Chemically tailored anionic antibiotic adjuvants targeting divalent cations to overcome carbapenem resistance in gram-negative bacteria

Abstract

No clinically available antibiotic adjuvants can overcome metallo-β-lactamase (MBL) resistance in carbapenem-resistant Gram-negative ESKAPE pathogens. Divalent cations in MBL and quasicrystalline lipopolysaccharides (LPS) are key determinants of acquired and intrinsic resistance in these bacteria. Chemically depriving these cations would block multiple resistance pathways simultaneously and restore carbapenem susceptibility. Here, we chemically engineer a class of anionic adjuvants to tailor their interactions with MBL, enhancing the enzymatic inhibition by up to ~5885-fold compared to parent compounds. An orthogonal screen identifies the tiopronin-engineered anionic adjuvant (TINA), which achieves complete rescue of all tested carbapenems against diverse New Delhi metallo-β-lactamase-1–producing Gram-negative ESKAPE bacteria. TINA targets both MBL and LPS, depriving divalent cations to inactivate MBL and damage membrane collaboratively while also preventing resistance evolution. Moreover, TINA’s anionic nature minimizes serum adsorption, enabling safe and effective in vivo treatment of bacterial pneumonia. This work provides an innovative chemical biology strategy leveraging anionic materials to combat superbugs globally.

Susceptibility to last-resort antibiotics is restored in carbapenem-resistant Gram-negative ESKAPE pathogens.

INTRODUCTION

Carbapenem-resistant (CR) Gram-negative ESKAPE pathogens (Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) represent extremely severe threats to global public health (1–4). These bacteria evolve resistance to make even the last-line antibiotics of carbapenems obsolete (5). Using adjuvants with antibiotics to overcome resistance is a feasible strategy to combat drug-resistant bacteria (6–11), but it is a great challenge to reverse metallo-β-lactamases (MBLs; class B carbapenemase) resistance in CR Gram-negative bacteria (1, 12–14). Presently, clinically available adjuvants are only active to inhibit serine hydrolases (class A carbapenemase, e.g., K. pneumoniae carbapenemase, class C carbapenemase, e.g., cephalosporinases, and class D carbapenemase, e.g., oxacillinases) (15, 16). There are no approved inhibitors of MBLs like New Delhi metallo-β-lactamase 1 (NDM-1), which is one of the most clinically relevant MBLs (17, 18). Inhibitors of class A, C, and D carbapenemase are also ineffective against NDM-1 (19). Besides, MBLs transfer among all CR Gram-negative ESKAPE strains, highlighting the need for broad-spectrum activities of adjuvants (20). Thus, new adjuvants are urgently needed.

Existing adjuvants cannot simultaneously overcome both acquired and intrinsic resistance in CR Gram-negative bacteria (21–24). In general, acquired MBLs hydrolysis β-lactam (25, 26), while intrinsic quasicrystalline lipopolysaccharides (LPS) hinder antibiotic penetration (27–29). Small molecules (e.g., ethylenediamine-N, N, N′, N′-tetraacetate, aspergillomarasmine A, and bismuth compounds) can inhibit MBL but are useless against LPS (19, 26, 30). Cationic adjuvants (e.g., antimicrobial peptides and polymers) lack the specificity to MBLs and are susceptible to increased cytotoxicity due to their severe nonspecific adsorptions with diverse biological components (e.g., serum proteins and mammalian membranes) (11, 31–35). Thus, overcoming both acquired and intrinsic resistance in MBL-producing Gram-negative bacteria is technically challenging, and devising new antibiotic adjuvant systems is highly desirable.

In this study, we chemically engineered a new class of anionic adjuvants targeting divalent cations of MBL and LPS to overcome both acquired and intrinsic resistance in NDM-1–producing Gram-negative ESKAPE pathogens (Fig. 1). Divalent cations play crucial roles in determining MBL activity (Zn²⁺) and LPS assembly (Ca²⁺ and Mg²⁺). We modulated the interaction between adjuvants and MBL toward efficient deprivation of divalent cations via a chemical tailoring strategy. We established a class of 18 effective anionic adjuvants, in which we identified the tiopronin-engineered anionic adjuvants (TINA) showing 100% reversal activities without inducing new resistance. We confirmed that TINA dually targeting NDM-1 and LPS and depriving multiple divalent cations. The anionic nature endowed TINA with high biosafety. The in vivo efficacy of TINA was also confirmed in multiple mouse models. Our study provided an innovative chemical biology strategy of overcoming both acquired and intrinsic resistance for combating superbugs.

Fig. 1. Schematic illustration.

Using a class of chemically tailored anionic antibiotic adjuvants that target divalent cations to address both intrinsic and acquired resistance in MBL-producing CR Gram-negative ESKAPE pathogens. Created in BioRender. Jiang, X. (2025) https://BioRender.com/d7xj9st.

RESULTS

Design of chemically tailored anionic adjuvants for depriving divalent cations of MBL and the identification of TINA

The dissociation behavior between adjuvants and MBL would control the Zn²⁺ deprivation capability of adjuvants. Upon binding to the Zn²⁺ of MBL, only if the adjuvants efficiently dissociate from MBL would Zn²⁺ possibly be deprived from MBL; thus, a faster dissociation speed would lead to more efficient deprivation of Zn²⁺. Because ligand-decorated gold scaffolds have surface chemical diversities (e.g., chemical conformation, surface acidity, and multivalent effect) (36–40) for versatile interactions with different substances (e.g., proteins, biopolymers, and ions), we selected five carboxylic-terminated thiols (penicillamine, thiorphan, 3-mercaptoisobutyric acid, tiopronin, and captopril) to engineer the gold nanoscaffolds toward tailoring diverse dissociation manners (Fig. 2A). These ligands can target MBL but have very weak enzymatic inhibition activity (41). To deprive Zn²⁺, we proposed to develop anionic adjuvants that exhibited a slow binding profile to MBL but followed a quick dissociation mode.

Fig. 2. The construction of the chemically engineered anionic antibiotic adjuvants and tailored interactions with MBLs.

(A) Workflow of the development of anionic adjuvants by chemically displaying carboxyl ligands on gold scaffolds, analyzing interaction modes using surface plasmon resonance (SPR), and performing in vitro enzymatic inhibition assays. (B) Dissociation constants (K_off) as shown in (D) to (H). (C) Retained Zn²⁺ (%) in MBLs after coincubating with diverse adjuvants compared to the blank untreated MBLs. (D to H) Relative response curves of (D) penicillamine-, (E) thipoprhan-, (F) 3-mercaptoisobutyric acid–, (G) tiopronin-, and (H) captopril-modified anionic adjuvants against MBL. Both the association and dissociation times were set to 100 s. (I to M) Enzymatic inhibition curves for different small molecules (black dots and lines) and their anionic adjuvant counterparts (red dots and lines) against MBL. The fold changes in half-maximal inhibitory concentration (IC₅₀) values were showed in each figure. (I to M) Curves for penicillamine (G), thipoprhan (J), 3-mercaptoisobutyric acid (K), tiopronin (L), and captopril (M) and their anionic adjuvant counterparts, respectively. In (D) to (M), all anionic adjuvants were prepared at the gold/ligands ratio of 1:1.

We used surface plasmon resonance (SPR) to study the interactions between MBL and anionic adjuvants. We synthesized the engineered anionic adjuvants through a sodium borohydride-mediated one-pot reduction approach (42). We aimed at finding anionic adjuvants that were capable of depriving Zn²⁺ of MBL to reverse the MBL-mediated acquired resistance. When these adjuvants interacted with MBL through Zn²⁺, we presumed that they would dissociate from MBL efficiently to separate Zn²⁺ from MBL. We used the dissociation constant (K_off) to characterize the dissociation rates. The surface chemical engineering strategy substantially increased the binding affinities between MBL and anionic adjuvants compared to ligands (Fig. 2, B and D to H, and fig. S1). The dissociation modes of anionic adjuvants differed in a ligand-dependent manner (Fig. 2B). Penicillamine-modified adjuvants did not dissociate (K_off = 1.30 × 10⁻⁵ s⁻¹). Thiorphan and 3-mercaptoisobutyric acid–modified anionic adjuvants dissociated slowly (K_off = 1.98 × 10⁻³ s⁻¹, K_off = 6.06 × 10⁻² s⁻¹). Notably, tiopronin and captopril-modified anionic adjuvants exhibited a unique immediate dissociation profile (K_off = 5.41× 10⁻¹ s⁻¹, K_off = 8.36 × 10⁻¹ s⁻¹), which had the potential to deprive Zn²⁺ quickly from MBL. We measured the Zn²⁺ deprivation activity of these anionic adjuvants. We incubated MBLs with these adjuvants, dialyzed the coincubated solutions, and measured the retained Zn²⁺ within MBLs by inductively coupled plasma mass spectrometry (ICP-MS). The K_off values of the adjuvants correlated with their Zn²⁺ deprivation efficiency toward MBLs. Tiopronin and captopril-modified anionic adjuvants exhibited higher K_off values, and they facilitated more efficient Zn²⁺ removal (30 to 40%) from MBLs. Penicillamine-, thiorphan-, and 3-mercaptoisobutyric acid–modified adjuvants exhibited lower K_off values, and they barely deprived Zn²⁺ from MBLs (Fig. 2C). We determined the enzymatic activity inhibition of these anionic adjuvants against MBL using an in vitro enzyme activity assay. The anionic adjuvants showed 121- to 5885-fold decreases in half-maximal inhibitory concentration (IC₅₀) compared to the ligands alone, demonstrating the great potential of reversing MBL-mediated carbapenem resistance (Fig. 2, I to M). Thus, chemically tailored anionic adjuvants showed enhanced inhibition of MBL.

We did not directly correlate the enzymatic inhibition activities of the adjuvants to their dissociation modes. These adjuvants showed activity in inhibiting MBLs (Fig. 2, I to M) while they had different K_off values (Fig. 2B) and Zn²⁺ deprivation efficiency (Fig. 2C). These results suggested that there would exist different enzymatic inhibition mechanisms among these adjuvants. Tiopronin and captopril-modified anionic adjuvants probably inhibited MBL by a Zn²⁺ deprivation mechanism, and the inhibition mechanism of penicillamine-, thiorphan-, and 3-mercaptoisobutyric acid–modified adjuvants would not involve Zn²⁺ deprivation. Considering that K_off was primarily reflective of Zn²⁺ deprivation, it was inappropriate to directly link K_off to the enzymatic efficiency of all mechanistically diverse adjuvants.

In sum, we found that tiopronin and captopril-modified adjuvants showed quick dissociation from MBL, most likely due to depriving Zn²⁺. We next extended the number of the anionic adjuvants and tested their synergy effects with carbapenems against MBL-related CR Gram-negative ESKAPE pathogens (Fig. 3A). We obtained a total of 18 anionic adjuvants by altering the feeding ratios (1:1, 1:2, 1:4, and 1:10; fig. S2). We did not obtain stable captopril-modified formulations at the gold/captopril feeding ratios at 1:4 and 1:10. We constructed a pairwise orthogonal high-throughput chessboard dilution assay over 552 combinations involving 23 adjuvants (18 anionic adjuvants and 5 ligands) and 6 carbapenems (imipenem, meropenem, ertapenem, biapenem, faropenem, and doripenem) on 4 clinically isolated NDM-1–producing CR Gram-negative ESKAPE pathogens (Fig. 3B and figs. S3 and S4 to S26; for the minimum inhibitory concentrations (MICs) of the used antibiotics against these pathogens, see table S1). The fractional inhibitory concentration index (FICI) was used to determine the synergistic efficacy, and an FICI lower than 0.5 was regarded as synergistic. Small molecular ligands were substantially less effective than their anionic adjuvant counterparts. The overall positive synergistic rate of each ligand alone was 0 to 20.83%, while anionic adjuvants showed remarkably high synergistic rates of 41.67 to 100% (Fig. 3B). This demonstrated that the surface chemical engineering strategy was a feasible approach to constructing the anionic adjuvants for combating Gram-negative CR bacteria. Notably, the TINA, which was formulated by feeding gold/tiopronin at 1:10, showed 100% broad-spectrum synergistic activity in 24 combinations with the FICI ranging from <0.031 to 0.5 (Table 1 and table S2). TINA also demonstrated the most potent synergistic effect compared to another three anionic adjuvants modified by tiopronin at Au/tiopronin ratios of 1:1, 1:2, and 1:4 (Fig. 3B and fig. S27). TINA had no synergistic antibacterial effect with imipenem against methicillin-resistant Staphylococcus aureus (fig. S28).

Fig. 3. The discovery of the TINA which potentiated the efficacy of carbapenems against CR Gram-negative ESKAPE bacteria and had high biosafety properties.

(A) Workflow of the discovery of TINA, including constructing a class of anionic adjuvants by varying the feeding ratios of different ligands on gold scaffolds, conducting high-throughput screening of synergistic efficacy with checkerboard assays, and validating its efficacy and biosafety properties. (B) Heatmap of the FICI for 552 combinations involving 23 adjuvants (18 anionic adjuvants and 5 surface ligands), 6 carbapenems, and 4 NDM-1–producing Gram-negative ESKAPE bacteria. The synergy (%) indicates the rate of positive synergy (FICI < 0.5) across 24 combinations of each adjuvant. (C) Double spherical aberration–corrected TEM images and size distribution analysis (n = 120) of TINA. (D) Zeta potential of TINA. (E) Gel analysis of serum protein adsorption capabilities of TINA. Cationic branch polyethyleneimine (PEI; 25,000 g/mol) was used as a control. NC, negative control. (F) The time-dependent killing curves for TINA (20 μg/ml) or imipenem (Imi; 8 μg/ml) monotherapy or their combination therapy against CRKP. The combination groups were set as TINA (20 μg/ml) and imipenem (2 μg/ml), TINA (10 μg/ml), and imipenem (8 μg/ml). These two groups were abbreviated as TINA + Imi (20 + 2) and TINA + Imi (10 + 8) in the figures of this article. (G) Resistance development curves during serial 20 passages with the subinhibitory concentration of imipenem and the combination of TINA and imipenem against CRKP. (H) Cytotoxicity of TINA toward HUVECs. (I) Hemolysis test of TINA. (J) H&E staining of organs including the heart, liver, spleen, lung, and kidney after treating mice with PBS and TINA.

Table 1. Potency of TINA in combination with different carbapenems against NDM-1–producing CR Gram-negative ESKAPE bacteria indicated by FICI.

Synergy was defined as an FIC index of ≤0.5.

Bacteria	Antibiotic
	Imipenem	Meropenem	Ertapenem	Biapenem	Faropenem	Doripenem
CRKP	<0.031	<0.280	0.270	0.190	<0.080	<0.090
CRAB	0.047	0.250	0.090	0.250	0.500	0.188
CRPA	0.375	<0.156	0.160	<0.140	0.250	0.313
CRE	0.156	<0.156	0.140	0.375	0.060	0.060

On the basis of the high potential in depriving Zn²⁺ and the potent synergy efficacy, we used TINA to demonstrate the strategy of addressing both acquired and intrinsic resistance by targeting divalent cations. Tiopronin was a Food and Drug Administration–approved medication to treat severe homozygous cystinuria (43). A high translation potential of TINA was expected provided the efficacy and biosafety were validated. TINA was to be a gold nanocluster (44) with a diameter of 2.14 ± 0.25 nm revealed by double spherical aberration–corrected transmission electron microscopy (TEM) images (Fig. 3C). In aqueous solutions, TINA had a hydrodynamic diameter of 7 ± 0.62 nm and an anionic zeta potential of −30.89 ± 1.40 mV (Fig. 3D). We confirmed the capping of tiopronin on the surface of TINA by x-ray photoelectron spectroscopy and Fourier transform infrared spectroscopy (figs. S29 and S30). Of note, TINA highly resisted serum protein adsorption due to its anionic characteristic, greatly benefiting its precise interaction with MBL to deprive divalent cations (Fig. 3E and fig. S31).

We validated the bactericidal capabilities of the combinations of TINA and imipenem on Gram-negative ESKAPE pathogens. TINA (20 μg/ml) plus imipenem (2 μg/ml) and TINA (10 μg/ml) plus imipenem (8 μg/ml) killed 99.99% of CR K. pneumoniae (CRKP) within 24 hours, respectively. In contrast, the control groups (TINA or imipenem alone) increased by two orders of magnitude in bacteria counts (Fig. 3F). Accordingly, the culturing media of CRKP remained clear in the TINA plus imipenem groups post–24-hour treatment, demonstrating the long-term synergistic strategy (fig. S32A). TINA also increased the antibacterial activity of imipenem and decreased the number of alive CRE and CRAB after combination treatments (fig. S32, B and E), showing potent and broad-spectrum synergistic bactericidal activities.

The ideal antibiotic adjuvants would prevent the bacterial evolution of new drug resistance. We investigated whether continued use of TINA and imipenem led the bacteria to generate new antibiotic resistance. We cultured CRKP with imipenem in the presence or absence of TINA at their subinhibitory concentrations for 20 passages (Fig. 3G). Bacterial resistance was prevented in CRKP cells cultivated with imipenem (1 μg/ml) and TINA (10 μg/ml) after 20 passages as the MIC was unchanged. In contrast, CRKP cells cultivated imipenem (32 μg/ml) alone for 20 passages became more resistant, and the MIC increased from 64 to 2048 μg/ml. Collectively, by chemically engineering the surface to achieve anionic adjuvants capable of depriving Zn²⁺ of MBL, we obtained TINA that exhibited potent, broad-spectrum synergistic efficacy and prevented the emergence of a new resistance.

Anionic TINA showed biocompatibility both in vitro and in vivo

We used human umbilical vein endothelial cells (HUVECs) and 3T3 to examine the cytotoxicity of TINA. More than 90% of cells stayed alive when we treated them with TINA at a remarkably high concentration of 400 μg/ml (Fig. 3H and fig. S33). Using phase-contrast microscopy, we found that TINA-treated cells showed identical morphology to the control cells (fig. S33C). We also confirmed the cell viability of TINA-treated cells by the calcein acetoxymethyl ester (Calcein-AM)/propidium iodide (PI) costaining assay. We did not find an obvious signal of PI when we incubated cells at 100 μg/ml for 48 hours (fig. S34). These results demonstrated that TINA was of low toxicity as it showed negligible cytotoxicity toward cells in vitro at a concentration far higher than the antibacterial concentration.

Before we evaluated the in vivo performances of TINA, we used the hemolytic assay to test the blood compatibility of TINA. Results showed that TINA did not lysis the erythrocytes at the high concentration of 400 μg/ml, indicating its biocompatibility to blood (Fig. 3I). Inspired by this, we administrated TINA intraperitoneally into mice (50 mg/kg) and comprehensively analyzed the in vivo toxicity of TINA. We analyzed blood biochemical parameters post-TINA injection. All hematological parameters we tested in the TINA-treated mice within a week postinjection were comparable to those of the control mice (fig. S35). We also did not observe signs of damage and inflammation in the hematoxylin and eosin (H&E) staining images of major organs (hearts, livers, spleens, lungs, and kidneys) in the TINA-treated mice (Fig. 3J and fig. S35). We studied the clearance of TINA from major organs over time. We found a clearance trend in the major organs from day 7 to day 14 post-TINA injection. The contents of TINA detected were all lower than 3 injected dose per gram of organ (%ID/g), suggesting that TINA had the ability for in vivo clearance (fig. S36). We also measured the clearance of TINA from blood upon intravenous injection (fig. S37). We found that the TINA within blood reduced to ~1 %ID/g at 2 hours postinjection. Thus, TINA showed negligible cytotoxicity to normal cells and did not induce acute toxicity to organs in mice, which collectively affirmed TINA to be a potent adjuvant with notable biosafety.

Anionic TINA dually targeted MBL and LPS to deprive divalent cations

Having demonstrated the synergistic efficacy and biosafety of TINA, we next used CRKP as the model pathogen to investigate the divalent cation-targeting mechanisms of TINA. We first confirmed that TINA demonstrated dose-dependent inhibition of MBL with an IC₅₀ of 0.80 μM (Fig. 4A). By ICP-MS, we determined that the addition of increasing amounts of Au resulted in the gradual deprivation of Zn²⁺ from MBL. For example, TINA at the 0.1 equivalent to MBL caused a 1.67-fold decrease of Zn²⁺ (Fig. 4B). However, the supplement of Zn²⁺ restored the hydrolysis activity. For example, the addition of 40 μM Zn²⁺ totally recovered the enzymatic activity to ~100% (Fig. 4C). Meanwhile, we found that adding Zn²⁺ (10 μM) also decreased the synergistic efficacy between TINA and imipenem (FICI = 0.531; Fig. 4D), suggesting that competing with Zn²⁺ of MBL would be responsible for the combination effects of TINA. These results proved that TINA acted against Zn²⁺ to inactivate the MBL, further driving us to gain insights into the detailed interaction at the adjuvant/protein interface. By isothermal titration calorimetry (ITC), we determined that TINA bound to MBL at the equilibrium dissociation constant (K_D) of 88 μM while changing entropy and enthalpy simultaneously (Fig. 4E), indicating a diversified interaction mode. Thermodynamic parameters during this process were similar to that between TINA and free Zn²⁺ (Fig. 4F), pointing out that Zn²⁺ might be the intermediate. The Lineweaver-Burk plot further confirmed that TINA was a competitive inhibitor for MBL with an inhibition constant value of 6 μM (Fig. 4G). Besides, SPR curves determined that TINA disassociated from MBL fastly (K_off = 6.94 × 10⁻¹ s⁻¹; fig. S38). We assumed that TINA would quickly deprive Zn²⁺ during the dissociation process. Next, we used SPR technology to conduct the antagonism binding experiment to prove this (Fig. 4H). We pretreated TINA with a series of concentrations of Zn²⁺ solutions (0, 100, 200, and 400 μM). If Zn²⁺ acted as the intermediary substance responsible for TINA/MBL interaction, then pretreating TINA with Zn²⁺ would decrease the binding of TINA to MBL. Results showed that Zn²⁺ abolished the binding of TINA to MBL in a dose-dependent manner, demonstrating that TINA interacted with MBL primarily driven by Zn²⁺. In short, we demonstrated that TINA interacted with MBL through binding Zn²⁺ and resulted in the deprivation of Zn²⁺ for enzymatic inhibition (Fig. 4I).

Fig. 4. TINA deprived Zn²⁺ of MBL and inhibited hydrolytic activity.

(A) TINA inhibited the hydrolytic activity of MBL with the IC₅₀ of 0.80 μM. (B) The deprivation of Zn²⁺ of MBL by TINA over equilibrium dialysis. The metal content was determined by ICP-MS. (C) Addition of excess Zn²⁺ restored activity postinactivation by TINA. (D) Exogenous addition of Zn²⁺ inhibited the combination effects of TINA and imipenem against CRKP. (E and F) ITC thermograms for the binding of TINA to MBL (E) and TINA to Zn²⁺ (F). The inserted data represented the thermodynamic constants during the binding between TINA to MBL or Zn²⁺. DP, differential power. (G) The double reciprocal plot of substrate-dependent enzyme kinetics on inhibition of MBL activity by TINA. (H) SPR-based antagonism binding curves of TINA and MBL by adding a series of concentrations of Zn²⁺ solutions (0, 100, 200, and 400 μM). (I) Schematic illustrations of the action mechanism of TINA inactivated MBL by depriving Zn²⁺.

We investigated the effects of TINA against LPS for depriving divalent cations. In Gram-negative bacteria, lipid A had anionic phosphate groups to form ionic bridges with divalent cations (e.g., Ca²⁺ and Mg²⁺) to achieve tight packing and low permeability of the outer membrane, exhibiting responsibility for the intrinsic resistance (45, 46). We found that either K. pneumoniae– or Escherichia coli–origin LPS significantly diminished the combination effects of TINA (20 μg/ml) and imipenem (2 μg/ml) treatments in a dose-dependent manner (Fig. 5A and fig. S39). Because TINA was anionic, we hypothesized that it competed with LPS for binding divalent cations, which differed from cationic adjuvants as they generally interacted with LPS directly (11). To verify the competition between TINA and the negatively charged domain of lipid A, we tested the effects of cations (Na⁺, K⁺, Ca²⁺, and Mg²⁺) on the antibacterial activity of the combination of TINA and imipenem (Fig. 5B). Only divalent cations (Ca²⁺ and Mg²⁺) inhibited the activity of synergistic formulation, whereas monovalent cations (Na⁺ and K⁺) had no effect. The synergistic effects of the combination of TINA and imipenem were diminished with the increase in the concentration of divalent cations. To compare the binding affinity of TINA and LPS for Ca²⁺ and Mg²⁺, we performed the ITC measurements. The K_D values of LPS for Ca²⁺ and Mg²⁺ were 34 × 10⁻⁸ M and 72.2 × 10⁻⁴ M, respectively (fig. S40). Compared with LPS, TINA can also bind to Ca²⁺ and Mg²⁺, with K_D values of 56.2 × 10⁻⁶ M and 17.3 × 10⁻⁵ M, respectively (Fig. 5, C and D). The affinity of LPS for Ca²⁺ was higher than that of TINA for Ca²⁺. However, the binding between LPS for Ca²⁺ mainly depended on the change in Gibbs free energy while that between TINA and Ca²⁺ depended on changing both enthalpy change and entropy. The affinity of TINA for Mg²⁺ was higher than that of LPS for Mg²⁺. Meanwhile, the binding of LPS to Mg²⁺ was an endothermic process, while the binding of TINA to Mg²⁺ was an exothermic process, which caused Mg²⁺ to tend to bind to TINA to reduce the system energy. These results indicated that TINA targeted LPS and competed with Ca²⁺ and Mg²⁺.

Fig. 5. TINA deprived Ca²⁺ and Mg²⁺ in LPS to disrupt the membrane and exert synergistic efficacy.

(A) Exogenous addition of KP-derived LPS (0 to 128 μg/ml) impaired the synergistic antibacterial activity of TINA (10 μg/ml) and imipenem (8 μg/ml) against CRKP. (B) Exogenous addition of Ca²⁺ or Mg²⁺ (25 μg/ml) abolished the synergistic antibacterial activity of TINA (10 μg/ml) and imipenem (8 μg/ml) against CRKP. (C and D) ITC thermograms for the binding of TINA for Ca²⁺ (C) and Mg²⁺ (D). The inserted data represented the thermodynamic constants. (E) The combination of TINA and imipenem disrupted the outer membranes of CRKP probed with N-phenyl-1-naphthylamine (NPN). (F) The combination of TINA and imipenem increased the permeability of the inner membrane of CRKP probed with PI. (G) TEM images of CRKP after exposure to TINA (10 μg/ml) or imipenem (8 μg/ml) monotherapy and their combination therapy. Arrows indicated the membrane disruption and perturbation of outer membrane vesicles (OMVs). (H) The combination of TINA and imipenem disrupted the cytoplasmic membrane potential probed with the 3,3-dipropylthiadicarbocyanine iodide [DiSC₃(5)]. (I) Inhibition of efflux pumps by combination treatment with TINA and imipenem in an ethidium bromide (EtBr) fluorescence intensity assay. In (E), (F), (H), and (I), T + I (20 + 2) and T + I (10 + 8) represented the combination treatments at concentrations of TINA (20 or 10 μg/ml) and imipenem (2 or 8 μg/ml), respectively. a.u., arbitrary units; ns, not significant.

Next, we investigated the impacts of targeting Ca²⁺ and Mg²⁺ of LPS packages by TINA. We tested the concentration of LPS, Ca²⁺, and Mg²⁺ in the bacterial supernatant. The combination treatments of TINA and imipenem increased the contents of LPS, Ca²⁺, and Mg²⁺ compared to those in the TINA or imipenem control groups (fig. S41). We used N-phenyl-1-naphthylamine (NPN), a hydrophobic and fluorescent molecule, to probe the integrity of the outer membrane (11). Combination treatments of TINA and imipenem enhanced the fluorescent intensity of NPN, verifying that TINA interfered with the outer membrane and increased the membrane permeability (Fig. 5E). We used PI, a nucleic acid dye that hardly entered into the intact membrane but easily entered into the broken membrane, to probe the integrity of the inner membrane (47). In flow cytometry assay, TINA in combination with imipenem induced the ratios of PI-positive cells to be 87.5% [TINA (10 μg/ml) + imipenem (8 μg/ml)] and 97.4% [TINA (20 μg/ml) + imipenem (2 μg/ml)], which were notably higher than that in TINA or imipenem groups (Fig. 5F and fig. S42). That was, the combination of TINA and imipenem disrupted the integrity of the bacterial outer and inner membranes, permitting PI to label the DNA presented in the bacterial cytoplasm. In addition, live/dead staining confirmed that the combination of TINA and imipenem substantially reduced the bacterial viability of CRKP (fig. S43). We also visualized the changes in bacterial morphology. By scanning electron microscopy, we observed that CRKP treated with the combination of TINA and imipenem exhibited severe damage to collapsed bodies and rough and irregular cell surfaces (fig. S44). TEM images also confirmed these effects on CRKP (Fig. 5G). The control cells exhibited a complete cell envelope, whereas bacteria treated with TINA and imipenem showed membrane damage and intracellular component loss. Notably, the outer membrane vesicles of CRKP attached to the surface of the bacterial membrane loosely upon the combination treatments, which indicated that the combination formulations induced the instability of the cell membrane of CRKP. Besides, we detected the increased leakage of lactate dehydrogenase (LDH) when treating CRKP with TINA and imipenem, which was in line with membrane damage of CRKP by the combination treatment (fig. S45). We used 3,3-dipropylthiadicarbocyanine iodide [DiSC₃(5)] and 2′, 7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF)–AM to determine the change in electric potential respectively (∆ψ) and transmembrane proton gradient (∆pH) of CRKP upon treatments of different formulations. Combining TINA and imipenem significantly destroyed the membrane potential (Fig. 5H) and caused ΔpH dissipation (fig. S46) of CRKP. In addition, the accumulation of ethidium bromide (EtBr) in CRKP bacteria increased substantially after adding TINA and imipenem, indicating that TINA combined with imipenem inhibited efflux pump activity (Fig. 5I). In short, anionic TINA competed with divalent cations (Ca²⁺ and Mg²⁺) against LPS, disrupting the membrane and facilitating the intracellular accumulation of antibiotics (fig. S47). Collectively, TINA dually targeted MBL and LPS to deprive divalent cations (Zn²⁺, Ca²⁺, and Mg²⁺) to inhibit MBL and disrupt the membrane system, thus rescuing carbapenem by simultaneously overcoming acquired and intrinsic resistance in CR Gram-negative pathogens.

Proteomic analysis validated the dual-targeting mechanisms of TINA

We performed the proteomic analysis on CRKP to reveal distinct expression patterns of proteins between imipenem (alone) and the synergy treatment of TINA and imipenem (10 μg/ml + 8 μg/ml). The volcano plot displayed 20 up-regulated and 42 down-regulated proteins in the TINA-imipenem combination group compared to the imipenem group, featuring a significant dissimilarity in the principal component analysis plot (Fig. 6, A and B). Notably, we found a pronounced down-regulation of MBL (fold change = 0.59, P = 0.02) (48), thus strongly suggesting that Zn²⁺ deprivation by TINA induced MBL destabilization (Fig. 6A). Besides, the up-regulated proteins were mainly linked to membrane damage of bacteria in cellular component including outer membrane, plasma membrane, and pilus, in line with the fact that TINA interrupted the membrane stability and permeability (Fig. 6, C and D). The combination of TINA and imipenem also substantially affected pathways associated with biological metabolism processes [e.g., two-component system, sulfur metabolism, carbon metabolism, adenosine 5′-triphosphate (ATP)–binding cassette transporters, and so forth] (Fig. 6, E and F). Since levels of ATP and reactive oxygen species (ROS) are closely related to metabolic pathways in bacteria, we next verified the effects of the combination of TINA and imipenem on disrupting the metabolism of CRKP by measuring them. We found that TINA increased the production of bactericidal ROS and declined the ATP levels in CRKP, enhancing oxidative stress and confirming the interruption in energy metabolism and in CRKP caused by TINA (Fig. 6, G and H). In short, we confirmed the reversal of both acquired and intrinsic resistance by TINA via proteomic analysis, as it down-regulated the expression of MBL and influenced the membrane integrity and biological metabolism.

Fig. 6. Proteomic analysis of CRKP after exposure to monotherapy of imipenem and the combination therapy of TINA and imipenem.

Volcano plot (A) and principal components (PC) analysis (B) in CRKP after exposing to imipenem alone (8 μg/ml, Imi) or in combination with TINA (10 μg/ml, TINA + Imi) for 3 hours. (C) Gene Ontology enrichment analysis of differentially expressed proteins in CRKP in biological processes, cellular components, and molecular function. (D) Heatmap of proteins that related to microbial resistance, membrane, and metabolism in CRKP after treating imipenem and its combination with TINA. CysA, sulfate/thiosulfate import ATP-binding protein; FtsB, cell division protein; IraP, anti-adapter protein; CbiG, cobalamin biosynthesis protein; RimO, ribosomal protein S12 methylthiotransferase; NTPH, nucleoside triphosphatase. (E and F) Kyoto Encyclopedia of Genes and Genomes enrichment analysis of up-regulated (E) and down-regulated (F) proteins. (G) ROS analysis in CRKP cells after treating TINA with different concentrations. (H) ATP levels analysis in CRKP after treating TINA with different concentrations. FDR, false discovery rate; NAD⁺, nicotinamide adenine dinucleotide; NADH, reduced form of NAD⁺; PAPS, 3′-phosphoadenosine 5′-phosphosulfate; ATPase, adenosine triphosphatase.

TINA showed potent synergistic efficacy with carbapenem in multiple bacterial pneumonia mouse models

Encouraged by the results that the biocompatible TINA effectively increased the susceptibility of imipenem against a wide range of NDM-1–producing Gram-negative CR pathogens, we evaluated the in vivo efficacy of TINA in acute bacterial pneumonia models. We first determined that 12.42 %ID/g of TINA accumulated in the lung through intranasal delivery (fig. S48). We challenged cyclophosphamide-induced immunosuppressive mice by lethal dose of CRKP [5 × 10⁷ colony-forming units (CFU) per mice, n = 10], followed by treatments: (1) control group [phosphate-buffered saline (PBS)]; (2) TINA (10 mg/kg, single dose); (3) imipenem (8 mg/kg, single dose); (4) TINA plus imipenem (10 + 8 mg/kg, single dose); (5) TINA plus imipenem (10 + 8 mg/kg, double dose) (Fig. 7A). A low survival rate of 20% was observed in the PBS group at day postinfection (dpi) 7. The monotherapy of TINA and imipenem did not notably increase the survival rate of pneumonia mice, as 30% survival rates were achieved in these two groups (Fig. 7B). This was consistent with the fact that TINA itself had no antibacterial activity and imipenem lost its effectiveness to the CRKP (Figs. 3F and 7E). In stark contrast, combination treatments of TINA and imipenem resulted in 60% (single dose) and 80% (double dose) survivals at dpi 7. Moreover, the body weights of surviving mice in groups of combination treatments recovered faster than that in other groups (Fig. 7C). We also measured the biosafety profiles of the combination of TINA and imipenem in mice. We found that treating mice with the combination of TINA (10 mg/kg) and imipenem (8 mg/kg) did not cause toxicity (fig. S49). These results demonstrated that TINA sensitized imipenem against CRKP in vivo as their combination increased the survival rate and quality of CRKP-infected pneumonia mice.

Fig. 7. The combination of TINA and imipenem showed potency in mouse pneumonia infection models caused by different NDM-1–producing CR Gram-negative bacteria.

(A) Experimental protocol scheme. Mice with cyclophosphamide-induced immunosuppression were challenged with a lethal dose of CRKP (5 × 10⁷ CFU per mouse, n = 10) and treated with: (1) PBS; (2) TINA (10 mg/kg, single dose); (3) imipenem (8 mg/kg, single dose), abbreviated as Imi; (4) TINA plus imipenem (10 + 8 mg/kg, single dose), abbreviated as TINA + Imi (1×); (5) TINA plus imipenem (10 + 8 mg/kg, double dose), abbreviated as TINA+Imi (2×). ip, intraperitoneal. (B) Survival rate curves of mice infected with CRKP after receiving different treatments. (C) Daily average weight of mice during the treatments. (D) Experimental timeline for the CR pathogens (CRKP, CRE, or CRAB, 5 × 10⁵ CFU per mouse, lower than the lethal dose, n = 5) pneumonia mode and randomly divided into five groups. (E) The bacterial loads in the lungs of mice treated with five groups of treatments after being challenged with CRKP, CRE, or CRAB (n = 5). (F) Representative images of CRKP-, CRE-, or CRAB-infected lung treated with TINA and imipenem. (G to I) Biochemical indicators of the blood indicators white blood cells (WBCs) (G), red blood cells (RBCs) (H), and PLT (I) after the intratracheal administration of TINA and imipenem in CRKP-, CRE-, and CRAB-infected mice. The ranges of WBC, RBC, and PLT of blank mice without any treatments were shown as the gray box in each figure. (J and K) Histopathology evaluation of lungs from different groups by H&E (J) and nuclear transcription factor-κB (NF-κB, K) expression in lungs using immunofluorescence. (L) Immunohistochemical staining images of IL-1β, IL-6, and TNF-α in the lungs after treatments. (M) Tissue injury scores measured in lung tissues after treatments. (N) Quantification of NF-κB expression in lungs posttreatments. (O to Q) Quantification of IL-1β (O), IL-6 (P), and TNF-α (Q) in lungs posttreatments.

To investigate the versatility of TINA in reactivating imipenem against CR pathogens, we established pneumonia models induced by CRKP, CRE, and CRAB in parallel to test the effects of TINA. Mice were intranasally injected with CR pathogens (5 × 10⁵ CFU per mice, lower than that of the lethal dose, n = 5) and randomly divided into five groups (Fig. 7D). We sacrificed the mice after 24 hours of treatments and performed the analysis of bacteria load and pathology in infected lungs. The standard plate dilution method determined that the infected lungs had bacteria loads of 9.5 CFU per lung (CRKP), 10.1 CFU per lung (CRE), and 9.7 CFU per lung (CRAB), respectively. Monotherapy of TINA and imipenem showed mere effects on eliminating live bacteria in the lungs. However, the single treatment of dual therapy of TINA plus imipenem significantly alleviated the lung infection by reducing 2.4-log (CRKP, P = 0.0134), 2.1-log (CRE, P = 0.0013), and 2.2-log (CRAB, P = 0.0004) bacteria loads, respectively. Double dosing of combination therapy further decreased the live bacteria loads in the lungs by 3.7-log (CRKP, P = 0.0002), 3.7-log (CRE, P = 0.0001), and 3.6-log (CRAB, P < 0.0001) (Fig. 7E). We visualized the excised lungs ex vivo. We found the dark-red infected lesions with varying sizes scattered across lung tissues in the PBS and monotherapy groups. In contrast, the lungs treated by the combination therapy, either by single or double dosing, notably decreased the infection locus in the lungs (Fig. 7F). Moreover, we also determined that the bacteria loads in major organs were markedly reduced in these three models (figs. S50 to S52). For example, in the CRKP-infected mice, we detected 5.8-, 5.3-, 4.2-, and 5.4-log reductions in the heart, liver, spleen, and kidney in combination groups (double dosing) compared to the PBS groups (fig. S50). Consistently, in the analysis of blood indicators, mice receiving combination therapy had the platelet (PLT), white blood cells (WBCs), and red blood cells (RBCs) within the normal ranges, while the control and monotherapy treatments were in abnormal status (Fig. 7, G to I). Thus, TINA increased the sensitivity of imipenem in vivo and broadly deterred resistance to a range of challenging NDM-1–producing CR Gram-negative bacteria.

We proceeded to conduct the histological analysis of lung injury. The histological examination of lung tissue revealed severe damage to the alveoli in mice treated with PBS or monotherapy of TINA and imipenem (Fig. 7, J and M). The airway wall exhibited noticeable infiltration of neutrophil cells and proliferation of goblet cells, accompanied by increased inflammatory secretions in the airway cavity. Both bronchial epithelial cells and lung cells displayed oval or irregularly shaped nuclei with abnormal signs of mitosis and the presence of vacuoles with varying diameters, indicating pathological phenomena. However, mice treated with TINA and imipenem did not exhibit evident vacuolation of epithelial cells. Instead, they displayed fewer symptoms of pulmonary fibrosis and lung dilatation, suggesting significant improvements in lung tissue damage resulting from the combined treatment. Bacterial clearance usually resulted in a decrease in local inflammation. We then investigated the inflammation effects of lung tissue following different treatments. We visualized that the signal nuclear factor κB (NF-κB) in the lungs was highly expressed along the bronchial tube wall in the control group but showed a significant reduction in combination therapy of TINA and imipenem (Fig. 7, K and N). These indicated that the combination treatments inhibited the activation of NF-κB protein and would lead to the reduction in the release of inflammatory factors and further improvement of lung injury (49, 50). We also accessed the levels of interleukin-1β (IL-1β), IL-6, and tumor necrosis factor–α (TNF-α) as indicators of biological inflammation in the lung (Fig. 7, L and O to Q, and figs. S53 and S54). IL-6, IL-1β, and TNF-α levels were significantly evaluated in the lungs of untreated mice. However, upon treatments with TINA and imipenem, these biomarkers decreased substantially relative to the control group, particularly in the TINA and imipenem combination group (P < 0.001). Thus, TINA restored imipenem activity in vivo, and their combination not only reduced bacterial loads in the CR bacteria-infected pneumonia but also decreased pulmonary inflammation to recover lung injury.

DISCUSSION

We leveraged the chemically engineered anionic adjuvants that deprive multiple divalent cations (Zn²⁺, Ca²⁺, and Mg²⁺) to overcome acquired and intrinsic resistance in NDM-1–producing Gram-negative ESKAPE bacteria. We tailored the dissociation speed of adjuvants against MBL toward efficient deprivation of divalent cations. We found the “100%-effective” TINA, which associated with MBL slowly to reach Zn²⁺ but dissociated quickly and deprived Zn²⁺ efficiently. Simultaneously, TINA competitively deprived Ca²⁺ and Mg²⁺ in LPS (Fig. 1). This was a notable example of using anionic adjuvants targeting multiple divalent cations to overcome both acquired and intrinsic resistance in superbugs. The anionic nature enabled TINA to resist serum absorption and exhibit high biosafety. Moreover, targeting multiple divalent cations lowered the resistance frequency. We did not observe new resistance during the continued use (Fig. 3G).

We used gold scaffolds in this study because of the feasibility in controlling the surface chemical properties (51). We reported a total of 18 anionic adjuvants rather than an individual case, suggesting that the chemical tailoring strategy would be general to other scaffolds. This class of adjuvants would be extended by expanding the scope of anionic ligands, scaffolds, and preparation parameters. In future studies, we will also investigate the potential of using adjuvants that have different interaction modes against MBL to overcome carbapenem resistance (Figs. 2B and 3B).

In conclusion, we reported a chemical strategy to develop new anionic adjuvants to tackle emerging antimicrobial resistance. This class of anionic adjuvants represents an innovative paradigm for combating carbapenem resistance and will extend the life span of carbapenems. The anionic adjuvants go beyond the traditional small molecules and cationic adjuvants against MBL resistance, guiding the chemical design of new antimicrobial agents with distinct mechanisms. The anionic characteristics enable the precise targeting of biological macromolecules, the fine regulation of interaction dynamics, and the significant enhancement in both specificity and biosafety. Our work paves the way for using the chemical biology approach to develop new anionic therapeutics with broad biomedical applications.

MATERIALS AND METHODS

Materials

N-(2-mercaptopropionyl)glycine (tiopronin), magnesium chloride (MgCl₂, 99.99%), calcium chloride anhydrous (CaCl₂), ertapenem, doripenem, faropenem, and nitrocefin were purchased from Aladdin (China). Hydrogen tetrachloroaurate (III) hydrate (HAuCl₄·3H₂O) was purchased from Sinopharm (China). Thiorphan and ertapenem were purchased from Yuanye Biotechnology (China). Sodium borohydride, imipenem, meropenem, and Calcein-AM/PI staining kits were purchased from Solarbio (China). Branch polyethylenimine and LPS (E. coli O111:B4 and K. pneumoniae) were purchased from Sigma-Aldrich (USA). Sodium chloride (NaCl) and potassium chloride (KCl) were purchased from Solarbio (China). Zinc sulfate heptahydrate and biapenem were purchased from Macklin (China). SYTOX Green nucleic acid stain and SYBR safe DNA gel stain were purchased from Invitrogen (USA). The ATP Assay Kit, 2,7-dichlorofluorescein diacetate (DCFH-DA) probe, BCECF AM probe, and LDH kit were purchased from Beyotime (China). Hepes buffer (1 M) was purchased from Coolaber (China). Cyclophosphamide was purchased from ACMEC (China). The Cell Counting Kit-8 (CCK-8) was purchased from DOJINDO Laboratories (Japan). Staining kits of H&E, IL-6, IL-1β, TNF-α, and NF-κB antibodies were purchased from Servicebio (China). Fetal bovine serum (FBS) was purchased from Gibco (USA).

Synthesis and characterization of the anionic adjuvants

We prepared ligand-protected anionic adjuvants in similar procedures. Briefly, HAuCl₄·3H₂O (1 mM) and penicillamine, thiorphan, 3-mercaptoisobutyric acid, tiopronin, or captopril (at concentrations of 1, 2, 4, and 10 mM, respectively) were dissolved in 10 ml of methanol in 100-ml flasks. The flasks were then placed in an ice bath for 20 min. Sodium borohydride was added dropwise to the stirred solution at a high speed (1500 rpm) for 15 min, followed by an additional 1 hour of stirring at 800 rpm. The resulting ligand-modified anionic adjuvants were dialyzed against double-distilled water [ddH₂O; molecular weight (MW) cutoff: 7000 to 14,000 Da, Solarbio] for 48 hours to remove untreated chemicals. These anionic adjuvants were then sterilized using a 0.22-μm filter (Millipore) and stored in the refrigerator for further use.

We calculate both the number of gold atoms and the MW. The number of gold atoms (N_Au) is determined as follows

$$\begin{array}{c}{N}_{\text{Au}}=\frac{V_{\text{NP}}\times \text{APF}}{V_{\text{Au}}}=\frac{\frac{4}{3}{\mathrm{\pi }r}_{\text{NP}}^3\times 0.7405}{\frac{4}{3}{\mathrm{\pi }r}_{\text{Au}}^3}=\frac{\left(\frac{d_{\text{NP}}}{2\text{nm}}\right)\times 0.7405}{{\left(0.144\text{nm}\right)}^3}={\left(\frac{d_{\text{NP}}}{\text{nm}}\right)}^3\times 31\hfill \end{array}$$

Here, V_NP is the volume of a sphere gold particle, V_Au is the volume of a single gold atom, APF is atomic packing factor, r_NP and d_NP are the radius and the diameter of the sphere gold particle, and r_Au is the radius of a gold atom. For penicillamine engineered adjuvants, d = 1.88 ± 0.49 nm, N_Au = (1.88)³ × 31 = 205.98, MW = N_Au × M_Au = 205.98 × 197 = 40,579 μg/μmol; for thiorphan engineered adjuvants, d = 4.38 ± 0.97 nm, N_Au = (4.38)³ × 31 = 2609, MV = N_Au × M_Au = 2609 × 197 = 514,161 μg/μmol; for 3-mercaptoisobutyric acid engineered adjuvants, d = 2.66 ± 0.49 nm, N_Au = (2.66)³ × 31 = 584.79, MV = N_Au × M_Au = 584.79 × 197 = 115,204 μg/μmol; for tiopronin engineered adjuvants, d = 2.91 ± 0.76 nm, N_Au = (2.91)³ × 31 = 761.24, MV = N_Au × M_Au = 761.24 × 197 = 149,963 μg/μmol; for captopril engineered adjuvants, d = 2.00 ± 0.26 nm, N_Au = (2.00)³ × 1 = 248.88, MV = N_Au × M_Au = 248.88 × 197 = 49,030 μg/μmol.

TEM images were captured by a double spherical aberration-corrected transmission electron microscope (Titan Themis G2). Nano-ZS Zetasizer dynamic light scattering (Malvern Instruments, UK) was used to measure the hydrodynamic diameter and zeta potential of the anionic adjuvants. The content of gold was quantitatively analyzed by ICP-MS (7700X, Agilent). Analysis of XPS was performed by Escalab 250XI (Thermo Fisher Scientific, UK). Ultraviolet (UV) absorption was determined by the UV-visible photometer (UV-2006i, Shimaszu). The molecular structure was analyzed by a Fourier transform infrared spectrometer (Spectrum One, PerkinElmer, USA).

Binding analysis by SPR

The binding assays for anionic adjuvants and small molecules to MBL (Macklin) were conducted using a Biacore 8K instrument (Cytiva) at 25°C. The CM5 sensor chip was activated using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide/N-hydroxysuccinimide coupling chemistry and immobilized with recombinant MBL at a concentration of 50 μg/ml in sodium acetate buffer (pH 4.5). To determine the binding affinity, MBL was immobilized in the channels to approximately 12,000 response units (RU), with a blank channel serving as the reference. Various anionic adjuvants modified with penicillamine, thiorphan, 3-mercaptoisobutyric acid, tiopronin, and captopril (prepared with a gold/ligand feeding ratio of 1:1), along with their corresponding small molecules, were injected at a flow rate of 30 μl/min in the PBS running buffer. The data were analyzed using Biacore Insight Evaluation software (version 3.0.12), and the equilibrium dissociation constants (K_off) were calculated using a steady-state affinity model.

In the antagonism binding experiments, TINA was preincubated with Zn²⁺ (at concentrations of 0, 100, 200, and 400 μM) at 4°C for 60 min before injection. All other parameters remained the same as described above.

Enzyme inhibition assay

MBL was dissolved in a buffer solution [20 mM Hepes and 100 mM NaCl (pH 7.4)] with varying concentrations of ligands and anionic adjuvants. The assay was initiated by adding nitrocefin to the mixture of MBL, ligands, and anionic adjuvants, resulting in a final enzyme concentration of 5 nM, a nitrocefin concentration of 30 μM, and different concentrations of ligands and anionic adjuvants. After 3 hours of incubation, absorbance was measured at 300 nm. IC₅₀ was determined as the half-maximal inhibitory concentration. Each measurement was performed in triplicate, and the individual values were used to calculate the SD.

Polymerase chain reaction

Overnight cultures of CR strains, including K. pneumoniae (CRKP), E. coli (CRE), A. baumannii (CRAB), and P. aeruginosa (CRPA), were diluted in LB medium and incubated overnight at 37°C with shaking at 225 rpm. To prepare the polymerase chain reaction (PCR) template, 200 μl of each overnight culture was centrifuged at 12,000 rpm for 1 min, and the supernatant was discarded. The resulting pellet was resuspended in 100 μl of ddH₂O and boiled at 100°C for 10 min in the presence of NP-40 (Solarbio, China), followed by cooling at 4°C for 30 min. After cooling, the suspension was centrifuged again at 12,000 rpm at 4°C for 10 min, and the supernatant, containing the PCR templates, was transferred to PCR tubes. The PCR reaction mixture (20-μl total volume) consisted of 10 μl of Fast Taq Plus Master Mix (Biosharp, China), 1 μl each of forward and reverse primers, 1 μl of the prepared template, and 7 μl of ddH₂O.

The primer sequences of E. coli and A. baumannii for NDM-1 were 5′-TGGCCCGCTCAAGGTATTTT-3′ (forward) and 5′-GTAGTGCTCAGTGTCGGCAT-3′ (reverse). The primer sequences of K. pneumoniae for NDM-1 were 5′-TTTGGCCCGCTCAAGGTATT-3′ (forward) and 5′-TTGATCAGGCAGCCACCAAA-3′ (reverse). The primer sequences of P. aeruginosa for NDM-1 were 5′-AACGGTTTGATCGTCAGGGA-3′ (forward) and 5′-GTCCAT-ACCGCCCATCTTGT-3′ (reverse). PCR amplification was carried out for 35 cycles. PCR products were analyzed by electrophoresis on a 2% agarose gel in 1× tris-acetate-EDTA (TAE) buffer at 120 V for 25 min. Positive PCR results indicated the presence of the target gene in the tested strains.

Bacteria strains and culture conditions

Clinical isolates, including CR Gram-negative ESKAPE bacteria (K. pneumoniae, A. baumannii, P. aeruginosa, and E. coli), and methicillin-resistant S. aureus (MRSA) are from local hospitals. The bacteria are cultured in LB liquid media and LB agar plates at 37°C.

The checkerboard assay and determination of FICI

The FICI between 23 adjuvants (18 nanoadjuvants and 5 ligands) and various carbapenems (imipenem, meropenem, biapenem, diorpenem, ertapenem, and faropenem) was measured using the checkerboard assay. A 96-well plate was prepared by adding 100 μl of LB medium to each well. Double dilutions of different antibiotics and nanoparticles were prepared, with adjuvants diluted along the Y axis and antibiotics along the X axis, respectively. Bacteria at a final concentration of 1 × 10⁵ CFU/ml were then added to each well. Following incubation at 37°C for 18 hours, absorbance at 600 nm was measured using a Spark microplate reader (TECAN, Switzerland). To determine the inhibitory effects of Zn²⁺ on the adjuvant effect of TINA, we used 10 μM Zn²⁺ as the working concentration.

Agarose gel electrophoresis

TINA and branhed polyethylenimine (bPEI; MW = 25,000 Da) at concentrations ranging from 3.13 to 200 μg/ml were dissolved in PBS containing 5% FBS and incubated at 37°C for 1 hour. Samples (10 μl of TINA or bPEI) were mixed with 2 μl of loading buffer and loaded onto the 1.5% agarose gel. PBS containing only 5% FBS served as the control. Gel electrophoresis was performed in 1× TAE buffer at 100 V for 40 min. Gels were stained with Coomassie blue to identify protein bands.

Time-dependent killing assay

Overnight cultures of CRKP, CRE, CRAB, CRPA, and MRSA were diluted 100-fold in LB medium and incubated at 37°C with shaking at 225 rpm for 4 hours to reach the logarithmic phase. Bacteria were then treated with TINA (20 or 10 μg/ml) or imipenem (2 or 8 μg/ml), alone or in combination. LB broth without drugs was used as a control. At 0, 1, 3, 6, 9, and 24 hours, the bacterial suspension was sampled, diluted in PBS, plated, and incubated overnight at 37°C. All measurements were performed in triplicate.

Resistance development assay

Log-phase CRKP was diluted in LB broth containing imipenem, with or without TINA, and cultured at 37°C. The MIC of the bacteria was determined. The culture was then diluted into fresh LB broth containing half the MIC of the drugs for the next passage, with in vitro passages repeated for 20 cycles. For imipenem, the 1× MIC was 64 μg/ml, while for the TINA and imipenem combination, 1× MIC was TINA (20 μg/ml) with imipenem (2 μg/ml).

Zn²⁺ restoration assays

MBL (4 μM), supplemented with 12 mM ZnSO₄, was incubated with 20 mM TINA for 15 min at 30°C. Following this, nitrocefin (30 mM) and ZnSO₄ at varying concentrations (500 nM to 40 mM) were added to reach a final volume of 200 μl, and absorbance at 490 nm was continuously monitored using a Spectramax reader (Molecular Devices) for 30 min at 30°C. Percent residual activity was calculated relative to the control without TINA. Any slightly negative values of percent residual activity were recorded as 0.

ITC analysis

ITC was performed using the MicroCal PEAQ-ITC Automated System (Malvern, England). ZnSO₄, TINA, and MBL were diluted in 10 mM PBS buffer (Gibco, pH 7.4). Each experiment involved titrating 25 injections of 0.5 mM ZnSO₄ into a sample cell containing either 0.4 mM TINA or 2 mM MBL. Each injection of 2 μl was administered at intervals of 400 s. Titrations were conducted with a stirring speed of 300 rpm at 25°C. The protein-to-buffer titration background was subtracted from the raw data to exclude the heat of dilution effects. To measure the binding affinity of LPS and TINA for Ca²⁺ and Mg²⁺, we injected 0.5 mM CaCl₂ or MgCl₂ into 0.5 mM LPS or TINA, respectively. ITC data were analyzed using MacroCal ITC Analysis Software (version 1.0), calculating adsorption strength and thermodynamic parameters. The data were fitted to a standard model.

Michaelis-Menten kinetics

The inhibition constant (K_i) was determined through the nonlinear fitting of the initial rates (V₀) of imipenem hydrolysis at varying concentrations (100, 200, 400, 800, and 1600 μM) by 1 nM MBL, in the absence or presence of three concentrations of TINA (0, 10, 80, and 200 μg/ml). The best fits were achieved using the competitive inhibition model, described by the equation V₀ = (V_max × [S])/(K_m × (1 + [I]/K_i) + [S]). K_m, [I], and [S] represented the concentration of TINA and imipenem, respectively. The kinetic parameters K_m/V_max/K_cat were determined by fitting the data into double reciprocal Lineweaver-Burk plots. GraphPad Prism software (version 9.5.1) was used for comprehensive data analysis.

Zn²⁺ displacement analysis

MBL (10 μM) was incubated with 25 μM ZnSO₄ in dialysis buffer [20 mM Hepes and 100 mM NaCl (pH 7.4)] overnight at 4°C. Following this, MBL was incubated with various concentrations of TINA by dialysis at 4°C for 12 hours with gentle shaking. The samples were then dialyzed in fresh dialysis buffer to remove unbound metal ions and subsequently acidified with concentrated HNO₃ at 60°C for 1 hour. The dialysate and original solutions were then analyzed using ICP-MS (7700X, Agilent).

To compare the Zn²⁺ deprivation efficiency among adjuvants with different K_off values, we incubated the MBL with adjuvants at a concentration of 20 μg/ml. The percentage of retained Zn²⁺ in MBLs after coincubation with adjuvants was calculated by dividing the Zn concentration in MBLs incubated with adjuvant by that in MBLs without adjuvant incubation. All measurements were performed in triplicate.

Antagonism assays between TINA and LPS and associated divalent cations

TINA (10 μg/ml), imipenem (8 μg/ml), and varying concentrations of LPS (0, 2, 4, 8, 16, 32, 64, and 128 μg/ml) were added to 96-well plates. CRKP bacteria at a final concentration of 1 × 10⁵ CFU/ml were then distributed into each well. After a 24-hour incubation at 37°C, absorbance at 600 nm was measured for each well using a microplate reader. In addition, the effects of different cations, including NaCl, KCl, CaCl₂, and MgSO₄, on the antibacterial activity of TINA and imipenem against CRKP were evaluated under similar conditions.

Measurement of LPS and divalent cations (Ca²⁺ and Mg²⁺) in bacterial supernatant

Overnight cultures of CRKP were washed three times with Ca²⁺ and Mg²⁺-free PBS. Bacterial cells (1 × 10⁵ CFU/ml) were then treated with the TINA (20 μg/ml), imipenem (8 μg/ml), and combinations of TINA plus imipenem (20 + 2 μg/ml and 10 + 8 μg/ml) in the same buffer for 12 hours at 37°C. The group that bacterial cells without treatments was set as the blank control. After incubation, we obtained the bacterial supernatant by centrifugation the bacterial solution at 7000 rpm for 5 min. The LPS concentration in bacterial supernatant was detected using the Chromogenic LAL Endotoxin Assay Kit (catalog no. C0276S, Beyotime, China) according to the manufacturer’s instructions. The concentration of Ca²⁺ and Mg²⁺ in the bacterial supernatant was detected by ICP-MS. The experiment was conducted in triplicate and repeated three times.

Outer membrane permeability assays

Overnight cultures of CRKP were washed three times in Hepes buffer (5 mM glucose). Bacterial pellets at a concentration of 1 × 10⁵ CFU/ml were resuspended in Hepes buffer containing NPN (10 μM) and incubated for 30 min at 37°C in the dark. TINA (20 μg/ml), imipenem (8 μg/ml), and combinations of TINA plus imipenem (20 + 2 μg/ml and 10 + 8 μg/ml) were added to the probe-labeled bacterial cells and coincubated for 3 hours. The negative control consisted of NPN with a bacterial suspension lacking TINA and imipenem. Fluorescence intensity was measured using a microplate reader at an excitation wavelength of 350 nm and an emission wavelength of 420 nm. Each experiment was conducted in triplicate and repeated three times.

Live and dead imaging

The LIVE/DEAD BacLight Bacterial Viability Kit (catalog no. L7012, Invitrogen) containing SYTO 9 and PI was used to assess the viability of bacteria treated with TINA and imipenem. Bacterial cells were cultured to the logarithmic phase in LB broth at 37°C and 200 rpm [OD₆₀₀ (optical density at 600 nm) = 0.5] and then diluted to 1 × 10⁷ CFU/ml in LB. The bacterial suspension was treated with PBS, TINA, imipenem, and combinations of TINA plus imipenem (20 + 2 μg/ml and 10 + 8 μg/ml) for 6 hours. Samples were visualized using confocal laser scanning microscopy (SP8, Leica, Germany).

Visualization of bacterial morphology by scanning electron microscopy

CRKP bacteria in the log phase were resuspended to an OD₆₀₀ of 0.5. They were coincubated with PBS, TINA, imipenem, and combinations of TINA plus imipenem (20 + 2 μg/ml and 10 + 8 μg/ml) for 24 hours. The bacterial suspension was then fixed in 2.5% glutaraldehyde for 3 hours and washed with PBS. For dehydration, bacteria were treated every 15 min with an ethanol gradient (30, 50, 70, 80, 90, 95, and 100%). Images were taken by scanning electron microscopy (Hitachi S4800, Japan).

Visualization of bacteria sections by TEM

CRKP bacteria were grown to the logarithmic phase at 37°C and 200 rpm until they reached an OD₆₀₀ of 0.5. The bacterial suspension was prepared by adding PBS, TINA, imipenem, TINA plus imipenem (20 + 2 μg/ml), or TINA plus imipenem (10 + 8 μg/ml) and incubated for 24 hours. After incubation, the bacterial precipitates were collected by centrifugation and stored at 4°C in a fixative solution suitable for electron microscopy. The bacteria were embedded in 1% agarose solution, rinsed, and fixed with 0.1 M phosphate buffer (pH 7.4). After dehydration at room temperature and embedding, ultrathin sections were prepared. The sections were stained with 2.6% lead citrate solution saturated with 2% uranyl acetate, and images were captured using the bio-TEM (HT7700, Japan).

Membrane depolarization assay

The membrane potential (∆Ψ) of CRKP treated with TINA and imipenem was determined using DiSC₃(5). Exponentially growing CRKP was washed three times in Hepes buffer (5 mM glucose) to obtain a bacterial suspension (OD₆₀₀ = 0.5). DiSC₃(5) was mixed with the bacterial suspension and incubated at 37°C in the dark for 30 min. Then, 100 μl of the DiSC₃(5)-treated bacterial mixture was added to 96-well plates along with 100 μl of PBS, TINA, imipenem, TINA plus imipenem (20 + 2 μg/ml), or TINA plus imipenem (10 + 8 μg/ml) and incubated for 1 hour. Fluorescence intensity was measured at an excitation wavelength of 622 nm and an emission wavelength of 670 nm using a Spark microplate reader (TECAN, Switzerland). The assay was performed in triplicate and repeated three times.

Transmembrane proton gradient assay

The transmembrane proton gradient (∆pH) was assessed using the pH-sensitive fluorescence probe BCECF-AM. CRKP was cultured overnight in LB at 37°C. The bacteria were harvested and resuspended in 2 μM BCECF-AM at a concentration of 5 × 10⁷ CFU/ml and then incubated at 37°C for 30 min. After washing three times with PBS to remove noninternalized BCECF-AM, 1 ml of PBS, TINA, imipenem, TINA plus imipenem (20 + 2 μg/ml), or TINA plus imipenem (10 + 8 μg/ml) was mixed with CRKP. Following a 3-hour incubation, fluorescence intensity was measured at an excitation/emission wavelength of 500/622 nm. The assay was performed in triplicate and repeated three times.

Efflux pump inhibition assay

CRKP bacteria were cultured to the logarithmic phase, washed, and resuspended in PBS. The bacterial suspension (1 × 10⁶ CFU/ml) was incubated with EtBr (16 μg/ml) for 30 min in the dark at 37°C. A mixture of 100 μl of EtBr-treated bacteria and 100 μl of PBS, TINA, imipenem, TINA plus imipenem (20 + 2 μg/ml), or TINA plus imipenem (10 + 8 μg/ml) was added to 96-well plates. Fluorescence intensity was measured every 2 min using a microplate reader with excitation and emission wavelengths of 530 and 600 nm, respectively. The assay was performed in triplicate and repeated three times.

Measurement of intracellular ROS

ROS was measured using DCFH-DA to assess ROS levels induced by TINA in CRKP. CRKP cells were cultured to the logarithmic phase in fresh LB medium (OD₆₀₀ = 0.5) and then resuspended in PBS at a concentration of 1 × 10⁸ CFU/ml after washing. The bacterial suspension was incubated with 10 μM DCFH-DA for 30 min. After centrifugation at 6000 rpm for 3 min and washing three times with PBS, the bacteria were treated with TINA at concentrations of 0, 2.5, 5, 10, 20, and 40 μg/ml for 6 hours in the dark. The fluorescent intensity of DCFH-DA was measured with excitation and emission wavelengths of 485 and 525 nm, respectively.

Detection of ATP concentration

Log-phase CRKP cells (1 × 10⁸ CFU/ml) were incubated with TINA at concentrations of 0, 2.5, 5, 10, and 20 μM for 3 hours at 37°C. The bacteria were then centrifuged and harvested in lysate for ATP measurement. Cell viability was assessed using the ATP Assay Kit (Beyotime, China). A standard curve was constructed using seven concentrations of ATP ranging from 0.006 to 0.36 μM.

LDH release assay

Overnight-cultured CRKP bacterial suspension was resuspended in PBS and added to 96-well plates at a final concentration of 1 × 10⁸ CFU/ml. PBS (positive control), TINA, imipenem, TINA plus imipenem (20 + 2 μg/ml), and TINA plus imipenem (10 + 8 μg/ml) were incubated for 3 hours. One hour before the end of the incubation, 2 μl of LDH release reagent was added to the PBS (positive control group). The LDH activity was determined by measuring the absorbance at 490 nm using the LDH Release Assay Kit (Beyotime, China) according to the manufacturer’s protocol.

Proteomic analysis

To perform proteomic analysis of the combination of TINA and imipenem-induced protein expression profiles, we treated overnight-cultured CRKP cells (~10⁹ CFUs) with imipenem (8 μg/ml) and the combination of TINA (10 μg/ml) and imipenem (8 μg/ml) at 37°C with shaking for 6 hours (n = 3). The data-independent acquisition (DIA) label-free quantitative proteomic analysis was performed at the National Center for Protein Science (Beijing, China). For protein extraction, bacterial samples were sufficiently lysed with the protein lysis buffer (containing 8 M urea and protease inhibitors), and the protein concentration of the samples was determined using a bicinchoninic acid assay. For enzymatic digestion and desalting, 10 mM dithiothreitol was added to the protein extract and incubated at 56°C for 1 hour; after returning to room temperature, 55 mM indole-3-acetic acid was added for incubation at room temperature in the dark for 45 min. On the basis of the protein concentration results, an appropriate amount of protein from each sample was subjected to enzymatic digestion using the filter-aided sample preparation method. On the following day, the ultrafiltration tube was centrifuged at 13,000g at room temperature for 10 min, the liquid in the collection tube was transferred to a new centrifuge tube and dried by vacuum heating, and lastly, the samples were desalted using a C18 desalting column. For mass spectrometry detection, the chromatographic mobile phases consisted of phase A [0.1% formic acid (FA)] and phase B (80% acetonitrile containing 0.1% FA). Lyophilized peptides were fully dissolved in phase A (0.1% FA) and centrifuged at 16,000g for 15 min; the supernatant was added to the inner tube and placed in the automatic sampler. Samples were introduced into a C18 analytical column (150-μm inner diameter, 25-cm length) via a UHPLC3000 liquid chromatograph (Thermo Fisher Scientific, USA) at a flow rate of 3 μl/min for elution, with liquid chromatography elution performed at 300 nl/min and phase B linearly increasing from 8 to 50% within 60 min. A Thermo Fisher Scientific Orbitrap Eclipse mass spectrometer equipped with a Nanospray Flex ion source was used, with an ion spray voltage of 2.2 kV and an ion transfer tube temperature of 320°C. The mass spectrometer operated in DIA mode, with a full scan resolution of 60,000, a scanning range of 350 to 1150 mass/charge ratio, and autoset maximum injection time; tandem mass spectrometry scans had a resolution of 30,000, 30 scanning windows, and a collision energy of 32%. DIA data were searched using Spectronaut 17.4 software (Biognosys) against the UniProt K. pneumoniae database, with parameters including trypsin digestion (maximum of two missed cleavages), fixed modification of carbamidomethyl (C), variable modifications of oxidation (M) and acetyl (N-terminal), peptide and protein false discovery rate < 1.0%, and each protein identified by at least one specific peptide.

Cell lines and culture

HUVECs and mouse embryonic fibroblasts (3T3) cell lines were cultured (37°C, 5% CO₂) in Dulbecco’s modified Eagle’s media (DMEM, Gibco) supplemented with 10% (w/v) FBS and 1% (w/v) penicillin-streptomycin (Gibco).

Cell cytotoxicity assay

The cytotoxicity of TINA was evaluated using the CCK-8 assay. HUVECs and 3T3 cells were cultured in DMEM medium supplemented with 10% FBS and 5% penicillin-streptomycin. The cells were seeded at a density of 8 × 10³ cells per well in 96-well plates and incubated at 37°C for 24 hours. Following this, the cells were treated with varying concentrations of TINA (0, 1.56, 3.13, 6.25, 25, 50, 100, 200, and 400 μg/ml) for 24 hours. Absorbance of 450 nm was measured to calculate the percentage of cell viability.

In vitro cell viability assay

To evaluate cell viability and mortality after TINA treatment, HUVECs and 3T3 cells were stained with Calcein-AM and PI. The cells were treated with TINA (100 μg/ml) for 12, 24, and 48 hours. After treatment, cells were visualized using confocal laser scanning microscopy (SP8, Leica, Germany).

Hemolysis assay

The hemolysis assay was conducted to confirm the blood biocompatibility of TINA. Blood from BALB/c mice was collected and centrifuged at 1500 rpm for 15 min. The supernatant was discarded, and the cells were resuspended in NaCl and centrifuged again for 15 min at 1500 rpm, with this process being repeated twice. The RBCs were then diluted to a 4% suspension in 0.9% NaCl. TINA at concentrations of 6.25, 12.5, 25, 50, 100, 200, and 400 μg/ml was added to the suspension. Water plus red blood cell suspension served as a positive control, and 0.9% NaCl plus red blood cell suspension served as a negative control. After incubation at 37°C for 2 hours, absorbance was measured at 540 nm. The assay was repeated three times.

Animals

All animal experiments are conducted with the permission of the Animal Ethics Committee of the Southern University of Science Technology (SUSTech-JY202407004) and Yishang Biotech (IACUC-2024-Mi-205). Female Blab/c mice (6 to 8 weeks old) weighed between 15 and 20 g and were purchased from Guangdong Medical Laboratory Animal Center or Shanghai Yishang Biotechnology Co. Ltd. Animal handling and experimental protocols adhered to the relevant national and international guidelines and regulations.

Biodistribution of TINA

To investigate the biodistribution of TINA following intraperitoneal and intranasal injection, TINA was administered to BALB/c mice (n = 3) at doses of 10 mg/kg (intranasal) and 50 mg/kg (ip). Organ samples (heart, liver, spleen, lung, and kidney) were collected 24 hours postadministration, weighed, and lysed overnight with 5 ml of fresh aqua regia. The tissue samples were then heated to completely evaporate the aqua regia and redissolved in 5 ml of 1% (v/v) HNO₃ by sonication. Gold concentration and the percentage of the injected dose per gram of organ (%ID/g) were measured using ICP-MS.

In vivo biosafety assay

An in vivo biocompatibility assay was performed using female BALB/c mice (6 to 8 weeks old). Mice were randomly assigned to three groups: (1) PBS (n = 3), (2) TINA–48 hours (n = 3), and (3) TINA–72 hours (n = 3). In groups 2 and 3, mice were intraperitoneally injected with TINA at 50 mg/kg. After indicated time points postinjection of TINA, mice were euthanized by cervical dislocation, and blood samples and major organs were collected for biochemical and H&E analysis.

Models of bacterial pneumonia

The mouse model of pulmonary bacterial infection was established in female BALB/c mice to evaluate the antimicrobial activity of TINA and imipenem. Neutrophil reduction was induced by intraperitoneal injection of cyclophosphamide twice (150 mg/kg at 4 days before infection and 100 mg/kg at 1 day before infection). Mice were unbiasedly challenged with bacteria and divided into five groups: (1) control (PBS) group; (2) TINA group; (3) imipenem group; (4) TINA plus imipenem (single dose) group; and (5) TINA plus imipenem (double dose) group. For the lethal high-dose CRKP infection model, BALB/c female mice (n = 10 per group) received 5 × 10⁷ CFU. Thirty minutes postinfection, mice were treated as (1) PBS, (2) TINA (10 mg/kg, single dose), (3) imipenem (8 mg/kg, single dose), (4) TINA plus imipenem (10 + 8 mg/kg, single dose), or (5) TINA plus imipenem (10 + 8 mg/kg, double dose). Survival was monitored for 7 days postinfection, and cumulative survival was calculated. In the acute pneumonia model, mice were intranasally infected with 1 × 10⁵ CFU of CRKP, CRE, or CRAB. Thirty minutes after infection, mice were treated with (1) PBS, (2) TINA (10 mg/kg, single dose), (3) imipenem (8 mg/kg, single dose), (4) TINA plus imipenem (10 + 8 mg/kg, single dose), or (5) TINA plus imipenem (10 + 8 mg/kg, double dose). Twenty-four hours after infection, each mouse was euthanized by cervical dislocation, and blood samples were collected for biochemical tests. Hearts, livers, spleens, lungs, and kidneys were harvested under aseptic conditions. Parts of these organs were washed with sterile PBS and fixed in 10% formaldehyde for 24 hours for morphological observation, while other portions were homogenized, serially diluted, and plated on LB agar to quantify CFU bacterial load.

H&E staining assay

Organ tissues (heart, liver, spleen, lung, and kidney) were fixed in paraformaldehyde, embedded in paraffin, and sectioned. The sections were dewaxed using graded alcohols, stained with hematoxylin for 3 min, and then differentiated twice with 1% hydrochloric acid for 2 s each time. After rinsing with running water for 5 min, the sections were stained with eosin for 5 min, dehydrated in graded alcohols, cleared in xylene for 15 min, and lastly sealed with neutral resin.

Immunofluorescence analysis

For immunofluorescence, the sections were first dewaxed and rehydrated, then followed by blocking using bovine serum albumin (45 min) at room temperature. Next, the slides were incubated overnight at 4°C with primary antibodies (NF-κB, GB119977). Afterward, they were incubated for 50 min at room temperature in the dark with a Cyanine 3–labeled goat anti-rabbit immunoglobulin G (IgG; GB21303, 1:100 dilution) secondary antibody. After washing with PBS, the samples were stained with 4′,6-diamidino-2-phenylindole.

Immunohistochemistry staining

The tissues were stained with specific antibodies directed against TNF, IL-6, and IL-1β to analyze the expression of TNF, IL-6, and IL-1β in the inflammatory tissue [anti–TNF-α antibody (1:100 dilution), anti–IL-6 antibody (1:100 dilution), anti–IL-1β antibody (1:100 dilution)]. Then, the slices were incubated with horseradish peroxidase (HRP)–conjugated secondary antibody goat anti-rabbit IgG H&L (HRP) (1:800 dilution) at 37°C for 1 hour. The sections were further stained using the 3,3ʹ-Diaminobenzidine substrate and then counterstained with hematoxylin. The average optical density of all images was measured by Image-Pro Plus.

Statistical analysis

Statistical analyses were performed using GraphPad Prism Software. Data were presented as the means ± SD with individual data points. The in vitro experiments were repeated at least three times with at least three repetitions for each group. The statistical differences were analyzed using a two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Supplementary Materials

This PDF file includes:

Figs. S1 to S54

Tables S1 and S2