DSS (36 to 50 kD) was from MP Biomedicals company (160110, CA, USA). The H&E staining kit was from Solarbio (G1120, China). The total RNA kit was from Omega Bio-Tek (GA, USA). PrimeScript RT Master Mix was from Takara (RR036A, Tokyo, Japan). Forget-Me-Not™ qPCR Master Mixes were from Biotium (CA, USA). Antibodies for Csf1r (sc-46662), Src (sc-8056), p-Src (sc-166860), GAPDH (sc-32233), and ACTB (sc-47778) were from Santa Cruz (CA, USA). Antibody for F4/80 (GTX41205) was from GeneTex (CA, USA). Goat anti-rabbit IgG (H+L)-HRP conjugate (#1706515) and goat anti-mouse IgG (H+L)-HRP conjugate (#1706516) were from Bio-Rad (Hercules, CA, USA). Alexa Fluor^® 647 AffiniPure Goat Anti-Rat IgG (H+L) was from Jackson ImmunoResearch (Pennsylvania, USA). Antifade Mounting Medium with DAPI (#P0131) was ordered from Beyotime Biotechnology (Shanghai, China). The Metal Enhanced DAB Substrate Kit was from Thermo Fisher Scientific (MA, USA). The Stool DNA Isolation Kit (27600) was from Norgen Biotek (Ontario, Canada). Cell Counting Kit-8 (CCK8, GK10001) was from GlpBio (CA, USA). Phorbol 12-myristate 13-acetate (PMA, HY-18739) was from MedChemExpress (NJ, USA). Fetal bovine serum (FBS, 10270106), penicillin–streptomycin (10,000 U/mL, 15140122), and 0.25%Trypsin-EDTA (25200–72) were from Gibco (NY, USA). PES filter (0.22 μm; SLGP33RB) was from Millipore (MA, USA). LPS (L6529) was ordered from Sigma (Wicklow, Ireland). All reagents and kits purchased were used according to the manufacturers’ recommendations.

C. chinensis Franch., S. baicalensis Georgi, P. chinense Schneid., and G. jasminoides Ellis were purchased from Henan Qianfang Pharmaceutical Industry Company (China). Preparation of the HLJD extract was conducted as below. Briefly, a mixture of four herbs at a weight ratio of 3:2:2:3 was boiled using distilled water twice (1:10 w/v and 1:8 w/v). The filtrates were collected and concentrated using a rotary evaporator at 60°C, followed by freeze drying in the vacuum freeze dryer. The main components of HLJD extracts were identified using Ultra-Performance Liquid Chromatography Quadrupole-Time-of-Flight Mass Spectrometry (UPLC-Q-ToF-HRMS) as described in our accepted paper (24).

C57BL/6J mice (male, approximately 6 weeks old) were obtained from Charles River (Zhejiang, China). Mice were kept under specific pathogen-free conditions in a 12-h light–dark cycle with standard rodent chow and drinking water ad libitum. All animals were housed and received humane care at The Hong Kong Polytechnic University Shenzhen Research Institute Centralized Animal Facility. Mice were allowed to acclimatize to the environment for 2 weeks before being randomly grouped into five groups: Normal control group (NC), HLJD group (2 g/kg, H), DSS-induced colitis group (DSS), DSS+HLJD-low dosage (2 g/kg, DHL), DSS+HLJD-high dosage (4 g/kg, DHH). DSS-induced colitis was triggered by the following protocol: 1% (w/v) DSS in drinking water for 5 days, replaced with normal drinking water for 2 days—this cycle was repeated for 8 weeks. From the second week, HLJD (low, 2 g/kg and high, 4 g/kg) was administrated by oral gavage. HLJD dosage was selected based on published papers (25, 26). At the experimental endpoint, face samples from each mouse were collected. When mice were sacrificed, colon and spleen samples were isolated for further analyses.

Mice from each group were observed and recorded weekly for body weight changes, stool blood, and stool consistency scores. Disease activity index (DAI) were calculated as previously described (27).

H&E staining was performed using both the distal and proximal end of colon tissues according to the manufacturer’s manual. The histological assessment was conducted according to Table 1 .

Paraffin sections of colon tissue (5 μm) were de-paraffinized, rehydrated, antigen retrieval solution processed, and blocked with 5% BSA at room temperature for 1 h, followed by incubation with the primary antibody of F4/80 at a dilution of 1:500 in a humidified chamber overnight at 4°C. After washing with PBST, the sections were incubated with Alexa Fluor^® 647 AffiniPure Goat Anti-Rat IgG (H+L) (1:500) for 2 h at room temperature, followed by PBST washing. All sections were counterstained with Antifade Mounting Medium with DAPI. Images were obtained using a Nikon confocal microscope (#FV3000, Olympus, Tokyo, Japan). Immunostaining quantification was completed using ImageJ software.

Total RNA was isolated and cDNA was synthesized using the Takara RT Master Mix, followed by quantitative real-time polymerase chain reaction (qPCR) using Forget-Me-Not™ qPCR Master Mixes in a CFX96 Touch™ Real-Time PCR (Bio-Rad, CA, USA). Expression of tumor necrosis factor alpha (Tnfa), interleukin-6 (Il6), transforming growth factor beta 1 (Tgfb), chemokine (C-X-C motif) ligand 1 (Cxcl1), chemokine (C-X-C motif) ligand 10 (Cxcl10), chemokine (C-C motif) ligand 2 (Ccl2), chemokine (C-C motif) ligand 3 (Ccl3), Csf1r, Csf1, and interleukin 34 (Il34) was quantified, and the relative expression of these genes was normalized to Actb. Cellular expression of CD11b, CD14, Csf1, Il34, Csf1r, Tnfa, Il6, iNOS, Ccl2, Tgfb, Arg-1, Vegf, CD206, and Cxcl10 was quantified and normalized to Gapdh. The primer sequences for mouse genes are shown in Table 2 while the primer sequences for human genes are shown in Table 3 .

Expression of Csf1r (1:1,000), p-Src (1:1,000), Src (1:2,000), Gapdh (1:3,000), or ACTB (1:3,000) was detected using a previously described Western blotting protocol (46). ImageJ was used to quantify the intensity of resulting bands.

Immunohistochemical (IHC) staining was performed as described previously (47). The primary Csf1r antibody and F4/80 antibody were used at a dilution of 1:200 and 1:500, respectively.

Stool samples were collected from each mouse at the experimental endpoint. Stool DNA was isolated and quantified as previously described (27). Hypervariable regions V3–V4 were amplified using primers 338F and 806R ( Table 2 ). Amplicon was sequenced using the Illumina MiSeq PE300 platform (Majorbio, China). The raw data are available on NCBI Sequence Read Archive database (SRA: PRJNA955002), with the link https://dataview.ncbi.nlm.nih.gov/object/PRJNA955002.

The human leukemic cell strain THP-1 and mouse macrophage cell line RAW264.7 were obtained from Prof. Yanxiang Zhao from The Hong Kong Polytechnic University. THP-1 cells were grown in RPMI-1640 medium supplemented with 10% FBS and 1% penicillin–streptomycin. RAW264.7 cells were cultured in high-glucose DMEM medium supplemented with 10% FBS and 1% penicillin–streptomycin. Cells were kept at 37°C with 5% CO₂.

HLJD freeze-dried powder was dissolved in DMSO and filtered with a 0.22-μm PES filter. For the cell viability assay, THP-1 or RAW264.7 cells were seeded in a 96-well plate at a density of 3 × 10⁵ and 1 × 10⁵ cells/mL, respectively. Different doses of HLJD (3.125/6.25/12.5/25/50/100/200 µg/mL) were exposed to cells for 24 h, followed by CCK8 incubation for 1 h at 37°C with 5% CO₂. Plates were read at 450 nm.

To induce differentiation of monocytes THP-1 into macrophages, PMA at the concentration of 80 ng/mL were administrated to THP-1 cells (48) with or without HLJD intervention. After 24-h PMA stimulation, cells were imaged using a phase contrast microscope (ECLIPSE Ts2-FL NIKON).

LPS (10 ng/mL) were exposed to RAW264.7 cells for 4 h to elicit M1 phenotype macrophage. M1 and M2 markers, as well as CSF1R-related molecules, were detected in HLJD (25 µg/mL) alone or in combination with LPS-treated cells.

Statistical analysis was performed with GraphPad Prism v 8.4.2. Data are presented as means ± SEM. Brown–Forsythe and Welch analysis of variance (ANOVA) followed by Dunnett’s multiple comparisons test or two-way ANOVA followed by Tukey’s multiple comparisons test were used to compare the mean values as well as significant difference among groups. Certain tests were also mentioned in the figure legends. p < 0.05 was considered statistically significant.

To evaluate the efficacy of HLJD formulation on colitis, a chronic colitis mouse model was established using 1% DSS in drinking water continuously for 8 weeks. HLJD decoction at both low and high doses (2and 4 g/kg, respectively) were administrated from the second week ( Figure 1A ). Mice treated with HLJD alone were included to evaluate any side effects of this drug (group H). In comparison with the NC group, there was significantly less body weight gain in the DSS group, while HLJD at a low dose (DHL) had a trend to rescue the phenotype ( Figure 1B ) (p < 0.0001 for DSS vs. NC, p = 0.0053 for DHL compared with DSS). Consistent with our ongoing anti-obesity study, HLJD at a high dose (DHH) could decrease the body weight gain. In addition, mice have significantly shorter colons in the DSS group compared with the NC group, while this phenotype was improved in DHL and DHH groups ( Figures 1C, D ) (p < 0.0001 for DSS vs. NC, p = 0.0547 and 0.0232 for DHL and DHH compared with DSS). DSS-induced mice have significantly higher stool consistency scores from the fifth week compared with the NC, DHL, and DHH group mice (p < 0.0001 for DSS vs. NC from week 5 to week 8, p < 0.001 for DHL compared with DSS on week 5 and week 7, p < 0.001 for DHH compared with DSS from week 5 to week 8) ( Figure 1E ). For the stool blood scores, the DSS group demonstrated significantly higher scores at week 8 (p < 0.001), and only DHH mice rescued this phenotype (p < 0.001 for DSS vs. NC on week 8, p < 0.001 for DHH compared with DSS on week 8) ( Figure 1F ). DSS-induced mice also have significantly higher DAI scores from the fifth week compared with the NC, DHL, and DHH group mice (p < 0.001 for DSS vs. NC from week 5 to week 8, p < 0.001 for DHL compared with DSS on week 5 and week 7, p < 0.001 for DHH compared with DSS from week 5 to week 8) ( Figure 1G ). Mice from the H group showed no difference with that in the NC group in body weight changes, colon length, and stool scores. Furthermore, HLJD treatment at a dose of 4 g/kg also showed no toxic effects in mice, which are consistent with published papers (25, 26). These results demonstrated that HLJD at both dosages show the trends to ameliorate DSS-induced symptoms, and a high dose of HLJD resulted in better outcome in DSS-stimulated colitis mice compared with a low dose of HLJD.

As shown in Figure 2A , distal ends of colonic tissue in NC and H group mice have intact mucosa, submucosa, and muscle layer. In contrast, DSS group mice were characterized by inflammatory cell infiltration, thickened mucosal layers, and crypt loss, resulting in significantly increased histological scores (p = 0.0046 for DSS vs. NC, p = 0.0868 and 0.0202 for DHL and DHH compared with DSS, respectively) ( Figure 2B ). Both doses of HLJD were able to partly restore crypt structures and inhibited inflammation ( Figures 2A, B ). Similarly, proximal ends of colonic tissue in DSS mice demonstrated severe inflammation with larger inflammatory area compared with that in NC mice ( Figure 2C ). DHH attenuated this phenotype and significantly decreased the histological scores but not in DHL mice (p = 0.0008 for DSS vs. NC, p = 0.3512 and 0.0056 for DHL and DHH compared with DSS, respectively) ( Figures 2C, D ), indicating that a high dose of HLJD could compromise the damage induced by DSS in both distal and proximal ends of the colon. The criteria for the histological assessment are shown in Table 1 .

During chronic colitis, homeostatic macrophages are massively replenished by classic monocytes and their macrophage progeny, which will migrate to inflamed mucosa in response to chemokines including CCL2, CCL7, and CCL8 (49). These monocyte-derived macrophages express inducible nitric oxide synthase (iNOS) and generate large amounts of IL-1β, TNF, IL-6, and IL-12, working as a rich source of cytokines and chemokines in the inflamed tissue (50, 51). Our data demonstrated that DSS-induced mice have significantly increased macrophage infiltration in the distal end of colon tissue, shown as increased F4/80-positive cells ( Figure 3A ). HLJD significantly reverted this phenotype (p = 0.0145 for DSS vs. NC, p = 0.0261 and 0.0188 for DHL and DHH compared with DSS, respectively) ( Figures 3A, B ). A similar result was obtained in IHC staining of F4/80-labeled macrophages in the colonic tissues of mice from each group ( Figure 3C ). Furthermore, HLJD (50 µg/mL) reduced PMA-induced cells attaching without affecting cell viability ( Supplementary Figures S1A, B ), indicating that HLJD may inhibit differentiation of monocytes to macrophages. This was supported by the decreased expression of the macrophage marker CD11b in PMA-HLJD50-treated cells compared with that in PMA-treated cells (p = 0.0134 for PMA vs. CON, p = 0.0425 for PMA-HLJD50 compared with PMA) ( Supplementary Figure S1C ). However, the monocyte marker CD14 was not significantly changed between PMA-HLJD50- and PMA-treated cells ( Supplementary Figure S1D ).

Expression of cytokines and chemokines in colon tissues was examined by qPCR. Tnfa was significantly increased in DSS-treated mice compared with NC, and DHH significantly reverted this trend ( Figure 3D ). There was a trend of higher expression levels of Il6, Tgfb, and Ccl2 in DSS mice compared with NC, while HLJD treatment tend to reduce their expression without statistical significance ( Supplementary Figures S2A–C ). The activated macrophages could secrete C-X-C motif ligand 1 (CXCL1) (52) and CXCL10 (53), which are chemoattractant chemokines that recruit neutrophils in response to inflammation. Cxcl1 was higher in the DSS group than in the NC group (p = 0.0328), which was significantly decreased in HLJD-treated mice at both doses (p = 0.0210 and p = 0.0430 for DHL and DHH compared with DSS, respectively) ( Figure 3E ). Similarly, for Cxcl10, expression levels in DSS-treated mice were significantly higher than in the NC group (p = 0.0461), but DHL and DHH showed trends of reduction ( Figure 3F ). In addition, expression of neutrophil recruitment genes Ccl3 was significantly increased in DSS-treated mice compared with NC (p = 0.0035), and DHL, but not DHH, could significantly reduce Ccl3 expression level when compared with DSS mice (p = 0.0369) ( Figure 3G ). DSS-induced mice had remarkably increased spleen weight percentages, while HLJD significantly reverted this phenotype (p = 0.0003 for DSS vs. NC, p = 0.0001 and 0.0010 for DHL and DHH compared with DSS, respectively) ( Supplementary Figure S2D ), further highlighting the anti-inflammatory role of HLJD in DSS-induced colitis. Consistent result was obtained using 10 ng/mL of the LPS-stimulated RAW264.7 cellular model. HLJD at the concentration of 100 µg/mL exerted no inhibitory effects on RAW264.7 cells ( Supplementary Figure S3A ). Expression of M1 macrophage markers, including Tnfa, Il6, and iNOS, was reduced at the mRNA level in cells treated with LPS and HLJD (25 µg/mL) compared with that in LPS-triggered cells (p = 0.0437, p = 0.0549, and p = 0.0491, respectively, for LPS-HLJD25 vs. LPS) ( Supplementary Figure S3B ). However, HLJD treatment could not affect M2 phenotype markers ( Supplementary Figure S3C ). Level of chemokines Cxcl10 and Ccl2 showed a trend to increase after LPS induction, but there was no significant difference ( Supplementary Figures S3D, E ).

In order to further clarify the molecular mechanism(s) underlying the effects of HLJD decoction on macrophage-regulated inflammation conditions in DSS-induced chronic colitis, we evaluated expression of the Csf1r/Src pathway, which has been reported to be a key player in contributing to mucosal inflammation during colitis development (18). The mRNA expression levels of Csf1r ligands Csf1 and Il34 (54) remained unchanged in the colon of DSS-treated mice ( Figure 4A, B ). Although there was a trend for reduced Csf1 level in the DHL-treated group, Il34 expression was significantly reduced. Both Csf1 and Il34 were significantly decreased in DHH-treated mice (p = 0.0388 and p = 0.0074, respectively) ( Figures 4A, B ). Csf1r expression was significantly lower in DHL- and DHH-treated mice (p = 0.0108 and p = 0.0068, respectively) ( Figures 4C ). Protein expression of Csf1r was enhanced in DSS-induced mice but significantly reduced after HLJD administration. Src level was induced in mice treated with DSS and no significant changes in expression were observed with HLJD treatment. However, elevated phosphorylated Src (p-Src) in DSS mice was significantly reduced in DHL and DHH mice (*p < 0.05, **p < 0.01, ***p < 0.001 vs. NC; #p < 0.05, ##p < 0.01, ###p < 0.001 vs. DSS) ( Figures 4D, E ). The expression of Csf1r among groups was further confirmed by IHC assay ( Figure 4F ), as the expression level of Csf1r was increased in DSS-treated mice and reduced as a result of HLJD treatment (p = 0.0003 for DSS vs. NC, p = 0.0001 and 0.0010 for DHL and DHH compared with DSS, respectively) ( Figure 4G ). CSF1R regulates the differentiation of most circulating macrophages through CSF1 and IL34 (55). Intriguingly, levels of Csf1, Il34, and Csf1r increased in PMA-induced THP-1 cells (p = 0.0567, p = 0.0448, and p = 0.0474, respectively, for PMA vs. CON), while additional HLJD treatment tend to reverse these alterations (p = 0.3805, p = 0.1499, and p = 0.4583, respectively, for PMA vs. CON) ( Supplementary Figures S1E–G ), indicating that HLJD has a trend to inhibit the CSF1R signaling pathway in macrophage differentiation triggered by PMA. Consistently, Csf1 was enhanced when RAW264.7 cells were triggered by LPS (p = 0.0064), but this trend was inhibited after treatment with HLJD ( Supplementary Figure S3F ). Expression of Il34 and Csf1r remains unchanged among groups ( Supplementary Figures S3G, H ).

To further confirm that Csf1r expression on macrophages was inhibited after HLJD treatment, F4/80 and Csf1r were co-labeled to show their expression and distributions using a confocal microscope imaging system. As shown in Figure 5 , expression of F4/80 and Csf1r was obviously expressed higher in DSS-induced colon tissue, which was compromised after HLJD administration. Colocalization of F4/80 and Csf1r was observed after DSS induction.

To explore the effects of HLJD intervention on the gut microbiota of DSS-induced colitis mice, the fecal samples (n = 6 for each group) from NC, H, DSS, and DHL (marked as DH) were analyzed by 16S rRNA sequencing. The overall differences in species numbers among each group were shown in a Venn image ( Figure 6A ). PCoA on OTU level plots indicated clustering of samples among groups ( Figure 6B ). The Sobs index of the OTU level is shown in Figure 6C , indicating a trend towards slightly lower community richness and evenness in DSS and DH groups. Differences in the community compositions in both phylum and genus levels were shown ( Figures 6D, E ). Moreover, LEfSe analyses were performed to identify bacterial groups that showed significant differences in different levels among groups ( Figure 7A ). As shown in Figures 7B–E , the abundance of Bacteroides, Odoribacter, Clostridium_sensu_stricto_1, and Parasutterella was dramatically enhanced in DSS mice, but was significantly suppressed after HLJD treatment (p = 0.009674 for Bacteroides; p = 0.01726 for Odoribacter; p = 0.0005901 for Clostridium_sensu_stricto_1; and p < 0.001for Parasutterella). Subsequently, the correlations between colitis symptom-related parameters such as body weight, colon length, spleen weight percentage, stool consistency scores, blood stool scores, DAI, and microbe abundance at the genus level were analyzed using Spearman correlation analyses ( Figure 8 ). The abundance of Roseburia, Eubacterium_xylanophilum_group, Clostridium_sensu_stricto_1, Romboutsia, Eubacterium_fissicatena_group, and Turicibacter significantly correlated with more than five symptom-related parameters with R- and p-values shown in Table 4 .