Genetically bio-engineered PD-L1 targeted exosomes for immunotherapy of resistant triple negative breast cancer

Abstract

Immunotherapy has transformed cancer treatment by harnessing the immune system to target tumor cells, with PD-L1 inhibition emerging as a promising strategy. Exosomes, which naturally function as nanocarriers, offer significant potential for delivering therapeutic payloads, while genetic engineering allows for improved cargo specificity and efficacy. Here, for the first time, we genetically engineered exosomes to express anti-PD-L1 (PDL E) on their surface, enabling targeted drug delivery and immunotherapeutic activity. These engineered exosomes were then loaded with STAT3 siRNA (PDL ESi) and evaluated against doxorubicin-resistant MDA-MB-231 cells in combination with paclitaxel. Both in vitro and in vivo studies demonstrated a pronounced reduction in tumor burden (P < 0.001) and progression. Mechanistic investigations revealed that these exosomes activated apoptotic pathways, including the PI3K/AKT/mTOR axis, while inhibiting survival signals such as BCL-2, thereby enhancing tumor cell apoptosis. Notably, PD-L1 expression was downregulated in tandem with modulation of the STAT3/Nrf2 signaling axis, further augmenting the anti-tumor immune response. Toxicity studies in MCF-10 A cells showed that PDL ESi was well-tolerated, with no off-target effects. Imaging analyses in both 3D spheroids and tumor xenograft models confirmed the efficient tumor targeting of PDL E, demonstrating their time-dependent accumulation at the tumor site. Collectively, these findings highlight the promise of PD-L1-targeted, genetically engineered exosomes as a versatile platform for combination cancer therapy, providing a multifaceted strategy to overcome therapeutic resistance in TNBC.

Introduction

In the last four decades, the incidence of breast cancer (BC) has increased by approximately 0.5% annually, largely due to localized-stage and hormone receptor–positive disease [1]. BC is the second-leading cause of cancer-related mortality among women worldwide—after lung cancer—with 685,000 recorded deaths in 2020 [2]. These numbers underscore the importance of proper clinical management and the development of more effective assessment methods [3]. Molecular profiling of BC subtypes, based on biomarkers such as estrogen receptor α (ERα), progesterone receptor (PR), and human epidermal growth factor receptor-2 (ERBB2/HER2), has led to the discovery of multiple effective treatments [4–6]. However, the heterogeneous nature of BC necessitates therapeutic strategies that account for the genetic variability of each subtype [7, 8]. Traditional management usually consists of surgery, radiotherapy, and chemotherapy, while more recently, targeted therapies and immunotherapies have shown considerable promise [9].

Immune checkpoint inhibitors (ICIs), a cornerstone of immunotherapy, act by harnessing the patient’s own immune cells to combat tumors. These agents shift the therapeutic focus toward the tumor microenvironment rather than the tumor itself, resulting in durable clinical responses [10–13]. Although BC is initially considered an “immune silent” tumor and thus less responsive to immunotherapy, growing evidence shows that BC actually spans a range of immunogenic profiles, with triple-negative breast cancer (TNBC) appearing more immunogenic than other subtypes [14–18]. Building on the success of ICI therapy, additional immune-based interventions are being explored to elicit robust anti-tumor responses in various cancers [19]. Monoclonal antibodies targeting programmed death-1 (PD-1), programmed death ligand-1 (PD-L1), and cytotoxic T-lymphocyte–associated antigen-4 (CTLA-4) have effectively lifted the inhibitory regulation of T cell activation, thereby reinvigorating anti-tumor immunity [20–24].

In BC, PD-L1 expression has been observed in up to 34% of tumors, especially in high-grade, hormone receptor–negative cases [25–29]. The clinical efficacy of ICIs targeting PD-1/PD-L1 and CTLA-4 has been demonstrated in several solid tumors, including melanoma, non–small cell lung cancer, and renal carcinoma [21, 30]. Despite the FDA approval of multiple ICIs—such as Pembrolizumab, Atezolizumab, and Nivolumab—only a subset of patients derives significant benefit, with response rates in TNBC ranging from 10 to 30% as monotherapy [31, 32]. A phase I clinical trial (NCT01375842) reported a 24% objective response rate (ORR) and a median overall survival (OS) of 17.6 months in metastatic TNBC patients treated with Atezolizumab, compared to an ORR of 6% in heavily pretreated patients [33]. However, limitations such as poor tissue penetration, off-target effects, and immunosuppressive TME components continue to constrain ICI efficacy [34–39].

Another potential approach for anti-cancer therapy involves reprogramming the TME. The signal transducer and activator of transcription (STAT) 3 is frequently hyperactivated in human cancers and is often linked to poor clinical prognosis [40]. STAT3 is a key oncogenic transcription factor frequently activated in triple-negative breast cancer, where it contributes to tumor growth, metastasis, and resistance to therapy. In TNBC, constitutive STAT3 signaling drives the expression of genes involved in cell survival (e.g., BCL-2), angiogenesis (e.g., VEGF), and oxidative stress defense (e.g., SOD2), thereby supporting aggressive tumor phenotypes. Moreover, STAT3 fosters an immunosuppressive tumor microenvironment by upregulating cytokines such as TGF-β and IL-10 and inhibiting the function of cytotoxic T lymphocytes and antigen-presenting cells. This immune evasion contributes to poor clinical outcomes and reduced responsiveness to conventional chemotherapy. In light of these roles, targeting STAT3—particularly in combination with immune checkpoint blockade—has emerged as a promising strategy to restore anti-tumor immunity and enhance therapeutic efficacy [41].

Recently, various extracellular vesicles (EVs), especially 50–150 nm exosomes secreted by immune cells, have been investigated as cell-free platforms for cancer immunotherapy [42, 43]. Exosomes facilitate essential intercellular communication in both normal physiology and cancer, influencing metastasis, tumor progression, and immune regulation [44]. Their inherent advantages, high bioavailability, stability, and relatively low production costs make them particularly appealing over cell-based therapies [45, 46]. Notably, their capacity to transport siRNA or microRNA directly into target cells [47], holds promise for engineering exosome-based anti-cancer therapeutics that carry immunostimulatory cargo.

Considering, the limitations of single-agent ICIs and the ongoing need for highly targeted, multifaceted therapies, we developed a novel “bioengineered targeted exosome” bearing an anti-PD-L1 sequence derived from Atezolizumab. This exosome is designed to target PD-L1–expressing cells while simultaneously delivering STAT3 siRNA to disrupt critical tumorigenic and immunosuppressive pathways, in combination with the chemotherapeutic agent paclitaxel (Pac). Our study examines how effectively this combined approach can overcome therapeutic resistance in MDA-MB-231 cells.

Materials and methods

Materials

pDisplay vector with Anti-PD-L1 construct was purchased from Invitrogen. Lipofectamine 3000 (Cat # L3000001, ThermoFisher Scientific, MA, USA) OptiMEM media (Cat # 11058021, Gibco, NY, USA). HEK 293T, MDA-MB-231 cells were purchased from American Type Cell Culture (VA, USA). STAT 3 SiRNA, Matrigel, Acridine Orange, Ethidium Bromide, Cy5-DSPE, Aldehyde/sulfate beads (Cat # A37307, Invitrogen, CA, USA). All the antibodies (pSTAT 3-Cat # 12446, mTOR-Cat # 2972, PD-L1-Cat # 13684, BCL-2-Cat # 15072, NF-κB-Cat # 8242, PARP-Cat # 9542, Cleaved PARP-Cat # 9541) were procured from Cell Signaling Technology (MA, USA). MDA-MB-231 DOX RT and Docetaxel (DTX) RT cells were cultured in DMEM-F12 media (Cat # M23450, R&D Systems, MN, USA). HEK 293 T were cultured in DMEM media (Cat # 25–501B, GenClone, MN, USA).

Selection of target protein

Western blot analysis on MDA MB 231 resistant cells (DOX RT, DTX RT) and wild type cells was done. Cells were lysed using RIPA buffer containing protease and phosphatase inhibitor. Lysates were mixed 75:25 with 2 × Laemmli buffer (Bio-Rad, CA, USA) and 2-mercaptoethanol (Cat # M6250, Sigma-Aldrich, MO, USA). Samples were run on SDS–PAGE gel (Bio-Rad, CA, USA) and transferred to nitrocellulose membranes before being blocked in a 5% BSA solution for 90 min and incubated overnight at 4 °C with their respective primary antibodies followed by washing and secondary antibody incubation for 2 h at 25 °C. Chemiluminescence was performed on nitrocellulose membranes using Western Plus-ECL Enhanced Chemiluminescence (Bio-Rad, CA, USA) substrate before exposing them to X-ray film and development.

Construction of vector

For expression of anti-PD-L1 scFV, a sequence optimized peptide of 241 amino acids was engineered from available heavy and light chain sequences of Atezolizumab, humanized anti-PD-L1 monoclonal antibody (Accession # DB11595; Ref 1). The optimized scFV sequence consisted of amino acids 1–118 that were derived from N-terminal region of the heavy chain of Atezolizumab, followed by a 15-amino acid linker, and amino acids 134–241 that were derived from N-terminal region of the light chain of Atezolizumab. This scFV sequence was then cloned into the pDisplay Vector (Invitrogen, CA, USA) to obtain a recombinant plasmid for expression of a scFV protein. This recombinant scFV will therefore have murine Ig kappa-chain V-J2-C signal peptide and Hemagglutinin A epitope positioned at its N-terminus, and myc epitope and PDGFR transmembrane domain positioned at its C-terminus. The recombinant pDisplay plasmid encoding tagged, anti-PD-L1 scFV was then transfected into HEK293T cells [48].

Stable transfection of HEK 293T cells

HEK-293T cells were stably transfected using Lipofectamine 3000 as per company protocol. Briefly, around 0.5 million cells were seeded and after 24 h when the cells reached 70–80% confluency the cells were transfected with pDisplay vector encoding Anti-PD-L1 sequence for 6–8 h. After 6–8 h, the transfection media was replaced with DMEM media and maintained for 3 days. Transfected clones were selected for 3 weeks using Geneticin. Post selection, the anti PD-L1 gene expression was induced with doxycycline. The expression of anti PD-L1 was confirmed by western blot analysis.

Isolation of PDL E from transfected HEK 293T cells

Briefly, 10 million transfected HEK-293T cells were transferred to 0.5 L PBS bioreactor along with 1.25 g of Cytodex I microcarrier beads (Cat # CLS3772–1EA, Corning, NY, USA) in exosome-free DMEM media. The bioreactor was set to run at 25 RPM for 5 min with a resting period of 15 min for 8 cycles. Around 40% of the media was replaced every 4 days with fresh media. Exosomes were isolated from the media collected using ultracentrifugation process. The media was subjected to serial centrifugation steps at 4° C for 5 min at 500×g, for 10 min at 2000×g and for 30 min at 10,000×g followed by overnight incubation with 16% PEG 6000 at 4°C on a shaker in the ratio 1:1. The supernatant was collected in each step and transferred to new centrifuge tubes for sequential steps. After incubation, the media was centrifuged at 3000×g for 70 min at 4°C followed by ultracentrifugation at 100,000×g for 70 min and the obtained pellet of exosomes was resuspended in PBS and stored at −80°C for further processing [49].

Characterization of PDL E

Nano tracking analysis

Exosomes were characterized using ZetaView instrument (ZetaView^® TWIN PMX-220) (Particle Matrix, NC, USA) for particle size and zeta potential at 25°C and 90° scattering angles. Exosomes were diluted in the ratio of 1:10,000 using PBS. Exosomes were also characterized for the expression of tetraspanins (CD81, CD63) using Fluorescent NTA (FNTA). Briefly, exosomes were tagged with fluorescently labelled CD81 and CD63 antibodies for 2 h as per the protocol given by the manufacturer and analyzed by FNTA using a 488 fluorescent filter [50].

Protein content

Protein content of exosomes was determined by BCA assay. Approximately, 5uL of sample was treated with 100 uL BCA reagents A&B (Cat # 23224, ThermoFisher Scientific, MA, USA) and the absorbance values were measured at 562 nm. The concentration was calculated from the standard curve of bovine serum albumin (Cat # A9418–50G, Sigma-Aldrich, MO, USA).

Expression of anti PD-L1 on exosomes

Expression of AntiPD-L1 on exosomes was confirmed using flow cytometer. Briefly, exosomes were incubated with Aldehyde/Sulfate beads overnight at 4°C on a shaker. The exosomes were blocked by 2% BSA for 30 min. Then they were incubated with fluorescent c-myc antibody (50:1) (Cat # 9402 S, CST, MA, USA) for 30 min followed by washing with PBS by centrifugation at 2000×g for 5 min. The pellet was resuspended in PBS and analyzed by using Sony SH800 Cell Sorter (San Jose, CA, USA).

NanoFCM

Exosomes were incubated with fluorescent c-myc antibody (50:1) for 30 min followed by washing with PBS by centrifugation at 2000×g for 5 min. The pellet was resuspended in PBS and analyzed by using NanoFCM instrument at the University of Florida core facility.

Knockdown of PD-L1 in different in-vitro and in-vivo tumors by PDL E and dose titration with atezolizumab

Briefly 0.5 million cells (MDA-MB-231 DOX RT, MIA-PACA 2) were seeded in a 6-well plate and treated with 1 × 10¹⁰ particles/mL of PDL E the following day for 24 h. Post treatment the cells were lysed using RIPA buffer and analyzed by western blotting for PD-L1 expression levels. For in-vivo studies two million MDA-MB-231 DOX RT and H1975 OSM RT cells were injected to BALB/c athymic nude mice to develop xenograft mouse models. PDL E (1 × 10¹¹ particles/animal) were injected intraperitoneally for one week on alternate days when the tumor size was 150mm3. Tumors were isolated and characterized for PD-L1 expression using western blotting. To compare the relative efficacy of atezolizumab (Catalog # MA5–42046, Thermofisher Scientific, MA, USA) and PDL E, a dose titration study was conducted using western blot analysis. MDA-MB-231 DOX RT cells were treated with different concentrations of Atezolizumab (25, 50, 100 nM) and PDL E (10⁹, 10¹⁰, 10¹¹ particles/mL) for 24 h and analyzed for PDL1 expression using western blotting.

Loading and quantification of STAT 3 SiRNA into PDL E

STAT 3 SiRNA was loaded into the PDL E using electroporation. The siRNA was prepared in an electroporation buffer and mixed with the PDL E. The mixture was then transferred to an electroporation cuvette, and an electrical pulse (typically 300–400 V for 5 milliseconds) was applied. The electroporated mixture was immediately transferred to PBS for stabilization. The siRNA-loaded exosomes were purified by ultracentrifugation at 100,000 ×g and washed with PBS to remove free siRNA. siRNA loading was validated via Quickdrop spectrophotometer (Molecular Devices, CA, USA) to ensure successful incorporation and functionality within target cells.

Cytotoxicity studies

2D cytotoxicity studies

A concentration-dependent cytotoxicity assay was conducted using MDA-MB-231 DOX RT cells. The cells were seeded at a density of 7,000 cells per well in a 96-well plate and maintained in an incubator at 37 °C with 5% CO₂. After 24 h, the media was replaced with experimental media containing different treatments (Doxorubicin, PDL E, STAT 3 SiRNA, Pac, PDL ESi, PDL ESi + Pac) followed by incubation for 48 h. Subsequently, the cells were treated with MTT solution (0.5 mg/mL) for 3 h to allow the formation of formazan crystals, which were then dissolved in DMSO. The absorbance was measured at 570 nm using a Tecan Infinite 200 PRO M Plex multimode microplate reader (Tecan, Switzerland) [49].

MTT assay was also conducted to assess the off-site toxicity of PDL E using MCF-10 A cell line. Combination index values were determined using Compusyn software which emphasizes if the treatment with combination of two therapies is additive, synergistic or antagonistic [51].

3D cytotoxicity studies

3D spheroids were cultured following established laboratory protocols using a magnetic nanoshuttle system. Briefly, MDA-MB-231 DOX RT cells were tagged with a nanoshuttle solution (10 μL for 10,000 cells; Greiner-Freickenhausen, Germany) by performing three rounds of centrifugation at 800 RPM for 7 min each. A total of 15,000 cells per well were seeded in cell-repellent surface 96-well plates and placed on a magnetic drive to facilitate the formation of 3D spheroids. The plate was maintained in an incubator at 37 °C with 5% CO₂ for 5 days, with a media change on the 3rd day. The spheroids were then treated with the above-mentioned treatments for 48 h. Subsequently, the media was replaced with 0.5 mg/mL MTT solution, and the spheroids were incubated for an additional 3 h. The resulting formazan crystals were dissolved in DMSO, and the absorbance was measured at 570 using a TECAN infinite M200 plate reader nm [52].

In-vitro mechanistic studies

Apoptosis assay

We assessed the apoptotic potential of treatment groups on MDA-MD-231 DOX RT Spheroids using the Acridine orange /Ethidium bromide double staining test. 3D spheroids were cultured using a low attachment plate for 3 days. On day 3, the spheroids were treated with sub-optimal concentrations of Pac, STAT 3 SiRNA, PDLE, PDLE Si, PDLE Si + Pac and incubated for 48 h. After 48 h, the spheroids were treated with acridine orange (10 μL of 100 μg/mL diluted to 100 μL using PBS) and ethidium bromide (10 μL of 100 μg/mL diluted to 100 μL using PBS) and incubated at room temperature for 30 min. After incubation the spheroids were washed thrice with PBS and fluorescence was measured using Olympus IX 73 fluorescence microscope (Olympus, Center Valley, PA) [53].

Western blotting

Briefly, the treated cells were lysed using a mixture of RIPA buffer (cat no.89900) (Thermo scientific, USA), protease (cat no. P8340), and phosphatase inhibitors (cat no. P2850) (1:100) (Sigma Aldrich’s. Louis, MO, USA), and the protein content was estimated by BCA assay. An equivalent volume of 30 μg of protein was loaded into SDS PAGE gel and run for protein separation. The proteins were transferred from gels to nitrocellulose membranes followed by blocking with 5% BSA. The blocked membranes were incubated with primary antibodies at 4° C overnight. The secondary antibodies corresponding to the primary antibodies were added and incubated for 2 h at room temperature. Following secondary antibody removal, the blots were washed thrice with TBST and developed with ECL substrate using ChemiDoc^™ XRS + Imaging system (BIO-RAD). The relative band densities were determined using densitometric analysis software (Image J 1.36; Wayne Rasband, National Institutes of Health, MD, USA).

qPCR

RNA was extracted from cells using TRIzol reagent (Invitrogen, CA, USA) and purified using a RNeasy Mini kit (Qiagen). A260/280 absorbance ratio of the extracted RNA was measured using a Quickdrop UV-Visible spectrophotometer (Molecular devices, San Jose, CA) to evaluate the quality and integrity of samples. To examine the mRNA levels of specific genes, cDNA synthesis from total RNA was performed according to the manufacturer’s instructions using the Maxima H Minus First-strand cDNA Synthesis Kit (Thermo Fisher Scientific, Lithuania). Various gene primers (STAT3, NF-κB), GAPDH were purchased from Integrated DNA Technologies, Inc (Table below). Quantitative PCR was used to detect gene expressions using SsoAdvanced^™ Universal SYBR Green Supermix (Bio-Rad) and the CFX96 Touch^™ Real-Time PCR Detection System (Bio-Rad Laboratories). Post amplification, a melt curve analysis was used to determine the reaction’s specificity. The whole mean expression level of GAPDH genes was used as a reference for comparison when assessing relative mRNA expression using the comparative Ct (ΔCt) technique [54].

GENE	PRIMER
STAT3	F: 5’-GAGGCATTCGGAAAGTATTGTC-3’ R: 5’-CATCGGCAGGTCAATGGTAT-3’
GAPDH	F: 5’-ACATCGCTCAGACACCATG-3’ R: 5’-TGTAGTTGAGGTCAATGAAGGG-3’
NF-κB	F: 5’-CAGGAGACGTGAAGATGCTG-3’ R: 5’-AGTTGAGAATGAAGGTGGATGA-3’

In-vivo studies

BALB/c athymic nude mice (male, 6 weeks old, Foxn1^nu) were purchased from Envigo and housed under stringent pathogen-free conditions in an American Association for Accreditation of Laboratory Animal Care-approved facility. The animals were kept in typical housing cages with access to food and water, at a temperature of 37 °C and a relative humidity of 60%. Institutional Animal Care and Use Committee (IACUC) regulations of Florida A&M University were followed throughout the study (Protocol Number for animal studies: 023 – 02). Mice were acclimatized for a week before inducing the tumors.

MDA-MB-231 DOX RT cells were mixed with vitrogel (TheWell Bioscience, NJ, USA) in the ratio of 1:1 to form a suspension. The suspension was kept at 37° C for 25 min. followed by 5 min. on ice. Subsequently, 2 × 10⁶ MDA-MB-231 DOX RT cells were subcutaneously injected into the right flank region of athymic nude mice. When the tumor volume reached 50–100mm³, the animals were randomly separated into six treatment groups, each with five animals after 10 days. The active treatment groups were administered with PDL E (1 × 10¹¹ particles/animal, i.p.), STAT 3 SiRNA (150 μg/mL, i.p.), Pac (5 mg/kg, i.p.), PDL ESi (150 μg/mL in 1 × 10¹¹ particles/animal, i.p.), PDL ESi + Pac (150 μg/mL in 1 × 10¹¹ particles/animal + 5 mg/kg, i.p) on alternate days for two weeks. The tumor dimensions were measured using vernier caliper throughout the treatment and the tumor volume was calculated using the tumor volume formula: 1/2 xy², where ‘x’ and ‘y’ represent the length and width of the tumors. Tumors were collected by euthanizing the animals and processed for further evaluation.

Western blotting in-vivo

Western blotting of tumor samples was carried out to evaluate the molecular mechanism as per the standard protocol. Briefly, the tumor samples were lysed using a mixture of T-PER^™ Tissue protein extraction reagent (Cat no.78510), Thermo Scientific, USA), protease (Cat no. P8340), and phosphatase inhibitors (Cat no. P2850) (1:100) (Sigma Aldrich, St. Louis, MO, USA), and the protein content was estimated by BCA assay. An equivalent volume of 40 μg of protein was loaded into SDS-PAGE gel and run for protein separation. The proteins were transferred from gels to nitrocellulose membranes followed by blocking with 5% BSA. The blocked membranes were incubated with primary antibodies at 4° C overnight. The secondary antibodies corresponding to the primary antibodies were added and incubated for two hours at room temperature. Following secondary antibody removal, the blots were washed thrice with TBST and developed with ECL substrate using ChemiDoc^™ XRS + Imaging system (BIO-RAD, CA, USA). The relative band densities were determined using densitometric analysis software (Image J 1.36; Wayne Rasband, National Institutes of Health, MD, USA).

Tumor targetability of PD-L1 E

3D spheroids

Cy5-DSPE-labeled exosomes were prepared by incubation of Cy5-DSPE (10μL of 1 mg/mL) with PDL E (1mL of 1 × 10¹¹ particles) at 4°C for 2 h. To evaluate the uptake of Cy5-DSPE-labeled exosomes (Cy5 PDL E) in MDA-MB-231 DOX RT cells, they were grown as 3D spheroids. The spheroids were prepared by seeding 10,000 cells per well in 96-well ultra-low attachment plates and were grown for three days at 37 °C with 5% CO₂. The labeled exosomes were added to the spheroids at a concentration of 10 μg/mL and incubated for 4–24 h. After incubation, the spheroids were washed with PBS to remove unbound exosomes and analyzed using confocal microscopy to assess fluorescence intensity. Uptake efficiency was determined by quantifying the mean fluorescence intensity of the spheroids using ImageJ analysis software (Image J 1.36; Wayne Rasband, National Institutes of Health, MD, USA) as compared to the untreated controls.

In-vivo imaging

Cy5-DSPE-Labeled exosomes were administered via intraperitoneal (IP) injection. At 2, 4 and 24 h post-injection, the mice were imaged using the Analytik Jena UVP iBox (Analytik Jena, CA, USA) instrument to assess the distribution and uptake of the Cy5 PDL E in real-time. Fluorescence signals were monitored in the tumor region to evaluate the uptake efficiency, the signal intensity was quantified using the instrument’s software. Control groups treated with Cy5-DSPE were used to compare the specific uptake in the tumor tissue.

Statistical analysis

To determine statistical significance, the values of each treatment group were compared to the respective controls either by One-Way ANOVA with Brown-Forsythe test or Two-Way ANOVA with Brown-Forsythe multiple comparisons test using GraphPad Prism version 8.0 for Windows (San Diego, CA) and P < 0.05 was considered significant. T-tests were used to compare two groups in some experiments.

Results

PD-L1 overexpression in drug-resistant TNBC cells and its potential as a therapeutic target

To justify the selection of PD-L1 as a target protein in this study, the expression levels of several key proteins, including PD-L1 (p < 0.001), STAT3 (p < 0.001), NRF2 (p < 0.001), and PD-1 (p < 0.001), were analyzed. Western blot results demonstrated that these proteins were significantly overexpressed in drug-resistant TNBC cell lines, such as MDA-MB-231 DOX RT and MDA-MB-231 DTX RT, compared to MDA-MB-231 wild-type cells (supplementary Fig. 1). Based on these findings, PD-L1 was chosen as the primary target due to its significant upregulation in the doxorubicin-resistant MDA-MB-231 cell line, suggesting a strong association with chemoresistance. These results indicate that PD-L1 inhibition could play a crucial role in sensitizing resistant TNBC cells to therapeutic intervention, particularly when combined with chemotherapy.

Characterization of PDL E

The NTA analysis revealed that the PDL E obtained showed a mean particle size of 100 ± 10 nm and a zeta potential of −25.81 ± 5 mV, with a particle concentration of 0.8–3.2 × 10¹¹ particles/mL. Control exosomes, (exosomes collected from un-transfected HEK293t cells), exhibited a mean particle size of 100 ± 10 nm and a zeta potential of −25 ± 5 mV, with a particle concentration of 0.7–2.5 × 10¹¹ particles/mL No significant difference was observed in the particle size and surface charge due to surface modification of exosomes as well as after loading STAT 3 SiRNA (Fig. 1). Fluorescent nanoparticle tracking analysis (NTA) confirmed that 81.25% of these exosomes expressed tetraspanins, (CD63 and CD81), emphasizing the purity of the exosomes. The expression of myc-Anti-PD-L1 was determined through western blotting. Flow cytometry also confirmed the presence of myc-Anti-PD-L1 revealing approximately 80% of exosomes were found to be stained by fluorescent myc antibody. Furthermore, NanoFCM analysis showed 80% of the exosomes expressed anti-PD-L1 on their surface.

Fig. 1.

Generation of stable anti-PD-L1 cell line, isolation, and characterization of transfected exosomes. A) Western blot represents expression of anti-PD-L1-Myc in HEK293T cells post transfection. B) Size and Zeta Potential of HEK-293 T exosomes C) Size and Zeta potential of Anti-PD-L1 exosomes D) Fluorescent NTA of transfected exosomes E) Expression of anti-PD-L1 on the surface of exosomes isolated from transfected HEK 293T cells. F) NanoFCM analysis for quantifying the number of copies of anti-PD-L1 expressed relative to the total number of exosomes. G) FACS analysis showing anti-PD-L1 presence on the surface of transfected exosomes

PDL E knocks down PD-L1 in vitro and in vivo

The ability of PDL E to target and knockdown PD-L1 was evaluated by western blot analysis. The western blot analysis showed significant knockdown of PD-L1 in both in-vitro and in-vivo in MDA-MB-231 DOX RT tumors as well as H1975 OSM RT tumors and MIA PACA 2 cells (pancreatic tumor) suggesting that these exosomes can be targeted to various cancers expressing PD-L1 (Fig. 2). The knockdown of PD-L1 by PDL E has significant therapeutic implications, particularly in the context of immune checkpoint inhibition. PD-L1 is a key immune checkpoint protein that binds to PD-1 on T cells, effectively “turning off” the immune response against tumor cells. In many cancers, upregulation of PD-L1 allows tumor cells to evade immune detection and destruction by suppressing T cell activity through the PD-L1/PD-1 axis [55].

Fig. 2.

PD-L1 knockdown in different tumors post treatment with PDL E. A) MDA-MB-231 DOX RT cells. B) MDA-MB-231 DOX RT xenograft tumors. C) H1975 OSM RT xenograft tumors. D) MIA PACA 2 cells

By knocking down PD-L1 in MDA-MB-231 DOX RT cells (P < 0.05), MDA-MB-231 DOX RT tumors (P < 0.05), H1975 OSM RT tumors (P < 0.05), and MIA PACA 2 cells (P < 0.05), PDL E could reverse this immune suppression, thus promoting T cell activation and enhancing the immune response against the tumor. This targeting approach could potentially sensitize tumors to immunotherapy by inhibiting the PD-L1/PD-1 interaction, a strategy already validated by the success of monoclonal antibodies against PD-1 and PD-L1 in cancer immunotherapy [56].

To compare the relative efficacy of atezolizumab, a PD-L1 inhibitor, and PDLE, a dose titration study was performed using western blot analysis. This study evaluated the effects of atezolizumab and PDLE on PD-L1 expression in MDA-MB-231 DOX RT cells. PD-L1 protein levels were quantified across various treatment doses, revealing differential modulation of PD-L1 expression. Notably, PDLE exhibited dose-dependent effects that were comparable to or exceeded those of atezolizumab. These findings underscore the therapeutic potential of PDLE as an innovative strategy to address PD-L1-mediated resistance mechanisms in triple-negative breast cancer (Supplementary Fig. 4).

PDL E inhibits viability of MDA-MB-231 DOX RT cells in 2D and 3D cultures

To investigate the cytotoxicity of the PDL E against the MDA-MB-231 DOX RT cells, the viability of the cells after PDL E treatment was examined by MTT assay and results showed that PDL E demonstrated 50 ± 0.5% cell kill against MDA MB-231 DOX RT cells. The cell line used was highly resistant to doxorubicin with an EC50 value of 47.21 ± 0.43μM. In comparison to PDL E-treated cells, STAT 3 siRNA was treated with a different group of cells and the EC50 value obtained was 4.54 ± 0.28nM. However, with PDL ESi, the EC50 value decreased by 2.65-fold indicating synergism (P < 0.0001). Moreover, when the combination of PDL ESi and Pac was used, the EC50 improved the cytotoxicity effect of Pac by 3.86-fold (P < 0.0001) (Combination Index, CI = 0.4, Supplementary Fig. 2) (Fig. 3).

Fig. 3.

In-vitro cytotoxicity studies of MDA-MB-231 DOX RT cells. A) 2D Cytotoxicity studies. B) 3D Cytotoxicity studies in spheroids. Results were expressed as μM/nM in Mean ± SD (n = 4). ****P < 0.0001

In 3D culture, PDL E treatment revealed 30 ± 0.7% of cell death in MDA-MB-231 DOX RT cells (Fig. 3). While PDL ESi treated cells, the EC50 was reduced significantly by 2.17-fold (P < 0.0001) and was observed to be 4.35 ± 0.51nM relative to STAT 3 SiRNA by itself (EC50 = 9.44 ± 0.85nM). The PDL E in combination with Pac reduced the EC50 value of Pac significantly by 7.6-fold indicating strong synergism (P < 0.0001) (Combination Index, CI = 0.6, Supplementary Fig. 2).

Based on the cytotoxicity results, treatment of the cells with PDL E showed prominent negative effect on the viability of the cells in 2D and 3D culture. A more pronounced cytotoxicity on the cells was observed in the cells treated with the combination, PDL ESi + Pac, implying the efficacy of a ‘cocktail’ effect for drug resistant breast cancer cells. To evaluate the off-target toxicity of PDL E, cytotoxicity assays were conducted using MCF-10 A cells. The results demonstrated that PDL E exhibited no detectable toxicity in these non-cancerous cells, supporting the safety profile of these exosomes for therapeutic applications (Supplementary Fig. 3).

PDL E induces apoptosis in MDA-MD-231 Dox RT spheroids

PDL E showed cytotoxic effects on the cells in both 2D and 3D cultures. To examine whether PDL E treatment plays a role in inducing apoptosis in the cells, cell staining and expression of several apoptotic proteins were investigated. The results of the apoptosis assay are presented in Fig. 4, illustrating the mean intensity, which is the ratio between intensity of the live cells (green fluorescence) and dead cells (red fluorescence), relative to the control for various treatment conditions. Pac treatment significantly reduced the mean intensity to 58.9 ± 1.1% (p < 0.0001), indicating a pronounced apoptotic response. Similarly, treatment with STAT 3 SiRNA decreased the mean intensity to 48.45 ± 1.89 (p < 0.0001). Treatments with PDL E and PDL ESi resulted in mean intensities of 62.65 ± 1.5 and 40.89 ± 1.83, respectively, both showing significant reductions compared to the control (p < 0.0001). The combination treatment, PDL ESi + Pac further led to substantial reduction in mean intensity to 15.48 ± 1.48 (p < 0.0001). These results suggest the role of PDL E in inducing apoptosis, which is significantly increased when combined with Stat3 siRNA and Pac (Fig. 4A).

Fig. 4.

Assessment of apoptosis and relevant apoptotic proteins in MDA-MB-231 spheroids and xenograft after respective treatment. A) Fluorescence microscopy images of MDA-MB-231 spheroids stained with ethidium bromide (top row), acridine orange (middle row), and merged images (bottom row). Ethidium bromide (red) stains dead cells, while acridine orange (green) stains live cells. The merged images indicate the relative proportions of live and dead cells in each spheroid, demonstrating increased cell death in the combination treatment group (PDL ESi + Pac). B) Western blot analysis of protein expression levels in MDA-MB-231 xenografts treated with the indicated treatment. Blots were probed for markers associated with apoptosis (Cleaved PARP, Cleaved caspase 3), survival (BCL 2, pAKT, SOD2), and key signaling pathways (pMTOR, and pSTAT 3). β-actin was used as a loading control. Quantification of protein expression was performed by normalizing the intensity of each target protein band to its corresponding β-actin

Apoptosis was further confirmed by assessing the expression of apoptotic proteins through western blot analysis. Our results demonstrated a significant downregulation of key regulators of the PI3/AKT/mTOR pathway, including p-AKT (P < 0.0001) and p-mTOR (P < 0.0001), in MDA-MB-231 DOX RT cells following treatment with PDL E Si + Pac (Fig. 4B). Additionally, other downstream proteins involved in apoptotic pathways, such as PARP (P < 0.0001) and BCL-2 (P < 0.0001), were also significantly downregulated. In contrast, apoptotic markers, including cleaved caspase-3 (P < 0.0001) and cleaved PARP (P < 0.0001), were upregulated, further confirming the induction of apoptosis. These findings suggest a strong synergistic effect between Pac and PDL E at the molecular level (Fig. 4B).

PDL ESi + Pac mediates Inhibition of PD-L1 and JAK/STAT pathway

PDL ESi when combined with Pac, showed significant effects on the PD-L1-PD-1 axis within the tumor microenvironment. Treatment with these exosomes resulted in a marked downregulation of PD-L1 (P < 0.0001) expression, indicating a successful blockade of this immune checkpoint. Additionally, key cytokines such as TGF-β (P < 0.0001) and IL-1β (P < 0.0001) were also downregulated, suggesting enhanced T-cell activation and a more favorable immune response against the tumor. Furthermore, the downregulation of pSTAT3 (P < 0.0001), a critical marker of the JAK/STAT signaling pathway, was observed, reinforcing the therapeutic potential of this combination strategy in reactivating anti-tumor immunity (Fig. 5).

Fig. 5.

Representative western blots showing activation of immune mediators. A) Treatment of MDA-MB-231 DOX RT cells with PD L1 targeting exosomes (PDL E) lead to activation of T cells and downregulation of IL-1β and TGF-β in tumor microenvironment which further acts on tumor cells and activates apoptotic pathways. β-actin was used as a loading control. Quantification of protein expression was performed by normalizing the intensity of each target protein band to its corresponding β-actin. B) qPCR depicting the mRNA expression levels

PDL E specifically binds to the MDA-MB-231 Dox RT cells

To evaluate the cellular uptake of exosomes by the cancer cells, exosomes were fluorescently labeled and incubated with MDA-MB-231 DOX RT cells in vitro. As shown in the fluorescent microscopy images (Fig. 6A), Cy5-DSPE exhibited low levels of internalization, indicated by sparse fluorescent signals within the cells. In contrast, PDL E significantly enhanced uptake, as evidenced by a more intense and widespread fluorescence distribution in the cytoplasm of MDA-MB-231 DOX RT cells. Quantification of the mean fluorescence intensity further confirmed this observation, with PDL E showing approximately a two-fold increase in cellular uptake compared to Cy5-DSPE (p < 0.05). These results suggest that PD-L1 targeting improves the selective uptake of exosomes by MDA-MB-231 DOX RT cells, which is crucial for the therapeutic efficacy of exosome-based delivery systems.

Fig. 6.

Evaluation of tumor-targeted exosome uptake and therapeutic efficacy in vivo. A) Fluorescent microscopy images of Cy5 PDL E (red) incubated with MDA-MB-231 DOX RT cells. (Top) Cy5-DSPE shows lower uptake by TNBC cells, while (Bottom) PDL E shows significantly enhanced cellular internalization. Scale bars represent 20 μm. Quantitative analysis of the mean fluorescence intensity in MDA-MB-231 DOX RT cells treated with Cy5-DSPE vs. Cy5 PDL E demonstrate enhanced uptake of targeted exosomes (p < 0.05). B) In vivo fluorescent imaging of mice with MDA-MB-231 DOX RT tumors after administration of Cy5 PDL E. (Top panel) Mice treated with Cy5-DSPE show minimal tumor accumulation, whereas (Bottom panel) mice treated with Cy5 PDL E exhibit pronounced exosome accumulation at the tumor site

PDL E targets TNBC xenografts in mice

To verify that PDL E specifically targets and binds to tumor cells only, we investigated the in vivo biodistribution and tumor-targeting ability of PDL E. MDA-MB-231 DOX RT xenograft tumor-bearing mice were administered with Cy5 PDL E, and whole-body imaging was performed (Fig. 6B). Mice treated with CY5-DSPE displayed minimal accumulation at the tumor site, as indicated by the weak fluorescence signal in the tumor region. In contrast, mice receiving Cy5 PDL E showed a strong fluorescence signal localized specifically to the tumor, suggesting effective targeting and retention of exosomes at the tumor site. Quantitative analysis of the exosome accumulation in the tumor region, based on fluorescent imaging, revealed a significant difference between the two groups (Fig. 6B). Mice treated with PDL E exhibited a three-fold higher accumulation of exosomes in the tumor compared to the Cy5-DSPE group (p < 0.01). These results demonstrate that PD-L1 targeting enables the preferential accumulation of exosomes in TNBC tumors, which has important implications for improving the efficacy of exosome-based drug delivery platforms.

PDL E impedes tumor growth in mice

Cell-based assays suggest the direct involvement of PDL E and PDL ESi + Pac in inducing apoptosis in MDA-MB-231 DOX RT cells. Thus, the effect of various treatments on tumor growth over a 15-day period was also investigated. The control group exhibited the highest tumor volume, reaching 3,612.16 ± 2 62.38 mm³. Treatments with STAT 3 siRNA and Pac moderately reduced tumor growth, with final volumes 2,560.95 ± 124.01 mm³ and 2,326.10 ± 287.54 mm³, respectively. The PDL E group showed similar results to Pac, while the PDL ESi treated group exhibited further reduced tumor volumes at 1,587.20 ± 301.63 mm³. Treatment with PDL ESi + Pac revealed significantly inhibited tumor growth with only 786.57 ± 43.74 mm³ (P < 0.0001) by day 15. These findings indicate a potential synergistic effect of combining PDL ESi + Pac in reducing tumor volume (Fig. 7). Western blotting analysis of tumors further validated the results obtained from in-vitro studies.

Fig. 7.

In-vivo anticancer activity of transfected exosomes. (A) Graph representing tumor volumes in MDA-MB-231 DOX RT Xenograft mouse model (p < 0.0001). Tumor growth curves represent the change in tumor volume (mm³) over a period of 15 days in mice treated with different therapeutic treatments (n = 5). PDL ESi + Pac results in the most significant reduction in tumor volume by four folds, indicating enhanced antitumor efficacy when combining immunotherapy and chemotherapy. (B) Representative western blots of MDA-MB-231 DOX RT xenograft tumors. β-actin was used as a loading control. Quantification of protein expression was performed by normalizing the intensity of each target protein band to its corresponding β-actin

Discussion

Cancer immunotherapy, exemplified by PD-1/PD-L1 immune checkpoint inhibitors, has significantly reshaped cancer treatment strategies over the past few decades. The key advantage of these “immunomodulatory drugs” lies in their capacity to activate a patient’s inherent immune defenses against tumor cells. Despite the impressive clinical outcomes associated with PD-1/PD-L1 inhibitors, the emergence of drug resistance remains a major challenge [57]. Mechanistically, such resistance typically involves T cell exhaustion, functional depletion of cytotoxic T lymphocytes, and the upregulation of secondary immune checkpoints [58–60]. To address this issue, novel treatment modalities are needed—particularly combination strategies that integrate immune checkpoint blockade with other potent anticancer agents [61–63].

PD-L1 upregulation has been widely associated with chemoresistance in various cancers [64–66]. Black et al. reported that PD-L1 expression was significantly increased in MDA-MB-231 cells upon doxorubicin exposure, demonstrating that the concurrent use of chemotherapy and immunotherapy was more effective than single-agent therapy [67]. Similarly, PD-L1 overexpression has been linked to cisplatin resistance and PI3K/AKT pathway activation in non-small cell lung cancer, correlating with poor prognosis, shorter survival, and lower response rates [68]. Notably, downregulation of PD-L1 improved cisplatin sensitivity via PI3K/AKT signaling blockade. Further supporting this, treatment of TNBC with auranofin, an FDA-approved thioredoxin reductase inhibitor, induced PD-L1 upregulation, and the combination of auranofin with an anti-PD-L1 antibody exhibited a synergistic anticancer effect [69]. These studies underscore PD-L1 as a key player in chemoresistance and highlight the potential benefits of targeting PD-L1 in combination with chemotherapy to overcome drug resistance in TNBC.

In this work, we explored exosome-based drug delivery to overcome the limitations posed by resistant tumors. We genetically engineered anti-PD-L1 exosomes (PDL E) to display an anti-PD-L1 moiety on their surface, enabling them to bind to and neutralize PD-L1 in tumor cells directly.

Following transfection of HEK293T cells with the pDisplay plasmid encoding the tagged anti-PD-L1 scFv, the recombinant protein is expressed as a membrane-anchored molecule by virtue of the PDGFR transmembrane domain at its C-terminus. This domain directs the scFv to the cell membrane, with the HA and myc epitopes serving as identifiable tags at either terminus. During the biogenesis of exosomes—small vesicles originating from inward budding of the endosomal membrane—the membrane-anchored scFv is incorporated into the exosomal membrane. Consequently, exosomes released from transfected cells bear the anti-PD-L1 scFv on their surface. This configuration preserves the functional antigen-binding site of the engineered scFv, enabling the resultant exosomes to potentially bind PD-L1 targets while displaying the HA and myc tags for detection or further analysis.

By disrupting the PD-1/PD-L1 axis, PDL E shifts the immune milieu toward an active anti-tumor state, allowing cytotoxic T lymphocytes to eliminate malignant cells more effectively. Indeed, PD-L1 expression often serves as an “adaptive mechanism” that helps tumor cells evade immune surveillance [70]. Our data, which demonstrate enhanced apoptosis following PD-L1 inhibition in breast cancer cells, are consistent with prior work indicating that PD-L1 silencing can increase both spontaneous and chemotherapy-induced cell death [71].

Building on PDL E’s capacity to target tumors, we further encapsulated STAT3 siRNA to create PDL ESi. To our knowledge, this is the first study to utilize STAT3 siRNA in PD-L1–targeted exosomes, providing a unique dual-action platform that addresses both immune evasion and intrinsic oncogenic signaling. STAT3 is frequently hyperactivated in human cancers and correlates with poor prognosis [40]. It interacts closely with NF-κB, together amplifying tumor growth, survival, and immune evasion [72, 73]. Accordingly, silencing STAT3 expression within tumor cells can be a key step toward suppressing malignancy. Our findings confirm that PDL ESi significantly diminishes both STAT3 and NF-κB activities, culminating in pronounced apoptosis. These effects are enhanced further by combining PDL ESi with paclitaxel (Pac), a widely used chemotherapeutic drug known to exert direct cytotoxicity and modulate various immune cell subsets, such as natural killer cells, dendritic cells, regulatory T cells, and macrophages [74]. Pac also interferes with tumor cell division by destabilizing microtubule dynamics and blocking mitosis [75]. Hence, the synergy observed with PDL ESi + Pac underscores the therapeutic advantages of merging exosome-based immunotherapy with established chemotherapy.

The PI3/AKT/mTOR signaling pathway, a central regulator of various cellular processes such as cell cycle regulation, apoptosis, growth, and proliferation, is implicated in breast cancer progression [76, 77]. Extensive studies have demonstrated the crosstalk between PD-L1 expression and activation of the PI3/AKT/mTOR pathway in different cancers. For instance, Zhao et al. reported that treatment with an anti-PD-L1 antibody (MIH5, rat IgG2a) in an orthotopic mouse pancreatic cancer model resulted in the downregulation of key markers of the PI3/AKT/mTOR pathway, including AKT, p-AKT, PI3K, mTOR, and p-mTOR, along with an upregulation of PTEN at both mRNA and protein levels [78].

Although the primary mechanism of action of Pac is microtubule disruption, it has also been shown to indirectly inhibit the PI3/AKT/mTOR pathway as well as BCL-2. Li et al. demonstrated that Pac effectively inhibited the proliferation and invasion of MCF-7 breast cancer cells while promoting apoptosis. This was evidenced by increased levels of cleaved caspase-3 and Bax and decreased expression of Bcl-2, MMP-9, and VEGF in a dose-dependent manner. Furthermore, the study noted that upregulation of PI3K reversed Pac’s effects by restoring AKT phosphorylation at both Thr308 and Ser473, highlighting the involvement of the PI3K/AKT signaling pathway [79].

Our findings align with previous reports, as we observed significant downregulation of p-AKT and p-mTOR, along with alterations in apoptotic markers, indicating the induction of apoptosis in MDA-MB-231 DOX RT cells. This suggests that the combination of Pac and PDL E exerts a potent synergistic effect by modulating the PI3/AKT/mTOR pathway and promoting apoptosis.

In cancer, PD-1 interacts with PD-L1, a ligand often overexpressed by tumor and stromal cells in response to inflammatory cytokines such as interferon-gamma (IFN-γ) secreted by cytotoxic T cells. Upon binding to PD-L1, PD-1 transmits an inhibitory signal by recruiting Src homology 2 domain-containing phosphatases (SHP1 and SHP2), which dephosphorylate downstream T-cell receptor (TCR) signaling elements. This suppresses T-cell proliferation, cytokine release, and cytotoxic activity, enabling tumor cells to evade immune surveillance. As a result, the PD-1/PD-L1 axis serves as a negative feedback loop that attenuates T-cell responses and facilitates immune escape within the tumor microenvironment [80]. Blocking the PD-L1/PD-1 axis can effectively restore cytotoxic T-cell activity and enhance anti-cancer immune responses.

Furthermore, the JAK/STAT signaling pathway has been shown to positively correlate with PD-L1 expression. A study by Moon et al. demonstrated a strong link between PD-L1 upregulation and JAK/STAT signaling in gastric cancers. Specifically, PD-L1 expression was significantly enhanced via JAK2/STAT1 signaling following IFN-γ treatment, which promoted immune resistance within the tumor microenvironment. This underscores the JAK/STAT pathway as a crucial regulator of PD-L1-mediated immune evasion and a potential therapeutic target in gastric cancer [81]. Our findings align with these reports, as the downregulation of PD-L1, pSTAT3, and key immunosuppressive cytokines (TGF-β and IL-1β) following treatment with PDL ESi + Pac suggests a reactivation of anti-tumor immunity. This highlights the potential of this combination strategy in overcoming immune evasion and enhancing anti-cancer immune responses.

Our findings demonstrate that PDL E significantly enhances cellular uptake and tumor-specific accumulation, reinforcing their potential as a precision drug delivery system for TNBC. In vitro studies showed that PDL E exhibited a two-fold increase in uptake by MDA-MB-231 DOX RT cells compared to Cy5-DSPE, likely due to the specific interaction between the anti-PD-L1 moiety and PD-L1 receptors on drug-resistant TNBC cells. Similarly, in vivo biodistribution analysis revealed a three-fold higher tumor accumulation of PDL E, suggesting that PD-L1-mediated targeting facilitates tumor retention, potentially enhancing therapeutic efficacy while minimizing off-target effects.

Our study aligns with a growing body of evidence advocating for exosome-mediated cancer therapies. Chen et al., for instance, modified exosomes derived from MDA-MB-231 cells to overexpress a high-affinity variant of human PD-1 (havPD-1), while knocking out PD-L1 and beta-2 microglobulin. These engineered exosomes restricted cancer cell proliferation, triggered apoptosis, and nullified PD-L1–mediated T cell suppression without eliciting antibody-dependent or complement-dependent cytotoxicity. In vivo experiments showed efficacy comparable to that of anti-PD1 monoclonal antibodies in xenograft models carrying human T cells, and loading these exosomes with a PARP inhibitor further enhanced anti-tumor activity by combining chemotherapy and immunotherapy [82]. These findings are consistent with our own results, highlighting the strong potential of exosome-based strategies for both immune modulation and targeted drug delivery. Also, Dai et al. demonstrated a novel exosomal vaccine strategy for enhanced cancer immunotherapy by simultaneously activating T cells and blocking the PD-1/PD-L1 immune checkpoint pathway. They engineered exosomes to express PD-L1 antibodies on their surface, enabling them to effectively bind and inhibit PD-1/PD-L1 interactions. This blockade helps restore T-cell activity against tumor cells. Additionally, the exosomes function as a vaccine by in situ activating T cells, further enhancing the immune response. In preclinical models, these engineered exosomes demonstrated significant antitumor efficacy by promoting T-cell activation and reducing tumor growth. This dual function exosome vaccine offers a promising approach to improving cancer immunotherapy outcomes [83]. Si et al. explored a novel strategy for cancer therapy by co-delivering a programmed death-ligand 1 (PD-L1)-blocking single-chain variable fragment (scFv) and chemotherapeutics using engineered exosomes. Exosomes were modified to encapsulate both scFv antibodies targeting PD-L1 and Paclitaxel, enhancing immune checkpoint blockade while simultaneously inducing cytotoxic effects on tumor cells. The engineered exosomes demonstrated efficient cellular uptake, immune activation, and tumor growth inhibition in preclinical models. This dual-function approach holds promise for overcoming immunosuppression in the tumor microenvironment and enhancing the efficacy of conventional chemotherapy, offering a potential translational strategy for colon cancer [84].

Our work advances these strategies by using a scFv derived directly from a clinically approved anti-PD-L1 antibody (Atezolizumab) and further embedding STAT3 siRNA within the same exosomes (PDL ESi). This dual-targeting mechanism tackles both immune checkpoint blockade and critical oncogenic signaling, culminating in synergistic tumor suppression when combined with conventional chemotherapy. Moreover, our study focuses on doxorubicin-resistant triple-negative breast cancer, underscoring the platform’s potential to overcome drug resistance and reinforcing the versatility of this exosome-based immunotherapeutic strategy.

In summary, we effectively reactivated antitumor immunity while also interfering with vital tumorigenic pathways, thereby increasing tumor cell apoptosis. When combined with Pac, a standard chemotherapeutic agent, these exosomes offer a synergistic approach that addresses both the immune-suppressive and proliferative facets of cancer. Given their capacity for selective cargo loading and favorable biocompatibility, exosomes may find wide applicability in diverse cancer types. Thus, exosome-based PD-L1 silencing—notably via the dual targeting of immune checkpoints and oncogenic signaling molecules like STAT3—presents a promising strategy to enhance and complement existing cancer therapies, ultimately improving patient outcomes.

Conclusion

We developed PD-L1-targeted exosomes that function as both immunomodulatory agents and chemotherapeutic drug delivery vehicles for the treatment of PD-L1-expressing tumors. The surface expression of PD-L1 on the exosomes not only facilitates tumor-specific accumulation, minimizing off-target toxicity, but also disrupts the PD-L1/PD-1 axis, thereby enhancing anti-tumor immune responses. In vitro and in vivo studies demonstrated that PDL E exhibited significant cytotoxicity and effectively downregulated PD-L1 expression in doxorubicin-resistant MDA-MB-231 cells. Moreover, when combined with Pac, PDL ESi induced a pronounced reduction in tumor volume. Mechanistically, these exosomes inhibit the PD-L1/NF-κB signaling pathway, ultimately leading to apoptosis in tumor cells. In conclusion, we have successfully engineered exosomes capable of delivering chemotherapeutic agents with active tumor targeting while also possessing inherent immunomodulatory properties. These exosomes hold potential for the treatment of various PD-L1-overexpressing malignancies, including colon, lung, pancreatic cancers, and melanoma.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s13346-025-01920-x.

Abbreviations

TNBC

Triple Negative Breast Cancer

PDL E

Anti-PD-L1 expressing exosomes/PD-L1 targeted exosomes

PDL ESi

Anti-PD-L1 exosomes loaded with STAT 3 SiRNA

Pac

Paclitaxel

PDL ESi + Pac

Anti-PD-L1 exosomes loaded with STAT 3 SiRNA + Paclitaxel

Cy5 PDL E

Anti-PD-L1 exosomes loaded with Cy5-DSPE

DOX RT

Doxorubicin Resistant

DTX RT

Docetaxel Resistant

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary Information files].