Multimodal hybrid nanozymes with antioxidant catalytic and antibiotic-free antibacterial activities for enhanced multi-target sepsis therapy

Summary

Sepsis, characterized by its complex pathophysiology, presents significant challenges for clinical treatment. An integrative approach combining highly effective antibacterial measures, immunomodulation, and organ protection is urgently needed to enhance the therapeutic efficacy. Here, we construct hybrid cerium-baicalein nanozymes (Ce-BE NZs), which exhibit broad-spectrum non-antibiotic antibacterial and redox enzyme-mimicking activities, effectively scavenging reactive oxygen species and reducing inflammatory mediators in lipopolysaccharide-stimulated macrophages. Ce-BE NZs also correct immune dysregulation, reduce liver injury, and extend survival in both cecal ligation and puncture and “two-hit” sepsis models. Mechanistically, Ce-BE NZs inhibit ferroptosis and mitigate mitochondrial dysfunction by promoting ferritin heavy chain-1 expression, thereby enhancing multi-target sepsis therapy. Additionally, they mitigate ferroptosis and cell damage in a macrophage-incorporating human liver-derived organoid model. Overall, Ce-BE NZs represent a promising multi-target therapy for sepsis and may pave the way for an antibiotic-free and transnational approach to treating other infectious diseases.

Graphical abstract

Highlights

• •Ce-BE NZs exhibit robust antibacterial and redox enzyme-mimicking activities
• •Ce-BE NZs exert a multi-target therapeutic efficacy in sepsis models
• •Ce-BE NZs mitigate ferroptosis and mitochondrial dysfunction via FTH1 upregulation
• •Ce-BE NZs mitigate ferroptosis and cell damage in the human liver organoid model

Zhu et al. develop Ce-baicalein nanozymes (Ce-BE NZs) with broad-spectrum antibacterial and redox enzyme-mimicking activities. These Ce-BE NZs show multi-target therapeutic efficacy in sepsis models, thus offering a strategy for the treatment of sepsis and other infectious diseases.

Introduction

Sepsis, a highly lethal syndrome caused by a complex dysregulated host response to infection (e.g., bacterial infection and COVID-19), imposes a staggering global burden, with approximately 48.9 million annual cases and 11 million deaths,¹ primarily due to sepsis-induced multiple organ failure. Among affected organs, the liver plays a dual role as both a frontline immune defender and a vulnerable target during sepsis.^2,3 Accumulating evidence has shown that sepsis-induced acute liver injury substantially impairs the prognosis of sepsis and is one of the leading causes of death in patients with sepsis.^4,5 Currently, sepsis management focuses on symptomatic treatments such as antibiotics, fluid resuscitation, and extracorporeal oxygenation, which fail to address the underlying pathophysiological complexity and reduce persistently high mortality rates.^6,7 The lack of specific interventions for sepsis and sepsis-induced liver injury highlights a critical gap in sepsis therapeutics.^8,9

The pathogenesis of sepsis involves three key dynamically interconnected events: pathogen invasion, dysregulated immune responses, and multiorgan damage.^10,11 In bacterial sepsis, infections provoke a systemic inflammatory response, leading to the release of pro-inflammatory cytokines from activated macrophages. These cytokines disrupt the mitochondrial respiratory chain, triggering the overproduction of reactive oxygen species (ROS), which results in mitochondrial dysfunction and ferroptosis,^11,12 ultimately leading to acute liver injury and even sepsis-related death. This multiphasic pathophysiology renders conventional single-target therapies, including anti-infection or anti-inflammation,¹³ insufficient to break the vicious cycle of infection, inflammation, and organ injury, resulting in limited therapeutic effects in clinical trials.¹⁴

Nanomedicine has emerged as a transformative approach for multifactorial diseases including sepsis.^15,16,17 Although nanomaterials with antioxidant/anti-inflammatory (e.g., CeO₂) or antibacterial properties show promise in sepsis treatment, their monotherapeutic design overlooks the interconnected nature of sepsis pathophysiology,^18,19 thereby hindering their potential clinical application in combating the multifaceted pathophysiology of sepsis.

Metal-coordination nanozymes, multifunctional nanomaterials formed through the coordination of metal ions with natural polyphenols, offer a promising integrated therapy due to their ease of preparation, stable physicochemical properties, and biocompatibility.^20,21 Baicalein (BE), a natural polyphenol flavonoid isolated from the root of Scutellaria baicalensis Georgi, exhibits antiviral, immunomodulatory, and antibacterial properties.²² However, its clinical translation is hindered by poor aqueous solubility and low bioavailability.^23,24 Therefore, we hypothesized that metal-BE nanozymes could boost the bioavailability and antibacterial activity of BE. Moreover, these nanozymes are endowed with multi-enzyme catalytic functions, thereby providing a multi-target therapeutic strategy by targeting the infection-inflammation-organ damage axis for sepsis and sepsis-induced liver injury.

Here, we present a promising and facile strategy to develop natural product-based metal polyphenol nanozymes (cerium-BE nanozymes [Ce-BE NZs]) with favorable biocompatibility by coordinating BE with Ce³⁺ ions based on antioxidant activity-guided screening. The Ce-BE NZs possess broad-spectrum non-antibiotic antibacterial and redox enzyme-mimicking catalytic activities including superoxide dismutase (SOD)- and catalase (CAT)-mimicking activities. It can scavenge overproduced ROS and inhibit the secretion of inflammatory mediators in lipopolysaccharide (LPS)-stimulated mouse macrophages. Meanwhile, Ce-BE NZs have demonstrated their potential as a promising multi-target therapy by effectively reducing oxidative stress, correcting dysregulated immune responses, alleviating liver injury, and prolonging the survival of septic mice in a clinically relevant sepsis model induced by cecal ligation and puncture (CLP). Furthermore, mechanism studies indicated that Ce-BE NZs inhibited ferroptosis and mitigated mitochondrial dysfunction by promoting ferritin heavy chain-1 (FTH1) expression. More importantly, Ce-BE NZs could mitigate cell damage and ferroptosis in a macrophage-incorporating human liver-derived organoid culture model. Collectively, our findings demonstrate that such accessible, safe, and multifunctional Ce-BE NZs might be a promising therapeutic choice for clinical synergetic treatment of sepsis and sepsis-induced liver injury.

Results

Synthesis and characterization of Ce-BE NZs

In this study, common metal ions such as iron (Fe), copper (Cu), zinc (Zn), manganese (Mn), and cerium (Ce) ions were coordinated with BE to form metal polyphenol nanozymes, and Ce-BE NZs, with the strongest antioxidative ability (Figures S1A–S1C), were used for further studies. CeCl₃ and poly(vinylpyrrolidone) (PVP) in methanol were mixed, and the methanol solution of BE was gradually added dropwise with thorough stirring at 25°C. Thereafter, the above-obtained liquid was stirred and dialyzed in 0.9% NaCl overnight to obtain the Ce-BE NZ aqueous solution, which changed from light yellow to dark brown in appearance significantly (Figure S1D). Fortunately, Ce-BE NZs had a water solubility of 24 mg/mL (Figure S1E), approximately 1,400 times higher than that of BE (16.82 μg/mL).²⁵ Transmission electron microscopy (TEM) showed a uniform, regular dot distribution with an average diameter of 3.4 nm (Figures 1A and S1F). Elemental mapping confirmed the presence of C, O, and Ce elements (Figure 1A), suggesting successful coordination of Ce³⁺ ions coordination with BE.

Figure 1.

Synthesis and characterization of Ce-BE NZs

(A) TEM images and elemental mapping images of Ce-BE NZs (scale bars, 100 nm, n = 3).

(B) FTIR spectra of BE and Ce-BE NZs (n = 3).

(C–E) XPS patterns of Ce-BE NZs and high resolution of C 1s, O 1s, and Ce 3d peaks (n = 3).

(F) UV-vis spectra of Ce³⁺ ions, BE, and Ce-BE NZs (n = 3).

(G) ROS-scavenging rate of Ce-BE NZs based on MB UV-vis absorbance (n = 3).

(H) DPPH radical-scavenging rate of Ce-BE NZs and Vc (n = 3).

(I) ABTS radical scavenging rate of Ce-BE NZs and Vc (n = 3).

(J) ESR spectra of DMPO indicating ⋅OH capture with or without Ce-BE NZs (n = 3).

(K) ESR spectra of TEMPO indicating ¹O₂ scavenging ability of Ce-BE NZs (n = 3).

(L and M) ESR spectra of BMPO and spectrophotometry indicating the SOD-mimicking activity of Ce-BE NZs (n = 3).

(N and O) Dissolved oxygen content and ESR spectra indicating the CAT-mimicking activity of Ce-BE NZs (n = 3).

(P) Michaelis-Menten kinetic analysis and Lineweaver-Burk plotting for Ce-BE NZs with H₂O₂ as a substrate (n = 5).

Data are represented as means ± SD. ∗∗p < 0.01, ∗∗∗p < 0.001 for the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test.

Dynamic light scattering (DLS) showed that the hydrodynamic size of Ce-BE NZs was approximately 14.2 nm (Figure S1G). Fourier transform infrared (FTIR) spectra (Figure 1B) revealed changes in the characteristic peak at 1,150–1,200 cm⁻¹ (C−OH), with an increase in infrared intensity. Meanwhile, X-ray photoelectron spectroscopy (XPS) also confirmed the successful synthesis of Ce-BE NZs, with peaks at 885, 531, and 285 eV corresponding to Ce, O, and C elements, respectively (Figure 1C). Specifically, Figure 1D presents the C1s XPS spectra of Ce-BE NZs, with the main peak at 284.8 eV corresponding to the unsubstituted carbons on the phenyl ring (C−C) and a shoulder at 285.8 eV corresponding to C−O/C=O bonds.²⁶ Notably, the Ce3d spectra of Ce-BE NZs showed four peaks after deconvolution (Figure 1E). The binding energies at 903.7 and 899.66 eV were assigned to the Ce⁴⁺ ion oxidation state, while the peaks at 884.89 and 880.76 eV were attributed to the Ce³⁺ ion oxidation state, and the ratio of Ce⁴⁺/Ce³⁺ in Ce-BE NZs was determined to be 0.57. Further, in comparison with the maximum peak of BE at 360 nm, the ultraviolet-visible (UV-vis) spectra of Ce-BE NZs exhibited a red shift, indicating the occurrence of coordination between the Ce³⁺ ions and BE (Figure 1F).

The resulting water-soluble Ce-BE NZs showed good dispersity and stability (Figure S1H), which was attributed to the PVP modification and a highly negative potential (about −25 mV) (Figure S1I). Next, the results of DLS and UV-vis analysis indicated the high stability of Ce-BE NZs in various physiological solutions for a duration of 7 days (Figures S1J and S1K). Taken together, our results confirm the successful synthesis of the Ce-BE NZs, which is advantageous for future biomedical applications.

Powerful free radical scavenging and SOD- and CAT-mimicking catalytic activities of Ce-BE NZs

Massive production of ROS is a key factor in organ injury during sepsis.²⁷ Thus, we first evaluated the in vitro antioxidant capacity of Ce-BE NZs. A methylene blue (MB) test was performed to determine the hydroxyl radical (⋅OH)-scavenging capacity of Ce-BE NZs. As expected, MB recovered the characteristic absorption wavelength of the double peaks by gradually increasing the concentration of Ce-BE NZs, indicating the strong scavenging capability of Ce-BE NZs for ⋅OH generated during the Fenton reaction (Figure 1G). Additionally, the 1,1-diphenyl-2-trinitrophenylhydrazine (DPPH) radical is stable and has a single electron. When antioxidants were present, DPPH free radicals were removed, the color of the solution became lighter, and the absorbance at 515 nm decreased. Our results show that when Ce-BE NZs were added at a concentration of 96 μg/mL, they scavenged approximately 75% of DPPH, comparable to that of vitamin C (Vc) at a concentration of 300 μg/mL (Figure 1H). Similarly, at a concentration of 96 μg/mL, Ce-BE NZs could almost eliminate 2, 2′-azobis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), which is comparable to that of Vc at a concentration of 350 μg/mL (Figure 1I). Furthermore, we detected ⋅OH and singlet oxygen (¹O₂) by electron spin resonance (ESR) spectroscopy using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) or 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) as free radical-trapping agents. The intensity of the spectral lines indicated that the ⋅OH/¹O₂ in the Ce-BE NZs treatment group decreased (Figures 1J and 1K), indicating the efficient scavenging ability of Ce-BE NZs against free radicals.²⁸ Overall, the results show that Ce-BE NZs have strong antioxidant properties and would act as highly effective ROS-scavenging agents.

SOD,²⁹ as one of the most vital intracellular enzymes, catalyzes the disproportionation of superoxide anion radicals (⋅${\mathrm{O}}_2^{-}$) to hydrogen peroxide (H₂O₂) and oxygen (O₂). The ESR spectra detected ⋅${\mathrm{O}}_2^{-}$ using 5-tert-butoxycarbonyl 5-methyl-1-pyrroline N-oxide (BMPO) to probe the intrinsic SOD-mimicking activity of Ce-BE NZs. These results suggest that Ce-BE NZs (1 mg/mL) had strong SOD-mimicking activity relative to BE (1 mg/mL), comparable to that of natural SOD at a concentration of 150 U/mL (Figure 1L). Meanwhile, the total SOD (T-SOD) activity of Ce-BE NZs was determined by the hydroxylamine method using a xanthine and xanthine oxidase reaction system, and the results were in general agreement with those of the ESR spectra (Figure 1M).

CAT,^30,31 another important intracellular enzyme, catalyzes the disproportionation of H₂O₂ to H₂O and O₂, facilitating the determination of the CAT-mimicking ability through oxygen production with reaction time after the decomposition of H₂O₂. The results of the oxygen release experiment suggest that compared with CeCl₃ or BE alone, Ce-BE NZs had robust CAT-mimicking activity, which was equivalent to 200 U/mL of natural CAT (Figure 1N); the ESR findings also confirm these results (Figure 1O). To further investigate the catalytic activity of Ce-BE NZs toward H₂O₂, the Michaelis constant (Km) and maximum reaction rate (Vmax) were determined to be 0.24 mM and 0.16 × 10⁶ M⋅ s⁻¹, respectively, based on the classical Michaelis-Menten equation and Lineweaver-Burk kinetic analysis (Figure 1P). The exceptional catalytic capabilities of Ce-BE NZs are derived from electron transfer and valence state alterations at their Ce ions’ active centers. Ce⁴⁺ is crucial for the CAT-mimicking activity in Ce-based nanozymes, and the SOD-like activity is closely linked to a high Ce³⁺/Ce⁴⁺ ratio of Ce-based nanozymes.^32,33 Taken together, our results revealed that Ce-BE NZs are endowed with SOD- and CAT-mimicking activities, converting toxic waste (⋅${\mathrm{O}}_2^{-}$ and H₂O₂) into nontoxic and useful products (H₂O and O₂).

Broad-spectrum antibacterial activities of Ce-BE NZs

We selected Staphylococcus aureus (S. aureus, a typical Gram-positive bacterium, ATCC 25923), Escherichia coli (E. coli, a typical Gram-negative bacterium, ATCC 25922), and Pseudomonas aeruginosa (PA., ATCC 27853) as typical bacteria for sepsis to characterize the antibacterial activities of Ce-BE NZs. Meanwhile, cefuroxime sodium (CM) or gentamicin (GM) was used as a positive control. Using the agar plate assay method, we found that the antibacterial rates of Ce-BE NZs (1 mg/mL) against S. aureus, E. coli, and PA. were 73.67% ± 5.51%, 65.33% ± 5.86%, and 39.67% ± 7.51%, respectively (Figures 2A and 2B), and the minimum inhibitory concentration of Ce-BE NZs against S. aureus, E. coli, and PA. were 1.2, 1.2, and 1.5 mg/mL, respectively (Figure 2C). Notably, the antibacterial activities of Ce-BE NZs against S. aureus and E. coli were generally reached at the same concentration of CM and GM, respectively, whereas BE did not significantly reduce the number of colonies at the same concentration. To visualize the antibacterial performance, the morphology of the bacteria was examined through scanning electron microscopy (SEM), which demonstrated that deformed, wrinkled surfaces appeared and broke the cell membrane with Ce-BE NZs (Figure 2D). The SEM results suggest that the destruction of the integrity of bacterial cell membranes is one of the important mechanisms of Ce-BE NZs against bacteria. Consistently, live-dead bacteria staining also indicated that Ce-BE NZs had substantial antibacterial activity (Figure 2E). Furthermore, the intracellular BE content in bacteria treated with Ce-BE NZs was significantly enhanced compared to that in bacteria treated with BE alone (Figures S2A–S2n), suggesting that the enhanced antibacterial activity of Ce-BE NZs over BE can be attributed to the improvement of BE’s solubility by metal-polyphenol coordination, thereby maximizing its antibacterial activity.

Figure 2.

Antibacterial activity of Ce-BE NZs

(A) Agar plate assay of S. aureus, E. coli, and PA. with different treatments (for the positive control groups, CM was used for S. aureus, while GM was employed for both E. coli and PA., n = 4).

(B) Antibacterial ratios of different kinds of bacteria after various treatments (n = 3).

(C) Determination of the minimum inhibitory concentration of Ce-BE NZs against S. aureus, E. coli, and PA. via the agar plate method (n = 5).

(D) Typical SEM images of different kinds of bacteria after various treatments. Red arrows point to the disrupted bacterial membrane (scale bars, 500 nm, n = 3).

(E) Live-dead bacteria staining, green: live bacteria; red: dead bacteria (scale bars, 200 μm, n = 3).

Data are represented as means ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 for the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. ns, not significant.

To further verify the antibacterial ability of Ce-BE NZs in vivo, we observed changes in intraperitoneal injection of PA. within 12 h using bioluminescence imaging. The results showed that the fluorescence signal was gradually weakened after Ce-BE NZ treatment, but there was no significant change in the treatment group with BE alone (Figure S2O). These results indicate that Ce-BE NZs exhibit superior antibacterial activity compared to BE, offering a promising avenue for antibiotic replacement therapy.

Ce-BE NZs alleviate the inflammatory response of LPS-induced macrophages

To evaluate the capacity of Ce-BE NZs to scavenge intracellular ROS, we initiated our study by pretreated mouse macrophages (RAW 264.7) with PBS or varying concentrations of Ce-BE NZs or BE for 1 h (Figure 3A). Subsequently, the cells were treated with 1 μg/mL of LPS for 24 h, and then analyzed for changes in intracellular ROS levels using DCF-DA staining. Flow cytometry results indicated that LPS significantly increased the percentage of ROS^high cells. Both Ce-BE NZs and BE significantly reduced intracellular ROS levels. Notably, Ce-BE NZs demonstrated their ROS-eliminating properties at lower concentrations, notably at 1 μg/mL (Figures 3B–3D).

Figure 3.

Effects of Ce-BE NZs on LPS-induced oxidative stress and inflammatory responses in macrophages

(A) Schematic representation of the experimental model.

(B) Flow cytometry analysis of intracellular ROS levels (n = 4).

(C) Statistical analysis of the percentage of ROS^high cells (n = 4).

(D) Statistical analysis of the mean fluorescence intensity (MFI) of intracellular DCF-DA (n = 4).

(E) Immunofluorescence of IL-6, iNOS, and PTGS2 protein (scale bars, 20 μm, n = 4).

(F–H) Statistical analyses of the fluorescence intensities of IL-6, iNOS, and PTGS2 (n = 3).

(I and J) Analyses of Il1b and Il6 gene expressions (n = 4).

(K) Analysis of NO concentration in cell supernatants (n = 5).

(L and M) ELISA analysis of MCP-1 and TNF-α concentrations in cell culture supernatants (n = 4).

Data are represented as means ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 for the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. ns, not significant.

To further assess immunomodulatory effects, we treated RAW 264.7 cells with Ce-BE NZs or BE. Immunofluorescence results indicate that LPS upregulated interleukin (IL) 6, inducible nitric oxide (NO) synthase (iNOS), and prostaglandin-endoperoxide synthase 2 (PTGS2) protein expression levels (Figures 3E–3H). Complementing these findings, qPCR analysis showed that treatment with 1 or 10 μg/mL Ce-BE NZs for 24 h in the presence of LPS significantly reduced Il1b and Il6 expression levels (Figures 3I and 3J). The NO concentration in the supernatant of LPS-treated cells was significantly elevated, while treatment with 1 or 10 μg/mL of Ce-BE markedly suppressed NO release (Figure 3K). ELISA analysis further showed that compared to BE, Ce-BE NZs significantly reduced monocyte chemotactic protein-1 (MCP-1) and tumor necrosis factor α (TNF-α) concentrations in the cell supernatant (Figures 3L and 3M). Importantly, Ce-BE NZ treatment effectively mitigated M1 marker expression, underscoring its ability to impede the LPS-driven shift of RAW 264.7 cells toward the M1 phenotype. In summary, these findings underscore that Ce-BE NZs can scavenge ROS in LPS-induced macrophages in vitro. Furthermore, they effectively inhibit the secretion of inflammatory factors and suppress the phenotypic polarization of macrophages toward the M1 subtype.

Ce-BE NZs exhibit multi-target therapeutic effect to rescue the CLP and “two-hit” sepsis models

Owing to the ROS-scavenging, anti-bacterial, and immunomodulatory effects of Ce-BE NZs in vitro, we further explored the therapeutic efficacy of Ce-BE NZs in septic models. A well-established septic mouse model was induced by CLP, and the effects of Ce-BE NZs on survival rate, hematological parameters, and microcirculatory dysfunction were examined (Figure 4A).

Figure 4.

Therapeutic efficacy of Ce-BE NZs in the CLP and “two-hit” sepsis mice models

(A) Schematic diagram depicting the establishment and study schedule of the CLP mice.

(B and C) Immediate (n = 40) and 2-h delayed (n = 30) administration of the 48-h survival rate of the CLP mice (∗∗∗p < 0.001 vs. CLP, ##p < 0.01 vs. CLP + Ce-BE).

(D) The 7-day survival of the CLP mice (∗∗∗p < 0.001 vs. CLP, ###p < 0.01 vs. CLP + Ce-BE, n = 20).

(E) Levels of WBC, Lym, and PLT in CLP mice at 24 h after receiving different treatments (n = 5).

(F and G) H&E staining results of liver tissues in CLP mice and liver index after different treatments (scale bars, 50 μm [left], and 10 μm [right], n = 5).

(H) Plasma levels of ALT and AST activities (n = 5).

(I) Immunofluorescence staining of CD86 and CD206 in liver tissue (scale bars, 10 μm, n = 3).

(J) Fluorescence intensity of CD86 (n = 3).

(K and L) Levels of IL-1β, IL-6, TNF-α, and MPO activity measured in the plasma (n = 5).

(M) Schematic diagram depicting the establishment and study schedule of the “two-hit” sepsis mice model.

(N) The 10-day survival rate of the “two-hit” sepsis mice model (∗∗p < 0.01 vs. CLP/PA., ###p < 0.001 vs. CLP/PA. + Ce-BE, n = 30).

(O) Representative H&E staining of liver tissues (scale bars, 10 μm, n = 3).

(P and Q) Plasma levels of ALT and LDH activities (n = 5).

(R) Plasma levels of IL-1β, IL-6, and TNF-α (n = 5).

Data are represented as means ± SD. ∗p < 0.05, ∗∗/##p < 0.01, ∗∗∗/###p < 0.001 or the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. ns, not significant.

We first observed 48-h survival in septic mice following a single intraperitoneal injection of Ce-BE NZs (10 mg/kg). The results showed that compared with the CLP model group, Ce-BE NZs significantly prolonged survival time of septic mice and notably increased the 48-h survival rate by 57.5% (Figure 4B). Whether administered immediately or after a 2-h delay, BE (10 mg/kg) did not significantly improve survival, while the survival rate of Ce-BE NZs group was 36.7% higher than that of the BE group (Figure 4C). As reducing mid- and long-term mortality is crucial in sepsis treatment, we further extended the observation period and found that all the sepsis models and BE-treated mice died by days 2 and 3, respectively, whereas 40% of Ce-BE NZs-treated mice survived at 3 days and 30% survived at 7 days (Figure 4D). Microcirculatory blood flow, a clinically validated predictor of multiple organ dysfunction syndrome severity and sepsis-related mortality,³⁴ was assessed using laser Doppler flowmetry. Sepsis markedly reduced the microcirculatory blood flow in the hind limbs of mice, which was largely reversed by Ce-BE NZs and not improved by BE treatment (Figures S3A–S3D). In addition, the hematological parameter results showed that compared with the CLP and BE groups, Ce-BE NZs effectively maintained the number of white blood cells (WBCs) and lymphocytes (Lym), increased the proportion of Lym, and increased the number of platelets (PLT) in septic mice (Figure 4E), indicating that Ce-BE NZs were beneficial for restoring the immune defense system during sepsis.

Considering the liver’s critical role during sepsis, it was chosen for the evaluation of Ce-BE NZs’ efficacy against organ damage. Plasma biochemistry, liver index, neutrophil infiltration, and H&E staining were performed to assess the liver condition.³⁵ Specifically, H&E staining of the hepatic tissue sections revealed notable inflammatory cell infiltration, death, and tissue edema, and there was no significant improvement in the BE group, whereas Ce-BE NZs significantly ameliorated these anomalies (Figures 4F, 4G, and S3E). Consistent with the above results, compared with the CLP and BE groups, Ce-BE NZ treatment significantly reduced the plasma levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) (Figure 4H).

Given the immunomodulatory effects of Ce-BE NZs in vitro, the inflammatory state of the liver was assessed by measuring macrophage polarization and cytokine secretion. To determine whether Ce-BE NZs could alter the M1-like (pro-inflammatory subtype) or M2-like (anti-inflammatory phenotype) polarization of macrophages, we used immunofluorescence to detect the macrophage markers in the liver, including CD86 (M1) and CD206 (M2). As illustrated in Figure 4I, the infection could shift macrophages to the predominant M1 phenotype. We found that Ce-BE NZs could significantly decrease the number of proinflammatory (M1-like) cells and increase the number of anti-inflammatory (M2-like) cells in the liver (Figures 4J and S3F). Consistently, Ce-BE NZs also significantly reduce sepsis-induced elevation of hepatic inflammatory cytokines including IL-1β, IL-6, TNF-α, and myeloperoxidase (MPO) activity (Figures 4K and 4L). Notably, no statistically significant differences were observed between the BE and CLP groups. These results suggested that Ce-BE NZs can significantly reduce the hepatic inflammatory response and liver injury, improve the microcirculatory condition, restore the immune system, and ultimately improve survival in CLP-induced sepsis. Moreover, we also found that Ce-BE NZs could reduce the neutrophil infiltration, promote M1- to M2-like macrophage polarization, and ultimately significantly ameliorate the pathological damage of lungs compared with model mice (Figure S4).

It is well established that patients with severe sepsis could rapidly progress to a prolonged immunosuppressive state, which significantly increases the risk of secondary infections, organ damage, and even mortality.³⁶ To further demonstrate the clinical translational potential of Ce-BE NZs, we established an additional clinically relevant “two-hit” sepsis model³⁷ through intratracheal inoculation of PA. after CLP (Figure 4M), subsequently evaluating the therapeutic potential of Ce-BE NZs. Post-CLP administration of Ce-BE NZs enhanced the 10-day survival rate of “two-hit” sepsis model by 30% compared to septic controls (Figure 4N), and 35% compared to administration of BE. Histopathological analysis revealed that Ce-BE NZs significantly attenuated CLP/PA.-induced hepatic lobule destruction (Figure 4O), which was corroborated by significant reductions in plasma ALT and LDH activities (Figures 4P and 4Q). Notably, compared with healthy controls, CLP-induced sepsis model and BE-treated mice exhibited markedly reduced plasma inflammatory factors (IL-1β, IL-6, and TNF-α) post-secondary PA. infection. However, Ce-BE NZ administration reversed these cytokine suppressions (Figure 4R), demonstrating its ability to alleviate CLP-initiated immunosuppression and exert balanced anti-infective effects during secondary bacterial challenge.

Collectively, our findings demonstrate that Ce-BE NZs exhibit therapeutic efficacy in both CLP and “two-hit” sepsis models, offering a nanotechnology-based solution to clinical management of heterogeneous sepsis trajectories.

Ce-BE NZs inhibited hepatic mitochondrial dysfunction in septic mice

Mitochondria have distinct functions in cellular homeostasis, including energy synthesis, tissue re-repair, and ROS regulation. Mitochondrial dysfunction participates in the pathophysiology of sepsis and sepsis-related organ damage and is closely associated with patient outcomes.^38,39 To verify whether Ce-BE NZs improve mitochondrial dysfunction, and thus sepsis and sepsis-induced liver injury, we further evaluated the effects of Ce-BE NZs on mitochondrial function, including mitochondrial morphology, the activities of key mitochondrial enzymes, and energy metabolism (Figure 5A).

Figure 5.

Therapeutic efficacy of Ce-BE NZs on hepatic mitochondria and mitochondrial proteomics of the liver

(A) Experimental design for the study of purified hepatic mitochondria.

(B) Ultrastructural changes of mitochondria by TEM. Red arrows point to the prominent cristae of mitochondria, while the yellow arrows point to the broken membranes (scale bars, 2 μm [left] and 500 nm [right], n = 5).

(C and D) Activities of Na⁺-K⁺ and Ca²⁺-Mg²⁺ ATPase (n = 5).

(E) The line plot of mitochondrial OCRs (n = 3).

(F and G) The OCRs for basal respiration and ATP-linked respiration (n = 3).

(H) Heatmap plots indicated the differentially expressed proteins between mitochondria from CLP and Ce-BE NZs treatment groups (n = 3).

(I) Sankey bubble diagram of KEGG enrichment analysis.

(J) Chord diagram of GO enrichment analysis.

(K) GSEA graph of the iron ion binding pathways.

(L) Cluster heatmap analysis indicated the differentially expressed proteins in the iron ion binding pathway. Data are represented as means ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 or the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test.

The mitochondrial membrane in hepatocytes from septic mice was significantly destroyed, downsized, and the cristae were broken compared with that in healthy mice. However, the Ce-BE NZs treatment significantly ameliorated this damage and prodigiously increased the number of mitochondria relative to the CLP model group, indicating that Ce-BE NZs could effectively alleviate mitochondrial dysfunction in septic mice (Figures 5B, and S5A and S5B). Additionally, BE did not significantly improve sepsis-induced alterations in mitochondrial morphology.

Na⁺-K⁺ and Ca²⁺-Mg²⁺ ATPases intricately regulate the ionic concentration gradients and simultaneously preserve the integrity and functionality of mitochondrial structures.⁴⁰ Thus, we extracted mitochondria from the livers of mice and measured the activities of Na⁺-K⁺ and Ca²⁺-Mg²⁺ ATPases. Our results show that Ce-BE NZs significantly retained the sepsis-induced decrease in activities of mitochondrial Na⁺-K⁺ and Ca²⁺-Mg²⁺ ATPases compared with the BE group (Figures 5C and 5D). Moreover, we measured the effect of Ce-BE NZs on mitochondrial energy metabolism and found that the mitochondria from the CLP and BE groups maintained reduced real-time oxygen consumption rates (OCRs), while Ce-BE NZs significantly upregulated basal respiration and ATP-linked respiration, respectively (Figures 5E–5G). Overall, our results reveal that Ce-BE NZs could rescue sepsis-induced hepatic mitochondrial dysfunction, which may be an important reason for the improvement in sepsis and sepsis-induced liver injury.

To further explore the specific mechanism underlying the improvement effect of Ce-BE NZs on sepsis, we performed a proteomic analysis of hepatic mitochondria collected from mice in the CLP and Ce-BE NZ groups. Principal-component analysis of CLP vs. Ce-BE NZ treatment revealed a separation between the two groups (Figure S5C). Cluster heatmaps can be used to observe sample quality, and clustering was performed according to protein expression levels, with each row indicating the expression level of each protein in different groups (Figure 5H). We observed that 88 proteins were upregulated and 149 proteins were downregulated, as shown in the bar graph and volcano plot (Figures S5D and S5E). Next, the Sankey bubble diagram of Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis and the chord diagram of Gene Ontology (GO) enrichment analysis of the differentially expressed proteins were performed. We found that Ce-BE NZs greatly enhanced pathways, including iron ion binding, glutathione peroxidase activity, and fatty acid oxidation, compared with the CLP model group (Figures 5I and 5J). Considering that enhanced iron ion binding and glutathione peroxidase activity in mitochondria are closely associated with ferroptosis inhibition, we speculated that Ce-BE NZs may improve mitochondrial dysfunction and inhibit ferroptosis, ultimately exerting a therapeutic effect on sepsis and sepsis-induced liver injury.

Subsequently, we conducted gene set enrichment analysis (GSEA) to further ascertain the major differentially expressed proteins in the iron ion binding pathways. The results of the enrichment score (ES) distribution map and cluster heatmap analysis showed that the expression of FTH1, an important active part of ferritin⁴¹ involved in promoting free iron binding and maintaining iron ion levels on the cellular level, was significantly upregulated by Ce-BE NZ treatment (Figures 5K and 5L). In summary, mitochondrial proteomic analysis revealed that Ce-BE NZs may potentially inhibit ferroptosis and ultimately exert a therapeutic effect on sepsis and sepsis-induced liver injury.

Ferroptosis inhibition mediates the therapeutic effect of Ce-BE NZs for sepsis and sepsis-induced liver injury

Ferroptosis, an iron-mediated form of regulated cell death characterized by lipid peroxidation, represents a pathway distinct from apoptosis, necrosis, and autophagy⁴² and is increasingly recognized as a critical factor in the pathogenesis of liver disorders.^43,44 The pathophysiological cascade of ferroptosis is initiated by iron overload and ROS production, leading to an escalation of polyunsaturated fatty acids within phospholipids and culminating in cell membrane rupture due to the accumulation of phospholipid hydroperoxides. To delineate the ferroptosis process, we quantified markers of lipid peroxidation, ROS levels, and free iron content. Notably, glutathione peroxidase 4 (GPX4), a key regulator in ferroptosis, was also measured.^45,46

Thus, to clarify the effect of Ce-BE NZs on ferroptosis, we first assessed ROS levels using dihydroethidium probes and the levels of the end products of lipid peroxidation, including 4-hydroxynonenal (4-HNE) and malondialdehyde (MDA). The immunofluorescence results showed that mice treated with Ce-BE NZs had lower levels of ROS and 4-HNE in the liver than those in the CLP and BE groups (Figures 6A and S6A and S6B). Consistently, hepatic MDA content was the lowest in the Ce-BE NZs group among septic mice (Figure 6B). Moreover, a Prussian blue-DAB staining showed the presence of iron accumulation in the CLP model group, which was remarkably reduced by Ce-BE NZs (Figure 6A). Similarly, Ce-BE NZs significantly reduced the sepsis-induced increase in the content of hepatic ferrous and ferric ions compared with BE (Figures 6C and S6C). In addition, the hepatic expression levels of GPX4 and FTH1 in septic mice were markedly decreased compared with those in healthy mice, whereas Ce-BE NZs largely reversed the sepsis-induced decrease in the expression of GPX4 and FTH1 (Figure 6A). Meanwhile, the hepatic GSH levels were significantly higher in the Ce-BE NZs group compared with the CLP model group (Figure 6D). Additionally, BE-Ce NZs can also reduce free iron deposition and upregulate the expression levels of GPX4/FTH1 in the liver of “two-hit” sepsis mice (Figures S6D and S6E).

Figure 6.

Ferroptosis inhibition mediates the therapeutic effect of Ce-BE NZs

(A) Levels of ROS (scale bars, 10 μm), 4-HNE (scale bars, 20 μm), Prussian Blue-DAB (scale bars, 20 μm), GPX4 (scale bars, 10 μm), and FTH1 (scale bars, 20 μm) expression. White circles point to the iron deposition, yellow circles point to the GPX4 expression, and orange circles point to the FTH1 expression (n = 3).

(B–D) The levels of MDA, Fe²⁺ ions, and GSH in tissue homogenates (n = 3).

(E) The 48-h survival rate (∗∗∗p < 0.001 vs. CLP, ###p < 0.001 vs. CLP + Ce-BE, n = 20).

(F–H) Levels of ALT and LDH activities in plasma and GSH levels in liver tissues (n = 3).

(I) H&E staining. Blue arrow points to the inflammatory cell infiltration, red arrows point to cell necrosis, and yellow arrows point to disordered hepatic lobules (scale bar, 20 μm, n = 3).

(J) CLSM images of GPX4/FTH1 expression (scale bar, 10 μm, n = 3).

Data are represented as means ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗/###p < 0.001 or the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. ns, not significant.

To further explore the role of ferroptosis in the therapeutic effect of Ce-BE NZs on sepsis-induced liver injury, we measured the effect of FINO2⁴⁷ (a canonical ferroptosis inducer) pretreatment on the therapeutic effect of Ce-BE NZs. Our results demonstrated that Ce-BE NZs significantly improved the 48-h survival rate of mice with sepsis by 45% and 40% relative to the CLP and BE groups, respectively (Figure 6E). In contrast, FINO2 pretreatment completely counteracted the therapeutic benefits of Ce-BE NZs. Consistently, the downward trend in plasma biomarkers of liver injury, ALT and LDH, was completely counteracted and unchanged in FINO2-pretreated mice (Figures 6F and 6G). The results of the H&E staining were consistent with the above results (Figure 6I). In addition, FINO2 pretreatment blocks the upregulation of GSH, GPX4, and FTH1 expression by Ce-BE NZs (Figures 6H and J). In summary, our findings indicate the ability of Ce-BE NZs to ameliorate ferroptosis by promoting mitochondrial FTH1 expression, reducing the levels of iron ions and ROS, and inducing the expression of GPX4, which in turn reduces liver injury and ultimately improves sepsis-related mortality.

Ce-BE NZs mitigated cell damage and ferroptosis in the LPS-induced macrophage-incorporating human liver-derived organoid

Considering organoid technology has received official Food and Drug Administration (FDA) approval for incorporation into pharmaceutical development pipelines and preclinical research applications, we used a liver organoid model, originating from the human healthy liver tissue^48,49 to investigate the therapeutic efficacy of Ce-BE NZs in the context of sepsis-associated liver damage. When cultivated in Matrigel, these organoids initially exhibited a hollow cystic morphology, characterized by the expression of ductal markers (EPCAM and SOX9), a hepatic progenitor marker (HNF-4α), and a proliferation marker (KI67). Upon maturation in a specialized differentiation medium for a period of 10–12 days, the organoids transitioned to a solid spherical form, concurrently augmenting the expression of mature hepatocyte markers, including ALB, P120, and F-actin (Figure S7). To emulate the complex in vivo immune microenvironment during sepsis, we implemented a co-culture system integrating mature liver organoids with macrophages differentiated from human embryonic stem cells (hESCs)^50,51,52,53 (Figures 7A and 7B). Using our established protocol for the directed differentiation of hESCs into hematopoietic progenitor cells, we fine-tuned the cytokine cocktail to efficiently generate macrophages from hESCs⁵⁴ (Figure S8). Following 2 days of co-culture, we introduced LPS, either in the presence or absence of BE or Ce-BE NZs, to assess the response of organoids to oxidative injury and markers indicative of ferroptosis.

Figure 7.

Exploring the protective effects of Ce-BE NZs on LPS-induced hepatocyte injury via a macrophage-incorporating human liver-derived organoid model

(A) Illustrative representation of the experimental setup. (MØ: macrophages).

(B) Co-culture morphology showcasing human primary liver tissue-derived organoids alongside macrophages originating from hESCs (scale bars, 200 μm, n = 3).

(C) Analysis of intracellular ROS using CellROX staining (scale bars, 200 μm, n = 3).

(D) Assessment of intracellular iron levels through iron orange staining (scale bars, 200 μm, n = 3).

(E) Immunofluorescence staining for GPX4 expression (scale bars, 20 μm, n = 3).

(F) Immunofluorescence staining for FTH1 expression (scale bars, 10 μm, n = 3).

(G) Evaluation of cell viability through live cell staining (scale bars, 200 μm, n = 3).

(H) Immunofluorescence staining for ALB expression (scale bars, 20 μm, n = 3).

(I–K) Levels of IL-1β, IL-6, and TNF-α in the culture supernatant (n = 5).

(L and M) Levels of ALT and LDH activities in culture supernatant (n = 4).

Data are represented as means ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 or the indicated comparisons. Differences between groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. ns, not significant.

Immunofluorescence staining revealed that Ce-BE NZs significantly mitigated ROS within the organoids, enhancing their viability compared to the BE-treated or control group (Figure 7C). Iron orange staining corroborated these findings, demonstrating a marked reduction in staining intensity in Ce-BE NZ-treated organoids (Figures 7D and S9A). Moreover, the upregulation of the ferroptosis-regulatory proteins GPX4 and FTH1, as indicated by immunofluorescence, suggests that Ce-BE NZs play a role in preventing ferroptosis (Figures 7E and 7F and S9B and S9C). Additionally, Ce-BE NZs counteracted the LPS-induced inflammatory response functional impairment in hepatocytes and stimulated ALB synthesis (Figures 7G–7M and S9D). Collectively, these findings underscore the capacity of Ce-BE NZs to suppress LPS-induced ROS production and ferroptosis in liver organoids, thereby potentiating the recovery of hepatocyte function during liver injury. This study highlights the immense potential of Ce-BE NZs in the clinical treatment of sepsis, owing to their demonstrated efficacy in a highly relevant pathological context.

In vivo biocompatibility and metabolism

Finally, the potential in vivo toxicity of Ce-BE NZs was evaluated. First, H&E staining revealed almost no toxicity of Ce-BE NZs to the heart, liver, spleen, lung, and kidney injected with Ce-BE NZs (10 mg/kg) at 1 and 7 days post-injection, which was not significantly different from the control group (Figure S10A). Next, the hematological parameters and plasma biochemistry data for mice treated with Ce-BE NZs were within the normal range. The results indicated no significant toxicity of Ce-BE NZs in the main organs of the mice (Figure S10B). Furthermore, a hemolytic test demonstrated that the hemolysis rate of red blood cells in mice remained lower than 2% with Ce-BE NZs ranging from 50 to 1,000 μg/mL (Figures S10C and S10D). Therefore, Ce-BE NZs demonstrate sufficient biocompatibility for application in sepsis treatment.

Furthermore, we evaluated the metabolism of Ce-BE NZs in vivo. The Ce ion content in the mouse blood and the main tissues was determined using inductively coupled plasma mass spectrometry. The results show that the content of Ce ions in the blood reached a peak in 6 h after the injection of Ce-BE NZs, was preferentially enriched in the liver of the mice at 6 h, and was enriched in the kidney in 12 h (Figure S10E and S10F). Furthermore, the Ce ion content in mouse feces and urine demonstrated that Ce-BE NZs were primarily excreted via urine within 48 h post-injection (Figure S10G).

Discussion

Sepsis is a complex clinical syndrome characterized by a dysregulated immune response to infection, leading to life-threatening multiple organ dysfunction. Bacterial infection, immune dysregulation, and organ damage represent three interconnected pathogenic factors that collectively determine clinical outcomes in sepsis.⁵⁵ Among these, bacterial infection serves as the primary trigger. Upon bacterial invasion, the immune system is triggered to eliminate the pathogens. In sepsis, however, this response often becomes exaggerated and dysregulated, culminating in dysfunction of multiple organs—including the liver and lungs—and ultimately contributing to patient mortality.⁵⁶ Thus, early intervention with effective measures to control bacterial infection and reduce the bacterial load forms the cornerstone of sepsis management. Building on this foundation, immunomodulatory strategies aimed at restoring immune homeostasis and mitigating organ injury may further improve overall prognosis. Our study demonstrates that cerium-BE coordination markedly enhances the solubility and bioavailability of BE, thereby unlocking its antibacterial potential. Concurrently, the Ce-BE NZs effectively restored immune balance, attenuated damage in organs such as the liver and lungs, and ultimately improved survival in both CLP and “two-hit” sepsis models. Our work not only establishes an effective strategy for functionalizing BE but also provides experimental evidence supporting the feasibility of targeting the infection-immune dysregulation-organ damage axis in sepsis treatment.

Ferroptosis, a form of programmed cell death driven by iron-dependent lipid peroxidation, is primarily characterized by iron accumulation and oxidative lipid damage.^57,58 Notably, growing evidence underscores the essential roles of mitochondria in both regulating and executing ferroptosis.^59,60 Beyond their classical role as cellular powerhouses, mitochondria are actively involved in critical processes including ROS generation, energy metabolism, and iron homeostasis.⁶¹ In particular, previous studies have established their involvement in iron-sulfur cluster assembly, heme synthesis, and iron storage.^61,62 Owing to this central role in iron metabolism, mitochondrial dysfunction can disrupt intracellular iron homeostasis, leading to excessive iron accumulation, which in turn promotes ferroptosis. Thus, a tight pathophysiological link exists between mitochondrial impairment and the induction of ferroptosis. In this study, we found that Ce-BE NZs effectively ameliorated mitochondrial dysfunction, upregulated the expression of FTH1 in mitochondria, and suppressed ferroptosis, ultimately attenuating sepsis-associated liver injury. Our findings provide further evidence supporting the interaction between mitochondrial dysfunction and ferroptosis. Additionally, based on the detection results of hepatic ferroptosis, we believe that ferroptosis and mitochondrial dysfunction in hepatocytes play critical roles in the therapeutic effects of Ce-BE NZs.

The widespread use of antibiotics has significantly improved the clinical management of bacterial infections. However, misuse and overuse of these drugs have accelerated the emergence and dissemination of multidrug-resistant bacteria, rendering antimicrobial resistance a pressing global health issue.⁶³ This trend is associated with increased incidence and mortality of bacterial infections, as well as elevated healthcare costs.⁶⁴ As a result, the development of promising non-antibiotic antimicrobial agents is urgently needed. Natural products derived from medicinal plants have long been valuable sources of therapeutic agents against infectious diseases, owing to their diverse biological activities and broad-spectrum antibacterial properties.^65,66 Among them, BE is a natural flavonoid isolated from the roots of Scutellaria baicalensis, which is a key active ingredient in several commercially available formulations widely used in clinical practice for their anti-infective and anti-inflammatory effects.⁶⁷ This well-documented clinical application not only attests to BEʼs favorable safety profile but also supports its potential for antimicrobial and immunomodulatory uses. Nevertheless, the therapeutic potential of BE is limited by its poor aqueous solubility and low permeability, which leads to inadequate systemic absorption and low bioavailability.⁶⁸ Therefore, in light of both the promise and the constraints associated with BE, this study focuses on this compound with the aim of advancing its clinical application and exploring potential therapeutic strategies for sepsis.

Human-derived organoid models are gaining increased attention as independent non-clinical evaluation systems for studying drug responses in human tissues, particularly in the fields of genetic disorders and critical illnesses.^69,70 They are emerging as alternatives to animal models in drug development and safety assessments,⁷¹ a shift further reinforced by recent FDA initiatives encouraging the use of human-relevant models like organoids for non-clinical evaluations.⁷² Building on the results of mouse models, we established a human liver-derived organoid model to evaluate the therapeutic potential of Ce-BE NZs in sepsis and sepsis-induced liver injury. The hepatic immune microenvironment plays a critical role in sepsis-associated liver injury, characterized by the rapid response of Kupffer cells and circulating immune cells to bacterial stimulation, followed by the production of pro-inflammatory cytokines such as IL-6, which exacerbate hepatocyte damage.⁵⁰ To more accurately mimic the interaction between hepatocytes and immune cells under septic conditions, we employed a co-culture system incorporating stem cell-derived macrophages with liver organoids. Experimental results demonstrated that Ce-BE NZs significantly alleviated hepatocyte injury through effective modulation of the inflammatory response, further validating their therapeutic potential in clinical sepsis treatment.

In conclusion, we successfully developed multifunctional hybrid Ce-BE NZs through metal-polyphenol coordination, which exhibited robust antioxidant and antibiotic-free antibacterial properties. Ce-BE NZs demonstrated stable ROS-scavenging and SOD- and CAT-mimicking abilities. In vitro and in vivo experiments showed that Ce-BE NZs can effectively promote the conversion of macrophages to an anti-inflammatory phenotype. Furthermore, mechanistic studies revealed that Ce-BE NZs can effectively inhibit mitochondrial dysfunction and ferroptosis to reduce liver damage and improve the survival and overall therapeutic effect in septic mice. Moreover, macrophage-incorporated human liver-derived organoid experiments further confirmed the therapeutic potential of Ce-BE NZs in sepsis. Overall, the synthesized hybrid Ce-BE NZs with multifunctional properties exhibited favorable biocompatibility and therapeutic effects in septic mice, making them promising targeted drugs for treating sepsis and other infectious diseases and facilitating their future clinical applications.

Limitations of the study

Although this study demonstrates that Ce-BE NZs effectively ameliorate hepatic mitochondrial dysfunction, suppress ferroptosis, and improve survival in sepsis-induced liver injury, it remains unclear whether the mitigation of ferroptosis is directly mediated through the improvement of mitochondrial function. While existing evidence supports a close association between mitochondrial dysfunction and ferroptosis, the precise mechanistic role of Ce-BE NZs in this pathway has not been fully elucidated. Future studies will be essential to determine whether Ce-BE NZs attenuate ferroptosis primarily via modulating mitochondrial function.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Gan Chen (chenlzu2005@163.com).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •The hepatic mitochondrial proteomics data have been deposited at the Mendeley Data (DOI: https://doi.org/10.17632/7dg6hrh8mv.1).
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (82572538).

Author contributions

K.Z., B.Z., and Y.L. performed most of the experiments. K.Z. and B.Z. analyzed the data and wrote the paper. S.Z., X.T., X.Z., and W.C. assisted with animal and cellular experiments. Y.P., H.T., Z.Y., Q.W., and S.M. assisted with analysis experiments and data analysis. L.Z., Y.Y., H.Z., and G.C. designed the experiments and supervised the project. All authors discussed the results and commented on the paper and approved of the final version of the paper.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-COX2 Antibody	Abcam	Cat# ab179800; RRID: AB_2894871
Anti-4 Hydroxynonenal Antibody	Abcam	Cat# ab48506; RRID: AB_867452
Anti-alpha 1 Fetoprotein Antibody	Abcam	Cat# ab54745; RRID: AB_2223910
Anti-F-actin Antibody	Abcam	Cat# ab205; RRID: AB_302794
iNOS Polyclonal Antibody	proteintech	Cat# 22226-1-AP; RRID: AB_2879038
p120 Catenin Polyclonal Antibody	proteintech	Cat# 12180-1-AP; RRID: AB_2086267
Anti-Glutathione Peroxidase 4 Mouse mAb	Servicebio	Cat# GB154327; RRID: AB_3714601
Anti-Ferritin Heavy Chain Rabbit pAb	Servicebio	Cat# GB115535; RRID: AB_3717704
Anti-Myeloperoxidase Mouse mAb	Servicebio	Cat# GB120016; RRID: AB_3717705
Anti-EpCAM Rabbit pAb	Servicebio	Cat# GB11274; RRID: AB_2941859
Anti-Ki67 Rabbit pAb	Servicebio	Cat# GB111141; RRID: AB_3096315
Anti-HNF-4-alpha Rabbit pAb	Servicebio	Cat# GB115549; RRID: AB_3717706
Anti-Albumin Mouse mAb	Servicebio	Cat# GB14005; RRID: AB_3717707
Anti-Cytochrome P450 3A4/CYP3A4 Rabbit pAb	Servicebio	Cat# GB112021; RRID: AB_3717708
SOX9 Rabbit pAb	ABclonal	Cat# A2479; RRID: AB_2764375
IL6 Rabbit pAb	ABclonal	Cat# A0286; RRID: AB_2757098
Human CD14 Microbeads	Miltenyi	Cat# 130-050-201; RRID: AB_2665482
BD Horizon BV786 Mouse Anti-Human CD45	BD Biosciences	Cat# 563716; RRID: AB_2716864
BD Pharmingen APC-Cy7 Mouse Anti-Human CD11b	BD Biosciences	Cat# 560914; RRID: AB_2033935
BD Pharmingen APC Mouse Anti-Human CD14	BD Biosciences	Cat# 555399; RRID: AB_398596
Bacterial and virus strains
Staphylococcus aureus	ATCC	Cat# 25923
Escherichia coli	ATCC	Cat# 25922
Pseudomonas aeruginosa	ATCC	Cat# 27853
Chemicals, peptides, and recombinant proteins
Baicalein	Macklin	Cat# B802462
Cefuroxime Sodium Salt	Macklin	Cat# C804336
Gentamycin Sulfate	Macklin	Cat# G810322
Dimethyl Sulfoxide	Macklin	Cat# D6258
Vitamin C	Macklin	Cat# S799300
Cerous Chloride	Aladdin	Cat# C102765
Methylene Blue	Aladdin	Cat# M196500
Poly(vinylpyrrolidone)	Sigma-Aldrich	Cat# 81420
Methanol Solution	Sinopharm	Cat# 40064260
Ferrous Sulfate	Sinopharm	Cat# 100121006
Sodium Shloride	Sinopharm	Cat# 69900795
RIPA Lysis Buffer	Beyotime	Cat# P0038
PMSF	Beyotime	Cat# ST505
DMPO	MCE	Cat# HY-107690
TEMPO	MCE	Cat# HY-W001187
Pentetic Acid	MCE	Cat# HY-B1335
BMPO	MCE	Cat# HY-121137
Xanthine Oxidase	MCE	Cat# HY-P2755
Hypoxanthine	MCE	Cat# HY-N0091
CHIR99021	MCE	Cat# HY-10182
SB431542	MCE	Cat# HY-10431
Monochlorobimane	MCE	Cat# HY-101899
Y-27632	MCE	Cat# HY-10071
FINO2	Selleck	Cat# E1244
Dihydroethidium	Servicebio	Cat# G1904-100T
PBS Buffer	Servicebio	Cat# G4202
mTeSR1 Complete Kit	STEMCELL	Cat# 85850
HepatiCult Organoid Growth Medium (Human)	STEMCELL	Cat# 100-0385
HepatiCult Organoid Differentiation Medium (Human)	STEMCELL	Cat# 100-0383
DMEM	Gibco	Cat# 11965092
FBS	Excell	Cat# FSP500
rhSCF	R&D systems	Cat# BT-SCF
rhTPO	R&D systems	Cat# BT-TPO
rhIL-3	R&D systems	Cat# 203-IL
rhFLT3L	R&D systems	Cat# 308-FKHB
rhGM-CSF	R&D systems	Cat# 7954-GM
rhBMP4	R&D systems	Cat# 314-BP
rh/m/r Activin A	R&D systems	Cat# 338-AC
rhbFGF	R&D systems	Cat# BT-FGFB
rhVEGF	R&D systems	Cat# 293-VE
rhM-CSF	novoprotein	Cat# C417
Critical commercial assays
Live/Dead Cell Viability Assay Kit	Abcam	Cat# ab287858
DCFDA/H2DCFDA-Cellular ROS Assay Kit	Abcam	Cat# ab113851
ABTS Free Radical Scavenging Capacity Assay Kit	Solarbio	Cat# BC4770
DPPH Free Radical Scavenging Capacity Assay Kit	Solarbio	Cat# BC4750
Live/Dead Bacteria Staining Kit	Solarbio	Cat# EX3000
Mitochondrial Extraction kit	Solarbio	Cat# SM0020
Total Superoxide Dismutase (T-SOD) Asay Kit	Nanjing Jiancheng	Cat# A001-1-2
Reduced Gutathione (GSH) Asay Kit	Nanjing Jiancheng	Cat# A006-1-1
Tissue Iron Assay Kit	Nanjing Jiancheng	Cat# A039-2-1
Mouse IL-1β ELISA Kit	Elabscience	Cat# E-EL-M0037
Human IL-1β ELISA Kit	Elabscience	Cat# E-EL-H0149
Mouse IL-6 ELISA Kit	Elabscience	Cat# E-EL-M0044
Human IL-6 ELISA Kit	Elabscience	Cat# E-EL-H6156
Mouse TNF-α ELISA Kit	Elabscience	Cat# E-EL-M3063
Human TNF-α ELISA Kit	Elabscience	Cat# E-EL-H0109
Alanine Aminotransferase (ALT/GPT) Activity Assay Kit	Elabscience	Cat# E-BC-K235-M
Lactate Dehydrogenase (LDH) Activity Assay Kit	Elabscience	Cat# E-BC-K046-M
Ferrous Iron Colorimetric Assay Kit	Elabscience	Cat# E-BC-K773-M
BCA Protein Colorimetric Assay Kit	Elabscience	Cat# E-BC-K318-M
Seahorse XF Cell Mito Stress Test Kit	Agilent	Cat# 103016-100
Seahorse XF Palm OX Stress Test Kit	Agilent	Cat# 103693-100
RNeasy Mini Kit	QIAGEN	Cat# 74104
ReverTra Ace™ qPCR RT Kit	TOYOBO	Cat# FSQ-101
SYBR® Green Real-time PCR Master Mix	TOYOBO	Cat# QPK-201
Deposited data
Hepatic mitochondrial proteomics data	This paper; Mendeley Data	https://doi.org/10.17632/7dg6hrh8mv.1
Experimental models: Cell lines
Human ESC Line WA09	Wicell	Cat# wb67844
RAW 264.7 Cells	Maintained in this laboratory	N/A
Human Liver Organoid	Maintained in this laboratory	N/A
Experimental models: Organisms/strains
Mouse: BALB/c	Charles River	Cat# 028
Oligonucleotides
Il1b, forward primer: GAAATGCCACCTTTTGACAGTG reverse primer: TGGATGCTCTCATCAGGACAG	This paper	N/A
Il6, forward primer: CTGCAAGAGACTTCCATCCAG reverse primer: AGTGGTATAGACAGGTCTGTTGG	This paper	N/A
Software and algorithms
GraphPad Prism (version 9.5.0)	GraphPad	https://www.graphpadprism.wxndl.cn/
ImageJ	NIH	https://imagej.nih.gov/ij/

Experimental model and study participant details

Mouse models

Male specific pathogen-free (SPF) BALB/c mice (19–20 g, 6–8 weeks old; quality certificate No.: SCXK (Jing) 2021-0006) were purchased from Charles River Experimental Animal Technology Co., Ltd. (Beijing, China). All animal experiments were performed in compliance with the Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of the Academy of Military Medical Sciences (Approval No.: IACUC-DWZX-2024-P671).

Human liver-derived organoid

Human liver-derived organoid was maintained in 24-well plates with Organoid Growth Medium (OGM), and was routinely passaged following the protocols outlined in the reagent manual. The medium was then switched to Organoid Differentiation Medium (ODM) for a period of 10–12 days to facilitate maturation into hepatocytes. Organoids were confirmed to be uncontaminated through mycoplasma testing. The experiment of constructing liver organoids with liver tissues from healthy liver tissues has passed the ethical review and been approved by the Ethics Committee of the Chinese People’s Liberation Army (PLA) General Hospital (Approval No.: S2018-111-01).

Cell culture

RAW 264.7 cells were maintained in DMEM supplemented with 10% FBS. Upon reaching 70–80% confluence, the culture medium was aspirated, and the cells were gently washed with PBS. Cells were detached using a cell scraper, collected by centrifugation at 300 × g, and resuspended in fresh complete medium for passaging at a 1:5 ratio. RAW 264.7 cells were confirmed to be uncontaminated through mycoplasma testing.

Human embryonic stem cells (hESC-WA09) were obtained from WiCell Research Institute and cultured under feeder-free conditions. When cell colonies reached 80% confluency, they were passaged with TrypLE Select and seeded into Matrigel-coated 6-well plates at a ratio of 1:20. The hESC were subjected to karyotypic analysis and uncontamination were tested routinely.

Microbe strains

S. aureus (ATCC 25923), E. coli (ATCC 25922), and PA. (ATCC 27853) were obtained from the China General Microbial Strain Collection Management Center (Beijing, China). The bacterial strains used in the experiment were cultured in Luria-Bertani (LB) medium at 37°C for 10 h with shaking at 200 rpm. Bacterial concentrations were determined by measuring optical density (OD₆₀₀ = 0.1, and colony counting is 10⁶ CFU/mL), and the strains were stored at −80°C.

Method details

Synthesis of Ce-BE NZs

CeCl₃ (18.4 mg) was dissolved in methanol (1 mL) and added dropwise to a solution of PVP (66 mg) in methanol (5 mL). The mixture was stirred vigorously to obtain Solution A. Subsequently, baicalein (14 mg) was dissolved in methanol (1 mL) and sonicated for 5 min to obtain Solution B. Solution B was then added dropwise to Solution A. The reaction was stirred until the color changed from yellow to fuchsia, yielding Solution C. Finally, Solution C was transferred to a dialysis bag (MWCO: 3500 Da) and dialyzed against 0.9% NaCl solution overnight to obtain the aqueous solution of Ce-BE NZs.

Characterization

TEM images and elemental mapping were obtained using a JEOL JEM-F200 TEM (Tokyo, Japan) at an accelerating voltage of 200 kV. FT‒IR imaging was obtained with a Nicolet iS20 wavenumber (400-4000 cm⁻¹). The DLS and zeta potential measurements of Ce-BE NZs were performed using the Zetasizer Nano-ZS instrument (Malvern, Panalytical-Zetasizer, UK). The cerium content was measured by inductively coupled plasma-atomic emission spectrometry (ICP–MS, Agilent-7800, USA). The stabilities of Ce-BE NZs were measured by an automatic microplate reader (Omega, FLUO star, Germany).

DPPH radical scavenging capacity of Ce-BE NZs assay

The DPPH free radical scavenging activity was assessed spectrophotometrically using a commercial assay kit, following the manufacturer’s instructions. Briefly, Ce-BE NZs at final concentrations of 0, 3, 6, 12, 24, 48, and 96 μg/mL were mixed with the DPPH working solution and incubated in the dark for 30 min. The absorbance of each sample was measured at 515 nm, and the radical scavenging efficiency was calculated.

ABTS radical scavenging capacity of Ce-BE NZs assay

The ABTS free radical scavenging activity was evaluated spectrophotometrically using a commercial assay kit according to the manufacturer’s protocol. Briefly, Ce-BE NZs at final concentrations of 0, 3, 6, 12, 24, 48, and 96 μg/mL were mixed with the ABTS radical solution and incubated in the dark for 6 min. Absorbance was measured at 405 nm, and the radical scavenging efficiency was calculated.

Methylene blue assay

The hydroxyl radicals (⋅OH) generated from the Fenton reaction decolorize methylene blue (MB). Therefore, the change in the characteristic absorption peaks of MB (at 610 nm and 660 nm) can be used to evaluate the ⋅OH radical scavenging ability of Ce-BE NZs. In this assay, Ce-BE NZs at final concentrations of 0, 3, 6, 12, 24, 48, and 96 μg/mL were mixed with an MB solution. Subsequently, a Fenton reaction solution containing H₂O₂ and Fe₂SO₄ was added, and the mixture was incubated in the dark for 30 min. The characteristic peaks of MB decreased with increasing concentrations of Ce-BE NZs, demonstrating their scavenging capacity.

The ⋅OH scavenging capacity of Ce-BE NZs

The ⋅OH scavenging capacity of Ce-BE NZs was further confirmed using electron spin resonance (ESR) spectroscopy with DMPO as a spin trap. Briefly, 20 μL of PBS, 10 μL of 100 mM H₂O₂, and 5 μL of DMPO were mixed with 5 μL of 0.1 mM Ce-BE NZs in a centrifuge tube. Then, 10 μL of 10 mM Fe₂SO₄ was added and mixed thoroughly (total volume: 50 μL). The reaction mixture was aspirated into a 50 μL glass capillary tube (1 mm inner diameter), which was then sealed and inserted into the ESR cavity. The spectrum was recorded after a 5-min reaction period. The ESR spectrum of the DMPO/⋅OH spin adduct exhibited a characteristic quartet signal with a relative intensity of 1:2:2:1. A decrease in the signal intensity indicated a reduction in ⋅OH content in the reaction system, confirming the radical scavenging activity of Ce-BE NZs. The ESR instrument settings were as follows: microwave power, 2 mW; modulation amplitude, 1 G; scan range, 100 G; and receiver gain, 20 dB.

The ¹O₂ scavenging capacity of Ce-BE NZs

A mixture was prepared in a centrifuge tube containing 20 μL of PBS, 10 μL of 100 mM H₂O₂, 5 μL of TEMPO, and 5 μL of 0.1 mM Ce-BE NZs. Then, 10 μL of 10 mM Fe₂SO₄ was added and thoroughly mixed, resulting in a total volume of 50 μL. The mixture was aspirated into a 50 μL glass capillary tube (1 mm inner diameter), which was sealed and inserted into the ESR cavity. After 5 min of reaction, the spectrum was recorded. The ESR spectrum of the TEMPO/¹O₂ spin adduct displayed a characteristic triplet signal with a relative intensity of 1:1:1. A decrease in signal intensity indicated a reduction in ¹O₂ content within the reaction system. The ESR instrument parameters were consistent with those previously described.

SOD-mimicking activity of Ce-BE NZs

The superoxide dismutase (SOD) -mimicking activity was evaluated spectrophotometrically using a commercial SOD Activity Kit according to the manufacturer’s instructions. In this assay, superoxide anions (⋅${\mathrm{O}}_2^{-}$) generated by the xanthine/xanthine oxidase system reduce azotetrazolium to form blue formazan products, which exhibit maximum absorption at 550 nm. For ESR detection, a mixture was prepared containing 15 μL of H₂O, 10 μL of 5 mM hypoxanthine, 10 μL of 0.25 mM DTPA, 5 μL of 250 mM BMPO, and 5 μL of 0.1 mM Ce-BE NZs. Then, 5 μL of 0.4 U/mL Xanthine Oxidase (XOD) was added and thoroughly mixed (total volume: 50 μL). The mixture was transferred to a glass capillary tube, sealed, and inserted into the ESR cavity. The spectrum was recorded after 5 min of reaction. The BMPO/⋅OOH adduct spectrum exhibited a quartet signal (1:1:1:1 intensity ratio), and a decrease in signal intensity indicated a reduction in ⋅${\mathrm{O}}_2^{-}$ content. The ESR instrument settings were the same as above.

CAT-mimicking activity of Ce-BE NZs

For dissolved oxygen measurement: A working solution was prepared by adding 300 μL of 100 mM H₂O₂ to 5 mL of PBS. The reaction was initiated by adding 5 μL of 1 mM Ce-BE NZs to the solution under slow magnetic stirring. Dissolved oxygen levels were monitored in the dark using an oxygen electrode (OXY-1 ST Trace, Germany), with readings taken every 30 s for 15 min. For ESR detection: A solution containing 200 μL of 0.1 M H₂O₂ in 2.8 mL PBS (pH 7.4) was prepared, and the reaction was initiated by adding 20 μL of Ce-BE NZs. ESR scanning was performed over a range of 3000–4000 nm at 4-min intervals for 16 min. The broadening of spectral lines and decreased signal intensity over time indicated continuous oxygen generation.

Antibacterial activity of Ce-BE NZs

The antibacterial activity of Ce-BE NZs was assessed using the plate colony-counting method. Bacterial suspensions (10⁶ CFU/mL) were co-cultured with 0.5 mL of each test compound-Ce-BE NZs (1 mg/mL), Ce³⁺ ions (96 μg/mL), baicalein (BE, 1 mg/mL), or a positive-control antibiotic (1 mg/mL)-for 8 h. The suspensions were then diluted, plated, and incubated overnight at 37°C. For morphological analysis, co-cultured samples were centrifuged, lyophilized, and examined using scanning electron microscopy (Regulus-8100, Japan). Bacterial viability was also assessed using a Live/Dead Bacterial Staining Kit and observed under an inverted microscope (RVL-100, USA). To ensure comparability, the Ce³⁺ ion control was used at 96 μg/mL, equivalent to the cerium content in 1 mg/mL Ce-BE NZs.

Determination of intracellular baicalein levels in bacteria

Bacteria co-incubated will be and Ce-BE NZs were centrifuged, washed, and then resuspended in methanol to a final volume of 3 mL. The mixture was sonicated for 30 min to ensure complete dissolution. After cooling, the sample was homogenized and filtered through a 0.22 μm organic membrane filter before analysis using a Thermo Scientific UltiMate 3000 HPLC system (USA). The operating parameters were as follows: flow rate, 1.0 mL/min; column temperature, 30°C; run time, 25 min; detection wavelength, 276 nm; column, AQ-C18 (250 × 4.6 mm).

Hemolysis ratio of Ce-BE NZs

Fresh mouse blood was centrifuged at 400 × g for 8 min. The pellet was gently washed and resuspended in 0.9% NaCl to obtain an RBC suspension. The RBC suspension was mixed with equal volumes of Ce-BE NZs solutions at various concentrations (25, 50, 75, 100, 250, 500, 750, and 1000 μg/mL). Negative and positive controls were prepared by mixing the RBC suspension with equal volumes of 0.9% NaCl and deionized water, respectively. After incubation at 37°C for 3 h, images were captured. The samples were then centrifuged at 400 × g for 8 min to collect the supernatant. The absorbance of each supernatant was measured at 540 nm using a microplate reader.

Cell treatment

RAW264.7 cells were treated for 1 h with either 1 μg/mL or 10 μg/mL of Ce-BE NZs, BE, or an equal volume of PBS as a control. Subsequently, LPS was added to the culture medium at a final concentration of 1 μg/mL. After 24 h of incubation, the cells were harvested for further analysis. For experiments, cells were seeded in six-well plates at a density of 1 × 10⁶ cells per well. After 24 h, the medium was replaced with DMEM containing 2% FBS, and the cells were cultured for an additional 6 h.

Flow cytometry

Following treatment, cells were collected, washed once with PBS, and resuspended in DMEM containing 10 μM DCF-DA. After incubation at 37°C for 30 min in the dark, the cells were harvested and washed three times with PBS. The cell suspension was adjusted to a final volume of 300 μL and filtered through a 70 μm cell strainer. Fluorescence intensity was analyzed using a flow cytometer, and data were processed with FlowJo software (BD, Aria Ⅱ, USA).

Real-time quantitative polymerase chain reaction (RT-qPCR)

After harvesting the cells, total RNA was extracted using the RNeasy Mini Kit. Reverse transcription was performed with the ReverTra Ace qPCR RT Kit, and the reaction mixture was prepared using SYBR Green Premix. RT-qPCR was carried out on the Bio-Rad CFX Connect System. The primer sequences used are as follows(primer sequences 5′→3′):

Il1b: forward primer: GAAATGCCACCTTTTGACAGTG, reverse primer: TGGATGCTCTCATCAGGACAG;

Il6: forward primer: CTGCAAGAGACTTCCATCCAG, reverse primer: AGTGGTATAGACAGGTCTGTTGG.

Cell immunofluorescence

Cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and blocked with PBS containing 10% donkey serum. Subsequently, the cells were incubated at 4°C overnight with the primary antibodies diluted in blocking buffer. After washing with PBS, the cells were incubated with fluorophore-conjugated secondary antibodies for 1 h at room temperature in the dark. Nuclei were counterstained with 4′,6-diamidino-2′-phenylindole (DAPI). All steps were followed by three washes with PBS. Images were acquired using a confocal laser scanning microscope (UNNY, CSIM110, Beijing, China).

Hematological parameter and plasma biochemistry

Hematological parameters were assessed using an automated hematology analyzer (Mindray Animal Care, Shenzhen, China). Plasma biochemical parameters were evaluated by biochemical autoanalyzer (MNCHIP, Tianjin, China).

Histological analysis

Tissue samples were sampled, fixed, and embedded to prepare sections, followed by H&E staining. Histopathological changes, such as edema, inflammatory cell infiltration, and necrosis, were observed under an optical microscope (NIKON, ECLIPSE-E100, Japan).

Immunohistochemistry and immunofluorescence staining

Paraffin-embedded and frozen tissue sections were processed for immunohistochemical and immunofluorescence analyses. Following deparaffinization and antigen retrieval, sections were incubated overnight at 4°C with primary antibodies. Subsequently, sections were incubated with corresponding horseradish peroxidase (HRP)-conjugated secondary antibodies. Nuclei were counterstained with DAPI. Imaging was performed using a Nikon Eclipse C1 microscope (Japan).

Mouse models of sepsis and “two-hit” sepsis models

A sepsis model was established via cecum ligation and puncture (CLP). Before surgery, mice were fasted for 12 h with free access to water. Anesthesia was induced by intraperitoneal injection of 2.5% pentobarbital sodium (30 mg/kg), followed by abdominal disinfection with 75% ethanol. A midline abdominal incision was made to expose the cecum. The distal portion of the cecum was ligated with a 3-0 sterile suture and punctured twice using a 20-gauge needle to allow fecal content leakage into the abdominal cavity. The cecum was then returned to the abdomen, and the incision was sutured layer by layer. To prevent shock, 1 mL of 0.9% NaCl was administered subcutaneously immediately after surgery.

Following CLP, the mice were randomly assigned to one of three experimental groups: (i) CLP group, (ii) CLP + BE group (treated with baicalein), and (iii) CLP + Ce-BE group (treated with Ce-BE NZs). Age- and gender-matched healthy mice without any surgical procedure served as the control group. Twenty-four hours after modeling, blood and tissue samples were collected from euthanized mice (via anesthetic overdose). All efforts were made to minimize suffering.

For “Two-hit” sepsis mice model, on day 3 post-CLP model establishment, PA. (1 × 10⁶ CFU/mL) was administered via intratracheal instillation followed by respective therapeutic interventions, and different therapeutic regimens were administered immediately postoperatively.

For ferroptosis induction, each mouse was given 0.2 mL 5 mM FINO2 by gavage once daily for 3 days. Subsequently, the CLP model was established.

Mitochondrial function assay

Hepatic mitochondria isolated from liver tissue were examined by transmission electron microscopy (TEM; Thermo Scientific, Talos F200S, USA) for structural, quantitative, and morphological analysis. ATPase activity was determined by quantifying inorganic phosphorus release.

Mitochondria were extracted from mouse liver tissue using a Mitochondrial Isolation Kit according to the manufacturer’s protocol. Oxygen consumption rates (OCRs) were assessed using the Agilent Seahorse XF Cell Mito Stress Test Kit and the Substrate Oxidation Stress Test Kit. Sequential injections of oligomycin, FCCP, and a mixture of rotenone and antimycin A were applied to evaluate basal respiration, maximal respiration, spare respiratory capacity, and ATP production-linked respiration.

Mitochondrial proteomic profiling was conducted on highly purified mitochondrial samples from the CLP and Ce-BE NZs treatment groups. Frozen samples were homogenized in low protein-binding tubes. Protein concentration was determined via BCA assay. Separation was performed using an EASY-nLTM 1200 system (Thermo Fisher Scientific, USA) equipped with a C18 column (25 cm × 75 μm) at a flow rate of 300 nL/min over a 75-min gradient. Mass spectrometric analysis was carried out on a Q-Exactive HF instrument coupled to a Nanospray Flex ion source (Thermo Fisher Scientific, USA).

Differentiation of hESCs into macrophages

The hESCs were cultured in mTeSR1 medium. On the day preceding differentiation, cells were seeded at a density of 1×10⁵ cells per well in a six-well plate and incubated overnight in mTeSR1 medium supplemented with 10 μM Y-27632. The following day, the medium was replaced with BEL medium containing 25 ng/mL BMP4, 25 ng/mL Activin A, 25 ng/mL bFGF, and 2 μM CHIR99021 (Stage 1 induction medium). After a 2-day induction period, the medium was refreshed with BEL medium containing 100 ng/mL VEGF, 20 ng/mL bFGF, and 5 μM SB431542 (Stage 2 induction medium). Post a 3-day induction, the cells were thoroughly digested into single cells, resuspended in the Stage 2 induction medium with 10 μM Y-27632, and replated at a density of 1.5×10⁵ cells per well in a low-adhesion six-well plate for further overnight culture. Thereafter, the medium was replaced with BEL medium containing 50 ng/mL SCF, 20 ng/mL TPO, 20 ng/mL IL-3, 20 ng/mL Flt3L, 50 ng/mL M-CSF, 20 ng/mL GM-CSF, and 10 μM SB431542 (Stage 3 induction medium). The culture was maintained for 6–8 days. Then, the suspended cells were harvested and passed through a 70 μm cell strainer. CD14⁺ cells were enriched and collected by magnetic bead sorting. Details regarding the formulation of BEL medium can be found in the relevant literature.

Human liver-derived organoid treatment

For the LPS-induced macrophage-incorporating human liver-derived organoid model, hESCs-derived macrophages were seeded into organoid-containing wells at a density of 1×10⁶ cells per well and co-cultured for 48 h. Subsequently, the wells were pretreated with 100 μg/mL of either Ce-BE NZs, BE, or PBS for 1 h, followed by the addition of LPS to a final concentration of 1 μg/mL. The organoids were further cultured for an additional 24 h prior to harvesting for subsequent analyses.

Quantification and statistical analysis

All experiments were performed with at least three independent replicates. Data are expressed as mean ± standard deviation (mean ± SD) and analyzed using GraphPad Prism (version 9.5.0). For statistical comparisons, one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison test was applied, while survival data were analyzed using the Log rank test. A significance threshold of ^∗p < 0.05, ^∗∗/##p < 0.01, and ^∗∗∗p < 0.001 was used to denote statistical significance.

Published: January 20, 2026