Development of an in situ small intestinal injection technique for targeted macromolecule delivery and in vivo functional studies in mice

Abstract

Background

Targeted delivery of biological macromolecules to the small intestine remains challenging due to their susceptibility to degradation in the hostile gastric environment.

Methods

This study introduces a minimally invasive, in situ injection technique for the murine small intestine that facilitates localized luminal delivery while circumventing gastric barriers. The procedure involves a small abdominal incision for direct injection into the duodenum near the pylorus. Postsurgical monitoring of physiological parameters, systemic inflammatory markers, liver function, and intestinal integrity was conducted over 72 h. Histopathological analysis was performed. The delivery of the functional protein TAT‐EGFP (Tat protein fused to enhanced green fluorescent protein) to intestinal epithelial cells was evaluated and compared with oral gavage. As a proof of concept, single‐cell RNA sequencing of the intestinal epithelium was performed after high‐mobility group box 1 administration.

Results

Postsurgical monitoring indicated only transient, anesthesia‐related hypothermia and minor behavioral alterations. No significant changes were observed over 72 h in body weight, core temperature, clinical severity scores, systemic inflammatory markers (C‐reactive protein and leukocytes), liver function (alanine aminotransferase), or intestinal integrity. Histopathological analysis confirmed preserved tissue architecture and normal digestive, absorptive, and barrier functions. The model successfully delivered TAT‐EGFP to intestinal epithelial cells, an outcome not achievable via oral gavage due to gastric degradation. Single‐cell RNA sequencing of the intestinal epithelium after high‐mobility group box 1 administration revealed inflammatory gene expression patterns in specific epithelial subpopulations.

Conclusions

Compared to traditional methods such as oral gavage or organoid culture, this technique offers precise, degradation‐resistant delivery of macromolecules in a physiological context. The model's versatility makes it a powerful platform for intestinal research, with applications in drug delivery assessment, gene therapy evaluation, and host–microbiota interaction studies.

This study developed a minimally invasive, in situ small intestinal injection technique for mice, enabling targeted delivery of bioactive molecules while bypassing gastric degradation. Validation using functional assays and single‐cell RNA sequencing reveals high‐mobility group box 1–mediated epithelial responses, offering a translational tool for gut‐targeted research.

1. INTRODUCTION

The gut is the primary organ for nutrient digestion and absorption in mammals and represents one of the largest immune and endocrine organs in the human body. ¹ , ² In recent years, with advances in research on gut microbiota, ³ the gut–brain axis, ⁴ intestinal immune regulation, ⁵ and the intestinal mucosal barrier, ⁶ the gut's role in metabolic diseases, ⁷ neurodegenerative disorders, ⁸ and immune dysregulation ⁹ has become increasingly evident. ¹⁰ , ¹¹ , ¹² The barrier function and signaling mechanisms of intestinal epithelial cells have emerged as key targets for drug development. ¹³ , ¹⁴ , ¹⁵ , ¹⁶

The study of intestinal function relies heavily on suitable animal models. Due to their well‐defined genetic background, short reproductive cycle, established genetic manipulation techniques, and low cost, mice have become the preferred model for intestinal research. ¹⁷ However, a major challenge in this field is the precise modulation of the local small intestinal microenvironment, particularly for bioactive molecules such as proteins, peptides, and nucleic acids, which are susceptible to degradation by gastric acid and digestive enzymes. ¹⁸ Traditional methods for administering substances in murine models—including oral gavage, intraperitoneal (IP), and intravenous (IV) injection—exhibit considerable limitations when used to study localized intestinal responses to macromolecules. Although oral gavage is technically straightforward, it subjects substances to the stomach's acidic pH and proteolytic enzymes, substantially compromising macromolecular integrity and bioavailability before intestinal arrival. ¹⁹ , ²⁰ , ²¹ In contrast, IP and IV injections bypass the gastrointestinal tract entirely by introducing substances directly into systemic circulation. Although suitable for whole‐body distribution, these systemic approaches fail to ensure intact macromolecule delivery to the intestinal lumen or regulate the specific dosage reaching it. ²² , ²³ , ²⁴ As a result, such methods are ill suited for achieving localized and controlled drug delivery to the intestinal lumen and cannot adequately support investigations into macromolecular interactions with epithelial cells in the authentic gut environment. Thus, developing a simple, minimally invasive, and direct drug delivery method is crucial for studying intestinal absorption, local drug efficacy, and epithelial function. This study aims to establish a technique for in situ small intestinal injection in mice. By precisely administering drugs near the pyloric end of the small intestine, this method effectively avoids gastric degradation and provides a reliable tool for localized intestinal intervention. This approach can be applied to protein‐based drug research or nucleic acid delivery, offering a novel experimental strategy for intestinal‐targeted therapy and fundamental research.

2. METHODS

2.1. Mouse strains

All experiments were performed on 10‐ to 12‐week‐old male C57BL/6 mice obtained from the Experimental Animal Management Center of Southern Medical University. All mice were housed in a specific pathogen‐free facility. All animal experiments were approved by the Institutional Animal Care and Use Committee of Southern Medical University (approval no. SMUL202404027).

2.2. Small intestinal injection

1. First, the mice were fasted for 12 h (with water provided ad libitum) prior to surgery. They were anesthetized by administering an IP injection of 2% sodium pentobarbital (5 mL/kg body weight), with anesthesia depth carefully monitored throughout the procedure to ensure a stable surgical plane.

2. Second, the abdominal area was shaved and disinfected with iodophor, and then swabbed with 70% alcohol (Figure 1B, panel 1). A sterile drape was used to maintain a clean surgical field.

3. Third, under sterile conditions, a 1‐cm midline abdominal incision was made 0.5 cm caudal to the xiphoid process. Segments of the liver and the small intestine were exposed. The liver lobe was gently elevated using forceps, and the pyloric–duodenal junction was identified. The small intestine was then carefully grasped with forceps, and the injection site was identified as the point 1 cm distal to the pylorus along the duodenum (Figure 1B, panel 2). The entire surgical procedure, from incision to closure, was typically completed within 10 min to minimize operative stress.

4. Fourth, a 26‐gauge needle fitted to a 1‐mL syringe was carefully inserted into the injection site ensuring no perforation of the intestinal wall. No more than 200 μL of solution was slowly administered at a controlled rate to avoid excessive pressure and minimize tissue damage (Figure 1B, panel 3). The injection site was closely monitored for signs of leakage during the injection. No significant resistance was encountered during plunger advancement. The needle was gradually withdrawn while maintaining elevation for 5 s to minimize backflow. Prior to suturing, 0.1 mL of 0.5% lidocaine hydrochloride was infiltrated locally along the incision margins. After the injection, the incision was sutured, and 1 mL of prewarmed saline (37°C) was administered subcutaneously for rehydration (Figure 1B, panel 4).

5. Finally, mice were placed on a heating pad immediately postsurgery and monitored continuously until full recovery from anesthesia, defined by the return of sternal recumbency and coordinated movement. Analgesia (0.1 mg/kg meloxicam) was administered immediately postprocedure and continued once daily for 3 days, or as required based on clinical assessment (e.g., clinical severity score [CSS] score >2). Upon recovery from anesthesia, mice were transferred to a warm, clean, dry, and quiet environment. They were provided ad libitum access to food and water, with intake observed during each monitoring session. Supplemental rehydration (e.g., subcutaneous saline) or soft food was provided if animals exhibited significantly reduced intake. Postoperative monitoring was conducted every 3 h for the initial 24 h, and subsequently every 6–12 h until 72 h postsurgery. This monitoring included an assessment of rectal temperature, body weight, CSS, and general appearance (e.g., posture, activity, piloerection).

FIGURE 1.

Small intestinal injection in mice and effects on mouse physiology. (A) Schematic diagram of the intestinal injection procedure. (B) Key surgical steps of small intestinal injection, including the initial abdominal incision, careful exposure of the specific intestinal segment, precise identification of the injection site, and subsequent abdominal closure. (C–E) Physiological parameters monitored in mice over a 72‐h period after PBS (phosphate‐buffered saline) injection: (C) rectal temperature, (D) body weight changes, and (E) clinical severity score (CSS). Data represent mean ± SD (standard deviation, n = 8 independent experiments). Statistical analysis for (C) and (D) was performed using two‐way repeated‐measures ANOVA (analysis of variance). For rectal temperature (C), the main effect of group was (F (1, 14) = 0.273, p = 0.609). For body weight changes (D), the main effect of group was (F (1, 14) = 1.172, p = 0.297). (F, G) Serum biomarkers detected using ELISA (enzyme‐linked immunosorbent assay) at various time points (0, 3, 12, 48 h) post‐PBS injection compared to normal control mice: (F) C‐reactive protein (CRP) concentration and (G) alanine aminotransferase (ALT) levels. Data represent mean ± SD (n = 6 independent experiments). Statistical analysis for (F) and (G) was performed using independent samples t‐tests, and no significant differences were observed (p > 0.05). (H) Representative H&E (hematoxylin and eosin) staining images of stomach and liver tissue sections at 3 and 12 h post‐PBS injection, illustrating tissue morphology compared to normal control tissue. Scale bar: 100 μm.

2.3. Establishment of cecal ligation and puncture model

The mouse cecal ligation and puncture (CLP) model was established as described. ²⁵ C57BL/6 mice were anesthetized via an IP injection of 2% sodium pentobarbital solution (50 mg/kg). Under aseptic conditions, a midline abdominal incision was made to expose the cecum, which was then ligated and punctured twice with an 18‐gauge needle. Gentle pressure was applied to extrude a small amount of fecal material through the perforation sites. The cecum was returned to the abdominal cavity, and the incision was closed with sutures. Postoperative fluid resuscitation was administered via subcutaneous injection of 1 mL of normal saline.

2.4. Vital sign monitoring

Mice were divided into two groups (n = 8 per group): untreated mice (normal group) and mice injected with 150 μL of phosphate‐buffered saline (PBS) into the small intestine (injection group). The CSS was evaluated at 0, 3, 6, 12, 24, 48, and 72 h postprocedure using a validated scoring system. ²⁶ The four‐point scale assesses disease severity based on spontaneous activity, response to exogenous stimuli, and posture. Rectal temperature and body weight were measured at the same time points.

2.5. Fecal water content measurement

After injection of 150 μL of PBS into the small intestine, fresh fecal pellets (three pellets per mouse) were collected at different time points and weighed to obtain the wet weight (W1). After the pellets were dried at 50°C for 4 h, the dry weight (W2) was determined. Fecal water content was calculated as [(W1 − W2)/W1] × 100%.

2.6. Blood collection and analysis

At designated time points after PBS injection into the small intestine, blood was collected via retro‐orbital puncture using Ethylenediaminetetraacetic acid (EDTA) tubes (Wuhan Servicebio Technology Co., Ltd, China) and processed for complete blood count analysis (Wuhan Servicebio Technology Co., Ltd). Serum levels of C‐reactive protein (CRP) and alanine aminotransferase (ALT) were measured by Wuhan Servicebio Technology Co., Ltd.

2.7. Histopathology

Intestinal, gastric, and hepatic tissues from treated and control mice were fixed in 4% paraformaldehyde (Biosharp, China), embedded in paraffin, and sectioned (3 μm thick). Tissue sections were stained with hematoxylin and eosin (H&E) (Guangzhou Haoke Biotechnology Co., Ltd., China) to assess histopathological damage and inflammation.

2.8. Intestinal transit rate measurement

Mice were administered 150 μL of PBS into the small intestine and were fasted 12 h later. After an additional 12 h of fasting (with free access to water), mice were administered an activated charcoal suspension (Ink, Hero Group, China) by oral gavage (10 mL/kg body weight). Twenty‐five minutes after administration, mice were killed by cervical dislocation. The small intestine was isolated from the pylorus to the cecum, and the migration distance of the charcoal front (L1) and the total intestinal length (L2) were measured using digital calipers. The intestinal transit rate was calculated as (L1/L2) × 100%.

2.9. Intestinal barrier permeability assessment

Mice were administered 150 μL of PBS into the small intestine and were fasted 12 h later. After a 4‐h fast (with water ad libitum), mice were gavaged with fluorescein isothiocyanate (FITC)–dextran (4 kDa, 500 mg/kg body weight, Beyotime, China). After 3 h, blood was collected from the retro‐orbital sinus and centrifuged to obtain serum. FITC–dextran concentration was quantified by fluorescence measurement (excitation/emission: 490/520 nm) using a microplate reader.

2.10. Bacterial RNA extraction from small intestinal contents

Bacterial RNA was extracted from small intestinal contents using the Bacteria RNA Extraction Kit (Vazyme Biotech, China). Briefly, small intestinal contents were collected and homogenized in cold, RNase‐free PBS. Host cells and debris were removed by centrifugation at 500 g for 5 min at 4°C. The supernatant was then centrifuged at 10 000 g for 10 min at 4°C to pellet bacteria.

The bacterial pellet (from ~1.5 mL of equivalent of bacterial solution) was resuspended in 200 μL of prewarmed (95°C) Bacteria RNA Plus Reagent and incubated at 95°C for 4 min for lysis. After lysis, 1 mL of RNA isolater Total RNA Extraction Reagent was added, and the mixture was incubated on ice for 5 min. Chloroform (200 μL) was then added and mixed vigorously, and the mixture was centrifuged at 13 800 g for 15 min at 4°C for phase separation.

The upper aqueous phase was transferred to a new RNase‐free tube. RNA was precipitated by adding an equal volume of cold isopropanol; then the mixture was incubated at −20°C for 10 min. The RNA pellet was obtained by centrifugation at 13 800 g for 10 min at 4°C, washed twice with 1 mL of 75% RNase‐free ethanol, and air dried. Finally, RNA was dissolved in RNase‐free ddH₂O. RNA concentration, purity, and integrity were assessed using spectrophotometry and agarose gel electrophoresis. Extracted RNA was stored at −80°C.

2.11. Reverse transcription and quantitative real‐time polymerase chain reaction

Total RNA extracted from small intestinal contents was reverse transcribed into complementary DNA (cDNA) using the HiScript IV All‐in‐One Ultra RT SuperMix for quantitative real‐time polymerase chain reaction (qPCR) (R433, Vazyme Biotech). A 20‐μL reaction mixture consisting of 5 μL of 4 × All‐in‐One Ultra qRT SuperMix, 1 pg. to 1 μg of total RNA template, and RNase‐free ddH₂O was prepared. The reaction was performed according to the following program: 50°C for 5 min and 85°C for 5 s. The synthesized cDNA was stored at −80°C until qPCR analysis.

qPCR was performed using the Taq Pro Universal SYBR qPCR Master Mix (Vazyme Biotech). Each 20‐μL reaction contained 10 μL of 2× Master Mix, 0.4 μL of each 10 μmol/L primer (forward and reverse), 1–2 μL of cDNA template, and RNase‐free ddH₂O.

Specific primers used for qPCR (5′–3′) were as follows: for total bacteria 16S rRNA gene (forward: GTGSTGCAYGGYTGTCGTCA, reverse: ACGTCRTCMCACCTTCCTC), for firmicutes (forward: GGCAGCAGTRGGGAATCTTC, reverse: ACACYTAGYACTCATCGTTT), for bacteroidetes (forward: CCGGWAWTYATGGGTTTAAAGGG, reverse: GGTAAGGTTCTCCGCGTA), for Bifidobacterium (forward: ATGCTGGTGTGGAAGAGA, reverse: TGCTCGGCCACTATCCAGT), for Enterobacteriaceae (forward: ATGTTACAACCAAAGCGTACA, reverse: TTACCTTGACGCTTAACTGCACG), and for Enterococcus (forward: AGAAATTCCAAAGAACTTGAG, reverse: CAGCTGCTCTACCTCCATCATT).

The cycling conditions were an initial denaturation at 95°C for 30 s and then 40 cycles of 95°C for 5–10 s and 60°C for 20–30 s. A melt curve analysis was performed at the end of the amplification to confirm primer specificity. Ct values were determined for each reaction. Relative gene expression was calculated using the $2^{-\mathrm{\Delta \Delta }{\mathrm{C}}_{\mathrm{t}}}$ method, with total 16S rRNA gene as the internal control and the normal control group as the calibrator.

2.12. Recombinant protein expression

The prokaryotic expression plasmids pET28a‐HTF‐EGFP (HTF: HIS‐TAT‐FLAG tag, EGFP: enhanced green fluorescent protein) ²⁷ and pET32a‐HMGB1 (high‐mobility group box 1) were constructed and verified in‐house. The plasmids were transformed into Escherichia coli strain BL21(DE3) (Invitrogen Life Technologies, USA). Protein expression was induced with 1 mmol/L isopropyl β‐d‐1‐thiogalactopyranoside (Solarbio Technology, China) at 18°C for 12 h. His‐tagged fusion proteins were purified using an Nickel‐nitrilotriacetic acid (Ni‐NTA) affinity chromatography system (GenScript Biotech, USA). Endotoxin was removed using an endotoxin removal kit (Pierce columns, Thermo Fisher Scientific, USA). After purification, the endotoxin levels of the recombinant proteins were determined using an enzyme‐linked immunosorbent assay (ELISA) kit (RuiXin Biotech, China) and confirmed to be within acceptable limits for in vivo experiments (Figure S1B). Protein purity was assessed using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) and Coomassie blue staining, demonstrating high purity suitable for experimental use (Figure S1D).

2.13. Fluorescence imaging assay

Intestinal segments were embedded in optimal cutting temperature compound (Wuhan Servicebio Technology Co., Ltd.) and sectioned (10 μm thick), and nuclei were stained with 4′,6‐diamidino‐2‐phenylindole (DAPI, Solarbio) for confocal microscopy analysis. Dual‐channel imaging was performed using endogenous EGFP autofluorescence (excitation/emission: 488/507 nm) and DAPI nuclear counterstaining (358/461 nm).

2.14. Isolation of intestinal epithelial samples

After the mice were killed via cervical dislocation under anesthesia, a midline abdominal incision was made to expose the abdominal organs. The mesentery surrounding the small intestine was carefully dissected, and the intestine was promptly transferred to an ice‐cold petri dish. A 6‐cm segment of the target intestinal region was isolated, longitudinally opened along its axis using tissue scissors, and placed luminal side up on the dish. Using a glass coverslip, the villous epithelium was gently scraped into a prechilled microcentrifuge tube, carefully avoiding the muscular layer. All procedures were performed under low‐temperature conditions to maintain tissue integrity.

2.15. Western blot analysis

Small intestinal epithelial tissues were collected 8 h after protein injection into the small intestine, and tissue lysate was added. The lysates were separated using 10% SDS‐PAGE and then transferred onto polyvinylidene fluoride membranes. After blocking, the blots were incubated with an anti‐EGFP antibody (1:2000, ABmart, China). The membranes were washed with PBS thrice and incubated with appropriate secondary antibodies. Protein bands were detected using Bio‐Rad ChemiDoc.

2.16. Single‐cell RNA sequencing

Intestinal epithelial samples from control mice, CLP mice, and HMGB1‐injected mice (n = 3 per group) were obtained as described previously and dissociated into single‐cell suspensions. Single‐cell libraries were prepared using the 10× Genomics Chromium system and sequenced on an Illumina Xplus platform (PE150 mode). Data were processed using Cell Ranger (version 7.1.0) and analyzed using Seurat (version 4). Quality control included the removal of doublets and low‐quality cells. Cell clusters were annotated using SingleR, and differential cell‐type analysis was performed using Wilcoxon rank‐sum tests. Reagents, instruments, and sequencing services were provided by Shanghai Majorbio Bio‐Pharm Technology Co., Ltd. (China).

2.17. Statistical analysis

Data are presented as mean ± standard deviation (SD) from the indicated number of independent experiments. Statistical analyses were conducted using IBM SPSS Statistics (version 30.0.0, IBM Corp., USA). Group comparisons were assessed using one‐way analysis of variance (ANOVA), two‐tailed Student's t‐tests, or two‐way repeated‐measures ANOVA, as appropriate. A p‐value <0.05 was considered statistically significant.

3. RESULTS

3.1. Small intestine injection in mice and effects on mouse physiology

As described in the Section 2, we established an in situ injection model in the mouse small intestine (Figure 1A,B). A 150‐μL PBS solution was injected into the intestinal lumen of C57BL/6 mice. Immediately after the procedure, 1 mL of saline was administered for rehydration, along with analgesia (0.1 mg/kg meloxicam) and thermal support. Core body temperature (rectal), body weight changes, and food intake were monitored continuously for 72 h postoperation in both experimental and control groups. Clinical assessments were quantified using the CSS system. ²⁶

Figure 1C shows that mice in the surgical group experienced a transient postoperative temperature decrease, attributable to anesthesia‐induced suppression of thermoregulation, which normalized within 3 h. Throughout the 72‐h observation period, temperature remained stable, and no progressive weight loss was observed (Figure 1D). Dynamic CSS assessments (Figure 1E) indicated mild, transient behavioral alterations (scores 1–2) in some mice at 3–12 h postsurgery, including reduced mobility and piloerection. However, all scores returned to baseline by 24 h, remaining comparable to controls. These early behavioral changes were temporally associated with physiological stress during anesthesia recovery.

To further assess systemic responses to surgical trauma, we analyzed peripheral blood leukocyte subsets and serum CRP levels in blood samples from mice. Table 1 indicates that inflammatory markers (i.e., lymphocytes and neutrophils) in the experimental group remained within physiological ranges at all postoperative time points. Particularly, CRP—a sensitive acute‐phase reactant—did not significantly elevate after the procedure (Figure 1F), confirming the absence of a systemic inflammatory response.

TABLE 1.

Routine blood examination in mice after PBS injection into the small intestine at different time points.

Parameter	Reference value	Group	Results (mean ± SD)
WBC counts (109/L)	0.8–10.6	Normal	5.53 ± 1.39
		PBS (0 h)	6.67 ± 1.55
		PBS (3 h)	4.62 ± 2.40
		PBS (12 h)	5.05 ± 0.83
		PBS (48 h)	5.05 ± 1.29
LYM counts (109/L)	0.6–8.9	Normal	4.05 ± 1.02
		PBS (0 h)	4.90 ± 1.34
		PBS (3 h)	2.85 ± 1.54
		PBS (12 h)	3.58 ± 0.77
		PBS (48 h)	3.45 ± 0.83
MON counts (109/L)	0.04–1.4	Normal	0.13 ± 0.08
		PBS (0 h)	0.20 ± 0.09
		PBS (3 h)	0.17 ± 0.12
		PBS (12 h)	0.17 ± 0.05
		PBS (48 h)	0.18 ± 0.04
NEU counts (109/L)	0.23–3.6	Normal	1.35 ± 0.36
		PBS (0 h)	1.55 ± 0.33
		PBS (3 h)	1.60 ± 0.86
		PBS (12 h)	1.30 ± 0.28
		PBS (48 h)	1.42 ± 0.72
LYM ratio (%)	40–92	Normal	72.79 ± 3.61
		PBS (0 h)	72.23 ± 5.80
		PBS (3 h)	61.35 ± 11.08
		PBS (12 h)	70.72 ± 6.53
		PBS (48 h)	69.2 ± 8.92
MON ratio (%)	0.9–18	Normal	3.27 ± 0.63
		PBS (0 h)	3.28 ± 0.29
		PBS (3 h)	3.72 ± 1.02
		PBS (12 h)	3.23 ± 0.52
		PBS (48 h)	3.62 ± 0.73
NEU ratio (%)	6.5–50	Normal	24.07 ± 3.34
		PBS (0 h)	24.48 ± 5.89
		PBS (3 h)	34.93 ± 10.28
		PBS (12 h)	26.05 ± 6.09
		PBS (48 h)	27.20 ± 8.71
RBC counts (1012/L)	6.5–11.5	Normal	6.68 ± 0.16
		PBS (0 h)	7.04 ± 0.15
		PBS (3 h)	5.80 ± 2.27
		PBS (12 h)	6.66 ± 0.46
		PBS (48 h)	6.48 ± 0.47
PLT counts (109/L)	400–1600	Normal	808.33 ± 139.77
		PBS (0 h)	816.83 ± 272.49
		PBS (3 h)	1211.50 ± 634.43
		PBS (12 h)	864.83 ± 55.00
		PBS (48 h)	817.67 ± 108.61

Abbreviations: LYM, lymphocyte; MON, monocyte; NEU, neutrophil; PBS, phosphate‐buffered saline; PLT, platelet; RBC, red blood cell; SD, standard deviation; WBC, white blood cell.

Postoperative serum ALT levels in mice exhibited transient, mild elevations at the immediate and 3‐h time points (Figure 1G). Such transient elevations are commonly observed in response to surgical stress and anesthesia; However, these changes were not statistically significant compared to controls. They exhibited a clear trend of recovery by 12 h and fully normalized by 48 h, confirming minimal impact on liver function. Histopathological evaluation demonstrated preserved hepatic lobular architecture and normally arranged gastric mucosal glands throughout the observation periods, with an absence of hemorrhagic lesions or inflammatory cell infiltration (Figure 1H). Collectively, these findings indicate that the surgical process and anesthesia had minimal impact on visceral organ function or structure in mice.

3.2. The effects of small intestinal injection on murine intestinal physiology

As an invasive surgical model for studying intestinal epithelial function, we evaluated the effects of the surgery on intestinal tissue structure and function. At different time points after small intestinal injection of PBS, mice were killed, and the entire small intestine was rapidly excised to assess gross morphological changes. Figure 2A shows no appreciable bleeding, exudation, or other signs of acute inflammatory injury in any of the postoperative time groups. Intestinal segments near the injection site, as well as from the jejunum and ileum, were then collected for histopathological examination. The results (Figure 2B) demonstrated that the small intestinal epithelium remained structurally intact with minimal damage.

FIGURE 2.

The effects of small intestinal injection on mice intestine physiology. (A) Gross morphological analysis of macroscopic intestinal changes at 0, 3, 12, and 48 h post‐PBS (phosphate‐buffered saline) injection compared to untreated control mice (n = 6). (B) Histopathological evaluation using H&E (hematoxylin and eosin) staining to determine injury scores at the injection site, as well as proximal (jejunum) and distal (ileum) intestinal segments, at 0, 3, 12, and 48 h post‐PBS injection, compared to normal control tissue. Scale bar: 100 μm (n = 6). (C) Motility monitoring through dynamic fecal water content quantification over 72 h (n = 8). Statistical analysis for (C) was performed using two‐way repeated‐measures ANOVA (analysis of variance). The main effect of group was (F (1, 14) = 0.102, p = 0.754). (D) Intestinal barrier permeability assay performed at 24 h post‐PBS injection using FITC (fluorescein isothiocyanate)–dextran (4 kDa) administered via oral gavage, with serum fluorescence levels measured (n = 6). Statistical analysis for (D) was performed using independent samples t‐tests, and no significant differences were observed (p > 0.05). (E, F) Intestinal transit rate measurement at 24 h post‐PBS injection, using ink to quantify the progression of luminal contents through the gastrointestinal tract (n = 6). (G–M) Quantitative real‐time polymerase chain reaction (qPCR) analysis of fecal microbial DNA from normal control and PBS‐treated mice (n = 8 per group) at 72 h postinjection. (G) Total bacteria (16S Ct value). Total bacterial load shown as 16S rRNA gene Ct values. (H–L) Relative abundance of major phyla and beneficial/potential pathogenic genera. Relative abundances are presented for (H) firmicutes and (I) bacteroidetes, (J) the beneficial genus Bifidobacterium, and potential pathogenic genera (K) Enterobacteriaceae and (L) Enterococcus. Relative abundance was calculated using the $2^{-\mathrm{\Delta \Delta }{C}_t}$ method, with total 16S rRNA as the reference gene and normalized to the normal group. (M) The ratio of firmicutes to bacteroidetes was calculated based on their respective Ct values. Statistical analysis for (F–M) was performed using independent samples t‐tests, and no significant differences (ns) were observed (p > 0.05). Data are presented as mean ± SD (standard deviation).

To further evaluate the effects of the surgical procedure on intestinal digestion and absorption, we measured fecal water content in mice at different time points after small intestinal injection of PBS. The results (Figure 2C) showed no significant increase or decrease in fecal water content compared to untreated mice, indicating normal intestinal absorption and excretory function.

To assess the impact of surgery on intestinal epithelial permeability and barrier function, we administered FITC–dextran by oral gavage to both untreated control mice and mice 12 h after PBS injection. Serum fluorescence intensity was measured 3 h later. Figure 2D shows that no discernible increase in FITC fluorescence intensity was observed in the PBS‐injected group compared to controls, suggesting that injection did not compromise intestinal barrier function.

Additionally, 12 h after PBS injection, mice were gavaged with ink, and intestinal transit was evaluated by measuring the ink migration distance after 25 min. The results (Figure 2E,F) revealed no significant difference in small intestinal ink progression between PBS‐injected and normal mice, indicating that the surgical procedure did not alter intestinal motility or emptying.

To further assess the long‐term physiological impact of the intestinal injection procedure, we analyzed the gut microbiota composition in small intestinal contents 72 h post‐PBS injection using qPCR targeting the 16S rRNA gene. Figure 2G shows that the cycle threshold (Ct) values for total bacteria (16S rRNA gene) did not differ significantly between the PBS‐injected group and the normal controls, indicating a stable total bacterial load. We further evaluated changes in specific bacterial taxa by calculating the relative abundance of major phyla and key bacterial groups. The relative abundances of firmicutes (Figure 2H) and bacteroidetes (Figure 2I) were comparable between the model and normal groups. Therefore, the firmicutes‐to‐bacteroidetes (F/B) ratio also remained stable and within the normal physiological ranges in both groups (Figure 2M), demonstrating that the procedure did not induce macroscopic structural shifts in the gut microbiota.

Furthermore, we examined the relative abundance of specific bacterial taxa. The relative abundance of Bifidobacterium, a representative beneficial genus, remained stable in the model group compared to normal controls (Figure 2J), indicating no adverse impact on this commensal population. Similarly, the relative abundances of Enterobacteriaceae and Enterococcus (Figure 2K,L, respectively) exhibited no significant changes. The consistently low levels of these potentially opportunistic pathogens provide strong evidence that the surgical intervention did not promote their expansion, suggesting an absence of procedure‐induced intestinal inflammation or conditions conducive to dysbiosis. Together, these findings demonstrate that the small intestinal injection model is well tolerated, compromising neither intestinal structure nor key physiological functions—including digestion, absorption, motility, and barrier integrity—and not inducing significant shifts in gut microbiota composition. Thus, this model is suitable for studying murine intestinal epithelial physiology and related functional assessments.

3.3. Application of the small intestinal injection model to study protein effects on epithelial cells

In subsequent experiments, we employed known functional proteins to validate their activity using the small intestinal injection model. To evaluate substance migration kinetics, we first injected ink into the small intestine and monitored its distribution. Figure 3A shows the indications: stomach (orange arrow), ileocecal valve (green arrow), and ink migration front (red arrow). Our findings demonstrated that ink injected 1 cm distal to the pylorus reached the ileocecal valve within 8 h and nearly completely cleared from the small intestine, confirming full intestinal transit.

FIGURE 3.

The small intestinal injection model was used to investigate protein effects on epithelial cells. (A) Intestinal ink transit dynamics: migration distance of a nonabsorbable ink marker within the intestinal lumen measured at 30 min and 2, 4, and 8 h after intestinal injection (10 mL/kg body weight, n = 3). (B) Serum fluorescence levels measured 3 h after FITC (fluorescein isothiocyanate)–dextran (4 kDa) administration to mice via oral gavage or intestinal injection (n = 6). Statistical analysis for (B) was performed using independent samples t‐tests, and no significant differences (ns) were observed (p > 0.05). (C) Western blot detection of enhanced green fluorescent protein (EGFP) in intestinal tissues 8 h after TAT‐EGFP intestinal injection (n = 3). (D) Confocal microscopic observation showing EGFP (green fluorescence) distribution in intestinal epithelial cells 8 h after TAT‐EGFP intestinal injection, compared to normal control and TAT‐EGFP oral gavage groups (n = 3). Scale bar: 100 μm. (E) Quantification of TAT‐EGFP fluorescence intensity in intestinal tissues from TAT‐EGFP injection and TAT‐EGFP oral gavage groups (n = 3). Statistical analysis for (E) was performed using independent samples t‐tests, showing a significant difference (***p < 0.001). Data are presented as mean ± SD (standard deviation).

Next, we compared the intestinal absorption of FITC‐labeled dextran (600 mg/kg) administered by either oral gavage or direct small intestinal injection. Serum FITC fluorescence measurements (Figure 3B) revealed no significant difference in absorption between the two delivery methods, suggesting that small intestinal injection does not alter intestinal absorption or barrier integrity relative to conventional gavage.

The human immunodeficiency virus transactivator of transcription (HIV‐Tat) protein, a key regulator of viral replication and pathogenesis, can translocate into mammalian cells via direct membrane penetration or endocytosis without intrinsic cytotoxicity. ²⁷ To investigate its uptake in the intestine, we expressed and purified a recombinant Tat protein fused to enhanced green fluorescent protein (TAT‐EGFP) and administered it via the small intestinal injection in C57 mice. After 8 h, Western blot (WB, Figure 3C) and confocal microscopy (Figure 3D,E) analyses were performed. WB detected EGFP in the intestinal epithelium of TAT‐EGFP‐injected mice. Confocal imaging further confirmed a fluorescent signal in both intestinal tissue and epithelial cells. In contrast, no epithelial fluorescence was observed in gavaged mice, likely due to gastric degradation of the protein and loss of cell‐penetrating activity.

These results establish small intestinal injection as a robust in vivo method for studying protein interactions with the intestinal epithelium, circumventing the degradation limitations associated with oral administration of macromolecules.

3.4. Example application of the small intestinal injection model: Single‐cell RNA sequencing analysis of HMGB1 effects on intestinal epithelium

HMGB1 is a 26‐kDa, highly conserved nuclear protein that functions as both a critical inflammatory cytokine and a damage‐associated molecular pattern (DAMP). ²⁸ During systemic inflammatory responses triggered by infection, tissue injury, or ischemia, HMGB1 is actively secreted or passively released into extracellular compartments, including the bloodstream, where it exacerbates inflammatory processes. In a murine model of polymicrobial sepsis induced by CLP, we observed significant accumulation of HMGB1 within the intestinal lumen. Our preliminary unpublished findings indicate that luminal HMGB1 interacts with intestinal epithelial cells, eliciting differential responses among distinct epithelial subpopulations.

To investigate the specific effects of intestinal HMGB1 on epithelial cells, in vivo delivery of exogenous HMGB1 to the intestinal epithelium is necessary. However, conventional administration methods are unsuitable for delivering large proteins like HMGB1. Using the small intestinal injection technique, we delivered HMGB1 (100 μg per mouse) directly into the intestinal lumen of C57BL/6 mice. At 8 h postinjection, intestinal epithelial tissues were scraped for single‐cell RNA sequencing (scRNA‐seq).

A total of 23 497, 25 462, and 23 037 cells were obtained from small intestinal epithelial samples of mice from the control groups, HMGB1 injection group, and CLP group, respectively. Uniform manifold approximation and projection visualization revealed that these cells clustered into eight distinct populations: B cells, epithelial cells, T cells, dendritic cells, monocytes/macrophages, endothelial cells, fibroblasts, and Schwann cells (Figure 4A). Based on marker gene expression, the epithelial cell cluster was further subdivided into six subclusters: enterocytes, enteroendocrine cells, goblet cells, Paneth cells, tuft cells, and intestinal stem cells (Figure 4B). To characterize these epithelial subpopulations, a dot plot was constructed showing the expression levels of key marker genes across these clusters (Figure 4C).

FIGURE 4.

Single‐cell RNA sequencing analysis of HMGB1 (high‐mobility group box 1) effects on intestinal epithelium. All analyses were performed on samples from n = 3 biological replicates per group (normal, CLP [cecal ligation and puncture], and HMGB1 treatment). (A) UMAP (uniform manifold approximation and projection) visualization of major cell‐type clustering based on single‐cell RNA sequencing data. All identified cells from the intestinal tissue were clustered into eight distinct populations (B cells, epithelial cells, T cells, dendritic cells, monocytes/macrophages, endothelial cells, fibroblasts, and Schwann cells) based on their gene expression profiles under normal, CLP, and HMGB1 treatment conditions. (B) UMAP visualization showing further subdivision of the epithelial cell cluster into six distinct subclusters: enterocyte cells, enteroendocrine cells, goblet cells, tuft cells, Paneth cells, and intestinal stem cells, across normal, CLP, and HMGB1 treatment groups. (C) Dot plot illustrating the expression levels (average expression, indicated by color intensity) and percentage of expressing cells (dot size) for key marker genes across the 36 initial cell clusters identified from the single‐cell RNA sequencing data. (D) MA plots (log ratio vs. average abundance) showing differentially expressed genes (DEG) in the epithelial cell population after HMGB1 treatment compared to normal control (HMGB1 vs. normal) and CLP conditions (HMGB1 vs. CLP). Red dots represent upregulated genes, blue dots represent downregulated genes, and gray dots indicate nonsignificant genes (log₂FC threshold: 0.5, *p < 0.05). (E) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Gene Ontology (GO) enrichment analysis of DEGs identified in epithelial cells after HMGB1 treatment compared to normal control conditions. Dot size represents the number of genes enriched in each term, and color intensity indicates the adjusted p‐value (p‐adjust).

Differentially expressed genes (DEG) were identified using MA plots (M‐versus‐A plots, where ‘M’ denotes the log ratio and ‘A’ denotes the average log expression), comparing HMGB1‐treated cells to normal control cells and HMGB1‐treated cells to CLP‐treated cells (Figure 4D). To further interpret the biological significance of the identified DEGs, functional enrichment analyses were performed. Figure 4E presents the results of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis and Gene Ontology term enrichment analysis for the DEGs.

These results demonstrate that HMGB1, as a DAMP molecule, influences cell‐type distribution and gene expression patterns across different populations and subsets of the small intestinal epithelium when delivered via the small intestinal injection technique. Compared to the CLP model, HMGB1‐induced DEGs included both injury‐related inflammatory genes and genes affecting intestinal function. These findings confirm that the small intestinal injection is an effective method for studying protein interactions with the intestinal epithelium. Further data and insights await subsequent in‐depth analysis.

4. DISCUSSION

In recent years, the critical role of the intestine in various diseases has garnered significant attention and been the subject of extensive research. ²⁹ , ³⁰ With the rapid development of gene editing technologies (e.g., CRISPR‐Cas9), ³¹ scRNA‐seq, ³² and multi‐omics, ³³ researchers have achieved an in‐depth understanding of intestinal epithelial signaling pathways. Consequently, intestinal epithelial cells have emerged as important targets for drug development. ³⁴ , ³⁵ Mice, due to their low cost, short reproductive cycle, and well‐established genetic manipulation techniques, remain a cornerstone animal model in medical research. Various specific gene knockout and reporter gene mouse models are widely utilized. However, achieving precise localized intestinal interventions in mice—particularly for delivering easily degradable molecules (e.g., proteins, nucleic acids)—to obtain samples under conditions that closely mimic physiology and pathology remains a key challenge.

In traditional animal experiments, oral gavage is the most common method for administering substances to the digestive system. ³⁶ Although simple to perform and minimally invasive, this approach fails to protect macromolecules such as proteins and nucleic acids from degradation by the low gastric pH and digestive enzymes, preventing functional molecules from reaching the intestine intact. Compared to gavage, ex vivo intestinal perfusion models allow precise control of experimental conditions and direct stimulation of the intestinal epithelium. However, these models cannot replicate the in vivo environment, have short experimental durations, and fail to maintain intestinal tissue viability long term. Intestinal catheter implantation enables chronic sampling or drug delivery but involves complex surgery and disrupts intestinal motility and normal physiology.

Intestinal organoid technology has gained increasing attention in recent years. ³⁷ As a three‐dimensional in vitro model, ³⁸ intestinal epithelial organoids serve as powerful tools for studying intestinal physiology, disease, and drug development. ³⁹ Compared to in vitro primary epithelial cell cultures, organoids better preserve the spatial architecture and cellular diversity of the intestinal epithelium while supporting cell–cell interactions and signaling. ⁴⁰ However, they lack systemic regulation from nerves, immune components, and blood flow, and cannot simulate intestinal peristalsis, complex epithelial–microbial interactions, or the gut microenvironment. Additionally, their high cost and technical complexity limit widespread adoption.

Thus, developing a gastric degradation‐resistant intestinal targeting method for mice is crucial for advancing intestinal research. The intestinal injection model described here offers distinct advantages. First, it bypasses gastric degradation, allowing proteins, peptides, and nucleic acids to be delivered intact to intestinal target sites. Although macromolecules remain susceptible to enzymatic degradation within the intestines, this method is highly valuable for studying their function or assessing the in vivo action of macromolecular drugs, as the intestinal environment represents a key physiological condition these molecules encounter. Second, excluding the effects of drug toxicity, mice exhibit long‐term survival and maintain normal physiological functions after the intestinal injection procedure. Third, it enhances the precision of controlling intestinal drug concentrations while reducing systemic exposure compared to gavage or systemic administration. In addition, the surgical procedure is not technically challenging. Personnel with standard animal experiment qualifications can successfully perform it independently after receiving protocol training and guidance. This intestinal injection model establishes an efficient and controllable experimental platform for localized intestinal drug delivery and functional studies.

Optimized for murine anatomy, this intestinal injection model addresses the high mortality risk associated with traditional intestinal surgery in small animals. Using a 1‐cm miniature incision combined with precise anatomical positioning and in situ injection, it minimizes trauma without displacing intestinal segments. To ensure the feasibility and standardization of the method, all surgical procedures were performed by multiple operators who had received systematic training in animal experiments. By strictly following the protocol described in this study, the model can be reliably established with a high success rate.

As an exploratory model for in vivo investigation of the intestinal epithelium in small animals, this model requires further optimization and systematic validation. Although no surgical mortality was observed during extended monitoring periods, its applicability for studies involving longer observation durations (>72 h) requires thorough evaluation. Future assessments should examine potential intestinal adhesions resulting from the surgical procedure at later postoperative time points. During model development and validation, considering the potential influence of sex hormones and age on animal sensitivity and stress tolerance, this study utilized 10‐ to 12‐week‐old male C57BL/6 mice, a choice consistent with prevailing practices in animal disease modeling. However, subsequent evaluation of this model in female mice and different age groups is necessary to ensure broader applicability. Furthermore, potential influences from variations in genetic background, specific dietary regimens, or subtle differences in housing environments should be considered for enhanced generalizability.

In future basic research, this model can be directly applied or combined with gene‐edited mouse models to study the functions of macromolecules (e.g., proteins, peptides) on the intestinal epithelium. It holds translational potential for evaluating the intestinal absorption efficiency of protein/nucleic acid–based drugs, nanoparticles, or liposomes. In gene therapy, it could be used to deliver CRISPR‐Cas9/siRNA to study gene function or treat inherited intestinal disorders. Antimicrobial peptides or probiotics could be precisely administered to investigate microbiota–host interactions. Methodologically, integration with miniature endoscopes could enable real‐time mucosal observation postinjection, whereas bioluminescent imaging could track distribution of fluorescently labeled proteins. These potential applications require further validation and parameter optimization.

In conclusion, this study establishes an effective and reliable intestinal‐targeted delivery method with broad applicability in intestinal physiology, pharmacology, and translational medicine research. It is valuable during drug development for evaluating candidates that require localized intestinal targeting or exhibit substantial systemic side effects. By clarifying the influence of administration routes on drug efficacy, this approach provides crucial insights into the selection of clinical dosing regimens, thereby supporting the translation of relevant therapeutics from basic research to clinical practice.

AUTHOR CONTRIBUTIONS

Yawen Lai: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; visualization; writing – original draft; writing – review and editing. Xintao Zhang: Conceptualization; data curation; investigation; validation; visualization; writing – original draft; writing – review and editing. Tingting Luo: Investigation; methodology. Wenhan Chen: Investigation; methodology. Chenyu Ma: Investigation; methodology. Haihua Luo: Methodology. Jinghua Liu: Conceptualization; project administration; resources; supervision. Jia Xu: Conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; project administration; resources; supervision; validation; visualization; writing – original draft; writing – review and editing.

FUNDING INFORMATION

This study was supported by the National Natural Science Foundation of China (NSFC, no. 82172140).

CONFLICT OF INTEREST STATEMENT

The authors have no relevant financial or nonfinancial interests to disclose.

ETHICS STATEMENT

All animal experiments were approved by the Experimental Animal Welfare and Ethics Committee of Southern Medical University (approval no. SMUL202404027). All procedures were conducted in strict accordance with the institutional guidelines for the care and use of laboratory animals, with all efforts made to minimize animal suffering.

Supporting information

Figure S1. (A) Step‐by‐step surgical procedure for small intestinal injection in mice. Panels 1–8 show the key surgical steps: (1) mouse fixation after anesthesia, (2) skin preparation, (3) abdominal incision, (4) careful exposure of the specific intestinal segment, (5) precise identification of the injection site and administration of the test solution, (6) organ repositioning, (7) incision closure, and (8) fluid resuscitation. (B) Endotoxin levels of recombinant proteins and PBS (phosphate‐buffered saline) vehicle. Endotoxin concentration (EU/mL) of PBS, HMGB1 (high‐mobility group box 1), and TAT‐EGFP (Tat protein fused to enhanced green fluorescent protein) solutions used for injection, determined using ELISA (enzyme‐linked immunosorbent assay). Data are presented as mean ± SD (standard deviation, n = 3 independent batches). No significant differences were observed among the groups (p > 0.05). Data are presented as mean ± SD. (C) Postoperative survival rate of mice. Kaplan–Meier survival curve showing the probability of survival for control (CTRL) mice and mice subjected to small intestinal PBS injection over a 15‐day observation period postsurgery (n = 10 mice per group). (D) Purity assessment of recombinant proteins. SDS‐PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) analysis with Coomassie blue staining demonstrated the purity of the recombinant proteins, TAT‐EGFP and HMGB1, used in this study. “M” indicates protein molecular weight marker. (E) Quality control metrics for single‐cell RNA sequencing data. Violin plots show the distribution of total RNA counts (nCounts_RNA), number of detected genes (nFeature_RNA), percentage of mitochondrial genes (percent.mito), percentage of ribosomal genes (percent.rb), and percentage of red blood cell contamination (percent.redcell) for cells from normal, CLP (cecal ligation and puncture), and HMGB1 treatment groups.

Lai Y, Zhang X, Luo T, et al. Development of an in situ small intestinal injection technique for targeted macromolecule delivery and in vivo functional studies in mice. Anim Models Exp Med. 2026;9:128‐141. doi: 10.1002/ame2.70123