
contribution of this virulence factor. LPS exerts its effects through
its lipid A moiety, which is buried in the cell wall of live bacteria.
During our infection model, cells would be stimulated by the lipid A of
LPS shed by live bacteria, extracellular bacteria killed by
antibiotics, or intracellular bacteria killed by macrophages. There was
a remarkable degree of overlap between genes induced by virulent
S. typhimurium and purified S. typhimurium LPS.
The 100-ng/ml dose of LPS was likely much higher than the amount of
free LPS that stimulated the cells during infection and caused
equivalent or higher alterations in gene expression when compared with
bacterial infection. The overlap in the macrophage expression data
following stimulation with virulent S. typhimurium or
purified S. typhimurium LPS suggests that there is
redundancy in host response to bacteria. Gene expression regulated by
LPS stimulation has also been shown to be altered by other bacterial
components. The ability of both S. typhimurium LPS and
flagellar proteins to trigger TNF- and IL-1ß release by
macrophages (20, 22) supports the concept that different
bacterial inputs can initiate a conserved program of macrophage
responses.
The remarkable overlap in macrophage gene expression induced by S. typhimurium or purified S. typhimurium LPS suggests that Salmonella specifically affects a relatively small subset of macrophage processes to secure their survival rather than completely dampening the inflammatory response. A number of host proteins and signaling cascades have been identified that are modified by specific bacterial virulence effectors. For example, the S. typhimurium virulence factor SopE up-regulates IL-8 production by epithelial cells (25), and SipB binds and activates caspase 1 (ICE) protease to promote macrophage apoptosis (24). The majority of these studies have used epithelial cells and have measured how S. typhimurium invasion and virulence factor expression specifically alter host protein abundance or activity. Our results in macrophages, at the level of altered gene expression, invites a comparative study in epithelial cells to identify similarities and differences in gene expression profiles between these two infection models. Because S. typhimurium resides within macrophages to cause systemic disease, bacterial factors independent of LPS likely specifically modulate macrophage phenotype at the levels of gene expression, protein abundance, and protein activity to secure this intracellular niche. We identified some genes induced to a higher extent by S. typhimurium infection compared with LPS stimulation and have confirmed this higher level of expression for tristetraprolin. While the differential increase in expression was small, it may be significant that another bacterial factor can produce a higher induction in gene expression compared with a relatively large dose of LPS. This raises the intriguing possibility that another virulence factor up-regulates tristetraprolin mRNA levels in macrophages. Our ability to confirm array data for differential tristetraprolin expression suggests that other differentially expressed genes identified by array hybridization may be altered by additional bacterial virulence factors acting synergistically or antagonistically with the effects of LPS. Experiments using killed bacteria or macrophages from LPS-hyporesponsive mice will more accurately quantify the contribution of LPS-independent factors in altering host gene expression. We are presently employing more quantitative techniques to determine whether macrophage genes shown in this study as being differentially expressed to a greater extent upon S. typhimurium infection than by 100 ng/ml LPS, such as tristetraprolin, are specifically responding to an active bacterial process. Array technology is also ideally suited to the study of host gene expression in response to characterized Salmonella mutants to address the contribution of other specific bacterial virulence factors in modulating host gene expression.
This application of array technology will provide insight into how pathogenic bacteria use some of their many virulence effectors to specifically alter host cell biology and secure their niche. Array technology is highly applicable to studying numerous host-pathogen interactions. Comparison of array data from host cells infected with a variety of pathogenic bacteria will likely reveal how specific virulence factors trigger a unique pattern of host gene expression in response to the particular pathogen. Comparison of these data sets with those obtained from LPS and other structural components will likely reveal an overall conserved host gene expression profile that serves as a common signature of infection. Gene array technology promises to provide much-needed insight into host cell gene expression during infection and to broaden our understanding of host-pathogen interactions.
Note added in proof.
Cohen et al. (72) recently profiled gene expression changes in human THP-1 cells following Listeria monocytogenes infection using gene arrays.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. B. Brett Finlay, Biotechnology Laboratory, University of British Columbia, Room 237 Wesbrook Building, 6174 University Boulevard, Vancouver, British Columbia, Canada, V6T 1Z3.
3 Abbreviations used in this paper: MOI, multiplicity of infection; MIP, macrophage inflammatory protein; Hox, homeobox transcription factor; DP-1, DRTF polypeptide-1; I-B, inhibitory B; iNOS, inducible NO synthase.
Received for publication December 20, 1999. Accepted for publication March 21, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
in murine Salmonella typhimurium infection. Microb. Pathog. 8:135.[Medline]
interferon and tumor necrosis factor in resistance to Salmonella typhimurium infection. Infect. Immun. 60:450., ß, or using oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:15623.: biology and role in pathogenesis. Adv. Immunol. 62:61.[Medline]
. Annu. Rev. Immunol. 15:749.[Medline]
in a human promonocytic cell line. Infect. Immun. 66:1127. and conservation of potential regulatory sequences with a human homolog, LD78. J. Immunol. 146:4031.[Abstract]
production by tristetraprolin. Science 281:1001.Bß and abrogation of NF-B activity in peritoneal macrophages stimulated with lipopolysaccharide. J. Biol. Chem. 272:23025.
) and following bacterial
infection (
) or LPS stimulation (
) is shown. Error bars represent
the SD from the mean of three intensity values for infected and
LPS-stimulated cells and the range of two intensity values for
unstimulated cells. Due to the extensive amount of data accumulated
from the gene array experiments, we have made the complete data sets
for all 588 genes available on our web pages
(http://www.cmdr.ubc.ca/salmonellaarray).
