Effective delivery of genome editor to cervical cancer targeting Mcl1 for cancer therapy

Gene editing could modify the targeted genome in a precisely and permanently manner through DNA-binding sequence-specific nucleases [1]. Gene editing allows for precise and permanent modifications to the targeted genome using DNA-binding, sequence-specific nucleases. Among the various tools developed for genome editing, the CRISPR/Cas system has emerged as the most widely adopted and influential technology worldwide [2]. Its ability to target specific genes and make alterations with remarkable precision has revolutionized fields such as genetics, biotechnology, and medicine. In recent years, CRISPR/Cas technology has gained significant traction, particularly with the approval of gene therapy methods based on this system to treat genetic diseases such as sickle cell anemia. The success of these therapies has sparked a renewed interest in the potential of CRISPR/Cas to treat a wide range of genetic disorders, highlighting its transformative potential in the realm of medical science. With approvement of the gene Therapy method based on CRISPR/Cas technology to treat patients with sickle cell disease, the delivery of CRISPR/Cas also drawn huge attention in the gene editing field [3, 4]. However, one of the key challenges that still persists in the field of CRISPR/Cas technology is the development of an ideal delivery system [4,5,6]. Efficient delivery is critical to ensuring that CRISPR components, such as Cas9 and the guide RNA, are successfully introduced into the target cells without causing damage or triggering unwanted immune responses. The ability to deliver these components with high efficiency and specificity remains a major bottleneck in the widespread clinical application of CRISPR-based therapies. Currently, viral vectors, such as adeno-associated viruses (AAVs), are the most commonly used method for delivering CRISPR/Cas9 into cells. These vectors have been approved for clinical trials in some cases and have shown promise in delivering gene editing tools effectively. However, despite their advantages, viral vectors still pose several risks and challenges. For example, there are concerns about potential integrated mutagenesis, where the viral DNA may insert itself into the genome at unintended locations, causing disruptions in the host DNA. Furthermore, off-target editing, where unintended genes are altered, remains a significant risk, leading to potential side effects that could undermine the safety and efficacy of CRISPR-based therapies.

There are increasing studies focus on the delivery systems, including liposomes [7], ligand-receptor mediated complexes [8], lipid nanoparticles (LNPs) [2], AAVs [9], self-assembled nanoparticles [10], magnetic nanoparticles [11], extracellular vesicles (EVs) and polymers [12]. EVs are nanoscale carriers secreted by various types of cells, and typically range in diameter from 30 nm to 150 nm, which play a crucial role in intercellular communication [13, 14]. As the understand of EVs biogenesis, EVs are becoming perfect candidates for loading and delivering cargoes between diverse cells [15] rather than vesicles for wastes. The biogenesis of EVs has garnered significant attention, and our growing understanding of their formation has made them ideal candidates for the delivery of therapeutic cargoes [16, 17]. Artificial EVs can be engineered in vitro and can be loaded with a variety of substances such as nucleic acids (e.g., mRNA, siRNA), proteins [18], lipids, and other bioactive molecules [19]. These engineered vesicles can then facilitate targeted delivery [20, 21], enabling the transfer of these cargos across cell membranes and enhancing intercellular communication [22, 23]. The potential applications of artificial EVs extend across multiple fields, particularly in the areas of gene therapy [24, 25] and cancer treatment [26, 27]. Depending on the type of cargo they carry, artificial EVs can activate specific signaling pathways, thereby modulating the tumor microenvironment [28, 29], suppressing inflammatory responses, and regulating cancer progression [30]. By tailoring the cargo inside EVs, it is possible to target particular pathways or even deliver gene-editing tools such as CRISPR/Cas9, enabling precise manipulation of the genome. With increasing studies to explore EVs scaffold proteins for EVs biogenesis [31, 32], there are more platforms for producing artificial EVs [33, 34] and more related strategies for delivering cargoes including the mRNAs [35, 36], siRNAs [37,38,39], bioactive proteins [40] and genome editing tools [41]. Besides, the photocleavable proteins, which were reported as protein linkers for the drug delivery field, lead to more methods for cargoes loading and releasing in the artificial EVs.

Mcl1 was reported that expressed highly in diverse cancers [42] and could mediate the cancer progression [43]. The MCL1 gene is frequently amplified in cancer and codes for the antiapoptotic protein myeloid cell leukemia 1 (MCL1), which confers cancer progression and drug resistance [44]. It was reported that the Mcl1 might be the therapeutic targeting sites for cervical cancer [42, 45,46,47].

Herein, in this present study, we produced artificial EVs to load and deliver Cas9 RNP through endogenous behaviors in producing cells. The artificial EVs could deliver the Cas9 RNP against Mcl1 gene (encoding anti-apoptotic Mcl1 protein) to inhibit the cervical cancer progression in vitro and in vivo.

The expression level of Mcl1 in cervical cancer

To clarify the vital role of Mcl1 in the progression of cervical cancer, we explored the Mcl1 expression in cervical cancer and the overall survival. As depicted in Fig. 1, the result derived from TCGA DataSet [48, 49] showed that the expression level of Mcl1in cervical cancer is remarkable higher (Fig. 1A). Furthermore, according to the overall survival analysis, the prognosis of patients with lower expression of Mcl1 superior to those with higher expression of Mcl1 (Fig. 1B). Also, we conducted the analysis in https://tnmplot.com/analysis/, which showed the Mcl1 expression in cervical cancer and non-cervical cancer (Supplementary Fig. S1). As the results from the immunohistochemistry assay, the Mcl1 expression in cervical cancer is higher than non-cervical cancer (which means cervix from women underwent total hysterectomy without cervical cancer) (Fig. 1C, D). Meanwhile, the mRNA expression level of Mcl1 in cervical cancer tissues is higher than that in non-cervical cancer tissues (Fig. 1E). What’s more, the Mcl1 expression in cervical cancer cells is higher than in cells from non-cervical cancer cervix (Fig. 1F, G). Above all, the Mcl1 expression in cervical cancer is higher, which results in poor prognosis.

The loading strategy of bioengineered artificial EVs and the characterizations of artificial EVs

We designed artificial bioengineered EVs-based platform to achieve both loading and delivering proteins of interests. The artificial EVs are isolated from cell culture supernatant after the plasmid transfection and changing medium through the ultra-centrifugation. The artificial EVs encapsulated Cas9-sgMcl1 RNPs (EVs^RNP (Cas9-sgMcl1)) exhibited classic cup-shaped structure under the TEM (Fig. 2A). The sizes of artificial EVs^RNP (Cas9-sgMcl1) ranged from 30 nm to 1000 nm (Fig. 2B). What’s more, the artificial EVs^RNP (Cas9-sgMcl1) could stably express CD63, Alix and Tsg101, which are acknowledged as EVs classic markers. Furthermore, the artificial EVs^RNP (Cas9-sgMcl1) could express the Cas9 and Lamp2b significantly. These results indicate that the Cas9 has been successfully loaded into the artificial EVs through the biogenesis (Fig. 2C and Supplementary Fig. S2). In addition, the bioluminescent signaling showed that EVs^nanoLuc could load nanoLuc, which could express bioluminescent signaling when incubated with substrates (Fig. 2D). Collectively, the delivery platform strategy could sort the protein of interest into the artificial EVs.

The uptake process of artificial EVs by recipient cells

To evaluate whether the artificial EVs could enter the recipient cells, we selected the PKH26 dye to stain EVs, then used the PKH26 labeled EVs to treat the HeLa cells. As demonstrated in Fig. 3A, after the treatment for 12 h, the FITC-Actin Tracker labeled Hela cells could uptake the PKH26 labeled EVs, which were captured by the confocal microscopy. Also, from the Z-stack imaging, the EVs could be internalized into the cytoplasm (Supplementary Fig. S3). Furthermore, we constructed the plasmid “Lamp2b-PhoCl-GFP” to load GFP into the EVs^GFP. The artificial EVs^GFP treated HeLa cells rather than PBS treated HeLa cells and blank EVs treated HeLa cells, could exhibit GFP expression (Fig. 3B–D). What’s more, the bioluminescent signaling showed that EVs^nanoLuc treated HeLa cells rather than PBS treated HeLa cells and blank EVs treated HeLa cells, could exhibit nanoLuc bioluminescent signaling (Fig. 3E), which exhibited a dose-dependent behavior (Fig. 3F). Collectively, the artificial EVs could be encapsulated by the recipient cells and deliver the loaded cargoes into the recipient cells.

Artificial EVs^RNP (Cas9-sgMcl1) could inhibit cell proliferation and migration in cervical cancer in vitro

Based on the above results, we used the plasmid “Lamp2b-PhoCl-Cas9” and “U6-sg-Mcl1” to produce EVs^RNP (Cas9-sgMcl1), which loads the Cas9-sgMcl1 RNPs into the artificial nanoparticles. We produced EVs^RNP (Cas9-sgMcl1), and EVs^RNP (Cas9-sgGFP) as control EVs. The artificial EVs^RNP were added into the cervical cancer cells to detect the effect on cell migration ability and cell viability. The migration ability of HeLa cells after the treatment of EVs^RNP (Cas9-sgMcl1) is lower than that treated with EVs^RNP (Cas9-sgGFP) and PBS (Fig. 4A, B). In addition, the viability of Hela cells after the treatment of EVs^RNP (Cas9-sgMcl1) is lower than that treated of EVs^RNP (Cas9-sgGFP) and PBS (Fig. 4C). When evaluated the DNA sequences of HeLa cells after the treatment of EVs^RNP, the EVs^RNP (Cas9-sgMcl1) could lead to significant indels while EVs^RNP (Cas9-sgGFP) and PBS could not lead to indels (Fig. 4D). Additionally, the western blot results indicated that EVs^RNP (Cas9-sgMcl1) could inhibit the protein expression of Mcl1 while EVs^RNP (Cas9-sgGFP) and PBS could not (Fig. 4E, Supplementary Fig. S4). Furthermore, we treated HeLa cells with diverse EVs to measure the apoptosis. The HeLa cells treated with EVs^RNP (Cas9-sgMcl1) exhibited higher apoptosis than HeLa cells treated with EVs^RNP (Cas9-sgGFP) and PBS (Fig. 4F, G). Collectively, the artificial EVs^RNP (Cas9-sgMcl1) could deliver the bioactive Cas9 RNP into both the HeLa cells and the inhibit the viability and migration ability of these cancer cells in vitro.

EVs^RNP (Cas9-sgMcl1) could inhibit cervical cancer xenograft models in vivo

To evaluate the therapeutic efficacy of EVs^RNP (Cas9-sgMcl1) in vivo, we performed the study in immunocompromised athymic nude mice bearing HeLa cells xenografts. The tumor-burden nude mice were grouped into three groups: PBS group, EVs^RNP (Cas9-sgGFP) group, EVs^RNP (Cas9-sgMcl1) group. The EVs^RNP were systemically injected via tail vein every 3 days and for six times totally. As depicted in Fig. 5A, B, Hela tumor–bearing mice treated with PBS and EVs^RNP (Cas9-sgGFP) showed rapid tumor growth, whereas EVs^RNP (Cas9-sgMcl1) treatment showed moderate antitumor activity. The results from the tumor wight (Fig. 5C) are similar to the results from tumor growth, while there is no significant difference among mice body weight (Fig. 5D). The immunohistochemistry assay results indicated that EVs^RNP (Cas9-sgMcl1) could inhibit the protein expression of Mcl1 while EVs^RNP (Cas9-sgGFP) and PBS could not (Fig. 5E). When evaluated the DNA sequences of tumor tissues after the treatment of EVs^RNP, the EVs^RNP (Cas9-sgMcl1) could lead to significant indels while EVs^RNP (Cas9-sgGFP) and PBS could not lead to indels (Fig. 5F). Above all, the artificial EVs could deliver the Cas9-sgMcl1 RNPs functionally into HeLa xenograft models in vivo with certain efficiency.

The targeting strategy of EVs in vivo

To design the EVs for targeting ability, we select the biotin streptavidin system. We engineered the Streptavidin-Lamp2b-GFP EVs, incubated with biotin-anti Her2 antibody, to form the anti Her2 EVs^GFP. We used the anti Her2 EVs^GFP to treat the Her2+ HeLa cells and HeLa cells. The anti Her2 EVs^GFP could increase the uptake of Her2+ HeLa cells (Fig. 6A, C), but not increase the uptake of HeLa cells (Fig. 6B, D). This result showed that it is feasible to design the EVs to confer the targeting ability.

The distribution and safety of EVs^RNP in vivo

To assess the distribution of EVs^RNP in vivo, we produced EVs^RNP-GFP and injected them into immunocompromised athymic nude mice bearing HeLa cell xenografts. As shown in Fig. 7A, EVs^RNP-GFP were detected in the liver, kidney, PBMCs, bone marrow cells, and the tumor. Approximately 50% of cells isolated from the liver showed GFP positivity, while about 40% of cells from tumor tissues exhibited GFP fluorescence. Notably, no indels were detected in liver cells, kidney cells, PBMCs and bone marrow cells, while approximately 10% of tumor cells showed indel formation. These findings suggest that EVs^RNP-GFP can mediate genome editing in tumor cells, but not in normal organ cells (Fig. 7B). Additionally, flow cytometry was employed to measure the Ki-67 expression in un-activated T cells, pre-activated T cells (with IL-2 + anti-CD3/CD28), and HeLa cells. The results revealed that HeLa cells exhibited a higher level of Ki-67 (Fig. 7C, D) without resulting in indels (Fig. 7E). This outcome could be attributed to the rapid replication of tumor cells, which may facilitate genome editing by Cas9 RNP more efficiently than in normal, slower-replicating cells.

Increasing evidence have witnessed the critical role of CRISPR/Cas9 as a promising method for therapy in diverse diseases. However, the application was limited due to lacking effective delivery platform. Previous studies showed that endogenous EVs could deliver Cas9/gRNA for gene-editing [50,51,52], but the efficiency of this delivery system still needs further improvement. Therefore, we selected the artificial EVs-based system as the delivery platform. In the present study, the artificial EVs-based system could load the Cas9 RNP into the EVs through the plasmids construction and transfection. Subsequently, the artificial EVs could deliver the Cas9 RNP into cervical cancer for potential therapy.

For verify the delivery system, a series of experiments were conducted to evaluate the loading ability and delivery ability of the artificial EVs. Firstly, in the biogenesis prat for loading cargoes, we constructed the plasmids for transfection. And the results from TEM, NTA and western blot showed that the Cas9 RNP were loaded into the artificial EVs, which is consistent to other studies [18, 50]. Then, we used the artificial EVs to load GFP protein and nanoLuc protein. The results showed that the artificial EVs could delivery GFP proteins, which was widely used to detect the genome editing approaches [53,54,55]. In addition, when we injected the EVs^RNP (Cas9-sgMcl1) into the cervical cancer bearing mice, EVs^RNP (Cas9-sgMcl1) treatment could inhibit the tumor growth through suppressing the Mcl1 expression level in the tumor tissues. Thus, the EVs^RNP (Cas9-sgMcl1) could exhibit greater anti-tumor effect in cervical cancer model in vivo. These results are consistence to some other reported studies [33, 41].

The targeting specificity of our EV-based CRISPR/Cas9 delivery system is influenced by several factors, with DNA replication frequency and chromosomal uncoiling being key determinants of editing efficiency. In normal, non-cancerous cells, DNA replication is typically less frequent compared to cancer cells. As a result, the Cas9 ribonucleoprotein (RNP) system faces limitations in efficiently editing these cells due to the lower occurrence of DNA replication and less active chromosomal uncoiling, both of which are essential for CRISPR-mediated genome editing.

On the other hand, tumor cells, such as cervical cancer cells, generally exhibit higher rates of DNA replication and chromosomal dynamics, which significantly enhance the efficiency of Cas9-mediated gene editing. This makes cancer cells more responsive to CRISPR/Cas9 editing, as the machinery is better able to access and modify the targeted DNA.

The Cas9-induced knockout efficiency depends on the cell proliferation status and might be associated to cell cycle phase. Proliferating cells are typically in active cell cycle phases (S, G2, M), where they are more likely to repair double-strand breaks. These repair mechanisms are more active during these phases, leading to higher chances of introducing mutations via non-homologous end joining (NHEJ). Non-proliferating cells (e.g., cells in the G0 phase or those undergoing senescence) may have a reduced capacity for DNA repair, particularly NHEJ, as the cell cycle machinery needed for efficient repair and proliferation may be less active. As a result, knockout efficiency might be lower in these cells, as the repair of Cas9-induced DNA breaks is impaired.

Proliferating cells generally show higher Cas9 RNP knockout efficiency because of active DNA repair processes and more dynamic cell cycle phases. Non-proliferating cells may exhibit lower knockout efficiency due to reduced DNA repair activity and a less favorable cell cycle state for Cas9-induced mutagenesis.

While surface modifications of EVs could potentially increase the binding and internalization efficiency of delivery particles into cancer cells, it is important to note that the success of the CRISPR/Cas9 system is not solely dependent on targeting but also on the intrinsic characteristics of the cells. Enhanced binding through surface modifications may improve delivery efficiency, but the true efficacy of Cas9 editing is more closely tied to the cell’s DNA replication and chromosomal conformation. Therefore, while EV targeting could play a supportive role in increasing the overall efficiency of delivery, the CRISPR editing efficiency remains fundamentally dependent on the tumor cell’s replicative state and chromosomal dynamics.

There are several limitations in the present study. Firstly, this study lacks related studies of different cargoes of EVs (such as chemotherapeutic drugs and mRNAs [56]), which might improve the efficiencies of cancer therapies. Secondly, the surface decoration of EVs (including click chemistry) for targeting delivery need more investigation in the future, which might enhance the delivery efficiency and safety [57, 58]. While we agree that off-target activity is an important consideration, our study primarily focuses on the efficacy and distribution of EVsRNP-mediated genome editing. Although we did not directly assess off-target effects in this work, we ensured a high degree of specificity through the careful selection of guide RNAs and optimized delivery systems. We also observed that genome editing was predominantly confined to the target tumor cells, with minimal indels detected in non-target tissues. Nevertheless, we acknowledge that further studies using methods like GUIDE-seq or deep sequencing could provide a more comprehensive analysis of potential off-target effects. We plan to explore these techniques in future work to enhance the safety profile of this platform.

Above all, our findings indicate that the bioengineered artificial EVs not only could load endogenous bioactive protein (including GFP protein, Cre recombinase and Cas9 RNP) into the EVs, but also could release the cargoes into the recipient cells in vitro and in vivo. The Cas9-sgMcl1 RNP could be delivered into cervical cancer cells and inhibit the viability of the cancer cell in vitro and in vivo. These results not only provide a practical strategy for Cas9 RNP delivery but also open a promising avenue for the bioactive protein therapy of cancers.

The EVs platform offers several advantages in gene delivery, particularly in terms of efficacy. EVs, which include exosomes and microvesicles, are naturally occurring biological carriers that have evolved to transfer molecular cargo between cells, providing them with inherent delivery capabilities. This makes EVs ideal for targeted delivery, especially in cancer therapies where the ability to direct payloads to specific cell types is crucial. EVs can encapsulate CRISPR/Cas9 components, RNA, and proteins, enhancing the precision of gene editing. Additionally, EVs possess natural surface proteins and lipids that allow for selective targeting and cellular uptake, which can be further enhanced through surface engineering for specific therapeutic applications. However, the efficacy of EVs in gene delivery may be less than that of viral vectors, especially when delivering larger genetic constructs or achieving high transfection efficiencies. In terms of safety, EVs provide distinct advantages over other delivery platforms. Derived from cells, EVs have a lower immunogenic profile and are less likely to provoke an immune response, making them a safer choice for repeated administration in clinical settings. Unlike viral vectors, EVs do not integrate into the host genome, significantly reducing the risk of insertional mutagenesis, which could lead to oncogenesis or other genetic disorders. This non-integrative nature makes EVs particularly suitable for gene therapies targeting sensitive tissues. However, while the safety profile of EVs is generally favorable, there is still a need for further research to fully understand the long-term effects and potential risks associated with their clinical use, such as their biodistribution and off-target effects. When considering scalability, EVs face challenges that hinder their widespread application. The isolation of EVs from biological fluids or cell cultures can be technically demanding and often yields relatively low quantities, which may not be sufficient for large-scale therapeutic production. The purification process can also lead to variability in EV quality and composition, which is a barrier to consistent and reproducible manufacturing. Despite these challenges, advances in EV isolation technologies and cell culture techniques are gradually improving the yield and purity of EVs, making them more feasible for clinical use. Moreover, EVs can be sourced from patient-derived cells, providing the added benefit of personalized medicine. In contrast, viral vectors and lipid nanoparticles (LNPs) offer established scalability, with viral vectors benefiting from decades of optimization in production processes. LNPs also present a more straightforward, scalable manufacturing process, particularly for RNA-based therapies, as evidenced by the mRNA COVID-19 vaccines. While EVs may lag in terms of scalable production, their safety, biocompatibility, and ability to be engineered for targeted delivery give them strong potential for specific therapeutic applications. In conclusion, the EV platform offers a promising, safe, and potentially effective delivery system for gene therapies, with significant advantages in biocompatibility and targeting specificity. However, challenges related to delivery efficiency and scalability need to be addressed before EV-based systems can achieve their full clinical potential. Ongoing research to optimize EV production, enhance cargo loading, and improve delivery efficiency will be critical in unlocking the full therapeutic potential of EVs for CRISPR/Cas9 and other genetic interventions.

Cell lines and animals

HeLa and HEK-293T cells were purchased from Chinese Academy of Sciences. HeLa^Cre-Loxp cells were awarded from the Obstetrics and Gynecology Hospital of Fudan University. These cells were supplied by DMEM containing 10% fetal bovine serum and 1% penicillin-streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂.

Female nude mice (six weeks) were fed in the Laboratory in Animal house (the Obstetrics and Gynecology Hospital of Fudan University). All the treatments were approved by the Laboratory Animal Welfare and Ethics Committee of the Obstetrics and Gynecology Hospital of Fudan University.

Producing, isolation and characterization of EVs^RNP

To load the bioactive proteins into the EVs, plasmids were constructed including (1) the transmembrane domain “Lamp2b”, to load the fused proteins, (2) photocleavable protein domain “PhoCl”, to help release the free protein, and (3) cargo protein “cargo” (including GFP, Cre or Cas9) for cell transfection and further producing EVs. The “Lamp2b” could act as transmembrane protein in EVs, which results in the followed “PhoCl-cargo” part could be sorted into the EVs. The light-cleavable “PhoCl” could be cleaved after the 405 nm violet light treatment, which results in releasing the “cargo” freely into the artificial EVs. Above all, the constructed plasmids could load the protein of interest into the EVs during the EVs biogenesis, then release the protein of interest after the 405 nm violet light treatment.

Firstly, the FBS were ultracentrifuged to exclude the FBS-EVs, then the EVs-free FBS were used for DMEM containing 10% EVs-free FBS. Briefly, the transfected HEK-293T cells were cultured in DMEM (containing 10% EVs-free FBS) for 48 h. Then the culture medium were collected to isolate the EVs according to Thery’s protocol as below [59]: (1) 300 g for 20 min; (2) 10,000 g for 30 min; and (3) 12,500 g for 75 min. The pellet was suspended and was filtered through a 0.22μm filter. The morphology of EVs^RNP were observed by transmission electron microscopy (FEI Tecnai G2 Spirit Twin, Philips, NL). The sizes of EVs^RNP were measured by nanoparticle tracking analysis through NanoSight NS300 (Malvern, Amesbury, GB). And the markers and cargoes of EVs^RNP including CD63, Alix, Tsg101, Lamp2b and Cas9 were detected by Western blot.

Transwell assay

To investigate the effect of EVs^RNP (Cas9-Mcl1) on the migratory capacity of HeLa cells, a Transwell assay was performed. HeLa cells were seeded in the upper chamber of the Transwell system at a density of approximately 1 × 10^5 cells per well. After 12 h, cells were treated with PBS (control), EVs^RNP (Cas9-GFP), or EVs^RNP (Cas9-Mcl1) at concentration of 4*10⁶ particles/well in 24-well-plate for 72 h. Following treatment, the non-migrated cells in the upper chamber were carefully removed using a cotton swab. The migrated cells on the lower surface of the membrane were fixed with 4% paraformaldehyde and stained with crystal violet. The number of migrated cells was counted under a light microscope in five random fields of view. The results were expressed as the average number of migrated cells per field.

Plasmids design and synthesis

Codon-optimized DNA sequences coding for the protein including “Lamp2b”, “PhoCl-Cre”, “Cas9” were ordered from IDT (Integrated DNA Technologies, USA). The constructs used in this study were then generated from the ordered fragments through restriction enzyme digestion and subsequent self-ligation. Plasmids are available from the corresponding author upon request.

Flow cytometry

Totally 5 × 10⁵ HeLa cells were seeded for EVs^RNP treatment for 48 h, then be collected for flow cytometry analysis. We used FITC-Annexin V and DAPI to stain the cells for apoptosis analysis. Data from at least 5 × 10³ cells were acquired using CytoFLEX and the FlowJo software.

For analyzing the delivery efficiency of artificial EVs, the EVs^GFP treated HeLa cells were measured by flow cytometry to calculate the percentage of GFP positive cells. Similarly, the artificial EVs^Cre treated HeLa^Cre-Loxp cells were measured by flow cytometry to calculate the percentage of GFP positive cells.

Establishment of cervical cancer mouse model and therapy

In the mice model, nude mice were divided as 3 groups (N = 4) with a blinded method, each mouse was injected HeLa cells (a total of 2 × 10⁵ cells in 80 µL per mouse) subcutaneously to create bearing-tumor mice model. Diverse EVs (1 × 10¹¹ particles per mice) were injected intravenously per 3 days in totally 21 days. Then, the tumor sizes and weights were measured, followed by the Western blot and IHC staining.

Western blotting

Cell lysates were treated by using RIPA buffer to extract the proteins. The proteins were separated by SDS–polyacrylamide gel according to the standard procedure. The polyvinylidene difluoride membrane was incubated with 5% BSA for 40 min and then incubated with the primary antibody at 4 °C overnight. After washing with 1 × TBST, the membrane was incubated with HRP conjugated secondary antibodies. The membranes were visualized with enhanced chemiluminescence. EVs were detected by CD63 (ab134045), Alix (ab275377), Tsg101 (ab133586), Lamp2b (ab18529) and Cas9 (ab189380).

Immunofluorescence assay

For tumor tissues, the tissues were embedded in paraffin and cut into 5μm sections. The slides were deparaffinized and rehydrated by xylene and different concentration of ethanol (100%, 95%, 80%, 70% and 50%). After the antigen repairment and blocking, anti-MCL1 (1:200, ab32087) antibody and secondary antibody were added at room temperature. Slides were examined under a confocal immunofluorescence microscope (Zeiss).

Sanger sequencing

The Sanger sequencing was used to evaluate the editing efficiency of target genomic loci. The cells or tissues were collected to extract the DNA through the FastPure Cell/Tissue DNA Isolation Mini Kit (Vazyme Biotech). The target genomic locus was amplified by PCR using FastPure Gel DNA Extraction Mini Kit (Vazyme Biotech). PCR products of the genomic region–flanking target sites of sgMcl1 were picked for Sanger sequencing.

PBMCs isolation

To isolate PBMCs, begin by collecting whole blood into an anticoagulant-containing tube, typically EDTA or citrate. Dilute the blood with PBS (1:1 ratio) to reduce viscosity and prevent clumping. Carefully layer the diluted blood onto a density gradient medium such as Ficoll-Paque in a centrifuge tube. Centrifuge the sample at 400–500 x g for 30 min at room temperature to separate the components based on density. The PBMCs will form a “buffy coat” layer just above the gradient medium, which can be carefully aspirated without disturbing the lower layers. Transfer the PBMCs to a new tube and wash them with PBS by centrifuging at 300 x g for 10–15 min to remove residual plasma and gradient media. After washing, resuspend the PBMCs in an appropriate buffer or medium for further experiment.

Statistical analysis

All data are presented as means ± standard deviation (SD). For statistical comparisons between two groups, a two-tailed unpaired t-test was performed, assuming equal variance. To assess differences among more than two groups, one-way analysis of variance (ANOVA) was employed, followed by a post-hoc Tukey’s test for pairwise comparisons when appropriate. Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. All statistical analyses were conducted using Prism 9.2.0 software (GraphPad, San Diego, CA, USA). Sample sizes for each experiment were determined based on power calculations to ensure adequate statistical power and precision in the results. Detailed information regarding the number of biological replicates for each experimental condition is provided in the figure legends.

Study approval

This study was conducted according to the approval from the Ethical Committee of Obstetrics and Gynecology Hospital of Fudan University, the approve number is 2019-77, with informed consent obtained from all patients. All animal procedures were performed in accordance with the protocol approved by the Institutional Animal Care and Use Committee at the Fudan University (SYXK2020-0032).

1. Pacesa M, Pelea O, Jinek M. Past, present, and future of CRISPR genome editing technologies. Cell. 2024;187:1076–100.

   Article  CAS  PubMed  Google Scholar 

2. Haley RM, Padilla MS, El-Mayta RD, Joseph RA, Weber JA, Figueroa-Espada CG, et al. Lipid Nanoparticles for In Vivo Lung Delivery of CRISPR-Cas9 Ribonucleoproteins Allow Gene Editing of Clinical Targets. ACS Nano. 2025;19:13790–804.

   Article  CAS  PubMed  Google Scholar 

3. Burbano DA, Kiattisewee C, Karanjia AV, Cardiff RAL, Faulkner ID, Sugianto W, et al. CRISPR Tools for Engineering Prokaryotic Systems: Recent Advances and New Applications. Annu Rev Chem Biomol Eng. 2024;15:389–430.

   Article  CAS  PubMed  Google Scholar 

4. Banskota S, Raguram A, Suh S, Du SW, Davis JR, Choi EH, et al. Engineered virus-like particles for efficient in vivo delivery of therapeutic proteins. Cell. 2022;185:250.e216.

   Article  Google Scholar 

5. de Jong, Murphy DE OG, Mäger I, Willms E, Garcia-Guerra A, Gitz-Francois JJ, et al. A CRISPR-Cas9-based reporter system for single-cell detection of extracellular vesicle-mediated functional transfer of RNA. Nat Commun. 2020;11:1113.

   Article  PubMed  PubMed Central  Google Scholar 

6. Geilenkeuser J, Armbrust N, Steinmaßl E, Du SW, Schmidt S, Binder EMH, et al. Engineered nucleocytosolic vehicles for loading of programmable editors. Cell. 2025;188:2637–2655.e31.

   Article  CAS  PubMed  Google Scholar 

7. Yuan Z, Gottsacker C, He X, Waterkotte T, Park YC. Repetitive drug delivery using Light-Activated liposomes for potential antimicrobial therapies. Adv Drug Deliv Rev. 2022;187:114395.

   Article  CAS  PubMed  PubMed Central  Google Scholar 

8. Wang QX, Li ZL, Gong YC, Xiong XY. The effects of ligand distribution and density on the targeting properties of dual-targeting folate/biotin Pluronic F127/Poly (lactic acid) polymersomes. Eur J Pharm Biopharm. 2025;206:114598.

   Article  CAS  PubMed  Google Scholar 

9. Ibraheim R, Tai PWL, Mir A, Javeed N, Wang J, Rodríguez TC, et al. Self-inactivating, all-in-one AAV vectors for precision Cas9 genome editing via homology-directed repair in vivo. Nat Commun. 2021;12:6267.

   Article  CAS  PubMed  PubMed Central  Google Scholar 

10. Krishnan N, Fang RH, Zhang L. Engineering of stimuli-responsive self-assembled biomimetic nanoparticles. Adv Drug Deliv Rev. 2021;179:114006.

    Article  CAS  PubMed  Google Scholar 

11. Wu D, Chang X, Tian J, Kang L, Wu Y, Liu J, et al. Bone mesenchymal stem cells stimulation by magnetic nanoparticles and a static magnetic field: release of exosomal miR-1260a improves osteogenesis and angiogenesis. J Nanobiotechnol. 2021;19:209.

    Article  CAS  Google Scholar 

12. Xin J, Lu X, Cao J, Wu W, Liu Q, Wang D, et al. Fluorinated Organic Polymers for Cancer Drug Delivery. Adv Mater. 2024;36:e2404645

    Article  PubMed  Google Scholar 

13. Kenific CM, Zhang H, Lyden D. An exosome pathway without an ESCRT. Cell Res. 2021;31:105–6.

    Article  PubMed  Google Scholar 

14. Wang G, Li J, Bojmar L, Chen H, Li Z, Tobias GC, et al. Tumour extracellular vesicles and particles induce liver metabolic dysfunction. Nature. 2023;618:374–82.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

15. Dar GH, Mendes CC, Kuan WL, Speciale AA, Conceição M, Görgens A, et al. GAPDH controls extracellular vesicle biogenesis and enhances the therapeutic potential of EV mediated siRNA delivery to the brain. Nat Commun. 2021;12:6666.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

16. Hamilton JR, Chen E, Perez BS, Sandoval Espinoza CR, Kang MH, Trinidad M, et al. In vivo human T cell engineering with enveloped delivery vehicles. Nat Biotechnol. 2024;42:1684–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

17. Raguram A, An M, Chen PZ, Liu DR. Directed evolution of engineered virus-like particles with improved production and transduction efficiencies. Nat Biotechnol. 2024;43:1635–47.

    Article  PubMed  PubMed Central  Google Scholar 

18. Yao X, Lyu P, Yoo K, Yadav MK, Singh R, Atala A, et al. Engineered extracellular vesicles as versatile ribonucleoprotein delivery vehicles for efficient and safe CRISPR genome editing. J Extracell Vesicles. 2021;10:e12076.

    Article  PubMed  PubMed Central  Google Scholar 

19. Ling S, Zhang X, Dai Y, Jiang Z, Zhou X, Lu S, et al. Customizable virus-like particles deliver CRISPR-Cas9 ribonucleoprotein for effective ocular neovascular and Huntingtonas disease gene therapy. Nat Nanotechnol. 2025;20:543–553.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

20. Jiang Y, Yu M, Song ZF, Wei ZY, Huang J, Qian HY. Targeted Delivery of Mesenchymal Stem Cell-Derived Bioinspired Exosome-Mimetic Nanovesicles with Platelet Membrane Fusion for Atherosclerotic Treatment. Int J Nanomed. 2024;19:2553–71.

    Article  CAS  Google Scholar 

21. Sahoo S, Kariya T, Ishikawa K. Targeted delivery of therapeutic agents to the heart. Nat Rev Cardiol. 2021;18:389–99.

    Article  PubMed  PubMed Central  Google Scholar 

22. Liu C, Wang Y, Li L, He D, Chi J, Li Q, et al. Engineered extracellular vesicles and their mimetics for cancer immunotherapy. J Control Release. 2022;349:679–98.

    Article  CAS  PubMed  Google Scholar 

23. Zhou G, Gu Y, Zhu Z, Zhang H, Liu W, Xu B, et al. Exosome Mediated Cytosolic Cisplatin Delivery Through Clathrin-Independent Endocytosis and Enhanced Anti-cancer Effect via Avoiding Endosome Trapping in Cisplatin-Resistant Ovarian Cancer. Front Med. 2022;9:810761.

    Article  Google Scholar 

24. Lin Y, Wu J, Gu W, Huang Y, Tong Z, Huang L, et al. Exosome-Liposome Hybrid Nanoparticles Deliver CRISPR/Cas9 System in MSCs. Adv Sci. 2018;5:1700611

    Article  Google Scholar 

25. Chen M, Lu Y, Liu Y, Liu Q, Deng S, Liu Y, et al. Injectable Microgels with Hybrid Exosomes of Chondrocyte-Targeted FGF18 Gene-Editing and Self-Renewable Lubrication for Osteoarthritis Therapy. Adv Mater. 2024;36:e2312559.

    Article  PubMed  Google Scholar 

26. Wang F, Yin L, Hu Y. Progress of extracellular vesicles-based system for tumor therapy. J Control Release. 2025;381:113570.

    Article  CAS  PubMed  Google Scholar 

27. Lin YY, Chen CY, Ma DL, Leung CH, Chang CY, Wang HD. Cell-derived artificial nanovesicle as a drug delivery system for malignant melanoma treatment. Biomed Pharmacother. 2022;147:112586.

    Article  CAS  PubMed  Google Scholar 

28. Peinado H, Zhang H, Matei IR, Costa-Silva B, Hoshino A, Rodrigues G, et al. Pre-metastatic niches: organ-specific homes for metastases. Nat Rev Cancer. 2017;17:302–17.

    Article  CAS  PubMed  Google Scholar 

29. Liu C, Lu J, Tian H, Du W, Zhao L, Feng J, et al. Increased expression of PD‑L1 by the human papillomavirus 16 E7 oncoprotein inhibits anticancer immunity. Mol Med Rep. 2017;15:1063–70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

30. Hoshino A, Costa-Silva B, Shen TL, Rodrigues G, Hashimoto A, Tesic Mark M, et al. Tumour exosome integrins determine organotropic metastasis. Nature. 2015;527:329–35.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

31. Larios J, Mercier V, Roux A, Gruenberg J. ALIX- and ESCRT-III-dependent sorting of tetraspanins to exosomes. J Cell Biol. 2022;219:e201904113.

    Google Scholar 

32. Salunkhe S, Dheeraj, Basak M, Chitkara D, Mittal A. Surface functionalization of exosomes for target-specific delivery and in vivo imaging & tracking: Strategies and significance. J Control Release. 2020;326:599–614.

    Article  CAS  PubMed  Google Scholar 

33. Han J, Sul JH, Lee J, Kim E, Kim HK, Chae M, et al. Engineered exosomes with a photoinducible protein delivery system enable CRISPR-Cas-based epigenome editing in Alzheimer’s disease. Sci Transl Med. 2024;16:eadi4830.

    Article  CAS  PubMed  Google Scholar 

34. Choi J, Chen W, Liao H, Li X, Shendure J. A molecular proximity sensor based on an engineered, dual-component guide RNA. Elife 2025,13:RP98110.

35. Costa Verdera H, Gitz-Francois JJ, Schiffelers RM, Vader P. Cellular uptake of extracellular vesicles is mediated by clathrin-independent endocytosis and macropinocytosis. J Control Release. 2017;28:100–8.

    Article  Google Scholar 

36. Bevers S, Kooijmans SAA, Van de Velde E, Evers MJW, Seghers S, Gitz-Francois J, et al. mRNA-LNP vaccines tuned for systemic immunization induce strong antitumor immunity by engaging splenic immune cells. Mol Ther. 2022;30:3078–94.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

37. Kooijmans SAA, Stremersch S, Braeckmans K, de Smedt SC, Hendrix A, Wood MJA, et al. Electroporation-induced siRNA precipitation obscures the efficiency of siRNA loading into extracellular vesicles. J Control Release. 2013;172:229–38.

    Article  CAS  PubMed  Google Scholar 

38. Wahlgren J, Statello L, Skogberg G, Telemo E, Valadi H. Delivery of Small Interfering RNAs to Cells via Exosomes. Methods Mol Biol. 2016;1364:105–25.

    Article  CAS  PubMed  Google Scholar 

39. Wahlgren J, De LKT, Brisslert M, Vaziri Sani F, Telemo E, Sunnerhagen P, et al. Plasma exosomes can deliver exogenous short interfering RNA to monocytes and lymphocytes. Nucleic Acids Res. 2012;40:e130

    Article  CAS  PubMed  PubMed Central  Google Scholar 

40. Shirmast P, Shahri MA, Brent A, Idris A, McMillan NAJ. Delivering therapeutic RNA into the brain using extracellular vesicles. Mol Ther Nucleic Acids. 2024;35:102373.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

41. Osteikoetxea X, Silva A, Lázaro-Ibáñez E, Salmond N, Shatnyeva O, Stein J, et al. Engineered Cas9 extracellular vesicles as a novel gene editing tool. J Extracell Vesicles. 2022;11:e12225.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

42. Sun W, Cai H, Zhang K, Cui H, Zhao E. Targeting MCL1 with Sanggenon C overcomes MCL1-driven adaptive chemoresistance via dysregulation of autophagy and endoplasmic reticulum stress in cervical cancer. Phytomedicine. 2024;133:155935.

    Article  CAS  PubMed  Google Scholar 

43. Terasaki T, Semba Y, Sasaki K, Imanaga H, Setoguchi K, Yamauchi T, et al. The RNA helicases DDX19A/B modulate selinexor sensitivity by regulating MCL1 mRNA nuclear export in leukemia cells. Leukemia. 2024;38:1918–28.

    Article  CAS  PubMed  Google Scholar 

44. Rauh U, Wei G, Serrano-Wu M, Kosmidis G, Kaulfuss S, Siegel F, et al. BRD-810 is a highly selective MCL1 inhibitor with optimized in vivo clearance and robust efficacy in solid and hematological tumor models. Nat Cancer. 2024;5:1479–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

45. Wang L, Xi C, Liu R, Ye T, Xiang N, Deng J, et al. Dual targeting of Mcl-1 and Bcl-2 to overcome chemoresistance in cervical and colon cancer. Anticancer Drugs. 2024;35:219–26.

    Article  CAS  PubMed  Google Scholar 

46. Zhao Y, Lan Y, Chi Y, Yang B, Ren C. Downregulation of Circ-CEP128 Enhances the Paclitaxel Sensitivity of Cervical Cancer Through Regulating miR-432-5p/MCL1. Biochem Genet. 2022;60:2346–63.

    Article  CAS  PubMed  Google Scholar 

47. Li Y, Li C, Yan J, Liao Y, Qin C, Wang L, et al. Polymeric micellar nanoparticles for effective CRISPR/Cas9 genome editing in cancer. Biomaterials. 2024;309:122573.

    Article  CAS  PubMed  Google Scholar 

48. Ghandi M, Huang FW, Jané-Valbuena J, Kryukov GV, Lo CC, McDonald ER 3rd, et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019;569:503–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

49. Cancer Cell Line Encyclopedia Consortium; Genomics of Drug Sensitivity in Cancer Consortium. Pharmacogenomic agreement between two cancer cell line data sets. Nature 2015,528:84–87.

50. Wan T, Zhong J, Pan Q, Zhou T, Ping Y, Liu X. Exosome-mediated delivery of Cas9 ribonucleoprotein complexes for tissue-specific gene therapy of liver diseases. Sci Adv. 2022;8:eabp9435.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

51. Zeng W, Zheng L, Li Y, Yang J, Mao T, Zhang J, et al. Engineered extracellular vesicles for delivering functional Cas9/gRNA to eliminate hepatitis B virus cccDNA and integration. Emerg Microbes Infect. 2024;13:2284286.

    Article  PubMed  Google Scholar 

52. Gee P, Lung MSY, Okuzaki Y, Sasakawa N, Iguchi T, Makita Y, et al. Extracellular nanovesicles for packaging of CRISPR-Cas9 protein and sgRNA to induce therapeutic exon skipping. Nat Commun. 2020;11:1334.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

53. Zomer A, Maynard C, Verweij FJ, Kamermans A, Schäfer R, Beerling E, et al. In Vivo imaging reveals extracellular vesicle-mediated phenocopying of metastatic behavior. Cell. 2015;161:1046–57.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

54. Bittel M, Reichert P, Sarfati I, Dressel A, Leikam S, Uderhardt S, et al. Visualizing transfer of microbial biomolecules by outer membrane vesicles in microbe-host-communication in vivo. J Extracell Vesicles. 2021;10:e12159.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

55. Zomer A, Steenbeek SC, Maynard C, van Rheenen J. Studying extracellular vesicle transfer by a Cre-loxP method. Nat Protoc. 2016;11:87–101.

    Article  CAS  PubMed  Google Scholar 

56. Luo X, McAndrews KM, Arian KA, Morse SJ, Boeker V, Kumbhar SV, et al. Development of an engineered extracellular vesicles-based vaccine platform for combined delivery of mRNA and protein to induce functional immunity. J Control Release. 2024;374:550–62.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

57. Strebinger D, Frangieh CJ, Friedrich MJ, Faure G, Macrae RK, Zhang F, et al. Cell type-specific delivery by modular envelope design. Nat Commun. 2023;14:5141

    Article  CAS  PubMed  PubMed Central  Google Scholar 

58. Stranford DM, Simons LM, Berman KE, Cheng L, DiBiase BN, Hung ME, et al. Genetically encoding multiple functionalities into extracellular vesicles for the targeted delivery of biologics to T cells. Nat Biomed Eng. 2024;8:397–414.

    Article  CAS  PubMed  Google Scholar 

59. Théry CA-O, Witwer KA-O, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina R, et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles. 2018;7:1535750.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Author notes

1. These authors contributed equally: Guannan Zhou, Xiaoyan Ying, Haiyan Zhang.

Authors and Affiliations

1. Department of Gynecology, Sir Run Run Hospital, Nanjing Medical University, Long Mian Avenue 109 Jiangning, Nanjing, Jiangsu, China

   Yue Wan, Xun Chen, Shengwu Wang, Guannan Zhou & Haiyan Zhang

2. Department of Obstetrics and Gynecology, Second Affiliated Hospital of Nanjing Medical University, No. 262 North Zhongshan Road, Nanjing, 210000, Jiangsu Province, China

   Yiming He & Xiaoyan Ying

3. Department of Gynecology, The Obstetrics and Gynecology Hospital of Fudan University, 419 Fang-Xie Road, Shanghai, 200011, China

   Guannan Zhou

Contributions

Yue Wan: extract raw data, draft manuscript; Yiming He: extract raw data; Xun Chen: extract raw data; Shengwu Wang: extract raw data; Guannan Zhou: review manuscript; Xiaoyan Ying: read and edit manuscript, read, and edit manuscript; Haiyan Zhang: review medical cord and data extraction, read, and edit manuscript.

Corresponding authors

Correspondence to Guannan Zhou, Xiaoyan Ying or Haiyan Zhang.

Competing interests

The authors declare no competing interests.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions