Purified baicalin (>95%, CAS No.: 21967-41-9), paeoniflorin (>98%, CAS No.: 23180-57-6), gallic acid (>98%, CAS No. 149-91-7), berberine (>98%, CAS No.: 633-65-8), coptisine (>98%, CAS No.: 6020-18-4), palmatine (>98%, CAS No.: 10605-02-4), jatrorrhizine (>98%, CAS No.: 222-817-3) and baicalein (>98%, CAS No.: 491-67-8) were purchased from Sigma (St. Louis, MO, USA). All chemicals were dissolved in sterile saline to the concentration of 10 mg/ml. Dextran sulfate sodium (DSS, Cat. No. 45-42867-5G) was purchased from Sigma and dissolved in 0.9% NaCl to final 1 mg/ml. KD was purchased from Tianjin Tasly Pharmaceutical Co., Ltd. (batch number 60060101, Tianjin, China) and prepared as follows (Voucher number): 30 g of A. mongholicus Bunge (20100478, Figure S1 ), 40 g of H. diffusa Willd. (PS1034MT01, Figure S2 ), 10 g of C. undulatumn (Nutt.) Spreng ( Figure S3 ), 10 g of C. setosum Willd. M. Bieb (PS0611MT04, Figure S4 ), 40 g of P. vulgaris Mill ( Figure S5 ), 10 g of P. vulgaris L. subsp. Vulgaris (MIB: ZPL:03560, Figure S6 ), 10 g of C. chinensis Franch (PS0915MT01, Figure S7 ), 15 g of P. cuspidatum Siebold & Zucc. ( Figure S8 ), 15 g of A. lancea (Thunb) DC (SKP 051011201, Figure S9 ), and 10 g of G. glabra L.) (8600100, Figure S10 ). The herbs were rinsed and soaked in 2-liter cold water for 30 min, and then boiled in water under reflux for 2 h. After the concentration and filtration, the medicinal solution was concentrated to 1.0 g/ml of crude drug, and stored in a refrigerator at 4°C. The required medicinal recipe was provided by the Chinese Medicine Bureau of the First Affiliated Hospital of Jinzhou Medical University.

The main ingredients in KD were detected by using a Shimadzu high pressure liquid chromatography (HPLC, Kyoto, Japan). The standards (0.002 g gallic acid, 0.002 g Paeoniflorin, 0.005 g baicalin, 0.003 g berberine, 0.004 g coptisine, 0.003 g palmatine, 0.004 g jatrorrhizine, 0.005 g baicalein and 0.005 g sulfasalazine) were placed in the same 20 ml of flask, dissolved in methanol and diluted to 0.01, 0.02 and 0.02 mg/ml, respectively. Two-gram KD was placed into a mortar and ground into fine powder, and placed in a 30 m flask. Thirty milliters of methanol was added and sonicated for 1 h. The dominant ingredients of KD were subsequently eluted by using the following chromatographic conditions: Column, Amethyst C18-H (4.6 mm × 250 mm, 5 µm, 120 Å); mobile phase: acetonitrile (A) −0.4% phosphoric acid aqueous solution (B), gradient elution (0–10 min, 4% A; 10–20 min, 4% A → 16% A; 20–35 min, 16% A → 55% A; Flow rate: 1.0 ml/min; Detection wavelength: 250 nm; Column temperature: 25 °C and Injection volume: 10 μl. Rat serum was also analyzed using the same condition to investigate the levels of sulfasalazine and the main chemicals of KD in the rats.

The UPLC-MS/MS analysis was performed using UPLC/MS/MS ACQUITY Xevo TQ from Waters (Milford,.MA, USA). The following UPLC condition was used: The chromatographic column a Fortis Xi C18 column (50 mm×2.1 mm, 1.7 μm); the mobile phase acetonitrile (A) −0.1% formic acid aqueous solution (B), gradient elution (0–1 min, 10%A; 1–4 min, 10% → 45%A; 4–7 min, 45% → 60%A; 7–8 min, 60%A; 8–8.5 min, 60% → 10%A); Flow rate: 0.3 ml/min; column temperature: 40°C; total running time 10 min; and injection volume 2 μl. The following MS conditions were used: dissolvent gas: nitrogen (600 L/h); capillary voltage: 2.5 kV (positive ion mode); ion source: ESI source; ion source temperature: 150°C and dissolvent temperature: 500°C. Multiple reaction monitoring (MRM) method for quantitative analysis. Multiple reaction monitoring (MRM) transitions m/z 171.1 → 125.0, m/z 399.2 → 367.1, m/z 481.2 → 198.0, m/z 271.2 → 240.1, m/z 337.0 → 320.1, m/z 353.0 → 336.1, m/z 339.0 → 322.1, m/z 321.0 → 292.1, m/z 271.1 → 123.0 and m/z 446.1 → 269.0 were selected for determining gallic acid, sulfasalazine, paeoniflorin, emodin, berberine, palmatine, jatrorrhizine, and coptisine, baicalein and baicalin.

All animal-related procedures were approved by the Institutional Animal Care and Use Committee of Jinzhou Medical University according to international, national and institutional rules considering animal experiments (Jinzhou, China). Sprague–Dawley (SD) rats (8 weeks old; weighing 200–220 g) were purchased from the animal center of Jinzhou Medical University (Jinzhou, China). The rats were housed in a standard cage under light/dark 12:12, with the light on from 6 am to 6 pm, and given a standard laboratory diet containing 23% protein and water ad libitum. UC model was induced by using the following protocol. The rats received 5% dextran sodium sulfate (DSS) in distilled water for 5 d, and changed to tap water for 10 d. The whole process was repeated three times and a chronic UC model was established. The general condition, body weight, stool characteristics, and occult blood or gross bloody stool of the rats were observed daily. The mortality rate reached up to 30% after three times of DSS treatments. The dead rats were excluded from the present study. After the modeling was completed, two rats in the normal group and the model group were sacrificed. The histopathological scores of the two groups of rats were compared and comprehensively analyzed to determine whether the model was successfully induced.

After 7-day model establishment, all rats administered twice daily by gavage and divided into 6 groups and treated orally, the vehicle (CG, 3-ml saline solution daily), UC model (UG, 3-ml saline solution daily), sulfasalazine [SG, 5 mg/kg sulfasalazine daily (Soliman et al., 2019)], low-dose KD (LG, 1 ml/kg daily), middle-dose KD (MG, 2 ml/kg) and high-dose KD (HG, 10 ml/kg) groups. After 15-day gavage, all rats were sacrificed.

Ten milligrams of colon tissue was ground in liquid nitrogen and lysed by using RIPA lysis (CST, Danvers, MA, USA). The supernatant was prepared via centrifugation at 12,000×g for 10 min and stored at −20°C for ELISA measurement. The levels of tumor necrosis factor α (TNFα) (ab100747), interleukin (IL)-1 β (ab100704), IL-6 (ab100713) and IL-10 (ab100697) were measured by using the ELISA kits from Abcam (San Francisco, CA, USA). The levels of superoxide dismutase (SOD) (#MBS080359) and catalase (CAT) (#MBS775862) and glutathione peroxidase (GSPx) (#MBS049725) were also evaluated using the kits from MyBioSource, Inc. (San Diego, CA, USA). All biochemical indexes were measured on an automatic chemical analyzer (Hitachi, Tokyo, Japan). MDA was measured by using UV–Vis spectroscopy at 267 nm in 10 mM Na₃PO₄ buffer pH 11 according to the molar absorbing coefficient 31,500 M⁻¹cm⁻¹ (Esterbauer et al., 1991). Wavelength scanning of MDA fractions was conducted via the wavelength against a control group without 1,3-propanediol and MDA.

The clinical assessment of disease severity was evaluated using a scoring procedure every day by combining scores of weight loss, stool consistency and fecal bleeding. DAI is usually calculated as the sum of the body weight loss (scored as: 0, none; 1, 1–5%; 2, 5–10%; 3, 10–20%; 4, over 20%), presence or absence of fecal blood (scored as: 0, negative hemoccult; 2, positive hemoccult; 4, gross bleeding) and stool consistency (scored as: 0, well-formed pellets; 2, loose stools; 4, diarrhea) (Schwanke et al., 2013). On day 15, the rats were sacrificed through intraperitoneal injection of phenobarbital sodium (50 mg/kg) and their colon length was compared among different groups. Colon tissue injuries were detected by following histopathological analysis.

For SEM observation, 5 mm² of pieces were cut from each rat after 15-day KD treatment and fixed with 1% osmium tetroxide for 2 h at 4°C. The tissues were rinsed, dehydrated in ethyl alcohol, dried with carbon dioxide, covered with gold, and examined under SEM JSM-6610lv (Jeol, Japan) with an INCA SDD X-MAX energy dispersive micro-analyzer.

UC or colon tissues were extracted after a 15-day model establishment. UC tissues were fixed in 4% paraformaldehyde and embedded in paraffin and remaining tissues were stored at −80 °C. The embedded pancreatic tissues were cut into 2–3 μm slices and stained with hematoxylin and eosin (H&E). The numbers of inflammatory cell infiltration, bleeding and necrotic cells were calculated. The severity of colon tissue damage was assessed by using pathological score = edema score + necrosis score + inflammatory cellular infiltration score + bleeding score. Five slices were evaluated in each group.

Immunohistochemistry analysis was conducted to evaluate the expression of TLR4, p-PI3K, p-AKT, and p-NF-κB. The paraffin-embedded tissue sections were deparaffinized and treated with hydrogen peroxide (3 m/v) for 15 min to remove endogenous peroxidase. Antigen retrieval was performed by blocking the samples in goat serum for 10 min at 22°C. The following antibodies were added and incubated 12 h at 4°C, including anti-TLR4 antibody (ab13867, 1:500), anti-p-PI3K p85 antibody (ab86714, 1:500), anti-p-AKT antibody (ab38449, 1:500), and or anti-p-NF-κB antibody (ab86299) from Abcam (Cambridge, MA, USA). Anti-PI3K antibody (ab133595), anti-AKT antibody (ab8805), and or anti-NF-κB antibody (ab28849) were also purchased from Abcam and used for the following Western Blot. A biotin-labeled goat anti-rabbit IgG secondary (1:1,000) was added followed by incubation at 37°C for 10 min. The slides were then incubated at 37°C for 10 min with peroxidase-conjugated streptavidin (Sigma, S5512). The sections were stained with 3,3’-diaminobenzidine (DAB, Sigma) and counterstained with hematoxylin (Sigma). The color separation was conducted by using 2% hydrochloride and alcohol, followed by 15-min washing. Each sample was observed in five fields and target protein signals were stained with brown. The positive rates were calculated as the number of positive cells/number of total cells by using an image analyzer (Image-Pro Plus 5.1, MediaCybernetics, MD, USA).

RNA was extracted from 5-mg colon tissues using TRIzol reagent (TIANGEN, Beijing, China). cDNA was prepared by using a reverse transcription kit (Bioteck, Beijing, China) according to the manufacturer’s instructions. The following primers were used: TLR4 forward primer 5’-CATGGCATTGTTCCTTTCCT-3’ and reverse primer 5’-CATGGAGCCTAATTCCCTGA-3’; PI3K forward primer 5’-TTAAACGCGAAGGCAACGA-3’ and reverse primer 5’-CAGTCTCCTCCTGCTGTCGAT-3’; AKT forward primer 5’-AAAGAGCGCATGAGTGGACG-3’ and reverse primer 5’-CGTGGTCCTCCTTGTAGTAG-3’; NF-κB forward primer 5’-AGAGCAACCGAAACAGAGAGG-3ʹ and reverse primer 5’-TTTGCAGGCCCCACATAGTT-3’ and β-actin forward primer 5’-AAGTCCCTCACCCTCCCAAAAG-3’ and reverse primer 5’-AAGCAATGCTGTCACCTTCCC-3’. qPCR was conducted using SYBR-Green (Invitrogen, USA) (dilution 1:1,000 with deionized water) for 5 min on an Applied Biosystems StepOne Plus real-time PCR machine (Applied Biosystems, Inc., CA, USA). The levels were detected and the relative mRNA levels were normalized to β-actin using the ΔΔCt method.

Ten milligram colon tissues were ground in liquid nitrogen and total protein was extracted by using RIPA lysis (CST, Danvers, MA, USA). Protein concentration quantified using the BCA kit (TaKaRa, Dalian, China). HRP-conjugated goat anti-rabbit IgG H&L (ab6721) secondary antibodies were from Abcam (Abcam, San Francisco, CA, USA). The proteins were separated by SDS-PAGE and transferred to the PVDF membrane in the transfer buffer at 100 V for 2–3 h. The membrane was blocked for 1 h at ambient room temperature in 10% non-fat milk, and probed with antibodies against the above primary antibodies for 2 h at 37°C, rinsed four times with PBTB, incubated 2 h at 37°C in secondary antibodies, washed extensively in PBS. Images were acquired on an Odyssey CLx infrared scanner (Li-Cor- Nebraska USA). Relative protein levels were calculated by using internal reference β-actin.

About 10-mg fresh feces were obtained from each rat after a 15-day KD administration. The genome of gut microbiota was isolated using a FastDNA Spin Kit (Qbiogen, Carlsbad, CA, USA). 16S rRNA was amplified by PCR using forward primer, 5’-GAGAGTTTGATCCTGGCTCAG-3’ and the reverse primer, 5’-GGTTACCTTGTTACGACTT-3’. Gut microbiota was analyzed by using 16S rRNA sequencing. Heatmap and taxon relative abundance bar diagram was created by using custom R scripts and ggplot2.

Data are presented as the means ± standard error of the mean (SEM) and analyzed using the SPSS 21.0 software (SPSS, Inc., Chicago, IL, USA). Student’s t-test and one-way analysis of variance (ANOVA) with post-hoc Tukey’s tests were used to evaluate the variables between groups. Paeoniflorin and baicalin occupies the main proportion of KD and the Pearson correlation coefficient test was used to explore the relationship between serum levels and oxidative, and or inflammatory factors. The statistical difference was significant if the value of P <0.05.

HPLC analysis indicated that elution time of standards gallic acid (8.26 min), sulfasalazine (10.01 min), paeoniflorin (11.05 min), emodin (13.98 min), berberine (15.41 min), coptisine (15.96 min), palmatine (16.38 min), jatrorrhizine (16.87 min), baicalein (25.91 min) and baicalin (28.68 min) ( Figure 1A ), respectively. The main components of KD consist of gallic acid (12.6 ± 1.0 mg/g), paeoniflorin (21.2 ± 1.8 mg/g), emodin (13.0 ± 1.2 mg/g), berberine (14.3 ± 1.1 mg/g), coptisine (13.4 ± 0.8 mg/g), palmatine (14.3 ± 0.7 mg/g), jatrorrhizine (15.7 ± 1.0 mg/g), baicalein (14.1 ± 0.4 mg/g) and baicalin (126.6 ± 2.1 mg/g) ( Figure 1B ). The results imply that KD possibly displays its function via these important compounds.

The HPLC results of standards and KD extracts were further confirmed by those from UPLC-MS/MS. UPLC-MS/MS analysis showed that elution time of standards gallic acid (2.12 min), sulfasalazine (2.81 min), paeoniflorin (3.24 min), emodin (3.82 min), berberine (4.31 min), coptisine (4.72 min), palmatine (4.93 min), jatrorrhizine (5.12 min), baicalein (7.03 min) and baicalin (8.25 min) ( Figure 2A ), respectively. The main components of KD consist of gallic acid, paeoniflorin, emodin, berberine, coptisine, palmatine, jatrorrhizine, baicalein and baicalin with the same elution time ( Figure 2B ).

Serum main chemicals (gallic acid, paeoniflorin, emodin, berberine, coptisine, palmatine, jatrorrhizine, baicalein and baicalin) of KD were detected in the LG, MG and HG groups, respectively. Meanwhile, the proportion of these chemicals were increasing from the LG to the HG group ( Figure 3 ). High-level of sulfasalazine (13.5 ± 1.8 µg/ml) was found in the SG group ( Figure 3 ). No main chemical of KD and sulfasalazine were found in the serum from other groups ( Figure 3 ).

After the establishment of the UC model, the levels of SOD ( Figure 4A ), CAT ( Figure 4B ), and GSPx ( Figure 4C ) were decreased while MDA ( Figure 4D ) was increased in the UG group when compared with the CG group (P <0.05). Sulfasalazine and KD treatment increased the levels of SOD ( Figure 4A ), CAT ( Figure 4B ), and GSPx ( Figure 4C ) and reduced the level of MDA ( Figure 4D , P <0.05). Especially, KD caused the changes of oxidative stress biomarkers in a dose-dependent way and the changes in the HG group were higher than those in the SG group ( Figure 4 , P <0.05).

After the establishment of UC model, the levels of TNFα ( Figure 5A ), IL-1 ( Figure 5B ), and IL-6 ( Figure 5C ) were increased while the level of IL-10 ( Figure 5D ) was reduced in the UG group when compared with the CG group (P <0.05). Sulfasalazine and KD treatment reduced the level of TNFα ( Figure 5A ), IL-1 ( Figure 5B ), and IL-6 ( Figure 5C ) and increased the level of IL-10 ( Figure 5D , P <0.05). Especially, KD caused the changes of inflammatory cytokines in a dose-dependent way and the IL-10 level in the HG group was increased more than that in the SG group ( Figure 5D , P <0.05).

After the establishment of the UC model, the DAI scores of UC rats were increased in the UG group when compared with those in the CG group ( Figure 6A , P <0.05). Sulfasalazine and KD treatment reduced their DAI scores when compared with those in the UG group ( Figure 6A , P <0.05). Especially, KD caused a reduction in the DAI scores in a dose-dependent way ( Figure 6A , P <0.05).

After the establishment of UC model, the colon length of UC rats was reduced in the UG group when compared with that in the CG group ( Figure 6B , P <0.05). Sulfasalazine and KD treatment increased their colon length when compared with that in the UG group ( Figure 6B , P <0.05). Especially, KD caused the increase in the colon length in a dose-dependent way ( Figure 6B , P <0.05).

SEM analysis showed that after the establishment of UC model, the intestinal barriers of UC rats were damaged with reduced amounts of intestinal villi in the UG group when compared with those in the CG group ( Figure 7A , P <0.05). Sulfasalazine and KD treatment increased the amounts of intestinal villi when compared with those in the UG group ( Figure 7A , P <0.05). Especially, KD caused an increase in the amounts of intestinal villi in a dose-dependent way ( Figure 7A , P <0.05). After the establishment of UC model, the structure of intestinal mitochondria in UC rats was damaged with the reduced number of ridges in the UG group when compared with that in the CG group ( Figure 7B , P <0.05). Sulfasalazine and KD treatment increased the amounts of mitochondrial ridges and repaired the mitochondrial structure when compared with those in the UG group ( Figure 7B , P <0.05). Especially, KD caused an increase in the number of mitochondrial ridges in a dose-dependent way ( Figure 7B , P <0.05).

After the establishment of the UC model, the infiltration of the colon by immature neoplastic myeloid ( Figure 8 ) and pathological scores ( Figure 8 ) were increased in the UG group when compared with those in the CG group (P <0.05). Sulfasalazine and KD treatment reduced the infiltration of pancreas by immature neoplastic myeloid and pathological scores ( Figure 8 , P <0.05). Especially, KD caused a reduction in the pathological scores in a dose-dependent way ( Figure 8 , P <0.05).

IHC analysis showed that the brown staining of TLR4 was increased in the UG group when compared with that in the CG group ( Figure 9A , P <0.05). Sulfasalazine and KD treatment reduced the brown staining and KD caused the changes in a dose-dependent way ( Figure 9A , P <0.05). Similarly, the brown staining of p-P3IK was increased in the UG group when compared with that in the CG group ( Figure 9B , P <0.05). Sulfasalazine and KD treatment reduced the brown staining and KD caused the changes in a dose-dependent way ( Figure 9B , P <0.05). In contrast, the brown staining of p-AKT increased in the UG group when compared with that in the CG group ( Figure 9C , P <0.05). Sulfasalazine and KD treatment reduced the brown staining and KD caused the changes in a dose-dependent way ( Figure 9C , P <0.05). The brown staining of p-NF-κB was increased in the UG group when compared with that in the CG group ( Figure 9D , P <0.05). Sulfasalazine and KD treatment reduced the brown staining and KD caused the changes in a dose-dependent way ( Figure 9D , P <0.05). Quantity analysis also showed that the UC model establishment reduced the protein levels of TLR4 ( Figure 9E ), p-P3IK ( Figure 9F ), p-AKT ( Figure 9G ) and p-NF-κB ( Figure 9H , P <0.05). Sulfasalazine and KD treatment reduced the protein levels of TLR4 ( Figure 9E ), p-P3IK ( Figure 9F ), p-AKT ( Figure 9G ) and p-NF-κB ( Figure 9H ) in a dose-dependent way (P <0.05).

The RT-qPCR analysis showed that the relative mRNA levels of TLR4 was increased in the UG group when compared with those in the CG group ( Figure 10A , P <0.05). Sulfasalazine and KD treatment reduced the levels and KD caused the changes in a dose-dependent way ( Figure 10A , P <0.05). Similarly, relative mRNA level of P3IK was reduced in the UG group when compared with that in the CG group ( Figure 10B , P <0.05). Sulfasalazine treatment reduced the level further while KD treatment increased the level ( Figure 10B , P <0.05). In contrast, relative mRNA level of AKT was reduced in the UG group when compared with that in the CG group ( Figure 10C , P <0.05). Sulfasalazine and high-dose KD did not change the level and low-dose and middle-dose KD treatment increased the level ( Figure 10C , P <0.05). Relative mRNA level of NF-κB was reduced in the UG and HG groups while sulfasalazine, low-dose and middle-dose KD treatment increased the level of NF-κB ( Figure 10D , P <0.05). KD treatment affected relative mRNA levels of TLR4/PI3K/AKT/NF-κB in the UC model.

Western Blot analysis showed that relative protein level of TLR4 was increased in the UG group when compared with the CG group ( Figure 11A , P <0.05). Sulfasalazine and KD treatment reduced the level of TLR4 and KD caused the changes in a dose-dependent way ( Figure 11A , P <0.05). Similarly, relative protein level of p-P3IK was increased in the UG group when compared with that in the CG group ( Figure 11B , P <0.05). Sulfasalazine and KD treatment reduced the level and KD caused the changes in a dose-dependent way ( Figure 11B , P <0.05) while relative protein level of P3IK was less significant among all groups ( Figure 11C ). In contrast, relative protein level of p-AKT increased in the UG group when compared with that in the CG group ( Figure 11D , P <0.05). Sulfasalazine and KD treatment reduced the level of p-AKT and KD caused the changes in a dose-dependent way ( Figure 11D , P <0.05) while relative protein level of AKT was less significant among all groups ( Figure 11E ). Relative protein level of p-NF-κB was increased in the UG group when compared with that in the CG group ( Figure 11F , P <0.05). Sulfasalazine and KD treatment reduced the level of p-NF-κB and KD caused the changes in a dose-dependent way ( Figure 11F , P <0.05) while relative protein level of NF-κB was less significant among all groups ( Figure 11G ). The results suggest that KD treatment reduces relative protein levels of TLR-dependent phosphorylated PI3K/AKT/NF-κB in the UC model. On the other hand, the changing trend for the ratios of p-P3IK/P3IK ( Figure 11H ), p-AKT/AKT ( Figure 11I ) and p-NF-κB/NF-κB ( Figure 11J ) was similar with the changing trend of phosphorylating status of p-P3IK ( Figure 11B ), p-AKT ( Figure 11E ) and p-NF-κB ( Figure 11G ).

For oxidative stress biomarkers, the increase in the serum levels of paeoniflorin would cause the increase in the levels of SOD ( Figure 12A ), CAT ( Figure 12B ) and GSPx ( Figure 12C ), and the reduction in the levels of MDA ( Figure 12D , P <0.001). Similarly, the increase in the serum levels of baicalin would cause an increase in the levels of SOD ( Figure 12E ), CAT ( Figure 12F ) and GSPx ( Figure 12G ), and the reduction in the levels of MDA ( Figure 12H , P <0.001). All these results suggest that the levels of paeoniflorin and baicalin had a strong correlation with the levels of inflammatory and oxidative stress biomarkers.

Correlation test showed that an increase in the serum levels of paeoniflorin would cause the reduction in the levels of TNFα ( Figure 13A ), IL-1β ( Figure 13B ) and IL-6 ( Figure 13C ), and the increase in the levels of IL-10 ( Figure 13D , P <0.001). Similarly, the increase in the serum levels of baicalin would cause a reduction in the levels of TNFα ( Figure 13E ), IL-1β ( Figure 13F ) and IL-6 ( Figure 13G ), and the increase in the levels of IL-10 ( Figure 13H , P <0.001).

Bar plot showed that the proportion of Alloprevotella, Treponema, Prevotellaceae, and Prevotella was decreased and the levels of Escherichia_Shigella and Desulfovibrio were increased in the UG group when compared with the CG group ( Figure 14A ). KD treatment increased the proportion of Alloprevotella, Treponema, Prevotellaceae, and Prevotella and decreased the proportion of Escherichia_Shigella and Desulfovibrio in the UC model ( Figure 14A ). Heatmap also showed that the proportion of Alloprevotella, Treponema, Prevotellaceae, and Prevotella were decreased and the proportion of Escherichia_Shigella and Desulfovibrio were increased in the SG group when compared with the CG group ( Figure 14B ). KD treatment increased the proportion of Alloprevotella, Treponema, Prevotellaceae, and Prevotella and decreased the proportion of Escherichia_Shigella and Desulfovibrio in the model ( Figure 14B ). The results suggest that KD treatment improved gut microbiota in the UC model.