AIE-pharmacology: Acacetin and NMN co-delivery rescues sperm motility

Abstract

The emergence of aggregation-induced emission (AIE) luminogens has revolutionized biomedical theranostics, yet the exploration of natural AIE-active therapeutics remains nascent. Herein, we propose a novel “AIE-Pharmacology” paradigm, employing the natural flavonoid acacetin as an intrinsic AIEgen with pharmacological activity for self-monitored therapy of asthenozoospermia. Overcoming traditional aggregation-caused quenching (ACQ), acacetin utilizes its aggregation-enhanced emission and excited-state intramolecular proton transfer (ESIPT) properties to enable simultaneous treatment, high-contrast imaging, and tracking. We developed an extracellular vesicles (EVs)-based co-delivery system loaded with acacetin and β-Nicotinamide mononucleotide (NMN). This nanoplatform (EVs@N + A) exhibits inherent affinity for the testicular microenvironment. Upon targeted accumulation, released acacetin triggers a potent anti-ferroptotic response by upregulating GPX4 and SLC7A11 and suppressing lipid peroxidation. In parallel, NMN elevates intracellular NAD⁺ levels, reprogramming cellular metabolism and restoring ATP production. This dual strategy—inhibiting ferroptosis and rejuvenating metabolism—effectively rescues sperm function, enhancing sperm count, motility, and progressive movement in vivo. Our work reports acacetin as a multifunctional AIEgen in reproductive medicine and establishes a traceable theranostic platform, introducing the “AIE-Pharmacology” paradigm as a promising approach for treating male infertility and other diseases.

Graphical abstract

Highlights

• •Acacetin, a novel AIEgen with ESIPT, enables self-monitored therapy for asthenozoospermia under the AIE-Pharmacology paradigm.
• •Engineered EVs enable BTB penetration and co-delivery of Acacetin+NMN, ensuring targeted testicular accumulation and therapy.
• •Dual therapy rescues sperm function by blocking ferroptosis (Nrf2/GPX4) and boosting ATP metabolism simultaneously.

1. Introduction

Asthenozoospermia is a condition characterized by reduced sperm motility and is one of the most common causes of male infertility, ranking as a leading contributor to infertility issues among men globally [[1], [2], [3]]. The pathophysiological landscape of this condition is increasingly recognized to be multifactorial, with recent investigations underscoring the pivotal roles of ferroptosis and metabolic dysregulation [[4], [5], [6]]. Ferroptosis, an iron-dependent form of regulated cell death driven by the lethal accumulation of lipid peroxides, has been implicated in the impairment of spermatogenic cells and the deterioration of sperm quality [[7], [8], [9], [10], [11]]. Concurrently, a compromised energy metabolism, marked by inadequate adenosine triphosphate (ATP) production, fails to fuel the high-energy demands of sperm motility, creating a vicious cycle of functional decline [[12], [13], [14], [15]]. While antioxidant therapies offer some benefit, their efficacy is often limited by a lack of specificity and an inability to simultaneously address the interconnected pathways of oxidative lipid damage and bioenergetic deficit [16]. Therefore, the development of a coordinated therapeutic strategy that concurrently suppresses ferroptosis and reprograms cellular metabolism presents a compelling yet largely unexplored avenue for the effective treatment of asthenozoospermia.

The advent of aggregation-induced emission luminogens (AIEgens) has opened new frontiers in biomedical theranostics [[17], [18], [19], [20], [21], [22], [23], [24], [25], [26]]. AIEgens are ideal for high-contrast bioimaging and tracking due to their enhanced emission upon aggregation, whereas traditional ACQ fluorophores see their fluorescence quenched in the aggregated state [[27], [28], [29], [30], [31], [32], [33], [34], [35], [36]]. While synthetic AIEgens have been widely explored, the reservoir of natural products with intrinsic AIE characteristics remains a nascent and promising frontier [37]. The development of such “AIE-Pharmacology” paradigms, where a single natural entity serves as both a therapeutic agent and a self-reporting imaging probe, could revolutionize targeted therapy by enabling real-time visualization of drug delivery and pharmacological activity [38]. Acacetin, a naturally occurring flavonoid found in various plants, has documented broad pharmacological activities, including potent antioxidant, anti-cancer and anti-inflammatory effects [39,40]. Its reported efficacy in mitigating oxidative stress-related pathologies suggests its potential as a ferroptosis inhibitor [41]. However, its intrinsic AIE property and its application in reproductive medicine, particularly for combating ferroptosis in spermatogenic cells, have never been investigated. We thus hypothesize that acacetin could serve as a natural AIEgen drug for theranostics in male infertility.

Despite its promise, the effective delivery of acacetin, along with essential metabolic modulators, to the testicular microenvironment remains a challenge [[42], [43], [44], [45], [46], [47], [48]]. Extracellular vesicles (EVs) are endogenous nanovesicles that have emerged as superior drug delivery vehicles, owing to their exceptional biocompatibility, low immunogenicity, and innate ability to cross biological barriers [49]. Their natural composition makes them an ideal platform for the coordinated delivery of multiple therapeutic cargoes [50]. β-Nicotinamide mononucleotide (NMN) is a key precursor to nicotinamide adenine dinucleotide (NAD⁺), an essential coenzyme serving fundamental roles in cellular metabolism [51]. By supplementing NMN to boost intracellular NAD⁺ levels, we can directly enhance the efficiency of ATP-producing pathways, such as glycolysis and oxidative phosphorylation [52]. This represents a logical strategy to reprogram the metabolic state of energy-starved spermatogenic cells, thereby directly addressing the bioenergetic aspect of asthenozoospermia.

Herein, we propose a novel “AIE-Pharmacology” strategy for the treatment of asthenozoospermia via an engineered extracellular vesicle-based platform for the co-delivery of the natural AIE drug acacetin and NMN (Scheme 1). This multifunctional nanoplatform (denoted as EVs@N + A) is designed to achieve synergistic therapeutic efficacy through a dual-pathway mechanism: i) the AIE-active acacetin inhibits ferroptosis by upregulating key defense proteins (GPX4, SLC7A11) and quenching lipid peroxidation, while simultaneously allowing for real-time, self-guided tracking of its accumulation and distribution; and ii) the co-delivered NMN replenishes the essential metabolic cofactor NAD⁺, thereby restoring cellular metabolism to rejuvenate the intracellular energy pool and restore ATP production. We demonstrate that this coordinated approach of ferroptosis suppression and metabolic rejuvenation leads to a remarkable recovery of sperm function in an in vivo model of asthenozoospermia. This work not only reports acacetin as a multifunctional natural AIEgen for reproductive medicine but also establishes a robust and traceable theranostic platform, heralding the “AIE-Pharmacology” paradigm as a promising strategy for treating male infertility.

Scheme 1.

This schematic illustrates an exosomal co-delivery platform (EVs@N + A) that utilizes the natural flavonoid acacetin, a newly discovered AIEgen, along with NMN to treat asthenozoospermia. Acacetin's intrinsic ESIPT-based AIE property enables self-tracked testicular accumulation and ferroptosis inhibition via Nrf2-mediated upregulation of SLC7A11/GPX4, while NMN replenishes NAD⁺ to restore ATP. This dual-action strategy synergistically improves sperm motility and function through combined ferroptosis inhibition and bioenergetic amelioration.

2. Experimental section

2.1. Materials

Acacetin, reduced glutathione (GSH), and dimethyl sulfoxide (DMSO) were purchased from Aladdin Biochemical Technology Co., Ltd. (China). All assay kits, including the BeyoExo™ Exosome Identification Kit (containing CD63 and HSP70 antibodies), as well as antibodies for the negative marker Calnexin, Reactive Oxygen Species Assay Kit, Lipid Peroxidation MDA Assay Kit, and NAD⁺/NADH Assay Kit, were obtained from Beyotime Institute of Biotechnology (Jiangsu, China). Cell culture media and supplements—DMEM, FBS, and penicillin-streptomycin solution—were acquired from Gibco (USA). Human Umbilical Cord Mesenchymal Stem Cells were sourced from Cyagen Biosciences (China). Milli-Q purified water (Millipore, USA) was used for the preparation of all aqueous solutions.

2.2. Preparation of UCMSC-derived EVs

A sequential ultracentrifugation protocol was employed to isolate EVs from UCMSC-conditioned media. Following a 48-h culture, the conditioned medium was collected and subjected to differential centrifugation: first at 300 × g for 10 min to pellet floating cells and apoptotic bodies, followed by 10,000 × g for 30 min to remove larger debris and microvesicles. The final supernatant was ultracentrifuged at 100,000 × g for 90 min to pellet the EVs. To ensure high purity, the EV pellet was washed in phosphate-buffered saline (PBS) and re-pelleted via a second ultracentrifugation step under the same conditions. The final purified EVs were resuspended in PBS for immediate use or stored at −80 °C.

2.3. Characterization of isolated UCMSC-EVs

A comprehensive characterization of the isolated UCMSC-EVs was performed using a multi-modal approach. The morphology and structural integrity of the EVs were examined by Transmission Electron Microscopy using an FEI Tecnai G2 S-Twin microscope, following negative staining with 2% uranyl acetate. The hydrodynamic diameter distribution and zeta potential of the EVs in suspension were determined by Dynamic Light Scattering (DLS) using a Malvern Zetasizer Nano ZS90 system. The particle concentration and size distribution profile were further quantified using Nanoparticle Tracking Analysis (NTA) on a NanoSight NS300 instrument (Malvern Panalytical). The presence of specific exosomal surface markers (Calnexin, CD63, and HSP70) was confirmed using the BeyoExo™ Exosome Identification Kit via western blot analysis. The total protein content of the EV preparations was quantified using a bicinchoninic acid protein assay kit (Beyotime) to standardize loading for functional experiments. Furthermore, the intrinsic optical properties of the EVs and any encapsulated cargo were assessed by UV-Vis absorption spectroscopy (Hitachi U-3100) and photoluminescence (PL) spectroscopy (Edinburgh Instruments F1000). The absolute photoluminescence quantum yield (PLQY) of acacetin powder was determined using a calibrated integrating sphere on an Edinburgh Instruments spectrometer.

2.4. Cell culture

Mouse spermatogonial-derived GC-1 spg cells, as well as mouse testicular somatic cell lines TM3 (Leydig cells) and TM4 (Sertoli cells), were cultured in high-glucose DMEM, supplemented with 10% (v/v) heat-inactivated FBS and 1% (v/v) penicillin-streptomycin. Human Umbilical Cord Mesenchymal Stem Cells (UCMSCs) were cultured and expanded under standard conditions using a proprietary, serum-free Mesenchymal Stem Cell Growth Medium (Cyagen Biosciences), as per the supplier's instructions.

2.5. Preparation of EVs@N + A

The co-delivery system EVs@N + A, loaded with β-Nicotinamide Mononucleotide (NMN) and Acacetin, was prepared using an ultrasound-assisted method. Briefly, UCMSC-EVs (6 × 10⁹ particles mL⁻¹) were mixed with NMN (0–3 mg mL⁻¹) and Acacetin (0–1 mg mL⁻¹) in PBS (pH 7.4). The mixture was probe-sonicated on ice (100 W, 2 min, 3s/5s pulse) to facilitate drug loading, then incubated at 4 °C for 0.5 h with gentle agitation to allow for equilibrium and stable incorporation of the drugs. Unencapsulated drugs were removed using a 100-kDa Amicon® Ultra centrifugal filter (4000 × g, 20 min, 4 °C). The product was washed twice with PBS and resuspended for use or storage. Drug loading was quantified by UV-Vis analysis of the filtrate, using standard curves. Final loading capacities (LC%) were ∼10% for Acacetin and ∼30% for NMN. The encapsulation efficiencies of acacetin and NMN were determined by quantifying unencapsulated drugs in the filtrate using high-performance liquid chromatography (HPLC). The encapsulation efficiencies (EE%) of acacetin and NMN were 36.7% ± 4.6% and 61.7% ± 2.3%, respectively. For in vitro release studies, EVs@N + A were incubated in PBS at pH 5.0 and 7.4 at 37 °C, and the cumulative release of both drugs was quantified at predetermined time points by HPLC and UV-Vis absorption spectroscopy.

2.6. Intracellular assays in GC-1 spg cells

To evaluate the therapeutic effects at the cellular level, mouse spermatogonial GC-1 spg cells were subjected to various treatments for 24 h. The experimental groups were as follows: PBS (control), 1.0 mM Busulfan (BU) to induce a cellular model of asthenozoospermia, 6 × 10⁹ particles of EVs, 10 μg mL⁻¹ Acacetin, 30 μg mL⁻¹ NMN, and a combination treatment of 6 × 10⁹ EVs containing 10 μg mL⁻¹ Acacetin and 30 μg mL⁻¹ NMN (EVs@N + A).

Following treatment, intracellular redox homeostasis was assessed by measuring the levels of GSH and GSSG using a commercial assay kit. To probe lipid peroxidation, a hallmark of ferroptosis, cells were treated for 12 h, then incubated with 10 μM BODIPY™ 581/591 C11 probe for 1 h at 37 °C. The fluorescence shift, indicative of lipid peroxidation, was visualized and quantified by confocal laser scanning microscopy.

For the assessment of key ferroptosis markers, the activity of GPX4 was measured using a specific peroxidase assay kit. Concurrently, after a 12-h treatment period, intracellular ferrous iron (Fe²⁺) levels were determined using a colorimetric iron assay kit. To identify the predominant cell death mode, busulfan-treated GC-1 spg cells were co-incubated with the apoptosis inhibitor Z-VAD-FMK, the necroptosis inhibitor Necrostatin-1, or the ferroptosis inhibitor Ferrostatin-1, and cell viability was measured by CCK-8 assay.

2.7. Trans-well assay for blood-testis barrier (BTB) permeability

A trans-well model was established to simulate the Blood-Testis Barrier (BTB). Mouse Sertoli cells (TM4 cell line) were seeded at a high density (1.0 × 10⁶ cells/cm²) on the upper chamber of polyester membrane transwell inserts. We monitored barrier integrity by measuring Transepithelial Electrical Resistance (TEER) daily with an epithelial voltohmmeter. Only inserts with stable TEER values exceeding 50 Ω cm² were used for experiments. To evaluate the transcytosis capability of the engineered exosomes, the EVs@N + A formulation was introduced into the upper chamber. After 12 h of incubation, the medium in the bottom chamber was carefully collected. The amount of EVs@N + A that had successfully translocated across the BTB was quantified by measuring the fluorescence intensity of the exosomal label using a microplate reader. To investigate the transcytosis mechanism, Sertoli cells in the transwell model were pre-treated with the clathrin inhibitor chlorpromazine, the macropinocytosis inhibitor EIPA, or the lysosome inhibitor bafilomycin A1 prior to EVs@N + A application, and EV translocation efficiency was quantified.

2.8. In vivo animal studies

All animal experiments were conducted under a protocol approved by the Zhejiang University Animal Care and Use Committee (Approval No. ZJU20251044) and in strict accordance with the institution's ethical guidelines for animal research.

Three-week-old male ICR mice (n = 5 per group) were obtained from Vital River Laboratory and randomly assigned to six experimental groups: 1) Control group: received PBS only; 2) BU model group: received a single intraperitoneal (i.p.) injection of busulfan (20 mg kg⁻¹) to induce asthenozoospermia; 3) BU + EVs group: BU-induced mice treated with 9 × 10¹⁰ particles of blank EVs via tail vein injection; 4) BU + NMN group: BU-induced mice treated with 22.5 mg kg⁻¹ NMN via tail vein injection; 5) BU + Acacetin group: BU-induced mice treated with 7.5 mg kg⁻¹ Acacetin via tail vein injection; 6) BU + EVs@N + A group: BU-induced mice treated with the co-delivery system (9 × 10¹⁰ particles containing 22.5 mg kg⁻¹ NMN and 7.5 mg kg⁻¹ Acacetin) via tail vein injection. All treatments were administered every two days for a total duration of one week, commencing three weeks post busulfan injection to allow for model establishment. At the endpoint, body weights, as well as testis and epididymis weights, were recorded. Spermatozoa were collected from the cauda epididymis by mincing the tissue in pre-warmed DMEM/F12 medium. Sperm kinematics were immediately analyzed using a computer-assisted sperm analysis system. For histological examination, testis and epididymis samples were stained with Periodic Acid-Schiff or Hematoxylin and Eosin for morphological and spermatogenic staging evaluation. Furthermore, immunofluorescence (IF) staining was performed on testicular sections to detect the expression and localization of key proteins involved in ferroptosis (e.g., GPX4, HO-1, Nrf2, SLC7A11, ACSL4, FTH1) and spermatogenesis (e.g., PGK2, DDX4, ACRV1, SYCP3, GFRA1). Fluorescence microscopy coupled with ImageJ software was utilized for image acquisition and semi-quantitative evaluation. To evaluate reproductive safety, treated male mice from each group were mated with healthy adult females after the treatment regimen. Offspring numbers were recorded at birth, and postnatal body weight of pups was monitored during 21 days.

2.9. Statistical analysis

All values represent the mean ± standard error of the mean (SEM). The experiments were conducted with a minimum of three independent replicates (n ≥ 3). Statistical analyses were conducted using a two-tailed Student's t-test or one-way ANOVA with Tukey's honestly significant difference (HSD) post hoc test, with significance denoted as ns (not significant), ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001.

3. Results and discussion

3.1. Acacetin: a natural flavonoid with intrinsic ESIPT-AIE characteristics

The exploration of natural products exhibiting intrinsic AIE characteristics represents an emerging frontier in theranostics. This study focuses on acacetin (5,7-dihydroxy-4′-methoxyflavone), a natural O-methylated flavone found in traditional medicinal plants including chrysanthemum (Dendranthema morifolium) and black locust [53]. Recognized in ethnopharmacology for its broad-spectrum bioactivities—including antioxidant, anti-inflammatory, and anticancer effects—acacetin's pharmacological potential stems from its classic flavonoid scaffold, which consists of a fused benzopyrone core [54]. This structural motif enables interactions with diverse cellular targets and pathways. Notably, the specific substitution pattern featuring hydroxyl groups at the 5- and 7-positions, along with a methoxy group at the 4′-position, is known to facilitate an ESIPT process. Inspired by this structural insight, we investigated the photophysical behavior of acacetin and, unexpectedly, uncovered its previously unreported AIE property.

We systematically examined the fluorescence behavior of acacetin in dimethyl sulfoxide (DMSO)/water mixtures with varying water fractions (f_w). As illustrated in Fig. 1A, in pure DMSO—a good solvent for acacetin—both enol and keto emissions were observed, though the keto emission was relatively weak. Upon incremental addition of water—a poor solvent that promotes molecular aggregation—the keto emission intensified significantly while the enol emission diminished, culminating in a dramatic overall emission enhancement at high f_w values. This behavior is characteristic of AIE-active systems. The observed spectral evolution can be attributed to the ESIPT process, which is inherent to many flavonoid structures. The inset in Fig. 1A illustrates the equilibrium between the enol and keto tautomer. In aggregated states, restriction of intramolecular motion activates the radiative decay of the photo-induced keto species, giving rise to a notably large Stokes shift of ∼275 nm and the enhanced AIE fluorescence centered around 600 nm. Collectively, these results confirm acacetin as a natural ESIPT-AIEgen, wherein aggregation restricts non-radiative relaxation and promotes efficient keto-form emission. The AIE phenomenon was quantitatively confirmed by plotting the relative fluorescence intensity at the emission maximum against the water fraction (Fig. 1B). The fluorescence intensity remained low in solutions with high DMSO content but experienced a steep, multi-fold increase once the water fraction reached a critical threshold for aggregation, definitively ruling out the ACQ effect that plagues conventional fluorophores. Furthermore, we examined the concentration-dependent emission behavior of acacetin in an aggregation-inducing solvent mixture (Fig. 1C). The PL intensity exhibited a significant surge as the concentration increased from 1 to 30 μg mL⁻¹, with the spectral profile maintaining its characteristic ESIPT shape. The corresponding plot of intensity versus concentration (Fig. 1D) demonstrates a clear positive correlation, underscoring that molecular aggregation, whether driven by poor solvation or high concentration, is the fundamental origin of its luminescence. The absolute photoluminescence quantum yield of acacetin powder was measured to be 6.80%, a notable value for a natural flavonoid AIEgen.

Fig. 1.

Investigation of the AIE characteristics and aggregate morphology of acacetin. (A) Photoluminescence (PL) spectra of acacetin in DMSO/water mixtures with varying water fractions (f_w). The inset shows the proposed ESIPT mechanism between the Keto and Enol forms. (B) Plot of relative fluorescence intensity versus water fraction, confirming the AIE effect with emission enhancement upon aggregation. (C) PL spectra of acacetin at different concentrations in an aggregation-inducing solvent mixture. (D) Relationship between fluorescence intensity and acacetin concentration, showing concentration-dependent emission enhancement. (E) Theoretical calculations elucidating the molecular conformation and electronic structure underlying the AIE and ESIPT properties. (F) Transmission electron microscopy (TEM) image of acacetin aggregates. Scale bar is 100 nm. (G) High-magnification TEM image of a single aggregate. Scale bar is 5 nm. (H) Selected area electron diffraction (SAED) pattern of the aggregate in (G), confirming its crystalline nature. The combined TEM and SAED data demonstrate that acacetin forms well-defined nanocrystals in the aggregated state, which contributes to its AIE properties. Scale bar is 5 1/nm.

To gain deeper insight into the photophysical mechanism, theoretical calculations were performed (Fig. 1E). The results confirmed that the molecular conformation of acacetin facilitates an efficient ESIPT process. A bandgap of 4.08 eV, calculated from HOMO/LUMO energies of −6.02 eV and −1.94 eV, is consistent with typical flavonoid AIE emitters and explains its emission in aggregates. Crucially, the calculations corroborated that the restricted intramolecular rotation in the aggregate state suppresses non-radiative decay, thereby promoting the radiative ESIPT transition and providing a solid theoretical foundation for the AIE activity. The morphological foundation of the AIE activity was directly investigated through transmission electron microscopy (TEM). As shown in Fig. 1F, acacetin self-assembled into nanoribbons upon aggregation in aqueous media. High-magnification imaging (Fig. 1G) revealed that these nanostructures consisted of numerous randomly oriented crystalline domains. Selected area electron diffraction (SAED) pattern obtained from the sample (Fig. 1H) displayed sharp diffraction spots, confirming the polycrystalline nature of these assemblies with distinct nanocrystalline grains. This polycrystalline architecture is crucial for the AIE effect, as the numerous nanocrystalline regions create a rigid microenvironment that effectively restricts intramolecular motion, thereby suppressing non-radiative decay and unlocking intense emission through the ESIPT pathway. Collectively, these findings establish acacetin as a novel natural AIEgen whose luminescence is activated by the formation of polycrystalline nanoribbons upon aggregation, where the network of nanocrystalline domains provides the necessary confinement for efficient light emission.

3.2. Engineering and validation of a biomimetic EV platform for co-delivery across the blood-testis barrier

A pivotal challenge in treating male infertility is the efficient delivery of therapeutic agents across the physiological barrier that isolates the seminiferous tubules – the blood-testis barrier (BTB). To overcome this hurdle, we engineered a biomimetic nanoplatform based on EVs for the co-delivery of NMN and acacetin (denoted as EVs@N + A). EVs were isolated from human umbilical cord mesenchymal stem cell conditioned media and characterized to confirm their identity. The purity and identity of the isolated EVs were validated by Western blot, as evidenced by the positive expression of canonical EV markers (CD63, HSP70) and the absence of the negative control Calnexin (Fig. S1). Transmission electron microscopy revealed the classic cup-shaped morphology of the isolated plain EVs (Fig. 2A) and the successful formation of the drug-loaded EVs@N + A (Fig. 2C). Size distribution analysis from TEM images indicated that the loading process did not significantly alter the vesicle size (Fig. 2B–D). The successful incorporation of the cargoes was verified spectroscopically; the UV-Vis spectrum of EVs@N + A exhibited the characteristic absorption peaks of both acacetin and NMN, confirming their co-loading (Fig. 2E). Crucially, the EVs@N + A formulation retained the strong fluorescence emission of aggregated acacetin (Fig. 2F), indicating that the AIE property was preserved within the EV lumen, which is essential for subsequent tracking. The minor shift in zeta potential of EVs@N + A to approximately −25 mV, compared to plain EVs (Fig. 2G), indicates successful drug loading while confirming that the colloidal stability of the vesicles was maintained. The biocompatibility of all components was confirmed using a CCK-8 assay on GC-1 spg cells, with no significant cytotoxicity observed (Fig. 2H). Additionally, the particle size of EVs@N + A remained stable during one week (Fig. S2), indicating good colloidal stability. The encapsulation efficiencies of acacetin and NMN were 36.7% ± 4.6% and 61.7% ± 2.3%, respectively. Cumulative release profiles revealed pH-dependent release behavior for both drugs, with faster release at pH 5.0 compared to pH 7.4 (Figs. S3 and S4), suggesting favorable intracellular liberation under acidic endolysosomal conditions.

Fig. 2.

Characterization and biological evaluation of the engineered extracellular vesicle-based co-delivery platform (EVs@N + A). (A) Representative TEM image of the isolated plain EVs. Scale bar: 100 nm. (B) Size distribution histogram of the plain EVs based on TEM analysis. (C) Representative TEM image of the EVs after co-loading with NMN and acacetin (EVs@N + A). Scale bar: 100 nm. (D) Size distribution histogram of the EVs@N + A based on TEM analysis. (E) UV-Vis absorption spectra of free acacetin (A), free NMN (N), plain EVs, and the final formulation EVs@N + A. (F) Fluorescence emission spectra of the corresponding components, confirming the successful loading of acacetin and its retained AIE property within the EVs. (G) Zeta potential measurements of plain EVs, free acacetin (A), and the EVs@N + A formulation. (H) Biocompatibility assessment of GC-1 spg cells treated with EVs, N, A, and EVs@N + A using CCK-8 assay. (I) Fluorescence imaging showing EVs@N + A distribution in testicular tissue sections after administration. Cell nuclei were counterstained with DAPI. Red fluorescence: acacetin. Scale bar = 20 μm. (J) Quantitative analysis of EVs@N + A fluorescence intensity in testicular tissue from panel (I). Data are presented as mean ± SEM; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

The fundamental rationale for using EVs was their reported potential to traverse the BTB. We first confirmed the cellular uptake of EVs@N + A in GC-1 spg cells, which was time-dependent (Fig. S5). The ability to penetrate the BTB was directly tested using an in vitro model with Sertoli cells (TM4 cell line) in a transwell system (Fig. S6). Quantification of the fluorescence signal in the bottom chamber demonstrated that a significant portion of EVs@N + A successfully translocated across this barrier (Fig. S6, T-EVs@N + A). Pharmacological inhibition studies further revealed that EVs@N + A internalization was significantly reduced by the macropinocytosis inhibitor EIPA and the lysosome inhibitor bafilomycin A1, but not by the clathrin inhibitor chlorpromazine, indicating a macropinocytosis-dependent, lysosome-mediated transcytosis mechanism (Fig. S7). Most importantly, following in vivo administration, fluorescence imaging of testicular tissue sections revealed a robust accumulation of EVs@N + A within the seminiferous tubules (Fig. 2I and J), providing direct visual evidence of their testicular targeting and BTB-crossing efficiency in vivo. In vivo fluorescence imaging confirmed that the AIE signal of acacetin could be detected in the testicular region following systemic administration of EVs@N + A. Ex vivo organ distribution analysis at 24 h revealed that fluorescence signal was clearly discernible in the testis, while other organs exhibited a typical nanoparticle biodistribution profile (Fig. S8). These results validate the testicular targeting capability of our platform and demonstrate the practical utility of acacetin's AIE property for label-free in vivo tracking. Collectively, these data substantiate that our engineered EV platform effectively co-loads acacetin and NMN while preserving the critical AIE feature, exhibits excellent biocompatibility, and, most critically, possesses the innate ability to bypass the major delivery obstacle of the BTB, thereby ensuring targeted drug delivery to the testicular microenvironment.

3.3. Synergistic restoration of cellular energetics and inhibition of ferroptosis in vitro

Based on the successful construction and demonstrated targeted delivery capability of the EVs@N + A platform, we proceeded to systematically evaluate its therapeutic efficacy at the cellular level using a busulfan (BU)-induced injury model in GC-1 spg cells [55]. BU treatment effectively recapitulated the core pathological features of asthenozoospermia, simultaneously inducing a severe metabolic crisis and activating the ferroptosis cascade. As shown in Fig. 3A–C, BU exposure resulted in a drastic depletion of the intracellular NAD⁺ pool, a reduced NAD⁺/NADH ratio, and consequently, a profound deficit in ATP production, collectively reflecting a critical breakdown in cellular energy metabolism.

Fig. 3.

In vitro efficacy of EVs@N + A in protecting GC-1 spg cells from BU-induced damage. GC-1 spg cells were treated with PBS, busulfan (BU) alone, or BU combined with free acacetin (A), free NMN (N), plain EVs, or the engineered EVs@N + A. (A) Intracellular NAD⁺ levels. (B) NAD⁺/NADH ratio. (C) ATP production. (D) Representative fluorescence images of intracellular ROS detected by a DCFH-DA probe (scale bar = 20 μm). (E) Quantified ROS levels. (F) Reduced glutathione (GSH) levels. (G) Malondialdehyde (MDA) levels. (H) GPX4 enzyme activity. (I) Representative fluorescence images of lipid peroxidation using BODIPY™ C11 probe (scale bar = 20 μm). (J) Quantified lipid peroxidation (LPO) levels. Data are presented as mean ± SEM; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the BU group.

The NMN component within the EVs@N + A platform directly and effectively counteracted this metabolic impairment. As a potent NAD ⁺ precursor, NMN significantly replenished intracellular NAD⁺ levels, restored the NAD⁺/NADH ratio, and thereby fueled a marked recovery of ATP generation (Fig. 3A–C). This metabolic rejuvenation provided the essential bioenergy required to support spermatogenic cell survival and function. Concurrently, acacetin conferred robust cytoprotection through potent anti-ferroptotic effects. This was evidenced by the significant reduction in global ROS levels (Fig. 3D and E), elevation of reduced glutathione (GSH, Fig. 3F), marked decrease in the lipid peroxidation product malondialdehyde (MDA, Fig. 3G), attenuation of BODIPY™ C11 fluorescence (Fig. 3I and J), and enhanced activity of the key ferroptosis repressor GPX4 (Fig. 3H). Further corroborating this, EVs@N + A effectively mitigated the accumulation of cytotoxic ferrous iron (Fe²⁺, Fig. S9) and oxidized glutathione (GSSG, Fig. S10). Pharmacological inhibition assays further revealed that the apoptosis inhibitor Z-VAD-FMK conferred only partial protection, whereas the ferroptosis inhibitor Ferrostatin-1 provided the most robust rescue. The necroptosis inhibitor Necrostatin-1 exhibited negligible effects, confirming ferroptosis as the predominant, though not exclusive, cell death pathway (Fig. S11).

Notably, while free acacetin (A) or free NMN (N) alone provided partial rescue in certain parameters, the EVs@N + A combination consistently yielded the most profound and comprehensive therapeutic outcome across all assays. This demonstrates a clear synergistic effect, wherein NMN-driven bioenergetic amelioration and acacetin-orchestrated ferroptosis inhibition work in concert to restore cellular homeostasis. Furthermore, this enhanced efficacy is attributable not only to the drug synergy but also to the EV-based delivery platform, which significantly improves drug stability, facilitates cellular uptake, and bioavailability. The integration of these two complementary therapeutic modalities within a single EV-based platform thus constitutes a potent and unified strategy to counteract the multifactorial pathology of asthenozoospermia.

3.4. In vivo restoration of spermatogenesis and sperm function

Based on the compelling in vitro evidence demonstrating the dual-pathway efficacy of EVs@N + A, we next evaluated its therapeutic potential in vivo using a well-established busulfan (BU)-induced mouse model of asthenozoospermia [56]. A single intraperitoneal injection of BU (20 mg kg⁻¹), a chemotherapeutic agent with known testicular toxicity, effectively induced a state of impaired spermatogenesis characterized by the hallmark features of ferroptosis and metabolic dysfunction, thereby providing a pathologically relevant platform for therapeutic assessment (Fig. 4A). We hypothesized that systemic administration of EVs@N + A would mitigate BU-induced testicular injury and restore sperm quality through its coordinated mechanisms of bioenergetic amelioration and ferroptosis inhibition.

Fig. 4.

EVs@N + A rescues busulfan-induced asthenozoospermia in vivo by inhibiting ferroptosis and bioenergetic amelioration. (A) Experimental timeline. (B) Body and testis weight of mice after busulfan (BU) administration. (C) Schematic overview of analyzed sperm parameters. (D–L) Key sperm parameters: (D) total count, (E) motile cells, (F) progressive cells, (G) static cells, (H) average path velocity (VAP), (I) curvilinear velocity (VCL), (J) straight-line velocity (VSL), (K) amplitude of lateral head displacement (ALH), and (L) linearity (LIN). Data represent mean ± SEM, n = 5. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus BU group.

We confirmed the successful induction of the asthenozoospermia model three weeks after BU administration using comprehensive computer-assisted sperm analysis (CASA). This revealed a severe impairment of sperm function, evidenced by a significant reduction in total sperm count (TC), the percentage of motile cells (MC), the proportion of progressively motile sperm (PC), and the average path velocity (VAP) (Fig. 4B and C). Notably, these functional deficits emerged despite the lack of significant changes in body or testicular weight (Fig. 4B), a phenotypic signature that closely mirrors the clinical presentation of asthenozoospermia, where functional deficiency often precedes gross morphological alteration.

Therapeutic efficacy was evaluated following a regimen of three intravenous injections of EVs@N + A on days 22, 24, and 26. This treatment resulted in a striking and statistically significant recovery of core sperm functional parameters. As quantified in Fig. 4D–H, the EVs@N + A group exhibited a near three-fold increase in progressively motile cells (PC) compared to the BU-injured group, restoring PC to approximately 90% of the level observed in healthy, PBS-treated controls. Concurrently, the percentage of motile cells (MC) and total sperm count were substantially enhanced, while the population of static cells (SC) was markedly reduced. The average path velocity (VAP), a key indicator of overall sperm motility vigor, also showed significant improvement, suggesting a recovery of the intracellular energy status necessary for sustained movement.

A more detailed kinematic analysis revealed that other parameters, including straight-line velocity (VSL), curvilinear velocity (VCL), amplitude of lateral head displacement (ALH), linearity (LIN) (Fig. 4I–L), as well as wobble (WOB), beat cross frequency (BCF), and straightness (STR) (Fig. S12–S14), did not exhibit statistically significant differences across the various treatment and control groups. This indicates that the therapy most robustly rescues the fundamental determinants of fertility—sperm production, initiation of motility, and progressive movement—while finer aspects of motion kinematics appear to be less susceptible to this particular intervention or are secondary to the primary metabolic and survival deficits.

Crucially, the EVs@N + A combination therapy consistently yielded superior recovery across all significantly improved parameters compared to treatments with free acacetin (A) or free NMN (N) alone. This clear synergistic effect underscores the necessity of the co-delivery platform. The in vivo functional rescue provides direct validation of our proposed dual-mechanism: the NMN component replenishes the NAD⁺ pool to rejuvenate cellular metabolism and ATP-dependent processes, while the acacetin component concurrently protects spermatogenic cells from BU-triggered ferroptotic death. In addition, the superior therapeutic outcome of EVs@N + A, compared to the free drugs, can be attributed to the EV-based delivery platform, which not only acts as a superior carrier to enhance drug stability and bioavailability but also critically enables efficient translocation across the blood-testis barrier (BTB), thereby ensuring targeted accumulation within the testicular microenvironment. Collectively, these results firmly establish EVs@N + A as a potent and synergistic nanotherapeutic strategy that effectively ameliorates asthenozoospermia by coordinately targeting the interconnected pathways of metabolic insufficiency and ferroptotic cell death.

To systematically evaluate the therapeutic effects of EVs@N + A at the histological level, we performed Periodic acid-Schiff (PAS) staining on mouse testicular and epididymal tissues. The results revealed significant pathological differences and restorative trends among the experimental groups. In testicular histology (Fig. S15), control group testes exhibited typical normal spermatogenic architecture with tightly packed seminiferous tubules, intact tubular walls, and clearly layered spermatogenic cells. All developmental stages—from spermatogonia at the base to mature spermatozoa in the lumen—were completely present, with abundant sperm released into the tubular lumina. In contrast, the busulfan (BU) model group showed severe spermatogenic disruption: disorganized seminiferous tubules with irregular diameters, markedly thinned seminiferous epithelium, extensive loss of germ cells, near-complete absence of late-stage spermatids, rare sperm in the lumina, accompanied by significant vacuolation and interstitial edema. Following treatment, the EVs, N, and A monotherapy groups all partially alleviated these pathological changes. However, the most robust restoration of spermatogenesis was observed in the EVs@N + A combination group—their seminiferous tubules were largely intact with increased layers of spermatogenic cells, all developmental stages were visible, and clusters of mature sperm reappeared in the lumina, demonstrating nearly complete reconstitution of the spermatogenic cycle. Epididymal histology findings were highly consistent (Fig. S17). Control epididymal ducts were filled with densely packed, morphologically normal sperm; the BU model group exhibited typical “empty tubule” phenomenon with almost no residual sperm in the lumina; whereas all treatment groups, particularly the EVs@N + A group, showed significantly restored sperm storage in the epididymal ducts. Although not fully reaching control levels, this confirmed effective recovery of spermatogenesis and subsequent sperm transport and storage in the epididymis.

Notably, despite clear histological damage and recovery, testicular and epididymal indices remained comparable across groups (Figs. S16 and S18). This indicates that BU-induced spermatogenic impairment and its recovery occurred while the organ structure remained largely intact, further validating the model's reliability in mimicking clinical asthenozoospermia—where functional deficits precede morphological atrophy—while also ruling out potential systemic toxicity from the treatment itself. In summary, histological evidence clearly demonstrates the superior efficacy of EVs@N + A in repairing the spermatogenic microenvironment, supporting the complete process of spermatogenesis, and improving epididymal sperm storage, providing strong morphological support for its therapeutic role in restoring sperm function.

To precisely delineate the restorative effects of EVs@N + A across the complete spermatogenic lineage, we performed comprehensive immunofluorescence analysis using a panel of stage-specific germ cell markers (Fig. 5A, Fig. S19). Quantitative evaluation (Fig. 5B–E, Fig. S20) demonstrated that BU administration caused a catastrophic depletion of germ cells at every developmental stage. Western blot analysis (Fig. S22) corroborated these findings at the protein level. The population of GFRA1⁺ undifferentiated spermatogonia—the foundational stem cell pool—was severely diminished, disrupting the very initiation of spermatogenesis. This depletion propagated through subsequent stages, with significant reductions in DDX4⁺ spermatogonia and primary spermatocytes, SYCP3⁺ meiotic spermatocytes, PGK2⁺ post-meiotic round spermatids, and ultimately ACRV1⁺ elongated spermatids and mature sperm.

Fig. 5.

EVs@N + A rescues busulfan-impaired spermatogenesis by acting on multiple stages of germ cell development. (A) Representative immunofluorescence images of mouse testicular sections from different treatment groups (PBS, BU, BU + free acacetin (A), BU + free NMN (N), BU + plain EVs, and BU + EVs@N + A). The sections were stained with specific markers for different stages of spermatogenic cells: GFRA1 (for undifferentiated spermatogonia), SYCP3 (for spermatocytes), PGK2 (for round spermatids), and ACRV1 (for elongated spermatids and mature sperm). Cell nuclei were counterstained with DAPI. Scale bar: 20 μm. (B-E) Quantitative analysis of the fluorescence intensity for each germ cell marker from panel (A). Data are presented as mean ± SEM (n ≥ 3). Significance versus the BU group is indicated as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

While treatments with free acacetin or NMN alone provided moderate protection, manifesting as partial but statistically significant increases in certain cell populations, the EVs@N + A combination therapy demonstrated unparalleled efficacy. It most effectively and comprehensively restored the populations across the entire lineage: GFRA1⁺ spermatogonia, DDX4⁺ spermatogonia and primary spermatocytes, SYCP3⁺ spermatocytes, PGK2⁺ round spermatids, and ACRV1⁺ elongated spermatids. This robust, multi-stage rescue conclusively demonstrates that the nanoplatform creates a supportive microenvironment that ensures the successful execution of the entire spermatogenic process, from stem cell maintenance and meiotic progression to terminal spermiogenic differentiation.

3.5. Molecular mechanisms of ferroptosis inhibition and in vivo biosafety

At the molecular level, our investigation revealed that busulfan (BU) challenge established a profoundly dysregulated testicular microenvironment characterized by extensive ferroptotic signaling (Fig. 6A, Fig. S21). The injury model demonstrated significant suppression of Nuclear factor erythroid 2-related factor 2 (Nrf2), the master regulator of cellular antioxidant responses, which initiated a cascade of detrimental molecular events. This Nrf2 downregulation directly impaired the expression of SLC7A11 (xCT), the critical subunit of system Xc-that mediates cystine uptake, thereby creating a fundamental deficiency in glutathione (GSH) biosynthesis precursors. The consequent GSH depletion synergized with the parallel downregulation of Glutathione Peroxidase 4 (GPX4), the central enzyme responsible for reducing cytotoxic lipid hydroperoxides, leaving spermatogenic cells defenseless against peroxidative damage. Compounding this antioxidant system collapse, diminished expression of Ferritin Heavy Chain 1 (FTH1), a crucial iron-sequestration protein, resulted in dangerous expansion of the labile iron pool that fueled iron-dependent Fenton reactions. Simultaneously, marked upregulation of Acyl-CoA Synthetase Long-Chain Family Member 4 (ACSL4) enhanced the incorporation of polyunsaturated fatty acids into membrane phospholipids, creating abundant substrates for peroxidation cascades. This integrated molecular signature—comprising compromised Nrf2-SLC7A11-GPX4 defense mechanisms coupled with ACSL4-mediated membrane vulnerability and iron overload—established a self-reinforcing cycle of oxidative damage that progressively drove spermatogenic cells toward ferroptotic death through multiple synergistic pathways.

Fig. 6.

EVs@N + A alleviates busulfan-induced testicular injury by activating the anti-ferroptotic pathway. (A) Representative immunofluorescence staining of testicular sections from different treatment groups (PBS, BU, BU + free acacetin (A), BU + free NMN (N), BU + plain EVs, and BU + EVs@N + A) for key ferroptosis-related proteins. The markers include the transcription factor Nrf2 and its downstream targets (HO-1, SLC7A11), and the key anti-ferroptosis effectors (GPX4). Cell nuclei were counterstained with DAPI. Scale bar: 20 μm. (B-E) Quantified fluorescence intensity for proteins shown in (A). Data represent mean ± SEM from at least three experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 compared to BU group.

Comparative analysis of therapeutic interventions demonstrated a clear efficacy hierarchy in counteracting the ferroptotic cascade (Fig. 6B–E, Fig. S21). Monotherapy with free acacetin (Group A) provided moderate protection through partial restoration of Nrf2 signaling and GPX4 expression, consistent with its known antioxidant properties, yet failed to comprehensively address SLC7A11 deficiency or significantly suppress ACSL4 upregulation, leaving critical vulnerabilities in the ferroptosis pathway. Free NMN treatment primarily functioned through metabolic support by enhancing NAD⁺ bioavailability, which indirectly stabilized redox homeostasis through improved NADPH generation and provided modest support to glutathione metabolism, but exhibited limited direct impact on core ferroptosis regulators like GPX4 and ACSL4. Plain EVs delivered baseline protection through their native bioactive cargo, including endogenous FTH1 that partially ameliorated iron dysregulation, yet this unsustained effect lacked comprehensive transcriptional modulation. The engineered EVs@N + A platform demonstrated unparalleled efficacy through a sophisticated dual-mechanism synergy: NMN component systematically rejuvenated cellular metabolism by replenishing NAD⁺ pools, which may indirectly support Nrf2 signaling and facilitate the generation of essential reducing equivalents (NADPH) for glutathione recycling, which effectively “primed” the cellular antioxidant capacity. Concurrently, EV-facilitated acacetin delivery ensured targeted accumulation and potent activation of Nrf2 signaling—likely through Keap1 inhibition—while directly neutralizing existing lipid radicals. This coordinated metabolic and transcriptional reprogramming resulted in robust upregulation of the complete cytoprotective network (HO-1, SLC7A11, GPX4, FTH1) with simultaneous suppression of pro-ferroptotic ACSL4, establishing a multi-layered defense strategy that enhanced iron sequestration, restored antioxidant substrate availability, and markedly enhanced peroxide clearance capacity. This reconstituted protective microenvironment uniquely supported spermatogenic lineage survival and maturation, directly correlating with the superior functional recovery observed exclusively in the combination treatment group. These immunofluorescence findings were further corroborated by Western blot analysis, which independently confirmed the upregulation of cytoprotective proteins (Nrf2, HO-1, SLC7A11, GPX4, FTH1) and suppression of the pro-ferroptotic marker ACSL4 in the EVs@N + A group (Fig. S22).

Finally, the in vivo safety profile of the EVs@N + A platform was rigorously assessed through histopathological examination of major organs via H&E staining (Fig. S23). Compared to the PBS control group, no apparent pathological lesions, inflammatory infiltrates, or signs of toxicity were observed in any of the organs from the EVs@N + A-treated mice. This indicates the excellent biocompatibility and high safety profile of our engineered therapeutic strategy, supporting its potential for further translational development. Furthermore, treated males successfully sired healthy offspring. The EVs@N + A group showed restored offspring numbers compared to the BU group, and postnatal body weight of pups was comparable to controls during 21 days, indicating no detectable adverse effects on fertility or offspring development (Figs. S24 and S25).

Notably, the AIE property is not the direct chemical executor of ferroptosis inhibition—the flavonoid pharmacophore independently mediates the antioxidant effect. Rather, the AIE feature is indispensable at the conceptual and translational levels: upon aggregation at the target site, acacetin molecules undergo restriction of intramolecular motion, which suppresses non-radiative decay and ignites a turn-on fluorescence with inherently low background. This aggregation-enhanced emission enables label-free, real-time self-monitoring of drug delivery, collapsing the therapeutic agent and the imaging reporter into a single molecule. Such unification of function fundamentally distinguishes the AIE-Pharmacology paradigm from conventional natural-product therapies and represents the defining innovation of this platform. Unlike conventional AIE theranostics that rely on a synthetic AIEgen as an imaging tag for a separate drug, acacetin functions as both the therapeutic agent and the self-reporting probe in a single natural molecule. This natural origin confers inherent advantages—well-established biosafety, abundant availability from medicinal plants, and no requirement for de novo synthesis or extensive toxicity evaluation—significantly accelerating its translational potential.

4. Conclusion

In summary, based on our comprehensive findings, we have successfully established a novel “AIE-Pharmacology” paradigm for the effective treatment of asthenozoospermia. This work reports the natural flavonoid acacetin as a multifunctional theranostic agent, leveraging its intrinsic AIE property for self-monitoring therapy and real-time tracking. We have engineered a biomimetic EVs-based co-delivery system (EVs@N + A) to encapsulate the AIE drug acacetin and the metabolic precursor β-Nicotinamide Mononucleotide (NMN). This platform demonstrates a synergistic therapeutic mechanism: acacetin potently inhibits testicular ferroptosis by activating the Nrf2 pathway and upregulating key defense proteins (GPX4, SLC7A11), while NMN directly reprograms cellular metabolism by replenishing the NAD⁺ pool and restoring ATP production. This dual-pathway strategy effectively rescues busulfan-impaired spermatogenesis at multiple stages, significantly improving sperm count, motility, particularly progressive motility in vivo. Our work not only provides a robust and traceable nanoplatform for male infertility treatment but also establishes the “AIE-Pharmacology” concept as a potentially promising theranostic strategy with broad potential for application in other biomedical fields beyond reproductive medicine.

CRediT authorship contribution statement

Yujun Zhang: Data curation. Fengfei Zhao: Data curation. Feiyi Sun: Data curation. Haiyan Chen: Formal analysis. Lingling Wang: Methodology. Rendy Hosea: Software. Zheng Zhao: Visualization. Ben Zhong Tang: Supervision. Fei Sun: Funding acquisition. Xinghua Yu: Data curation, Supervision, Visualization, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.