Differential regulation of lipopolysaccharide and Gram-positive bacteria induced cytokine and chemokine production in macrophages by Gα_i proteins

Abstract

Heterotrimeric G_i proteins play a role in signalling activated by lipopolysaccharide (LPS), Staphylococcus aureus (SA) and group B streptococci (GBS), leading to production of inflammatory mediators. We hypothesized that genetic deletion of G_i proteins would alter cytokine and chemokine production induced by LPS, SA and GBS stimulation. LPS-induced, heat-killed SA-induced and heat-killed GBS-induced cytokine and chemokine production in peritoneal macrophages from wild-type (WT), Gα_i2^–/– or Gα_i1/3^–/– mice were investigated. LPS induced production of tumour necrosis factor-α (TNF-α), interleukin-6 (IL-6), IL-10 and interferon-γ-inducible protein-10 (IP-10); SA induced TNF-α, and IL-1β production; and GBS induced TNF-α, IL-6, IL-1β, macrophage inflammatory protein-1α (MIP-1α) and keratinocyte chemoattract (KC) production were all decreased (P < 0·05) in Gα_i2^–/– or Gα_i1/3^–/– mice compared with WT mice. In contrast to the role of G_i proteins as a positive regulator of mediators, LPS-induced production of MIP-1α and granulocyte–macrophage colony-stimulating factor (GM-CSF) were increased in macrophages from Gα_i1/3^–/– mice, and SA-induced MIP-1α production was increased in both groups of Gα_i protein-depleted mice. LPS-induced production of KC and IL-1β, SA-induced production of GM-CSF, KC and IP-10, and GBS-induced production of IL-10, GM-CSF and IP-10 were unchanged in macrophages from Gα_i2^–/– or Gα_i1/3^–/– mice compared with WT mice. These data suggest that G_i2 and G_i1/3 proteins are both involved and differentially regulate murine inflammatory cytokine and chemokine production in response to both LPS and Gram-positive microbial stimuli.

Introduction

Sepsis is a systemic response to bacterial, fungal, or viral infections that may result in septic shock, multiple organ failure and death.¹ Lipopolysaccharide (LPS) from Gram-negative bacteria, and peptidoglycan, lipoteichoic acid and lipoproteins from Gram-positive bacteria are potent inducers of the pro-inflammatory responses.² The interaction of these microbes and/or their cellular components with macrophages, monocytes, or other host cells induces the release of inflammatory mediators that play a major role in the pathophysiology of septic shock.³

LPS-induced signalling is mediated by Toll-like receptor 4 (TLR4) coupled with CD14 and MD-2.⁴ The Gram-positive bacteria Staphylococcus aureus (SA)-induced signalling pathways are mediated, in part, through TLR2 and other receptors.^5–8 Group B streptococci (GBS)-induced tumour necrosis factor-α (TNF-α) production is mediated by myeloid differentiation protein 88 (MyD88), suggesting that TLRs also mediate GBS signalling.⁹ Stimulation of TLR2 or TLR4 and MyD88 dependent and independent signalling pathways results in the activation of a series of signalling proteins leading to expression of pro- and anti-inflammatory cytokine and chemokine genes. The cytokines include the pro-inflammatory mediators TNF-α, interleukin-1β (IL-1β), IL-6, IL-12 and interferon-γ (IFN-γ) and the anti-inflammatory cytokine IL-10.¹⁰ The CXC chemokine family includes KC (the murine homologue of IL-8), IP-10 and monokine induced by IFN-γ (MIG). The CC chemokine family includes monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein-1α (MIP-1α).^11,12 These cytokines and chemokines contribute to the complex inflammatory milieu of sepsis. However, it remains uncertain to what extent common or discrete signalling pathways regulate cytokine and chemokine production.

Heterotrimeric guanine nucleotide binding regulatory (G) proteins of the G inhibitory class (G_i) are involved in LPS signalling. It has been shown that Gα_i proteins are coimmunoprecipitated with the CD14 receptor.¹³ Molecular approaches have demonstrated that TLR4-induced extracellular signal-regulated kinase 1/2 activation was blocked by dominant negative G_i protein constructs in HEK293 cells.¹⁴ Mastoparan, a G_i protein agonist/antagonist, suppressed LPS-induced and SA-induced cytokine production in human monocytes.^13,15,16 Studies using pertussis toxin, which inhibits receptor-G_i coupling, have demonstrated inhibition of LPS-induced mediator production in different cell types.^17–19 We also demonstrated that pertussis toxin inhibited not only LPS-induced, but also SA- and GBS-induced TNF-α production in murine J774 and human THP-1 cell lines.²⁰ The latter observations suggest that one or more G_i proteins serve as common post-receptor-signalling proteins for both TLR4 and TLR2 ligands. Our previous studies have demonstrated that LPS-, SA- and GBS-induced TNF-α, and thromboxane B₂ production were decreased in peritoneal macrophages from Gα_i2^–/– mice compared with wild-type (WT) mice.²¹ The potential role of Gα_i protein in the regulation of other important cytokines and chemokines induced by LPS and Gram-positive bacteria remains to be investigated.

To further examine potential cellular phenotype and isoform specificity of Gα_i proteins in cytokine and chemokine expression, the effects of Gα_i gene deletion on chemokine and pro-inflammatory and anti-inflammatory cytokine production in peritoneal macrophages following stimulation with the TLR ligands LPS, SA and GBS were investigated. Genetic deletion of Gα_i2 and Gα_i1/3 in mice and luminex analysis offered a unique opportunity to investigate the role of Gα_i proteins in LPS-, SA- and GBS-induced cytokine and chemokine release. We hypothesized that genetic deletion of specific G_i protein isoforms differentially altered cytokine and chemokine production in peritoneal macrophage stimulated by LPS, SA and GBS.

Materials and methods

Mice

Gα_i2^–/– mice and littermate WT mice with 129Sv background were generated by breeding heterozygous and homozygous knockout mice. Gα_i1/3^–/– mice with 129Sv background were generated by breeding homozygous double knockout mice. Studies employed 5- to 8-week-old Gα_i2^–/–, Gα_i1/3^–/– and age-matched WT mice for all the experiments. The original knockout mice were obtained from Dr Lutz Birnbaumer (NIH, Research Triangle Park, NC). The investigations conformed to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health and had the approval of the institutional animal care and use committee.

Genotyping

Polymerase chain reaction (PCR) was performed with genomic DNA from the tails of 4-week-old Gα_i2^–/–, mice. We used the following primer pairs: WT (+), forward, 5′-GAT CAT CCA TGA AGA TGG CTA CTC AGA AG-3′; reverse, 5′-CCC CTC TCA CTC TTG ATT TCC TAC TGA CAC-3′. Knockout (–), forward, 5′-CAG GAT CAT CCA TGA AGA TGG CTA C-3′; reverse, 5′-GCA CTC AAA CCG AGG ACT TAC AGA AC-3′. The reactions were run for 35 cycles. Amplified sequences were 805 base pairs for the WT allele and 509 base pairs for the targeting construct.

Cell stimulation

Peritoneal macrophages were harvested from Gα_i2^–/–, mice, Gα_i1/3^–/– mice and littermate WT mice as described previously²¹ using peritoneal lavages with 10 ml of ice-cold RPMI-1640 medium. The cells were maintained in RPMI-1640 medium (Cellgro Mediatech Inc., Herndon, VA), supplemented with heat-inactivated 1% fetal calf serum (Sigma, St Louis, MO), 50 U/ml penicillin, 50 μg/ml streptomycin (Cellgro Mediatech Inc.) at 37° for 2 hr. Non-adherent cells were washed off. Typically, more than 95% of macrophages were obtained as examined by stain with fluorescein isothiocyanate-conjugated anti-mouse F4/80 antibody (Caltag, Burlingame, CA) and analysed by flow cytometry. Peritoneal macrophages were stimulated with 10 μg/ml LPS (from Salmonella enteritidis, Sigma), or heat-killed SA or GBS (heat-killed SA and GBS were prepared as described previously²²), and protein-free Salmonella enterica minnesotaR595 LPS (provided by Dr Ernst Reitschel, Borstel Germany, 1 μg/ml) for 18 hr. The concentration of agonist and the time of stimulation were selected based on our previous time–course and dose–response studies.²¹ The supernatants were collected for assay of mediator production.

Luminex analysis

The samples were stored in liquid nitrogen until assayed. Cytokine levels were assayed with a Biosource murine 10-plex cytokine and 6-plex chemokine assay (Biosource, Camarillo, CA) on a Luminex 100 instrument. Specifically, we assayed the sample for IL-1β (sensitivity 8·7 pg/ml), IL-4 (sensitivity 5 pg/ml), IL-5 (sensitivity 6·5 pg/ml), IL-6 (sensitivity 2·7 pg/ml), IL-10 (sensitivity 35 pg/ml), IL-12(p40/p70) (sensitivity 11 pg/ml), IL-17 (sensitivity 5 pg/ml), TNF-α (sensitivity 4·5 pg/ml), granulocyte–macrophage colony-stimulating factor (GM-CSF; sensitivity 10 pg/ml), MIP-1α (sensitivity 15·2 pg/ml), MIG (sensitivity 3 pg/ml), IP-10 (sensitivity 46 pg/ml) and KC (sensitivity 11·5 pg/ml). Cytokine levels were established by comparison to a standard curve as per the manufacturer's instructions.

Statistical analysis

Data are expressed as the mean ± standard error of the mean. Statistical significance was determined by analysis of variance (anova) with Fisher's probable least-squares difference test using Statview software (SAS Institute Inc., Cary, NC). A value of P < 0·05 was used to reject the null hypothesis.

Results

Gα_i2 and Gα_i1/3 differentially regulate cytokine production induced by microbial stimuli in macrophages

We determined the effect of deletion of Gα_i2 or Gα_i1/3 on LPS-, SA- and GBS-induced cytokine production in macrophages from Gα_i2^–/– or Gα_i1/3^–/– mice. Macrophages from Gα_i2^–/– and Gα_i1/3^–/– and age-matched WT mice were stimulated in vitro with LPS, SA or GBS for 18 hr after which the media were subjected to luminex analysis for TNF-α, GM-CSF, IL-1β, IL-4, IL-5, IL-6, IL-10, IL-12 and IL-17. LPS-, SA- and GBS-induced TNF-α production was decreased in macrophages from Gα_i2^–/– mice (42 ± 13%, 69 ± 11%, and 39 ± 7%, respectively, n = 3, P < 0·05) in comparison with WT mice (Fig. 1a). Similarly, LPS-, SA- and GBS-induced TNF-α production was also decreased in macrophages from Gα_i1/3^–/– mice (51 ± 10%, 59 ± 13%, and 54 ± 14%, respectively, n = 3, P < 0·05) compared with WT mice (Fig. 1a). LPS- and GBS-induced IL-6 production was decreased in both Gα_i2^–/– and Gα_i1/3^–/– mice compared to WT mice (46 ± 13%, 51 ± 5% and 46 ±17%, 45 ± 21%, respectively, n = 3, P < 0·05) (Fig. 1b). SA-induced IL-6 production was only decreased in Gα_i1/3^–/– mice (84 ± 9%, n = 3, P < 0·05) (Fig. 1b). LPS-induced IL-10 production was decreased in both Gα_i2^–/– and Gα_i1/3^–/– mice (64 ± 12% and 56 ± 8%, respectively, n = 3, P < 0·05). SA-induced IL-10 production was decreased in both Gα_i2^–/– and Gα_i1/3^–/– mice (73 ± 4% and 86 ± 3%, respectively, n = 3, P < 0·05) (Fig. 1c). LPS-induced IL-1β production was unchanged in Gα_i2^–/– and Gα_i1/3^–/– mice. However, SA- and GBS-induced IL-1β were decreased in both Gα_i2^–/– and Gα_i1/3^–/– mice compared to WT mice (7 ± 4%, 15 ± 3% and 27 ± 2%, 21 ± 0%, respectively, n = 3, P < 0·05) (Fig. 1d). LPS-induced GM-CSF production was increased in Gα_i1/3^–/– mice (fold increase 3·0 ± 0·7; n = 3, P < 0·05), while SA- and GBS-induced GM-CSF production was unchanged in both Gα_i2^–/– and Gα_i1/3^–/– mice compared to WT mice (Fig. 1e). None of LPS, SA and GBS stimulated detectable IL-4, IL-5, IL-12 or IL-17 production in macrophages (data not shown). Stimulation of TNF-α of Gα_i2^–/– and WT macrophages was repeated with protein-free S. minnesota R595 LPS. Similar results were found with these TLR4 or TLR2 ligands as with Salmonella enteritidis LPS and SA (data not shown).

Figure 1.

Effect of deletion of Gα_i2 or Gα_i1/3 genes on LPS-, SA- and GBS-induced TNF-α, IL-6, IL-10, IL-1β, and GM-CSF production in macrophages. Peritoneal macrophages were isolated from Gα_i2^–/– or Gα_i1/3^–/– and age-matched 129sv WT mice and stimulated in vitro with LPS, SA or GBS (10 μg/ml). LPS-, SA- and GBS-induced TNF-α (a), IL-6 (b), IL-10 (c), IL-1β (d), and GM-CSF (e) production were determined. Data represent means ± SE from three independent experiments. *P < 0·05 compared to basal; ^#P < 0·05 compared to stimulated WT group; ND, not detectable.

Gα_i2 and Gα_i1/3 differentially regulate chemokine production induced by microbial stimuli in macrophages

We sought to determine the effect of deletion of Gα_i2 or Gα_i1/3gene on chemokine production in response to LPS, SA and GBS stimulation. Macrophages from Gα_i2^–/–, Gα_i1/3^–/– and age-matched WT mice were stimulated in vitro with LPS, SA or GBS for 18 hr. Luminex analysis was performed for MIP-1α, KC and IP-10 and MIG production. LPS induction of the CC chemokine MIP-1α was increased in the macrophages isolated from Gα_i1/3^–/– mice compared to WT mice (fold increase 7·1 ± 2·9 n = 3, P < 0·05). SA-induced MIP-1α production was significantly increased in the macrophages isolated from Gα_i2^–/– and Gα_i1/3^–/– mice compared to WT mice (fold increases 2·4 ± 0·4 and 5·9 ± 1·5, respectively, n = 3, P < 0·05). In contrast, GBS-induced MIP-1α production was significantly suppressed in both groups of Gα_i-deficient mice (Fig. 2a). LPS significantly induced the production of CXC chemokines KC and IP-10. However, LPS-induced KC was unchanged (Fig. 2b) and IP-10 was decreased in the Gα_i2^–/– and Gα_i1/3^–/– mice compared to WT mice(54 ± 12% and 86 ± 0·2%, respectively, n = 3, P < 0·05) (Fig. 2c). SA-induced production of KC and IP-10 was unchanged. GBS-induced KC production was decreased in the Gα_i2- and Gα_i1/3-deficient mice (45 ± 20% and 48 ± 21%, respectively, n = 3, P < 0·05) whereas GBS induced no stimulation on IP-10 production (Fig. 2b,c). Not LPS, SA nor GBS stimulated MIG production in macrophages (data not shown).

Figure 2.

Effect of deletion of Gα_i2 or Gα_i1/3 gene on LPS-, SA- and GBS-induced MIP-1α, KC and IP-10 production in macrophages. Peritoneal macrophages were isolated from Gα_i2^–/– or Gα_i1/3^–/– and age-matched 129sv WT mice and stimulated in vitro with LPS, SA or GBS (10 μg/ml). LPS-, SA- and GBS-induced MIP-1α (a), KC (b) and IP-10 (c) production was studied. Data represent means ± SE from three independent experiments. *P < 0·05 compared to basal; ^#P < 0·05 compared to stimulated WT group.

Discussion

Our studies demonstrate that Gα_i proteins differentially regulate cytokine and chemokine production in murine peritoneal macrophages in response to microbial stimuli. TNF-α induced by LPS, SA and GBS; IL-6 induced by LPS and GBS; IL-10 and IP-10 induced by LPS; IL-1β induced by SA; and IL-1β, MIP-1α and KC induced by GBS were all decreased in macrophages from Gα_i2^–/– and Gα_i1/3^–/– mice compared with WT mice. However, LPS-induced production of GM-CSF and MIP-1α was increased only in the macrophages from Gα_i1/3^–/– mice whereas SA-induced MIP-1α production was increased in macrophages from both groups of Gα_i-deficient mice. Some cytokines and chemokines induced by the TLR ligands were Gα_i independent. LPS-induced production of IL-1β, KC, SA-induced GM-CSF, KC and IP-10 production, and GBS-induced IL-10 and GM-CSF production were all unchanged in macrophages from Gα_i2^–/– or Gα_i1/3^–/– mice compared with WT mice. Collectively, the data suggest that G_i proteins play both positive and negative roles in the production of specific cytokines and chemokines. The postulated roles of Gα_i protein regulation of cytokines and chemokines induced by LPS, SA and GBS in peritoneal macrophages are listed in Tables 1 and 2.

Table 1.

Postulated Gα_i2 and Gα_i1/3 regulation of LPS-, SA- and GBS-induced cytokine¹ production in peritoneal macrophages

Stimuli	G protein	TNF-α	IL-6	IL-10	IL-1β	GM-CSF
LPS	Gαi2	+	+	+	=	=
SA	Gαi2	+	=	+	+	+
GBS	Gαi2	+	+	=	+	=
LPS	Gαi1/3	+	+	+	=	−
SA	Gαi1/3	+	+	+	+	+
GBS	Gαi1/3	+	+	=	+	=

¹Based upon results with peritoneal macrophages from Gα_i2^–/– and Gα_i1/3^–/– mice.

+, mediates increased production; −, negatively modulates production; =, does not participate in production.

Table 2.

Postulated Gα_i2 and Gα_i1/3 regulation of LPS-, SA- and GBS-induced chemokine production in peritoneal macrophages

Stimuli	G protein	MIP-1α	KC	IP-10
LPS	Gαi2	=	=	+
SA	Gαi2	−	=	=
GBS	Gαi2	+	+	=
LPS	Gαi1/3	−	=	+
SA	Gαi1/3	−	=	=
GBS	Gαi1/3	=	+	=

+, mediates increased production; −, negatively modulates production; =, does not participate in production.

The finding that mice deficient in Gα_i proteins exhibited predominantly decreased LPS-, SA- and GBS-stimulated TNF-α production in peritoneal macrophages is consistent with our previous studies that pretreatment with pertussis toxin, an inhibitor of G_i proteins, inhibited LPS-, SA- and GBS-induced TNF-α production in J774 cells and THP-1 cells.²⁰ Adib-Conquy et al. demonstrated that heat-killed SA-induced IL-10 production in whole blood samples from trauma patients was suppressed by pertussis toxin pretreatment.²³ In human monocytes, the G_i protein agonist/antagonist mastoparan also inhibited LPS-induced cytokine production and signalling.^13,16 The latter data are similar to the observation of Bocker et al. who observed that mastoparan inhibited SA-induced IL-10, IL-12 and IL-18 production in peripheral blood mononuclear cells.¹⁵ In general, previous studies suggest that microbial factor activation of G_i proteins up-regulates cytokine and chemokine production in macrophages and monocytes. In addition to peritoneal macrophages, G_i protein signalling may also up-regulate cytokine production in B cells, because SA-induced IL-10 production is suppressed in B cells from Gα_i2^–/– mice.²⁴

Although most cytokines and chemokines were suppressed following LPS, SA and GBS stimulation in peritoneal macrophages from G_i-deficient mice, there were notable exceptions. In particular, MIP-1α production was augmented in response to LPS stimulation in Gα_i2^–/– macrophages and in response to SA in both Gα_i2^–/– and Gα_i1/3^–/– macrophages. LPS stimulation of Gα_i1/3^–/– macrophages also augmented GM-CSF relative to WT macrophages. In contrast, although LPS and SA were robust inducers of KC production, a deficiency of either Gα_i2 or Gα_i1/3 had no effect. Our data also suggest that in general Gα_i2 and Gα_i3 proteins play similar roles in the regulation of LPS and SA gene expression in macrophages. However, Gα_i1/3 deficiency had a greater effect. The data suggest that Gα_i3 proteins play a predominant role in the regulation of SA-induced IL-6 and IL-10 production, and of LPS-induced GM-CSF and MIP-1α production. These findings underscore the multiplicity of Gα_i protein functions in regulating cytokine and chemokine production. In addition to G_i isoform specificity, stimulus specificity was also seen in the cytokine/chemokine profile. SA was a more potent inducer of MIP-1α, which is consistent with the findings of Parker et al.²⁵ who demonstrated that the TLR2 ligand Pam(3) CSK(4) was a more potent inducer of MIP-1α than LPS. Stimulus specificity is also demonstrated by the finding that LPS, but not SA, induced the chemokine IP-10.²⁶ Other receptors such as CD36 and MARCO also interact with SA and mediate SA signalling, suggesting that SA activates other signalling besides TLR2.^5,6 TLR4 signalling can be induced by MyD88-independent signalling leading to activation of IFN regulatory factor-3, which stimulates expression of IFN-inducible genes including IP-10.²⁷ This MyD88-independent signalling pathway is not present in the TLR2 signalling cascade. These findings add to a growing list of cytokines and chemokines differentially induced by TLR2 and TLR4 ligands.^28,29

Although GBS shared common features with LPS and SA, e.g. suppressed TNF-α, IL-6 and IL-1β in macrophage from Gα_i-deficient mice, there are notable exceptions. GBS-induced MIP-1α was inhibited rather than augmented in macrophages from G mice and GBS-induced KC production was reduced in both groups of G_i-deficient mice. IL-10 was unaffected by GBS. Also there were fewer Gα_i isoform-specific effects that were evident in response to SA and LPS. These distinguishing features of GBS may be a consequence of different receptor/signalling pathways activated by GBS. Two membrane proteins, the integrin β₂ CD11b/CD18 (CR3) and CD14 have been suggested as integral parts of the innate immune response to GBS.^22,30 However, CD11b/CD18 has no known inflammatory signalling capabilities.³¹ CD14 is known to interact with TLRs. Heat-killed and washed GBS cell wall identical to the preparation employed in the present study stimulated TNF-α production and was shown to be dependent on MyD88.⁹ However, targeted deletion of the TLR1, -2, -4, -6 or -9 did not alter the GBS-induced cytokine response.^9,32 Thus the identity of the TLR or TLRs for GBS cell wall binding remains uncertain.

In macrophages, the G_i protein appears to mediate predominantly pro-inflammatory gene expression. However, there is a remarkable cellular phenotype specificity since we have shown that splenocytes from G mice exhibit augmented mediator production relative to wild-type mice.²¹ Thus the role of G_i proteins in regulating innate immune responses may be cell/organ specific. G_i proteins typically couple to heptahelical receptors and TLRs are non-heptahelical. One possibility is that the TLR signalling pathway may trans-activate other receptors that are G_i coupled.³³ Of interest in this context, Triantafilou et al.³⁴ proposed that LPS interacts with a cluster of receptors in lipid rafts including G_i-protein-coupled receptors, e.g. CXCR4,³⁵ growth factor receptors³³ and integrin β₂.³⁶ Also, disruption of the cholesterol content of lipid rafts results in loss of expression of Gα_i signalling and induces an anti-inflammatory phenotype.³⁷ Another possibility is that G_i proteins are not coupled directly to TLR or TLR postreceptor signalling proteins but rather to heptahelical receptors that are activated by autocrine mediators.³⁸

These composite findings support a convergent role of G_i proteins in the regulation of many cytokines and chemokines in response to Gram-negative and Gram-positive bacterial stimuli. Also importantly, this study provides evidence of G_i isoform specificity in TLR activation. Since the incidence of Gram-positive sepsis is increasing³⁹ with the increased risk of polymicrobial sepsis, modulation of G_i protein function may afford novel approaches for regulating the Th1/Th2 response. Understanding the role of G_i proteins in the regulation of TLR activation in response to Gram-negative and Gram-positive microbial stimuli will provide important insights into the regulation of innate immunity.

Abbreviations

anova

analysis of variance

GBS

group B streptococci

GM-CSF

granulocyte–macrophage colony-stimulating factor

IFN

interferon

IL

interleukin

IP

IFN-γ-inducible protein

KC

keratinocyte chemoattract

LPS

lipopolysaccharide

MCP-1

monocyte chemoattractant protein-1

MIG

monokine induced by IFN-γ

MIP

macrophage inflammatory protein

MyD88

myeloid differentiation protein 88

SA

Staphylococcus aureus

TLR

Toll-like receptor

TNF

tumour necrosis factor

WT

wild-type