Nano-NucleOTAC-mediated Z-DNA transformation disrupts mycobacterium tuberculosis biofilms and activates antibacterial immunity

Abstract

The biofilms of Mycobacterium Tuberculosis (Mtb) use extracellular DNA (eDNA) as a protective structural scaffold that shields resident bacteria from host immunity and antibiotics, promoting persistent infection. An FDA-approved DNase can partially hydrolyze eDNA but has limited affinity and nucleolytic activity against Z-form DNA (Z-DNA). After confirming the abundance of Z-DNA in Mtb biofilms, we develop Nano-NucleOTAC, a nano-nucleolysis targeting chimera that combines an eDNA-binding ligand (lactoferrin), a DNA intercalator (chloroquine) and DNase. Nano-NucleOTAC anchors to Z-DNA in Mtb H37Ra biofilms, enabling effective biofilm degradation via Z-DNA transformation and enzymatic cleavage. Following treatment, dispersed Mtb H37Ra clusters induce neutrophil extracellular trap (NET) formation and cytokine secretion (TNF, IL-1β, and IL-10) to recruit macrophages, while NET remodeling and nucleolysis enhance macrophage bactericidal effect vs. DNase alone. In mouse lung infection models, Nano-NucleOTAC improves lung drug retention, reduces pathology, and reactivates host immune responses. It also demonstrates high efficacy against Mtb strains of varying virulence (H37Ra, H37Rv) and non-tuberculous mycobacteria (Mycobacterium smegmatis). In clinical sputum samples, Nano-NucleOTAC potentiates rifampicin efficacy, and enhances neutrophil and macrophage antibacterial immunity in bronchoalveolar lavage fluid (BALF) and pus. This safe, versatile nucleic-acid-targeting strategy modulates bacterial biophysical barriers and host immunity to address biofilm-associated multidrug-resistant infections.

Extracellular DNA (eDNA) from tuberculosis can protect bacteria from host immune responses and antibiotics and can contain Z form eDNA which is harder for DNase to hydrolyze. Here the authors use a nanoparticle system to more specifically target eDNA and the Z form and show cleavage of Z-DNA and improvement of immune responses against TB infection in mice.

Introduction

Bacterial biofilms are intricate microbial communities embedded in a self-produced matrix of extracellular polymeric substances (EPS)^1–4. This matrix, rich in extracellular DNA (eDNA), proteins, polysaccharides and lipids, plays a pivotal role in protecting bacteria from antibiotics and host immune response^5,6. The highly pathogenic Mycobacterium tuberculosis (Mtb), responsible for life-threatening tuberculosis (TB), is known to form dense biofilms within human pulmonary granulomas, contributing to the TB persistence and resurgence⁷. DNase, approved by the FDA for degrading eDNA in biofilms^8,9, faces limitations due to insufficient biofilm penetration and incomplete eDNA degradation. Concurrently, eDNA comprises both Z-form and B-form DNA (B-DNA), and it is noteworthy that the DNase-resistant Z-form DNA (Z-DNA) accumulates as biofilms mature^4,10. The steric hindrance posed by Z-DNA significantly impedes DNase from accessing and cleaving eDNA¹¹. This Z-form DNA confers nuclease resistance to the biofilm structures, and thereby substantially contributes to the persistence of eDNA-dependent bacterial biofilms⁴. Therefore, overcoming the challenge of degrading Z-DNA and eradicating established biofilms is crucial for combating refractory TB infections.

During TB infection, the host immune response mediated by neutrophils and macrophages is critical for infection control. Neutrophils, as the front line of defense, are rapidly recruited to infection sites, where they undergo NETosis and release neutrophil extracellular traps (NETs) that capture and immobilize Mtb^12,13. These NETs are composed of DNA and peroxidase (an antibacterial protein)¹⁴. The reticular DNA structure captures bacteria through charge interactions, and peroxidase then kills the trapped bacteria¹⁵. In addition, NETs act as pseudo-opsonins to enhance phagocytic activity, thereby promoting the engulfment of peroxidase from NETs by macrophages to efficiently kill ingested bacteria¹⁶. However, it is worth mentioning that NETs also contain Z-DNA structures, which are resistant to DNase degradation^17,18. This resistance hinders the breakdown of NETs into fragments and obstructs the bactericidal activity of macrophages. Therefore, effective degradation of Z-DNA is crucial for eliminating established biofilms and modulating the host immune response against drug-resistant Mtb. Since DNase-resistant Z-DNA stabilizes biofilms and enables bacterial immune evasion, there is an urgent need for technologies to target and degrade Z-DNA. Targeting chimeras (TACs) consists of a targeting moiety that directs them to specific molecular targets, coupled with an effector domain that mediates degradation, making them a powerful tool to precisely target and eliminate harmful substances.

In this work, we design Nano-NucleOTAC, a nano-nucleolysis targeting chimera technology. Utilizing this technology, DNase-chloroquine nanoparticles (DCNP) are designed by loading DNase (a DNA-degrading enzyme), chloroquine (a DNA-intercalating agent), with disulfide bonds-remodeled lactoferrin (LF, the chimeric ligand) (Supplementary Fig. S1). The efficacy of DCNP treatment is assessed by evaluating the physical characteristics of biofilms, including biofilm eradication, DNA ratio, biofilm thickness, deformation stress, and bacterial dispersion from the biofilm. In vitro and in vivo immunobactericidal effects after DCNP-mediated Z-DNA transformation are evaluated by NET production, pathogen capture and macrophage reactivation. The presence of Z-DNA in Mtb-infected mouse lungs and clinical samples (sputum, bronchoalveolar lavage fluid (BALF), and pus) is confirmed by immunofluorescence microscopy. When treated with DCNP, these samples show enhanced bactericidal sensitivity to anti-TB drugs and display aggravated innate immune response against Mtb, demonstrating significant potential for enhancing antibiotic efficacy and antibacterial immunity. Thus, our Nano-NucleOTAC technology may provide a promising therapeutic solution for multidrug-resistant and biofilm-related Mtb infections.

Results

Z-DNA discovery in Mtb biofilms

Z-DNA acts as a key structural element within biofilms of Escherichia coli, Klebsiella pneumoniae, and Haemophilus influenzae, enhancing their resilience to both immune defenses and antibacterial treatments, thereby promoting recurrent infections^19–22. However, it is still unknown whether the presence of Z-DNA in biofilms or infected tissues plays a pivotal role in refractory TB infection. Studies have shown that, in severe cases, high concentrations of Z-DNA may indicate high bacterial load and rapid TB spread^4,23. Hence, we evaluated the presence of both B-DNA and Z-DNA in cultured biofilms, infected mouse lungs and clinical specimens. Interestingly, Mtb H37Ra biofilm formation was initiated on day 4, and both B-DNA and Z-DNA were found to be abundant within the extracellular polymeric substances (EPS) of the Mtb H37Ra biofilms (Fig. 1A). Biofilm thickness and the content of B-DNA/Z-DNA increased rapidly, reaching a plateau on day 10. Of note, the mature biofilm on day 10 showed an obvious presence of Z-DNA (Fig. 1B, C and Supplementary Fig. S2). These results confirmed the presence of Z-DNA within Mtb H37Ra biofilms and showed that the Z-DNA content increased over time, achieving a stable level in vitro. Subsequently, eDNA composition in Mtb H37Ra-infected mice lungs was investigated by immunofluorescence staining. As shown in Fig. 1D, the lung section of Mtb H37Ra-infected mouse displayed abundant B-DNA and Z-DNA, confirming the presence of Z-DNA after intranasal administration of Mtb H37Ra.

Fig. 1. Presence of Z-DNA in all tested Mycobacterium tuberculosis biofilms, involving Mtb H37Ra-infected mouse lung and clinical samples (sputum, bronchoalveolar lavage fluid (BALF) and pus).

A Mtb H37Ra biofilms were stained with Calcofluor White (CW) (white), antibodies against B-DNA (blue) and Z-DNA (yellow), and visualized by confocal laser scanning microscopy (CLSM). B, C Trends of biofilm thickness and relative fluorescence intensity of B-DNA and Z-DNA (n = 3 independent samples; mean ± s.d.). D Immunofluorescence staining of B-DNA (blue) and Z-DNA (yellow) in Mtb H37Ra-infected mouse lung. Scale bars = 50 µm. E Immunofluorescence staining of B-DNA (blue) and Z-DNA (yellow) in clinical samples (sputum, BALF and pus of Mtb-infected patients). Scale bars = 10 µm. F Z-DNA/eDNA ratio in Mtb H37Ra-infected mouse lung and clinical samples (sputum, BALF and pus) (n = 3 independent mice or clinical samples; mean ± s.d.). Source data are provided as a Source Data file.

The presence of Z-DNA within the Mtb infected clinical samples, such as sputum, BALF, and pus, was further evaluated by immunofluorescence. As depicted in Fig. 1E, we identified Z-DNA in sputum, bronchoalveolar lavage fluid (BALF), and purulent samples from patients with pulmonary tuberculosis. This finding parallels the observations by Buzzo et al., published in Cell, who detected substantial amounts of Z-DNA in the sputum of patients infected with Staphylococcus aureus (MSSA) and Burkholderia cenocepacia. Our results demonstrated that the presence of Z-DNA in sputum was found to be higher than the Mtb H37Ra-infected mice lungs, BALF and pus samples (Fig. 1F). Despite this observation in Mtb patient’s sputum, the infected mouse lung showed low levels of Z-DNA, suggesting differential occurrence of Z-DNA in infected tissues. Overall, our results demonstrate that Z-DNA exists in the biofilms of various Mtb samples.

DCNP-mediated biofilm disruption and Z-DNA elimination in Mtb H37Ra

FDA-approved DNase can hydrolyze eDNA matrix but lacks high affinity and nucleolytic activity against Z-DNA^8,9,24. Utilizing Nano-NucleOTAC technology, we synthesized DCNP with eDNA structure-switching functions. DCNPs loaded with DNase and chloroquine enhanced biofilm targeting and nucleic acid degradation. The DCNP exhibited cationic surface potential (+ 27.3 mV) with a size distribution of 37.4 ± 3.5 nm (Supplementary Fig. S3), and showed a satisfactory batch-to-batch consistency (coefficient of variation < 10) and long-term particle stability (Supplementary Table S1). In addition, DCNP with a DNase loading efficiency of 21.8% showed redox-responsive drug release in the presence of extracellular glutathione (GSH, about 500 μM in infected lung)²⁵ (Supplementary Fig. S4). Chloroquine-free nanoparticles (DNP) were prepared as a control, whose surface potential and size distribution were comparable to DCNP (Supplementary Fig. S5). To investigate biofilm anchoring and structural transformation of eDNA, biofilms were stained with CW (red) and incubated with DCNP. Both DCNP and DNP achieved effective accumulation in biofilms (Fig. 2A). Moreover, DCNP demonstrated marked co-localization with the biofilm-associated bacteria and substantially higher accumulation (~ 90%) than free DNase (Fig. 2B). In DNase-treated biofilm, only 4% of the free DNase was retained in the biofilm, with the majority distributed on the surface (Supplementary Fig. S6). As shown in Fig. 2C, DCNP-embedded biofilm anchoring was achieved through a comprehensive hydrogen bonding network between lactoferrin and eDNA. Key lactoferrin residues, including R21, K282, S283, Q287, and R296, were involved in these intricate interactions. R21 formed a crucial hydrogen bond with DT20, while the hydrogen bonding between DA14 and K282 enhanced the stability of the complex. Notably, the hydrogen bonding of S283 and S303 with their DNA counterparts reinforced the structural integrity of the assembly. The hydrogen bonding of Q287 with DC13 and R296 with DA12 further strengthened the intermolecular hydrogen bond network within the complex. Finally, the affinity between eDNA and DCNP was − 15.7 kcal/mol, stronger than the conventional binding strength of antibody-antigen interactions (~ − 12 kcal/mol)²⁶. To validate the efficacy of DCNP in transforming Z-DNA and disrupting biofilm structure, biofilm thickness and eDNA components were quantified by confocal microscopy. As shown in Fig. 2D, E and Supplementary Fig. S7, free DNase removed only 35% of B-DNA from the eDNA, resulting in a 20% reduction in biofilm thickness. The eDNA-anchoring chimeric DNP slightly promoted B-DNA nucleolysis (50%). Particularly, DCNP with its special biofilm penetration and nucleic acid targeting abilities, efficiently disintegrated the biofilm (91% clearance rate) and achieved a marked reduction in both B-DNA and Z-DNA. The results were further validated by crystal violet staining (Supplementary Fig. S8).

Fig. 2. Binding, mechanical and transcriptomic analysis of DCNP (NucleOTAC) treated biofilms.

A Confocal images demonstrating the targeted accumulation of DCNP in Mtb H37Ra biofilms. DNase alone or DNase encapsulated in DCNP or DNP were labeled with fluorescein (green), and biofilms were stained with CW (red). B Biofilm accumulation ability of DCNP was analyzed by measuring the fluorescence intensity at different depths of biofilm. Dotted lines in panel (A) represent the region of interest chosen for the biofilm accumulation analysis. C Schematic representation of hydrogen bond-mediated DCNP (Nano-NucleOTAC) embedded biofilm anchoring. Residues involved in eDNA anchoring were shown as cyan sticks. D Biofilm thickness attenuation and structural component variation after DCNP treatment. The biofilms were stained with CW (white), antibodies against B-DNA (blue) and Z-DNA (yellow). E Fluorescent intensity of B-DNA and Z-DNA within biofilms after DCNP treatment (n = 3 independent samples; mean ± s.d.). B-DNA: Overall, F = 80.615, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: mean difference (MD) = − 34.000, 95%CI (− 52.08–− 15.92), adjusted P = 0.0010; DNP vs. Control: MD = − 50.000, 95%CI (− 68.08–− 31.92), adjusted P < 0.0001; DCNP vs. Control; MD = −86.333, 95%CI (− 104.42–− 68.25), adjusted P < 0.0001]. Z-DNA: Overall, F = 267.099, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: MD = − 10.000, 95%CI (− 21.02–− 1.02), adjusted P = 0.0759; DNP vs. Control: MD = − 15.667, 95%CI (− 26.68–− 4.65), adjusted P = 0.0081; DCNP vs. Control; MD = − 87.000, 95%CI (− 98.02–− 75.98), adjusted P < 0.0001]. F Graphical illustration of B-DNA to Z-DNA transformation followed by nucleolysis and decreased deformation stress (DS). Created in BioRender. Deliang, L. (2026) https://BioRender.com/5wi4wtq. G DS analysis of DCNP-treated biofilms measured after Z-DNA transformation and biofilm destruction (n = 3 independent samples; mean ± s.d.). Overall, F = 91.565, P < 0.0001; Post-hoc Tukey multiple comparisons: [Control vs. DCNP: MD = 0.483, 95%CI (0.38–0.59), adjusted P < 0.0001; DNase vs. DCNP: MD = 0.420, 95%CI (0.32–0.52), adjusted P < 0.0001; DNP vs. DCNP; MD = 0.190, 95%CI (0.09–0.29), adjusted P = 0.0018]. H Analysis of Young’s modulus from the G’ (real part of the complex modulus), indicating DS change in DCNP-treated biofilms (n = 3 independent samples; mean ± s.d.). Overall, F = 96.670, P < 0.0001; Post-hoc Tukey multiple comparisons: [Control vs. DCNP: MD = 13.780, 95%CI (11.07–16.49), adjusted P < 0.0001; DNase vs. DCNP: MD = 8.860, 95%CI (6.15–11.57), adjusted P < 0.0001; DNP vs. DCNP; MD = 4.697, 95%CI (1.99–7.40), adjusted P = 0.0024]. I CLSM images showing the morphology of biofilms after DNase, DNP and DCNP treatment. Biofilms were stained with CW (red), and Mtb H37Ra was marked with the metabolic probe (blue)⁵⁵. The inset showed bacterial colonies on agar plates, which were released from the biofilms after different treatments. Scale bars = 5 µm. J Scanning electron microscopy showing biofilm morphology after DCNP treatment. Scale bars = 5 µm. K Radar map showing the systemic performance of DCNP (Nano-NucleOTAC) in Accumulation (normalized accumulation in biofilm), Biomass⁻¹ (reciprocal of normalized biofilm thickness), Z-DNA⁻¹ (reciprocal of normalized fluorescence intensity), DS⁻¹ (reciprocal of normalized deformation stress), and RB (normalized released bacterial count). L Volcano plot analysis depicting DEGs post-DCNP treatment, with gray showing no change, blue for down-regulation, and red for up-regulation. M KEGG pathway analysis of DEGs between the control and DCNP group. N Heatmap of up-regulated DEGs in arginine biosynthesis and histidine metabolism, created by an online website: https://www.omicstudio.cn/. O mRNA expression analysis of essential genes related to arginine and histidine biosynthesis following DCNP treatment (n = 3 independent samples; mean ± s.d.) (argC: t = − 30.855, P < 0.0001, argB: t = −4.337, P = 0.0123, argF: t = − 5.771, P = 0.0045, hisH: t = − 7.628, P = 0.0016, hisB: t = −6.983, P = 0.0022, hisD: t = − 6.522, P = 0.0029). P Potential immune activation mechanisms achieved by DCNP-induced Mtb metabolic modulation. Created in BioRender. Deliang, L. (2026) https://BioRender.com/v0rl3l4. Statistical significance was determined using the Student’s t test for two-group comparisons, and (E, G, H) were analyzed using one-way ANOVA for multiple-group comparisons, followed by Tukey’s two-tailed post-hoc test for pairwise analysis, all tests were two-sided. Source data are provided as a Source Data file.

To investigate whether the transformation and nucleolysis of eDNA influence the physical properties of biofilms, we conducted rheological analysis of DCNP-treated biofilms using axial mechanical indentation. Compared with other treatments, DCNP demonstrated a significant decrease in deformation stress (DS) and the minimum Young’s modulus (an indication of material stiffness in response to a normal force) (Fig. 2F–H and Supplementary Fig. S9). By analyzing these mechanical properties of Mtb H37Ra biofilm, we hypothesized that DCNP may reduce bacterial aggregation and retard the structural stability of biofilm. CLSM and scanning electron microscopy (SEM) results showed that DNP treatment effectively reduced biofilm density, evidenced by increased hollows and decreased bacterial load. Notably, DCNP treatment resulted in the complete disintegration of the biofilm, releasing 95% of the bacteria (Fig. 2I, J and Supplementary Fig. S10). By evaluating multiple datasets from different experiments, a comparative analysis was performed for DNase, DNP and DCNP-treated Mtb H37Ra biofilms. From this analysis, the DCNP-treated biofilm showed marked biofilm biomass reduction (Biomass⁻¹), increased Z-DNA transformation (Z-DNA⁻¹), decreased deformation stress (DS⁻¹), and increased bacterial release (RB). These improvements were approximately 120-fold when compared to the control. The radar map depicted the enhanced biofilm eradication and heightened bacterial release from DCNP-treated Mtb H37Ra biofilm (Fig. 2K). These indicators supported that the structure-switching functional Nano-NucleOTAC represented a robust technology to destroy biofilm and expose resident bacteria, promoting the bacteria clearance potential by immune cells and antimicrobial agents.

Nano-NucleOTAC modulated Mtb H37Ra arginine and histidine metabolism

The Nano-NucleOTAC induced biofilm degradation and potential immune activation may alter the gene expression pattern in biofilm-associated bacteria. Hence, RNA isolated from DCNP-treated Mtb H37Ra biofilms was subjected to RNA sequencing (RNA-Seq) and multiple parameters were evaluated from the RNA-Seq data. Volcano plots and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed Differentially Expressed Genes (DEGs) were enriched in pathways related to arginine biosynthesis and histidine metabolism (Fig. 2L, M). Heatmap of DEGs and mRNA expression analysis using qPCR showed that NucleOTAC intervention significantly increased the expression of genes related to arginine biosynthesis (argC, argB, argF) and histidine metabolism (hisH, hisB, hisD) (Fig. 2N, O). These results further demonstrated that DCNP treatment activated the pathways of arginine biosynthesis and histidine metabolism. Arginine, as an essential substrate, could be converted into citrulline by systemic peptidylarginine deiminase-4 to accelerate NETosis, while bacterial histidine could be catalyzed into histamine by macrophage histidine decarboxylase to maintain phagocytic activity^27,28. Therefore, the regulation of Mtb H37Ra metabolism by DCNP may enhance innate immunity and advancing TB immunotherapy (Fig. 2P).

Nano-NucleOTAC triggered cascaded anti-TB immune response

Early recruitment and activation of innate immune cells such as neutrophils plays a pivotal role in the clearance of pathogenic microorganisms²⁹. In response to pathogens, neutrophils release tumor necrosis factor (TNF), interleukin-1β (IL-1β) and interleukin-10 (IL-10), and produce NETs, acting as essential mediators in antibacterial immunity²⁹. Since DCNP treatment of Mtb H37Ra biofilms resulted in the formation of Mtb H37Ra clumps in the bottom chamber. To investigate the influence of Mtb H37Ra clumps on neutrophil recruitment and NETs production, a transwell migration assay was performed. Results showed that neutrophils migrated towards the bottom chamber containing Mtb H37Ra clumps. These clumps are the dispersed biofilm fragments which were released post DNase or DNP or DCNP treatments. DCNP-treated group showed a 75-fold increase in neutrophil migration compared to the control group (Fig. 3A, B). Upon activation, neutrophils secreted eDNA and antimicrobial peptides such as myeloperoxidase (MPO) through NETosis to capture pathogens. In DCNP-treated cells, MPO expression was markedly up-regulated, showing a 10-fold increase compared to untreated cells and surpassed both DNase and DNP groups (Fig. 3C, D). These results highlighted the enhanced antimicrobial response facilitated by DCNP treatment. NETs may enhance the phagocytosis of bacteria by macrophages¹⁶ and thus act as pseudo-opsonins. However, Z-form NETs resist DNase-mediated hydrolysis, impeding macrophage activation and the eradication of Mtb H37Ra. As shown in Fig. 3E, DCNP treatment not only induced effective neutrophil NETosis, but also transformed Z-form NETs into B-form NETs (Supplementary Fig. S11). The bacteria capturing capacity of NETs was enhanced by 10 folds compared to the control group (Fig. 3F). Thus, from these findings, we confirmed the DCNP’s ability in enhancing biofilm dispersion, neutrophil migration and Z to B-DNA transformation during NETosis, leading to the successful capture of Mtb.

Fig. 3. Nano-NucleOTAC stimulated neutrophils to produce neutrophil extracellular traps (NETs) for pathogen capture and macrophage reactivation via B-DNA transformation and nucleolysis.

A Mtb H37Ra clump activated neutrophils to release NETs. Following DCNP intervention, Z-form NETs were converted to B-form NETs, which were susceptible to degradation by DNase. Created in BioRender. Deliang, L. (2026) https://BioRender.com/5a877h4. B DCNP-mediated biofilm eradication enhanced neutrophil migration into the bottom chamber having Mtb H37Ra. C Confocal laser scanning microscope (CLSM) images showed that Mtb H37Ra clumps, dispersed by DCNP, enhanced neutrophil NETosis. Neutrophils (red) were labeled with CD66b antibody, and myeloperoxidase (MPO) was profiled by MPO antibody (green). Scale ba = 50 µm. D Quantitative immunofluorescence intensity of MPO (n = 3 independent samples; mean ± s.d.). Overall, F = 80.100, P < 0.0001; Post-hoc Tukey multiple comparisons: [Control vs. DCNP: MD = − 88.843, 95%CI (− 108.66–-69.02), adjusted P < 0.0001; DNase vs. DCNP: MD = − 75.133, 95%CI (-94.95–-55.31), adjusted P < 0.0001; DNP vs. DCNP; MD = − 60.630, 95%CI (− 80.45–-40.81), adjusted P < 0.0001]. E DCNP treatment eased the conversion of Z-form NETs into DNase-degradable B-form NETs. NETs were stained with antibodies against B-DNA (blue) and Z-DNA (yellow). Scale bars = 10 µm. F Bacterial load in the bottom chamber after different treatments (n = 3 independent samples; mean ± s.d.). Overall, F = 41.648, P < 0.0001; Post-hoc Tukey multiple comparisons: [Control vs. DCNP: MD = 89.333, 95%CI (69.70–115.97), adjusted P < 0.0001; DNase vs. DCNP: MD = 65.667, 95%CI (39.03–92.30), adjusted P = 0.0002; DNP vs. DCNP; MD = 44.667, 95%CI (18.034–71.30), adjusted P = 0.0030]. G Neutrophils in the upper chamber were replaced with macrophages, and the impact of NET fragments on macrophage phagocytosis and sterilization capabilities was examined. Tumor necrosis factor (TNF), Interleukin-1 beta (IL-1β) and Interleukin-10 (IL-10) released by neutrophils stimulated macrophage migration, promoting NET fragments internalization and bacterial elimination. Created in BioRender. Deliang, L. (2026) https://BioRender.com/vgsl9nn. H Cytokine expression (TNF, IL-1β, and IL-10) determined by qPCR (n = 3 independent samples; mean ± s.d.) (TNF: t = − 7.687, P = 0.0015, IL-1β: t = -8.97, P = 0.0009, IL-10: t = -9.65, P = 0.0006). I Histogram analysis of migrated macrophages assessed by flow cytometry (n = 3 independent samples; mean ± s.d.) (t = − 14.245, P = 0.0001). J Size distribution of NET fragments in control and DCNP-treated samples. K CLSM images depicted macrophages engulfing eDNA after DCNP treatment. Macrophages were labeled with CD86 antibody (red), and eDNA were profiled with antibodies against B-DNA and Z-DNA (blue and yellow). Scale bars = 5 µm. L CLSM images showed that DCNP treatment promoted the expression of iNOS in macrophages. Inducible Nitric Oxide Synthase (iNOS) was profiled by an iNOS antibody (red). Scale bars = 50 µm. M Agar plates and bacterial survival graph showing bacterial colonies from DCNP-treated suspension (n = 3 independent samples; mean ± s.d.) (t = 11.434, P = 0.0003). Statistical significance was determined using the Student’s t test for two-group comparisons, and (D, F) were analyzed using one-way ANOVA for multiple-group comparisons, followed by Tukey’s two-tailed post-hoc test for pairwise analysis, all tests were two-sided. Source data are provided as a Source Data file.

To evaluate macrophage recruitment and its phagocytic ability, the Mtb H37Ra supernatant with activated neutrophils and NETs were subjected to macrophage treatment using transwell membranes (Fig. 3A, G). Expression levels of TNF, IL-1β, and IL-10, along with their impact on macrophage migration, were assessed by quantitative polymerase chain reaction (qPCR) analysis. In the DCNP-treated group, there was a remarkable increase in TNF, IL-1β, and IL-10 expression, which corresponded to a 60-fold enhanced macrophage migration versus the control group (Fig. 3H, I). Interestingly, the recruited macrophages in the DCNP treatment group showed increased internalization of small-size NETs fragments (Fig. 3J, K and Supplementary Fig. S12). In addition, macrophages treated with DCNP exhibited significant activation of Inducible Nitric Oxide Synthase (iNOS), led to a 35-fold increase in iNOS expression compared to the control group (Fig. 3L and Supplementary Fig. S13). From these findings, we confirmed that the transformation and nucleolysis of NETs during innate immune response was mediated by DCNPs. The antibacterial immunity activated by DCNP was further determined by Colony Forming Unit (CFU) analysis. The immune bactericidal effect induced by DCNP treatment showed 30, 21 and 20 times increased potent bactericidal activity when compared to DNase, DNase + chloroquine or DNP treatments, respectively (Fig. 3M and Supplementary Fig. S14). Collectively, DCNP triggered a systemic cascade of innate immune responses specific to Mtb by not only recruiting the neutrophils to form NETs for trapping the bacteria but also activated macrophages to phagocytose the trapped bacteria.

Inhibition of Z-DNA transformation impeded biofilm degradation and immunotherapy efficacy by DCNP

To determine the impact of Z-DNA transformation on DCNP-mediated biofilm degradation and to validate the activation of host immune response against TB, cerium (III) chloride (CeCl₃), a Z-DNA inducer, was employed to promote the formation of self-assembled Z-DNA aggregates within biofilms⁴. Through this intervention, we aimed to inhibit DCNP-induced Z-DNA transition, thereby hindering biofilm disintegration and immune response modulation. As shown in Fig. 4A, CeCl₃ effectively promoted Z-DNA formation within biofilms, increased biofilm thickness, and impaired the Z-DNA to B-DNA transformation induced by DCNP. This further prevented the Mtb H37Ra release from the biofilm (Fig. 4A–C) and affected the activation of NETosis and DNA transformation within NETs (Fig. 4D–F). Consequently, this neutrophil inactivation led to a substantial decrease in cytokine secretion (TNF, IL-1β, and IL-10), and blocked the recruitment of macrophages (Fig. 4G, H and Supplementary Fig. S15). Thus, from these reverse validation experiments, we confirmed the DCNP-mediated biofilm degradation and cascaded anti-TB innate host immune response through Z-DNA to B-DNA transformation within biofilms and NETs.

Fig. 4. Blocked Z-DNA transformation impaired NucleOTAC-mediated biofilm degradation and bactericidal immune response.

A Altered biofilm thickness and structural components following treatments. The Mtb H37Ra biofilms were stained with CW (white), antibodies against B-DNA (blue) and Z-DNA (yellow). B CLSM microscopy images illustrating the morphology of Mtb H37Ra biofilms after treatments. Biofilms were stained with CW (red), and bacteria was marked with a metabolic probe in blue. Scale bars = 5 µm. C Quantification of released bacteria from biofilms (n = 3 independent samples; mean ± s.d.) (t = 0.241, P = 0.8212). D Inhibition of Z to B-DNA transformation by CeCl₃. NETs were stained with antibodies against B-DNA (blue) and Z-DNA (yellow). Scale bars = 10 µm. E Quantification of B-DNA and Z-DNA contents in NETs (n = 3 independent samples; mean ± s.d.) (t = -4.466, P = 0.0111). F Released bacterial load in the bottom chamber after treatments (n = 3 independent samples; mean ± s.d.) (t = -0.093, P = 0.9302). G CLSM images depicting macrophages did not engulf eDNA after treatments. Macrophages were labeled with CD86 antibody (red), and eDNA were labeled with antibodies against B-DNA (blue) and Z-DNA (yellow). Scale bars = 5 µm. H Bacterial survival in suspensions after various treatments (n = 3 independent samples; mean ± s.d.) (t = 1.031, P = 0.3607). Statistical significance was determined using the Student’s t test for two-group comparisons. Source data are provided as a Source Data file.

Nano-NucleOTAC eradicated biofilms and activated anti-TB host-immune response in vivo

To evaluate the pulmonary accumulation of DCNP, we prepared nanoparticles using Cy5.5-labeled DNase to track the distribution of nanoparticles in vivo through real-time fluorescence imaging. Increased accumulation of DCNP was observed detected in mouse lungs 12 h after intranasal administration (Fig. 5A). At 48 h, DCNP and DNP exhibited a significant increase of 30% and 40% drug retention compared to free DNase. These results indicated that nano-packaging facilitated DNase retention in the lungs, and identically increased DNase bioavailability owing to enhanced permeability and retention (EPR) effect of nanoparticles³⁰. Interestingly, DCNP concentration in the lung of Mtb H37Ra-infected mice was 30% higher than that of non-infected mice (Fig. 5B, C and Supplementary Fig. S16). This key finding highlighted the targeted accumulation of DCNPs in Mtb H37Ra biofilms, thereby elevating its therapeutic efficacy.

Fig. 5. Nano-NucleOTAC facilitated systemic and cascaded immune response in vivo via Z-DNA transformation.

A Time-lapse fluorescence images of Mtb H37Ra-infected mice. B, C The ex vivo fluorescence images showing DCNP biodistribution in mouse lungs 48 h after intranasal administration (n = 3 independent mice; mean ± s.d.). Overall, F = 22.023, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. DNP: MD = − 7733333.33, 95%CI (-16329459.95–862793.28), adjusted P = 0.0787; DNase vs. DCNP: MD = − 10800000.00, 95%CI (-19396126.61– -2203873.39), adjusted P = 0.0161; DCNP vs. Infected mice, DCNP; MD = − 10700000.00, 95%CI (− 19296126.61–-2103873.39), adjusted P = 0.0170]. D Timeline showing Mtb H37Ra infection (on day 0), drug administration (on day 1) and histopathological analysis (on day 3). Created in BioRender. Deliang, L. (2026) https://BioRender.com/eqtsdq8. E Mouse lung sections stained with antibodies against Z-DNA (yellow), B-DNA (blue) and Mtb H37Ra (red). Scale bars = 50 μm. F Mouse lung sections stained with CW (green) for biofilm observation. Scale bars = 50 μm. G Immunofluorescence staining of lung sections with MPO antibody (green), Ly-6G antibody (red, neutrophil) and DAPI (blue, nucleus). Scale bars = 50 μm. H CLSM images showing phagocytosis of eDNA by macrophages in the mouse lung. Lung sections were stained with antibodies against B-DNA (blue), Z-DNA (yellow), and macrophages with CD86 antibody (red). Scale bars = 5 µm. I Flow cytometry analyzing the percentage of macrophages (CD86) phagocytic bacteria (Mtb H37Ra). J Immunofluorescence staining of lung sections with Inducible Nitric Oxide Synthase (iNOS) antibody (red). Scale bars = 50 μm. K Illustration showing the in vivo systemic and cascaded activation of host immune response against tuberculosis by DCNP. NucleOTAC-mediated Z-DNA transformation not only promoted biofilm matrix eradication to stimulate neutrophil NETosis, but also facilitated the conversion of Z-form NETs into DNase-degradable B-form NETs, enhancing pathogen capture and macrophage reactivation. Created in BioRender. Deliang, L. (2026) https://BioRender.com/1qvoeyb. L Intracellular B-DNA content in macrophages (Mφs) (n = 5 independent mice; mean ± s.d.) (t = − 36.754, P < 0.0001). M Bactericidal effect of DCNP in lung (n = 5 independent mice; mean ± s.d.) (t = 24.186, P < 0.0001). N, O H&E staining and pathological scoring of lung sections. Scale bars = 50 µm (n = 5 independent mice). Statistical significance was determined using the Student’s t test for two-group comparisons, and (C) were analyzed using one-way ANOVA for multiple-group comparisons, followed by Tukey’s two-tailed post-hoc test for pairwise analysis, all tests were two-sided. Source data are provided as a Source Data file.

After Mtb H37Ra inoculation, intranasal administration of different therapeutic agents was conducted in mice on day 1, followed by tissue sectioning and immunofluorescence analysis on day 3 (Fig. 5D). In Fig. 5E, F, using Mtb H37Ra-infected mouse model, minimal reduction in B-DNA/Z-DNA content and bacterial load were observed in the DNase treatment group, demonstrating that free DNase in mouse lung was insufficient to degrade biofilms or induce innate antibacterial immunity against Mtb H37Ra. In contrast, DNP degraded 55% B-DNA and 20% Z-DNA, and showed increased bacterial clearance in the infected lungs. Notably, DCNP achieved nearly complete clearance of both eDNA and biofilm, facilitated by NucleOTAC-mediated Z-DNA transformation and nucleolysis (Supplementary Fig. S17). As shown in Fig. 5G, H, DCNP treatment induced robust neutrophil NETosis with abundant MPO release and substantial internalization of B-DNA by macrophages (Supplementary Fig. S18). This ultimately led to a 14-fold increase in the phagocytosis of Mtb H37Ra by macrophages and a 44-fold increase in iNOS expression compared to the control group (Fig. 5I, J and Supplementary Fig. S19).

Through these in vivo findings, we confirmed the Nano-NucleOTAC-mediated eDNA degradation, biofilm disruption, and bacteria release under pathogenic conditions. Similar to the in vitro findings, the in vivo mouse infection studies employing DCNPs demonstrated the activation of neutrophils, which captured the released bacteria through the formation of NETs. Meanwhile, Nano-NucleOTAC cleaved NETs into small fragments, and accelerated macrophage recruitment and bacterial clearance through phagocytosis, achieving 94% reduction in pulmonary bacterial load, significantly higher than other treatment groups. This systemic and cascaded immune response in vivo resulted in a remarkably enhanced eDNA internalization by macrophages when compared to the control group (50-fold) (Fig. 5K–M and Supplementary Fig. S14). After DCNP treatment, the pro-inflammatory factors (TNF, IL-1β, and IL-6) returned to normal levels, and the concentration of anti-inflammatory IL-10 increased significantly, with no hepatic or renal function impairment or hematological changes detected (Supplementary Figs. S20, S21). H&E staining of mouse lung sections and pathological score analysis revealed no obvious tissue lesions or immune inflammation in the lung, heart, liver, spleen, or kidney. In addition, after 60-day DCNP treatment, mice showed no biofilm relapse or residual infection in lungs with low PD-1 levels, indicating no T cell exhaustion (Supplementary Figs. S22, S23). Notably, intervention with chloroquine or lactoferrin alone did not cause changes in inflammatory factors or tissue damage (Supplementary Figs. S24, S25). These data demonstrated that DCNP elicited a transient, self-limiting innate immune response balanced by IL-10, without the risk of immune overactivation, thereby highlighting the acceptable therapeutic efficacy and biosafety of Nano-NuleOTAC therapy for pulmonary TB treatment (Fig. 5N, O and Supplementary Figs. S20, S26).

To further determine whether DCNP-mediated biofilm disruption and immune activation were specific to pathogenic Mtb, we evaluated the virulent reference strain Mtb H37Rv under identical DCNP treatment conditions. In vitro, Z-DNA also existed in Mtb H37Rv biofilms (Fig. 6A). DCNP removed B/Z-DNA, disintegrated the biofilm and released most Mtb H37Rv as clumps, which activated neutrophils to release chemokines that recruit macrophages with enhanced bactericidal activity (Fig. 6A–H and Supplementary Figs. S14, S27A, S28A). In virulent Mtb H37Rv-infected mouse model, a single intranasal dose of DCNP cleared approximately 90% of pulmonary B-DNA and Z-DNA within 48 h through Nano-NucleOTAC mediated Z-DNA transformation and degradation (Fig. 6I, J and Supplementary Fig. S29A). This robust eDNA removal triggered a 23-fold increase of neutrophil myeloperoxidase (MPO) release and 32-fold increase of macrophage iNOS expression in comparison to healthy controls (Fig. 6K, L and Supplementary Fig. S29B, C), indicating explosive reactivation of innate immunity. Lung-homogenate colony counts dropped 8- and 4-fold compared with untreated infected mice and DNase + chloroquine-treated mice (Fig. 6M and Supplementary Fig. S14), while histopathology showed markedly reduced inflammatory infiltration, preserved alveolar architecture, and significantly improved pathological scoring (Fig. 6N and Supplementary Fig. S30A). Together, these findings demonstrated that DCNP dismantled the Z-DNA-rich biofilm scaffold of virulent Mtb H37Rv and unleashed a systemic and cascaded immune response that achieved rapid bacterial clearance and substantial tissue protection in vivo.

Fig. 6. Nano-NucleOTAC facilitated biofilm disruption and immune activation against Mtb H37Rv.

A Biofilm thickness attenuation and structural component variation after DCNP treatment. The Mtb H37Rv biofilms were stained with CW (white), antibodies against B-DNA (blue) and Z-DNA (yellow). B Quantitative biofilm thickness after different treatments (n = 3 independent samples; mean ± s.d.) (t = 20.069, P < 0.0001). C Fluorescent intensity of B-DNA and Z-DNA within Mtb H37Rv biofilms after DCNP treatment (n = 3 independent samples; mean ± s.d.). B-DNA: Overall, F = 244.479, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: MD = − 37.667, 95%CI (-48.42–-26.91), adjusted P < 0.0001; DNP vs. Control: MD = -59.667, 95%CI (-70.42–-48.91), adjusted P < 0.0001; DCNP vs. Control; MD = − 88.000, 95%CI (− 98.75–-77.25), adjusted P < 0.0001]. Z-DNA: Overall, F = 362.638, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: MD = − 10.667, 95%CI (-19.91–-1.42), adjusted P = 0.0252; DNP vs. Control: MD = -20.667, 95%CI (− 29.91–-11.42), adjusted P = 0.0004; DCNP vs. Control; MD = −86.333, 95%CI (− 95.58–-77.09), adjusted P < 0.0001]. D CLSM images showed that Mtb H37Rv clumps, dispersed by DCNP, enhanced neutrophil NETosis. Neutrophils (red) were labeled with CD66b antibody, and MPO was profiled by MPO antibody (green). Scale bars = 50 µm. E CLSM images showed that DCNP treatment promoted the expression of iNOS (red) in macrophages. Scale bars = 50 µm. F Quantitative immunofluorescence intensity of MPO (n = 3 independent samples; mean ± s.d.) (t = − 11.475, P = 0.0003). G Quantitative immunofluorescence intensity of iNOS (n = 3 independent samples; mean ± s.d.) (t = -24.527, P < 0.0001). H Bacterial survival graph showing bacterial colonies from DCNP-treated suspension (n = 3 independent samples; mean ± s.d.) (t = 21.602, P < 0.0001). I Timeline showing Mtb H37Rv infection (on day 0), drug administration (on day 1) and histopathological analysis (on day 3). Created in BioRender. Deliang, L. (2026) https://BioRender.com/4v0nshb. J Mouse lung sections stained with antibodies against Z-DNA (yellow), B-DNA (blue) and Mtb H37Rv (red). Scale bars = 50 μm. K Immunofluorescence staining of lung sections with MPO antibody (green), Ly-6G antibody (red, neutrophil) and DAPI (blue, nucleus). Scale bars = 50 μm. L Immunofluorescence staining of lung sections with iNOS antibody (red). Scale bars = 50 μm. M Bactericidal effect of DCNP in lung (n = 5 independent mice; mean ± s.d.) (t = 24.032, P < 0.0001). N Pathological scoring of lung sections (n = 5 independent mice). Statistical significance was determined using the Student’s t test for two-group comparisons. Source data are provided as a Source Data file.

Non-tuberculous mycobacteria (NTM) are increasing challenges in clinical settings due to their growing drug resistance and pathogenicity³¹, so we used Mycobacterium smegmatis (M.smeg) as a model strain to evaluate DCNP’s efficacy against pulmonary NTM infections. Under identical experimental conditions, DCNP exhibited comparable biofilm-dismantling activity and immune-activating effects against M. smeg (Fig. 7 and Supplementary Figs. S14, S27B, S28B, S30B, S31). Altogether, validated in three complementary mouse models (Mtb H37Ra, Mtb H37Rv and M.smeg), Nano-NucleOTAC demonstrated robust broad-spectrum anti-mycobacterial activity against virulent Mtb and clinically refractory NTM by targeted biofilm disruption and immune activation.

Fig. 7. Nano-NucleOTAC facilitated biofilm disruption and immune activation against M.smeg.

A Biofilm thickness attenuation and structural component variation after DCNP treatment. The M.smeg biofilms were stained with CW (white), antibodies against B-DNA (blue) and Z-DNA (yellow). B Quantitative biofilm thickness after different treatments (n = 3 independent samples; mean ± s.d.) (t = 14.511, P = 0.0001). C Fluorescent intensity of B-DNA and Z-DNA within M.smeg biofilms after DCNP treatment (n = 3 independent samples; mean ± s.d.). B-DNA: Overall, F = 94.821, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: MD = − 41.000, 95%CI (− 58.01–-23.99), adjusted P = 0.0003; DNP vs. Control: MD = − 55.000, 95%CI (− 72.01–-37.99), adjusted P < 0.0001; DCNP vs. Control; MD = -88.333, 95%CI (-105.35–-71.32), adjusted P < 0.0001]. Z-DNA: Overall, F = 242.275, P < 0.0001; Post-hoc Tukey multiple comparisons: [DNase vs. Control: (MD) = -11.000, 95%CI (− 22.40–-0.40), adjusted P = 0.0585; DNP vs. Control: MD = − 16.667, 95%CI (− 28.06–-5.27), adjusted P = 0.0069; DCNP vs. Control; MD = − 86.333, 95%CI (− 97.73–-74.94), adjusted P < 0.0001]. D CLSM images showed that M.smeg clumps, dispersed by DCNP, enhanced neutrophil NETosis. Neutrophils (red) were labeled with CD66b antibody, and MPO was profiled by MPO antibody (green). Scale bars = 50 µm. E CLSM images showed that DCNP treatment promoted the expression of iNOS (red) in macrophages. Scale bars = 50 µm. F Quantitative immunofluorescence intensity of MPO (n = 3 independent samples; mean ± s.d.) (t = − 15.628, P < 0.0001). G Quantitative immunofluorescence intensity of iNOS (n = 3 independent samples; mean ± s.d.) (t = − 16.351, P < 0.0001). H Bacterial survival graph showing bacterial colonies from DCNP-treated suspension (n = 3 independent samples; mean ± s.d.) (t = 10.459, P = 0.0005). I Timeline showing M.smeg infection (on day 0), drug administration (on day 1) and histopathological analysis (on day 3). Created in BioRender. Deliang, L. (2026) https://BioRender.com/6099cm2. J Mouse lung sections stained with antibodies against Z-DNA (yellow), B-DNA (blue) and M.smeg (red). Scale bars = 50 μm. K Immunofluorescence staining of lung sections with MPO antibody (green), Ly-6G antibody (red, neutrophil) and DAPI (blue, nucleus). Scale bars = 50 μm. L Immunofluorescence staining of lung sections with iNOS antibody (red). Scale bars = 50 μm. M Bactericidal effect of DCNP in lung (n = 5 independent mice; mean ± s.d.) (t = 18.563, P < 0.0001). N Pathological scoring of lung sections (n = 5 independent mice). Statistical significance was determined using the Student’s t-test for two-group comparisons. Source data are provided as a Source Data file.

DCNP potentiated TB biofilm disruption and immune activation in clinical samples

During infection, planktonic bacteria are susceptible to antibiotics and the host immune response. However, within the EPS-based biofilm matrix, the eDNA affects the neutrophil-mediated bacterial capture and inhibits the phagocytosis by macrophages^32–34. To assess the therapeutic potential of DCNP for clinical tuberculosis, we collected expectorated sputum, BALF and pus samples from confirmed TB patients. The representative images of the sputum solids disruption assay demonstrated complete disintegration of TB sputum solids after DCNP treatment (Fig. 8A). Consistent with earlier results, DCNP treatment effectively eradicated almost all Mtb biofilms, B-DNA and Z-DNA in sputum (Fig. 8B). This was achieved by disrupting the steady-state Z-DNA and B-DNA by DCNPs loaded with DNase. We also examined the synergistic bactericidal ability of DCNP in combination with anti-TB drugs. DCNP intervention significantly enhanced the therapeutic effect of RIF against Mtb. DCNP intervention significantly enhanced the efficacy of rifampicin (RIF) against Mtb, reducing bacterial load by 97%, whereas RIF alone and RIF + DNase + chloroquine achieved bacterial clearance rates of 16% and 62%, respectively, and the standard drug regimens (HRZE combination, ‘isoniazid, rifampicin, pyrazinamide and ethambutol’) achieved a 61% bacterial clearance rate (Fig. 8C and Supplementary Figs. S14, S32). Using clinical samples obtained from Mtb patients, we further confirmed the efficacy of Nano-NucleOTAC technology. This eDNA targeting technology may open opportunities for the treatment of drug-resistant TB, which forms recalcitrant biofilms in infected patients.

Fig. 8. Nano-NucleOTAC enhanced the bactericidal efficacy of anti-TB drugs or immune cells in clinical samples (sputum, BALF and pus).

Biofilms in sputum, BALF or pus were stained with CW (green), antibodies against B-DNA (blue) and Z-DNA (yellow). Created in BioRender. Deliang, L. (2026) https://BioRender.com/ei64kzz. A Photographs and CLSM images of solids-disrupted sputum. DCNP completely disintegrated the TB sputum solids. B CLSM images showing the contents of biofilms, B-DNA and Z-DNA in sputum after different treatments. Scale bars = 10 μm. C Enhanced bactericidal effect of rifampicin (RIF) following DCNP intervention (n = 3 independent sputum samples; mean ± s.d.). Overall, F = 38.005, P < 0.0001; Post-hoc Tukey multiple comparisons: [Untreated vs. DCNP (RIF treated): MD = 97.067, 95%CI (66.25–127.88), adjusted P < 0.0001; Control vs. DCNP (RIF treated): MD = 81.280, 95%CI (50.46–112.10), adjusted P = 0.0002]. D CLSM images showing the contents of biofilms, B-DNA and Z-DNA in BALF after different treatments. Scale bars = 10 μm. E, F Flow cytometry quantifying the percentage of macrophages (CD86) and neutrophils (CD66b) in BALF (n = 3 independent BALF samples; mean ± s.d.). G Measurement of MPO secretion and macrophage phagocytosis in BALF. BALF samples were stained with MPO antibody (green), CD66b antibody (red, neutrophils) and CD86 antibody (red, macrophages). Scale bars = 50 μm and 5 μm. H Bactericidal effect of DCNP in BALF (n = 3 independent BALF samples; mean ±  s.d.) (t = 18.805, P = 0.0005). I CLSM images showing the contents of biofilms, B-DNA and Z-DNA in pus after different treatments. Scale bars = 10 μm. J, K Flow cytometry quantifying the percentage of macrophages (CD86) and neutrophils (CD66b) in pus samples (n = 3 independent pus samples; mean ± s.d.). L MPO secretion and macrophage phagocytosis in pus. Pus was stained with MPO antibody (green), CD66b antibody (red, neutrophils) and CD86 antibody (red, macrophages). Top scale bars = 50 μm. Bottom scale bars = 5 μm. M Bactericidal effect of DCNP in pus (n = 3 independent pus samples; mean ± s.d.) (t = 27.285, P < 0.0001). Statistical significance was determined using the Student’s t test for two-group comparisons, and (C) were analyzed using one-way ANOVA for multiple-group comparisons, followed by Tukey’s two-tailed post-hoc test for pairwise analysis, all tests were two-sided. Source data are provided as a Source Data file.

BALF and lung pus contain abundant neutrophils and macrophages. Thus, they were collected from TB patients to validate the function of Nano-NucleOTAC in eradicating hard-to-treat biofilms and activating innate immunity. Experimental results showed that biofilm content and B-DNA/Z-DNA expression gradually decreased across the control group, DNase group and DNP group, and almost disappeared in the DCNP group (Fig. 8D, I). Phagocytic immune cells, including neutrophils and macrophages, are important innate immune cells that phagocytose and degrade bacteria for antimicrobial defense^35,36. Flow cytometry results demonstrated high abundance of neutrophils and macrophages in both BALF and pus, with a ratio of 7:1 and 1:6, respectively (Fig. 8E, F, J, K). DCNP treatment induced effective neutrophil NETosis and enhanced macrophage phagocytosis in these samples (Fig. 8G, L). Particularly, the bactericidal efficacy of DCNP in BALF and pus significantly reached 85% and 94%, respectively, when compared to the control group (Fig. 8H, M and Supplementary Fig. S14). Thus, in clinical samples, Nano-NucleOTAC technology disrupted Z-DNA skeleton, collapsed refractory biofilms, exposed resident bacteria, and facilitated its clearance by anti-TB drugs or activated immune cells. These findings confirmed that Nano-NucleOTAC therapy could be used for treating lesions in the respiratory tract, oral cavity and urinary tract, where biofilm-associated infections are common.

Discussion

Considering TB causes lethal respiratory failure and systemic inflammation, the WHO has established a goal to eradicate TB by 2035³⁷. Therefore, developing safe and effective strategies for Mtb elimination without drug resistance are urgently needed. However, recent studies reveal that Mtb forms biofilms, made up of eDNA, polysaccharides, proteins, and lipids⁷. Z-DNA is a left-handed eDNA, with its multifaceted implications in biofilm integrity and immune modulation^6,17. It poses a significant challenge to medical and health fields by bolstering biofilm resistance, undermining antimicrobial defenses, and complicating therapeutic interventions^38–40, thereby intensifying the complexity of combating biofilm-associated infections. Secreted by Mtb, the eDNA matrix coalesced into a membranous structure that adhered to lesion surfaces in infection, such as throat, lungs, and pleural space^7,41,42. Our study discovered the presence of Z-DNA within biofilms in Mtb H37Ra-infected mice, sputum, BALF and pus samples from confirmed TB patients. Notably, elevated Z-DNA levels in sputum suggest that sputum accumulation in the respiratory tract creates an optimal environment for biofilm formation, providing nutrients and protection for bacterial survival. In contrast, lower Z-DNA levels in infected mouse lungs are likely due to efficient Mtb clearance by the mouse immune system⁴³. Whatever, Z-DNA within the biofilms served as a crucial scaffold to induce antibiotic resistance, immune tolerance and persistent infections. By revealing this biological principle, we aim to apply nanotechnology to break biofilms and facilitate bacterial clearance by combining antimicrobial agents and host immune response.

Targeting Chimera (TAC) represents a sophisticated class of molecular tools that degrade specific substances, providing a formidable platform for the clearance of pathogenic factors⁴⁴. For instance, Proteolysis Targeting Chimeras (PROTACs) leverage the ubiquitin-proteasome system to dismantle target proteins, while RNase Recruiting Chimeras (RIBOTACs) harness the RNA quality control pathways to neutralize toxic transcripts^45,46. Building on these principles, we have developed Nano-NucleOTAC technology, a nano-nucleolysis targeting chimera that integrates an eDNA-binding ligand, such as lactoferrin, alongside a self-carried nuclease equipped with DNase. To further enhance the degradation of nuclease-resistant Z-DNA, we incorporated chloroquine into this system, as it has been shown to revert Z-form DNA back to the B-form conformation, thereby restoring its susceptibility to DNase I degradation⁴⁷. This innovative approach synergizes the targeted eDNA binding with the inherent clearance mechanisms. Importantly, unencapsulated chloroquine readily accumulates in erythrocytes, liver, kidney, heart, and retina, causing prolonged retention and off-target toxicity⁴⁸. To overcome this, Nano-NucleOTAC leverages three strategies: first, GSH-responsive release ensures on-demand delivery to pulmonary biofilms; second, an ultra-low treatment dose (0.02 mg/kg in mice, approximately 1/1600 of human toxic dose)⁴⁹; moreover, intranasal administration bypasses systemic absorption and first-pass metabolism, maximizing lung retention. Notably, DCNP caused no inflammatory response or tissue damage in major organs, confirming its optimized safety. Acting as a molecular sculptor, Nano-NucleOTAC embeds DNase within biofilm Z-DNA and transforms the nuclease-resistant Z-DNA into nuclease-degradable B-DNA, achieving effective Z-DNA degradation and biofilm clearance. The intricate relationship between Z-DNA’s structure with gene regulation, tumorigenesis and viral infections, underscores its role in various pathologies^4,50. Thus, developing modulators of Z-DNA as therapeutic interventions is highly promising. However, controlled release of DNase is critical to limit undesired tissue damage and requires more precise dosimetry and spatial control mechanism⁵¹. Using DCNP, we studied its efficacy and mode of action under in vitro, in vivo and real-world conditions. It provides a readily deployable strategy that disrupts Z-DNA structures and acts potently against various virulent Mtb clades and recalcitrant NTM, curbing emerging mycobacterial threats and accelerating the global eradication of drug-resistant pulmonary infections. Moreover, Nano-NucleOTAC electrostatically adsorbs to any eDNA-rich, negatively charged biofilms, enabling broad-spectrum accumulation on biofilms not only in Mtb but also in NTMs and other bacterial species.

Neutrophils and macrophages play a critical role in pathogen elimination⁵². But immunomodulation strategies to enhance bactericidal activity against Mtb are rarely explored. Consistent with prior reports that Mtb cells released from biofilms trigger neutrophil NETosis, releasing myeloperoxidase (MPO), chemotactic mediators and NET-associated cues to recruit and activate macrophages for enhanced iNOS expression and bactericidal efficacy^29,53, our experimental results confirmed that DCNP treatment likewise elicited this sequential immune activation cascade. More importantly, we further demonstrated that following DCNP-mediated nucleolysis, NET fragments derived from neutrophil NETosis enhance the phagocytic uptake of Mtb by macrophages. Thus, multi-functional Nano-NucleOTAC DCNP has the potential to eradicate biofilms and reactivate host immune response against Mtb. Clinical sample intervention results emphasized Nano-NucleOTAC’s capability to fortify antibacterial immunity and enhance drug susceptibility against resistant infections. The clinical findings in this study are still preliminary due to the small sample size. A large-scale, multi-centric, and well-designed trial is anticipated to validate the efficacy and safety of DCNPs. Setting operatable quality control standards and processes for scale-up production are also crucial for its actual application. Nevertheless, these have not yet been implemented in the clinic, and may require further characterization and optimization in vitro and in vivo. The modulation mechanisms of Nano-NucleOTAC on immune cells such as T-lymphocytes, B-lymphocytes, natural killer cells, and dendritic cells remain poorly understood, as TB infection affects diverse immune responses, including innate and adaptive immunity⁵⁴. To further investigate whether DCNP treatment can prevent TB recurrence or reinfection, clarifying whether DCNP affects T-cell activation, granuloma remodeling, and immune memory in latent/reactivation tuberculosis models is necessary. In addition, to advance the application of Nano-NucleOTAC in clinical settings, further efforts are needed to overcome challenges related to long-term toxicity assessments in large animal models, and large-scale production for GMP and regulatory approval.

In summary, our study identified biofilms enriched with DNase-resistant Z-DNA in clinical samples from TB patients, which contributed to Mtb drug resistance and immune evasion. To address this challenge, we synthesized DCNP using Nano-NucleOTAC technology, incorporating an eDNA-binding ligand, DNA intercalator and proteasomes. The structure-switching functional DCNP was designed to enhance biofilm eradication and reactivate systemic, cascading anti-bacterial immunity via Z-DNA structure transformation and nucleolysis. Acting as an eDNA-targeted degradation agent, DCNP effectively accumulated within biofilms by the EPR effect, degraded the Z-DNA scaffold, decreased deformation stress to disrupt biofilms (Mtb H37Ra, Mtb H37Rv and even M.smeg) and release resident bacteria. Moreover, DCNP cleaved Z-form NETs produced by neutrophils into small fragments, activating macrophages for efficient bacterial phagocytosis and elimination. Importantly, DCNP significantly increased the susceptibility of biofilm-residing Mtb to anti-TB drugs. Overall, DCNP demonstrated robust efficacy in eradicating biofilms across various clinical samples, promoting the Mtb clearance potential of anti-TB drugs, and eliciting systemic, cascaded immune responses against Mtb. The Nano-NucleOTAC technology represents a potent tool for advancing therapeutic strategies in the treatment of clinical recalcitrant tuberculosis biofilm infections and enhancing drug discovery and antimicrobial therapeutics.

Methods

Ethics approval

All the animal experiments were conducted in compliance with the Chinese National Standard and under the protocols that were approved by the Animal Care and Use Committee at Shenzhen Third People’s Hospital (protocol number 2022-018). Mice were euthanized by carbon dioxide inhalation. Clinical sample studies were conducted according to the ethical principles of the Declaration of Helsinki. Patient recruitment and BALF/sputum/pus sample collection were approved by the Research Ethics Committee of Shenzhen Third People’s Hospital (clearance number 2024-195). All participants provided written informed consent for sample collection and the subsequent analyses.

Bacterial growth

Mycobacterium tuberculosis H37Ra (Mtb H37Ra), Mycobacterium tuberculosis H37Rv (Mtb H37Rv) or Mycobacterium Smegmatis (M.smeg) strain were cultured in Middlebrook 7H9 broth supplemented with 10% (w/v) OADC (Oleic Acid-Albumin-Dextrose-Catalase), 0.5% (w/v) glycerol, and 0.1% (w/v) Tween 80. For liquid culture, bacteria-containing flask were shaken at 200 rpm and 37 °C on an Innova 42 shaker incubator (GER). For solid culture, bacteria were grown on solid Middlebrook 7H10 agar plates supplemented with 0.5% (w/v) glycerol and 10% (w/v) OADC.

Characterization of B-DNA and Z-DNA in Mtb biofilms

Mtb H37Ra, Mtb H37Rv or M.smeg were inoculated in 7H9 medium to an initial optical density (OD₆₀₀) of 0.3 and cultured in 24-well plates at 37 °C for 10 days without shaking. After formation, the biofilms were washed with phosphate-buffered saline (PBS) and probed with monoclonal antibodies against B-DNA (Abcam, ab27156, 1:1000), Z-DNA (Absolute Antibodies, Ab00783-3.0, 1:1000) in PBS supplemented with 3% (w/v) bovine serum albumin (BSA) and incubated overnight at 4 °C. In this process, the anti-B-DNA and Z-DNA monoclonal antibodies have been verified by Goodman et al. to specifically label B-DNA and Z-DNA in the biofilm, respectively¹⁰. Following incubation, the biofilms were washed again with PBS and then incubated with PBS containing 3% (w/v) BSA and diluted secondary antibodies (anti-rabbit ab150077, anti-mouse ab175660) and Calcofluor White (CW, a reagent that labels Mtb biofilm cellulose⁷) (Sigma, 3 μM) for 1 h at room temperature. The biofilms were finally observed using a Zeiss LSM 980 confocal microscope (Carl Zeiss Inc. GER). For quantitative analysis of Mtb biofilms in vitro, after removing the supernatant, biofilms were stained with 1% crystal violet for 10 min, gently washed twice, and finally solubilized with ethanol. Absorbance was measured at 600 nm with a microplate reader (BioRad, USA) to determine biofilm biomass.

Mtb H37Ra, Mtb H37Rv or M.smeg suspensions at a concentration of 5 × 10⁷ CFU/mL in physiological saline (0.9%, w/v) were prepared for animal studies. Eight-week-old BALB/c mice were administered with 50 μL of suspension via intranasal injection. All mice used for animal experiments are female with a gentle character, and the sex of the mice does not affect the results of the experiments. After sacrifice by carbon dioxide, the lungs of infected mice were dissected and frozen in optimal cutting temperature compound (OCT). Sections of 6 μm were cut and processed using the standard protocol³⁸. The study ensured the highest standards of animal welfare, with all experimental procedures conducted in strict accordance with the ethical guidelines and regulations for animal research. Archived clinical specimens from a confirmed TB patient, including sputum, bronchoalveolar lavage fluid (BALF) and pus, were frozen in OCT immediately upon collection. The presence of B-DNA, Z-DNA and biofilms were detected using the same antibodies, CW, and visualization techniques as for in vitro biofilms.

Preparation and characterization of DCNP

A total of 60 mg lactoferrin (Yuanye Bio-Technology, CHN) was reduced with 30 mg GSH in 10 mL of deionized water, stirred for 1 h, then purified using ultrafiltration (MW 3500) to remove excess GSH. Reduced lactoferrin was mixed with 1 mL DNase (Deoxyribonuclease I from bovine pancreas, 15 mg/mL, Sigma-Aldrich USA) and 100 μL chloroquine (600 μg/mL), followed by the addition of 3 mL ethanol to precipitate DCNP. After reconstructing disulfide bonds by stirring for 3 h, the mixture was subjected to ultra-filtration (MW 10 kDa, Millipore, USA) to purify DCNP. This enabled the removal of ethanol, free lactoferrin, DNase, and chloroquine.

Size distribution, surface potential and polydispersity index (PDI) of DCNP were characterized using a Zetasizer Nano ZS (Malvern, UK). The batch-to-batch consistency was evaluated by calculating the coefficient of variation (CV) of particle size, PDI, and zeta potential across six batches of DCNP, and the stability was assessed by comparing these parameter changes after 30 days. For in vitro drug release analysis, DNase was labeled with FITC-NHS at a 10:1 molar ratio, stirred for 1 h, and unbound FITC-NHS was removed by ultrafiltration (MW 10 kDa). FITC-DNase and lactoferrin were formulated into nano-NucleOTAC. DCNP solutions were then dialyzed (MW 10 kDa) against deionized water at 37 °C with shaking. At set intervals, samples were taken and replaced with fresh water. DNase release from DCNP, in the presence or absence of 0.5 mM GSH, was monitored using a fluorescence microplate reader (Ex/Em, 540 nm/590 nm). To determine the enzyme loading efficiency of DCNP, a standard calibration curve was generated by plotting enzymatic activity (U/mL) of serially diluted DNase standards against known protein concentrations (mg/mL) to quantify DNase mass. Subsequently, the enzyme loading efficiency was calculated using the formula: Enzyme loading efficiency (%) = [Mass of DNase bound to nanoparticles / Mass of nanoparticles] × 100%.

The binding interactions between DCNP and DNA were investigated using the AutoDock Vina software. The three-dimensional structures of lactoferrin and DNA were obtained from the RCSB Protein Data Bank (PDB: 1BLF) and incorporated into the docking workflow. AutoDock Vina was employed to predict the binding affinity and preferred molecular orientation between the two biomolecules. To facilitate a comprehensive understanding of the lactoferrin-DNA molecular complex, we utilized the PyMOL molecular visualization software. PyMOL was used to generate high-resolution graphical representations of the docked complex, enabling detailed examination of the intermolecular interactions and spatial arrangements of the key residues involved.

For the biofilm accumulation assay, FITC-labeled DNase and DCNP were prepared. Mature Mtb H37Ra biofilms were grown in 24-well plates at 37 °C for 10 days, which were then incubated with FITC-DNase, DNP or DCNP (all containing 0.4 mg/mL DNase) for 2 h, followed by washing with PBS and stained with CW. The images were captured using a Zeiss LSM 980 microscope (Carl Zeiss, GER) to observe CW (Ex/Em, 355 nm/435 nm) and FITC (Ex/Em, 488 nm/525 nm). Axial mechanical indentation was applied to biofilms using a Discovery Hybrid Rheometer-2 (DHR-2; TA instruments, USA) with a sand-blasted parallel plate geometry. The FluoroDish with biofilms was placed on a Peltier plate, and the plate geometry was lowered at 1 mm/s. Data were collected using TRIOS v5 software and converted to stress-strain curves in GraphPad 8.0, with Deformation Stress (DS) indicated by the Young’s modulus. Differently treated biofilms were fixed in 2.5% (w/v) glutaraldehyde, washed, dehydrated through ethanol series, treated with hexamethyldisilane, freeze-dried, sputter-coated and examined via SEM (Thermo. USA).

For RNA sequencing, mature Mtb H37Ra biofilms were co-cultured with medium or DCNP (containing 0.4 mg/mL DNase) at 37 °C for 24 h. Post-treatment biofilms were harvested, and RNA was isolated for RNA-seq by BGI (Beijing Genomics Institute) on the DNBSEQ-T7 platform. Differentially expressed genes (DEGs) were identified with a false discovery rate (FDR) below 0.05, and an absolute log2 fold change greater than or equal to 0.5. The expression of genes related to arginine biosynthesis (argC, argB, argF) and histidine metabolism (hisH, hisB, hisD) were detected by qPCR (Analytik Jena, GER), and all primers are described in Supplementary Table S2.

DCNP-mediated neutrophil response profiling

The activation of neutrophil NETosis by DCNP treatment was investigated. Neutrophils were isolated from human peripheral blood using an extraction kit (Solarbio, P9040, CHN). In order to obtain Mtb H37Ra clumps, bacteria were cultured in 7H9 medium for 10 days in vitro to form mature biofilms. The biofilms were pretreated with DCNP for 24 h to allow for biofilm disruption. Then 5 × 10⁵ neutrophils were placed in the transwell upper chamber, while DCNP-treated bacteria clump-containing biofilms were added to the bottom chamber with cell-climbing slices.

For flow cytometry analysis, neutrophils were gated by removing debris and adherent cells using FSC-A/SSC-A and FSC-A/FSC-H at 3 h. Monocytes were labeled with CD45 (Biolegend, 368504, 1:200), and neutrophils were labeled with CD66b (Biolegend, 392911, 1:200). Neutrophils were then identified as CD45⁺CD66b⁺ (Supplementary Fig. S33A). The cell climbing slices were collected for immunofluorescence staining with primary antibodies specific to B-DNA, Z-DNA, myeloperoxidase (MPO, Abcam, ab208670, 1:1000), and CD66b (a neutrophil surface marker, Abcam, ab25377, 1:1000), incubated overnight at 4 °C. This was followed by a 1 h incubation with secondary antibodies. The samples were examined using a CLSM, and fluorescence intensities were quantified using ImageJ software. Samples were analyzed using a BD LSRFortessa flow cytometer (BD Biosciences, USA) and FACSDiva v8.0.1 (BD Biosciences, USA). Meanwhile, cytokines (TNF, IL-1β or IL-10) in the supernatant were detected by qPCR (Analytik Jena, GER). All the primers were designed and synthesized by Sangon Biotech (Sangon Biotech Co., Ltd., CHN), and the primer sequences are listed Supplementary Table S2. The bacteria-capturing capacity of NETs was evaluated by harvesting the supernatant from the bottom chamber and performing serial dilutions. Subsequently, 50 μL of each dilution was spotted onto 7H10 agar plates. After incubation for 2 weeks at 37 °C, the number of colony-forming units (CFUs) was determined. The viability of Mtb was compared with the control group.

Macrophage migration and phagocytosis

Human macrophages were isolated from human peripheral blood using an extraction kit (Solarbio, P9040, CHN). Then macrophages were cultured in RPMI 1640 medium containing 1000 IU/mL recombinant human GM-CSF (rhGM-CSF) for 3 days, with half of the medium replaced every 48 h. Finally, the macrophages were scraped off and counted. To estimate macrophage functions after DCNP-mediated neutrophil stimulation, the upper chamber was replaced by macrophages (5 × 10⁵ cells). The size distribution of NET fragments in control and DCNP-treated samples was characterized using a Zetasizer Nano ZS.

After 12 h, the cell climbing slices were collected and incubated overnight at 4 °C with primary antibodies specific to B-DNA, Z-DNA, iNOS (Proteintech, 18951-1, 1:1000), and CD86 (Abcam, ab119857, 1:1000) for immunofluorescence staining. This was followed by a 1 h incubation with secondary antibodies. The samples were examined using a CLSM, and fluorescence intensities were quantified using ImageJ software. For flow cytometry analysis, macrophages were gated by removing debris and adherent cells using FSC-A/SSC-A and FSC-A/FSC-H. Monocytes were labeled with CD45 (Biolegend, 368504, 1:200), and macrophages were labeled with CD86 (Biolegend, 374206, 1:200). Macrophages were then identified as CD45⁺CD86⁺ (Supplementary Fig. S33B). Samples were analyzed using a BD LSRFortessa flow cytometer (BD Biosciences, USA) and FACSDiva v8.0.1 (BD Biosciences, USA). The bacterial load was evaluated by harvesting the supernatant and performing serial dilutions, and 50 μL of each dilution was plated onto 7H10 agar. After 2 weeks of incubation at 37 °C, CFUs were counted and compared to the control group.

In vivo assessment of DCNP biodistribution and therapeutic efficacy

The biodistribution of DCNP in Mtb H37Ra-infected mice was monitored by labeling DNase, DNP, and DCNP with Cy5.5 probes. In vivo fluorescence images of mice were recorded from 0 h to 48 h post-intranasal injection using in vivo imaging system (IVIS Spectrum, USA). The fluorescent signals were analyzed using the Living Image 4.5 software (PerkinElmer, USA). To monitor DCNP accumulation in the mouse lung, mice were euthanized 48 h after the last administration, and major organs were collected for ex vivo imaging.

The therapeutic effect of DCNP against infected mice was monitored using 8-week-old BALB/c mice, which were infected with 2.5 × 10⁶ CFU Mtb H37Ra, Mtb H37Rv or M.smeg via intranasal injection. Then mice were treated with 50 μL DNase, DNP or DCNP (all containing 0.4 mg/mL DNase), via intranasal administration on day 1. On day 3, all mice were euthanized, and major organs and peripheral blood were harvested for further analysis. Mtb H37Ra-infected mice were euthanized, and major tissues were collected 60 days after DCNP treatment to evaluate the long-term safety and efficacy of DCNP. To investigate the immunological effects of chloroquine and lactoferrin, mice were administered with 50 μL of chloroquine (8 μg/mL) or lactoferrin (8 mg/mL) via intranasal injection for 3 days. For immunofluorescence staining of mouse lung tissues, primary antibodies specific to B-DNA, Z-DNA, Mtb antibody (Abcam, ab20962, 1:1000), myeloperoxidase (MPO, Abcam, ab208670, 1:1000), Ly-6G antibody (Biolegend, 127605, 1:1000) and CW were used. The samples were incubated with these primary antibodies overnight at 4 °C, followed by a 1 h incubation with secondary antibodies. The samples were examined using a confocal laser scanning microscope (CLSM), and fluorescence intensities were quantified using ImageJ software.

The function of macrophages was validated by flow cytometry and immunofluorescence staining. Mouse lung tissues were dissociated into single-cell suspensions using a digestion solution for flow cytometry. Debris and adherent cells were removed by gating on FSC-A/SSC-A and FSC-A/FSC-H. Monocytes were labeled with CD86 antibody (Biolegend, 159215, 1:200) and Mtb with Mtb antibody (Abcam, ab20962, 1:200). Mtb-infected macrophages were identified as CD86⁺Mtb⁺ by flow cytometry (Supplementary Fig. S33C). Lung tissues were also stained with primary antibodies against CD86 (Abcam, ab119857, 1:1000) and iNOS (Proteintech, 18985-1, 1:1000) overnight at 4 °C, followed by secondary antibodies and imaged using CLSM.

To assess immune exhaustion, lung tissues were lysed into single-cell suspensions, stained with CD4 (BD Biosciences, 550954, 1:50), CD8 (BD Biosciences, 553030, 1:50) and PD-1 (Biolegend, 109110, 1:50) antibodies (T cell exhaustion biomarker), and analyzed via flow cytometry (BD Biosciences, USA, 1:50) to determine the proportions of CD4⁺PD-1⁺ and CD8⁺PD-1⁺ cells. The bacterial load following DCNP activation of immune cells in vivo was assessed by homogenizing lung tissues and serially diluting and spotting the supernatants onto 7H10 agar plates. After 2 weeks culture at 37 °C, CFUs were counted and compared to the control group. H&E staining was performed using the H&E Staining kit (Beyotime, C0105S, CHN). The pathological score was determined based on the INHAND Proposal (International Harmonization of Nomenclature and Diagnostic Criteria for Lesions in Rats and Mice). The levels of inflammatory factors (TNF, IL-1β, IL-6, and IL-10) in serum of peripheral blood were measured using an ELISA kit (Biodragon, BDEL-0070 (TNF), BDEL-0057 (IL-1β), BDEL-0061 (IL-6), BDEL-0062 (IL-10), CHN) according to the manufacturer’s manual. The concentration of serum ALB (albumin), ALP (alkaline phosphatase), TBA (total bile acids), Crea (creatinine), UA (uric acid) and Urea and counts of blood WBC (white blood cells), NEUT (neutrophils) and RBC (red blood cells) in peripheral blood were measured using an automatic biochemistry analyzer (Chemray 240).

Assessing clinical specimen susceptibility to DCNP

A clinical outcome of TB diagnosis was based on AFBs, Mtb culture and GeneXpert MTB/RIF. Immunocompromised individuals or those with prior anti-TB therapy were excluded. Information on sputum, BALF, and pus samples from TB patients can be found in Supplementary Table S3. The cohort used in this study includes participants of both sexes. The ‘Male’ reported in Supplementary Table S3 is based on the standard of ‘sex’, i.e., defined based on currently understood biological differences between females and males, including chromosomes, sex, organs, and endogenous hormonal profiles.

The patient’s mouth was rinsed with water or cleaned with a toothbrush prior to sputum collection. Then, individuals were instructed to cough up sputum from the deep part of the respiratory tract and expel it directly into a sterile, leak-proof, wide-mouth container with a cap. The sample volume should be at least 1 mL. BALF was collected with three instillations of sterile physiological saline (50 mL) through a flexible bronchoscope. Pus samples were collected by performing pleural fibrous dissection to remove the abscess cavity from the thickened fibrous layers of the parietal and visceral pleura. The samples were then aspirated using a needle or syringe and transferred to sterile containers.

The susceptibility of clinical specimens (fresh sputum, BALF, or pus) was assessed by treating each with DNase (0.4 mg/mL), DNP, and DCNP, followed by incubation for 24 h at 37 °C. Cells and biofilms were snap-frozen in OCT for B-DNA/Z-DNA/biofilms/MPO labeling and neutrophil/macrophage visualization. For flow cytometry, debris and adherent cells were removed by gating on FSC-A/SSC-A and FSC-A/FSC-H. M1 macrophages were marked with CD86 (Biolegend, 374206, 1:200), monocytes with CD45 (Biolegend, 368504, 1:200), and neutrophils with CD66b (Biolegend, 392911, 1:200). CD66b⁺ and CD86⁺ cells were gated from CD45⁺ cells (Supplementary Fig. S33D). Bacterial survival was evaluated by CFU plate counting. For sputum, specimens were further incubated with rifampicin (1 µg/mL) or standard drug regimens (HRZE combination: isoniazid 0.2 µg/mL, rifampicin 1 µg/mL, pyrazinamide 50 µg/mL, and ethambutol 2 µg/mL) for 12 h at 37 °C to assess bactericidal effects with anti-TB drugs. Supernatants were serially diluted, and 50 μL of each dilution was spotted onto 7H10 agar plates. After 2 weeks at 37 °C, CFUs were counted and compared to the control group.

Statistical analysis

All the statistical data were presented as the mean ± standard deviation (s.d.) from at least three independent experiments. We used a Student’s t test for comparisons with only two groups and a one-way ANOVA with Tukey’s two-tailed post-hoc test for multiple groups. In all tests, P-values < 0.05 was considered statistically significant. *indicated P < 0.05, **indicated P < 0.01, ***indicated P < 0.001 and ns indicated no significant statistical difference. The statistical analysis was performed using SPSS (IBM SPSS 23.0, USA).

Reporting summary

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

Author contributions

M.B.Z., H.Z.L., D.L.L., P.F.Z. and T.A. designed the overall study. D.L.L., P.F.Z., Yang.Z., T.A., H.Z.L. and M.B.Z. designed all experiments. Z.J.H., Z.Q.L., G.Q.D., D.L.L., Yang.Z., P.F.Z., P.Q.Z., D.C.R. and S.L.S. performed the experiments. D.L.L., P.F.Z., S.L.S., P.Q.Z., Yao.Z., J.Z.P., X.F.D. and Y.Y.C. analyzed data. P.F.Z., M.B.Z., H.Z.L. and T.A. supervised the project. D.L.L., P.F.Z., M.B.Z. and Yang.Z. wrote the manuscript. P.F.Z., M.B.Z., T.A. and H.Z.L. reviewed and edited the manuscript. All authors have read and approved the final manuscript.

Peer review

Peer review information

Nature Communications thanks Derek Fleming, Jichan Jang, Rikke Meyer, David Pires, Rahul Verma and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was financially supported by the National Key R&D Program of China (2023YFA0915600), Shenzhen Medical Research Fund (D250402006, A2503081), Natural Science Foundation of China (82372271), Major Project of Guangzhou National Laboratory (GZNL2024A01009), Projects of International Cooperation and Exchanges NSFC (82561128247), Key Area Projects for Universities in Guangdong Province (2022DZX2022), China Postdoctoral Fund (2023M731525), Guangdong Province Medical Science and Technology Research Fund (A2024400), Guangdong Basic and Applied Basic Research Fund (2025A1515140076), Shenzhen Science and Technology Program (JCYJ20240813102021028, JCYJ20240813102012017, JCYJ20250604143832042), Shenzhen Clinical Medical Center for Emerging infectious diseases (LCYSSQ20220823091203007), and Shenzhen High-level Hospital Construction Fund (23274G1001, 23250G1005, 24250G1027, 24250G1019).

Data availability

All raw data generated for the figures in this study are provided in the source data file. The raw RNA-sequencing data generated in this study have been deposited in the Sequence Read Archive (SRA) under accession code PRJNA1181643. The remaining data are available within the Article, Supplementary Information or Source Data file. Source data are provided in this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77825-1.