Synthesis and characterization of LD₄ was previously reported by our laboratory (Meng et al., 2015). The structure of LD₄ is shown in Supplementary Figure S1A. The sample was excited by a 650 nm semiconductor laser (WSLS-650-500m-200M-H4; Wave spectrum Laser; China) through a columnar fiber. The energy density of the spot was measured using a light power meter (LM1; Carl Zeiss).

Male Sprague Dawley rats aged 6–8 weeks were purchased from HFK Biosciences Experimental Animals Company (SCXK 2019-0008, China) and raised in specific pathogen-free conditions. All animal experiments procedures were experimented according to the National Institutes of Health Guide for Care and Use of Laboratory Animals, and the protocol was approved by the Laboratory Animal Management Committee/Laboratory Animal Welfare Ethics Committee, Institute of Radiation Medicine, Chinese Academy of Medical Sciences (Approval No. IRM-DWLL-2017092). The ambient temperature was 22 ± 2°C, with relative air humidity of 40–70%, and food and sterile water were fed according to the experimental requirements.

Rats were fasted for 24 h then anesthetized with 10% chloral hydrate before colitis induction. The UC model was induced using the method described by Morris et al. (1989), Wirtz et al. (2007). Briefly, 30 mg of TNBS (A28757; Innochem) in 0.25 ml of 50% ethanol was injected into rat colon through a 3 mm diameter polyethylene rubber catheter (inserted 8 cm into the proximal anal rectum). Rats were then maintained in a head low, tail high position for 1 min. Occult blood test paper was used to detect feces every day.

The rats were randomly divided into six groups (10 rats per group) and treated as follows: 1, control group (normal saline); 2, TNBS group (TNBS enema); 3, LD₄-PDTL group [TNBS and low-dose LD₄ (60 μg/kg), both via enema]; 4, LD₄-PDTM group [TNBS and medium-dose LD₄ (120 μg/kg), both via enema]; 5, LD₄-PDTH group [TNBS and high-dose LD₄ (240 μg/kg), both via enema]; 6, SASP group [TNBS enema and positive control drug, SASP (500 mg/kg), gavage]. With the day of TNBS enema injection as day 0, treatment was initiated at day 7. LD₄ was administered via enema every other day, while SASP was administered via gavage at the same time points, for a total of four treatments. Thirty minutes after each treatment, the colon was irradiated with an intensive 650 nm PDT system at an energy density of 25 J/cm², excluding the SASP group, which was not irradiated. Body weight and food intake were daily measured. After all treatments were completed, fecal samples from each group were collected and stored at −80°C for further analysis of 16S rRNA. After 24 h of fasting, all rats were sacrificed and whole colons and blood were collected. The colon tissue was weighed, dissected and stored at −80°C. Parts of the tissue were also fixed with 4% paraformaldehyde. All rats in the experiment were treated according to guidelines set out in the National Institutes of Health’s Laboratory Animal Care and Use Guidelines.

Intestinal permeability can be semi-quantitatively measured by detecting fluorescence intensity in serum using fluorescent tracers. Two hundred micrograms of FITC-dextran (FD40S; Sigma-Aldrich) powder were weighed and dissolved in 5 ml rat serum, diluted by doubling dilution method for 10 dilutions, and then fluorescence intensity was detected with a Varioskan Flash 3001 enzyme plate analyzer (Thermo Fisher, Waltham, MA) to obtain a standard curve. The standard curve were shown in Supplementary Figure S1C. Following LD₄-PDT treatment, six rats from each group were fasted for 4 h before intragastric administration of FITC-dextran at a dose of 0.6 mg/g. Blood samples were collected before the animals were sacrificed and serum without hemolysis was collected. Sera were added to a 96-well plate, at 100 μl per well. Fluorescence intensity (Ex/Em: 488/520 nm) was measured with a Varioskan Flash 3001 enzyme plate analyzer. The FITC-dextran content in the rat sera was then calculated from the established standard curve.

Cytokines was correlated tightly with the occurrence of the infection, interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α) and interleukin-1 (IL-1), while myeloperoxidase (MPO), malondialdehyde (MDA), glutathione (GSH) and superoxide dismutase (SOD) were correlated with oxidative stress reaction, which was often occurred in UC. So here we measured this cytokinesis to evaluate the efficacy of the treatment. IL-6 (SEA079Ra; Cloud-Clone Corp) and TNF-α (SEA133Ra; Cloud-Clone Corp) were quantified using commercially available ELISA kits. Standard curves are shown in Supplementary Figure S1D,E. MPO (A044-1-1; Nanjing Jiancheng Bioengineering Institute), GSH (A006-1-1; Nanjing Jiancheng Bioengineering Institute), MDA (A003-1-2; Nanjing Jiancheng Bioengineering Institute) and T-SOD (A001-1-2; Nanjing Jiancheng Bioengineering Institute levels in sera were determined by ELISA assay kits according to the manufacturer’s instructions.

Harvested colon tissue was dehydrated, embedded and sliced for histopathological analysis. Hematoxylin and eosin staining was performed as previously described (Zhang et al., 2020), and the extent of inflammation was scored according to the literature (Rong et al., 2018).

Fresh, uncontaminated feces were collected from six rats in different cages and stored at −80°C until use. DNA was extracted from feces and measured by Qubit Fluorometer. The mass concentration of the DNA library was greater than 1.0 ng/μl, which was of sufficient quality for use in subsequent experiments. Illumina MiSeq technology was used to amplify and sequence the V3–V4 region of the bacterial 16S rRNA gene. The bacterial 16S ribosomal (r) RNA forward primer sequence was 5′-CCTACGGGNGGCWGCAG-3′ and the reverse primer sequence was 5′-GACTACHVGGGTATCTAATCC-3′. USEARCH (http://www.drive5.com/usearch/7.0) was used to analyze the data. Bioinformatics analysis was performed according to the operational taxonomic unit (OTU). Similarity greater than 97% sequence clustering represented an OTU. The ribosomal database project (RDP) classifier was used to systematically classify OTU sequences with reference to the Silva database.

The colon tissue samples were lysed by addition of 600 μl of 8 M urea (lysate: protease inhibitor, 50:1), sonicated for 1 s, stopped for 2 s; this procedure was repeated for a total of 120 s. Samples were centrifuged at 14,000 × g for 20 min at 4°C. Protein was quantified by sodium dodecyl sulphate-polyacrylamide gel electrophoresis. The protein solution rapid prototype high performance liquid chromatograph (RP-HPLC) was separated using an RIGOL L-3000 system (Rigol Technologies, inc.; China) according to the manufacturer’s protocol.

The colon tissue samples were lyophilized and ground into powder before being dissolved in 10 µL of 0.1% formic acid solution, centrifuged at 14,000 × g for 20 min at 4°C, then a 1 µg sample was taken for LC–MS/MS measurement. The label-free mass spectrum was analyzed by MaxQuant software and the protein data were screened by Beijing QLBio Company using the Uniprot database.

HCoEpiC was cultured in RPMI 1640 (C11875500BT; Gibco) and supplemented with 10% fetal bovine serum (S711-001S; Lonsera) at 37°C in a humidified atmosphere with 5% CO₂. The following viability experiment was performed in the same manner in all cell lines, with HCoEpiC cells described here as an example. HCoEpiC cells were divided into six groups randomly, marked as control, LPS (model), LD₄-PDTL, LD₄-PDTM, LD₄-PDTH and Dexamethasone (DXMS) groups. All groups were cultured with 2 ml serum-free 1640 medium. Except for the control group, 10 μg/ml LPS was added to each group, and the cells were cultured for another 24 h. The culture medium was replaced with fresh 1640 medium for the control and model groups, while 2 ml serum-free 1640 medium containing 1.9, 3.8, or 7.5 μM LD₄ was added to LD₄-PDTL, LD₄-PDTM and LD₄-PDTH groups, respectively; 2 ml serum-free 1640 medium containing 7.5 μM DXMS was added to the DXMS positive control group. Cells were cultured with LD₄ for 30 min. The time was sufficient for bacteria to take up LD₄, as demonstrated in the bacterial strain deactivation previously reported (Xu et al., 2016; Zhao et al., 2021), so we chose this time to investigate the cellular damage occurring under the same conditions. Cells were irradiated with an energy density of 6 J/cm² or kept in the dark for 30 min, following which the cells were cultured for a further 24 h before MTT assay.

Total RNA in HCoEpiC cells was extracted using the TRIzol Reagent (15596018; Invitrogen) according to the manufacturer’s protocol. qRT-PCR was performed using a UltraSYBR mixture kit (CW0957H; Kangwei Biotech) according to the manufacturer’s instructions. The IL-1β, IL-6, TNF-α, AOC ₁ , AKT, and NF-κB gene sequences were synthesized by Sangon Biotech (Shanghai, China). Primers are shown in Supplementary Table S1. qRT-PCR was performed using a LightCycler^® real-time PCR assay (Roche, China).

HCoEpiC cells were grown on glass bottomed cell culture dishes (801001; NEST) and treated with LPS for 24 h. HCoEpiC cells were then cultured with LD₄ for 30 min and irradiated with 650 nm laser light at an energy density of 6 J/cm². Follow-up experiments were conducted at 24 h. Anti-AOC₁(A6249; ABclonal) and fluorescein-conjugated goat anti-rabbit IgG (ZF-0311; ZSGB-BIO) at 1:100 dilution was added as previously described (Liu et al., 2019).

RIPA (CW2333; CWBio) protein lysate was added to the culture plates containing treated cells as required to extract proteins. According to the molecular weight of the target protein, the corresponding prefabricated glue (C35502009; GenScript) was used, and the loading volume of the protein sample to be tested was 30–60 μg. Electrophoresis was performed at 150 V constant pressure for 50 min. A constant pressure of 100 V was set using the wet rotation method, and the film was transferred by ice bath for 2 h. The membrane was immersed in western blot blocking solution (232100; BD) and shaken gently at room temperature for 2 h. The following rabbit primary antibodies were diluted with TBST and prepared according to the manufacturer’s instructions; AOC ₁ (16338-1-AP; Proteintech), AOC ₁ (A6249; ABclonal), p-NF-κB (3033; CST), NF-κB (8242; CST), p-IκB (2859S; CST), IκB (4812; CST), IKK (2697; CST), AKT (4691; CST), p-AKT (4060; CST), p-IKK (ab178870; Abcam), IL-6 (WL02841; Wanleibio) and TNF-α (WL01581; Wanleibio).

HCoEpiC cells were cultured in six-well plates for 24 h and transfected with short hairpin (sh) RNA (Supplementary Table S2) and overexpressing plasmid vector. Sh-AOC ₁ inserted into the pGPU6/GFP/Neo vector and the total nucleotide sequence of AOC ₁ inserted into pEX-1 to obtain the plasmid. All plasmid were synthetized by Gene Pharma (Suzhou, China). Cells were cultured in Opti-mem medium (11058021; Gibco). All transfections were performed using lipofectamine 2000 (11668019; Invitrogen) according to the manufacturer’s protocol.

Data analysis were conducted using SPSS 18.0 and GraphPad Prism 6. SPSS 18.0 was data statistical analysis software, used to analyze the differences between data. GraphPad Prism 6 was used to plot the corresponding statistics. All data were expressed as mean ± standard deviation (SD). Significant differences were determined using one-way analysis of variance (ANOVA). Statistical differences in 16S rRNA high-throughput sequencing were assessed using Tukey’s HSD (Li H. et al., 2020)^, Chen et al., 2020). p < 0.05 was considered statistically significant.

Intestinal epithelial cells were the first protective barrier for the intestinal, its weakness in function would cause the illness in intestinal. So here HCoEpiC cells and LPS-stimulated HCoEpiC cells were chosen as normal and infectious cells in vitro model to evaluate the LD₄-PDT in IBD. DXMS a medicine with efficacy like anti-inflammatory, immunosuppressive and other pharmacological effects, is widely used in the treatment of autoimmune diseases, allergies, inflammation and other diseases in clinics. Therefore, DXMS is often selected as a positive control drug in UC in vitro and in vivo. Exposure to less than 30 μM LD₄ in either the dark or the light had no effect on the growth of HCoEpiC cells, indicating that LD₄-PDT had no obvious toxicity to intestinal tissue cells (Supplementary Figure S1B). While LPS-stimulated HCoEpiC cells were sensitive to LD₄-PDT, and their proliferation was inhibited in a LD₄ dose-dependent manner (Figure 1A). Proinflammatory cytokine expression is often changed with the progress of UC. We therefore determined the levels of IL-6, TNF-α and IL-1 in HCoEpiC cells by western blotting and qRT-PCR. The protein and gene expressions of IL-1, IL-6 and TNF-α were significantly decreased in the LD₄-PDT-treated group compared with control group (Figures 1B–F).

Our previous research has shown that LD₄ could inactivate microbial pathogens in traumatic infection ((Meng et al., 2015; Xu et al., 2016; Zhao et al., 2021). We have proved that the three different doses of LD₄ used in this paper have no significant effect on normal rats and PDT alone did not significantly relieve UC (Supplementary Figure S1F,G). The occurrence and development of IBD is also closely related to the change of gut microbes. On the basis of these results, we suggested LD₄-PDT could be used to treat UC. UC model was established using TNBS, and the treatment was started from day 7 post-induction, LD₄ at doses of 60, 120, and 240 μg/kg was administrated via enema every second day, 30 min later after which the colon was irradiated with an intensive 650 nm PDT system at an energy density of 25 J/cm², SASP, a standard UC drug as a positive control, was administered to the control group. The treatment was preformed four times. A steady increase in body weight in control group and body weight decrease in TNBS model group was observed, while body weight increasing in LD₄-PDT or SASP treated rats (Figure 2A). After LD₄-PDT treatment the animals were sacrificed and the colon was harvested. Its inner wall in the normal control group was complete, with regular folds and clear vascular texture, no obvious erosion, ulcers, or granuloma. In the TNBS model group, the colon intestine was shortened, the mucosa was marked with hyperemia and edema, scattered erosion or ulcers with hemorrhage, and large ulcerated areas. While the length of the colon and the thickness of the intestinal wall were improved to varying degrees in LD₄-PDT-treated or SASP control groups (Figures 2B,C). Colon histopathology showed that in TNBS group, epithelial cells shedding off, inflammatory cell infiltration in the sub-membrane, crypt abscess and ulcer formation were observed locally, mucosal glands were disorganized, destroyed or absent, goblet cells were reduced or absent, and histological injury score was significantly increased. Compared with the model group, LD₄-PDT-treated rats showed significantly alleviated pathological symptoms in the colon, and the colonic mucosal glands were arranged neatly, with few edemas and almost no ulcerative exudate (Figures 2D,E). The intestinal epithelial barrier is an important part of the intestinal innate immunity, and so measuring the permeability of the intestinal wall could determine disease activity. Following administration of FITC-dextran by intragastric gavage, the fluorescence intensity in serum indirectly reflects the intestinal permeability, the higher fluorescent intensity, the more severe the intestinal damage. Results showed that the levels of FITC-dextran in the serum of normal control rats were very low but significantly increased in TNBS groups, indicating that intestinal permeability was increased, and the intestinal wall was damaged following TNBS exposure. The content of FITC in serum of LD₄-PDT- and SASP-treated groups was significantly decreased, indicating that these drugs had a protective effect on intestinal integrity (Figure 2F). In accordance with the results above, the expression of IL-1, IL-6 and TNF-α either in serum or in colon tissue were lower in LD₄-PDT-treated rats compared with the TNBS group (Figures 2G–J).

Because there is a close relationship between oxidative stress and UC (Morris et al., 1989), the level of MPO and MDA in serum as indicators of oxidative stress were investigated. MPO and MDA were significantly increased in the TNBS group but decreased in LD₄-PDT-treated rats in a dose dependent manner (Figures 3A,B), although levels in these animals remained higher than those of the normal control rats. Furthermore, the levels of two antioxidants GSH and SOD in serum were significantly decreased after TNBS exposure but increased significantly after LD₄-PDT treatment (Figures 3C,D). These changes all indicated that the LD4-PDT treatment could improve the living state of UC rats.

In order to investigate the mechanism of LD₄-PDT in the treatment of UC, the related protein profile was analyzed based upon proteomics. Label-free proteomics was used to detect the protein profiles in colon tissues of rats in the control, TNBS model and LD4-PDT treatment groups. A t-test was used directly for differential analysis, and the differentially expressed proteins meeting the criteria of p ≤ 0.05 and fold change ≥1.5 times were screened. Scatter plots showed changes in protein expression between the TNBS model and LD₄-PDT groups (Figure 4A). Compared with LD₄-PDT-treated rats, there were 176 differentially expressed sites in the TNBS model group, among which 116 were highly expressed (Figure 4A, red) and 60 showed low expression (Figure 4A, green). Differentially expressed proteins are shown in Supplementary Figure 2A for comparison with other groups. Gene ontology (GO) is a standard vocabulary to describe the function, location and activity of genes, which covered three aspects of biology, namely biological process, cellular component, and molecular function. GO analysis of biological processes showed that differentially expressed proteins between the control and TNBS model groups were significantly related to collagen fibril organization, collagen biosynthesis, peptide acetyl threonine phosphorylation and negatively regulated biosynthesis by cell adhesion (Supplementary Figure S2B). However, treatment with LD₄-PDT changed the colonic protein expression profiles (Figure 4B). In terms of cellular component, the differentially expressed proteins were related to intracellular organelles, collagen and immunoglobulin complexes (Supplementary Figure S2B), indicating that treatment with LD₄-PDT resulted in changes in the cell cortex, intracellular organelles, and major histocompatibility complex class II proteins (Figure 4B). In terms of molecular function, the differentially expressed proteins were mainly derived from structural components of the extracellular matrix and the activities of ligase and transferase (Supplementary Figure S2B), while treatment with LD₄-PDT changed the activities of oxygenase and hydrolase (Figure 4B). We screened 40 significantly different proteins using heat maps to illustrate the differing protein expression profiles in the colon samples. The results showed that expression of AOC ₁ in model rats was obviously increased (Supplementary Figure S2C), while this upregulation was eliminated following treatment with LD₄-PDT (Figure 4C). The role of AOC ₁ in immune cell transport made it a target for autoimmune and inflammatory diseases, so we selected AOC ₁ for further study (Peet et al., 2011).

Western blotting was used to further verify the expression of AOC ₁ protein. Compared with the control group, the expression of AOC ₁ was significantly increased in the TNBS model group but was significantly decreased after LD₄-PDT treatment (Figure 5A). The expression of AOC ₁ in HCoEpiC cells was detected by western blot (Figure 5B), qRT-PCR (Figure 5C) and immunofluorescence (Figure 5D). Consistent with the results of the animal experiments, AOC ₁ protein and gene expression were up-regulated after LPS treatment in HCoEpiC cells but were then reduced upon LD₄-PDT treatment. Because AOC ₁ can activate AKT and downstream pathways (Xu et al., 2020), the key factors interacting with AOC ₁ were screened. Western blotting showed that compared with the control group, the expression of p-AKT, p-NF-κB, p-IKK, and p-IκB was increased significantly in the TNBS group, while LD₄-PDT treatment significantly decreased the expressions of p-AKT, p-NF-κB, p-IKK, and p-IκB compared with the TNBS model group (Figures 6A–D). Similarly, the expressions of p-AKT, p-NF-κB, p-IKK, and p-IκB were significantly decreased after LD₄-PDT treatment compared with LPS stimulation alone in HCoEpiC cells (Figures 6E–I). Thus, our findings suggest that treatment with LD₄-PDT may decrease or block the expression of AOC ₁ /AKT/IKK/NF-κB in UC.

The above results indicated that AOC ₁ played an important role in LD₄-PDT treatment of UC, so the impact of AOC ₁ protein in LD₄-PDT treatment was further investigated. We knocked down the AOC ₁ gene in HCoEpiC cells using a specific shRNA then induced the inflammatory model with LPS before treating the cells with LD₄-PDT. Western blot verified that more than 70% AOC₁ was silenced (Supplementary Figure S3A), and knockdown AOC ₁ led to reduce the levels of IL-1, IL-6 and TNF-α in the cells to a varying degree compared with mock-transfected cells. Following LD₄-PDT treatment, the expression levels of these cytokines in all groups was reduced remarkably, and there was no obvious difference among knockdown AOC ₁ group and other group (Figures 7A,B). On the contrary, overexpressed AOC ₁ cells was also built, the same treatment was conducted as that of the down-regulated cells (Supplementary Figure S3B). Western blot assay showed that compared with mock-transfected cells, protein levels of IL-1, IL-6 and TNF-α in the AOC ₁ overexpression cells were increased, while these cytokines all reduced following LD₄-PDT treatment in all cells. Among all LD₄-PDT treatment cells, AOC ₁ overexpression HCoEpiC cells exhibited the higher protein levels of IL-1, IL-6 and TNF-α (Figures 7C,D). We set AOC₁-knockdown group and AOC₁-overexpression group in HCoEpiC cells without LPS stimulation, as well as control cells group, to examine whether AOC₁ is critical for cytokine production at the basal conditions. Our results demonstrated that AOC₁ plays an important role in basal condition (Supplementary Figure S3C,D).Summary above, we found that the expression of AOC ₁ was up-regulated in the model group, while LD₄-PDT treatment reduced the expression of AOC ₁ at the gene and protein levels, indicating that LD₄-PDT exerted its effects in UC via AOC ₁ , thus mediating expression of downstream cytokines. These results suggested that LD₄-PDT treatment may work via inhibition of AOC ₁ to mediate AKT/IKK/NF-κB pathways, and thus knockdown of AOC ₁ protein either by LD₄-PDT treatment or other means could alleviate UC symptoms.

Effect of LD₄-PDT on the gut microbes among the control, TNBS and LD₄-PDT-treated groups were studied using the 16S rRNA method. The Venn diagram (Figure 8A) shows that a total of 2472 OTUs were obtained from all samples. A total of 589 OTUs were appeared in all three groups, while 990 OTUs were shared between the control and LD₄-PDT groups, 64 OTUs were shared between the TNBS model and control groups, the TNBS model and LD₄-PDT groups shared 60 (Figure 8A). The Shannon curve showed that sequencing depth covered rare new phylotypes, with both diversity and Shannon index in the TNBS model group being lower than those in the control and LD₄-PDT groups (Supplementary Figure S4A). Weighted UniFrac-based principal coordinates analysis (PCoA) indicated the unique intestinal microbiota composition clustering of individual groups (Figure 8B). The α-diversity of intestinal bacteria decreased significantly in TNBS model rats, but it was recovered to an almost normal state after LD₄-PDT treatment (Figure 8C, Supplementary Figure S4B). Unweighted pair group method with arithmetic mean analysis showed that there was a significant difference in intestinal flora between the TNBS model group and LD₄-PDT-treated rats, and LD₄-PDT treatment made the overall composition of intestinal flora in UC model rats return to a state similar with the control group (Figure 8D). Taxonomic bins at the phylum level indicated that gut bacterial composition patterns in model animals were obviously different from the control groups (Figure 8E): the proportions of Firmicutes in fecal stool of rats in the TNBS model group were higher than those in the normal control group (Figure 8F), while the abundance of Bacteroidetes (Figure 8G) and Verucomicrobia (Figure 8H) were lower than in control rats, the Firmicutes/Bacteroidetes (F/B) value in the TNBS model group was higher than that of the control group (Figure 8I); however, LD₄-PDT treatment reduced Firmicutes, increased Bacteroidetes and Verucomicrobia in fecal samples, the F/B value was reduced after LD₄-PDT treatment, and the gut bacterial composition pattern was comparable with that of the control group following LD₄-PDT treatment (Figures 8F–I).