The wild-type AB zebrafish and transgenic lines, including Tg(lyz:DsRED2) (labeling neutrophils (28), https://zfin.org/ZDB-TGCONSTRCT-071109-3), Tg(rag2:DsRed) (labeling lymphocytes (29), http://zfin.org/ZDB-TGCONSTRCT-131022-4), Tg(lck:EGFP) (labeling T lymphocytes (30), https://zfin.org/ZDB-TGCONSTRCT-070117-48), Tg(mpeg1:EGFP) (labeling macrophages (31), http://zfifin.org/ZDB-TGCONSTRCT-120117-1),were purchased from the China Zebrafish Resource Center (CZRC). Then, the double labeled fishTg(lyz:DsRED2);Tg(mpeg1:EGFP) were obtained by crossing related strains. Zebrafish was maintained, raised, and reproduced as previously described (21).

The feeding of zebrafish larvae began with the feeding of Paramecium caudatum at 6 dpf (day post fertilization) until they were able to survive by feeding on live brine shrimp. Thereafter, the zebrafish were fed brine shrimp for at least 3 months. To model SBMIE (soybean meal-induced enteritis) in the zebrafish, the experimental diets (soybean meal diet (50SBM) or fish meal diet) were fed as previously described (21, 32). 3-month-old zebrafish (n = 150) were kept in 6 tanks (3L volume). The fish were fed fishmeal for a fortnight. Afterwards, the fish were fed twice daily with soybean meal or fish meal (as control) for one week.

For sampling, all reagents and tools were pre-sterilized or disinfected with 75% ethanol. Zebrafish were euthanized with 0.2mg/ml tricaine methanesulfonate (MS -222, Sigma-Aldrich) solution after fasting for 24 hours to promote intestinal emptying (33). The tail of the zebrafish was cut off with a sharp scalpel and then the body dissected to obtain the internal organs. Separate the whole intestine with its orientation, stretch it out and roll it gently on the absorbent paper to loosen the sticky mesentery (34) with tweezers. The operation to divide the zebrafish intestine was performed as previously described (35, 36), with some modifications. The hindgut tissue (last third segment) was prepared for haematoxylin-eosin staining (HE), meanwhile cell samples were collected for quantitative real-time PCR (qPCR) or preparation of gut cell suspensions for flow cytometry analysis (e.g. fluorescence activated cell sorting, FACS). Blood was collected from the tail and 0.1 ml of 10mg/ml heparin sodium was added to prevent clotting. Single cell suspensions from the internal tissues, including intestine, liver, kidney and spleen, were obtained by filtering the minced tissue through a 300 mm nylon mesh (37). For liver, kidney and spleen, the tissue was rubbed several times in a small amount of cold PBS with 1% FBS (fetal bovine serum) on the nylon mesh before filtration.

The hindgut was used for HE staining. One half was fixed directly with paraformaldehyde, while the other half was used to prepare a cell suspension. The remaining tissue after preparing the intestinal cell suspension was also fixed with paraformaldehyde. All samples were cut into sections (10μm) and stained as previously described (23).

First, the hindgut (last third of the intestine) were dissected from adult zebrafish (about 3 mouth-old), and then was lengthwise cutting with a micro-eye scissor (Zrbiorise). The intestinal tissue was transferred to 400μl of cold PBS (4°C) in a 1.5ml pre-cold microtube (on ice) with vortexing (on the Vortex Mixer, Shanghai Huxi, WH-2) for 2 seconds to dissociate the attaching mucus. The tissue was then transferred to 1.5ml cold PBS (4°C) containing 1% FBS (fetal bovine serum). Meanwhile, 200µl micropipette was adjusted to the volume of 150µl, and the tip was cut off at the front end with a sharp scalpel. This was to ensure that the intestinal tissue could easily pass though. Finally, to release the immune cells in epithelial and LP layers, the hindgut tissue was blowing repeatedly, until the tissue looked semitransparent. Additionally, for the liver, kidney and spleen of the zebrafish, this step ended with the tissue being minced without any obvious tissue fragments being present.

Only the cell suspension was kept to be filtered through a 300 mesh nylon sieve and lightly centrifuged (600 g for 5 min at 4°C). After discarding the supernatant, the cells were resuspended in the flask in an additional 300μl of cold PBS containing 1% FBS. The cells could be washed a second time by gentle centrifugation. 300μl cold PBS with 1% FBS was added to dilute the suspension and prevent the cells from becoming entangled by mucus secretion. During all steps, the cells were kept on ice at all times to ensure cell viability. Finally, the cells could be immediately analyzed with a cytometer.

The properties of the intestinal cell suspension obtained, including cell viability, morphology and fluorescence, were evaluated using both an automatic cell counter (Countstar) and a fluorescence microscope (Olympus). To check cell viability, acridine orange/propidium iodide (AO/PI) was used to label dead cells with red fluorescence, while living cells were genetically visualized with green fluorescence. 10 ul AO/PI were added to the same volume of homogenised cell suspension. After mixing, the cell suspension was immediately transferred to the slide to carefully observe the cell size and quantity. To determine the fluorescence properties, the cell suspension samples of the genetically fluorescently labelled transgenic strains were examined under the fluorescence channels.

Total RNA extraction of intestinal tissue and mucosal cells after centrifugation was performed with Trizol (Invitrogen). RNA was quantified using Nanodrop2000 (Thermo Fisher) and its integrity was checked on a 1% agarose gel. cDNA was synthesized from RNA samples (600 ng) using HiScript^® II QRT SuperMix for qPCR with gDNA wiper (Vazyme). Quantitative real-time PCR (qPCR) was performed on a Bio-Rad CFX Maestro™ (Bio-Rad) instrument with 6ng cDNA per well using Hieff^® qPCR SYBR^® Green Master Mix (Yeasen). The run was 95°C for 5 minutes, then 95°C for 10 seconds, 60°C for 15 seconds, 72°C for 60 seconds with 41 cycles. Each reaction was performed in triplicate. The qPCR analysis was performed according to previously published procedures (38, 39) with rpl13a serving as the endogenous control. The average ΔCt value was calculated by subtracting the control ΔCt value from the treated ΔCt value. The relative amount of mRNA was calculated as 2^-ΔΔCt. The primers used in the current study were partially cited from published articles or designed and tested using the primer BLAST in NCBI. The primers used are summarized in Table 1 .

To count dead cells, after centrifugation and removal of supernatant, cells were diluted with 100 ul of cold 1×DAPI (Coolaber) for 5 minutes. The suspension was then immediately diluted with 300 μl cold PBS containing 1% FBS to prevent cell adhesion caused by mucus formation and then filtered through a 300 mesh nylon to prevent clogging of the tubes for the flow cytometer. The specific workflow refers to the previously published protocol (40). Briefly, the definition of the different cell types was done by a large number of events, and the identification of live or dead cells was visualized by iodopropylidine (PI) staining. On the flow cytometry (CytoFLEX S), the parameters, such as the number of cells, the region of interest, voltage and balance, selection of detectors, were set to analyze all samples. In addition, the cell suspension stained and filtered with DAPI was kept on ice during all manipulations to ensure cell viability.

After staining with DAPI and filtration, the samples of cell suspension from multiple organs, including blood, gut, liver, kidney and spleen, were analyzed using flow cytometry. The intestinal mucosal cell suspension from four zebrafish were mixed and sorted for immune fluorescence-labeled cells on a high-speed flow sorting cytometry (Becton Dickinson). Two operation repetitions were carried out for each sample.

Transcriptomic analysis was used to systemically disclose the expression of intestinal RNA (n = 3) in both the intestinal tissue (IT group) and current prepared intestinal mucosa cell suspension (IMC group). The process for preparing the gene library and sequencing the transcriptome was done according to previously described methods (41). In brief, sequencing libraries were prepared using the NEBNextRUltra™ RNA Library Prep Kit for Illumina R (NEB, USA), and the quality of the libraries was determined using the Agilent Bioanalyzer 2100 system. On an Illumina platform (NovaSeq 6000), the library preparations were sequenced and 150bp paired-end reads were produced.

We used Hisat2 (42) to map the clean reads to the zebrafish genome (GRCz11, https://www.ncbi.nlm.nih.gov/genome/?term=txid7955[orgn]). The lots of reads that were mapped was counted by verse (version 0.1.5). Imputing the reads counts to DESeq2 (version 1.24.0) and analyze it DEGs. To select and analyze DEGs, we use DESeq2 (43) with | log2FoldChange | > 1 and padj 0.05. Then, for annotation, cluster Profiler (version 3.12.0) (44) was used to perform enrichment analysis of GO (Gene Ontology) terms and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways, with p < 0.05 considered significant enrichment. When the parameters used were not listed, default parameters were used. Afterwards, according to FishSCT (http://bioinfo.ihb.ac.cn/fishsct/), the immune cell, neuron, and muscle related marker genes were selected, to infer cell types. Further, the RPKM of related DEGs were applied in Visual Omics (http://bioinfo.ihb.ac.cn/visomics) (45), to generate a heatmap.

For all data, statistical analyses and imaging were performed using the computer program GraphPad Prism version 8.0 (GraphPad Software Inc. CA, United States). For comparison between two group, data were analyzed with paired student’s T-test with qPCR data analysis and unpaired student’s T-test with others experiment’s results. The bar diagram represented mean ± SD. Differences were considered statistically significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).

After anaesthesia, the intestine of the zebrafish was removed to prepare an intestinal cell suspension, as the intestine is relatively simple in structure and there is no submucosa between the mucosa and muscle. The primary steps are shown in ( Figure 1A ), so the immune cells in the zebrafish intestinal mucosa were released by the currently developed blowing method ( Figure 1B ). Specifically, first, an adult zebrafish (with a length of approximately 3.5cm) was removed from its intestinal tract after anaesthesia. Secondly, the front of the pipette (200μl) was cut off by approx. 5 mm to reduce shear forces and then, after blowing out the intestinal tissue several times with the pipette (approx. 2min), the cell suspension was obtained by filtering the cell debris.

The round cells in the intestinal cell suspension were counted using an automated cell counter ( Figure 2A ). Most cells had diameters ranging from 5 to 15μm. There were around 25% of cells with diameters greater than 5μm, and only 0.12% of cells with diameters greater than 15μm ( Figure 2A ). As a result, the residual of cell debris and clumps was minimal. Notably, both flow cytometry and an automatic cell counter revealed that the vitality of intestinal cell suspension generated using the current blowing approach was 70%-85%. ( Figure 2B ). Using the population features of zebrafish immune cells (such as lymphocytes and neutrophils) (46, 47), we chose the gate of common target cells in the FSC-A and SSC-A channels for further investigation. Cell viability increased significantly, with cells in the selected gates having vitality more than 90%. In addition to physiological properties, immune cells from present genetically labelled zebrafish retained significant fluorescence ( Figure 2C ), which is essential for FACS.

According to HE stained zebrafish intestinal structure, the zebrafish gut consisted primarily of mucosa and muscularis, with the mucosa containing epithelium and lamina propria ( Figure 3A ). The current technique of blowing the intestinal epithelium has almost completely separated the mucosa from the muscle layer ( Figure 3A ). The cells in mucosal layers, including both intestinal epithelial layer and lamina propria, were then taken away by the PBS-FBS buffer ( Figure 3A ). Moreover, as shown in qPCR result ( Figure 3B ), the enrichment of immune cells in the cell suspension was also proved by the significantly increased (P < 0.05) expression of the immune genes, including markers of immune cells (neutrophil function related mpx, macrophages related mpeg1, B-cell receptor signaling related blnk, Th cell’s surface marker cd4-1, initiating T-cell responses related zap70, T cell activation related lck) and immune regulation related transcriptional factors (foxp3a and foxp3b).

The pattern recognition-related “C-type lectin receptor signaling pathway,” “Toll-like receptor signaling pathway,” “NOD-like receptor signaling pathway,” as well as cytokine signaling-related “Cytokine-cytokine receptor interaction,” etc., were among the DEGs that showed an enrichment between the IMC and IT groups. Whereas barrier function-related KEGG pathways such as “vascular smooth muscle contraction,” “ECM-receptor interaction,” and “focal adhesion” were revealed for IT advantages ( Figure 4A ).

Regarding immune cell activity, a “phagosome” associated to macrophages was revealed for IT advantage while a B cell-related “intestinal immune network for IgA synthesis” was revealed solely for IMC benefit. Chemotaxis, cytokine-mediated signaling pathways, inflammatory responses, responses to cytokines, cellular responses to cytokine stimuli, etc. were among the immune-related enriched GO terms ( Figure 4B ). Regarding to the expression of cell type related genes, most immune genes were revealed for IMC advantage, while marker genes of neuron and muscle were IT advantage ( Figure 4C ). Meanwhile, the inner mucus related genes muc2.1 and muc21 were IMC advantage, yet the gel-forming (outer) mucus related genes muc13b, muc5.1, muc5d, et al. were IT advantage ( Figure 4D ). In addition, the mast cell’s marker genes cd63, fcer1g1, itga7, kita, et al. were IMC advantage ( Figure 4E ). Specifically, as to genes involved in innate immune related pathways, the pathogen sensing and chemotaxis related genes, such as tlr, fcer, ccl, and cxcl were IMC advantage ( Figure 5 ).

The aggregation of immune cells in the recently generated intestinal cell solution was examined using flow cytometry. Neutrophils, macrophages, immature lymphocytes, and T lymphocytes that were genetically fluorescently marked could be distinguished in significant numbers ( Figure 6A ). The cells from transgenic zebrafish have a completely distinct fluorescence expression profile from the wild type fish sample, and they can be specifically distinguished by flow cytometry in subgroups. Also, compared to the way of just rubbing tissue on the mesh, the proportion of immune cells from the present blowing method was significantly higher ( Figure 6B ).

As controls, systemic immune organs in zebrafish were evaluated in parallel with intestine samples obtained using the optimized approach. The proportion of lyz⁺ cells and Rag2⁺ cells in the kidney and spleen was found to be relatively high, but only detectable in the gut. Nevertheless, Lck⁺ T cells were detected in approximately equal numbers in the gut, liver, and kidney. Meanwhile, mpeg+ macrophages were found in large concentrations in the spleen, kidney, and gut. In particular, the intestinal cell suspension had a significant concentration of activated T cells (~3%), as well as macrophages (2%). Non-fluorescence labelled controls were also performed on wide type zebrafish cell suspension samples ( Figure 7 ).

To test the pathological use of the current blowing approach for creating intestinal cell suspension, SBMIE modelling was used to induce aggregation of intestinal macrophages and neutrophils, which was then released by the blowing method. HE staining confirmed the pathophysiology of the middle intestine due to intestinal inflammation. Condensed blue spots in the LP layer revealed immune cell aggregation, and intestinal villi length was considerably (p = 0.0295) shorter in the FM group ( Figure 8A ). Further, the qPCR result of inflammation related genes showed that cytokine genes il8 and il10 was significantly upregulated, and proinflammatory factors il1 and nfkb was also with the trend of upregulation in soybean (SBM) group against the fish meal (FM) group ( Figure 8B ). Meanwhile, the expression of cell maker genes mpeg1 and mpx rose considerably in the SBM group ( Figure 8B ). At the cellular level, the proportions of macrophages and neutrophils in the intestinal cell suspension were considerably higher in the SBM group compared to the FM group ( Figure 8C ).