Glycyrrhizin Protects γ-Irradiated Mice from Gut Bacteria–Associated Infectious Complications by Improving miR-222–Associated Gas5 RNA Reduction in Macrophages of the Bacterial Translocation Site

High mortality rates associated with gastrointestinal acute radiation syndrome (GIARS) are a serious concern in victims exposed to 6–15 Gy of γ-rays (1–4). In our previous studies (5, 6), the mortality of mice exposed to 6–8 Gy of whole-body γ irradiation was associated with sepsis caused by bacterial translocation. Generally, gut bacteria–associated infectious complications do not develop in healthy individuals (5–7) because pathogens that invade from the gastrointestinal (GI) tracts are rapidly eliminated by the host defense effector cells at the bacterial translocation sites (mesenteric lymph nodes [MLNs] and lamina propria) (3). M1Mϕ (IL-12⁺CD38⁺iNOS⁺ Mϕ) have been identified as major host defense antibacterial effector cells against bacterial translocation (5–7). However, M2bMϕ (CCL1⁺IL-10⁺LIGHT⁺miR-27a⁺ Mϕ), which appear in association with irradiation at the bacterial translocation sites, suppress the M1Mϕ polarization (5). As such, GIARS mice are much more susceptible to bacterial translocation and subsequent sepsis because of their weakened and/or lack of antibacterial innate immunity.

Molecular mechanisms and signaling pathways involved in the phenotypic polarization of macrophages (Mϕ) have been well described (8–10). Recently, we have reported (11) that decreased expression of noncoding RNA Gas5 is linked to M2bMϕ polarization influenced by LPS and immune complex or IL-1β. The expression of Gas5 RNA level is shown to be minimal in MLN-Mϕ isolated from GIARS mice (6). M1Mϕ are readily obtained by Ag stimulation from MLN-Mϕ isolated from GIARS mice transduced with the Gas5 gene using lentiviral vector (Gas5 lentivirus) (6). Also, the MLN-Mϕ isolated from GIARS mice transduced with Gas5 gene did not polarize to M2bMϕ again, although they were stimulated with a combination of LPS and immune complex (6). Ccl1 gene transcription is silenced by Gas5 (12), and CCL1 has been well known to be an essential chemokine required for M2bMϕ survival (13).

The effects of glycyrrhizin (GL) on various opportunistic infections have been well described (14–20). GL is an extract from licorice roots with a structure of 20β-carboxy-11-oxo-30-norolean-12-en-3β-yl-2-O-β-d-glucopyranuronosyl-α-d-glucopyranosiduronic acid (21). GL has been used clinically for >40 y in patients with chronic hepatitis in Japan (22, 23). Through the induction of antisuppressor cells or inhibition of CCL2 production, GL suppresses M2a/cMϕ polarization (20, 24, 25), myeloid-derived suppressor cell function (18, 19), and Th2 cell generation (14–17, 26, 27). Through the modulation of these suppressor cell polarizations, GL protects severely burned mice from infections stemming from Candida albicans (15, 16), Staphylococcus aureus (20), and Pseudomonas aeruginosa (18–20). Through the induction of IFN-γ, GL protects healthy individuals infected with the hepatitis virus (22, 28) and influenza virus (29). Also, the antitumor activities of GL have been well described (24, 30–32). GL suppresses the growth of Meth A solid tumor (24) and B16 melanoma (30) through the stimulation of IFN-γ–associated host antitumor immunities. Furthermore, various immunomodulating activities of GL have been reported, including inhibition of inflammation (33, 34), augmentation of NK cell activity (35), induction of antimicrobial peptide production by keratinocytes (18, 19, 36), and induction of IL-12, CCL3, and CCL5 from Mϕ and/or T cells (17, 20, 37).

In the current study, the effect of GL on the mortality of mice exposed to 7 Gy of γ-rays (7 Gy GIARS mice) with gut bacteria–associated sepsis was investigated. Bacterial translocation and subsequent sepsis were demonstrated in 7 Gy GIARS mice starting 1 wk after irradiation, and all of these mice died within 3 wk of irradiation. However, the mortality rates of GIARS mice treated with GL were dramatically reduced. GL reduced the mortality of 7 Gy GIARS mice by controlling the M2bMϕ polarization in the bacterial translocation site. Host antibacterial effector cells against bacterial translocation and subsequent sepsis have been identified to be M1Mϕ (5, 7), and M2bMϕ have been characterized as inhibitor cells for the M1Mϕ polarization (5, 38). Decreased expression of Gas5 RNA, a silencer of Ccl1 gene transcription (39), was restored in MLN-Mϕ of 7 Gy GIARS mice treated with GL. CCL1 is an essential chemokine for M2bMϕ survival (13). The MLN-Mϕ from 7 Gy GIARS mice treated with GL were shown to be non-M2bMϕ and hence had the ability to easily polarize to M1Mϕ with Ag stimulation. A decrease in Gas5 RNA expression in MLN-Mϕ from GIARS mice was shown to be associated with a high expression of miR-222 in the same Mϕ, but a high level of miR-222 expression was not demonstrated in MLN-Mϕ of GIARS mice treated with GL. These results indicate that GL protects 7 Gy GIARS mice from bacterial translocation and subsequent sepsis through the modulation of M2bMϕ polarization in the bacterial translocation site by inhibiting miR-222 expression followed by the restoration of the decreased expression of Gas5 RNA.

BALB/c mice (9–12 wk old, specific pathogen free) were purchased from The Jackson Laboratory (Bar Harbor, ME). Data shown in survival experiments in Fig. 1A are representative of two independent experiments using 44 male mice (10–12 mice per group), and data shown in the others are mean ± SD from three independent experiments performed with male mice (two experiments) and female mice (one experiment), consisting of six to nine mice per group. These mice were exposed to 7 Gy of whole-body irradiation with a ¹³⁷Cs-ray (0.662 MeV) irradiator (Mark I Model 30; J.L. Shepherd and Associates, San Fernando, CA) at a dose rate of 1.05 Gy/min, which was reduced from 5.08 Gy/min via lead attenuators. All animal studies were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Texas Medical Branch at Galveston (Institutional Animal Care and Use Committee approval no. 1305020).

Advanced DMEM/F12 medium, murine CCL1 ELISA kit, TRIzol reagent, Ambion mirVana miRNA Isolation Kit, PrimeScript RT reagent kit, streptavidin magnetic beads, PEG-it Virus Precipitation Solution, Lipofectamine 2000 and RNAi MAX transfection reagents, miR-222 mimic, negative control (NC) miRNA, and TaqMan microRNA probes for miR-222 and miR-361 quantification were purchased from Thermo Fisher Scientific (Waltham, MA). iTaq Universal SYBR Green Supermix was purchased from Bio-Rad Laboratories (Hercules, CA). Biotin-conjugated anti-mouse F4/80 Ab was obtained from eBioscience (San Diego, CA). MagCollect buffer was purchased from R&D Systems (Minneapolis, MN). Anti-Ly6G Ab was purchased from BioLegend (San Diego, CA). Recombinant murine M-CSF was purchased from PeproTech (Rocky Hill, NJ). RPMI 1640 medium (Life Technologies, Grand Island, NY) supplemented with 10% heat-inactivated FBS (GE Healthcare Life Sciences, Pittsburgh, PA) and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin; Life Technologies) (complete medium) was used for cultivation of Mϕ. NMDI14, murine IgG, and LPS were purchased from Sigma-Aldrich (St. Louis, MO).

GL (ammonium salt) was supplied from Minophagen Pharmaceutical (Tokyo, Japan). GL was dissolved in 5% ethanol containing saline or serum-free medium (10 mg/ml), diluted with saline at appropriate concentrations, and 0.2 ml of the solution was administered i.p. to 7 Gy GIARS mice. Final concentration of ethanol was <0.1%. In our early studies (29), the effect of various doses of GL on influenza A₂ virus infection in mice was examined. The protective effect of GL was demonstrated when infected mice were treated with >2.5 mg of the compound per kilogram. This effect was dose dependent and observed with doses ranging from 2.5 to 10 mg/kg. The maximum protection of mice exposed to the virus was shown when doses of GL of 10 mg/kg or more were administered. Similar dose-dependent protective/therapeutic effects of GL have been demonstrated in various immunocompromised mice infected with C. albicans (15, 16) and P. aeruginosa (18, 19). In these experiments, GL protected the infected mice through the modulation of impaired production of IL-12 (17) and IFN-γ (15) or excessive CCL2 production (25, 27). Based on our accumulated results, in this study, a 10 mg/kg dose of GL was administered to 7 Gy GIARS mice.

In the majority of experiments, Mϕ (F4/80⁺ cells) were prepared from the MLNs of 7 Gy GIARS mice 10 d after γ irradiation. As previously described (5–7), single-cell suspensions were obtained by gently pressing MLNs in PBS supplemented with 2% FBS using a cell strainer, adjusted to 5 × 10⁶ cells/ml in MagCollect buffer, and biotin-labeled anti-F4/80 Ab was added. Fifteen minutes after the incubation on ice, cells were washed twice, resuspended with the same cold buffer, and streptavidin-coated magnetic beads were added. Mϕ were separated from the cell suspension by a positive selection technique. A Mϕ-enriched population (>97% pure as F4/80⁺ cells) was consistently obtained using this technique. In some experiments, MLN-Mϕ were cultured in complete medium supplemented with 10 μg/ml of GL for 24 h. For the functional analysis of miR-222, MLN-Mϕ (5 × 10⁵ cells/ml) were transfected with miR-222 mimic or control miRNA using Lipofectamine RNAi MAX, according to the manufacturer’s protocol (reverse transfection method).

Bone marrow–derived Mϕ were prepared as previously described (11). Briefly, bone marrow cells were isolated from mouse femoral and tibial bones by flushing the marrow out with PBS. Then, these cells (1 × 10⁵ cells/ml) were cultured for 7 d in advanced DMEM/F12 medium supplemented with 10% FBS and 25 ng/ml M-CSF on a 6-cm petri dish. The medium was changed every 2 d during the cultivation. The purity of F4/80⁺ cells 7 d after the cultivation was routinely >98%. For the preparation of Mϕ stimulated with LPS and immune complex in combination (M[LPS + IC]), bone marrow–derived Mϕ (5 × 10⁵ cells/ml) were harvested and recultured for 3 d in media supplemented with 100 ng/ml of LPS on culture plates, previously coated with 100 μg/ml of murine IgG for 2 h at 37°C (38). These Mϕ preparations were harvested, and total RNAs extracted from these cells were assayed for the gene expression by real-time PCR, as described below.

Murine Gas5 cDNA was amplified from pCMV-Sport6-Gas5 plasmid and cloned into pLenti7.3/V5-TOPO vector (pLenti7.3-Gas5). Lentiviruses were prepared using HEK293FT cells as described in the manufacturer’s protocol. In brief, 3 μg of pLenti-Gas5 vector and 9 μg of packaging mix were cotransfected into HEK293FT cells using transfection reagent Lipofectamine 2000. Seventy-two hours after transfection, supernatants were filtered (0.45-μm filter). To concentrate produced viruses in the supernatants, PEG-it Virus Precipitation Solution was added to every 4 vol of the supernatants, and the mixture was incubated for 16 h at 4°C. Virus in the mixture was precipitated by centrifugation at 1500 × g for 30 min, resuspended in PBS, and stored at −80°C until further use. For the titration of the viruses, the suspension was serially diluted with DMEM supplemented with 10% FBS and added to HEK293FT cells. Forty-eight hours after the cultivation, GFP⁺ cells were counted by flow cytometry, which was expressed from pLenti7.3/V5-TOPO vector. The titers (transducing units [TU] per milliliter) were calculated with the following formula: TU/ml = (% GFP⁺ cells/100 × total number of cells in the well)/culture volume (milliliter) × dilution factor. Mock viruses were generated by the same procedure using an otherwise identical vector lacking Gas5 cDNA (NC lentivirus).

For the gene expression analysis, the total RNA was extracted from MLN-Mϕ with TRIzol reagent, according to the manufacturer’s instruction. The cDNAs were synthesized with SuperScript III First-Strand Synthesis System. Quantitative real-time PCR was performed using an iTaq Universal SYBR Green Supermix on a ViiA 7 Real-Time PCR System (Thermo Fisher Scientific) with the specific primer pairs to murine Ccl1, Tnfsf14 (LIGHT), Gas5, and Gapdh, as listed in Supplemental Table I. The expression levels of Ccl1 and Gas5 were normalized to that of housekeeping gene Gapdh. For a quantification of miR-222, the total RNA was extracted from the same cells using Ambion mirVana miRNA Isolation Kit. The expression levels of miR-222 and miR-361 were quantified using TaqMan microRNA Assay. The kit uses gene-specific stem-loop reverse-transcription primers and TaqMan probes to detect mature microRNA transcripts. PCR was carried out on the ViiA 7 Real-Time PCR System. The expression levels of miR-222 were normalized to that of miR-361.

Data were summarized as mean ± SD. Results were statistically analyzed by a Student t test or a one-way ANOVA. Survival curves were compared using a log-rank test. Differences were considered significant at the 0.05 level of significance.

All GIARS mice treated with saline died within 3 wk of 7 Gy γ irradiation (7 Gy GIARS mice). When GL (10 mg/kg) was administered i.p. to these mice 1, 3, 5, 7, and 9 d after exposure to 7 Gy of γ rays, all of them survived for >30 d after the irradiation (Fig. 1A). The same protective effect of GL on 7 Gy GIARS mice was demonstrated when it was administered only on days 7 and 9 postirradiation. When treated only on day 9 postirradiation, 70% of 7 Gy GIARS mice survived (Fig. 1A). In our previous studies (6), the mortality of 7 Gy GIARS mice was associated with infectious complications stemming from gut microbiota. Therefore, we next examined the bacterial growth in the primary bacterial translocation site (MLNs) and other organs (liver, spleen, and kidneys) in 7 Gy GIARS mice treated with or without GL. As shown in Fig. 1B, progressive growth of bacteria was shown in all tested organs of GIARS mice 8–12 d postirradiation. However, the pathogen did not grow significantly in these organs of GIARS mice treated with GL 7 and 9 d postirradiation (Fig. 1B). These results indicate that bacterial translocation and subsequent sepsis do not develop in 7 Gy GIARS mice treated with GL.

MLN-Mϕ isolated from GIARS mice have been previously characterized as M2bMϕ, which have a function to suppress the polarization of M1Mϕ (a major host defense antibacterial effector cell against bacterial translocation and subsequent sepsis) (5, 6). Therefore, the effect of GL on the M2bMϕ polarization in the MLNs of 7 Gy GIARS mice was examined. CCL1⁺IL-10⁺LIGHT⁺miR-27a⁺ Mϕ (M2bMϕ) were detected in the MLNs of these mice 10 d postirradiation. However, M2bMϕ were not isolated from the MLNs of these GIARS mice treated with GL. CCL1⁻IL-10⁻LIGHT⁻miR-27a⁻ cells (non-M2b phenotype) were isolated from the MLNs of GIARS mice treated with GL (Fig. 2A–C).

The importance of Gas5 RNA for avoiding the M2bMϕ polarization has been described in our previous paper (11). Gas5 is a long noncoding RNA with a function to silence Ccl1 gene transcription (12), and CCL1 is an essential cytokine for the M2bMϕ survival (13). Gas5 was consistently expressed in the MLN-Mϕ of normal mice, whereas it was not in the MLN-Mϕ of GIARS mice. When 7 Gy GIARS mice were treated with GL, the Gas5 RNA levels in their MLN-Mϕ was restored to a normal level (Fig. 3A). Similar effect of GL on Gas5 RNA expression was demonstrated in cultures of MLN-Mϕ from 7 Gy GIARS mice. Thus, MLN-Mϕ of 7 Gy GIARS mice, treated with 10 μg/ml of GL, expressed Gas5 RNA at the level shown in normal mouse MLN-Mϕ (Fig. 3B). This effect of GL was also seen in M2bMϕ generated from bone marrow–derived M(LPS + IC). M(LPS + IC) have been described as a standard M2bMϕ preparation (38). The expression of Ccl1 mRNA (Fig. 3C) and production of CCL1 (Fig. 3D) were markedly decreased in M(LPS + IC) 24 h after cultivation with 10 μg/ml of GL. Similarly, mRNA expression and production of CCL1 were decreased in M(LPS + IC) transduced with the Gas5 gene via a lentiviral vector (Gas5 lentivirus).

Because miR-222 is known to directly bind to Gas5 RNA and inhibit this RNA expression (39), we focused on this miRNA expression in MLN-Mϕ from 7 Gy GIARS mice treated with or without GL. miR-222 was increasingly expressed in MLN-Mϕ of GIARS mice 1–15 d after the γ irradiation (Fig. 4A). When GL was administered to 7 Gy GIARS mice 7 and 9 d postirradiation, miR-222 expression dropped to the level shown in normal mouse MLN-Mϕ (Fig. 4A). A similar effect of GL on miR-222 expression was demonstrated in cultures of MLN-Mϕ isolated from 7 Gy GIARS mice 10 d postirradiation (Fig. 4B, 4C). Gas5 RNA reduction was shown to be mediated by miR-222, as evident from the Gas5 RNA reduction seen in MLN-Mϕ from normal mice after the transfection of miR-222 (Fig. 4D). The results shown in Fig. 4A–D indicate that GL improves Gas5 RNA expression in MLN-Mϕ of 7 Gy GIARS mice through the reduction of miR-222 expression.

Although there are some measures of preventing mortality of victims with hematopoietic acute radiation syndrome exposed to 2–5 Gy of γ-rays, similar measures are less effective in preventing mortality of victims with GIARS exposed to 6–15 Gy of γ-rays (1–4). Thus, there is an urgent need for innovative strategies to treat GIARS patients. In our previous paper (6), the mortality of mice exposed to 6–9 Gy of γ-rays was shown to be associated with infectious complications stemming from gut microbiota translocation. Also, we have demonstrated that M1Mϕ located in the bacterial translocation sites (MLNs and lamina propria) of unirradiated mice are characterized as a major host defense antibacterial effector cell. However, Mϕ isolated from the MLNs of GIARS mice were shown to be mostly M2bMϕ (inhibitor cells for M1Mϕ polarization) (5). Reduced Gas5 expression and increased miR-222 expression in the MLN-Mϕ of mice exposed to 6–9 Gy of γ-rays were shown to be associated with M2bMϕ polarization (6). In the current study, the effect of GL on the mortality of 7 Gy GIARS mice was investigated. Furthermore, to make clear the mode of action of GL, the effects of the compound on M2bMϕ polarization in the MLNs of these mice was studied. Bacterial translocation from the gut to the MLNs was observed in 7 Gy GIARS mice 1 wk after irradiation, and all of these mice died within 3 wk of irradiation. However, all of the 7 Gy GIARS mice treated twice with GL (7 and 9 d postirradiation) survived. The growth of bacteria in the MLNs and other organs was minimal in these mice treated with GL. M1Mϕ (IL-12⁺CD38⁺ iNOS⁺ Mϕ) in the bacterial translocation site (from the intestinal tracts to MLNs) have been described as major host defense antibacterial effector cells against invading pathogens (5–7). However, M2bMϕ (IL-10⁺LIGHT⁺CCL1⁺ miR-27a⁺ Mϕ) were predominant in the MLNs of 7 Gy GIARS mice because of the M1Mϕ polarization being suppressed by M2bMϕ (5, 6). As shown in Fig. 2, GL was shown to be an inhibitor of M2bMϕ polarization. After GL treatment, MLN-Mϕ isolated from 7 Gy GIARS mice did not show any properties of M2bMϕ. Also, MLN-Mϕ isolated from GL-treated GIARS mice easily polarized to M1Mϕ with bacterial Ag stimulation (data not shown). In contrast, GL was not effective in controlling the mortality for GIARS mice exposed to 10 or more Gy (data not shown), which is an indication that for severe intestinal damage induced by high doses of γ-rays, GL treatment may need to be modified to account for the higher dose of radiation. In fact, 10 or more Gy of γ-rays suppresses the self-renewing function of intestinal stem cells that are important for the maintenance of crypt integrity and regeneration (40). In our recent studies (6), intestinal damages (detected by fewer crypts and diminished length of villi in the ileum) were markedly healed in mice exposed to 10 Gy of γ-rays after transplantation of intestinal lineage cells differentiated from embryonic stem cell–derived definitive endoderm. Thus, as our next steps, we are currently trying to control the mortality of mice exposed to 10–15 Gy of γ-rays by the transplantation of intestinal lineage cells combined with the GL treatment.

Some adverse effects (e.g., sodium retention, hypokalemia, decreased ion aldosterone and renin) of GL have been known in humans (41); however, we have not seen abnormal electrolyte or hormone levels in 7 Gy GIARS mice treated i.p. with 10 mg/kg of GL. According to the US Food and Drug Administration guideline, a murine dose of 10 mg/kg is calculated as 0.8 mg/kg for a human-equivalent dose. For >40 y, GL has been widely used for chronic hepatitis patients in Japan by drip infusion of 40–60 ml/d of Stronger Neo-Minophagen C (Minophagen Pharmaceutical), a GL-containing preparation. Because 20 ml of Stronger Neo-Minophagen C injection contains 40 mg of GL, the doses used for these patients are 1.3–2.0 mg/kg/d. Therefore, a dose of GL used in our murine studies is considered to be not high. In the following studies, we will determine the effect of GL on the M2bMϕ polarization in the human experimental system.

Now we discuss how GL suppresses the M2bMϕ polarization in MLNs of GIARS mice. CCL1 is an essential cytokine for the maintenance of M2bMϕ survival (13). The expression of Gas5 RNA, a long noncoding RNA to silence Ccl1 gene expression, was decreased in MLN-Mϕ of 7 Gy GIARS mice, and it was restored in MLN-Mϕ of these mice after treatment with GL. The recovery of the Gas5 gene expression in MLN-Mϕ of 7 Gy GIARS mice treated with GL was confirmed when a standard preparation of M2bMϕ polarized from bone marrow–derived Mϕ (M[LPS + IC]) after being treated with GL. Also, CCL1 was not expressed/produced in M(LPS + IC) and the MLN-Mϕ of 7 Gy GIARS mice after transduction with the Gas5 gene via a lentiviral vector. These results indicate that GL suppress the M2bMϕ polarization by improving the Gas5 level. miR-222 is shown to directly bind to Gas5 and reduce its RNA level (39). We also demonstrated that Gas5 RNA level was reduced in MLN-Mϕ from normal mice transfected with miR-222 mimic. These results indicate that miR-222 plays a significant role in the reduction of Gas5 RNA expression during the M2bMϕ polarization process. miR-222 expression that was increased in MLN-Mϕ of 7 Gy GIARS mice was restored to a level shown in normal mice after the GL treatment. Also, reduced miR-222 expression and increased Gas5 RNA expression in MLN-Mϕ from GIARS mice were demonstrated when these cells were cultured with GL. These results suggest that the effect of GL on the increased Gas5 RNA expression and decreased miR-222 expression in MLN-Mϕ from GIARS mice may be the driving force that is suppressing the M2bMϕ polarization. Generally, the Gas5 RNA level is regulated by nonsense-mediated RNA decay (NMD) (42, 43). miR-222–induced reduction of Gas5 RNA was not seen in MLN-Mϕ treated with NMDI14, an inhibitor of NMD pathway (data not shown). In contrast, certain cytokines (such as CCL2, IL-10, etc.) have been described to be involved in the polarization of M2a/M2cMϕ (44, 45). In the previous reports, GL has been reported as an inhibitor of M2a/M2cMϕ polarizations (18, 25, 27). Thus, GL suppresses M2a/M2cMϕ polarizations by inhibiting CCL2 and IL-10 production. In this paper, the inhibitory effect of GL on M2bMϕ polarization and its mode of action was demonstrated, as follows: 1) severe sepsis stemming from bacterial translocation was not shown in 7 Gy GIARS mice treated with GL, and all these mice survived for 30 d or more after the irradiation; 2) a major antibacterial effector cell (M1Mϕ) against bacterial translocation appeared in MLNs of the irradiated mice treated with GL, and the polarization of M2bMϕ (an inhibitor cell on M1Mϕ polarization) in MLNs of these mice was not demonstrated; and 3) GL improved miR-222–associated Gas5 RNA reduction in MLN-Mϕ of γ-irradiated mice. In addition to miR-222, certain miRNAs (miR-21, miR-34a, and miR-135b) have been reported to bind to Gas5 (46–48). In 7 Gy GIARS mice, these miRNAs have a possibility to be involved M2bMϕ polarization. This possibility will be tested in the following studies.

In recent 16S rRNA gene amplicon sequencing analysis (49), abundance of Proteobacteria has shown to be increased almost 1000-fold in the feces of mice 4 d after 10 Gy of total body γ irradiation, and then it was returned to normal levels within 7 d of the irradiation. Abundance of Clostridia and Bacteroidetes was less affected in the feces of these irradiated mice. Because M2bMϕ polarization was started in the MLNs of 7 Gy GIARS mice 7 d after the irradiation (5), radiation-associated microbiome changes in the intestine (during 4–7 d after the irradiation) seem to be not directly linked to this Mϕ polarization (seven or more days after the irradiation).

GL interacts with various immunocompetent cells. In 7 Gy GIARS mice, however, neutrophils and lymphocytes are dramatically decreased within a few days of the irradiation. Therefore, the interaction of GL with these cells in GIARS mice is considered to be not practical. In contrast, Mϕ and some groups of innate lymphoid cells are known to be resistant against this dose of γ irradiation. Whereas NSG mice adoptively transferred with Mϕ from 7 Gy GIARS mice died within 3 d of Enterococcus faecalis oral infection, NSG mice exposed to the same pathogen survived after inoculation with Mϕ from 7 Gy GIARS mice treated with GL (F. Suzuki and M. Kobayashi, unpublished data). Therefore, we considered that Mϕ were responsible cells when 7 Gy GIARS mice survived after the GL administration. In our recent studies, Mϕ isolated from various organs including MLNs of 7 Gy GIARS mice did not show any M2bMϕ properties, and peripheral blood monocytes and peritoneal Mϕ from GIARS mice treated with GL were shown to be IL-12⁺ cells. These results indicate that various Mϕ derived from 7 Gy GIARS mice that were treated with GL are convertible to be M1Mϕ.

Because a high dose of GL has been reported to protect GI mucosal injuries in mice with nonalcohol fatty liver diseases (50), the effect of GL on GI damages of 7 Gy GIARS mice was examined in our laboratory. In the results, similar GI mucosal injuries (shortening of villus length and reduced number of crypts in the intestine) were seen in 7 Gy GIARS mice before and after treatment with the compound at a dose of 10 mg/kg. From these results, we assume that the mitigative effect of GL on GI mucosal injuries is not directly involved in the reduced mortality rate of 7 Gy GIARS mice treated with GL.

For M2bMϕ polarization in GIARS mice, we have hypothesized the following five steps: 1) high-mobility group box 1 (HMGB1) is released from damaged or dying cells influenced by the irradiation (51, 52), 2) this protein induces miR-222 expression in Mϕ (53–55), 3) increased miR-222 activate the NMD pathway (39), 4) the activation of this pathway causes decreased Gas5 RNA expression (Fig. 4D), and 5) Mϕ with reduced Gas5 RNA polarize to the M2b phenotype (11). Thus, it is indicated that resident Mϕ are polarized into M2bMϕ through the few steps after initiation by the increased level of HMGB1. GL has been proved to directly bind to HMBG1 and interfere with the binding of HMGB1 to DNA in living cells (56, 57). Also, the interaction of HMGB1 with its receptors (RAGE and TLR4) is shown to be blocked by GL (58). We have also demonstrated that serum HMGB1 levels were not increased in 7 Gy GIARS mice treated with GL (F. Suzuki and M. Kobayashi, unpublished data). When 7 Gy GIARS mice were treated i.p. with anti-HMGB1 Ab, Mϕ from these mice did not show any M2bMϕ properties, just as GIARS mice treated with GL. Therefore, we think that the effect of GL on the improved antibacterial resistance against bacterial translocation is exhibited through the inhibition of M2bMϕ polarization, which is initiated by HMGB1 released from damaged or dying cells influenced by irradiation, and GL suppresses M2bMϕ polarization through the HMGB1 blocking.

The antibacterial effects of GL against various bacterial infections through the modulation of mannose receptor expressing M2Mϕ (e.g., M2aMϕ and M2cMϕ) have already been demonstrated in mice 2–5 d after burn injury (20, 25). By controlling M2a/cMϕ polarization, the suppression of CCL2 and/or IL-10 production from neutrophils, Mϕ, and/or T cells have been shown to be involved in the antibacterial effects of GL (20). Because of the plasticity of Mϕ, M2a/cMϕ switch to a non-M2a/cMϕ when IL-4 or IL-10 is absent (59). Therefore, the lifespan of M2a/cMϕ is relatively short. In contrast, M2bMϕ live longer without any exogenous growth factors because M2bMϕ themselves produce CCL1 (13). In this study, the M2bMϕ polarization was shown to be inhibited by GL through the restoration of impaired Gas5 RNA expression in MLN-Mϕ of 7 Gy GIARS mice. So far, the pathologic roles of M2bMϕ have been positively and negatively described in various diseases, including infectious diseases (bacterial, viral, and parasitic infection), autoimmune diseases (systemic lupus erythematous, etc.), diseases in nervous systems (spinal code injury and Alzheimer disease), glycolipid metabolic disorders (obesity and type 2 diabetes mellitus), cardiovascular diseases (atherosclerosis and myocardial disease), and cancer (hepatocellular carcinoma) (60). Based on these findings, GL may have a function to control these diseases.

The authors have no financial conflicts of interest.