Bioengineered extracellular vesicles escape lysosomal degradation and deliver Tet-PKM2 for macrophage immunometabolic reprogramming and periodontitis treatment

Abstract

Modulating macrophage phenotype and function via immunometabolic reprogramming represents a new therapeutic paradigm to combat chronic inflammatory diseases such as periodontitis. Tetrameric pyruvate kinase M2 (Tet-PKM2), a highly active metabolic enzyme involved in the flux of glucose-derived carbons into the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS), was found to be dramatically decreased in response to inflammation, rendering a potential immunometabolic target for developping new therapeutics. Hence, we report a large extracellular vesicle (LEV) that is bioengineered to intracellularly deliver Tet-PKM2 for the reprogramming of proinflammatory macrophages and the restoration of their aberrant immunometabolism. We engineered Tet-PKM2-enriched LEVs modified by tannic acid (LEVs^Tet−PKM2@TA) that can intracellularly deliver Tet-PKM2 and increase their ability to escape lysosomal degradation for the intracellular delivery of Tet-PKM2. In vitro, LEVs^Tet−PKM2@TA were able to rescue aberrant pyruvate metabolism in lipopolysaccharide (LPS)-activated macrophages by increasing TCA cycle activity and enhancing mitochondrial OXPHOS metabolism. In vivo, LEVs^Tet−PKM2@TA exerted robust immunomodulatory effects by increasing pyruvate kinase (PK) activity and coaxing macrophages toward the M2 phenotype, ultimately resulting in robust periodontal tissue regeneration in a mouse ligature-induced periodontitis model. This study provides a versatile and safe method for the targeted delivery of Tet-PKM2 via EVs to modulate macrophage phenotype and function. Our work demonstrates a new concept for immunometabolic reprogramming to treat chronic inflammatory diseases.

Graphical abstract

Schematic illustration of (A) the yield of bioengineered extracellular vesicles (LEVs^Tet−PKM2@TA) and (B) following LEVs^Tet−PKM2@TA uptake, cellular events (lysosomal degradation escape and cell immunometabolic reprogramming) occurred within macrophages. The up (down) arrows indicate increases (decreases) in the corresponding molecules. Abbreviations: Tet-PKM2, tetrameric pyruvate kinase M2; LEV, large extracellular vesicle; TA, tannic acid; PEP, phosphoenolpyruvate; LA, lactate; MMP, mitochondrial membrane potential; ATP, adenosine triphosphate; TCA, tricarboxylic acid; OXPHOS, oxidative phosphorylation.

Highlights

• •Tet-PKM2 downregulation causes aberrant immunometabolism and hyperinflammation in gingival macrophages during periodontitis.
• •TEPP-46 treatment of PKM2-overexpressing cells enables enrichment of Tet-PKM2 in their large extracellular vesicles (LEVs).
• •Tannic acid (TA) modification of Tet-PKM2-enriched LEVs (LEVs^Tet−PKM2@TA) can achieve lysosomal escape within macrophages.
• •LEVs^Tet−PKM2@TA coax LPS-activated M1 macrophages toward anti-inflammatory phenotypes through metabolic reprogramming.
• •LEVs^Tet−PKM2@TA combat periodontitis via immunometabolic reprogramming and promotion of M2 polarization in vivo.

1. Introduction

Chronic inflammation is a long-term pathophysiological state in response to bacteria, viruses, toxins, infections or maladaptive lifestyle factors [1,2]. Periodontitis is a chronic inflammatory disease characterized by the progressive destruction of tooth-supporting tissues [3,4]. The primary driver of periodontitis-induced tissue destruction is dysbiotic polymicrobial communities, which trigger dysregulated host immune responses [5]. As fundamental innate immune effector cells, macrophages perform diverse functions in host defense and inflammation resolution through the modulation of host‒microorganism interactions [[6], [7], [8]]. However, during the pathogenesis of periodontitis, a pronounced shift in macrophage polarization is evident: proinflammatory (M1) macrophages exhibit hyperactivation and substantial tissue infiltration, whereas anti-inflammatory (M2) macrophages are suppressed [9,10]. Accumulating evidence has indicated that fine-tuning the polarization of macrophages is a promising strategy for restoring tissue homeostasis and mitigating inflammatory tissue destruction [10,11]. Nevertheless, persistent M1 macrophage overactivation impedes repolarization toward the M2 phenotype by compromising metabolic reprogramming [12], thereby fostering the chronicity of inflammatory pathologies such as periodontitis.

Metabolic reprogramming is a fundamental mechanism that governs macrophage polarization and effector functions [[13], [14], [15]]. M1 macrophages exhibit increased glycolytic metabolism coupled with impaired tricarboxylic acid (TCA) cycle activity and mitochondrial oxidative phosphorylation (OXPHOS), enabling rapid effector responses to various proinflammatory stimuli. In contrast, M2 macrophages are characterized by an intact TCA cycle and enhanced OXPHOS, contributing to increased production of anti-inflammatory cytokines [16,17]. Although immune effector responses involve complex metabolic pathways, a shift from glycolysis toward the TCA cycle and OXPHOS is widely regarded as a hallmark of the transition from the M1 to the M2 macrophage phenotype. A growing body of evidence suggests that small molecules or metabolic intermediates may offer a promising therapeutic paradigm for metabolic reprogramming of macrophages [18]. Typical examples of metabolic-regulating agents include metformin (an FDA-approved antidiabetic drug that regulates fatty acid oxidation and glucose metabolism), 2-deoxyglucose (2-DG, a well-established glycolytic inhibitor), and NX-13 (a novel regulator of OXPHOS). These agents can effectively reprogram the metabolism of immune cells [[19], [20], [21]]. Nevertheless, the current applications of these agents are limited due to off-target effects, potential biosafety concerns, and low bioavailability attributable to lysosomal degradation [22,23]. Therefore, it is imperative to develop a more favorable and clinically promising strategy for immunometabolism regulation. The glycolytic enzyme pyruvate kinase (PK) is the final rate-limiting step of glycolysis, facilitating the generation of pyruvate from phosphoenolpyruvate (PEP) to fuel the TCA cycle [24,25]. In macrophages, the predominant expression of the PK isoform pyruvate kinase M2 (PKM2) orchestrates metabolic reprogramming by altering its conformational state [26]. Monomeric and dimeric PKM2 isoforms lack enzymatic activity but mediate nonmetabolic functions. Notably, dimeric PKM2 is translocated to the nucleus, where it interacts with hypoxia-inducible factor 1-alpha (HIF-1α) and subsequently promotes inflammatory responses or enhances the translation of proglycolytic enzymes [[26], [27], [28]]. In contrast, tetrameric PKM2 (Tet-PKM2) possesses highly active metabolic enzymatic activity, which can promote the flux of glucose-derived carbons into the TCA cycle and OXPHOS [29,30]. The allosteric activation of PKM2 by TEPP-46 significantly increases adenosine triphosphate (ATP) production in LPS-stimulated macrophages and promotes IL-10 production by activating adenosine [31]. These findings strongly suggest the critical role of PKM2 in regulating the metabolic reprogramming of macrophages and that increasing the expression of Tet-PKM2 in M1 macrophages may be an effective strategy for modulating the immune functions of macrophages during the pathogenesis of chronic inflammation.

Cell-derived extracellular vesicle (EV)-based carrier systems have attracted much interest in periodontitis treatment [[32], [33]]. EVs are nanoscale, phospholipid bilayer-enclosed structures that are extensively utilized as delivery vehicles to transport bioactive cargoes (e.g., proteins, lipids, and nucleic acids) to recipient cells [34,35]. Owing to their inherent targeting specificity toward macrophages, coupled with advantageous properties such as high versatility, low immunogenicity, and favorable safety profiles, EVs represent a promising platform for modulating macrophage polarization and effector functions via the targeted delivery of bioactive agents [32,34,36], which can avoid the undesirable off-target effects associated with systemically administered therapeutics. Furthermore, compared with small EVs (SEVs, 50–150 nm), large EVs (LEVs, 50–1000 nm) possess greater cargo content, endowing them with enhanced therapeutic potential for clinical translation [37,38]. Despite these advantages, only a few EVs can penetrate the cytoplasm of recipient cells, and most of their cargoes are internalized and degraded in the lysosome [39,40], leading to decreased therapeutic effects of EVs. Therefore, engineering EVs with lysosomal escape ability is imperative for improving their therapeutic performance.

In the present study, we aimed to engineer Tet-PKM2-enriched LEVs (LEVs^Tet−PKM2) with the ability to lysosomal escape to ensure the cytoplasmic transport of Tet-PKM2, thereby modulating the metabolic reprogramming of macrophages. LEVs^Tet−PKM2 were isolated from PKM2-overexpressing cells stimulated by TEPP-46, while the lysosomal escape ability of LEVs^Tet−PKM2 was enhanced by tannic acid (TA) modification, which can trigger surface charge reversal and the pH-responsive degradation of lysosomes [[41], [42], [43]]. The influence of LEVs^Tet−PKM2 on the immunometabolic reprogramming of macrophages in vitro and in vivo was further determined. We hypothesized that LEVs^Tet−PKM2@TA could escape lysosomal degradation and deliver Tet-PKM2 to macrophages to modulate immunometabolic reprogramming (Scheme 1), thereby providing a promising strategy for the treatment of chronic diseases such as periodontitis.

Scheme 1.

Schematic illustration of (A) the yield of bioengineered extracellular vesicles (LEVs^Tet−PKM2@TA) and (B) following LEVs^Tet−PKM2@TA uptake, cellular events (lysosomal degradation escape and cell immunometabolic reprogramming) occurred within macrophages. The up (down) arrows indicate increases (decreases) in the corresponding molecules. Abbreviations: Tet-PKM2, tetrameric pyruvate kinase M2; LEV, large extracellular vesicle; TA, tannic acid; PEP, phosphoenolpyruvate; LA, lactate; MMP, mitochondrial membrane potential; ATP, adenosine triphosphate; TCA, tricarboxylic acid; OXPHOS, oxidative phosphorylation.

2. Materials and methods

2.1. Human gingival tissue (healthy and periodontitis) samples

All the donors without periodontitis (healthy) who underwent third molar extraction and periodontitis patients who underwent flap debridement surgery signed informed consent forms. They voluntarily contributed their resected gingival tissue during surgery. The inclusion criteria and exclusion criteria for participant recruitment are described in detail in the supporting information. The gingival tissue samples were harvested during trimming of the surgical incisions for satisfactory wound suturing at the Stomatological Hospital, Fourth Military Medical University (FMMU). The use of human gingival tissue was conducted in accordance with a protocol approved by the Ethics Committee of the School of Stomatology, FMMU (KQ-YJ-2025-226). To determine metabolic and tetrameric PKM2 (Tet-PKM2) changes in response to periodontitis, the harvested gingival tissue samples were either freshly subjected to hematoxylin and eosin (H&E) staining, immunoelectron microscopy (IEM), or immunofluorescence analysis, or were rapidly frozen in liquid nitrogen and stored at −80 °C for metabolomic analysis, quantitative real-time polymerase chain reaction (qRT–PCR), and Western blot analysis.

2.2. Histological analysis

The regeneration of periodontal tissues and infiltration of immune cells in gingival tissues were observed via histological analysis. Human gingival tissue samples were collected and fixed in 4 % paraformaldehyde (PFA; Coolaber, China). After micro-CT scanning, the mouse maxillary samples were decalcified in 10 % EDTA solution (Proandy, China) for 2 weeks, and the solution was replaced every 3 days. Fixed human gingival tissues, decalcified maxillae and fixed organs from mice were subsequently dehydrated in graded ethanol solutions, embedded in paraffin, and sectioned at a thickness of 2 μm using a microtome (Leica, Germany). Each section was subsequently subjected to H&E staining or Masson's trichrome staining (Masson) to histologically evaluate the newly formed hybrid periodontal tissues.

2.3. Tissue sample IEM

IEM of tissue samples was conducted to investigate the subcellular distribution of PKM2 in gingival tissues collected from periodontally healthy donors (GT-H) and periodontitis patients (GT-P). Fresh gingival tissues were dissected into approximately 1 mm³ blocks within 1–3 min post-excision and immediately immersed in fresh IEM fixative at 4 °C (Servicebio, China). After being rinsed with phosphate-buffered saline (PBS) 3 times, the gingival tissue was dehydrated with graded ethanol, and resin penetration and polymerization were performed. Ultrathin sections (70–80 nm) were subsequently prepared using an ultramicrotome (Leica) and collected on nickel grids. Grid-mounted sections were blocked with 1 % BSA for 30 min at room temperature, followed by incubation with primary antibody against PKM2 (1:200; Invitrogen, PA5-29339, USA) overnight at 4 °C. After being rinsed with Tris-buffered saline (TBS) 3 times, the sections were incubated with the secondary antibody for 20 min at room temperature, then for 1 h at 37 °C, and finally for an additional 30 min at room temperature. After immunolabeling, the sections were sequentially stained with 2 % uranium acetate-saturated alcohol solution for 8 min and rinsed in 70 % ethanol or ultrapure water 3 times. After drying in a 37 °C oven for 10 min, representative IEM images were captured using transmission electron microscopy (TEM; Hitachi HT7800, Japan). Positive immunoreactivity was identified by the presence of 10 nm colloidal gold particles.

2.4. Metabolomic analysis

Targeted metabolomic analysis was employed to investigate metabolic activity in gingival tissues and the metabolic-modulatory effects of LEVs^Tet−PKM2@TA on LPS-pretreated macrophages in vitro. Metabolites were extracted from gingival tissues or cells standardized by either mass (∼50 mg) or cell count (∼5 × 10⁶ cells) via lysis in 500 μL of extractant and vortexed to mix. The lysates were cryogenically ground at 60 Hz for 1 min and then agitated at 1500 rpm for 10 min at 4 °C on a thermostatic metal shaker to maximize metabolite release. The samples were subsequently centrifuged at 12,000×g for 10 min at 4 °C, after which 100 μL of the supernatant was transferred to fresh tubes for evaporation to dryness under a gentle stream of nitrogen. Dried residues were redissolved in 50 % acetonitrile/aqueous solution and vortexed briefly to yield the alkaline loading reagent. Subsequently, an aliquot of the original supernatant was diluted 1:20, with 10 μL of the dilution evaporated to dryness and redissolved in 0.1 % formic acid-40 % methanol/aqueous solution (v/v) as an acid loading reagent, which was also centrifuged at 12,000×g for 10 min at 4 °C. The resulting supernatants were analyzed by liquid chromatography‒mass spectrometry (LC‒MS) for comprehensive metabolomic analysis. LC separations were performed on an ExionLC ultrahigh-performance LC system (AB Sciex, USA) equipped with an Agilent Eclipse XDB-C18 column (Agilent Technologies, USA). MS analysis was performed using an AB Sciex Triple Quadrupole 6500 plus mass spectrometer (AB Sciex). The software SCIEX OS (AB Sciex) was used to collect raw data and qualitatively and quantitatively analyze the results.

2.5. Immunofluorescence analysis

Immunofluorescence staining was performed to investigate the expression of PKM2 in CD68⁺ macrophages in gingival tissues and evaluate the immunomodulatory effect of LEVs^Tet−PKM2@TA on macrophages in vitro and in vivo. For PKM2 detection in gingival tissues and macrophage polarization analysis in vivo, tissue samples were collected and fixed in 4 % paraformaldehyde (PFA; Coolaber). Mouse maxillae were subsequently decalcified in 10 % EDTA (Proandy) for 2 weeks. The gingival and decalcified tissues were subsequently paraffin-embedded, cut into 2-μm-thick sections using a microtome (Leica), deparaffinized and subjected to antigen retrieval using 10 mM sodium citrate buffer. The tissue sections were subsequently permeabilized and blocked with 5 % bovine serum albumin (BSA; Beyotime, China) for 1 h at room temperature, after which each section was incubated with the following primary antibodies overnight at 4 °C: mouse anti-CD68 (1:200, Santa Cruz, SC-20060, USA) plus rabbit anti-PKM2 (1:200, Cell Signaling Technology, 4053S, USA), mouse anti-CD68 (1:200, Santa Cruz, SC-20060) plus rabbit anti-iNOS (1:200, Proteintech, 80517-1-RR, China), mouse anti-CD68 plus rabbit anti-CCR7 (1:200, Abcam, ab32527, UK), mouse anti-CD68 plus rabbit anti-CD163 (1:200, Proteintech, 16646-1-AP), mouse anti-CD68 plus rabbit anti-CD206 (1:200, Abcam, ab64693), or mouse anti-CD68 plus rabbit anti-Hif-1α (1:300, Cell Signaling Technology, 36169T). After being rinsed with PBS, the sections were incubated with the following corresponding secondary antibodies for 1 h at room temperature: donkey anti-rabbit IgG Alexa Fluor 488 (1:500; Abbkine, A23220, China) and donkey anti-mouse IgG Alexa Fluor 594 (1:500; Abbkine, A23410). The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Coolaber). Slides were mounted in anti-fade medium (Beyotime). Representative immunofluorescence images were obtained using a confocal laser scanning microscope (CLSM; Nikon, Japan), and quantitative analysis of the fluorescence intensity was performed using ImageJ software (NIH, USA). For macrophage polarization analysis in vitro, the same procedure was performed with slight modifications. Briefly, cells were fixed in 4 % PFA (Coolaber) for 30 min, rinsed, permeabilized, blocked, and incubated overnight with the following primary antibodies: rabbit anti-iNOS (1:200, Proteintech), rabbit anti-CCR7 (1:200, Abcam), rabbit anti-CD163 (1:200, Proteintech), rabbit anti-CD206 (1:200, Abcam), or rabbit anti-Arg-1 (1:200, Cell Signaling Technology, 93668T). After the sections were rinsed, they were incubated with the corresponding secondary antibodies, and immunofluorescence images were captured using CLSM (Nikon).

2.6. Quantitative real-time polymerase chain reaction (qRT–PCR)

qRT‒PCR was performed to quantify the expression levels of PKM2 in gingival tissue samples or M1 polarization-related genes (IL-6 and IL-1β) and M2 polarization-related genes (IL-4 and Arg-1) in macrophages. Briefly, total RNA was isolated from cells and tissues using TRIzol (Yeasen, China), and cDNA was synthesized using Hifair® III 1st Strand cDNA Synthesis SuperMix for qPCR (Yeasen) following the manufacturer's protocol. The cDNA samples were subjected to RT‒qPCR using qPCR SYBR Green Master Mix (Yeasen) according to the manufacturer's guidelines on a CFX96 real-time PCR system (Bio-Rad, USA). The data were normalized to β-actin values, and relative gene expression was calculated using the 2^−ΔΔCt method. The primers used were purchased from TSINGKE, and detailed primer information for every gene is listed in Table S1.

2.7. PKM2 crosslinking assay and western blot analysis

PKM2 crosslinking assay was employed to stabilize PKM2 in its monomeric, dimeric, and tetrameric states. Western blot analysis was performed to determine the expression levels of Tet-PKM2 in gingival tissues, LPS-activated macrophages, and EVs. In addition, the expression of EV-related proteins in SEVs and LEVs, and the expression levels of Hif-1α in LPS-activated macrophages after various treatments were analyzed by Western blot. For the PKM2 crosslinking assay, cells and tissues were collected and washed twice with PBS (pH 8.0). The crosslinking reaction was performed with 5 mM disuccinimidyl suberate (DSS; Thermo Scientific) at room temperature for 30 min, followed by termination with 20 mM Tris-HCl (pH 7.5; 1 M) for 15 min at room temperature. The samples were then washed and lysed for Western blot analysis. Cellular and tissue proteins were extracted using RIPA lysis buffer (Beyotime) supplemented with 1 % protease inhibitor (Beyotime). Protein concentrations were determined via a BCA assay. Equal amounts of protein lysates were separated by sodium dodecyl sulfate‒polyacrylamide gel electrophoresis (SDS‒PAGE; Epizyme, China) and electrophoretically transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, USA). The membranes were then blocked with 5 % skim milk dissolved in Tris-buffered saline containing 0.1 % Tween 20 (TBST) for 2 h at room temperature. After being washed, the membranes were incubated with the following primary antibodies overnight at 4 °C: anti-PKM2 (1:1000; Cell Signaling Technology, 4053S), anti-HSP70 (1:1000; Cell Signaling Technology, 4872T), anti-TSG101 (1:1000; Abcam, ab125011), anti-CD63 (1:1000; Cell Signaling Technology, SC55275), anti-calnexin (1:1000; Cell Signaling Technology, 2433T), anti-GM130 (1:1000; Cell Signaling Technology, 12480T), anti-Hif-1α (1:1000; Cell Signaling Technology, 36169T), and anti-β-actin (1:5000; Proteintech, 66009-1-Ig). The membranes were subsequently washed and probed with appropriate (anti-rabbit or anti-mouse) HRP-conjugated secondary antibodies (1:5000; Proteintech). Immunoreactive bands were visualized using an enhanced chemiluminescence (ECL) super kit (ABclonal, China) and imaged with a chemiluminescence imaging system (Bio-Rad). The level of β-actin was used as a control for total cell protein. Densitometry of the bands was quantified using ImageJ software (NIH).

2.8. Generation of PKM2-overexpressing cells

Cell lines of human embryonic kidney cells (HEK293T) and mouse macrophages (RAW 264.7) were purchased from Procell Life Science & Technology (China).

Cells were cultured in DMEM (Gibco, USA) supplemented with 10 % fetal bovine serum (FBS; Natocor, Argentina) and maintained in an incubator at 37 °C and 5 % CO₂. To overexpress the target PKM2 in HEK293T cells, the sequence of the target PKM2 was cloned and inserted into the lentiviral vector pLV-ZsGreen(2A) PURO-CMV (TSINGKE, China). HEK293T cells were then seeded in 6-well plates at a density of 4 × 10⁵ cells per well. When the percentage of cells reached 60 %, the culture medium of cells was replaced by DMEM medium with packaged lentiviruses containing the sequence of the target PKM2 (ov-PKM2, TSINGKE) (MOI = 20) overnight following the manufacturers’ protocols. The medium was then replaced with fresh complete DMEM medium. The nonspecific lentiviral vector (ov-NC) served as a negative control vector. Cells transfected with the lentiviruses were selected with 1 μg/mL puromycin (Thermo Scientific, USA). The culture medium was replaced with puromycin every two days until the death of cells was no longer observed. The overexpression of PKM2 was confirmed by qRT‒PCR and Western blotting.

2.9. Generation of Tet-PKM2-overexpressing cells

Tet-PKM2 in HEK293T cells was allosterically activated by TEPP-46 (MCE, HY-18657, USA). TEPP-46 was first dissolved in dimethyl sulfoxide (DMSO; MP Biomedicals, USA) to a stock concentration of 200 mM and then aliquoted and stored at −20 °C for further use. PKM2-overexpressing HEK293T cells generated via lentiviral transduction were seeded into 6-well plates at a density of 4 × 10⁵ cells per well. To determine the optimal TEPP-46 concentration, cells were treated with different concentrations of TEPP-46 (0, 10, 20, 40, and 70 μM) in complete medium for 24 h, after which pyruvate kinase (PK) activity was assessed. For time-course optimization, cells were treated with 40 μM TEPP-46 for 8 or 24 h, after which the conformational states of PKM2 were determined by Western blot analysis. According to the data obtained, incubation with 40 μM TEPP-46 for 24 h resulted in maximal Tet-PKM2 expression.

2.10. PK activity assays

The PK activity assay was performed to determine the optimal TEPP-46 concentration for the administration on the PKM2-overexpressing HEK293T cells, to measure the PK activity of EVs, and to quantify the PK activity in the gingival tissues of mice subjected to various treatments. The PK activity was determined using a pyruvate kinase assay kit (Solarbio, China) following the manufacturer's instructions. Briefly, 100 μL of extract solution was added to 5 × 10⁵ cells or 10 mg of tissue, followed by centrifugation at 8000×g for 10 min at 4 °C. The supernatant was subsequently collected for analysis. 10 μL of supernatant or EVs, 10 μL of detection solution III, and 180 μL of solution II were added to 96-well plates, followed by immediate mixing. The absorbances at a wavelength of 340 nm at 20 s and 2 min 20 s were recorded as A₁ and A₂, respectively, using a microplate reader (BioTek, USA). The ΔA value was calculated as A₂-A₁. The PK activity was calculated via the following equation:

For cellular PK activity:

PK (U/10⁴ cells) = [ΔA × V₁÷ (ε × d) × 10⁹] ÷ (500 × V₂ ÷ V₃) ÷ T = 3.216 × ΔA

For EVs and tissue PK activity:

PK (U/mg prot) = [ΔA × V₁ ÷ (ε × d) × 10⁹] ÷ (Cpr × V₂) T = 2680 × ΔA ÷ Cpr

V₁: total reaction volume, 2 × 10⁻⁴ L; ε: molar extinction coefficient of NADH, 6.22 × 10³ L/mol/cm; d: optical path length in 96-well plates, 0.6 cm; V₂: sample volume, 0.01 mL; V₃: extract solution volume, 0.1 mL; T: reaction time, 2 min; Cpr: sample protein concentration, mg/mL.

2.11. Isolation and characterization of the SEVs^Tet−PKM2 and LEVs^Tet−PKM2 from Tet-PKM2-overexpressing cells

Tet-PKM2-enriched SEVs (SEVs^Tet−PKM2) and LEVs^Tet−PKM2 were isolated by differential velocity centrifugation according to previously published methods [44,45]. First, PKM2-overexpressing HEK293T cells were seeded in 10 cm dishes overnight. Subsequently, the culture medium was replaced by EVs-depleted culture medium (DMEM plus 10 % EVs-depleted FBS) containing 40 μM TEPP-46 for another 24 h. EVs-depleted FBS was obtained by ultracentrifuging FBS at 100,000×g for 14 h using an SW 32 Ti Rotor Swinging Bucket rotor (Beckman Coulter, USA). Then the culture medium of Tet-PKM2-overexpressing HEK293T cells was collected and centrifuged at 300×g for 10 min at 4 °C to remove dead cells. The supernatant was subsequently centrifuged at 2000×g for 10 min at 4 °C to remove apoptotic bodies or cell debris. Next, the supernatant was centrifuged at 10,000×g for 50 min at 4 °C to pellet the LEVs^Tet−PKM2. Afterward, the LEV^Tet−PKM2-depleted supernatant was filtered through a 0.22 μm disposable membrane (Millipore) and further ultracentrifuged at 100,000×g for 70 min at 4 °C using an SW 32 Ti Rotor Swinging Bucket rotor (Beckman Coulter) to pellet the SEVs^Tet−PKM2. The LEVs^Tet−PKM2 and SEVs^Tet−PKM2 were resuspended in sterile PBS and stored at 80 °C before use. To characterize the SEVs and LEVs, the morphologies of the SEVs and LEVs were confirmed by TEM (Hitachi HT7800). The size distributions and particle concentrations were analyzed by nanoparticle tracking analysis (NTA) using a ZetaView instrument (Particle Metrix, Germany). The expression of specific EV-related proteins and Tet-PKM2 in SEVs and LEVs was analyzed by Western blotting, and the cellular uptake of SEVs and LEVs by RAW 264.7 cells was determined by immunofluorescence staining.

2.12. Cellular uptake of SEVs and LEVs by RAW 264.7 cells

A PKH 67 Green Fluorescent Cell Linker Kit (Sigma‒Aldrich, USA) was used to label the obtained SEVs and LEVs following the manufacturer's protocol. Briefly, the pelleted SEVs and LEVs were resuspended and incubated in 1 mL of diluent C containing 2 μL of PKH67 dye for 10 min at room temperature. The labeling reaction was terminated by the addition of 1 mL of FBS. Unbound dye was subsequently removed by centrifugation at 10,000×g for 50 min to obtain PKH67-labeled LEVs or at 100,000×g for 70 min to obtain PKH67-labeled SEVs at 4 °C. PKH67-labeled SEVs and LEVs resuspended in complete medium were used to treat RAW 264.7 cells for 24 h. The cells were then fixed with 4 % PFA (Coolaber), permeabilized, and stained for cytoskeletal visualization using fluorescein phalloidin (1:1000 dilution; MCE) for 30 min. The nuclei were counterstained with DAPI for 5 min. Representative fluorescence images were then captured using CLSM (Nikon).

2.13. Bioengineering LEVs through TA modifications

TA can be anchored to the phospholipid bilayer via hydrogen bond formation between its polyphenolic moieties and the phosphatidyl groups of cellular membranes or extracellular vesicles (LEVs). This interaction facilitates the generation of TA-modified cells or TA-modified LEVs (LEVs@TA) [41,46]. The HEK293T cell membrane was first modified with CY5-TA (QIYUE BIOLOGY, China). Powders of CY5-TA were dissolved in 2 mM DMSO and stored at −20 °C. The CY5-TA working solutions were prepared by dilution with PBS. The cells were incubated with various concentrations of CY5-TA (0, 0.1, 1, 5, and 10 μM) at 4 °C for 24 h. After incubation, the excess CY5-TA was removed from the cells by rinsing them twice with PBS. A NovoCyte flow cytometer (Agilent Technologies, USA) was subsequently used to determine the proportion of cells positively modified with CY5-TA. CY5-TA-labeled cells were also confirmed by CLSM (Nikon). Upon confirming successful phospholipid bilayer modification in cells, CY5-TA was further used to modify LEVs. LEVs were first incubated in various concentrations of CY5-TA working solution (0, 10, 20, 50, and 100 μM) at 4 °C for 24 h. Subsequently, the excess unbound CY5-TA was removed by centrifugation at 10,000×g for 50 min at 4 °C. The percentage of CY5-TA-labeled LEVs was subsequently measured using a NovoCyte flow cytometer (Agilent Technologies). Similarly, the CY5-TA-modified LEVs were evaluated by CLSM (Nikon), following the same procedure as previously described, except that only 50 μM CY5-TA was used. Therefore, for following TA modification of LEVs, TA (Merck, 403040, Germany) was dissolved in PBS with a concentration of 50 μM and incubated with LEVs at 4 °C for 24 h. After incubation, excessive TA were removed from LEVs by centrifugation at 10,000×g for 50 min at 4 °C. The interaction between TA and the cellular membrane during the uptake of LEVs and LEVs@TA by macrophages was evaluated by coarse-grained molecular dynamics (CGMD) simulations.

2.14. Lysosomal escape capacity of bioengineered LEVs@TA

Like in the cellular uptake assays, both the LEVs and LEVs@TA were fluorescently labeled using a PKH 67 Green Fluorescent Cell Linker Kit (Sigma‒Aldrich, PKH67GL). RAW 264.7 cells were seeded in confocal dishes (DEEYEE, China) at a density of 2 × 10⁵ per well and subsequently treated with 100 ng/mL LPS for 24 h. After LPS stimulation, the cells were incubated with PKH67-labeled LEVs or PKH67-labeled LEVs@TA. After an incubation period of 6 h, the cells were washed with PBS and stained with 75 nM LysoTracker Deep Red probe (Thermo Scientific, L12492) for 45 min in an incubator at 37 °C and 5 % CO₂. After being washed, the nuclei were counterstained with Hoechst 33342 (DEEYEE, DY20108) for 10 min. Immunofluorescence images were captured using lattice-structured illumination microscopy (Lattice SIM, ZEISS, Germany), and colocalization between LEVs and lysosomes was quantified by calculating the Pearson correlation coefficient using ImageJ software (NIH). For the calcein leakage assay, RAW 264.7 cells were seeded in a 24-well glass-bottom plates (Thermo Scientific) at a density of 4 × 10⁴ per well and stimulated with 100 ng/mL LPS for 24 h. After LPS stimulation, the cells were incubated with 150 μg/mL calcein alone (Merck, C0875) or combined with LEVs or LEVs@TA for 6 h. After the cells were washed with PBS 3 times, representative fluorescence images were immediately acquired using CLSM (Ex/Em = 495/515 nm) (Nikon). For bio-TEM analysis of morphology of endo/lysosomes in macrophages, RAW 264.7 cells were seeded in a 6-well plate at a density of 5 × 10⁴ cells per well and stimulated with 100 ng/mL LPS for 24 h. After stimulation, the cells were incubated with 100 μg/mL LEVs or LEVs@TA for 6 h. After incubation, ∼10⁷ cells were harvested and immediately fixed in 2.5 % glutaraldehyde solution at 4 °C for 12 h. After fixation, samples were rinsed with PBS and then fixed with 1 % OsO₄ + 2 % K₃[Fe(CN)₆] solution for 1 h. After being rinsed with PBS 3 times, the samples were dehydrated with graded ethanol, and resin penetration and polymerization were performed. Ultrathin sections (70–90 nm) were subsequently prepared using an ultramicrotome (Leica) and collected onto copper grids. The sections were put on copper grids and stained with 2 % uranyl ethanol solution for 4 min and lead citrate solution for another 4 min. After being rinsed with deionized H₂O, the grids were dried and applied for TEM analysis on a TEM (HITACHI). The interaction between TA and the lysosomal membrane after the uptake of LEVs and LEVs@TA by macrophages was evaluated by CGMD simulations.

2.15. Coarse-grained molecular dynamics (CGMD) simulations

CGMD simulations were performed to mimic the assembly process of TA onto membranes of LEVs, the stability of LEVs@TA following uptake by macrophages, or the disassembly of TA from the LEV membrane when they were entrapped in endo/lysosomes. For all coarse-grained MD simulations, the Martini3.0 force field [47] and GROMACS software [48] (version 2019.6) were used. The LEVs were constructed according to Zhuo's work [49], which consisted of 40.5 % DOPC, 32.7 % DOPE, 15.5 % PSM, and 11.4 % CHOL. The plasma membrane consisted of 40 % POPC, 24 % POPE, 8 % POPI, 9 % POPS, 1 % POPA, and 17 % PSM [50]. The lysosomal membrane comprised 39 % POPC, 14 % POPE, 5 % POPI, 2 % POPS, 1 % POPA and 20 % PSM [50]. TA can be fully protonated at pH < 4 and fully deprotonated at pH = 7 [51]. Deprotonated TA can form hydrogen bonds with mainly PE lipids. Insane.py [52] was used to construct the planar lipid bilayer systems, and bumpy.py was used for the spherical lipid bilayer systems. To determine the amount of TA molecules adsorbed onto the LEVs, we set up a system with one LEV and excess (145) TA molecules, which was run for 160 ns. The final TA-modified LEVs (128 TA molecules) were used for subsequent interactions with the plasma membrane and the lysosomal membrane. LEVs with/without TA molecules were initially placed close to the plasma and lysosomal membrane. The simulation box sizes for the plasma and lysosomal membranes were 39.4 nm × 39.4 nm × 28.2 nm and 39.3 nm × 39.3 nm × 28.6 nm, respectively. For the plasma membrane systems, LEV with/without TA molecules was pulled across the plasma membrane at 1 nm/ns (force constant k_B = 8000 kJ mol-1 nm-2) to model the endocytosis process of the LEVs. For all coarse-grained MD simulations, the recommended standard parameters for lipid membrane systems were used [47]. Pull simulations of plasma membrane systems were run for 20 ns each, while the lysosomal membrane systems were run for 200 ns. All the system snapshots were rendered by VMD [53].

2.16. Immunomodulatory effects of LEVs^Tet−PKM2@TA on LPS-activated macrophages

Based on the TA modification method established above, we bioengineered the LEVs^Tet−PKM2@TA using Tet-PKM2-enriched LEVs. RAW 264.7 cells were stimulated with 100 ng/mL LPS for 24 h. Then, the cells were treated with 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for 24 h. The metabolic states of the macrophages were analyzed by metabolomic analysis and a Seahorse XF test. Moreover, the mitochondrial functions of macrophages were analyzed by assessing the mitochondrial membrane potential (MMP) and ATP production. In addition, the polarization of macrophages was analyzed by qRT‒PCR, ELISA, and immunofluorescence staining.

2.17. Seahorse XF test

To investigate the metabolism of macrophages after various treatments, a Seahorse XF test was performed. Extracellular acidification rates (ECARs) and oxygen consumption rates (OCRs) were measured using a Seahorse XF96 Analyzer (Agilent Technologies, USA) according to the manufacturer's guidelines. Briefly, RAW 264.7 cells were seeded at a density of 5 × 10⁴ cells per well in a 96-well Seahorse plate. After being stimulated with 100 ng/mL LPS for 24 h, the cells were then incubated in 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for another 24 h. Before ECAR or OCR analysis, the cells were washed with XF medium (nonbuffered DMEM containing 2 mM L-glutamine and 1 mM sodium pyruvate) and incubated in a CO₂-free incubator for 45 min. The real-time ECAR was measured under basal conditions and after the following drugs were sequentially added: 10 mM glucose, 1 mM oligomycin, and 50 mM 2DG (XF Glycolysis Stress Test Kit, Agilent Technologies). The respiratory parameters were calculated as follows: Glycolysis: basic ECAR value prior to the oligomycin injection; Glycolysis capacity: difference between the maximum rate measurement after the Oligomycin injection and the non-glycolytic acidification rate; Glycolysis reserve: difference between the Glycolytic capacity and Glycolysis. OCR was measured under basal conditions and after the following drugs were added: 1 mM oligomycin, 1.5 mM FCCP, and 0.5 mM Rot/AA (XF Cell Mito Stress Test Kit, Agilent Technologies). The respiratory parameters were calculated as follows: Basal Respiration: basic OCR value prior to the oligomycin injection; Maximal Respiration: differences between the maximum rate measurement after FCCP injection and non-mitochondrial oxygen comsumption; ATP Production: difference between the final rate measurement prior to the oligomycin injection and the minimum rate measurement after the oligomycin injection.

2.18. Intracellular ATP concentration

The intracellular ATP concentration of RAW 264.7 cells was quantified using a commercial ATP assay kit (Beyotime). Following the prescribed experimental treatments, 1 × 10⁵ cells were lysed and centrifuged at 12,000×g for 10 min at 4 °C. 20 μL of the supernatant was then added to a black 96-well plate containing detection solution and mixed quickly before chemiluminescence was measured with a multimode microplate reader (TECAN, Switzerland).

2.19. Mitochondrial membrane potential (MMP)

The MMP of the macrophages was assessed using a JC-1 MMP Assay Kit (MCE). Briefly, RAW 264.7 cells were seeded in a 24-well glass bottom plate (Thermo Scientific) at a density of 4 × 10⁴ cells per well. After 24 h of stimulation with 100 ng/mL LPS, the cells were incubated with 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for an additional 24 h. After being washed with PBS twice, the cells were incubated in JC-1 staining working solution at 37 °C for 20 min. Representative fluorescence images were captured immediately using CLSM (Nikon). The fluorescence intensity was quantified using ImageJ software (NIH). The relative red/green fluorescence ratio was acquired as an index of the MMP (green, Ex/Em = 510/527 nm; red, Ex/Em = 585/590 nm).

2.20. Inhibition of LEVs^Tet−PKM2@TA-induced metabolic reprogramming in LPS-activated macrophages

To verify whether metabolic reprogramming is caused by LEVs^Tet−PKM2@TA promoting the generation of pyruvate and its entry into the TCA cycle, we inhibit the mitochondrial pyruvate carrier 1(MPC1) by a specific inhibitor, UK-5099. UK-5099 (MCE) was dissolved in DMSO, with a storage concentration of 100 mM. RAW 264.7 cells were stimulated with 100 ng/mL LPS for 24 h. Then, the cells were treated with 10 μM UK-5099, 100 μg/mL LEVs^Tet−PKM2@TA, or 10 μM UK-5099 plus 100 μg/mL LEVs^Tet−PKM2@TA for another 24 h. The metabolism states of macrophages were analyzed by a Seahorse XF test (OCR) and an ATP assay as described before.

2.21. Enzyme-linked immunosorbent assay (ELISA)

ELISA was performed to investigate the concentrations of pro-/anti-inflammatory cytokines secreted by macrophages. RAW 264.7 cells were seeded in 24-well plates at a density of 1 × 10⁵ cells per well and stimulated with 100 ng/mL LPS for 24 h. Next, the cells were treated with 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA. After a 24-h incubation period, the culture supernatants were collected and centrifuged at 3000×g for 10 min at 4 °C to remove cellular debris. The concentrations of M1-related cytokines (IL-6 and TNF-α) and M2-related cytokines (IL-4 and IL-10) in the supernatants of RAW 264.7 cells were determined using ELISA kits according to the manufacturer's instructions. Kits were sourced as follows: TNF-α, IL-4, and IL-10 from Fankew (Shanghai Kexing Trading Co., Ltd., China) and IL-6 from Proteintech. The absorbance was measured at a wavelength of 450 nm using a microplate reader (BioTek).

2.22. Administration of bioengineered LEVs^Tet−PKM2@TA for periodontitis treatment

A ligature-induced periodontitis model in murine was employed to evaluate the therapeutic effect of LEVs^Tet−PKM2@TA on periodontitis [54]. Male C57BL/6J mice (aged 7–8 weeks) sourced from the Animal Center of Air Force Medical University were used. Ethical approval for this animal study was granted by the Air Force Medical University Ethics Committee (AFMU-20240018). The mice were randomly divided into 5 groups (Control, PBS, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA). A murine experimental model of periodontitis was established by placing a ligature (a sterile 4-0 silk suture) into the interdental space between the maxillary first and second molars in the upper jaw to facilitate microbial colonization and periodontal inflammation for 9 days in mice except those in the Control group [55]. Afterward, the ligature was removed, and 10 μL of PBS, 100 μg of LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA dissolved in 10 μL of PBS was injected into the gingival sulcus between the maxillary first and second molars every other day. At 24 h post-injection, LEV uptake in vivo was analyzed by immunofluorescence staining (for details, see the supporting information). On day 15 after surgery, the animals were euthanized by cervical dislocation. maxillae were subsequently obtained and subjected to PK activity assays, micro-CT analyses, histological analysis, and immunofluorescence staining. The systemic organs (heart, liver, spleen, lungs, and kidneys) were harvested for evaluation of histocompatibility by histological analysis, and the blood was collected for blood chemistry analysis.

2.23. Micro-CT analyses

The fixed maxillary samples were first analyzed by micro-CT scanning using a high-resolution micro-CT scanner (AX2000 CT, Always Imaging, China). Three-dimensional (3D) reconstructed models were generated and analyzed quantitatively using VG Studio MAX 3.5.1 software (Volume Graphics, Germany). The distance from the cementoenamel junction to the alveolar bone crest (CEJ-ABC) of the maxillary first molar was quantified to evaluate periodontal bone loss. The trabecular microarchitecture within the interdental alveolar bone between the first and second molars was assessed through morphometric parameters, including trabecular separation (Tb.Sp), trabecular number (Tb.N), and bone volume fraction (BV/TV).

2.24. Evaluation of in vivo stability and biodistribution of LEVs@TA

In vivo fluorescence imaging was performed to trace the stability of LEVs@TA, and to assess the biodistribution of LEVs. Firstly, LEVs were stained with DIO dye (DEEYEE) at 4 °C for 10 min. After incubation, the excess DIO dye was removed from the LEVs by centrifugation at 10,000×g for 50 min at 4 °C. DIO-labeled LEVs were then incubated with 50 μM CY5-TA (QIYUE BIOLOGY) at 4 °C for 24 h. Subsequently, the excess unbound CY5-TA was removed by centrifugation at 10,000×g for 50 min at 4 °C. 100 μg of DIO-labeled-LEVs@CY-5-TA were subcutaneously injected into the back of mice. At different time points (6, 12, 24, and 48 h) after injection of DIO-labeled-LEVs@CY5-TA, the fluorescence distribution in vivo was visualized and imaged using an in vivo fluorescence imaging system (DIO: Ex/Em: 465/540 nm; CY5: Ex/Em: 640/700 nm; AniView100/600 DXA, Guangzhou Biolight Biotechnology, China). The fluorescence intensity was analyzed by AniView X Software (Guangzhou Biolight Biotechnology).

2.25. Statistical analysis

All experiments were performed with a minimum of three biological replicates. Data analysis was performed using GraphPad Prism version 8 (GraphPad Software Inc., USA). For comparisons among more than two groups, we used one-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test or Dunnett's multiple comparisons test. Two-group comparisons were conducted using unpaired 2-tailed Student's t tests. All values are expressed as the mean ± SEM. P values less than 0.05 were considered to indicate statistical significance (∗P < 0.05, ∗∗P < 0.01, or ∗∗∗P < 0.001). Detailed method information and the number of biological replicates are provided in the corresponding figure legends.

3. Results

3.1. Metabolic activity and tetrameric PKM2 (Tet-PKM2) expression changes in periodontitis gingival tissues

To investigate the metabolic changes in gingival tissues during the development of periodontitis, gingival tissues obtained from 4 periodontally healthy donors (GT-H) and periodontitis patients (GT-P) were subjected to targeted energy metabolomics analysis (Fig. 1A). Representative intraoral photographs showing the clinical appearance of periodontium in periodontally healthy donors and periodontitis patients (Fig. S1A), and the infiltration of immune cells in the GT-P was further confirmed by H&E staining (Fig. S1B). Partial least squares discriminant analysis (PLS-DA) revealed significant metabolic variations between the GT-H and GT-P groups (Fig. 1B). To functionally characterize the differentially detected metabolites, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was conducted, and the results revealed that the differentially detected metabolites were enriched in the TCA cycle and glycolysis/gluconeogenesis pathways (Fig. 1C). The metabolites involved in glycolysis and the TCA cycle are further shown in Fig. 1D. Quantitative analysis of glycolysis indicated that phosphoenolpyruvate (PEP) and lactate (LA) levels were significantly higher in GT-P than in GT-H, although the production of pyruvate remained unchanged (Fig. 1E), indicating enhanced glycolytic flux in GT-P [56]. Pyruvate is a key metabolite that links both glycolysis and the TCA cycle, and its conversion to LA rather than entering the TCA cycle can disrupt the TCA cycle and OXPHOS [26]. As expected, the production of fumarate (FA) and malate (MA), which may represent disrupted the TCA cycle [26,57], increased significantly in GT-P (Fig. 1E and F). The conversion of PEP to pyruvate has been reported to be related to the conformational state of PKM2. Specifically, Tet-PKM2 catalyzes the conversion of PEP to pyruvate and enhances the TCA cycle, whereas dimeric or monomeric forms are translocated to the nucleus and promote glycolysis and inflammatory responses in macrophages [31]. To investigate the expression and conformational states of PKM2 in GT-H and GT-P, immunofluorescence staining, qRT‒PCR, and DSS crosslinking followed by Western blot analysis and IEM were conducted. Representative immunofluorescence images revealed increased expression levels of the PKM2 protein in CD68⁺ macrophages in the GT-P group compared with those in the GT-H group (Fig. 1G). A similar phenomenon could also be observed in ligature-induced periodontitis (LIP) in mice (Fig. S2). Furthermore, the relative mRNA expression levels of PKM2 in the GT-P group were significantly higher than those in the GT-H group (Fig. 1H). However, compared with those in GT-H, the expression levels of Tet-PKM2 in GT-P decreased significantly (Fig. 1I and J). Representative IEM images clearly revealed the presence of PKM2-immunoreactive particles in the cytoplasm of macrophages in GT-H compared with those in GT-P (Fig. 1K). To investigate the changes in PKM2 expression levels in macrophages during proinflammatory responses, macrophages were activated by LPS, after which the expression and conformational states of PKM2 were investigated. Consistently, LPS treatment significantly decreased Tet-PKM2 expression (Fig. S3B and C), although the total expression level of PKM2 increased (Fig. S3A). Taken together, these findings indicated that tetrameric PKM2 expression levels were lower in the GT-P than in GT-H, especially in macrophages, which might be associated with metabolic dysfunctions.

Fig. 1.

Metabolic dysfunction and decreased Tet-PKM2 expression in periodontitis gingival tissue. (A) Schematic illustration of gingival tissue sample collection from periodontally healthy donors (GT-H) and periodontitis patients (GT-P). (B) Metabolites differ between GT-H and GT-P (PLS-DA). (C) Altered metabolism from GT-H to GT-P, including the TCA cycle and glycolysis (KEGG enrichment analysis). (D) Heatmap representing differentially detected metabolites in GT-H and GT-P involved in glycolysis and the TCA cycle. (E) Quantitative analysis of the changes in the expression of key glycolysis and TCA metabolites from GT-H to GT-P. (F) Schematic illustration of increased glycolysis and disruption of the TCA cycle. (G) Increased protein levels of PKM2 in CD68⁺ macrophages in the GT-P group (immunofluorescence assay). (H) Relative mRNA expression levels of PKM2 in GT-H and GT-P (qRT‒PCR) (n = 3). (I) Representative immunoblot bands of PKM2 conformational states in GT-H and GT-P (DSS cross-linking assay). (J) Quantitative analysis of Tet-PKM2 expression in GT-H and GT-P (n = 7). (K) Decreased cytoplasmic distribution of PKM2 in macrophages from GT-P compared with those from GT-H (immunoelectron microscopy). The data are expressed as the mean ± SEM. Statistical analysis was performed with Student's t- test. ns indicates no significant difference between the indicated columns; ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

3.2. TEPP-46 treatment results in abundant intracellular Tet-PKM2 expression and facilitates the generation of Tet-PKM2-enriched LEVs

To increase Tet-PKM2 expression in cells, lentiviral transduction was first performed to overexpress PKM2 in HEK293T cells, followed by allosteric activation of Tet-PKM2 by TEPP-46 treatment (Fig. 2A). Both qRT‒PCR and Western blot analyses revealed that lentiviral transduction significantly increased the expression level of PKM2 (Fig. 2B–D). To determine the optimal concentration of TEPP-46, various concentrations of TEPP-46 (0, 10, 20, 40, and 70μM) were used to allosterically activate Tet-PKM2 for 24 h. The results of the PK activity analysis indicated that compared with the Control, 40 μM TEPP-46 significantly increased PK activity (Fig. 2E), thereby serving as the optimal concentration. To assess the effect of treatment duration on the conformational state of PKM2, a DSS cross-linking assay followed by Western blotting was performed. This assay revealed that 24 h of TEPP-46 (40 μM) stimulation led to the highest expression levels of Tet-PKM2 (Fig. 2F and G).

Fig. 2.

Generation of PKM2-overexpressing cells and generation of Tet-PKM2-enriched LEVs. (A) Schematic illustration depicting the overexpression of PKM2 in HEK293T cells by lentiviral transduction and subsequent allosteric activation of Tet-PKM2 via TEPP-46 treatment. Cells transfected with ov-NC or ov-PKM2 were named the NC or OV, respectively. (B) Relative mRNA expression levels of PKM2 in the NC and OV groups (qRT‒PCR, n = 4). (C) Representative immunoblot bands of PKM2 in the NC and OV groups. (D) Semiquantitative analysis of the expression levels of PKM2 in the NC and OV groups (n = 6). (E) PK activity in PKM2-overexpressing HEK293T cells in response to treatment with various concentrations of TEPP-46 (0, 10, 20, 40, and 70 μM) (n = 6). (F) Representative immunoblot bands of PKM2 conformational states in PKM2-overexpressing cells after treatment with TEPP-46 (40 μM) for 0, 8, and 24 h (DSS cross-linking assay). (G) Semiquantitative analysis of the expression levels of Tet-PKM2 in PKM2-overexpressing cells after treatment with TEPP-46 (40 μM) for 0, 8, and 24 h (n = 4). (H) Schematic illustration of the isolation of LEVs and SEVs by differential velocity centrifugation. (I) Representative TEM images showing the morphology of SEVs and LEVs. (J) Size distributions of SEVs and LEVs (NTA). (K) Particle counts of SEVs and LEVs (NTA). (L) Particle-to-protein ratios of SEVs and LEVs. (M) Cellular uptake of SEVs and LEVs (labeled with PKH67; green) by macrophages (immunofluorescence assay); cell skeletons were stained with phalloidin (red), and nuclei were stained with DAPI (blue). (N) Expression of CD63, HSP70, TSG101, calnexin, and GM130 in whole-cell lysates (Cells), SEVs, and LEVs (Western blot). (O) Representative immunoblot bands of PKM2 conformational states in LEVs and SEVs. (P) Semiquantitative analysis of the expression levels of Tet-PKM2 in LEVs and SEVs (n = 3) (Q) Conformational states of PKM2 in LEVs^Tet−PKM2 in response to TEPP-46 treatment. (R) Semiquantitative analysis of the expression levels of Tet-PKM2 in LEVs following TEPP-46 treatment (n = 3). The data are expressed as the mean ± SEM. Statistical analysis was performed by one-way ANOVA (E and G) and Student's t-test (B, D, K, L, P, and R). ns indicates no significant difference between the indicated columns; ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

SEVs and LEVs derived from HEK293T cells were subsequently isolated via differential velocity centrifugation, respectively (Fig. 2H). The obtained EVs were subsequently characterized by TEM, NTA, and Western blot. Representative TEM images revealed that LEVs were much larger than SEVs, although they exhibited similar spherical vesicle structures (Fig. 2I). Consistently, NTA indicated that compared with the diameter of the SEVs (131.8 nm), that of the LEVs (218.8 nm; Fig. 2J) increased. The particle counts showed no difference between SEVs and LEVs (Fig. 2K). Meanwhile, the particle-to-protein ratios of the LEVs and SEVs both exceeded 1 × 10⁸ particles/μg, indicating the purity of the isolated EVs (Fig. 2L) [58,59]. Furthermore, representative immunofluorescence images revealed that PKH67-labeled SEVs and LEVs could both be internalized by RAW 264.7 cells after an incubation period of 24 h (Fig. 2M). Moreover, Western blot analysis revealed that both SEVs and LEVs expressed the CD63, HSP70, and TSG101 proteins but did not express the calnexin or GM130 protein (Fig. 2N). To assess the relative expression levels of Tet-PKM2 in SEVs and LEVs, DSS crosslinking followed by Western blot analysis was performed. Representative immunoblot bands and quantitative analysis revealed higher expression levels of Tet-PKM2 in LEVs than in SEVs (Fig. 2O and P). Moreover, TEPP-46 stimulation further increased Tet-PKM2 expression in LEVs (Fig. 2Q and R). To evaluate the storage stability of LEVs^Tet−PKM2@TA, we assessed the physical integrity and biological activity of freshly prepared LEVs after 1 week and 1 month of storage at −80 °C using TEM and pyruvate kinase (PK) activity assays, respectively. TEM analysis revealed that most LEVs^Tet−PKM2@TA showed similar morphology comparable to freshly prepared LEVs^Tet−PKM2@TA after 1 week of storage. However, the morphology of LEVs^Tet−PKM2@TA particles changed after 1 month of storage, with irregular membrane edges, suggesting that prolonged storage compromised the structural integrity of the membrane (Fig. S4A). PK activity assays demonstrated no significant differences between freshly prepared LEVs^Tet−PKM2@TA and those after 1 week of storage. However, storage for 1 month significantly decreased the PK activity of LEVs^Tet−PKM2@TA compared with freshly prepared samples or storage for 1 week (Fig. S4B). Based on these findings, we conclude that the biological activity of LEVs^PKM2@TA is maintained for at least one week under −80 °C storage but notably decreased after one month. Moreover, to exclude any potential influence of the lentiviral vector system on PKM2 enzymatic activity of LEVs, we investigate the influences of empty lentivirus vector transfection on PK activity of LEVs from TEPP-46 treated HEK293T cells. The PK activity result showed no significant difference between two groups (Fig. S5). Taken together, these findings indicated that lentiviral transduction combined with TEPP-46 stimulation increased the enrichment of Tet-PKM2 in LEVs, named LEVs^Tet−PKM2, which could be used for further investigations.

3.3. Bioengineered LEVs through in situ TA modifications (LEVs@TA) exhibit a robust capacity for lysosomal escape following uptake by macrophages

To avoid the degradation of LEVs@TA by the lysosomal system, we modified the surface of the LEVs with TA. TA molecules can bind reversibly to the phospholipid bilayer, which is facilitated by the formation of hydrogen bonds between the phosphate groups of the phospholipid bilayer and the polyphenol groups of the TA molecule (Fig. 3A) [41,46]. When HEK293T cells were incubated with 0, 0.1, 1, 5, and 10 μmol/L CY5-TA for 24 h, flow cytometry analysis revealed that CY5-TA could self-assemble on the membrane of HEK293T cells, yielding 0.017 %, 3.05 %, 57.6 %, 96.5 %, and 99.4 % CY5-positive cells, respectively (Fig. 3B). Consistent with these findings, representative fluorescence microscopy images revealed that most cells displayed CY5-TA on their surface after incubation with 10 μmol/L CY5-TA (Fig. 3C). These results suggested that TA could directly bind to the phospholipid bilayer of the cell membrane. Given that the phospholipid bilayer of LEVs is similar to that of their parent cells, we hypothesized that TA could also interact with the phospholipid bilayer of LEVs. Flow cytometry analysis revealed that 0.63 %, 86.8 %, 90.8 %, 98.9 %, and 99.9 % of the LEVs were labeled with CY5-TA after incubation with 0, 10, 20, 50, and 100 μmol/L CY5-TA, respectively, at 4 °C for 24 h (Fig. 3D). In line with the results of the flow cytometry analysis, representative fluorescence microscopy images revealed that most LEVs were labeled with CY5-TA after incubation with 50 μmol/L CY5-TA at 4 °C for 24 h (Fig. 3E). The assembly of TA and LEVs was further simulated by CGMD, which visually showed the assembly process of TA onto the LEV membrane (Movie S1). Collectively, these findings confirmed that TA successfully bound to the membrane of the LEVs. Given that the cellular membrane may interact competitively with the TA during endocytosis, which might lead to the disassembly of LEVs@TA, we further investigated the stability of LEVs@TA when they were internalized by macrophages based on CGMD simulations. Dynamic changes were recorded every 5 ns over a total simulation period of 20 ns. As shown in Fig. 3F and Movies S2 and S3, both LEVs and LEVs@TA were successfully internalized by macrophages. Crucially, LEVs@TA maintained structural integrity during the uptake process (Fig. 3F). Furthermore, in vivo fluorescence imaging was used to determine the stability of LEVs@TA. We labeled the LEVs with DIO dyes and modified them with CY5-TA. After subcutaneous injection of LEVs@TA into the backs of mice, representative images revealed that the amounts of both DIO-labeled LEVs and CY5-labeled TA decreased with increasing incubation time, and the co-expressing fluorescence signals could be observed on 48 h post injection (Fig. 3G), which indicating good in vivo stability of LEVs@TA for at least two days. In conclusion, these findings indicated successful TA modification on the surface of LEVs and excellent structural integrity of LEVs@TA during both the cellular uptake process and under physiological in vivo conditions. (Fig. 3A).

Fig. 3.

Generation of bioengineered LEVs (LEVs@TA) through in situ TA modifications. (A) Schematic illustration showing the interaction of TA with the phospholipid bilayer of LEVs via hydrogen bonding and the uptake of LEVs@TA by macrophages. (B) The percentages of CY5-TA-modified cells following incubation with gradient concentrations of CY5-TA (0, 0.1, 1, 5, and 10 μM) for 24 h (flow cytometry assay). (C) Colocalization of CY5-TA on HEK293T cells following incubation in 10 μM CY5-TA for 24 h (fluorescence microscopy). The nuclei were stained with DAPI (blue). (D) The percentages of CY5-TA-modified LEVs following incubation with gradient concentrations of CY5-TA (0, 10, 20, 50, and 100 μM) for 24 h (flow cytometry assay). (E) Colocalization of CY5-TA and PKH67-labeled LEVs (green) following incubation with 100 μM CY5-TA for 24 h (fluorescence microscopy). (F) Snapshots of CGMD simulations depicting the uptake of LEVs and LEVs@TA by macrophages at 0, 5, 10, 15, and 20 ns. (G) Representative in vivo fluorescence images showing good stability of DIO-labeled-LEVs@CY5-TA in vivo.

It has been reported that acidic environments in endo/lysosomes (pH 4.5–7.0) trigger protonation of TA, which then induces lysosomal swelling through the proton sponge effect [43,46]. To verify whether TA could disassemble from LEVs in endo/lysosomes, CGMD simulations were performed to investigate the changes in LEVs@TA in an acidic environment. As shown in the snapshots, TA disassembled from the surface of the LEVs when they were entrapped in endo/lysosomes due to the protonation of TA (Fig. 4B and Movie S4). To investigate whether TA modification could enhance the endo/lysosomal escape capability of LEVs, the intracellular distribution of both LEVs and LEVs@TA was tracked using a LysoTracker Deep Red probe. Representative fluorescence microscopy images showing that the fluorescence signals of the PKH67-labeled LEVs overlapped with the LysoTracker fluorescence signals. In contrast, only a small portion of the PKH67-labeled LEVs@TA fluorescence overlapped with the LysoTracker signal (Fig. 4C). The Pearson correlation coefficient of the LEVs@TA was also significantly lower than that of the LEVs (Fig. 4D), suggesting that the colocalization of the LEVs@TA and endo/lysosomes was weaker than that of the LEVs. Furthermore, the stability of the endo/lysosomal membranes was investigated via a calcein leakage assay. When entrapped within intact endo/lysosomes, calcein displays punctate fluorescence because of self-quenching at high local concentrations, whereas endosomal membrane rupture facilitates calcein release, resulting in a homogeneous fluorescence distribution by dequenching. (Fig. 4E). Representative fluorescence microscopy images revealed punctate fluorescence in the cytoplasm of macrophages incubated with PBS or LEVs (Fig. 4F). In contrast, diffuse and enhanced fluorescence throughout the cells was observed in macrophages incubated with LEVs@TA (Fig. 4F), suggesting that the membranes of the endo/lysosomes were destabilized. The escape process of LEVs@TA from lysosomes was also confirmed by TEM analysis. Compared to LEVs group, the endo/lysosomes of macrophages in LEVs@TA group exhibited marked swelling and loss of membrane integrity, with visible signs of destabilization, including membrane discontinuities and partial disintegration. These morphological changes are indicative of lysosomal membrane permeabilization, which facilitates the escape of LEVs@TA from the lysosomal compartment (Fig. 4G). Collectively, these findings indicated that TA modification promoted the endo/lysosomal escape of LEVs in macrophages, which might have enhanced the cytosolic delivery of Tet-PKM2.

Fig. 4.

Endo/lysosomal escape capacity of bioengineered LEVs@TA following uptake by macrophages. (A) Schematic illustration showing the endo/lysosomal escape process of LEVs@TA within the cytoplasm of macrophages. After uptake by macrophages, LEVs@TA were entrapped within endo/lysosomes, and then TA underwent protonation and disassembled from LEVs in an acidic environment, leading to rupture of the endo/lysosomal structure. (B) Snapshots of CGMD simulations showing the disassembly of TA and LEVs in the lysosomal environment. (C) Colocalization of LysoTracker-labeled endo/lysosomes (violet) and PKH67-labeled LEVs or LEVs@TA (green) (fluorescence microscopy). The nuclei were stained with Hoechst (blue). (D) Quantification of the colocalization of endo/lysosomes and LEVs or LEVs@TA using the Pearson correlation coefficient (n = 12). (E) Schematic illustration showing the leakage of calcein into the cytosol when TA diffused from LEVs@TA and destabilized the endo/lysosomal membranes. (F) The distribution of calcein (green) in macrophages treated with PBS, LEVs, and LEVs@TA (fluorescence microscopy). (G) Representative TEM images of macrophages showing the structure of lysosomes in macrophages treated with LEVs and LEVs@TA. The data are expressed as the mean ± SEM. Statistical analysis was performed with Student's t-test (D). ∗∗∗p < 0.001 indicates significant differences between the indicated columns.

3.4. Bioengineered LEVs^Tet−PKM2@TA deliver Tet-PKM2 for macrophage metabolic reprogramming

To validate the cell viability of the LEVs^Tet−PKM2@TA, a CCK-8 assay was performed to evaluate the proliferation of macrophages treated with PBS (Control), LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA. The CCK-8 results showed an increasing trend in cell growth in all the tested groups, indicating that the bioengineered EVs elicited no obvious toxic effects on cell viability (Fig. S6). Given that Tet-PKM2 facilitates glycolytic flux into the TCA cycle and OXPHOS [29,30], we further investigated the effects of LEVs^Tet−PKM2@TA on macrophage metabolism. The macrophages were first stimulated with 100 ng/mL LPS for 24 h, followed by a 24-h incubation in PBS, LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA. Targeted metabolomic analysis was subsequently performed to determine the concentrations of metabolites involved in glycolysis and the TCA cycle (Fig. 5A). Quantitative analysis revealed that both LEVs^Tet−PKM2 and LEVs^Tet−PKM2@TA significantly decreased the levels of PEP, which is the direct catalytic substrate of Tet-PKM2 [24]. Moreover, the concentration of 2-phosphoglycerate (2-PG), an upstream glycolytic metabolite, decreased significantly in the LEVs^Tet−PKM2 and LEVs^Tet−PKM2@TA groups, indicating that the delivery of Tet-PKM2 by LEVs promoted the consumption of glycolytic metabolites. With respect to the TCA cycle, the levels of energy production, including those of guanosine triphosphate (GTP) and ATP, increased significantly following incubation with LEVs^Tet−PKM2@TA. Moreover, the production of succinate (SA), which accumulates in macrophages in response to LPS stimulation and acts as an inflammatory signal contributing to increased IL-1β expression [60], significantly decreased following treatment with LEVs^Tet−PKM2@TA. Moreover, the production of FA significantly decreased in the LEVs^Tet−PKM2@TA group compared with that in the LEVs^Tet−PKM2 group, indicating that the disruption of the TCA cycle was reversed by LEVs^Tet−PKM2@TA treatment (Fig. 5B). The spatial distribution of significantly changed glycolysis and TCA cycle metabolites in the LEVs^Tet−PKM2@TA group compared with those in the Control group are shown in Fig. 5C. These data suggest that LEVs^Tet−PKM2@TA increased the immunometabolic energy of LPS-pretreated macrophages and prevented the accumulation of glycolytic or TCA cycle intermediates related to proinflammatory responses. The immunometabolic profile of LPS-pretreated macrophages was further verified by extracellular flux analysis. An evaluation of glycolytic activity revealed an increased extracellular acidification rate (ECAR) in LPS-pretreated macrophages following incubation in LEVs^Tet−PKM2@TA (Fig. 5D). Quantitative analysis of the ECAR revealed that, compared with the Control and LEVs^PKM2 groups, the LEVs^Tet−PKM2@TA group had significantly elevated levels of glycolytic parameters, including glycolysis, glycolysis capacity, and glycolysis reserve (Fig. 5E). An evaluation of the OXPHOS process revealed an increased real-time oxygen consumption rate (OCR) in both the LEVs^Tet−PKM2 and the LEVs^Tet−PKM2@TA groups compared with that in the Control or LEVs^PKM2 groups (Fig. 5F). Quantitative analysis of the OCR indicated that, among all the tested groups, the LEVs^Tet−PKM2@TA treatment resulted in the greatest increase in basal respiration. Compared with those in the Control or LEVs^PKM2 groups, maximal respiration in the LEVs^Tet−PKM2@TA group was significantly greater. Moreover, ATP production was significantly greater in both the LEVs^Tet−PKM2 and the LEVs^Tet−PKM2@TA groups than in the Control or LEVs^PKM2 groups (Fig. 5G). Collectively, these results indicated that LEVs^Tet−PKM2@TA promoted the metabolic reprogramming of LPS-activated macrophages.

Fig. 5.

Metabolic reprogramming and enhanced mitochondrial function in LPS-activated macrophages in response to LEV^Tet−PKM2@TA treatment. The macrophages were pretreated with 100 ng/mL LPS for 24 h and then treated with PBS (Control), 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for another 24 h. (A) Heatmap representing differentially detected metabolites involved in glycolysis and the TCA cycle in the Control, LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA groups (n = 4). (B) Concentrations of key glycolysis and TCA metabolites in Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 4). (C) Schematic illustration revealing changes in key glycolysis and TCA metabolites in the LEVs^Tet−PKM2@TA group versus the Control group. The up (down) arrows indicate increased (decreased) levels of metabolites in macrophages. (D) Kinetic profile of the ECAR in LPS-activated macrophages in response to sequential injections of glucose, oligomycin, and 2-DG in various groups (Seahorse XF test) (n = 4). (E) Quantification of glycolysis, glycolytic capacity and glycolytic reserve in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 4). (F) Kinetic profile of the OCR in LPS-activated macrophages in response to sequential injections of oligomycin, FCCP, and Rot/AA in various groups (Seahorse XF test) (n = 4). (G) Quantification of basal respiration, ATP production, and maximal respiration in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 4). (H) JC-1 aggregation (red fluorescence) in healthy mitochondria and cytosolic JC-1 monomers in compromised mitochondria (green fluorescence) (immunofluorescence assays). (I) Quantitative analysis of MMP levels determined by the relative ratio of red/green fluorescence intensity in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 4). (J) Intracellular ATP levels of LPS-activated macrophages in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 3). (K-M) The macrophages were pretreated with 100 ng/mL LPS for 24 h and then treated with PBS (Control), 10 μM UK-5099, 100 μg/mL LEVs^Tet−PKM2@TA, or 10 μM UK-5099 plus 100 μg/mL LEVs^Tet−PKM2@TA for another 24 h. (K) Schematic illustration revealing mechanism of LEVs^Tet−PKM2@TA promotes macrophage metabolic reprogramming depending on pyruvate influx into the TCA cycle. (L) Kinetic profile of the OCR in LPS-activated macrophages in response to sequential injections of oligomycin, FCCP, and Rot/AA in various groups (Seahorse XF test) (n = 3). (M) Quantification of basal respiration, ATP production, and maximal respiration in the Control, UK-5099, LEVs^Tet−PKM2@TA, and UK-5099 + LEVs^Tet−PKM2@TA groups (n = 3). The data are expressed as the mean ± SEM. Statistical analysis was performed with one-way ANOVA (B, E, G, I, J, and M). ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

Considering that the TCA cycle and OXPHOS activity in cells are closely related to mitochondrial function [61,62], we further investigated whether LEVs^Tet−PKM2@TA could rescue mitochondrial function by assessing mitochondrial membrane potential (MMP) and intracellular ATP levels. Representative fluorescence images revealed that compared with those in the Control group, JC-1 aggregates (indicated by red fluorescence) in healthy mitochondria were more frequently observed in the LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (Fig. 5H). Quantitative analysis of MMPs calculated by the ratio of red to green fluorescence intensity confirmed that the LEVs^Tet−PKM2@TA group presented the highest MMP values among the tested groups, and both the LEVs^Tet−PKM2 and the LEVs^PKM2 groups presented increased MMP levels in LPS-pretreated macrophages compared with Control (Fig. 5I). Consistent with the results of the targeted metabolomic and OCR analyses, the ATP levels significantly increased in the cells of the LEVs^Tet−PKM2@TA group compared with those in the Control and LEVs^PKM2 groups (Fig. 5J). Taken together, these findings revealed that LEVs^Tet−PKM2@TA enhanced the mitochondrial function of LPS-activated macrophages. To test whether Tet-PKM2 drives metabolic reprogramming in macrophages by generating pyruvate and facilitating pyruvate entry into the TCA cycle, we inhibited mitochondrial pyruvate carrier 1 (MPC1) using the specific inhibitor UK-5099. Evaluation of OXPHOS activities revealed that UK-5099 effectively suppressed the increased OCR induced by LEVs^Tet−PKM2@TA in LPS-activated macrophages (Fig. 5L and M). Consistent with OCR analysis, a decrease in ATP levels was also observed in the cells treated with UK-5099 (Fig. S7). Taken together, LEVs^Tet−PKM2@TA effectively reprograms LPS-activated macrophages toward a metabolic state of enhanced glycolysis and OXPHOS, rescues mitochondrial function, and diminishes pro-inflammatory metabolites accumulation, through a mechanism dependent on Tet-PKM2-mediated pyruvate entry into the TCA cycle through MCP1 (Fig. 5K).

3.5. Bioengineered LEVs^Tet−PKM2@TA coax macrophage polarization in response to metabolic reprogramming

Immunometabolic reprogramming is a fundamental mechanism that governs macrophage polarization and effector function [[13], [14], [15]]. Metabolic rewiring from OXPHOS to aerobic glycolysis (Warburg effect) is an important mechanism modulating the M2/M1 balance [12]. Considering that LEVs^Tet−PKM2@TA can restore the TCA cycle and OXPHOS in macrophages, we further investigated whether LEVs^Tet−PKM2@TA could induce the repolarization of M1 macrophages, the plasticity of which is impaired, and whether persistent proinflammatory responses occur [12]. LPS-activated macrophages were incubated with PBS (Control), LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for 24 h. Subsequently, the polarization states of the macrophages were characterized using qRT–PCR, ELISA, and immunofluorescence staining. qRT‒PCR revealed that incubation with LEVs^Tet−PKM2@TA significantly decreased the expression levels of M1 polarization-related genes (IL-6 and IL-1β) but increased the expression levels of M2 polarization-related genes (IL-4 and Arg-1) (Fig. 6A). Similarly, ELISA revealed that LEVs^Tet−PKM2@TA significantly decreased the concentrations of proinflammatory cytokines (IL-6 and TNF-α). In contrast, compared with those in the Control group, the concentrations of anti-inflammatory cytokines (IL-4 and IL-10) in the LEVs^Tet−PKM2@TA group increased significantly. Moreover, LEVs^Tet−PKM2 did not affect IL-4 or IL-10 production, and the concentration of TNF-α significantly differed between the LEVs^Tet−PKM2 group and the LEVs^Tet−PKM2@TA group (Fig. 6B). Consistent with these results, representative immunofluorescence images and quantifications revealed that incubation with LEVs^Tet−PKM2@TA significantly decreased the expression levels of M1-related proteins (iNOS and CCR7) but increased the expression levels of M2-related proteins (Arg-1, CD206, and CD163). Moreover, significant differences in the expression levels of iNOS between the LEVs^Tet−PKM2 and LEVs^Tet−PKM2@TA groups were also observed (Fig. 6C). Furthermore, the representative bands and quantitative analysis showed that the expression levels of Hif-1α in the LEVs^Tet−PKM2@TA group were significantly lower compared with those in the Control group (Fig. S8), which indicated that Tet-PKM2 inhibited LPS-induced Hif-1α expression. To exclude the possibility that TA alone contributes to the observed macrophage reprogramming and therapeutic outcomes, we further analyzed the influence of LEVs@TA on macrophage repolarization. The qRT-PCR analysis indicated that LEVs@TA had similar effects on the expression levels of M1 polarization-related genes (IL-6 and IL-1β) and M2 polarization-related genes (IL-4 and Arg-1) (Fig. S9), indicating that TA alone had little effect on macrophage reprogramming and therapeutic outcomes.

Fig. 6.

Immunomodulatory effects of the bioengineered LEVs^Tet−PKM2@TA in terms of their ability to modulate macrophage polarization in vitro. The macrophages were treated with 100 ng/mL LPS for 24 h and then treated with PBS (Control), 100 μg/mL LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA for another 24 h. (A) The relative mRNA expression levels of M1 polarization-related genes (IL-6 and IL-1β) and M2 polarization-related genes (IL-4 and Arg-1) in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (qRT‒PCR) (n = 3). (B) Concentrations of M1-related cytokines (IL-6 and TNF-α) and M2-related cytokines (IL-4 and IL-10) in the supernatants of the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (ELISA) (n = 3). (C) Representative immunofluorescence images and quantification of the expression levels of M1-related proteins (iNOS and CCR7) and M2-related proteins (CD163, CD206, and Arg-1) in the Control, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 3). The data are expressed as the mean ± SEM. Statistical analysis was performed with one-way ANOVA (A, B, and C). ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

3.6. Administration of LEVs^Tet−PKM2@TA combat periodontitis and promote inflammation resolution in mice

To investigate the therapeutic effect of LEVs^Tet−PKM2@TA on periodontitis, a ligature-induced periodontitis model was established in mice. After the silks were removed on day 9 post-ligature, PBS, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA were injected into the gingival tissue on days 10, 12, and 14 post-ligature, respectively. On day 15 post-ligature, the samples were harvested, and the histocompatibility of LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA was analyzed by H&E staining. No signs of pathological abnormalities in the major organs (heart, liver, spleen, lungs, or kidneys) of the mice were observed, indicating favorable in vivo biocompatibility of LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA (Fig. S10). Meanwhile, the levels of serum biochemical indices [e.g., alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine (CREA), and urea (UREA)] were analyzed by an automatic biochemistry analyzer to evaluate systemic toxicity. The results showed similar levels of ALT, AST, CREA and UREA among all the tested groups, indicating the good biocompatibility of the LEVs^Tet−PKM2@TA (Fig. S11). Furthermore, the regeneration outcomes of hybrid periodontal tissues were analyzed by microcomputed tomography (micro-CT), H&E staining, and Masson staining. (Fig. 7A). Representative micro-CT scanning and 3D-constructed images viewed from the buccal sides of the maxillary alveolar bone revealed severe alveolar bone resorption in defects that received PBS injections, whereas much more newly formed bone was observed in periodontal defects that received either LEVs^Tet−PKM2 or LEVs^Tet−PKM2@TA injection (Fig. 7B). Quantitative analysis of the micro-CT scanning revealed that compared with the other treatments, LEVs^Tet−PKM2@TA injection resulted in the lowest distance from the cemento-enamel junction to the alveolar bone crest (CEJ-ABC) but the highest BV/TV among all the tested groups. Moreover, significant differences in the Tb.N and Tb.Sp were observed only between the PBS group and the LEVs^Tet−PKM2@TA group (Fig. 7C). These data indicated that TA modification enhanced the therapeutic performance of LEVs^Tet−PKM2 and that LEVs^Tet−PKM2@TA significantly promoted periodontal bone regeneration. Representative images of H&E staining revealed much less infiltration of immune cells in the LEVs^Tet−PKM2@TA groups than in the control groups (Fig. 7D). Similarly, representative images of Masson staining revealed that the elastic and collagen fibers were denser and arranged in a more orderly manner in the LEVs^Tet−PKM2@TA groups. These results strongly demonstrated that the LEVs^Tet−PKM2@TA promoted inflammation resolution in mice and the repair of periodontal tissues.

Fig. 7.

The therapeutic effects of LEVs^Tet−PKM2@TA on ligature-induced periodontitis in mice. (A) Schematic illustration showing the experimental timeline of silk ligation and injections of EVs. 10 μL of PBS or 100 μg of LEVs^PKM2, LEVs^Tet−PKM2, or LEVs^Tet−PKM2@TA dissolved in 10 μL of PBS was injected into the gingival tissue using a microsyringe on days 10, 12, and 14 post-ligature. (B) View of the buccal sides of the maxillary alveolar bone (micro-CT scanning). The yellow and red dashed lines in the 3D-constructed images indicate the boundaries of the CEJ and ABC, respectively. The linear distance from the CEJ to the ABC is indicated by the red lines in the coronal scanning of the micro-CT images. (C) Statistical analysis of the average linear distance of CEJ-ABC, BV/TV, Tb.N, and Tp.Sp in the Control, PBS, LEVs^PKM2, LEVs^Tet−PKM2, and LEVs^Tet−PKM2@TA groups (n = 6). (D and E) Representative H&E (D) and Masson staining (E) images showing the infiltration of immune cells and newly formed periodontal tissues. The data are expressed as the mean ± SEM. Statistical analysis was performed with one-way ANOVA (C). ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

3.7. LEVs^Tet−PKM2@TA modulate the in vivo macrophage phenotype and function to facilitate tissue repair

Considering that LEVs^Tet−PKM2@TA facilitate the repolarization of M1 macrophages toward the M2 phenotype in vitro (Fig. 6), the immunomodulatory effects of LEVs^Tet−PKM2@TA on macrophages were further investigated in vivo. First, the in vivo phagocytosis of LEVs by CD68⁺ macrophages was confirmed by immunofluorescence staining (Fig. S12). Representative immunofluorescence images revealed that M1 polarization-related proteins (iNOS and CCR7) in CD68⁺ macrophages were highly expressed in gingival tissues from mice receiving PBS or LEVs^PKM2 injections (Fig. 8A). In contrast, more M2 polarization-related proteins (CD163 and CD206) were observed in CD68⁺ macrophages in the LEVs^Tet−PKM2 and LEVs^Tet−PKM2@TA groups than in the PBS and LEVs^PKM2 groups (Fig. 8B). Quantitative analysis of immunofluorescence staining revealed that the ratios of iNOS⁺/CD68⁺ and CCR7⁺/CD68⁺ cells in the PBS group were significantly greater than those in the LEVs^Tet−PKM2@TA group, while the ratios of CD163⁺/CD68⁺ or CD206⁺/CD68⁺ cells in the LEVs^Tet−PKM2@TA group were significantly greater than those in the PBS and LEVs^PKM2 groups (Fig. 8C). There were no significant differences in the ratio of CCR7⁺/CD68⁺ or CD206⁺/CD68⁺ cells between the PBS and LEVs^Tet−PKM2 groups. These data indicated that among all the tested groups, LEVs^Tet−PKM2@TA exerted the strongest immunomodulatory effects on macrophages. Considering that the therapeutic effects of LEVs^Tet−PKM2@TA are derived mainly from the enzymatic activity of Tet-PKM2, we further investigated PK activity in the gingival tissues of mice in all the investigated groups. As expected, compared with that in the PBS group, PK activity significantly increased in the LEVs^Tet−PKM2 and LEVs^Tet−PKM2@TA groups (Fig. S15), indicating that the delivered Tet-PKM2 was enzymatically activated. It has been reported that dimeric PKM2 can translocate to the nucleus of LPS-activated macrophages, where it interacts with HIF-1α and subsequently promotes inflammatory responses or enhances the translation of proglycolytic enzymes [[26], [27], [28]]. The representative immunofluorescence images showed that the expression of Hif-1α and PKM2 in the nuclear of macrophages decreased significantly in the LEVs^Tet−PKM2@TA group (Figs. S13 and 14), which indirectly demonstrated that LEVs^Tet−PKM2@TA could attenuate glycolysis and inflammatory responses in vivo. Taken together, these results confirmed that LEVs^Tet−PKM2@TA had robust immunomodulatory effects across healing cascades and could be used to coordinate the polarization of macrophages toward the M2 phenotype in vivo.

Fig. 8.

Immunomodulatory function of the LEVs^Tet−PKM2@TA in terms of their ability to modulate macrophage polarization in vivo. (A-B) The expression of M1 polarization-related markers (iNOS/CD68 and CCR7/CD68) (A) and M2 polarization-related markers (CD163/CD68 and CD206/CD68) (B) in gingival tissues on day 15 post-ligature (immunofluorescence assay). (C) Quantitative analysis of the ratios of M1-polarized macrophages (iNOS/CD68-and CCR7/CD68-positive cells) and M2-polarized macrophages (CD163/CD68-and CD206/CD68-positive cells) among total macrophage populations (CD68-positive cells) based on immunofluorescence staining (n = 3). The data are expressed as the mean ± SEM. Statistical analysis was performed with one-way ANOVA (C). ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between the indicated columns.

4. Discussion

During the development of chronic inflammatory diseases, metabolic reprogramming from OXPHOS toward glycolysis triggers uncontrolled overactivation of macrophages, which is fundamental for sustaining the chronicity of inflammatory pathologies such as periodontitis [17,63,64]. Modulating metabolic reprogramming represents a promising novel therapeutic strategy for resolving pathological inflammation. However, current immunomodulatory strategies focus primarily on modulating the production of proinflammatory cytokines and the activation of transcription factors responsible for inducing the M1 phenotype[65], which overlooks the role of immunometabolism in orchestrating macrophage polarization. Our study revealed aberrant immunometabolism and significantly decreased expression of Tet-PKM2 in macrophages during the development of periodontitis, suggesting the pivotal role of Tet-PKM2 in the metabolic reprogramming of macrophages. Therefore, we engineered the LEVs^Tet−PKM2 with lysosomal escape ability to ensure the cytoplasmic transport of Tet-PKM2, thereby modulating the metabolic reprogramming of macrophages. As expected, LEVs^Tet−PKM2@TA successfully escaped endo/lysosome-mediated degradation, increased the expression level of Tet-PKM2, and elicited a notable immunomodulatory effect by modulating the metabolic reprogramming of macrophages. Ultimately, treatment with LEVs^Tet−PKM2@TA significantly promoted inflammation resolution in periodontitis models. Our findings demonstrate that the delivery of Tet-PKM2 by LEVs^Tet−PKM2@TA represents a potent and effective strategy for immunometabolic reprogramming of macrophages, thereby alleviating inflammation in chronic inflammatory diseases.

Abnormal metabolic states often perturb normal physiology and even lead to severe tissue dysfunction. As they are metabolically sensitive cells, the metabolic states of immune cells are closely related to their effector functions [14,66,67]. In the present study, our results revealed increased glycolysis and an interrupted TCA cycle, and pyruvate was predominantly converted to lactate instead of being metabolized to acetyl-CoA (Fig. 1), indicating reduced PKM2 enzymatic activity (Fig. 1). PKM2 is a well-established rate-limiting enzyme that is recognized as a key regulator of metabolic reprogramming associated with immune responses [24]. In recent years, emerging evidence has indicated the pivotal role of PKM2 in driving the development of periodontitis, highlighting its specific functions in modulating stem cell and osteoclastic activity [23,68,69]. However, the specific role of PKM2 in immune cells, particularly macrophages, during periodontitis pathogenesis remains poorly defined. The functional versatility of PKM2 is governed by its conformational shift between monomeric/dimeric and tetrameric forms [70]. In the present study, we found significant downregulation of the expression of enzymatically active Tet-PKM2, which catalyzes the conversion of PEP to pyruvate and the subsequent entry of pyruvate into the TCA cycle, in gingival tissues from periodontitis patients and periodontitis models (Fig. 1 and Figs. S1-2). Therefore, pyruvate is preferentially converted to LA rather than entering the TCA cycle under these conditions [71,72]. Our data indicate that PKM2 is a pivotal metabolic regulator of macrophage metabolism during the development of periodontitis, suggesting that the restoration of Tet-PKM2 expression might represent a viable therapeutic strategy for modulating macrophage effector functions in chronic inflammation.

Current strategies for modulating immunometabolism include the use of small molecules and nanoparticles, which have demonstrated potential in rescuing impaired mitochondria, altering cellular metabolism, and eliciting anti-inflammatory effects [73]. However, direct administration of small molecules and nanoparticles tends to lead to off-target effects and low bioavailability attributable to lysosomal degradation [74,75]. TEPP-46 has been confirmed to play multiple roles in different immune cells. For example, the activation of PKM2 with TEPP-46 in neutrophils increases ROS production and enhances microbial killing [76]. PKM2 activation has been shown to prevent CD4 T cell activation, proliferation, and cytokine production, whereas it promotes CD8 T cell activation and enhances effector functions [30,77]. Hence, direct administration of TEPP-46 in vivo may have additional cellular targets and trigger complex reactions. In contrast, bioactive agents such as Tet-PKM2 delivered via EVs have multiple advantages in immunomodulating macrophages, including low immunogenicity, favorable safety profiles, and inherent targeting specificity toward macrophages [34,36,78]. Nevertheless, the application of EVs in targeting the energy metabolism of macrophages, which play a key role in inflammatory diseases, is limited. To address this issue, we successfully generated Tet-PKM2-enriched EVs to target the immunometabolism of macrophages. HEK293T cells were used to generate genetically engineered EVs because of their high transfection efficiency and robust EV yield [79,80]. In the present study, HEK293T cells were first transfected with lentivirus to increase the expression level of PKM2 and then stimulated with TEPP-46 to activate Tet-PKM2 (Fig. 2A). Interestingly, our findings revealed that compared with SEVs, Tet-PKM2 was more enriched in LEVs (Fig. 2). Our results aligned with the findings of Ye et al., who reported that compared with SEVs, LEVs exerted enhanced therapeutic efficacy both in vitro and in vivo [37]. This phenomenon may be attributable to the greater internal volume of LEVs, which allows for the incorporation of a greater quantity of substances from donor cells, potentially leading to enhanced biological responses [81]. Differences in contents, including protein profiles, between LEVs and SEVs have been reported in mounting studies using proteomic analysis [[82], [83], [84]]. In the future, we plan to utilize proteomic analysis in subsequent studies to systematically elucidate the distinct cargo profiles of these EV subtypes, which will provide a more in-depth understanding of their specific biogenesis and functional roles. Similar to traditional SEVs, LEVs^Tet−PKM2 could be successfully internalized by macrophages (Fig. 2). LEVs^Tet−PKM2@TA treatment not only restored the impaired mitochondrial function of LPS-stimulated macrophages, including ATP levels and the MMP but also reprogrammed cellular energy metabolism (Fig. 5). The targeted energy metabolomics analysis revealed that LEVs^Tet−PKM2@TA rescued LPS-induced metabolic dysfunction, which significantly decreased the accumulation of glycolysis intermediates involved in proinflammatory responses [60,85]. In contrast, LEVs^Tet−PKM2@TA significantly increased the production of terminal metabolites from the TCA cycle and OXPHOS, such as GTP and ATP, indicating that higher amounts of pyruvate enter the TCA cycle. Consistent with the results of the metabolomics analysis, those of the extracellular flux analysis verified the increased OXPHOS levels, as indicated by the OCR results. In addition, an increased ECAR results was observed in LPS- induced macrophages after LEVs^Tet−PKM2@TA treatment. The increase in glycolysis may be due to the high enzymatic activity of Tet-PKM2, which can promote the production of pyruvate and subsequent entry of pyruvate to the TCA cycle. Some studies have also indicated that glycolysis can be also crucial for M2 macrophages, highlighting the complexity of this regulation [16,86]. Therefore, further investigations based on metabolic flux analysis may be required to confirm metabolic reprogramming by increased glycolytic flux into the TCA cycle. When mitochondrial pyruvate uptake was blocked by MPC1, the LEVs^Tet−PKM2@TA failed to promote metabolic programming in macrophages (Fig. 5K−M), indicating that LEVs^Tet−PKM2@TA promotes macrophage metabolic reprogramming through a mechanism dependent on pyruvate influx into the TCA cycle. Overall, these findings indicated that the delivery of Tet-PKM2 successfully promoted metabolic reprogramming in LPS-activated macrophages, which is in line with the findings of previous studies [26,30].

Previously published studies have indicated that metabolic reprogramming toward OXPHOS guides the repolarization of proinflammatory macrophages into anti-inflammatory cells [12]. As expected, our data revealed that LEVs^Tet−PKM2@TA successfully repolarized LPS-treated macrophages toward the M2 phenotype (Fig. 6). Our previously published studies have indicated that macrophage polarization plays key roles in the repair cascades of periodontal tissues [87,88]. Therefore, we further explored the influence of the LEVs^Tet−PKM2@TA on periodontal regeneration. In ligature-induced periodontitis models in mice, our data indicated that LEVs^Tet−PKM2@TA successfully induced M2 macrophage polarization, significantly relieved tissue inflammation and promoted periodontal tissue regeneration (Fig. 7, Fig. 8). Given its efficacy as a versatile and targeted delivery vehicle for Tet-PKM2 to macrophages, the LEVs^Tet−PKM2@TA presents a potential universal strategy for immunometabolic reprogramming of macrophages. Hence, LEVs^Tet−PKM2@TA may also have therapeutic effects on other chronic inflammatory diseases, such as arthritis and diabetes. Despite the promising potential for the clinical translation of LEVs^Tet−PKM2@TA in the treatment of periodontitis, several major challenges remain to be addressed. Current laboratory-scale production methods involve low-yields, are time-consuming, and lack the standardization required for Good Manufacturing Practice. Furthermore, ensuring the long-term stability of these protein-rich, membrane-bound vesicles presents a critical hurdle. Our stability assessment revealed that the biological activity of LEVs^Tet−PKM2@TA is preserved for at least one week at −80 °C but undergoes a notably decreased after one month (Fig. S4). Scalable production processes must also guarantee rigorous batch-to-batch consistency in key parameters, including vesicle size, zeta potential, surface protein composition, and drug loading efficiency.

In the present study, we provide a simplified and biosafe engineering strategy to facilitate the lysosomal escape of EVs, thereby ensuring efficient, targeted delivery of Tet-PKM2 to macrophages (Fig. 3, Fig. 4). Various materials, including polyethylenimine (PEI) and cell-penetrating peptides (CPPs), have been developed to promote the endosomal escape capability of nanoparticles[89]. High concentrations of PEI can compromise cell viability because of its high amine group density, which causes excessive endo/lysosomal charge disruption [90,91]. Therefore, these polycationic materials need to be appropriately engineered to balance their escape efficiency and inherent toxicity. Similarly, many CPPs necessitate the use of auxiliary charged polymers, leading to increased potential toxicity [92]. Furthermore, functionalizing EVs with such agents often involves sophisticated and time-consuming reaction steps [93,94]. In contrast, TA, a natural polyphenol, can directly bind to the EV membrane and promote endo/lysosomal escape owing to its pH-responsive behavior in acidic environments [41,43]. Previous TA-based modified metal‒phenolic networks (MPNs), such as TA–Fe³⁺ films have been utilized to confer lysosomal escape capability to nanoparticles. However, these approaches typically involve multistep reactions and the addition of ancillary chemical reagents[95]. Our study presents a more direct method by incubating the TA with the LEVs. The interaction of the EV phospholipid bilayer and TA via hydrogen bonding facilitates endo/lysosomal escape of LEVs@TA owing to the pH-responsive behavior of TA in acidic environments. Flow cytometry and immunofluorescence images suggested successful TA modification of the LEV membrane (Fig. 3). Enhanced endo/lysosomal escape ability was further confirmed through colocalization analysis of LEVs@TA with endo/lysosomal markers coupled with calcein leakage assays (Fig. 4). Moreover, we applied CGMD simulation to mimic the uptake of EVs by macrophages and observed the disassembly of TA from the surface of EVs in an acidic environment due to the “proton-sponge effect” of TA. This computational approach provides a novel methodology for investigating the dynamic interactions of EVs and cells, and the dynamic subcellular processes, such as lysosomal escape of EVs, in the future. The underlying mechanism of TA or metal‒phenolic coating-based modifications for lysosomal escape has been extensively studied [[41], [42], [43],46], which has indicated that acidic environments in endo/lysosomes (pH 4.5–7.0) trigger protonation of TA, which then induces lysosomal swelling through the proton sponge effect [43,46]. In summary, this work presents a direct and efficient membrane modification approach to facilitate the escape of EVs from endo/lysosomal compartments, increasing the translational potential of this lysosomal-escape approach for LEV^sTet−PKM2@TA.

There were several limitations of this study that remain to be addressed in the future. Owing to technical constraints, the CGMD simulations revealed the disassembly of TA from LEVs before the endo/lysosomal escape of LEVs@TA. However, the simulation could not fully replicate the ionic osmotic pressure gradients or electrical potential changes across lysosomal membranes. To address this limitation, future studies should investigate lysosomal ion gradients using live-cell imaging with LysoTracker and the chloride-specific dye MAQE-MP [96], thereby providing direct experimental validation of the escape effect. In addition, no model that is suitable for studying all aspects of human periodontal disease is available thus far. In our study, we performed ligature-induced periodontitis models in mice, which could simulate most aspects of periodontal disease, including bacterial interactions and dysbiosis, the periodontal inflammatory response and bone biology [55]. Future studies employing large animal models, such as dogs or pigs, as well as extended disease induction periods, would be invaluable for more accurately mimicking the chronic nature of human periodontitis and enhancing clinical translatability.

5. Conclusion

In summary, we successfully engineered Tet-PKM2-enriched LEVs that can escape lysosomal degradation and modulate the metabolic reprogramming of macrophages. During the pathogenesis of periodontitis, the expression of Tet-PKM2 in macrophages is significantly downregulated, leading to aberrant metabolism in the gingiva. To address this dysregulation, Tet-PKM2-enriched extracellular vesicles modified with TA were successfully generated. LEVs^Tet−PKM2@TA could escape endo/lysosome-mediated degradation and significantly rescue the PK activity of macrophages. In vitro studies further revealed that LEVs^Tet−PKM2@TA promoted a metabolic shift from glycolysis toward the TCA cycle and mitochondrial OXPHOS and improved the mitochondrial function of LPS-pretreated macrophages. The immunomodulatory and repair-promoting functions of LEVs^Tet−PKM2@TA were subsequently tested in ligature-induced periodontitis models in mice. LEVs^Tet−PKM2@TA treatment significantly promoted the repair of periodontal tissues and exerted robust immunomodulatory effects, facilitating M2 macrophage polarization throughout the healing process. Our study highlights the potential of Tet-PKM2 as a metabolic regulator to modulate the immunometabolism and polarization of macrophages, which may provide promising nanotherapeutic approaches for the treatment of chronic inflammatory diseases.

CRediT authorship contribution statement

Wen-Jie Zhang: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation. Bei-Min Tian: Validation, Supervision, Funding acquisition, Formal analysis. Fang Li: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Xuan Li: Methodology, Investigation, Formal analysis, Data curation. Rui-Xin Wu: Software, Methodology, Investigation. Yuan Yin: Visualization, Formal analysis. Jia Wang: Software, Methodology, Investigation. Dao-Kun Deng: Methodology, Investigation. Yu-Zhe Chen: Methodology, Investigation. Hong-Yu Wang: Software, Methodology, Investigation. Yu-Xuan Du: Visualization, Formal analysis, Data curation. Xuan Wang: Visualization, Formal analysis. Yin Xiao: Validation, Supervision, Formal analysis, Data curation. Xiao-Tao He: Writing – review & editing, Visualization, Validation, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization. Fa-Ming Chen: Writing – review & editing, Visualization, Supervision, Funding acquisition.

Ethics approval and consent to participate

All animal procedures were approved by the Air Force Medical University Ethics Committee (AFMU-20240018). Human sample collection was approved by the Ethics Committee of the School of Stomatology, Fourth Military Medical University (KQ-YJ-2025-226). Informed corwas obtained from all participating subjects prior to the initiation of the study, and allprocedures involving human participants adhered to the Declaration of Helsinki and relevant ethical standards.

Declaration of competing interest

The authors declare that they have no competing interests.

Appendix A. Supplementary data

The following are the Supplementary data to this article: