Sealing the Pandora’s vase of pancreatic fistula through entrapping the digestive enzymes within a dextrorotary (D)-peptide hydrogel

Abstract

A variety of therapeutic possibilities have emerged for skillfully regulating protein function or conformation through intermolecular interaction modulation to rectify abnormal biochemical reactions in diseases. Herein, a devised strategy of enzyme coordinators has been employed to alleviate postoperative pancreatic fistula (POPF), which is characterized by the leakage of digestive enzymes including trypsin, chymotrypsin, and lipase. The development of a dextrorotary (D)-peptide supramolecular gel (CP-CNDS) under this notion showcases its propensity for forming gels driven by intermolecular interaction. Upon POPF, CP-CNDS not only captures enzymes from solution into hydrogel, but also effectively entraps them within the internal gel, preventing their exchange with counterparts in the external milieu. As a result, CP-CNDS completely suppresses the activity of digestive enzymes, effectively alleviating POPF. Remarkably, rats with POPF treated with CP-CNDS not only survived but also made a recovery within a mere 3-day period, while mock-treated POPF rats had a survival rate of less than 5 days when experiencing postoperative pancreatic fistula, leak or abscess. Collectively, the reported CP-CNDS provides promising avenues for preventing and treating POPF, while exemplifying precision medicine-guided regulation of protein activity that effectively targets specific pathogenic molecules across multiple diseases.

Therapeutic possibilities have emerged for rectifying abnormal biochemical reactions by modulating intermolecular interactions. Here, the authors present a regulatory strategy for capturing biomolecules and preventing enzyme leakage using a chiral-engineered peptide supramolecular gel.

Introduction

Postoperative pancreatic fistula (POPF) is the most prevalent and potentially life-threatening complication that can occur after pancreatic surgeries such as pancreatoduodenectomy or distal pancreatectomy^1,2. This complication arises due to the leakage of pancreatic exocrine secretions, which contain digestive enzymes like trypsin, chymotrypsin, and lipase, at the anastomosis or closure of the pancreatic remnant³. In this state, these digestive enzymes erode both the pancreatic and peripancreatic tissues, leading to the formation of intra-abdominal abscesses and subsequent fatal sepsis^1,2. Although various techniques have been proposed to prevent POPF, such as reconstructing with a technique called pancreaticogastrostomy reconstruction or placing a stent in the pancreatic duct or using fibrin glue formation sealing and somatostatin analog octreotide utilization, it remains an arduous challenge to fundamentally suppress the hyperactivity of digestive enzymes during POPF and consequently effectively neutralize its inherent threat⁴ Therefore, this work endeavors to introduce an efficient approach of insulating these enzymes from the solution and preventing their affinity with their substrate, thereby curtailing their functional activity and ultimately mitigating POPF.

To achieve this objective, two formidable obstacles must be overcome: (1) it remains a daunting task to create trappable functional molecules of digestive enzymes from scratch that exhibit no inclination towards self-aggregation or binding to molecular scaffolds, and (2) it continues to be challenging to completely restrict the interaction between digestive enzymes and their substrates due to the absence of membranes and the consequent dynamic exchange with their counterparts in the external milieu. To simultaneously tackle both challenges, we employ a computer simulation-driven screening process to identify the D-enantiomeric peptide Ac-(rada)₄-NH₂ motif, which exhibits affinity towards trypsin, chymotrypsin, and lipase. Subsequently, a CP-CNDS hydrogel was developed from the Ac-(rada)₄-NH₂ peptide, showcasing its capacity to ensnare these enzymes within its gel. Moreover, CP-CNDS effectively immobilizes the three digestive enzymes within its internal gel matrix while preventing their dynamic interaction with substrates in the external environment. As a result, CP-CNDS completely suppresses trypsin, chymotrypsin, and lipase activity leading to potent alleviation of postoperative pancreatic fistula (POPF). Remarkably, with the administration of CP-CNDS, all rats suffering from POPF not only survived but also made a recovery within a mere 3 days. In stark contrast, without any treatment, the occurrence of postoperative pancreatic fistula, leak, or abscess resulted in an alarming survival rate of less than 5 days. The collective findings of the reported CP-CNDS not only offered promising avenues for preventing and treating POPF, but also exemplify the precision medicine-guided regulation of protein activity that effectively targets specific pathogenic molecules across multiple diseases.

Results and discussion

The motif screening of CP-CNDS

The inspiration of design for gel motif sequences arise from the fact that some proteins regulate protein function via intermolecular interaction to highly order amyloid fibrils with a β-sheet aggregation structure, such as Tau protein and spider silk protein⁵. Based on this, this kind of peptide hydrogel sequences are given priority which L-enantiomers possess the ability to form self-assembled β-sheet structures^6–8. Wherein the selected 12 L-peptide sequences with β-sheet structure in Fig. 1B were widely used in the literature⁶. To acquire CP-CNDS capable of capturing these digestive enzymes and facilitating intermolecular interaction, a library of D-enantiomeric peptide motifs was established for the preplanned self-healing hydrogel (Fig. 1B and Supplementary Fig. 1)^6,9–12. The subsequent step entailed subjecting all the D-peptide motifs in this library to a examination of their binding affinity towards trypsin, chymotrypsin, and lipase. Moreover, the interface area of the five potential simulation structures was calculated using PDBePISA, and their means were sorted from highest to lowest in each digestive enzyme (Fig. 1B). After summing up the three ranks (as shown in the rightmost panel of Fig. 1B), it was found that the Ac-(rada)₄-NH₂ motif, also known as code name A, exhibited the minimum value, indicating its superior affinity potential for all three enzymes.

Fig. 1. The screening of D-enantiomeric peptide motifs for capturing digestive enzymes.

A Schematic illustration for the outcomes and potential solution of postoperative complication of pancreatic surgery. B The screening for β-sheet peptide hydrogel motif by the interaction evaluation between candidates and pancreas enzyme. (n = 5 biological replicates in each group, means ± SD, statistical analysis in (B) was performed using two-sided t-test.). C The molecular dynamics simulation of Ac-(rada)₄-NH₂ with typical enzymes, which are trypsin, lipase and chymotrypsin. The evaluation of binding affinity between Ac-(rada)₄-NH₂ and different enzymes. The TEM images and SEM images of Ac-(rada)₄-NH₂ incubating with different enzymes. (n = 3 biological replicates in each group, means ± SD).

Furthermore, to explore the potential of the Ac-(rada)₄-NH₂ motif for capturing the three enzymes, a 500 ns coarse-grained molecular dynamics approach is employed to simulate the intermolecular interaction of Ac-(rada)₄-NH₂ in conjunction with trypsin, lipase, or chymotrypsin (Fig. 1C). After 500 ns of molecular random motion, the superimposition of the Root Mean Square Fluctuation (RMSF) unveiled three composite structures of Ac-(rada)₄-NH₂ engaging with three enzymes exhibiting low RMSF values (Fig. 1Ca), thereby indicating a high degree of structural stability and good repeatability after three times of random initial positions and random initial velocities. These high levels of structural stability were further substantiated by the stabilizing solvent-accessibility surface area (SASA) and 3D radius of gyrations (Supplementary Fig. 2A–C) observed at a subsequent time point, precisely 350 ns later. The free energy landscape (FEL) of the three simulation structures is noteworthy for its abundance of local energy basins (Fig. 1Cb), implying the dynamic cross-linking or growth properties inherent in the Ac-(rada)₄-NH₂ motif. Meanwhile, we synthesized Ac-(rada)₄-NH₂ and assessed its binding affinity with the aforementioned enzymes to enhance the reliability of the simulation conclusions through experimental data. As illustrated in Supplementary Fig. 3, LC-MS analysis and HPLC confirmed precise molecular weights and purity exceeding 95%, indicating successful synthesis of Ac-(rada)₄-NH₂¹³. Consistent with simulation trend, Ac-(rada)₄-NH₂ possess affinities with binding constants K_d of ~342 nM, ~452 nM, and ~227 nM, implying that there is co-assembly potential between Ac-(rada)₄-NH₂ and enzymes (Fig. 1Cc). Furthermore, as depicted in Fig. 1Cd, SEM and TEM data demonstrated that Ac-(rada)₄-NH₂ can co-assemble with multiple enzymes to form nanofibers, consistent with the simulated findings. The further analysis of the bonding interface between Ac-(rada)₄-NH₂ and the three enzymes unveiled an abundance of hydrogen bonds, serving as a testament to the high structural stability fortified by these intermolecular interactions (Supplementary Fig. 4). The combined findings collectively affirm the potential of the Ac-(rada)₄-NH₂ motif in facilitating the capture of alongside the three enzymes.

The design and construction of CP-CNDS

To further enhance the intermolecular cross-linking of the Ac-(rada)₄-NH₂ motif^14–17, we introduced a blood coagulation factor FXIIIa-responsive crosslinkable L-peptide motif GGQQLK and a calcium ion-responsive crosslinkable D-peptide motif gsvlgyiqir at the C-terminal of the Ac-(rada)₄-NH₂ motif (Fig. 2A). Subsequently, both (rada)₄-gsvlgyiqir (^DCP1) and (rada)₄-GGQQLK (^DCP2) were synthesized via HOBT/HCTU condensation utilizing the FMOC chemistry^18–22. Similarly, the synthesis and characterization method employed for the production of ^LCP1 and ^LCP2 was identical. The LC-MS analysis demonstrated the precise molecular weights and purity exceeding 95% of the peptides (Fig. 2B and Supplementary Figs. 5 and 6). The NMR signals also revealed the presence of rich side-chain functional groups (Supplementary Figs. 7 and 8)⁹. Further, the ratio of the two peptides was confirmed. ^DCP1 peptide can form a stable hydrogel state at the mass fraction of 5% (Supplementary Fig. 9A). Based on the 5% concentration of ^DCP1, the same volume of ^DCP2 peptide containing different masses were mixed with mass ratios of 2:1, 2:2, 2:3, 2:4, respectively. The hydrogel starts to form a stable state when the mass ratio is 2:3 (Supplementary Fig. 9B). Rheological data show no significant change in storage modulus (G′), loss modulus (G″), or loss tangent from this ratio (Supplementary Fig. 9C). Therefore, a 2:3 mass ratio of peptides was selected for further research. By combining these two D-peptides in ultrapure water, an injectable hydrogel called CP-CNDS was created, demonstrating rapid gel formation and post-injection plasticity (Fig. 2C). Furthermore, the CD spectrum of CP-CNDS exhibited a striking mirror symmetry along the X-axis compared to its L-enantiomer CNDS (Fig. 2D), providing evidence for accomplished chiral engineering. The mechanical properties of the CP-CNDS were evaluated by conducting rheological measurements. The storage modulus (G′) of the CP-CNDS hydrogel ranged from 1403 to 2926 Pa, while the dissipated energy (G″) ranged from 292 to 431 Pa (Fig. 2E). Additionally, the loss tangent ranged from 0.13 to 0.21 (Fig. 2E). Furthermore, shear rate sweep measurements unequivocally demonstrated that CP-CNDS exhibited a significant shear-thinning behavior (Fig. 2E). It is evident from Fig. 2E that following high strain (100%), the storage modulus (G’) of the CP-CNDS decreased below the loss modulus (G″), which subsequently recovered upon strain withdrawal (1%). After undergoing eight cycles of high-to-low strain, the regenerated hydrogel exhibited nearly identical G’ and G” values. These findings substantiate the self-healing properties of CP-CNDS hydrogel.

Fig. 2. The design, synthesis and characterization of CP-CNDS hydrogel.

A Schematic diagram of the design of CP-CNDS hydrogel. B Characterization of synthesized CP-CNDS by LC-MASS and HPLC. C The injectability of CP-CNDS. D The CD spectrum of CP-CNDS and CNDS. E Rheological properties of the CP-CNDS hydrogel. The storage modulus (G’) and loss modulus (G”) against frequency ranging from 0.1–100 rad/s (strain = 1%, 37 °C). Shear viscosity-shear rate of the CP-CNDS hydrogel with the shear rate ranging from 0.1–100 rad/s at 25 °C. The shear recovery property of the CP-CNDS hydrogel when the alternate step strain was switched from 1 and 1000% periodically (frequency = 10 rad/s, 37 °C). F The gelatinization image of CP-CNDS hydrogel in response to Ca²⁺ and Ca²⁺ + FXIIIa. G Gelatinization of CP-CNDS hydrogel quantized by the absorbance at 600 nm (n = 3 biological replicates in each group, means ± SD). H The TEM images and SEM images of CP-CNDS. I The evaluation of resistance to degradation of CP-CNDS in vitro. J The percentage of residue of CP-CNDS in enzyme solution (means ± SD). The experiments in (B–F, H–J) were independently replicated three times, yielding consistent results.

The response of CP-CNDS to Ca²⁺ ions and blood coagulation factor FXIIIa was confirmed through gelation analysis, as demonstrated by the captivating photograph (Fig. 2F) and the peak absorbance at 600 nm corresponding to the most densely formed gel (Fig. 2G). The functional motif (vlgyiqir in ^DCP1) around the self-assembling nanofibers is capable of binding to Ca²⁺ with coordination number 3, driving the self-assembly into a mesh-like network⁹. The FXIIIa acts as a catalyst, facilitating the cross-linking of nanofibers and fibrin through an acyl transfer reaction involving the flanking motif (QQLK in ^DCP2), in order to improve the adhesion between the hydrogel and the postoperative wound⁹. The addition of Ca²⁺ and FXIIIa significantly augmented the gel formation (Fig. 2F, G). The mechanism underlying the responsiveness to Ca²⁺ and FXIIIa was substantiated through the utilization of transmission electron microscope (TEM) images and scanning electron microscopy (SEM) images (Fig. 2H), wherein the incorporation of Ca²⁺ and FXIIIa augmented the extent of cross-linking amidst gelatinous fibers. Additionally, the resistance of CP-CNDS to degradation was put to the test by a pancreatic mixture comprising trypsin, chymotrypsin, and lipase (Fig. 2I, J). After 24 h incubation, the SEM images revealed a pristine microstructure of CP-CNDS in stark contrast to the impaired microstructure of CNDS (Fig. S9D). Similarly, the CD signal of CP-CNDS remained largely unchanged after the mixed enzyme treatment, while the CD signal of CNDS significantly decreased (Fig. S9E). Consequently, CP-CNDS exhibited a preservation rate exceeding 90% even after enduring these digestive enzymes for 7 days, while its L-enantiomer CNDS underwent near-complete degradation within just 2 days. The resistance to enzymolysis displayed by CP-CNDS grants it exceptional ability to suppress the activity of these digestive enzymes.

CP-CNDS led to the transition of pancreatic digestive enzymes from a soluble state to an interior gel matrix

The design of CP-CNDS enables the transition of digestive enzymes from a solution state to the gel, thereby inducing the isolation between digestive enzymes and their substrates, encompassing proteins and lipids within napes or internal abdominal organs during POPF (Fig. 3A). To accomplish this task, the indispensable capability of CP-CNDS to interact with trypsin, chymotrypsin, and lipase was evaluated through structural analysis using Discovery Studio and experimentally validated via fluorescence polarization (Fig. 3B). The binding areas of CP-CNDS to trypsin, lipase, and chymotrypsin were discovered to be 984.9 Ų, 881.2 Ų, and 865.7 Ų respectively, accompanied by high affinities with binding constants K_d of ~102 nM, ~251 nM, and ~243 nM (Fig. 3B), implying a high affinity. The subsequent step entailed the dissolution of trypsin, lipase, and chymotrypsin in PBS buffer, either individually or in combination, marked with Coomassie Brilliant Blue to create a visually striking solution. To conduct a more detailed investigation into the interaction process, we utilized a low-concentration CP-CNDS solution with a mass fraction of 0.5%. These enzyme solutions were then carefully added to the tube containing CP-CNDS solution. The CP-CNDS hydrogel, as illustrated in Supplementary Fig. 10, encapsulated all dyed blue digestive enzymes, indicating a remarkable transformation of these enzymes from their soluble state to the interior of the hydrogel. By documenting the entire process, a discernible inclination towards accumulation within the CP-CNDS hydrogel can be observed for all three digestive enzymes, thereby further substantiating their transition from solution to gel interior (Fig. 3C). Furthermore, the fluorescence recovery after photobleaching (FRAP) experiment of Cy5-labeled digestive enzymes performed by scanning confocal microscopy (LSCM) imaging once again supported their entrapment within the CP-CNDS hydrogel. As depicted in Fig. 3D, CP-CNDS effectively suppressed the fluorescence recovery of the three Cy5-labeled digestive enzymes, suggesting that they are confined within the gel interior.

Fig. 3. The presence of CP-CNDS led to the transition of pancreatic digestive enzymes from a soluble state to an interior gel matrix, consequently impeding their enzymatic activity.

A Schematic diagram of the biofunction of CP-CNDS hydrogel. B The docking, interactions and evaluation of fluorescence polarization affinity for the (rada)₄ motif of CP-CNDS with trypsin, lipase and chymotrypsin respectively, means ± SD. The experiment in (B) was independently replicated three times, yielding consistent results. C The protein concentration evaluation of solution or hydrogel for CP-CNDS incubating with trypsin, lipase or chymotrypsin in vitro respectively (n = 3 biological replicates in each group, means ± SD, statistical analysis was performed using two-sided t-test.). D Representative images obtained from Laser Scanning Confocal Microscopy (LSCM) during FRAP experiments are presented, illustrating the effects of 5-min incubation with trypsin^Cy5, lipase^Cy5, or chymotrypsin^Cy5 in vitro respectively. The scale bars indicate a length of 5 μm. The bleached area is indicated by circles. E The inhibition evaluation of enzyme bioactivity for trypsin, lipase and chymotrypsin by CP-CNDS or CNDS. The experiments depicted in (D, E) were independently replicated three times, resulting in consistent findings.

The CP-CNDS hydrogel holds the promise of impeding the interaction between the intramellar trio of digestive enzymes and their substrate in the external milieu, thereby presenting a potential avenue for suppressing their activity. To validate this, the enzymatic activity of trypsin, lipase, and chymotrypsin was monitored through the chromogenic reaction of their respective enzyme digestion products. It is apparent that CP-CNDS evinces near-total inhibition of the activity of three digestive enzymes (Fig. 3E), thereby indicating the practicability of employing interaction-derived enzyme inhibition strategy. Furthermore, CP-CNDS outperforms CNDS in curtailing the bioactivity of trypsin and chymotrypsin (Fig. 3E). The activation of trypsin or chymotrypsin enables the enzymatic hydrolysis of L-type proteins or peptides, while D-type peptides are not susceptible to enzymolysis. As depicted in Fig. 2I, J, when enzymes are activated and coexist for an extended period, CNDS undergoes gradual enzymolysis, leading to the deterioration of gel structure and the loss of its ability to continuously inhibit enzyme activity. The enzyme resistance of CP-CNDS, composed of D-enantiomer amino acids, is superior to that of CNDS, suggesting the a more durable ability to inhibit the bioactivity of trypsin and chymotrypsin. Collectively, the presence of CP-CNDS led to the transition of pancreatic digestive enzymes from a soluble state to an interior gel matrix, consequently impeding their enzymatic activity.

The administration of CP-CNDS leaded to a potent alleviation of POPF

The challenge posed to CP-CNDS for POPF therapy was met by establishing a classical rat model of POPF, wherein the main pancreatic duct was deliberately disrupted, leading to the leakage of pancreatic fluid into the abdominal cavity (Fig. 4A)^23,24. After 4 h of constructing the model, the main pancreatic duct, pancreas, and surrounding tissues were coated with normal saline (Ctrl), CP-CNDS or CNDS respectively. The occurrence of a peculiar event unfolded, wherein the suppressive effect on digestive enzyme activity by CP-CNDS surpassed that of CP-CNDS itself, as substantiated by the fluorescence indicator depicting chymotrypsin activity (Fig. 4B). After 1 day of surgery, the abdominal adhesions and pancreatitis were evaluated through laparotomy, revealing that CP-CNDS effectively suppressed the development of abdominal adhesions and pancreatitis (Supplementary Fig. 11), thus demonstrating its anti-inflammatory properties following POPF. Furthermore, the inflammatory response was evaluated by quantifying splenic immune cells using ssGSEA analysis of RNA-seq results from the spleen. Compared to the mock treatment, CP-CNDS was found to significantly down-regulate almost all immune cells that have been observed, including B lymphocytes (B cells), T lymphocytes (T cells), natural killer cells (NK), dendritic cells (DCs), macrophages (Mø) and granulocytes (GRs) surpassing its L-enantiomer CNDS in efficacy as shown in Fig. 4C. More significantly, the administration of CP-CNDS treatment exhibited high efficacy in inhibiting abscess of abdominal tissues and organs as well as subsequent inflammation, surpassing its L-enantiomer CNDS in terms of action (Fig. 4D). The H&E staining of organ slices from the Ctrl group revealed evident damage to the three layers of abdominal wall muscle fibers, characterized by degeneration, atrophy, irregular arrangement, and peripheral edema. Additionally, hepatocytes in the liver exhibited noticeable atrophy in the Ctrl group. The white pulp structure of the spleen in the Ctrl group displayed signs of atrophy with reduced lymphocyte count, while congestion was observed in the red pulp. Local sections of duodenal villi showed heart-shaped changes along with necrosis and submucosal edema in the Ctrl group. Furthermore, localized degenerative alterations were found in gastric mucosa accompanied by interstitial congestion in the Ctrl. Pancreatic acinar cells demonstrated degeneration and necrosis alongside proliferation of interstitial connective tissue and infiltration of numerous inflammatory cells. At the same time, bleeding was also observed within the pancreatic tissue of Ctrl group. In the CP-CNDS group, a significant reduction in inflammatory cells was observed in the abdominal interstitial tissue, which exhibited a scattered distribution. The liver, stomach, and spleen displayed normal tissue structure without any apparent abnormalities within the CP-CNDS group slices. Compared to the Ctrl group, there was a notable decrease in congestion and interstitial edema of the pancreas as well as inflammatory cell infiltration within the lamina propria of the duodenum in the CP-CNDS group. In CNDS group, slight atrophy of muscle fibers in the third layer of abdominal wall along with surrounding edema occurred, which was more severe than that observed in CP-CNDS. A small number of inflammatory cells infiltrated around hepatic sinuses within CNDS group. There were also signs of congestion in lamina propria of duodenum and infiltration of inflammatory cells present within CNDS group. Additionally, gastric mucosa atrophy accompanied by inflammatory cell infiltration was evident in CNDS group. The tissue structure of spleen remained normal within CNDS group. Furthermore, compared to CP-CNDS, pancreatic interstitial edema and increased inflammatory cell infiltration were observed to be more pronounced in CNDS. The organ tissue damage in the Ctrl group was observed to be the most severe, followed by the CNDS group, while the CP-CNDS group exhibited the least significant damage. The findings demonstrated that the administration of CP-CNDS resulted in a substantial alleviation of POPF.

Fig. 4. The administration of CP-CNDS for POPF (postoperative pancreatic fistula) in vivo.

A Diagrammatic sketch of the classical rat model for POPF. B The pancreatic chymotrypsin fluorescence evaluation for CP-CNDS and CNDS in vivo, depicting CP-CNDS leading to a better alleviation of POPF. Scale bars indicate 1 cm. C The ssGSEA analysis of RNA-seq for CP-CNDS and CNDS, including B lymphocytes (B cells), T lymphocytes (T cells), natural killer cells (NK), dendritic cells (DCs), macrophages (Mø) and granulocytes (GRs). D The macroscopic organs, histological morphology and H&E staining evaluation for CP-CNDS and CNDS in vivo. Green arrows indicate abnormal damage. The red dotted line indicates the inflammatory damage resulting from the pancreatic fistula. The pink dotted line indicates the organ hemorrhage resulting from the pancreatic fistula. Scale bars indicate 200 μm. The experiment depicted in (D) was independently replicated a minimum of three times, resulting in consistent findings.

The therapeutic effect of CP-CNDS against POPF was further demonstrated by monitoring the rats for 7 days after surgery, revealing a contrast in survival rates. All rats treated with CP-CNDS survived, while the median survival times for mock-treated and CNDS-treated mice were only 56 and 108 h respectively (Fig. 5A). Furthermore, the levels of ascitic amylase and lipase in CP-CNDS-treated rats exhibited a significant reduction compared to those in mock-treated rats. Conversely, both ascitic enzymes displayed an elevation at day 5 post-surgery in CNDS-treated rats, presumably attributed to the rapid degradation of CNDS (Fig. 5B). Additionally, the serum levels of amylase and lipase in CP-CNDS-treated rats continued to decline throughout the 7-day postoperative monitoring period, while both remained elevated in PBS-treated mice until their demise (Fig. 5C). Besides, a comprehensive analysis of 20 inflammatory indicators in the blood revealed that CP-CNDS treatment exhibited the most important reduction in inflammatory markers compared to the persistently elevated levels observed with PBS treatment and the rapid escalation witnessed on the fifth day in the CNDS treatment group (Fig. 5D). Further analysis of the eight inflammatory markers with the most pronounced differences, including TNF α, IL-6, IL-8, IL-33, sST2, CPR, HMGB1 and MPO revealed that treatment with CP-CNDS not only promoted their decline but also restored them to normal levels (Fig. 5E), thus once again demonstrating the efficacy of CP-CNDS in POPF therapy.

Fig. 5. The therapeutic effect of CP-CNDS against POPF in vivo.

A The survival rates of rats treating with CP-CNDS or CNDS. B The evaluation of ascitic amylase and ascitic lipase with the treatment of CP-CNDS or CNDS in vivo (n = 5 biological replicates in each group, means ± SD, statistical analysis was performed using ANOVA LSD test). C The evaluation of serous amylase and serous lipase with the treatment of CP-CNDS or CNDS in vivo (n = 5 biological replicates in each group, means ± SD, statistical analysis was performed using ANOVA LSD test). D The comprehensive analysis of 20 inflammatory indicators in the blood treated with CP-CNDS or CNDS in vivo. E The detained analysis of the eight inflammatory markers with the most pronounced differences, including TNF α, IL-6, IL-8, IL-33, sST2, CPR, HMGB1 and MPO (n = 5 biological replicates in each group). All box-whisker plots (E) center on the median; the bounds of the boxes mark the upper and lower quartile.

CP-CNDS possessed the favorable biosafety profiles

The in vitro and in vivo biosafety of CP-CNDS or CNDS was further investigated to substantiate its potential for clinical translation. The 5% CP-CNDS or CNDS exhibited no toxicity towards human vascular endothelial cells (HUVEC), as demonstrated by the results of live-dead staining (Fig. 6A) and apoptosis evaluation (Fig. 6B). Furthermore, the intraperitoneal administration of CP-CNDS or CNDS at a dosage ten times higher than the therapeutic level exhibited no discernible toxicity towards healthy rats. The administration of CP-CNDS or CNDS, in comparison to the isopycnic NS, yielded negligible effects on both routine blood analysis (Fig. 6C) and immunotoxicity-related blood biochemical indexes (Fig. 6D). Furthermore, no signs of hepatotoxicity or nephrotoxicity were observed in CP-CNDS/CNDS-treated rats, as demonstrated by the histopathological examination, liver and renal function biochemical tests, and immune factors analysis in the liver and kidney (Fig. 6E, F). Additionally, the histopathological examination and analysis of immune factors in each organ have substantiated the pulmonary, cardiac, and splenic safety of CP-CNDS or CNDS (Fig. 6G–I). Moreover, the histopathological examination of the stomach and intestine provided additional evidence supporting the safety of CP-CNDS or CNDS (Supplementary Fig. 12). Collectively, CP-CNDS or CNDS possessed the favorable biosafety profiles, thus significantly bolstering its potential for clinical translation.

Fig. 6. The favorable biosafety of CP-CNDS in vitro and in vivo.

A The Live/Dead staining of human vascular endothelial cells (HUVEC) treated with CP-CNDS, CNDS or PBS. B The cell viability test of HUVEC co-cultured with 5% CP-CNDS, CNDS or PBS (n = 3 biological replicates in each group, statistical analysis was performed using two-sided t-test). The apoptosis evaluation of HUVEC co-cultured with 5% CP-CNDS, CNDS or PBS by flow cytometry. The gating strategy for apoptosis evaluation involves two sequential steps: (1) identification of the cell cluster, and (2) characterization of apoptosis or necrosis as cells positive for FITC+ or PI+, cell survival as cells negative for FITC− and PI−. C, D The routine blood analysis and immunotoxicity-related blood biochemical indexes of rats treated with the CP-CNDS, CNDS or PBS using intraperitoneal administration (n = 5 biological replicates in each group). E Hepatotoxicity measured by pathological section of liver, ALT (glutamic-pyruvic transaminase), AST (glutamic-oxalacetic transaminase), total bilirubin (TBIL) and inflammatory cytokines (IL-6, TNF-α, MCP-1and NF-κB) of liver (the H&E of liver: n = 3 biological replicates; ALT, AST, TBIL and inflammatory cytokines: n = 5 biological replicates in each group). F Nephrotoxicity test, including pathological section of kidney, CR (creatinine), BUN (Blood Urea Nitrogen), albumin (ALB) and inflammatory cytokines (IL-6, TNF-α, MCP-1and NF-κB) of kidney (the H&E of kidney: n = 3 biological replicates; CR, BUN, ALB and inflammatory cytokines n = 5 biological replicates in each group). G Toxicity evaluation of lung displayed by H&E staining images of lung sections and IL-6, TNF-α, MCP-1and NF-κB cytokines content in alveolar lavage fluid. H Cardiotoxicity reflected by pathological section and inflammatory cytokines (IL-6, TNF-α, MCP-1and NF-κB) infiltrating in heart. I The H&E staining of spleen sections, and IL-6, TNF-α, MCP-1and NF-κB cytokines were involved in safety detection of CP-CNDS, CNDS or PBS in spleen. The data from each group were presented as the mean ± SD. Scale bar: 100 μm. All box-whisker plots (C, E, F) center on the median; the bounds of the boxes mark the upper and lower quartile. Statistical analysis in (C–I) was performed using two-sided t-test.

Here, we show a cleverly devised regulatory strategy for enzyme leakage, achieved through the biomolecule capture via intermolecular interaction with a chiral-engineered peptide supramolecular gel (CP-CNDS). In detail, the present study encompasses the development of a chiral-engineered peptide (CP), which exhibits a propensity for forming gels CP-CNDS driven by intermolecular interaction, while also demonstrating satisfactory affinity towards trypsin, chymotrypsin, and lipase. Moreover, CP-CNDS possesses the ability to capture these digestive enzymes from solution into gel, effectively entrapping them within the internal gel and preventing their dynamic exchange with their counterparts in the external milieu. The activity of trypsin, chymotrypsin, and lipase was completely suppressed by CP-CNDS, resulting in a potent mitigation of POPF. With the administration of CP-CNDS, all rats afflicted with POPF not only survived but also experienced a recovery within a mere 3 days. In stark contrast, the absence of any treatment led to an alarming survival rate of less than 5 days due to the occurrence of postoperative pancreatic fistula, leak or abscess. The collective findings presented by CP-CNDS not only offer promising avenues for preventing and treating POPF but also exemplify the precision medicine-guided regulation of protein interaction that effectively targets specific pathogenic molecules across multiple diseases.

Methods

Molecular dynamics (MD) simulation

Using the GROMACS software package (version 2021.4) combined with the MARTINI force field (version 2.2) to construct a coarse-grained model of peptides, a 500 ns MD was performed on an aqueous system consisting of 100 × RADA and 5 typsin or lipase or chymotrypsin peptides simulation, which PDB ID are 1H4W, 1LPA and 4H4W respectively. In the initial state, peptide molecules are randomly dispersed in a cubic box of 20 × 20 × 20 nm³ and filled with coarse-grained water. Cl⁻ ions and Na⁺ ions are added to the box to neutralize the system charge. The cutoff for calculation of electrostatic interactions and van der Waals forces was set at 1 nm. The peptide group and the water/ion group were coupled to an external temperature bath using velocity, respectively. Rescaling method and pressure bath using the Parrinello-Rahman method. Update the neighbor list every 10 steps using a Verlet buffer with a cutoff distance of 1 nm. Trajectory analysis using Gromacs internal tools. Coarse-grained remapping to an all-atom model is achieved through the Backward method developed by Tsjerk Wassenaar. The hydrogen bond number statistics are realized by the hydrogen bond analysis tool developed by Gromacs. The two-dimensional free energy landscape (FEL) is calculated as: RTln[P(X-/Z-axis gyration radius, RMSD)] + C, where P(X-/Z-axis gyration radius, RMSD) is RMSD and Probability distribution function for X-/Z-axis gyration radius, C is a constant used to shift the highest free energy to zero. Pymol and VMD software were used for trajectory visualization and peptide conformation analysis.

The evaluation of binding affinity via fluorescence polarization

Using the solid-phase peptide synthesis (SPPS) method, the N-terminal cysteine of peptide ^DCP1 or ^DCP2 was labeled with fluorescein isothiocyanate. The labeled peptide was purified by preparative C4 reversed-phase high-performance liquid chromatography (RP-HPLC) and lyophilized. CP-CNDP with fluorescence was obtained by mixing ^DCP1 and ^DCP2with mass ratio 2:3. Fluorescence polarization experiments of CP-CNDS with trypsin, lipase, and chymotrypsin were conducted in Corning® 96-well black plates. SpectraMax M5 multi-mode microplate reader was used for reading. A series of enzyme solutions with varying concentrations (maximum concentrations of trypsin, lipase, and chymotrypsin were 80 μM, 10 μM, and 10 μM, respectively) were prepared in Tris-HCl buffer (10 mM Tris, 150 mM NaCl, 1 mM EDTA, pH 7.0). These enzyme solutions were added to the 96-well plate and incubated with 100 nM CP-CNDS for 30 min at room temperature, with triplicates. Fluorescence polarization was measured under excitation at 470 nm and emission at 530 nm. The fluorescence polarization value (Fp) was calculated using the formula:

$$Fp=\frac{\left(S-P\times G\right)}{\left(S+P\times G\right)}\times 1000$$

where S is the perpendicular fluorescence intensity, P is the parallel fluorescence intensity, and G is the instrument correction factor, with Fp expressed in mP units.

For data analysis, the concentration-fluorescence polarization data were fitted using the logistic function in Origin 2021 software. The calculation formula for K_d value is:

$$K_d={\mathrm{EC}}_{50}-\frac{c_{\mathrm{CP}-\mathrm{CNDS}}}{2}$$

where EC₅₀ is the EC₅₀ value obtained from the fitting curve, and $c_{\mathrm{CP}-\mathrm{CNDS}}$ is the concentration of peptide CP-CNDS, i.e., 100 nM.

The main pancreatic duct transection model of rat

Animal studies were performed according to the protocols approved by the Institution Guidelines and were approved by the Laboratory Animal Center of Xijing Hospital of The Fourth Military Medical University. The animal ethics number is No. IACUC-20220308. ALL rats were purchased from the Laboratory Animal Center of The Fourth Military Medical University. The mice were housed under standard specific pathogen-free conditions with a 12 h–12 h light–dark cycle. All SD rats used in the experiment were male. This choice was made due to the consistent pancreatic tissue structure between male and female rats, as well as the absence of sex-dependent conclusion in pancreatic fistula formation. Therefore, male SD rats were randomly selected as experimental subjects.

Forty-five male SD rats aged 8–9 weeks were selected to establish the pancreatic fistula model by the pancreatic main transection method. Before the surgery began, the rats’ abdominal skin was prepared. On the day of surgery, rats were anesthetized with isoflurane, the skin was sterilized with 75% alcohol, and the pancreatic heads were separated and exposed by layer-by-layer laparotomy. After locating the branches of the pancreatic charge, the rat pancreatic charge was transected without damaging the capillaries surrounding the duct. Immediately after that, the transected main catheter was filled with 0.3 mL of normal saline, CP-CNDS, and CNDS, respectively. The 1.875 mg CP-CNDS or CNDS was dissolved in 0.3 ml aqueous solution of calcium chloride (0.1 mg/ml). Subsequently, the abdominal cavity opening of the homemade drainage tube was placed around the transcribed pancreatic duct. It was fixed to the abdomen with sutures, and the abdomen was closed layer by layer, then the rats were placed in a rewarming box until awakening. The postoperative condition of certain SD rats was monitored (n = 5). Rat survival in each group was recorded and a survival curve was drawn (n = 10).

After repeating the model and treatment as described above (15 male SD rats aged 8–9 weeks), sterile gauze soaked in a small amount of normal saline was applied around the abdominal cavity (to prevent dehydration of abdominal organs) while the rats remained anesthetized. After 30 min, 200 μL of the prepared MeO-Succ-Arg-Pro-Tyr-AMO solution (ATT Bioquest) was added to the pancreatic stump and surrounding tissues. In the dark field, API filters were used to observe and photograph in real-time. The real-time extent of the pancreatic fistula was determined by observing the fluorescence area.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

W. He. and J. Yan. designed the study and analyzed the data. Y. Wang., X. Li., Y. Ji., J. Yuan., W. Yang. and S. Yan. performed the experiments. W. He. and Y. Wang. wrote the manuscript. J. Yan. revised the manuscript. J. Yan. supervised the project.

Peer review

Peer review information

Nature Communications thanks Jakob Troppmair and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences, GSA: CRA017925. All the other data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-024-51734-7.