BALB/c mice were purchased from KOATECH (DaeJeon, Korea). Coral green fluorescent protein (cGFP) transgenic mice in a BALB/c background were kindly provided by Dr. Kwon (International Vaccine Institute, Seoul, Korea). The mice were subsequently bred at Chonnam National University in chambers maintained at 22~23℃ with equal 12 h periods of light and dark each day. The Cyclooxygenase (COX)-1-specific inhibitor SC-560 (Cayman Chemical, Ann Arbor, MI, USA) and the non-specific COX-1 inhibitor acetylsalicylic acid (ASA; Sigma-Aldrich, St. Louis, MO, USA) dissolved in ethanol were used. TRIzol reagent was purchased from Invitrogen (Carlsbad, CA, USA) and primers were obtained from Bionics (Seoul, Korea). All experiments using mice were followed the guideline approved by the Committee for the Care and Use of Laboratory Animals at Chonnam National University.

The ginsan polysaccharide was purified from the ethanol insoluble fraction of the aqueous P. ginseng extract as described previously (15). Further purification was carried out successively using size exclusion and ion exchange column chromatography. Nuclear magnetic resonance analysis revealed that the purified fraction was composed of α(1→6) glucopyranoside and β(2→6) fructofuranoside in a 5:2 molar ratio. The endotoxin level in the purified ginsan preparation was less than 0.03 EU/mg as measured using the Endosafe™ Limulus amebocyte lysate assay (Charles River Laboratories, Wilmington, MA, USA) according to the manufacturer's instructions. The purified ginsan was dissolved in phosphate buffered saline (PBS; pH 7.4) and filtered through a 0.22µm Millipore membrane (Millipore, Billerica, MA, USA).

Ginsan (100 mg/kg) was intraperitoneally introduced to the mice once a day for two days before oral antigen administration. COX inhibitors, SC-560 (5 mg/kg) and ASA (100 mg/kg), were orally administered on the same day ginsan treatment was done (16). To continuously suppress COX, orally vaccinated mice were fed with indomethacin (30µg/ml) dissolved in drinking water.

Salmonella typhimurium were grown to late log phase in LB broth at 37℃ with vigorous aeration. For immunization, S. typhimurium was fixed with 4% paraformaldehyde (Sigma-Aldrich) for 30 min and washed three times with sterile PBS. Mice fasted overnight were orally immunized with 1×10⁹ colony forming units of S. typhimurium suspended in 500 µl PBS using oral zonde needle. Mice were immunized twice during a two week interval.

Samples were taken from each group of immunized mice. Stool samples collected by isolating each mouse in a dark box for 10 minutes before other procedures. Blood samples were collected by cardiac puncture of mice anaesthetized with ketamine (100 mg/kg) 14 days after the second immunization. Vaginal lavage fluids were obtained by gentle flushing of the mouse vagina with 100 µl of sterile PBS. Lavage fluids were then centrifuged and the supernatants were analyzed for antigen-specific antibody content. Serum, saliva, stool and vaginal lavage samples were stored at -20℃ until ELISA. Briefly, 96-well microplates (Corning, Corning, NY, USA) were coated with 50 µl of whole S. typhimurium cell lysate dissolved in PBS overnight. The plates were blocked with 1% BSA (Sigma-Aldrich) in PBS to avoid non-specific binding. A volume of 50 µl of each diluted sample was added to the plates, incubated at room temperature for 2 h, and then the plates were washed. Horseradish peroxidase conjugated goat anti-mouse IgA secondary antibody (Sigma-Aldrich), IgG1 (BD Pharmingen) or IgG2 (BD Pharmingen) was then added and incubated for 1 h at room temperature. o-Phenylenediamine dihydrochloride (0.4 mg/ml; Sigma-Aldrich) in 0.05 M phosphate citrate buffer was used as the substrate. The absorbance was measured at 490 nm in an ELISA reader (Molecular Devices, Sunnyvale, CA, USA).

The activation of COX-1 and COX-2 in Peyer's patch and ileum was observed 24 h following ginsan treatment by Western blot. Proteins were probed with monoclonal anti-COX-1, anti-COX-2, or anti-beta-actin antibody at a 1 : 1,000 or 1 : 5,000 dilution. Horseradish peroxidase-linked secondary antibodies were used at a 1:1,000 dilution. Signal detection was carried out using a LAS-3000 scanning imager (Fujifilm, Tokyo, Japan).

Expression levels of RNA transcripts were quantified using RT-PCR. Total RNA from the Peyer's patches, spleen and ileum were isolated using TRIzol reagent (Invitrogen). RNA (2 µg/tube) was reverse transcribed into cDNA for 45 min at 40℃ using Superscript RT (Promega, Madison, WI, USA) in the presence of 1 µl random primer (Promega). The cDNA was used as a template for PCR amplification (Eppendrof, Westbury, NY, USA). The reaction mixture consisted of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl₂, 200 µM dNTP (TaKaRa Bio, Shiga, Japan), 0.5 U Taq DNA polymerase (TaKaRa Bio), and 0.05 mM of each primer. The primers used were as follows: CCL3 sense, 5'-AACATCATGAAGGTCTCCAC-3', CCL3 antisense, 5'-CCAAGACTCTCAGGCATTCA-3', CCL4 sense, 5'-GTTCTCAGCACCAATGGGCTCTGA-3', CCL4 antisense, 5'-CTCTCCTGAAGTGGCTCCTCCTG-3', CCL5 sense, 5'-GGTACCATGAAGATCTCTGCA-3', CCL5 antisense 5'-AAACCCTCTATCCTAGCTCAT-3', CCL7 sense, 5'-TCTGTGCCTGCTGCTCATAG-3', CCL7 antisense, 5'-CTTTGGAGTTGGGGTTTTCA-3', CCL8 sense, 5'-CCTTCAACATGAAGATCTACGCA-3', CCL8 antisense, 5'-AACTCAGGTGTGAAGGTTCAAG-3', CCL9 sense, 5'-ATCACACATGCAACAGAGACAAAA-3', CCL9 antisense, 5'-TTATTGTTTGTAGGTCCGTGGTTG-3', CCL20 sense, 5'-TGCTCTTCCTTGCTTTGGCATGGGTA-3', CCL20 antisense, 5'-TCTGTGCAGTGATGTGCAGGTGAAGC-3', CCL21 sense, 5'-ATGGCTCAGATGATGACTCT-3', CCL21 antisense, 5'-TACTGGGCTATCCTCTTGA-3', CCL25 sense, 5'-TCACTATGAAACTGTGCCTTTTTG-3', CCL25 antisense, 5'-CTTAATTGTTGGTCTTTCTGGGC-3', COX-1 sense, 5'-AGGAGATGGCTGCTGAGTTGG-3', COX-1 antisense, 5'-AATCTGACTTTCTGAGTTGCC-3', COX-2 sense, 5'-ACACACTCTATCACTGGCACC-3', COX-2 antisense, 5'-TTCAGGGAGAAGCGTTTGC-3', β-actin sense, 5'-TCATGAAGTGTGACGTTGACATCCGT-3', and β-actin antisense, 5'-CTTAGAAGCATTTGCGGTGCACGATG-3'.

The thermal cycle conditions were 94℃ for 1 min, 55℃ for 30 s, and 72℃ for 30 s for 35 cycles. PCR products were electrophoresed on 2% agarose gels.

Splenocytes were isolated from cGFP-positive mice. Single cell suspensions were obtained by gently chopping the tissues with the needles of a 1 ml syringe and suspending the cells in RPMI 1640 containing 10% fetal bovine serum (FBS). After washing, each cell suspension was subjected to gradient centrifugation by overlaying 5 ml volumes on 5 ml Lymphoprep™ (1.077 g/ml; Nycomed Pharma, Oslo, Norway) and spinning at 400 ×g for 15 min at room temperature as recommended by the manufacturer. Interface cells were collected, washed in RPMI 1640, resuspended, and the number of viable cells was determined using the Trypan blue exclusion assay. Splenocytes (1×10⁷) were transferred to the syngenic Balb/c mice through the caudal vein in 100µl of RPMI 1640.

Tissues were embedded in Tissue-Tek OCT compound (Miles Scientific, Naperville, IL, USA) and stored at -80℃. Cryosections 10 µm in thickness were fixed in ethanol and co-stained for phenotypic markers. Antibodies used for the phenotypic markers were as follows: CD11c (BD Pharmingen), CD3 (R&D Systems, Minneapolis, MN, USA), B220 (R&D Systems), and CCL3 (R&D Systems). All were visualized with Alexa-488 labeled or Alexa-596 labeled secondary antibody (Molecular Probes, Eugene, OR, USA). The tissues were mounted with a Slowfade™ antifade kit (Molecular Probes) and visualized using a Radiance 2100 confocal laser scanning microscope (Bio-Rad, Hercules, CA, USA).

Peyer's patches were removed from small intestines. Single cell suspensions were obtained as described earlier in the text. The cells were pooled and poured through a 100 µm nylon mesh (Millipore) prior to centrifugation for 15 min at 400 ×g. Cells were incubated on ice for 30 min in Fc block consisting of 5% mouse serum (Sigma) and 5 µg/ml anti-mouse CD32/CD16 (BD Pharmingen) in the presence of phycoerytherin-conjugated CD11c (BD Pharmingen). After staining, cells were washed and analyzed by flow cytometry with FACScan (Becton Dickson, Franklin Lakes, NJ, USA) and analyzed using CellQuest software (Becton Dickson).

Data analysis was performed with SPSS software (SPSS, Version 17, SPSS Inc., Chicago, IL, USA). Student's t-test and ANOVA were used to compare the parameters between groups. Values of p<0.05 were considered significant.

Given the descriptions of ginseng extracts as being a potent vaccine adjuvant (6-8), we initially tested whether ginsan was likewise a potent adjuvant to orally administered antigens. Mice were pretreated twice with ginsan before oral immunization twice weekly for two weeks. Two weeks after the second immunization, specific antibodies against Salmonella were determined from serum, saliva, vaginal fluid and stool. Ginsan treated mice produced significantly higher amounts of serum IgG1 and IgG2 (Fig. 1A), as well as secretory IgA in all samples (Fig. 1B).

Recently, we have reported that ginsan modulates the expression of COX. Ginsan treatment suppressed allergic reaction in murine asthma model (16) and it was also effective in protecting the host from pathogenic bacterial infection via nasal route (17). Since both models were related with mucosal organs, we speculated that ginsan may also modulate COX expression in the Peyer's patch. Ginsan also specifically enhanced the mRNA expression of COX-1 and COX-2 in Peyer's patches (Fig. 2A) and it was confirmed by western blot examination (Fig. 2B).

Next, we examined various genes to elucidate the mechanism of ginsan. Among various genes, ginsan specifically increased the expression of CCL3 in the Peyer's patches, while expression was decreased both in the spleen and nearby ileum. Expression of some chemokine ligands were rather decreased following ginsan treatment (Fig. 3). Since CCL3 is related with dendritic cell (DC) migration (18,19), other chemokine ligands related with DC migration were screened. There were no changes in the expression of other genes related with DC migration in the Peyer's patches.

To determine CCL3 expressing cells within the Peyer's patches, frozen sections were stained with anti-CCL3 in combination with CD11c. CCL3 was present within the subepithelial dome region and CCL3 expression was markedly increased following ginsan treatment (Figs. 4A~D), consistent the RT-PCR results.

To determine the effect of ginsan on the DC population in the Peyer's patch, mice were sacrificed 24 h after ginsan treatment and the Peyer's patches were collected for flow cytometry analysis. Ginsan treatment significantly increased CD11c positive cells in the Peyer's patch compared to the control group (Fig. 5A). To determine if ginsan modulated DC migration, splenocytes from cGFP transgenic mice were intravenously introduced 24 h after ginsan treatment. The following day, Peyer's patches were prepared for immunofluorescence analysis of the GFP splenocyte migration. Consistent with the flow cytometry results, immunofluorescence revealed ginsan-mediated enhanced migration of GFP splenocytes to the Peyer's patches (Fig. 5B). Most of the migrated cells were found near the subepithelial dome of the Peyer's patch where DC are more prevalent (Fig. 5C). The results are consistent with the suggestion the ginsan can enhance the migration of CD11c-positive splenocytes to the Peyer's patches.

To test the role of COX on CCL3 expression regulation, mice were treated in combination with ginsan and COX inhibitors. Semi-quantitative RT-PCR for CCL3 mRNA in the Peyer's patch was determined following various treatments. The specific COX-1 inhibitor SC-560 moderately reduced the ginsan-enhanced CCL3 expression (Fig. 6A). ASA, a nonspecific COX inhibitor, mostly reduced the CCL3 expression induced by ginsan (Fig. 6B). However, single SC-560 or ASA did not totally suppress CCL3 expression suggesting that physiological CCL3 expression may be regulated by some other mechanism.

To determine the effect of ginsan and COX inhibitor on splenocyte migration, GFP splenocytes were intravenously introduced to mice pretreated with ginsan or ASA. Since COX-1 inhibition can upregulate COX-2 expression (20), ASA was used to suppress both COX-1 and COX-2. Twenty-four hours after GFP splenocyte delivery, mice were sacrificed and Peyer's patches were collected for flow cytometry analysis. Ginsan treatment increased both GFP positive and CD11c positive cells compared to the control group. ASA antagonized the effect of ginsan, reducing both CD11c cells and the migration of GFP positive cells (Fig. 7). Consistent with the CCL3 mRNA expression result, treatment with just ASA did not totally reduce the population of CD11c or the migration of DCs to the Peyer's patches.

Finally, antibody responses to oral immunization were determined. Ginsan treated mice produced significantly higher antibody titers both in stool and serum compared to that the control group. ASA suppressed antibody production induced by ginsan. Treatment just with ASA did not influence systemic antibody response, but decreased the amount of secretory IgA antibody compared to other groups (Fig. 8).