To investigate whether all classes of CpG ODN can induce PAF-AH, we examined the effect of CpG-A₁₅₈₅, -B₁₈₂₆, and -C₂₃₉₅ on induction of paf-ah2 mRNA expression and PAF-AH activity in splenocytes in vitro. Interestingly, we found that only CpG-A₁₅₈₅ significantly induced paf-ah2 mRNA expression compared to other ODNs (p < 0.01) (Figure 1A; Figure S1A in Supplementary Material). Inhibitory ODN H154 (iODN_H154), a TLR9-specific antagonist, significantly suppressed CpG-A₁₅₈₅-induced paf-ah2 mRNA expression and PAF-AH activity (p < 0.01) (Figure 1B). Since CpG-A₁₅₈₅ containing unmethylated CpG dinucleotides triggers the vertebrate immune response through TLR9 activation (23, 24), we examined whether a CpG motif in CpG-A₁₅₈₅ would affect CpG-A₁₅₈₅-induced paf-ah2 mRNA expression using non-CpG ODN of CpG-A₁₅₈₅ (non-CpG-A₁₅₈₅). We found that Crt₁₆₁₂ and non-CpG-A₁₅₈₅ did not induce paf-ah2 mRNA expression (Figure 1C). In addition, CpG-A₁₅₈₅ showed stronger PAF-AH activity compared with control (Ctr₁₆₁₂), whereas pretreatment of iODN_H154 and non-CpG-A₁₅₈₅ showed insignificant PAF-AH activity compared with Ctr₁₆₁₂ (Figure 1D). Similar results were observed in peritoneal macrophages (Figures 1E,F). CpG ODNs, especially CpG-A, are known as IFN-α inducers (12). To investigate whether IFN-α is involved in induction of paf-ah2 mRNA expression, we examined the response of polyinosinic-polycytidylic acid (poly(I:C)) as an IFN-α inducer and recombinant mouse IFN-α (rmIFN-α) on induction of paf-ah2 mRNA expression. Poly (I:C) and rmIFN-α did not induce paf-ah2 mRNA expression (Figures S1B,C in Supplementary Material). Taken together, these data suggest that only CpG-A₁₅₈₅ strongly induces PAF-AH via TLR9 but not IFN-α, and its production is derived from macrophages.

It has been reported that PAF administration alone induces sudden death in mice (20). Moreover, PAF is known to be involved in DIC as a symptom of sepsis (17, 25). Therefore, we examined the ability of CpG-A₁₅₈₅ to protect mice from PAF-induced lethal shock and sudden death. We administered 50, 100, and 300 µg CpG-A₁₅₈₅ intraperitoneally (i.p.) 1, 3, and 5 days before PAF challenge (Figure 2A). CpG-A₁₅₈₅ treatment at the highest dose (300 µg) significantly improved 20-min mortality in all mice compared with those in the PAF group (p < 0.0001), whereas 50 and 100 µg CpG-A₁₅₈₅ treatment did not advance mortality (Figure 2B).

We next applied methyl arachidonyl fluorophosphonate (MAFP) treatment, a PAF-AH inhibitor, to determine whether the effect of CpG-A₁₅₈₅ is dependent on PAF-AH. It has been shown that monocytes and thrombin-stimulated human coronary artery endothelial cells accumulated PAF following decreased PAF-AH when treated with MAFP in vitro (26, 27), and LPS-induced plasma PAF-AH activity was inhibited by MAFP in vivo (28). As expected, MAFP reduced the survival from 100 to 40% in the 300 µg CpG-A₁₅₈₅ group (Figure 2B). Taken together, CpG-A₁₅₈₅ protects mice from PAF-induced sudden death, and this effect is dependent on PAF-AH.

As our studies demonstrated that CpG-A₁₅₈₅ was effective for PAF-induced lethal shock, we sought to clarify the mechanism by which CpG-A₁₅₈₅ protects mice from PAF-induced lethal shock. Severe congestion was confirmed in the PAF group compared with the non-treated (NT) group, which was improved by CpG-A₁₅₈₅ treatment (Figure 3A). It was reported that PAF promotes platelet aggregation (15). Therefore, we investigated platelet aggregation by measuring heart blood volume. Heart blood volume in the PAF group was significantly decreased compared with the NT group (p < 0.01), whereas CpG-A₁₅₈₅ significantly improved heart blood volume (p < 0.01) (Figure 3B). In hematoxylin-eosin (HE) and phosphotungstic acid hematoxylin (PTAH)-stained sections (Figures 3C,D), increased fibrin in a fibrin thrombus and an interlobular vein to renal corpuscle was observed in the PAF group compared with the NT group. In the CpG-A₁₅₈₅ + PAF group, fibrin thrombus size and fibrinosis were significantly decreased compared with the PAF group (Figure 3E). Serum analysis showed that PAF-AH activity was significantly increased in the CpG-A₁₅₈₅ + PAF group compared with the PAF group (p < 0.01) (Figure 3F). These data suggest that CpG-A₁₅₈₅ ameliorates PAF-induced fibrin formation via PAF-AH activity.

Platelet-activating factor has been shown to play a fundamental role in the regulation of TNF secretion in vitro (29). Here, we examined the effects of CpG-A₁₅₈₅-induced PAF-AH on inflammatory responses generated by LPS stimulation. CpG-A₁₅₈₅ reduced LPS-induced TNF-α mRNA and protein expression levels (Figures 4A,B). These data suggest that CpG-A₁₅₈₅-induced PAF-AH can regulate inflammatory responses.

In previous studies on the use of CpG ODNs for endotoxin shock and sepsis, Class B CpG ODN was confirmed to reduce survival in an experimental murine model of endotoxin shock or prevent sepsis-induced mortality (5, 30, 31). However, the effect of CpG-A ODN on endotoxin shock remains unclear. In addition, LPS is involved in PAF synthesis and PAF-induced diseases in vitro and in vivo (32, 33). Therefore, to determine whether CpG-A ODN could protect mice from LPS-induced endotoxin shock, mice were administered CpG-A₁₅₈₅ via i.p. injection 3 h before LPS challenge (Figure 5A). CpG-A₁₅₈₅ significantly improved 24-h mortality in the CpG-A₁₅₈₅ + LPS group compared with the LPS group (p < 0.0005) (Figure 5B). In addition, CpG-A₁₅₈₅ treatment improved hypothermia (Figure 5C). However, MAFP worsened hypothermia (Figure S2 in Supplementary Material). In addition, PAF-AH activity was significantly increased in the CpG-A₁₅₈₅ group compared with the control group at 12 h after LPS challenge (at 15 h after CpG-A₁₅₈₅ treatment) (p = 0.0199). IL-6 and TNF-α expression levels were significantly inhibited in the CpG-A₁₅₈₅ group compared with the control group at 12 h after LPS challenge (p = 0.0001 and p = 0.0001, respectively) (Figures 5D–F). Particularly, TNF-α expression levels in mice pretreated with CpG-A₁₅₈₅ (45.4 ± 17.6 pg/ml) were significantly lower than those in mice that received LPS only (190 ± 53.6 pg/ml) (p < 0.001). These results suggest that CpG-A₁₅₈₅ pretreatment effectively protects mice from LPS-induced endotoxin shock.

Endotoxin-free desalted PS-ODNs were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA, USA) or Gene Design, Inc. (Osaka, Japan). PS-ODNs were reconstituted in phosphate-buffered saline (PBS) and passed through a 0.22-µm pore microfilter (Nihon Millipore K.K., Tokyo, Japan). ODN sequences are shown in Table 1: CpG-A₁₅₈₅ (48), CpG-B₁₈₂₆ (49), CpG-C₂₃₉₅ (50), Ctr₁₆₁₂ (51), non-CpG-A₁₅₈₅, and iODN_H154 (13). LPS from Escherichia coli 0127:B8 was purchased from Sigma-Aldrich (St. Louis, MO, USA). MAFP and PAF C-16 were purchased from Cayman Chemical Co. (Ann Arbor, MI, USA). Poly(I:C) was purchased from InvivoGen (San Diego, CA, USA). rmIFN-α was purchased from BioLegend (San Diego, CA, USA).

Female BALB/c and male ICR (Swiss albino) mice (6 weeks of age) were purchased from Japan SLC (Shizuoka, Japan), housed under temperature- and light-controlled conditions, and fed a standard diet (MF, Oriental Yeast Co. Ltd., Tokyo, Japan) and sterile water ad libitum. Mice were used for experiments after preliminary housing for 1 week.

Splenocytes from female BALB/c mice (8 weeks of age) were prepared using standard methods. Cells were seeded onto 24-well plates (Nalge Nunc International K.K., Tokyo, Japan) at a final concentration of 1 × 10⁶ cells/well in complete RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% fetal calf serum (Sigma-Aldrich), 100 U/ml of penicillin, 100 mg/ml of streptomycin, 25 mM HEPES, 1.0 mM sodium pyruvate, non-essential amino acids, and 0.0035% 2-ME. Cells were then treated with 0.01–10 µM CpG ODN, 0.01–10 µg poly(I:C), or 0.01–1 ng rmIFN-α for 24 h (total 1 ml/well) (Figure 1; Figure S1 in Supplementary Material). After treatment, the cells and supernatant were collected for real-time quantitative PCR (qPCR) and PAF-AH enzymatic activity analysis, respectively. Cells were treated with 3 µM CpG-A₁₅₈₅ for 24 h, followed by stimulation with 10 ng/ml LPS for 6 h for qPCR analysis or for 24 h for enzyme-linked immunosorbent assay (ELISA) (Figure 4). Peritoneal macrophages were collected from mercy-killed female BALB/c mice (8 weeks of age) by peritoneal lavage with 5 ml of cold PBS, centrifuged at 1,500 rpm for 5 min and resuspended in medium. Murine peritoneal macrophages were then preincubated and adhered to dishes in medium for 24 h before exposure to CpG ODNs. Cells were seeded onto 24-well plates at a final concentration of 1 × 10⁵ cells/well and then treated with 3 µM CpG ODN for 24 h (total 1 ml/well) for qPCR analysis and PAF-AH enzymatic activity analysis (total 1 ml/well).

A schematic schedule of the experimental procedure is shown in Figure 2A. To determine the effect of PAF on the survival of ICR mice, we divided the animals (6 weeks of age) into the following three groups: PAF group, CpG-A₁₅₈₅ + PAF group, and CpG-A₁₅₈₅ + MAFP + PAF group (N = 8 mice per group). ICR mice were i.p. injected with 50, 100, or 300 µg CpG-A₁₅₈₅ 1, 3, and 5 days before PAF challenge. MAFP (1 mg/kg) was i.p. injected 20 min before CpG-A₁₅₈₅ injection. A stock solution of PAF was made in methanol, and the required aliquot was dried under a stream of nitrogen. PAF was then reconstituted in 0.5 ml PBS containing 0.1% human serum albumin (Wako Pure Chemical Industries, Ltd., Osaka, Japan) before use and administered i.p. into Swiss albino mice. After each treatment, animals were monitored for up to 60 min for survival.

A schematic schedule of the experimental procedure is shown in Figure 5A. BALB/c mice (7 weeks of age) were divided as follows: LPS group vs. CpG-A₁₅₈₅ + LPS group (Figure 5) and CpG-A₁₅₈₅ + LPS group vs. CpG-A₁₅₈₅/MAFP (5 mg/kg) + LPS group (Figure S2 in Supplementary Material) (N = 8 mice per group). BALB/c mice were i.p. injected with 300 µg CpG-A₁₅₈₅ 3 h before LPS challenge. Mice receiving MAFP (5 mg/kg) were i.p. injected 20 min before CpG-A₁₅₈₅ injection. The LPS-induced endotoxin shock model was established by i.p. injection of LPS (750 µg; Sigma-Aldrich). Survival was recorded at 0, 3, 6, 9, 12, 15, 18, 21, and 24 h. Body temperature (degree Celsius) was measured at 3-h intervals using an NTC thermistor (Tateyama Kagaku Industry Co., Ltd., Toyama, Japan).

Mice were euthanized by cervical dislocation 10 min after PAF injection. The maximum amount of blood was collected from the heart using a 10-ml syringe and 18 G needle within 5 min of death. The blood was weighed by ELECTRONIC BALANCE IBA-200 (AS ONE Corporation, Osaka, Japan). The volume (in milliliters) was calculated using the specific gravity (1.035) of mouse blood. The progression of blood clotting was evaluated by the quantity of collected blood.

The kidney, liver (quadrate lobule), and spleen were fixed with 10% formalin neutral buffer solution (Wako Pure Chemical Industries, Ltd.), embedded in paraffin, sliced, and stained with HE and PTAH. Slicing and staining of embedded blocks were performed by Biopathology Institute (Oita, Japan). Histological pathology was evaluated under light microscopy. The area of fibrin accumulation was determined by Image Processing Software (Media Cybernetics Inc., Bethesda, MD, USA) and collected from three mice per group at adjacent axial locations.

Total RNA from the cells stimulated with CpG ODN and LPS was isolated using NucleoSpin^® RNA (TaKaRa Bio Inc., Tokyo, Japan). cDNA was prepared by reverse transcription from 1 µg of total RNA per sample using PrimeScript^® RT Master Mix (TaKaRa Bio Inc.). Equal volumes of cDNA were used for quantification of various cytokine cDNAs via qPCR using the Thermal Cycler Dice^® Real Time system (TaKaRa Bio, Inc.). qPCR analyses were performed with SYBR Premix Ex Taq (TaKaRa Bio, Inc.) using specific primers. Primers for β-actin, PAF-AH2, and TNF-α were purchased from TaKaRa Bio, Inc. For cross-sample comparison of results obtained following various treatments, cytokine mRNA levels were first normalized to those of β-actin mRNA. Data are shown as the mean ± SD of one representative experiment of three independent experiments with similar results.

TNF-α level in sera or cell culture supernatants was quantified using a commercially available ELISA kit (TNF-α, eBioscience Inc., San Diego, CA, USA) according to the manufacturer’s instructions.

Platelet-activating factor acetylhydrolase activity in supernatant or sera was measured using the PAF-AH assay kit (Cayman Chemical Co.) (28). Briefly, 10 µl sample and 10 µl 5,5′-dithiobis (2-nitrobenzoic acid) were added to wells of a 96-well plate. The reactions were initiated by adding 200 µl substrate solution (2-thio-PAF). The absorbance at 405 nm was read every minute using a plate reader (iMark™ Microplate Reader, Bio-Rad, Hercules, CA, USA).

Statistical analyses were performed using a statistical software package (ystat2004.xls, Igaku Tosho Shuppan, Tokyo, Japan) or GraphPad Prism7 (GraphPad Software, Inc., La Jolla, CA, USA). All data were analyzed by one-way analysis of variance with the post hoc Student–Newman–Keuls test, except for survival analyses, in vitro qPCR analysis, PAF-AH activity assay, and all data from the LPS-induced endotoxin shock model. Survival analyses were performed using the log-rank test. In vitro qPCR analysis, PAF-AH activity assay, and all data from the LPS-induced endotoxin shock model were analyzed using Student’s t-test and the Dunnett test. Differences were considered significant at p < 0.05. Values for in vitro data are expressed as the mean ± SD. Other values are expressed as the mean ± SE.