|Home | About | Journals | Submit | Contact Us | Français|
The innate immune system is triggered when pathogen associated molecular patterns (PAMPs) expressed by infectious microorganisms interact with toll-like receptors (TLR) present on immune cells. Individual TLRs signal through distinct molecular pathways. For example, TLR9 interacts with unmethylated CpG motifs expressed by bacterial DNA and triggers via a MyD88 dependent pathway whereas TLR3 recognizes viral RNA through a MyD88-independent pathway. Bioinformatic analysis of microarray data was used to identify the regulatory patterns underlying changes in gene expression induced when RAW 264.7 macrophages were stimulated via TLR9 by CpG oligonucleotides (ODN) and/or via TLR3 by poly (I:C). While the genes activated by each ligand mediated similar functions, poly (I:C) elicited a larger and more diverse change in gene expression. Co-stimulation with both ligands accelerated gene expression and synergistically activated genes primarily associated with immune function. This is the first work to compare global changes in gene regulation triggered by distinct TLR pathways and clarify their impact on gene expression.
The innate immune system is triggered by the recognition of pathogen-associated molecular patterns expressed by a variety of infectious microorganisms. Members of the toll-like receptor family contribute to the recognition of these PAMPS. For example, TLR9 is stimulated by the unmethylated CpG motifs present at high frequency in bacterial DNA (an effect mimicked by synthetic ODN expressing identical CpG motifs) (Hemmi et al., 2000; Klinman et al., 1996; Krieg et al., 1995) while TLR3 recognizes double-stranded viral RNA (an effect mimicked by poly (I:C)) (Alexopoulou et al., 2001; Kulka et al., 2004; Matsumoto et al., 2002; Yamamoto et al., 2003). Microarrays provide a powerful tool for analyzing changes in gene expression induced when the innate immune system is activated. Previous work from this laboratory examined the network of gene interactions induced when CpG ODN were used to stimulate murine spleen cells. Results showed that a small group of regulatory genes mediated global changes in gene expression over time (Klaschik et al., 2007). The current work extends those findings by comparing the effects of poly (I:C) vs CpG ODN on gene regulation in a murine macrophage cell line and delineating the synergy resulting from the co-administration of both ligands.
In addition to responding to distinct ligands, individual TLRs use different adaptor proteins. For example, TLR9 signals via MyD88 whereas TLR3 signals via TRIF (Boonstra et al., 2006; Hoebe et al., 2003; Honda et al., 2005; Latz et al., 2004) Individual pathogens frequently express multiple PAMPs and thus can trigger the innate immune system via multiple discrete TLR pathways, potentially culminating in a synergistically enhanced immune response (Gautier et al., 2005; Napolitani et al., 2005; Trinchieri and Sher, 2007; Zhu et al., 2008). Consistent with such a possibility, previous studies showed that the expression of IL6, IL12B, TNF and NOS2 were synergistically enhanced by co-stimulation with CpG DNA plus poly (I:C) (He et al., 2007; Whitmore et al., 2004; Zheng et al., 2008). However, neither the full impact on gene up-regulation nor basis of this synergy is known (Whitmore et al., 2007). The current work examines genome-wide changes in gene expression induced by treating RAW264.7 macrophages with CpG ODN and/or poly(I:C). Results provide novel insights into the activation patterns and regulatory gene networks induced with both ligands and identifies that subset of genes whose expression is synergistically enhanced.
Unmethylated CpG-containing phosphorothioate oligonucleotides 1555 (5′-GCTAGACGTTAGCGT-3′) and 1466 (5′-TCAACGTTTGA-3′) were synthesized at the CBER core facility (CBER/FDA, Bethesda, MD) and used as an equimolar mixture (Takeshita and Klinman, 2000). Polyinosine-polycytidylic acid (poly I:C) was purchased from InvivoGen (San Diego, CA). The mouse RAW 264.7 macrophage-like cell line was obtained from ATCC (catalogue number TIB-71, Manassas, VA).
Bone marrow derived dendritic cells (BMDC) were produced by culturing BALB/c BM in RPMI 1640 medium supplemented with 10% FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 20 ng/ml GM-CSF for 6-7 days at 37° in a 5% CO2 in air incubator. BM-derived macrophages were produced by culturing BALB/c BM in DMEM supplemented with 10% FCS, 50 U/ml penicillin, 50 μg/ml streptomycin, 1 mM sodium pyruvate and 50 μg/ml M-CSF (R&D, Minneapolis, MN) for 5 days.
Cultured DC and macrophages were washed and stimulated in vitro for 20 h with CpG ODN and/or poly(I:C). Following surface staining of CD11c and MHC class II cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences, San Jose, CA) and then incubated with antibodies against IL-12p70/40 (BD Biosciences, San Jose, CA). Sample data were acquired on a FACSCalibur or LSR II (BD, Sunnyvale, CA) and analyzed with FlowJo software (TreeStar Inc, Ashland, OR). IL12, IL-6, and IL-10 protein levels were quantified by ELISA as described previously (Verthelyi et al., 2001).
RAW 264.7 cells were cultured for 5 days in DMEM supplemented with 10% FCS, 50 U/ml penicillin, 50 ug/ml streptomycin, and 1mM sodium pyruvate at 37C C in a 5% CO2 in air incubator. Cells were washed, transferred into new flasks, and rested for 16 h before being stimulated with 3.2 ug/ml CpG ODN and/or 32 ug/ml poly(I:C) for 4 - 12 h. These ligand concentrations were identified in preliminary dose-ranging experiments as representing the minimal concentration of each ligand capable of inducing optimal RAW cell stimulation (data not shown). All experiments were independently repeated three times, and data from all experiments combined for analysis.
Total RNA was extracted from resting and stimulated RAW 264.7 cells using TRIzol Reagent (Invitrogen, Carlsbad, CA) as specified by the manufacturer. The RNA was quantified using an ND-1000 (NanoDrop Tecnologies, Wilmington, DE) and qualified for the absence of degradation by electrophoresis through a 1% agarose denaturing gel. Twenty micrograms of total RNA was reverse-transcribed using 3 ul 10× first strand buffer (Stratagene, La Jolla, CA), 2 ul (1 ug) oligo(dT)12-18 (Invitron, Carlsbad, CA), 3 ul (150 U) of AffinityScript Reverse Transcriptase, 2 ul 20× aminoallyl-dUTP/dNTP mix, and 3 ul 0.1 M DTT in a final volume of 30 ul at 42° C for 1 h as previously described (Klaschik et al., 2007). In all studies, a reference murine RNA (Stratagene, La Jolla, CA) sample was processed in parallel. Both cDNA's were purified using a MinElute PCR Purification Kit (Qiagen Sciences, Valencia, CA) and labeled with Cy5 (sample cDNA) or Cy3 (universal reference cDNA) as described (Klaschik et al., 2007). The probes were mixed, diluted in 5 ul DMSO plus 1.7 ul of 1 M NaHCO3 and hybridized to 36K mouse 60-mer oligonucleotide array slides (NCI Mouse Array Set Mm-MEEBO-v1.3) at 42CC for 18 h in a MAUI hybridization system (BioMicro Systems, Salt Lake City, UT) followed by washing, centrifugation and air drying.
Arrays were scanned using a GenePix 4000B Scanner (Axon Instruments, Union City, CA), and analyzed using the GenePix Pro 6.0 Software Tool (Axon Instruments/Molecular Devices, Union City, CA) using GAL files provided by the manufacturer at http://nciarray.nci.nih.gov. Data were uploaded to CIT/BIMAS-NCI/CCR Microarray Database (mAdb; http://nciarray.nci.nih.gov) and formatted via export function to BRB Array Tools. Gene expression analysis was performed using BRB Array Tools Version 3.6.0 developed by the NCI Biometric Research Branch (Bethesda, MD). Expression ratios were log base 2 transformed, lowess normalization was used to adjust for differences in labeling intensities and analyses restricted to genes present in >50% of the arrays. Based on this filtering, 30,703 features could be analyzed.
Ingenuity Pathway Analysis (IPA, Ingenuity Systems, Inc) was used to identify the regulatory interactions and functional characteristics of all significantly up-regulated genes. IPA maps each gene using a global molecular network developed from information contained in the Ingenuity Pathways Knowledge Base. Gene networks were generated algorithmically based on their connectivity in terms of expression, activation, transcription, and/or inhibition. A ‘network’ in IPA is defined as a graphical representation of the molecular relationships between genes. Genes are represented as nodes, and the biological relationship between nodes is shown by a connecting line. All connections are supported by published data stored in the Ingenuity Pathways Knowledge Base or PubMed. IPA ranks all genes based on their connectivity, using a generalization of the concept of node degree, which measures the number of single genes to which a gene is connected. (see https://analysis.ingenuity.com/pa/info/help/Ingenuity_Network_Algorithm_Whitepaper_FINAL(2).pdf; and (Calvano et al., 2005) for details). Genes identified by IPA as influencing the expression of >10% of the genes up-regulated at any given time point were defined as “regulators”.
Microarray data were deposited in NCBI's Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo) and are accessible through GEO series number GSE15416.
Total RNA was extracted and first-strand cDNA synthesized as described above for microarray studies. Quantitative-PCR (Q-PCR) analysis was performed using TaqMan probes (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions in 384-well microtiter plates (Applied Biosystems, Foster City, CA) in a final volume of 10 ul. Specific primers for Q-PCR of GAPDH and IL33 genes were purchased from Applied Biosystems (Foster City, CA). Thermocycling conditions using an Applied Biosystems ABI-7900 SDS were as follows; 50CC for 2 minutes, 95CC for 10 minutes and 40 cycles of 95CC for 10 seconds and 60CC for 1 minute. mRNA levels were quantified by TaqMan PCR. Relative mRNA levels were then determined using the comparative cycle threshold method. This approach calculates the amount of target mRNA by normalization to an endogenous reference gene (GAPDH). Microarray studies showed that GAPDH mRNA levels did not differ significantly between stimulated vs unstimulated cells (data not shown). mRNA levels in stimulated cells were then compared to those of non-stimulated cells.
The gene expression profile of treatment and control groups was compared using a random variance t-test. The random-variance t-test permits sharing information among genes about within-class variation without assuming that all genes have the same variance (Wright and Simon, 2003). Genes were considered significant if their p value was <0.000001. Differences between groups were measured using Student's t test.
RAW 264.7 macrophages were stimulated with CpG ODN and/or poly (I:C). mRNA was isolated from these cells after 4 - 12 h of culture, converted in Cy5-labeled cDNA, and hybridized to murine microarrays in the presence of a Cy3-labeled cDNA reference standard (facilitating inter-sample comparison). Reproducible patterns of gene up-regulation were observed in similarly treated cells from three independent experiments (R2 = 0.941 ± 0.02).
Data from all experiments were combined, and genes whose level of expression increased significantly (p <10-6) identified. RAW 264.7 cells stimulated with CpG ODN activated 190 genes at 4 h and 105 genes at 12 h (Table I). Poly (I:C) treatment led to the stimulation of 574 genes at 4 h and 1,051 at 12 h. Thus, the number of genes triggered by poly (I:C) significantly exceeded the number stimulated by CpG ODN at both time points (p <.01, Table I).
Additional analysis focused on those genes most strongly stimulated at each time point by each treatment. Whereas 97% of the genes most strongly activated by CpG ODN were also stimulated by poly (I:C), only 66% of the genes most strongly activated by poly (I:C) were triggered by CpG ODN (Table II). The magnitude by which the most strongly up-regulated genes was activated also differed significantly between groups, with poly (I:C) triggering a 37.2 ± 3.1 fold increase in mRNA level vs only 15.4 ± 2.8 fold in the CpG ODN treated group (p. <.001). These findings are consistent with the observation that poly (I:C) induced broader gene up-regulation than CpG ODN.
Ingenuity Pathway Analysis (IPA) was used to identify the regulatory pathways underlying the patterned changes in gene expression induced by CpG ODN and/or poly (I:C) stimulation. Results show that TNF was the dominant regulator of gene expression at both 4 and 12 h in RAW 264.7 cultures stimulated with CpG ODN (Table III). IFNg and IL1B were also important regulators of CpG-driven macrophage activation, as together with TNF they impacted the expression of ≈95% of all genes at both time points (Table III). Several other genes, such as IL1A and CSF2, had detectable but limited effects on the CpG-induced response of these cells (Table III).
The pattern of gene regulation observed following poly (I:C) stimulation of RAW 264.7 cells differed from that induced by CpG ODN. Poly (I:C) significantly increased mRNA levels for many type I interferons, an effect that was absent from CpG ODN stimulated cultures. TNF and IFNG were the strongest regulators of poly (I:C) induced gene activation (Table IV), together accounting for 84% of the up-regulated network. Regulators other than IFNG and TNF uniquely accounted for the stimulation of ≈10% of the poly (I:C) induced network (data not shown). Of interest, nearly half of the genes strongly stimulated by poly (I:C) but not CpG ODN were controlled by IFNb1, suggesting that this gene has a greater impact on the breadth and magnitude of the poly (I:C) response than other minor regulators (Table II and data not shown). The fold rise in mRNA level of genes stimulated by multiple regulators significantly exceeded that of genes triggered by a single regulator (p = .033), suggesting that these regulatory networks interacted synergistically. TNF continued to dominate poly (I:C) driven gene regulation at 12 h whereas the role of IFNG declined (impacting only 41% of genes at 12 h vs 58% at 4 h, Table IV).
IPA was then used to examine the function of the genes up-regulated by CpG ODN and poly (I:C). Despite differences in the regulatory patterns noted above, both ligands activated genes in similar functional categories, specifically immune response, immune system development and cell death (Table V).
Bone marrow derived macrophages and dendritic cells (BMDC) express both TLR9 and TRL3, and thus can be stimulated by both CpG ODN and poly (I:C). Preliminary experiments were conducted to confirm previous reports showing that co-stimulation by these two ligands synergistically enhanced immune activation. These cells were cultured in vitro with CpG ODN and/or poly (I:C) for 20 h and analyzed for the production of IL-12, IL-6, and/or IL-10. As seen in Fig 1, the production of all three cytokines was synergistically enhanced (p < 0.01).
RAW cells stimulated simultaneously with CpG ODN plus poly (I:C) up-regulated significantly more genes than cells treated with either ligand alone at 4 h (p<0.001, Table I). The gene networks activated by the ligand combination included elements of the regulatory networks triggered by each ligand independently, and mirrored the extent of that activation (Table III). Of interest, IL-6 was identified as contributing to the regulation of gene expression triggered by the combination of CpG ODN plus poly (I:C) at 4 h, yet this regulator was not induced by either CpG ODN or poly (I:C) alone at that time point. IL-6 was up-regulated by poly (I:C) at 12 h, suggesting that the synergy observed when RAW cells were co-stimulated via TLR3 plus TLR9 reflects an acceleration of the gene expression profile induced by each ligand independently.
Co-stimulation with CpG ODN plus poly (I:C) also resulted in a significant increase in the magnitude of gene activation. Restricting the analysis to those genes up-regulated by the combination of CpG plus poly (I:C) at 4 h, synergistic stimulation increased mRNA level by 5.9-fold on average versus only 4.2- and 2.7-fold for the same set of genes when stimulated by poly (I:C) or CpG ODN alone, respectively (p<0.001). Having found in Table III that mRNA levels increased significantly when genes stimulated by multiple regulators (p = 0.003), this outcome may reflect the contribution of IL-6 to the other regulators activated by each ligand. At 12 h, there were fewer genes up-regulated in cultures stimulated with the combination of CpG ODN plus poly (I:C) than with poly (I:C) alone. This is consistent with the decrease in gene activation mediated by CpG ODN alone at the same time point.
Evidence that the combination of CpG ODN plus poly (I:C) synergistically enhanced the secretion of certain cytokines led us to examine whether this ligand combination had broader synergistic effects on mRNA expression. To insure that inter-experimental variability in mRNA levels was not misinterpreted as evidence of synergy, only those genes whose level of expression when treated with CpG ODN plus poly (I:C) exceeded the mean + 3 SD of the sum induced by each ligand individually were included in this analysis. By this criterion, 60 genes were identified as being synergistically up-regulated by CpG ODN plus poly (I:C). IPA analysis of these genes showed that 80% fell into the “immune response” category and 22 encoded proteins of immunological interest (cytokines or chemokines, Table VI).
At 4 h, the synergistically up-regulated genes included those encoding IL-1A, CSF3, PTGS2, several type I IFN's and the chemokine CCL17. At 12 h, the synergistically up-regulated group included the interleukin genes IL33, IL12A, IL12B, IL10, IL19, and IL1F6, the adipokine genes LCN2 and SAA3, the neuropeptide-encoding genes NPY and NTS, and the chemokine genes CCL7 and S100A8. IL6 and TSLP were the only 2 cytokines/chemokines synergistically up-regulated at both 4 and 12 h. Of interest, neither CpG ODN nor poly (I:C) alone effected IL33, IL19, IL12A, S100A8, DUSP1, TIMP1, SCN9A, CCL17, HAMP, MARCKS, PFKFB3, CSN1S1, VAV2 or NPY expression, yet all of these genes were significantly up-regulated by this ligand combination.
RT-PCR was performed to confirm that the increases in cytokine mRNA levels identified by microarray were valid. As seen in Fig 2, the concentration of IL-33 mRNA increased 60 fold after 12 h of stimulation with CpG ODN plus poly (I:C), consistent with the 46 fold increase predicted by the microarray data (Fig 2 vs Table VI). This is the first evidence of a synergistic increase in IL-33 after dual TLR stimulation.
Microarrays facilitate the evaluation of global changes in gene expression induced by immune stimuli (Huang et al., 2006; Jenner and Young, 2005; Klaschik et al., 2008; Klaschik et al., 2007). By combining microarray with IPA analysis, this work examined the effect of stimulating the RAW 264.7 macrophage cell line via TLR9 and/or TLR3. Results indicate that the pattern of signaling triggered via these distinct TLR receptors are broadly similar but include reproducible differences. In general, poly (I:C) stimulation via TLR3 activated a larger and broader regulatory network than CpG stimulation of TLR9. Gene activation was accelerated when poly (I:C) was co-administered with CpG ODN, and triggered the synergistic up-regulation of many immune-related genes including a subset that was not activated by either ligand alone.
Simultaneous exposure to TLR9 and TLR3 ligands is likely to occur under physiologic conditions. For example, an infected host would recognize the CpG motifs expressed by adenovirus, herpesvirus and other DNA viruses via TLR9 while the double-stranded stem-loop RNA structures produced during the replication of the same viruses would be recognized via TLR3 (Jenner and Young, 2005; Sorensen et al., 2008; Tabeta et al., 2004; Wells et al., 2008).
RAW 264.7 cells were selected for this study because they respond to stimulation via both TLR3 and TLR9 and can be expanded to yield a large and uniform macrophage population. The genes triggered by key regulators following CpG-driven activation of RAW cells in the current study represented a consistent subset of those detected in an earlier study of CpG-stimulated whole spleen cells (Klaschik et al., 2007). In both cases TNF, IFNG, and IL1B contributed to global changes in gene expression. These observations suggest that i) data generated from the study of RAW cells appears to be reliable and ii) while macrophages comprise only a small fraction of the mixed cell population present in whole spleen (Johansson et al., 2006) the regulatory elements that dominate CpG mediated signaling are conserved among cell types. However, several “minor regulators” were induced by CpG ODN treatment of whole spleen that were not present after the stimulation of RAW cells, including NFKB1, IL6, IL12B, MYC, and STAT1. As mouse spleen contains multiple cell types capable of responding to CpG stimulation, these differences may reflect cell-specific patterns of gene activation, the impact of cell-cell interactions on global mRNA expression, or underlying differences in the behavior of RAW cells vs resting macrophages. Ongoing studies of a B cell line are consistent with the conclusion that individual phenotypes contribute to the overall pattern of CpG-induced gene regulation.
Poly (I:C) treatment resulted in the up-regulation of nearly 3-fold more genes at 4 h and 10-fold more at 12 h than did CpG ODN. Several notable differences in the pattern of gene activation induced by these distinct TLR ligands were identified. The CpG-mediated response was primarily regulated by TNF, IFNG and IL1B. IL1B played a lessor role in the regulation of the poly (I:C) dependent response while a number of other inducers (including IFNB1, IFNA2, IL15, IL18, and IL3) played more important roles. The poly (I:C) response was more complex and persistent than that elicited by CpG ODN. At 12 h, the number of genes up-regulated by CpG ODN fell by nearly half whereas the number of poly (I:C) stimulated genes doubled over the same period. The temporal pattern of gene regulation elicited by CpG treatment of RAW cells was entirely consistent with that observed in studies of whole spleen cells in vivo (Klaschik et al., 2008). Analysis of those genes most highly up-regulated by each treatment (Table II) also showed that components of the poly (I:C) induced response were not activated by CpG ODN. This effect could be largely attributed to the impact of IFNB1 on the breadth and magnitude of the poly (I:C) response. Despite these differences, both CpG ODN and poly (I:C) triggered genes within the same broad functional groups: immune response, immune system development and cell death (Table V), consistent with both contributing to the induction of an innate immune response.
Co-administering poly (I:C) with CpG ODN led to the activation of more genes than either ligand alone at 4 h. The regulatory networks triggered when RAW cells were treated with CpG ODN plus poly (I:C) primarily reflected the combined effect of each ligand individually. Thus, IFNB1, IL18, IFNA2 and IL3 were stimulated by poly (I:C) but not CpG ODN, CSF2 was stimulated by CpG ODN but not poly (I:C), yet all were stimulated by combination treatment. IL6 was unique in being activated by combination treatment at 4 h but neither ligand alone until 12 h. In addition, the magnitude of activation (fold increase in mRNA level) of those genes stimulated by each ligand separately was significantly lower than the same genes stimulated by the combination of CpG ODN plus poly (I:C). These findings suggest that exposure to both ligands magnified and accelerated the response elicited by each individual ligand.
Previous studies show that certain PAMP combinations synergistically enhance cytokine secretion ((Napolitani et al., 2005; Zhu et al., 2008) and Fig 1). New insight into the extent of this synergy was gained by monitoring global changes in gene expression by microarray. We found that the expression level of 60 genes was significantly elevated following treatment with CpG ODN plus poly (I:C) vs either ligand alone. Studies of protein production by ELISA and mRNA levels by Q-PCR confirmed the synergistic activation of members of this group (including IL12a, IL12b, IL6, IL10 and IL-33, Figs 1, ,22 and Table VI (Napolitani et al., 2005; Zhu et al., 2008). IPA analysis showed that 87% of the synergistically activated genes were immune related. Indeed, over one-third of all synergistically up-regulated genes, including the most highly up-regulated, were cytokines or chemokines (Table VI). These findings support the hypothesis that stimulation via multiple PAMPs synergistically broadens and increases the magnitude of the host's protective immune response.
The mechanism(s) underlying the synergistic up-regulation of immune genes by PAMP combinations is of interest. Napolitani et al. concluded that TLR synergy occurred primarily at late time points and involved only a small number of genes (Napolitani et al., 2005). They hypothesized that synergy resulted from a prolongation of signaling due to dual TLR engagement. Our results suggest that synergy is instead characterized by the early and persistent activation of one or more regulatory pathways, resulting in a prolonged signaling cascade that increases, magnifies, and diversifies gene expression. Our work establishes that the pattern of gene regulation triggered by the interaction of CpG ODN with TLR9 and poly (I:C) with TLR3 are distinct despite sharing several features in common. The level of gene activation triggered by poly (I:C) exceeds that induced by CpG ODN. The resulting expansion in gene activation reflects both the acceleration and prolongation of gene up-regulation, processes that are initiated shortly after ligand administration.
Analyses were preformed using BRB Array Tools developed by Richard Simon and Amy Peng Lam. The assertions herein are the private ones of the authors and are not to be construed as official or as reflecting the views of the NCI at large.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.