Berberine ameliorates aGVHD by gut microbiota remodelling, TLR4 signalling suppression and colonic barrier repairment for NLRP3 inflammasome inhibition

Abstract Berberine (BBR), an isoquinoline alkaloid, is used to treat gastrointestinal disorders as an herbal medicine in China. The aim of this study was to investigate the anti‐inflammatory activities of BBR in a mouse model with acute graft‐versus‐host disease (aGVHD). Mice were intravenously injected with bone marrow cells from donors combined with splenocytes to develop aGVHD. The body weight, survival rate and clinical scores were monitored. Then the levels of inflammatory cytokines, histological changes (lung, liver and colon), colonic mucosal barrier and gut microbiota were analysed. Moreover, the toll‐like receptor 4 (TLR4)/myeloid differentiation primary response gene 88 (Myd88)/nuclear factor‐κB signalling pathway, NLRP3 inflammasome and its cytokines’ expressions were determined. The results showed that the gavage of BBR lessened GVHD‐induced weight loss, high mortality and clinical scores, inhibited inflammation and target organs damages and prevented GVHD‐indued colonic barrier damage. Additionally, BBR modulated gut microbiota, suppressed the activation of the TLR4 signaling pathway and inhibited NLRP3 inflammasome and its cytokine release. This study indicated that BBR might be a potential therapy for aGVHD through NLRP3 inflammasome inhibition.


| INTRODUC TI ON
Allogeneic haematopoietic stem cell transplantation (allo-HSCT) is effective for many haematologic and hereditary diseases as a treatment strategy. 1 During allo-HSCT, donor-derived immunocompetent cells may cause the organs damage in recipient, which is known as GVHD. The major cause of non-relapse mortality is acute GVHD (aGVHD). 2 The pathophysiology of aGVHD involves the dysregulation of inflammatory cytokine cascade. 3 Studies have confirmed the relationship between gut microbiota and the development of aGVHD. 4 Gut microbiota plays extensive roles containing immunological functions. Gut microbiota can interact with the immune components of microbial immunity, although it is a non-immune component. The microbiota composition dysregulation or alterations would be resulted in the pathogenesis of GVHD. Before allo-HSCT, recipients undergo a conditioning regimen, which induces damage in tissue, allowing translocation of bacterial products from mucosa into the internal milieu where a "cytokine storm" is provoked. 5 However, the molecular mechanism inducing cytokine production is still poorly understood.
Increasing studies have suggested that NLRP3 inflammasome plays a key role in the shaping of gut microbiota composition. NLRP3 inflammasome, as the essential step for the activation of caspase-1, controls the activation of inflammatory cytokines. 6 The NLRP3 inflammasome activation causes the active form of the precursor protein pro-IL-1β by cleavage, and the release of inflammatory factors, such as TNFα, IFNγ and IL-1, can induce "cytokine storm" and altered gut microbiota composition, which causes vicious spiral. The clinical use of immunosuppressive agents, cytotoxic drugs, or in vitro/in vivo T-depletion are used in aGVHD, but it also results in a high relapse rate after transplantation. 7 Hence, other valid strategy for the relieve of GVHD-induced damages is necessary, NLRP3 would be the therapeutic target of GVHD.
Berberine (BBR) is extracted from the Chinese herb Coptis chinensis (Huang-Lian, a traditional Chinese herb in medicine). 8 As the major pharmacological component of Huanglian, BBR and its containing herbs are used to treat bacterial diarrhoea and other intestinal infections in China for thousands of years.
Recently, some researchers found that BBR was clinically effective in type 2 diabetes treatment and significantly total cholesterol and low-density lipoprotein cholesterol levels decreasing. 9 However, the mode of action of BBR remains a paradox due to its absorbed poorly into the blood from the intestine, gut microbiota modulation is the most likely the mechanisms of its pharmacological effect, 10 similarly to many other traditional Chinese herbal medicines. 11 In this study, we aimed to explore whether BBR had any influences on inflammatory cytokines, the pathological changes of target organs, especially gut microbiota and NLRP3 inflammasome in mice with aGVHD. Notably, this study was conductive to elucidate the pharmacological mechanisms of BBR and suggested that BBR may be a possible clinical gut microbiota modulator for the treatment of aGVHD.

| Animals
Female BALB/c and male C57BL/6 mice (5-6 weeks, weight 20-  Sigma-Aldrich). Splenocytes were isolated by a cell strainer from the spleen tissue. In GvHD induction process, after the adjustment of determined optimal dose, 2 × 10 7 BMCs combined with 2 × 10 7 splenocytes that both from either allogeneic (C57BL/6) donors were injected intravenously into the recipients (BALB/c) which treated with 9 Gy TBI, at day 0.

| aGVHD monitoring and scoring
The body weight and clinical symptoms were recorded every 4 days, and the survival was monitored every day. A clinical scoring system, described previously, was applied to evaluate the severity of GVHD, including weight loss, activity, posture (hunching), fur texture and skin integrity. 12 Experimental mice, placed in coded cages, were assessed once a week. Correspondingly, scores range from 0 to 2 were used to record the gradational changes which according to each criterion, and then created the clinical index through the sum of 5 criteria scores above (maximum, 10).

| Cytokine measurements
Serum samples were collected from whole blood without anticoagulants and centrifuged at 4℃/1000 g for 10 min to collect the serum supernatant for analysis. Then, the level of murine TNFα, IL-1β, IFNγ, MCP-1, IL-6, IL-18 (CUSABIO) and LBP (Abcam, ab269542) were assayed quantitatively via ELISA kits (CUSABIO) between day 4 and 6 after bone marrow transplantation (BMT), following the manufacture protocols. Next, we analysed microwell strips by using a microwell reader (Molecular Devices), weighed the extracted colon tissues, homogenized tissues in the tissue extraction reagent for 3 min on ice and acquired the supernatants after 4℃/10,000 g centrifugation for 5 min, then, stored at −20℃. Finally, we detected the levels of IL-1β, TNFα and IL-6 in colon homogenates through ELISA assay.

| Histology
Mice were sacrificed to harvest aGVHD target organs (lung, liver, colon) from recipients on day 25, then immediately fixed with 4% formalin and embedded into paraffin following standard procedures.
Afterwards, we mounted the sections (5 μm) on the slide for H&E staining and estimated histologic changes of liver and small intestine by using the semiquantitative scoring system (SSS) mentioned above. 13 Interstitial/alveolar inflammation and periluminal bronchial/ vascular infiltration are histological changes compatible with GVHD in the lungs. Grading categories of interstitial/alveolar inflammation: none = 0, mild = 1 (<25%), moderate = 2 (25%-50%) and severe =3 (>50%). Liver pathology severity was evaluated with GVHD grading system recomposed from Cooke, 14 which includes portal infiltrates, cell apoptosis, vascular endothelialitis, bile duct injury and lobular infiltrates (normal = 0, mild = 1, moderate = 2 and severe = 3). the histological changes in colon also evaluated by SSS as well. Changes of compatible with GVHD was defined by lamina propria inflammation, goblet cell depletion and ulceration of the colonic mucosa. All slides were coded and a Nikon Eclipse E400 microscope was utilized for visualization. Ultimately, a Zeiss Axiom camera was applied for image acquisition and analysed with software AxioVision 3.0.6 SPZ (Zeiss).
Bacterial species were identified and classified based on 16S rRNA gene sequencing. 15,16 Briefly, samples were centrifuged at 20,000 g for 30 min at 4℃ to extract the total genomic DNA. The integrity, concentration and quality of the total DNA were determined by agarose gel electrophoresis through A260 and A260 to A280 ratio.
After extraction, pooled and individual DNA samples were quantified by QuantiGene 2.0 Reagent System (Panomics/Affymetrix), following protocols.

| Western Blot analysis
After extraction of total protein with RIPA and PMSF (Beyotime),

| Real time-PCR analysis
We extracted total RNA from colonic tissues by using Trizol rea-

| Statistical analysis
Statistical analysis was performed by the SPSS 13.0 software.
All data were presented as mean values ± standard deviation (mean ± SD). Statistical analysis of the original data was evaluated with the Student's t-test. One-Way ANOVA was used for single factor analysis of variance. When variance was uniform, S-N-K method was used for the comparison among groups. As for heterogeneity of variances, it would be compared with Dunnett's method. A p value of <0.05 was taken as statistically significant.

| The administration of berberine significantly alleviates the severity of aGVHD
The administration of berberine prevented significant weight loss, while the model group body weight decreased until day 8, which may be the side-effects of TBI. Then, the weight increased until day 28 and decreased once more. The mice body weight in berberine group kept higher than the model group with statistical significance at day 32 (p = 0.0075), day 36 (p = 0.0007) and day 40 (p < 0.0001) ( Figure 1A). 100% of the mice died within 40 days after BMT, which is only 10% with treatment of berberine ( Figure 1B). Clinical symptoms of GVHD were quantified by the scores summed for each of five parameters (0-2/each parameter): weight loss, posture, skin integrity, fur texture and activity, as previously reported. 17 The higher scores mean the clinical symptoms are more severe. The scores of model group increased until day 12 and reduced until day 16, then the score displayed a stable increase. The rise in clinical score after day 28 was declined in berberine group. There were significant differences at  Figure 1C).

| Administration of berberine decreases proinflammatory cytokines in serum of aGVHD
Some studies have demonstrated that proinflammatory cytokine release plays critical role in the development of aGVHD. 18 We meas-

| Berberine attenuates the pathological changes of target organs
We also examined whether any improvement of GVHD pathology is by berberine. Since the mice in model group died mostly from day 25 after BMT, the mice were sacrificed on that day. The target organs (lung, liver and colon) were prepared as tissue sections. As shown in Figure 3A-C, the inflammatory infiltration was observed in interstitial and perivascular of lung, the alveolar spaces were filled with the exudate in the model group; however, the inflammatory infiltration decreased in the berberine group. Figure 3D

| Berberine prevented GVHD-induced colonic barrier dysfunction
To evaluate the colonic tight junction function, the expression  LBP is a major transporter of proinflammatory lipopolysaccharides (LPS) in the plasma which is often used as an indicator of colonic barrier integrity. The plasma LBP concentration increased in aGVHD, indicating colonic barrier impairment. However, berberine group was comparable with the control values, suggesting that berberine prevented colonic barrier degradation ( Figure 4I).
These data disclosed that berberine could prevent the colonic barrier impairment caused by aGVHD.

| Berberine partly recovered the intestinal dysbiosis in GVHD mice
As shown in Figure 5A, 40  were less abundant in the GVHD model mice, but only Akkermansia were more abundant in the GVHD model mice.
After berberine treatment, the gut microbiome was changed in GVHD model mice. As shown in Figure 5B, phylum Actinobacteria

| Berberine reduced the expression of NLRP3 inflammasome and TLR4 signalling
We collected colon samples (N = 3/group) to detect the expression of the NLRP3 inflammasome, TLR4 signalling and its inflammatory cytokines. In the colon, NLRP3 inflammasome was activated and its inflammatory cytokines were released in the mice model of GVHD, while berberine could inhibit the NLRP3 inflammasome ( Figure 6A-C), TLR4 signalling ( Figure 6D) and suppressed inflammatory cytokines except IL-6 ( Figure 6E,F).

| DISCUSS ION
Graft-versus-host disease, as a main complication of allo-HSCT, usually causes in severe injuries of organs and the significant mortality.
Although some strategies are applied in the GVHD therapy, it remains the major cause of mortality which approaches nearly 15% of patients after allo-HSCT. 19 Berberine was proven to be safe for human and used for treating type 2 diabetes, NAFLD (non-alcoholic fatty liver disease) and gastrointestinal disorders. 20 It has been reported that BBR, be used alone or in combination with other compound, contributes to the attenuation of GVHD in mice, 21 but its mechanism is still unclear. Previous studies had proved that BBR could alleviate colitis by suppressing the release of proinflammatory cytokines. 22 Inflammatory cytokines are considered to be important in the pathogenesis of aGVHD. 23 The strong release of proinflammatory cytokines such as TNFα, MCP-1 and IFNγ is often characterized in the transplantation model, which contributed to target organ injury in GVHD significantly. 24 As we all know, the gut microbiota plays a crucial Yet, it is believed that conditioning before allo-HSCT and translocation of bacterial products in the recipients triggers the activation of NLRP3 inflammasome located in the cytoplasm in an inactive form. The basic cellular expression of NLRP3 inflammasome elements is regulated by a priming signal described as "signal 1". 32,33 This signal is continuously delivered to the cells by LPS released from intestinal Gram-negative bacteria after engaging TLR4 and its downstream effectors, including MyD88 and NF-κB, which leads to the transcription of NLRP3 inflammasome components in an NFκb transcription factor-dependent manner. 34 In our study, we found that BBR can downregulate TLR4 signalling and remodel gut microbiota which NLRP3 inflammasome would not be primed by "signal 1". In contrast to priming "signal 1", functional activation of the synthesized NLRP3 inflammasome is mediated by "signal 2", which is related to infection, cell activation, or cell/tissue damage. 35,36 Otherwise, we disclosed the dys- microbiota over the impaired intestinal barrier provokes inflammation. 37,38 The mucosal damage is the bystander effect of cytokines released in syngeneic implants. 39 Animal studies have showed that the mortality of mice decontaminated the gut by antibiotics after BMT obviously decreased; the bystander effect was abrogated by the gut microbiota normalization. 40 Protection of the mucosal barrier was also demonstrated to result in a lower incidence of GVHD. 41 BBR was reported to protect the intestinal mucosal barrier, 42 and might inhibit NLRP3 inflammasome by "signal 2", which were also demonstrated in our results.
Through our work, we discovered that BBR significantly inhibited NLRP3 inflammasome activation and its inflammatory cytokines expression by gut microbiota remodelling and intestinal mucosal barrier protection, which involved in both "signal 1" and "signal 2" of NLRP3 inflammasome activation in aGVHD mice.
However, there are some limitations of this study, for example, the mechanism needs to be shown and the gut microbiota change in GVHD mice after BBR treatment is still to be explored and clarified in future study.

| CON CLUS ION
Our results demonstrated that BBR significantly remodelled the gut microbiota to suppress TLR4 signalling and impaired gut barrier, which inhibited the vicious circle of NLRP3 activation and inflammatory cytokines release induced by GVHD. We have suggested a new molecular mechanism of BBR on GVHD protective through its inhibition of NLRP3 inflammasome primed and activated by "signal 1" and "signal 2". Our finding indicated that BBR is a potential candidate for the GVHD treatment.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I LI A B I LIT Y S TATEM ENT
The raw data of Miseq sequences from 15 mice can be checked on