Microarray Analysis of Aeromonas hydrophila Cytotoxic Enterotoxin-Treated Murine Primary Macrophages

Abstract

We performed microarray analyses of murine peritoneal macrophages to examine cellular transcriptional responses to a cytotoxic enterotoxin of Aeromonas hydrophila. While 66% of altered genes were common to both primary macrophages and the murine macrophage cell line RAW 264.7, Act caused expression changes of 28 genes specifically in murine peritoneal macrophages.

Aeromonas species are significant human pathogens that cause both gastrointestinal and nonintestinal diseases in children and adults (11, 22). Our laboratory isolated and molecularly characterized a cytotoxic enterotoxin (Act) from A. hydrophila that caused lysis of red blood cells, was cytotoxic to intestinal and nonintestinal cells, evoked intestinal fluid secretion, and was lethal when injected intravenously into mice (11, 15, 38). Recently, we performed microarray analyses of Act-treated RAW 264.7 murine macrophages in order to obtain a global view of transcriptional responses and to identify genes previously not known to be involved in Act-induced host cell signaling (18). This study directly led to the discovery that Act induced apoptosis in murine macrophages (18). Because signaling pathways in the transformed cell line could differ from those that occur in primary macrophages recruited during infection (3, 20, 33, 34), we performed microarray analysis of Act-treated murine primary (peritoneal) macrophages and compared that to previous data obtained with RAW 264.7 cells (18) in order to identify common and unique Act-induced genes in these cells.

Transcriptional alteration of genes in murine primary macrophages.

Thioglycolate-induced peritoneal macrophages (9) were treated with a sublethal dose of Act (6 ng/ml) for 0 and 2 h (in triplicate), and the RNA was isolated and applied to MGU74Av2 GeneChips (18). In one experiment, the macrophages were treated with Act for 45 min, and the RNA was applied to a single array. The data were analyzed separately by using four different techniques: MAS 5.0, significance of analysis of microarrays (SAM; software program from Stanford), Spotfire 7.1, and analysis of variance (ANOVA) (18). Because only one experiment was performed for Act treatment at 45 min, a change in gene expression from 0 to 45 min was considered significant only if it also occurred between 0 and 2 h. Genes that were deemed significant by all four analysis techniques were compiled into a list that included 90 probe sets representing 81 genes for 0- versus 2-h treatments (data not shown). The majority of the significant genes were associated with inflammation, immune responses, or apoptosis, as we previously demonstrated in microarray findings from Act-treated RAW 264.7 cells (18). However, out of the 81 altered genes, there were 22 genes that were up-regulated by Act in only primary macrophages, such as the interferon regulatory factor 1 (IRF-1) gene, myeloid differentiation primary response gene 116 (MyD116), the chemokine receptor-like 2 (L-CCR) gene, the neutrophil chemoattractant growth-related oncogene 1 (GRO1), the apoptosis-associated tumor necrosis factor (TNF) receptor superfamily member 5 (TRAP) gene, a cDNA representing Ras-related protein Rab-20, and the GTP binding protein (GEM) gene (Table 1).

TABLE 1.

Most significant genes up-regulated or down-regulated by Act in murine peritoneal macrophages from 0 to 2 h as determined by four separate analysis methods

Functiona and GenBank accession no.	Gene	Fold changeb in peritoneal macrophages from 0 to 2 h as determined by:			P value determined by ANOVA
		MAS 5.0	Spotfire	SAM
Adhesion (3)
U68267	Myosin binding protein H	2.2	2.3	2.4	0.029771
Apoptosis (12)
L31532	B-cell leukemia/lymphoma 2 (Bcl-2)	2	2.7	2.2	0.004427
AI324342	Hematopoietic-specific early-response A1-d protein gene (cDNA)	5.9	7.7	7.1	0.004103
X51829	Myeloid differentiation primary response gene 116 (MyD116)	2	2.2	2.1	0.001671
M83312	TRAP	2	2.4	2.4	0.002112
Cell growth or differentiation (7)
Y17860	Ganglioside-induced differentiation-associated protein 10	2.7	2.8	2.7	0.000299
Cytokines or chemokines (13)
J04596	GRO1 oncogenec	18.9	23.2	21.7	0.000226
		22.4	24.7	21.2	0.000138
Immune response or inflammation (24)
AV370035	Chemokine (C-C motif) receptor 5 (cDNA)	−2	−2.2	−2	0.000205
AF030185	L-CCR gene	5	4.7	4.4	0.000388
L22181	Formyl peptide receptor 1	2.8	3	2.8	0.000439
M32489	Interferon consensus sequence binding protein (IRF-8)	−2.6	−2.7	−2.6	0.004545
M21065	IRF-1	3.1	3	3	1.47E-05
AF071180	N-formylpeptide receptor-like 2	3.8	4.1	4.1	9.88E-05
AI060798	Prostaglandin E synthase (cDNA)	3.2	2.9	2.9	0.001989
Immune regulation (3)
AV376312	SPRY domain-containing SSB-1 (cDNA)	−3.0	−3.4	−2.5	0.035525
Signaling (12)
AF030896	1κB epsilon	2.3	2.9	2.9	3.36E-06
L07924	Ral guanine nucleotide dissociation stimulator	2.4	2.7	2.7	0.000583
AF084466	Ras-like GTP binding protein Rad	2	2.7	2.7	8.49E-06
AI854462	Ras-related protein Rab-20 (cDNA)	3.2	3.9	3.4	0.048888
U67187	Regulator of G-protein signaling 2	−3.8	−4.1	−4	0.00014
AI047791	Zinc finger protein 131 (cDNA)	2	2.2	2.1	0.002173
Transcription factors (4)
AB009694	Musculoaponeurotic fibrosarcoma oncogene family, protein F (MafF)	7.5	6.3	3.1	0.006866
L36435	Transcription factor MafB	−3.3	−3.3	−3.2	0.005224
Other or unknown (12)
Z48670	ATP binding cassette (ABC), subfamily D (ALD), member 2	−2.1	−2.6	−2.5	0.001232
AF059213	Cholesterol 25-hydroxylase	2.3	3	2.9	0.039138
U10551	GTP binding protein (GEM)	3.6	4.9	4.4	0.02785
AI850484	Potassium voltage-gated channel subfamily A member 3 (cDNA)	3.1	3.6	3.2	0.001877
AI647003	Spermatogenesis-associated 13 (Spatal13)	3.2	3.3	3.2	2.79E-05

^aNumbers in parentheses indicate the total number of genes in that category that were altered by Act in RAW 264.7 cells (18) and murine peritoneal macrophages. Only those genes in each category that were unique to primary macrophages are listed.

^bA negative sign before the number indicates down-regulation of the gene.

^cThis gene was represented on the GeneChips by more than one probe set, and each was determined separately to be significantly up-regulated by Act.

IRF-1 is an interleukin-6 (IL-6)-induced transcriptional activator of the beta interferon (IFN-β) gene that plays a central role in host immune and inflammatory responses (42). IRF-1 has also been implicated in p53-independent, DNA damage-induced apoptosis of activated T cells and in macrophage-mediated tumor cell apoptotic death (41, 43). Like IRF-1, MyD116 is rapidly induced by IL-6 and expressed in bone marrow cells (27, 29). MyD116 has been reported to be induced by growth arrest and DNA damage stimuli, such as alkylating agents and irradiation, and by diverse apoptotic stimuli (2, 16, 17, 21).

L-CCR, a C-C chemokine receptor originally isolated as a lipopolysaccharide (LPS)-inducible gene in RAW 264.7 cells (39), was also up-regulated by Act only in murine peritoneal macrophages (Table 1). L-CCR is thought to be activated by monocyte chemotactic protein-1 (MCP-1) (45), which has been suggested to play a role in control of bacterial infections and neuroinflammatory processes. While the function of L-CCR has yet to be elucidated, a study of the receptor in HEK 293 cells, a human kidney cell line, demonstrated pertussis toxin-sensitive chemotaxis and small, transient calcium fluxes in response to MCP-1, MCP-2, MCP-3, and RANTES (regulated on activation, normal T-cell expressed) (7). L-CCR has also been implicated in the recruitment and activation of macrophages in the lung, possibly in the early phase of airway inflammation in response to allergens (32). It is therefore likely that L-CCR is a functional chemokine receptor, which may play a role in leukocyte recruitment during Aeromonas-associated infections and likely contributes to the acute inflammatory response of the host.

GRO1 (also called KC), is an inflammatory chemokine involved in neutrophil recruitment (8, 36). TRAP (also called CD40) is a TNF receptor superfamily member which mediates the activation of multiple signaling pathways, including NF-κB, c-Jun NH₂-terminal kinase (JNK), and p38 kinase (5, 6, 26, 44). Rab-20 is a GTP binding protein involved in the membrane trafficking mechanism (30), and GEM binds Ca²⁺ and calmodulin and has been proposed to regulate Ca²⁺ channel expression at the cell surface (4). While the involvement of L-CCR, GRO1, MyD116, IRF-1, and TRAP in immune-related responses seems clear, how Rab-20 and GEM might contribute to inflammation or other host response signaling involved in the mechanism of action of Act requires further study.

There were also six genes that were down-regulated in Act-treated murine primary macrophages (Table 1) but not in RAW 264.7 cells (18), such as the MafB and suppressor of cytokine signaling (SOCS) box protein (SSB-1) genes. MafB is an essential transcription factor involved in the differentiation of multipotent myeloid progenitors to macrophages (23), and Act-induced down-regulation of this gene could conceivably interfere with the ability of host macrophages to respond effectively during Aeromonas-associated infection. While the function of SSB-1 is not known, it does contain a SOCS box domain typical of proteins that suppress the production of cytokines (1, 24). It is possible that Act-induced down-regulation of SSB-1 leads to an inability of host macrophages to dampen cytokine production, which would lead to increased inflammation. Act-induced alteration in the expression of the SSB-1 gene and other genes with unknown function (Table 1) are prospects for future research aimed at characterizing the mechanism of action of Act.

The biological consequences of differences between Act-induced responses of RAW 264.7 cells and murine peritoneal macrophages also require further investigation, as subtle differences may not affect the overall outcome where redundancy occurs. For instance, macrophage inflammatory proteins (Mip-1α, Mip-1β, and Mip-2), which were up-regulated by Act in both RAW 264.7 cells (18) and peritoneal macrophages (data not shown), function similarly to GRO1, in that they are known to recruit neutrophils and have been associated with inflammatory diseases (14, 31, 37, 40). Also, MyD116 functions similarly to growth arrest and DNA damage-inducible 45α (GADD45α) and MyD118/GADD45β (27), the latter two of which were induced upon Act treatment in both RAW 264.7 cells (18) and murine peritoneal macrophages (data not shown).

Confirmation of up-regulated genes in murine peritoneal macrophages by real-time RT-PCR.

In order to confirm the microarray data, real-time reverse transcriptase-PCR (RT-PCR) was performed on selected genes (Table 2). Experiments were run in parallel, and fold change values were determined after normalization of each gene to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene by using the comparative threshold method (28). We chose 7 genes (out of 22) for verification (Tables 1 and 2), based on their known involvement in stress and/or apoptosis and intracellular signaling related to immune responses. As shown in Table 2, L-CCR and GRO1 were up-regulated 4- to 4.5-fold and 18.9- to 24.7-fold, respectively, according to microarray analyses, and we verified up-regulation of L-CCR (59.3-fold) and GRO1 (49.2-fold) by real-time RT-PCR. Likewise, real-time RT-PCR verified Act-induced up-regulation of the other five genes. (Table 2).

TABLE 2.

Confirmation by real-time RT-PCR of selected genes determined to be up-regulated by Act in murine peritoneal macrophages by using microarrays

GenBank accession no.	Gene	Fold change from 0 to 2 h as determined by:
		Microarray	RT-PCR
AF030185	L-CCR	4-4.5	59.3
J04596	GRO1	18.9-24.7	49.2
X51829	MyD116	2-2.2	4.0
M83312	TRAP	2-2.4	4.7
U10551	GEM	3.6-4.9	10.3
M21065	IRF-1	3-3.1	6.3
AI854462	Rab-20	3.2-3.9	2.6

Because previous studies of Act-treated RAW 264.7 macrophages and the present study using microarrays indicated that Act's mode of action involved primarily up-regulation of host signaling molecules, rather than suppression, we did not choose to verify down-regulated genes by RT-PCR. Additionally, it is unclear how these down-regulated genes might contribute to the alteration of host cell signaling mediated by Act in the context of Aeromonas-associated diseases. Nonetheless, our laboratory plans to investigate these genes further in future.

Up-regulation of early inflammatory molecules in Act-treated murine peritoneal macrophages.

We investigated Act-mediated gene expression changes in primary macrophages at 45 min to identify earlier transcriptional responses to Act, which yielded 15 genes. Six of these genes (Table 3) were for cytokines (TNF-α, Mip-2, IL-1α, IL-1β, IL-6, and GRO-1). Immune response molecules were also up-regulated by 45 min, including SOCS3, a molecule possessing ankyrin repeats induced by LPS (MAIL), a member of the nuclear IκB family that has been shown to induce the production of IL-6 in response to LPS (25), and TNF receptor (ligand) superfamily member 9 (also called 4-IBB ligand or CD137 ligand), which upon binding to its receptor, CD137, stimulates both primary and secondary responses of CD4⁺ and CD8⁺ T cells (10).

TABLE 3.

Most significant genes up-regulated or down-regulated by Act in murine peritoneal macrophages from 0 to 45 min, with comparison to genes up- regulated from 0 to 2 h and with Act-treated RAW 264.7 cells

GenBank accession no.	Gene	Peritoneal macrophages					RAW 264.7 cells, fold change from 0 to 2 h (MAS 5.0)
		Fold changea				P value (ANOVA)
		0 to 45 min (MAS 5.0)	0 to 2 h
			MAS 5.0	Spotfire	SAM
D84196	TNF-α	19.7	68.7	73.9	45.3	5.32E-06	7.7
AV374591	MAIL (cDNA)	18.4	9.3	4.9	3.2	0.015896	5.5
L15435	CD137 ligand	9.8	41.6	45.3	31.8	0.002085	2.2
X53798	Mip-2	7	13.9	14.1	13.6	0.000338	60.8
J04596	GRO1b,c	5.7	18.9	23.2	21.7	0.000226
		5.7	22.4	24.7	21.2	0.000138
AA614971	MAIL	4.3	7.7	8.1	8.1	0.000631	14.2
M15131	IL-1β	3.7	41.9	46.1	27.3	0.001014	164.9
X54542	IL-6d	3.5	12.1	21.9	11.3	0.001477
U88328	SOCS-3	3.5	5.5	9	8.9	0.003707	27.5
M14639	IL-1α	2.1	7.4	8.7	8.7	0.035821	15.7
C78513	CDNA	2.1	5.9	12.6	9.4	0.001337	7.5
AF030185	L-CCRc	2	5	4.7	4.4	0.000388
AV374868	SOCS-3 (cDNA)	2	4.7	4.4	4.2	0.000381	11.1
AA204579	Major histocompatibility complex region	−2.1	−2.2	−2.5	−2.1	0.00743	−2.0

^aA negative sign before the number indicates down-regulation of the gene.

^bThis gene was represented on the GeneChips by more than one probe set, and each was determined separately to be significantly up-regulated by Act.

^cThis gene was unique to primary macrophages.

^dAct-induced up-regulation of IL-6 in RAW 264.7 cells was previously demonstrated by enzyme-linked immunosorbent assay (12).

While microarray analyses of RAW 264.7 cells at 2 h did not reveal up-regulation of IL-6 (Table 3), our laboratory has previously demonstrated Act-induced production of IL-6 at the protein level (12). We therefore considered only two of the genes that were up-regulated by 45 min as unique to primary macrophages (the GRO1 and L-CCR genes) (Table 3). Future studies will be aimed at identifying which transcription factors or other signaling molecules are directly responsible for induction of these early genes in murine macrophages.

Act-induced genes follow a similar pattern of expression in primary macrophages.

To identify groups of genes with similar expression patterns, we employed three separate software programs to perform hierarchical clustering: Cluster/Treeview, CLUSFAVOR 6.0, and Spotfire DecisionSite 7.1. Example clusters are shown in Fig. 1A (CLUSFAVOR) and B (Spotfire). There were 21 genes specifically altered in primary macrophages that consistently clustered together and that were also considered significant by four analysis techniques: MAS 5.0, SAM, Spotfire DecisionSite, and ANOVA (Fig. 1C). The similar expression patterns of these genes, as determined by clustering, suggested their involvement in the same signaling pathway or in separate but coregulated pathways.

FIG. 1.

Hierarchical clusters of genes from Act-treated murine peritoneal macrophages. Hierarchical cluster analysis was performed on signal values from seven arrays (0- and 2-h replicate experiments and a single experiment for 45 min), using CLUSFAVOR 6.0 and Spotfire DecisionSite 7.1. The Cluster/Treeview software programs were used to cluster fold change values generated from four experimental comparisons with murine peritoneal macrophages (three replicates of 0 versus 2 h and a single experiment for 0 versus 45 min). (A) Graphical representation of a cluster generated by using CLUSFAVOR 6.0 that represents a set of genes up-regulated by 2 h in Act-treated murine peritoneal macrophages. Normalized intensity values are displayed on the ordinate, and the abscissa represents each experiment. G451, G455A, and G455B represent the three independent experimental array sets (0 h, 45 min, and 2 h) with murine peritoneal macrophages. Error bars indicate standard deviations. (B) Clustering generated by using Spotfire DecisionSite 7.1, showing a set of up-regulated genes (similar to that generated by CLUSFAVOR 6.0). Higher signal values are shown in red, and lower signal values are shown in green. Black represents median signal values. (C) List of genes up-regulated by 2 h specifically in Act-treated murine peritoneal macrophages that clustered together for all three clustering programs used (Cluster/Treeview, CLUSFAVOR 6.0, and Spotfire DecisionSite 7.1).

To reduce the dimensions of the data, we also performed principal component analysis, which could group similarly expressed genes into functional sets (13, 18). Based on principal component analysis, nearly half of the Act-induced gene expression changes in murine peritoneal macrophages occurred by 45 min (data not shown). The majority of these genes were for cytokines and other immune response molecules.

Cytokine profile of Act-treated murine peritoneal macrophages.

There were four cytokines, i.e., Mip-α, Mip-1β, Mip-2, and TNF-α, that were induced by Act in both peritoneal macrophages (data not shown) and RAW 264.7 cells (18) according to microarray analyses, which we confirmed in both cell types by real-time RT-PCR or enzyme-linked immunosorbent assay. Additionally, Act induced up-regulation of IL-1β and IL-6 in murine peritoneal macrophages. Although IL-1β and IL-6 were not determined to be significantly up-regulated in RAW cells by stringent microarray analysis methods (18), we have shown by traditional laboratory techniques that Act induced these genes in RAW 264.7 cells (12, 18, 35). There were several cytokines that were significantly up-regulated in murine peritoneal macrophages but were not included in our original microarray study of Act-treated RAW cells, because they were not consistently altered by all of the statistical methods used. These cytokines included IFN-β, IL-1α, macrophage interferon inducible protein 10 (IP-10), MCP-1, and RANTES. GRO1, on the other hand was considered specific for murine peritoneal macrophages (Tables 1 and 2).

To determine which cytokines were indeed produced in response to Act, we performed a broad cytokine profile analysis on supernatants of Act-treated murine peritoneal macrophages. We chose later time points than those used for microarray analyses (4 and 8 h) in order to examine levels of secreted proteins. There was no significant production of IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, IL-12 (p40), IL-17, granulocyte colony-stimulating factor, granulocyte-macrophage colony-stimulating factor, or IFN-γ in response to Act. However, Act induced production of IL-1α, IL-1β, and RANTES by 4 h and of TNF-α by 8 h, based on a multiplex cytokine assay (Fig. 2A), which supported the microarray analysis results. Act also up-regulated the bioactive heterodimer of IL-12 (p70), demonstrating for the first time the production and secretion of IL-12 by macrophages in response to Act (Fig. 2A). IL-12 (p40), which acts as an antagonist of IL-12 signaling (19), was not detected (data not shown).

FIG. 2.

Cytokine profiling for Act-treated murine peritoneal macrophages. Multiplex cytokine assays (Bio-plex; Bio-Rad, Richmond, Calif.) were performed on the supernatants of Act-treated murine peritoneal cells (0, 4, and 8 h). The concentrations of various cytokines are shown on the ordinate, cytokine names and time points are displayed on the abscissa, and a table below each graph displays the actual, normalized concentration of each cytokine. Results from a representative experiment from a total of three are shown.

A substantial increase in the production and secretion of Mip-1α was observed in response to Act (Fig. 2B), which was expected since Act-induced up-regulation of the transcript for Mip-1α was demonstrated by real-time RT-PCR (data not shown). Act also induced production and secretion of GRO1 (Fig. 2C), which was discovered in this study of primary macrophages by using microarrays and confirmed by real-time RT-PCR (Tables 1 and 2) but which was not demonstrated for RAW 264.7 cells (18).

This study showed, for the first time, Act-induced host transcriptional changes in murine primary macrophages. Act-induced genes in primary macrophages that were discovered by using microarrays, such as those for GRO1, L-CCR, and IRF-1, are prospects for future research and may aid in the further elucidation of the mechanism of action of Act.