h3B7A-LDM, a novel anti-mesothelin antibody-drug conjugate with the potential to induce antitumor immunity, shows potent efficacy against solid tumors

Abstract

Antibody-drug conjugates (ADCs) are a promising class of cancer therapeutics that enable the targeted delivery of highly cytotoxic payloads to cancer cells. Mesothelin (MSLN) is an attractive therapeutic target in cancer treatment. Lidamycin (LDM), an enediyne-containing antibiotic with potent antitumor effects, has potential as ADC payload. To generate an ADC targeting MSLN, we first produced a novel anti-MSLN antibody, 3B7A, using hybridoma technology. We then obtained the humanized version, h3B7A, via complementarity-determining region (CDR) grafting. This was followed by fusion with LDM through genetic recombination and molecular assembly to create the ADC h3B7A-LDM. h3B7A-LDM undergoes efficient internalization and lysosomal trafficking in MSLN-positive cancer cells. It demonstrates strong tumor-targeting capability and long-term persistence in tumor-bearing mice. In vitro, it exhibits potent antitumor effects, suppressing the proliferation and migration of cancer cells with sub-nanomolar half maximal inhibitory concentration (IC₅₀) values. Mechanistically, h3B7A-LDM induces cell cycle arrest and apoptosis, triggers immunogenic cell death (ICD), and may elicit antitumor immunity. In vivo, h3B7A-LDM significantly inhibits tumor growth in multiple cancer xenograft models. Together, these findings support h3B7A-LDM as a promising drug candidate for treating MSLN-positive cancer.

Graphical abstract

Highlights

• •A novel antibody-drug conjugate, h3B7A-LDM, was constructed by conjugating the anti-MSLN antibody h3B7A with the potent cytotoxin lidamycin.
• •h3B7A-LDM shows potent efficacy against MSLN-overexpressing solid tumors.
• •h3B7A-LDM inhibits tumor via anti-proliferation/migration, pro-apoptosis/cycle arrest, and immune activation.
• •Lidamycin-based ADCs, along with lidamycin itself, can induce antitumor immunity in solid tumors.

1. Introduction

As the second leading-cause of death, cancer remains a major public health challenge and causes one in six deaths globally. In 2022, approximately 20 million new cancer cases were reported globally, with cancer-related deaths reaching nearly 9.7 million [1]. Surgery, radiotherapy, and chemotherapy, either individually or in combination, remain the three cornerstones of cancer treatment [2]. However, these three treatment modalities have limitations. For example, surgery is only feasible for patients with localized cancer without metastasis [3], whereas chemotherapy and radiation therapy not only kill both cancer cells and normal cells, but also cause systemic toxicities and compromise the patient's immune system [4,5]. Targeted therapy, which can specifically identify and kill cancer cells, plays an increasing important role in the treatment of malignant tumors due to its precise targeting and minimal side effects [6]. Various targeted therapy strategies have been developed, including monoclonal antibodies (mAbs), small-molecule inhibitors, immunotherapy, and antibody-drug conjugates (ADCs) [7]. ADCs have been one of the fastest-growing oncology therapeutics in recent years [8]. They consist of mAbs conjugated to cytotoxic payloads via chemical linkers, combining the target specificity and long circulation half-life of mAb with the high cytotoxic potency of antitumor agents that are too toxic to be used alone, thereby exhibiting strong selective killing efficacy against cancer cells [9]. Since the U.S. Food and Drug Administration (FDA) approval of the first ADC (gemtuzumab ozogamicin) in 2000, 15 ADCs have been approved for clinical application [10]. The first step for an ADC to enter tumor cells is the identification and binding of an antibody to antigens on the surface of cancer cells; thus, selecting an appropriate target is the primary consideration to prepare new ADCs.

Mesothelin (MSLN), a glycosylphosphatidylinositol (GPI)-anchored glycoprotein, is physiologically restricted to mesothelial cells lining the peritoneum, pleura, and pericardium [11]. However, the exact physiological role of MSLN remains unknown [12]. A previous study revealed that MSLN-knockout mice developed and reproduced normally, which indicates that MSLN is dispensable for normal development and survival [13]. In contrast, MSLN exhibits significant upregulation in multiple malignancies, including mesotheliomas as well as ovarian, pancreatic, pulmonary, thymic, and gastric cancers [14]. Moreover, MSLN plays an important role in shaping the tumor microenvironment, promoting cancer cell survival, proliferation, migration/invasion, and drug resistance. High MSLN expression is linked to unfavorable survival rates among patients [15]. The distinct overexpression of MSLN in malignant cells compared to healthy tissues, coupled with its negligible expression in essential organ parenchyma and pivotal role in tumorigenesis, has positioned it as an attractive target for anticancer drug development [16]. Multiple MSLN-targeted therapeutics, including mAbs [17], ADCs [18], anti-MSLN vaccines [19], and chimeric antigen receptor (CAR) T cell therapies [20], are currently under evaluation in clinical trials, but none have yet received clinical approval. Most MSLN-directed ADCs utilize microtubule inhibitors as payloads; for example, RC-88 and DMOT4039A use monomethyl auristatin E (MMAE) [21,22], BMS-986148 uses tubulysin [23], and anetumab ravtansine uses maytansinoid DM4 [24]. Unfortunately, although tubulin inhibitors have greater toxicity against rapidly dividing tumor cells, they are much less effective against quiescent cancer cells (e.g., cancer stem cells), a feature that readily leads to drug resistance and tumor recurrence [25]. Therefore, there is an unmet clinical need for MSLN-targeting ADCs that utilize payloads with different mechanisms of action compared to microtubule inhibitors.

Lidamycin (LDM) (C-1027) is an enediyne antibiotic that exhibits potent cytotoxicity against cancer cells. LDM entered phase-II clinical trials several years ago but was discontinued because of its narrow therapeutic index and significant toxicity concerns. The LDM molecule consists of an apoprotein of LDM (LDP) and an active enediyne chromophore of LDM (AE), where LDP acts as a protective carrier, while AE is the active component responsible for DNA damage [26]. Intriguingly, AE and LDP can be isolated and reconstituted in vitro, with the reconstructed LDM molecules retaining biological activity comparable to that of the native form. Its potent antitumor activity, along with its reversible disassembly and reassembly, makes LDM a promising payload for ADC development.

In this study, we first produced a novel anti-MSLN antibody, h3B7A, and assessed its activity. The fusion protein h3B7A-LDP, comprising the antibody h3B7A and LDP of LDM, was subsequently prepared and evaluated. Finally, we constructed the ADC h3B7A-LDM, investigated its antitumor efficacy and molecular mechanism, and demonstrated its promise as a treatment for MSLN-positive malignancies.

2. Materials and methods

2.1. Reagents

Freund's complete adjuvant (F5581), non-complete adjuvant (F5506), and Triton X-100 (T8787) were obtained from Sigma (New York, NY, USA). Hypoxanthine-aminopterin-thymidine (HAT) supplement (21060017), hypoxantin-aminopterin (HA) supplement (11067030), Lipofectamine™ (15338100), bicinchoninic acid (BCA) protein assay kit (23227), Alexa Fluor™ 555 labeled donkey anti-rabbit IgG (1:500, A-31572), and DyLight 680 Antibody Labeling Kit (53056) were purchased from Thermo Fisher Scientific Inc. (Waltham, MA, USA). Mouse mAb isotyping kit (BF16002) and hybridoma feeder (CM-2001) were sourced from Biodragon (Beijing, China), while 3,3′,5,5′-tetramethylbenzidine (TMB) substrate (PA107) was from TIANGEN Biotech (Beijing) Co., Ltd. (Beijing, China). Radio-immune precipitation assay (RIPA) lysis buffer (R0010) and horseradish peroxidase (HRP)-labeled secondary antibodies, i.e., goat anti-mouse IgG (1:2000, SE131) and goat anti-human IgG (SE101, 1:2000), as well as fluorescein isothiocyanate (FITC)-labeled goat anti-human IgG (SF101, 1:200) were provided by Solarbio (Beijing, China). Antibodies targeting cyclin D1 (#2922s, 1:1000), caspase-3 (#9662s, 1:1000), cleaved caspase-3 (#9661s, 1:1000), cyclin-dependent kinase (CDK) inhibitor 1A (P21) (#2947s, 1:1000), CDK6 (#3136s, 1:1000), apoptosis-related cysteine peptidase (caspase-9) (#9508s, 1:1000), cleaved caspase-9 (#9507s, 1:1000), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (#97166, 1:2000), lysosomal-associated membrane protein 1 (LAMP1) (#9091s, 1:100), and calreticulin (CRT) (#62304, 1:100) were obtained from Cell Signaling Technology (Boston, MA, USA). 4′,6-Diamidino-2-phenylindole (DAPI) (ZLI-9557), diaminobenzidine (DAB) substrate solution (ZLI-9018), and HRP-labeled goat anti-rabbit IgG (ZB-2301, 1:2000) were provided by ZSGB-BIO (Beijing, China). Cycle kit (CCS012) and apoptosis kit (AP101) were sourced from Multi Sciences (Hangzhou, China). Enhanced adenosine triphosphate (ATP) assay kit (S0027) was obtained from Beyotime Biotechnology (Shanghai, China). Mouse tumor necrosis factor alpha (TNF-α) enzyme-linked immunosorbent assay (ELISA) kit (JL10484), mouse macrophage inflammatory protein 1 alpha (MIP-1α) ELISA kit (JL20414), mouse interferon gamma (IFN-γ) induced protein 10 kDa (IP10) ELISA kit (JL13372), and human high mobility group box 1 (HMGB-1) ELISA kit (JL13693) were purchased from Jianglai Industry Co., Ltd. (Shanghai, China). Alexa Fluor 700 anti-mouse Ly-6G (1:100, 127621), fluorescein isothiocyanate (FITC) anti-mouse leukocyte common antigen (CD45) (1:100, 147709), allophycocyanin (APC) anti-mouse Integrin αX (CD11c) (1:100, 117310), phycoerythrin (PE) anti-mouse lymphocyte antigen 6 complex, and PE anti-mouse lymphocyte antigen 6 complex, locus G (Ly-6G) (Ly-6G) (1:100, 127607) were obtained from BioLegend (San Diego, CA, USA). MF-T antibody (the anti-MSLN antibody used in the preparation of anetumab ravtansine) was generously gifted by Nanjing Huayan Biotechnology Co., Ltd.. The ADC hIMB1636-LDM, which consists of anti-trophoblast cell surface antigen 2 (anti-Trop2) antibody hIMB1636 and LDM, was prepared in our laboratory.

2.2. Cell lines

Mouse myeloma cell SP2/0, ovarian cancer (OV) cell SKOV3, pancreatic cancer cell SW1990, lung cancer cells H460 and H1975, and mouse Lewis lung carcinoma (LLC) cells were preserved by our laboratory. OV cell OVCAR3 and pancreatic cancer cell AsPC-1 were obtained from Shanghai Jin Yuan Biotechnology Co., Ltd. (Shanghai, China). OVCAR3 and SW1990 cell lines were cultured in Dulbecco's modified Eagle medium (DMEM) complete medium (Shanghai BasalMedia Technologies Co., Ltd., Shanghai, China). Other cell lines were maintained in Roswell Park Memorial Institute 1640 (RPMI-1640) complete medium (Shanghai BasalMedia Technologies Co., Ltd.). All cells were maintained at 37 °C in a humidified 5% CO₂ atmosphere.

2.3. Human tissue microarray

Paired paraffin-embedded tissue sections of cancerous and adjacent non-cancerous tissue from ovarian and pancreatic cancer patients were purchased from Shanghai Outdo Biotech (Shanghai, China).

2.4. Animals

Female BALB/c mice (eight weeks old) were provided by SPF Biotechnology Co., Ltd. (Beijing, China, SCXK (Beijing) 2019-0010). BALB/c nude mice (six weeks old) were obtained from Beijing HFK Bioscience Co., Ltd. (Beijing, China, SCXK (Beijing) 2020-0004) and GemPharmatech Co., Ltd. (Nanjing, China, SCXK (Jiangsu) 2023-0009). The C57BL/6‌ mice were purchased from GemPharmatech Co., Ltd.. All mice were bred and housed under specific pathogen free conditions at the Institute of Medicinal Biotechnology (Beijing, China), and all related experiments were approved by the Institute's Ethics Committee of Institute of Medicinal Biotechnology (Approval Nos.: IMB-20230513D101 and IMB-20250326D602), following the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines [27].

2.5. MSLN expression analysis using bioinformatics

The gene expression profiling interactive analysis (GEPIA) was used to analyze the RNA sequencing expression data of bulk tumorous and normal samples from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) databases [28]. We firstly analyzed the expression levels of MSLN in a variety of tumors and their corresponding adjacent normal tissues, such as breast invasive carcinoma (BRCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), OV, pancreatic adenocarcinoma (PAAD), lung adenocarcinoma (LUAD), rectum adenocarcinoma (READ), gastric adenocarcinoma (STAD), uterine corpus endometrial carcinoma (UCEC), uterine carcinosarcoma (UCS), and cholangiocarcinoma (CHOL). The correlation between MSLN messenger RNA (mRNA) expression and survival was then analyzed in OV, LUAD, STAD, and PAAD.

2.6. Immunization of mice and generation of hybridoma cells

BALB/c mice were immunized by subcutaneous injection with 25–50 μg of MSLN protein in Freund's adjuvant (complete or incomplete) every three weeks for a total of 4–6 times. Orbital vein blood was collected from mice one week after the third immunization to measure serum antibody titers. A final intraperitoneal booster was administered three days before fusion. Splenocytes were fused with myeloma SP2/0 cells according to a previous report [29]. Briefly, cells were mixed, washed, and fused with 50% polyethylene glycol 1450. The fused cells were resuspended and seeded in 96-well plates with HAT supplemented RPMI-1640 medium. After seven days, the medium was replaced with HT supplemented RPMI-1640 medium, and positive hybridoma clones were screened by ELISA (Section 2.15). After three rounds of subcloning, a stable hybridoma cell line, designated 3B7, was isolated, and its secreted antibody was named 3B7A. The heavy chain and light chain genes of 3B7A were sequenced from the hybridoma cells, and their nucleotide/amino acid sequences were analyzed using National Center for Biotechnology Information (NCBI) BLAST database for homology assessment.

2.7. Antibody subtype detection

The subtype of 3B7A antibody was determined using mouse mAb isotyping kit (BF16002). Briefly, a 96-well plate was coated with MSLN protein (0.2 μg/well) and incubated at 4 °C overnight. After washing, 3B7A antibody (0.1 μg/well) was added and then the plate was incubated for 2 h. Following another wash, goat anti-mouse Ig-HRP was added, and the plate was incubated at 37 °C for 1 h. Then, TMB solution was added, and the plate was incubated at room temperature for 30 min. The reaction was stopped with 2 M H₂SO₄. The optical density (OD) at 450 nm was measured using a Spectra Max i3X microplate reader (Molecular Devices, Sunnyvale, CA, USA).

2.8. Preparation of anti-MSLN mAb

The BALB/c mice were injected intraperitoneally with hybridoma cells secreting 3B7A antibody to produce ascitic fluid. The antibody was isolated by (NH₄)₂SO₄ precipitation, followed by purification with a HiTrap Protein G HP column (GE Healthcare, Freiburg, Germany). The purified samples were then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under non-reducing and reducing conditions. Size exclusion-high performance liquid chromatography (SEC-HPLC) was used to detect antibody purity (column, BioSep-SEC-s3000; mobile phase, phosphate-buffered saline (PBS); and detection wavelength, 280 nm).

2.9. Binding activity of 3B7A to recombinant MSLN antigen

The antigen-binding kinetics of 3B7A antibody was investigated using surface plasmon resonance (SPR) with a Biacore T200 instrument (GE Healthcare). A CM5 sensor chip was immobilized with 8.5 μg/mL of 3B7A antibody. Serially diluted MSLN protein (Sino Biological, Beijing, China) in PBS with Tween-20 (PBST) buffer (concentration range, 60 to 3.75 nM) was injected over the sensor chip at a flow rate of 10 μL/min. Sensor surface regeneration was achieved with 60-s injections of 10 mM glycine-HCl (pH 2.5). The results were analyzed by the Biacore evaluation software. The measured SPR sensorgrams were obtained by subtracting reference cell (channel 1) signals from the active flow cell (channel 2). SPR signals are reported in resonance unit (RU). The data were fitted to a 1:1 binding model to determine the binding rate constant (k_a, M⁻¹s⁻¹), dissociation rate constant (k_d, s⁻¹), and equilibrium dissociation constant (K_D = k_d /k_a, M). A smaller K_D value indicates higher binding affinity.

2.10. Western blot analysis

The treated cells were harvested and lysed to extract cellular protein. Protein concentration was determined using a BCA protein assay kit. Samples were separated by SDS-PAGE, transferred onto polyvinylidene difluoride (PVDF) membranes, and blocked. Membranes were then incubated overnight at 4 °C with primary antibody (3B7A), followed by incubation with the corresponding secondary antibody for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) system (Tanon, Shanghai, China).

2.11. Binding affinity of 3B7A to natural MSLN antigen

2.11.1. Flow cytometry analysis

The binding of 3B7A to the natural MSLN antigen on the tumor cell surface was evaluated by flow cytometry. Harvested cells (2.5 × 10⁶ per tube) were treated with 10-fold serial dilutions of 3B7A (10–0.01 μg/mL) and incubated at 4 °C for 1 h. Following washing, cells were incubated with FITC-conjugated secondary antibodies at 4 °C for 1 h in the dark. Following another wash, the cell-associated fluorescence signal was analyzed by flow cytometry (ACEA Biosciences Inc., San Diego, CA, USA).

2.11.2. Immunohistochemistry (IHC)

For IHC analysis, the paraffin-embedded tissue sections were dewaxed in xylene and rehydrated through graded alcohol solutions, as previously described [29]. After antigen retrieval, the sections were permeabilized with 0.5% Triton X-100 in PBS, treated with 3% hydrogen peroxide (H₂O₂) solution for 15 min to quench endogenous peroxidase activity, and blocked with 3% bovine serum albumin (BSA) in PBS for 15 min. The sections were then incubated with the primary antibody (3B7A) at 4 °C overnight. After washing, they were incubated with secondary antibodies (30 min, at room temperature). DAB development was followed by hematoxylin counterstaining, dehydration through a graded ethanol series, xylene clearing, and resin mounting. Images were captured using an Olympus DP72 microscope (Olympus, Tokyo, Japan).

2.12. Internalization analysis by confocal microscopy

Cells were grown on glass coverslips in two six-well plates and cultured for 48 h. Following treatment with 3B7A (10 μg/well) at 4 °C for 30 min, one plate was placed in a 37 °C, 5% CO₂ incubator for 2 h and the other were kept at 4 °C. Subsequently, cells were successively fixed, permeabilized, and blocked as previously reported [29]. Next, the cells were incubated with anti-LAMP1 antibody overnight at 4 °C, and then with secondary antibodies at 37 °C for 30 min in the dark. The cell nuclei were stained with DAPI (blue). Images were captured using a laser scanning fluorescent confocal microscope (Olympus).

2.13. In vivo fluorescence imaging

The in vivo tumor-targeting capability of 3B7A was evaluated in OVCAR3, AsPC-1, and SKOV3 xenograft tumor models. For each model, 5.0 × 10⁶ cells were subcutaneously inoculated into the right flank of each BALB/c nude mice. When tumor volumes reached approximately 200–300 mm³, the mice were intravenously injected with DyLight 680-labeled 3B7A (20 mg/kg). The mice were then anesthetized and imaged using the Xenogen IVIS-200 imaging system (Xenogen Inc., Alameda County, CA, USA) at the indicated time points. Finally, the mice were humanely euthanized via CO₂ asphyxiation, and the tumors along with other normal tissues (heart, liver, spleen, lungs, kidneys, pancreas, large intestine, small intestine, femur, and stomach) were excised and photographed.

2.14. Humanization of mouse antibody 3B7A

Humanization of 3B7A antibody was performed using the complementarity-determining region (CDR)-grafting method [30]. In brief, the sequences of heavy- and light-chain variable regions of 3B7A were analyzed using the international ImMunoGeneTics information system (IMGT, www.imgt.org) and the VQuest sequence alignment tool [31]. Three different human variable heavy chain (V_H) as well as variable light chain (V_L) framework acceptors were identified as templates for humanization. Homology modeling was performed using Discovery Studio 2019. Subsequently, murine-derived CDRs were grafted into the human variable-domain framework to minimize the murine component in the antibodies, and framework residues potentially critical for antigen-binding were identified and backmutated to the murine sequence. All humanized V_H and V_L sequences were synthesized, cloned into the mammalian expression vector containing human IgG1 isotype constant domains, and expressed and purified as full-length IgG. Antibodies with high antigen-binding activity were identified by ELISA, and their K_D values were determined by SPR assay. Finally, the degree of antibody humanization was quantified using online analysis software T20 score and IMGT Domain GapAlign Tool. The final humanized and optimized antibody was named h3B7A.

2.15. Binding activity and internalization rate assay

MF-T is a fully human anti-MSLN antibody used in the preparation of ADC anetumab ravtansine (BAY 94–9343). To characterize h3B7A, we compared the binding activity and internalization rate of MF-T and h3B7A in AsPC-1 cells. For the binding analysis, harvested cells were treated with 10 nM of either h3B7A or MF-T and incubated at 4 °C for 1 h. After three washes with PBS, the cells were incubated with FITC-conjugated secondary antibodies at 4 °C for 1 h in the dark. Following additional washes to remove unbound secondary antibodies, cell-associated fluorescence signals were analyzed by flow cytometry.

The internalization rate assay was performed according to the procedure of Zhao et al. [32]. Briefly, harvested cells were incubated with 10 nM of either h3B7A or MF-T for 1 h at 4 °C. After washing, control cells (untreated) were maintained at 4 °C, while the other cells were transferred to 37 °C and incubated for 1, 2, 4, or 6 h to allow internalization. The process was terminated by washing with cold PBS. Cells were then stained with a goat anti-human IgG (H&L)-Alexa Fluor 647 antibody for 1 h at 4 °C. After a final wash, cells were analyzed by flow cytometry. The internalization rate (%) was calculated using the formula: [1 − M/T] × 100%, where M means mean fluorescence intensity (MFI) of the sample at each incubation time and T means total MFI of the non-treated cells maintained at 4 °C.

2.16. Generation, expression, and purification of the h3B7A-LDP fusion protein

The expression vector pIZDHL (containing Bleo and Dhfr genes as selection marker), which harbors a human IgG1-Fc domain, was preserved in our laboratory [33]. We fused C-terminus of the LDP gene (from LDM) to the N-terminus of V_L gene of antibody h3B7A via a peptide linker (SGGPEGGS) to generate fusion gene fragment LDP-V_L. The gene fragments of h3B7A V_H and LDP-V_L were cloned into the vector pIZDHL using standard sub-cloning methods, generating the recombinant expression vector pIZDHL-h3B7A-LDP. Additionally, V_H and V_L sequences were separately cloned into the pIZDHL vector to construct the antibody h3B7A expression vector pIZDHL-h3B7A. Both pIZDHL-h3B7A-LDP and pIZDHL-h3B7A vectors were linearized prior to transfection, respectively. All gene sequences were synthesized and provided by GenScript Company (Nanjing, China).

The linearized vectors were transfected into CHO/dhFr-cells and the single-cell clones with high-level expression of antibody and fusion protein were screened as previously reported [34]. The recombinant proteins h3B7A-LDP and h3B7A were purified using HiTrap™ Protein G columns (GE Healthcare) following the manufacturer's recommended protocol. Protein concentration was determined using a BCA assay kit, and the purified proteins were stored at −80 °C.

2.17. ELISA

The binding affinity of h3B7A-LDP and h3B7A to recombinant MSLN protein were determined by indirect ELISA. The MSLN protein (0.2 μg/well) was coated onto a 96-well plate and incubated overnight at 4 °C. After blocking with 5% BSA/PBS for 2 h, serially diluted h3B7A-LDP or h3B7A were added to the wells, followed by incubation at 37 °C for 2 h. The plate was then washed with PBST, and HRP-labeled goat anti-human IgG was added to each well for 1 h. After another washing step, TMB substrate was added to initiate colorimetric reaction, which was stopped by adding 2 M H₂SO₄. Finally, the OD was measured at 450 nm using a microplate reader.

2.18. Biolayer interferometry (BLI) assay

The binding affinity of h3B7A-LDP or h3B7A protein to MSLN antigens was measured by BLI using Gator™ Label-Free Bioanalysis Instrument (GatorBio, Suzhou, China). Anti-human IgG Fc biosensors were pre-incubated in Q buffer (10 mM PBS, 0.02% Tween 20, 0.2% BSA; PH 7.4) for at least 300 s before use. Next, 100 nM h3B7A-LDP or h3B7A antibody was captured on the anti-human IgG Fc biosensors for 120 s. After equilibration in Q buffer for 120 s, the sensors were exposed to MSLN antigen solutions (400 nM in 2-fold serial dilutions) for 120 s (association) followed by immersion in Q buffer for 300 s (dissociation). Data were analyzed using GatorBio's data analysis software.

2.19. Preparation of the h3B7A-LDM

The AE was prepared as previously reported [34]. To generate enediyne-containing ADC h3B7A-LDM, free AE was added to h3B7A-LDP protein solution at a 3:1 molar ratio, followed by overnight incubation at 4 °C with gently shaking. Uncoupled AE in the solution was removed by ultrafiltration centrifugation at 4000 rpm and 4 °C. The components of h3B7A-LDM were finally identified using a Delta Pak C4-300A column (Waters Corportion, Milford, MA, USA). All operations were performed under dark conditions.

2.20. Cell viability assay using Cell Counting Kit-8 (CCK-8)

The cytotoxicity of h3B7A-LDM to MSLN-positive (MSLN⁺) tumor cells was analyzed using CCK-8 assay. In brief, OVCAR3, H460, SW1990, and AsPC-1 cells were seeded into 96-well plates (5 × 10³ cells/well) and incubated for 24 h at 37 °C in 5% CO₂ atmosphere. Next, the cells were treated with h3B7A-LDM or LDM at varying concentrations (0, 0.001, 0.003, 0.01, 0.03, 0.1, 0.3, and 1 nM) for 72 h. After removing the culture supernatant, 90 μL of complete culture medium plus 10 μL of CCK-8 reagent were added to each well, followed by incubation at 37 °C for 1 h. The absorbance at 450 nm was read on a microplate reader.

2.21. Real-time cell proliferation and migration assays

RTCA E−16 plates were used for cell proliferation assay. At first, 50 μL of serum-free medium was added to each well of the plate, followed by measurement of baseline impedance to ensure a good connection. After the cells (OVCAR3, H460, SW1990, and AsPC-1) were collected, counted, diluted, and resuspended, 100 μL of cell suspension (containing 5 × 10³ cells) was seeded in each well of the E-16 plates and allowed to attach for 30 min at room temperature. Different concentrations of h3B7A-LDM were added to the wells, while the negative control was treated with an equal volume of PBS. Finally, the cell index (CI) was monitored by the xCELLigence RTCA system (ACEA Biosciences Inc.) every 15 min for approximately 72 h.

Cell migration was assessed using a 16-well cell invasion and migration (CIM)-plate, which features a dual-chamber design: the upper chamber contains the cell suspension in serum-free medium, while the lower chamber is filled with medium supplemented with fetal bovine serum (FBS) or a chemoattractant. First, to measure the background impedance, the plate was installed in the xCELLigence instrument with 165 μL of cell complete culture medium (containing 10% FBS) in the lower chamber and 30 μL of serum-free medium in the upper chamber. After the cells (H460, SW1990, and AsPC-1) were harvested, counted, diluted, and resuspended, 5 × 10⁴ cells in 100 μL of serum-free medium were seeded in the upper chamber. Then, different concentrations of h3B7A-LDM were added to wells, while the negative control was treated with an equal volume of PBS. The plate was placed onto the xCELLigence RTCA instrument to monitor the CI every 15 min for approximately 48 h.

2.22. Cell cycle and apoptosis assays

H460, SW1990, or AsPC-1 cells were cultured in six-well plates and treated with 1 nM h3B7A-LDM for 24 h, respectively. Then, the cells were harvested and washed with PBS. For cell cycle distribution assessment, cells were stained with 0.5 mL of propidium iodide (PI)/ribonuclease (RNase) solution and incubated in the dark for 30 min at room temperature. The DNA content was then analyzed by flow cytometry, and data were processd using NovoExpress software. Cell apoptosis was evaluated by double staining with 10 μL of annexin V-FITC and 5 μL of PI for 15 min in the dark at room temperature. Subsequently, 400 μL of binding buffer was added, and samples were analyzed by flow cytometry within 1 h.

2.23. Immunogenic cell death (ICD) hallmark assessment

2.23.1. CRT exposure

MSLN-positive cells (H460, SW1990, and AsPC-1) were treated with 1 nM of h3B7A-LDM or LDM for 24 and 48 h. To assess surface-exposed CRT, cells were collected and double-stained with a CRT-specific antibody and PI, following the manufacturer's protocol. CRT on the cell surface was analyzed by flow cytometry via gating on PI-negative cells.

2.23.2. ATP and high mobility group protein B1 (HMGB1) release assays

H460, SW1990, and AsPC-1 cells were plated on six-well dishes and allowed to reach 70% confluence. The cells were then treated with 1 nM h3B7A-LDM or LDM for 48 h. For ATP secretion measurement, 100 μL of ATP test solution was added to 96-well plate. Then, 100 μL of cell culture medium (from different treatment groups) was collected, centrifuged, and transferred to the 96-well plate containing the ATP solution. Relative light unit (RLU) values were measured using a luminometer Spectra Max i3X (Molecular Devices).

Extracellular HMGB1 in the supernatants of drug-treated cells was quantified using an HMGB1 ELISA kit.

2.24. Immune cell activation and in vivo vaccination studies

Five million H460 cells suspended in 20% Matrigel (Corning Inc., Corning, NY, USA) were subcutaneously inoculated into the right flank of female BALB/c nude mice. Once the tumor volume reached approximately 150 mm³, the mice were randomly divided into different groups (n = 5 per group) and intravenously administered 0.8 mg/kg h3B7A-LDM, 0.045 mg/kg LDM, or PBS (control) every four days for two doses. Three days after the last injection, tumors were harvested and weighed. Each tumor was then divided into two portions: one portion was mechanically dissociated into single-cell suspensions for flow cytometric analysis after staining with fluorescent antibodies against murine CD45, CD11c, and Ly6G; the other portion was lysed in RIPA buffer for cytokine analysis, including TNF-α, MIP-1α, and IFN-γ-induced protein 10/C−X−C motif chemokine ligand 10 (IP-10/CXCL10), using ELISA.

C57BL/6 mice were randomly allocated into two groups (n = 6 per group). For vaccine preparation, LLC cells were treated with 1 nM LDM for 72 h before collection. Group 1 (treatment group) was subcutaneously vaccinated with 2 × 10⁶ LDM-treated LLC cells, while group 2 (control group) received necrotic LLC cells prepared by three freeze-thaw cycles. One week later, all animals were rechallenged subcutaneously in the opposite flank with 5 × 10⁶ viable LLC cells. Tumor emergence and growth were subsequently monitored in both groups.

2.25. In vivo antitumor effect of h3B7A-LDM

For tumor xenograft model, H460 cells (5.0 × 10⁶) were subcutaneously injected into the right flank of six-week-old BALB/c nude mice. Once the tumor volumes reached approximately 100 mm³, the mice were randomly allocated into five groups (n = 6 per group): control group, 0.8 mg/kg h3B7A-LDM-treated group, 0.4 mg/kg h3B7A-LDM-treated group, 0.8 mg/kg h3B7A-treated group, and 0.045 mg/kg LDM-treated group. Treatments were administered intravenously every five days for a total of four doses. Tumor growth was monitored every five days using vernier calipers to measure the length (L) and width (W), and tumor volume was calculated as: tumor volume = 0.5 × L × W². Body weights were recorded concurrently to assess treatment tolerability. At the experimental endpoint, mice were euthanized, and tumors were excised and weighed. Major organs (heart, liver, spleen, lungs, kidneys, bone, and intestine) were harvested and fixed for histopathological analysis by hematoxylin and eosin (H&E) staining. Similarly, we also established SW1990 and AsPC-1 xenograft models to further evaluate the antitumor efficacy of h3B7A-LDM.

Moreover, we detected the tumor-specificity of h3B7A-LDM with H460 cell-derived xenograft model. Once the tumor volumes reached approximately 100 mm³, the mice were randomly allocated into three groups (n = 6 per group): control group, 0.8 mg/kg h3B7A-LDM-treated group, and 0.8 mg/kg hIMB1636-LDM-treated group (as a non-targeting IgG-LDM control group). Treatments were administered intravenously every four days for a total of three doses. Tumor growth was monitored every three days and body weights were recorded as mentioned above.

2.26. Hematological parameters

Blood samples were collected from mice in the control group, the h3B7A-LDM-treated group, and the LDM-treated group, and then placed in ethylenediaminetetraacetic acid dipotassium salt (EDTA-K2) anticoagulant tubes. Hematological parameters, including hemoglobin (g/L), erythrocytes (×10¹²/L), leukocytes (×10⁹/L), neutrophils (×10⁹/L), and platelets (×10⁹/L), were measured using a Hitachi Automatic Analyzer (Hitachi, Tokyo, Japan).

2.27. Statistical analysis

The data were analyzed using GraphPad Prism 7 software, and the results are expressed as mean ± standard deviation (SD). Differences between groups were assessed by two-way analysis of variance (ANOVA), followed by Tukey's test. All experiments were done at least three times. P < 0.05 was considered to be statistically significant.

3. Results

3.1. MSLN gene expression analysis

Gene expression analysis of MSLN were conducted using GEPIA. The results revealed that MSLN mRNA levels were greater in various tumors compared to corresponding normal tissues (Fig. 1A). Moreover, patients with low MSLN expression exhibited a higher survival rate than those with high expression in multiple cancer types, including OV, LUAD, STAD, and PAAD (Fig. 1B). These findings demonstrate that MSLN is overexpressed in multiple cancer types and is strongly associated with poor patient prognosis, underscoring its potential as a promising therapeutic target.

Fig. 1.

Mesothelin (MSLN) gene expression analysis and generation of the novel antibody 3B7A. (A) MSLN messenger RNA (mRNA) expression in tumor tissues (red) versus normal tissues (gray) from the gene expression profiling interactive analysis (GEPIA) database. ^∗P < 0.05 vs. the control. (B) Analysis of overall survival in cancer patients with high versus low MSLN expression using the GEPIA database. (C) Schematic diagram of anti-MSLN antibody production via hybridoma technology. (D) Purity assessment of 3B7A by size exclusion-high performance liquid chromatography (SEC-HPLC). (E) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of purified 3B7A antibody under reducing and non-reducing conditions. (F) 3B7A antibody isotype identification using a commercial detection kit. Data are presented as mean ± standard deviation (SD) (n = 3). (G) Binding kinetics of 3B7A to the MSLN antigen analyzed by surface plasmon resonance (SPR) analysis. TPM: transcripts per million; BRCA: breast invasive carcinoma; CESC: cervical squamous cell carcinoma and endocervical adenocarcinoma, COAD: colon adenocarcinoma; ESCA: esophageal carcinoma; OV: ovarian cancer; PAAD: pancreatic adenocarcinoma; LUAD: lung adenocarcinoma; READ: rectum adenocarcinoma; STAD: gastric adenocarcinoma; UCEC: uterine corpus endometrial carcinoma; UCS: uterine carcinosarcoma; CHOL: cholangiocarcinoma; HR: hazard ratio; UV-vis-WVL: ultraviolet-visible spectroscopy-wavelength; OD: optical density; RU: response unit; K_D: equilibrium dissociation constant.

3.2. Development of 3B7A, a novel anti-MSLN mAb

To obtain a promising candidate molecule for MSLN-targeting therapy, a hybridoma cell clone (3B7) stably secreting an anti-MSLN mAb (mAb 3B7A), was established through mouse immunization, cell fusion, and hybridoma screening (Figs. 1C and S1). The 3B7A antibody, with a purity of >95%, was obtained by purifying ascites from mice inoculated with the hybridoma cells (Fig. 1D). SDS-PAGE confirmed correct antibody assembly, revealing expected molecular weights (MWs): 150 kDa for intact antibody molecules, 50 kDa for heavy chains, and 25 kDa for light chains (Fig. 1E). Isotyping identified 3B7A as a mouse IgG1 subtype with a kappa chain (Fig. 1F). Hybridoma sequencing and subsequent alignment revealed that 3B7A is a novel antibody with unique amino acid sequence. K_D is a critical determinant of therapeutic antibody efficacy. SPR assay revealed high-affinity binding between 3B7A and MSLN, with a K_D of 3.96 nM (Fig. 1G).

3.3. Specific binding, endocytosis, and tumor-targeting capability of 3B7A

MSLN expression levels in different cells were determined by Western blot analysis. The results showed that OVCAR3 cells exhibited the highest level of MSLN expression, followed by AsPC-1 cells, whereas SKOV3 cells minimal or no MSLN expression (Fig. 2A). Binding of the 3B7A antibody to natural MSLN protein on the surface of these cells was evaluated by flow cytometry. As shown in Fig. 2B, binding to OVCAR3 and AsPC-1 cells with increasing concentration of 3B7A antibody resulted in a shift in emission peak with a corresponding increase in fluorescence intensity. In contrast, no notable fluorescent signal shift was observed in MSLN-negative SKOV3 cells. These results were consistent with the Western blot data (Fig. 2A), confirming the specific binding between 3B7A antibody and natural MSLN antigen on the surface of cancer cells.

Fig. 2.

Characterization and humanization of 3B7A monoclonal antibody (mAb). (A) The binding activity of 3B7A to mesothelin (MSLN) in different cells was assessed by Western blot. Band intensities were quantified by densitometry and normalized to GAPDH. Data are presented as mean ± standard deviation (SD) (n = 3). (B) Binding of 3B7A to OVCAR3, AsPC-1, and SKOV3 cells at varying concentrations was analyzed by flow cytometry. (C) Representative image of 3B7A binding to ovarian cancer (OV), pancreatic cancer, and adjacent tissues. (D) Binding, internalization, and lysosomal localization of 3B7A in OVCAR3 and SKOV3 cells were detected using a laser scanning confocal microscope. (E) In vivo fluorescence imaging was used to analyze 3B7A in OVCAR3 and SKOV3 cell-derived xenograft models. 3B7A was labeled with DyLight 680 and administered intravenously via tail vein injection to mice bearing tumor xenografts. Serial in vivo fluorescence imaging was performed at predetermined time points to monitor tumor targeting in both OVCAR3 and SKOV3 xenograft models. The color scale represents photon flux intensity (photons/s/cm²/steradian). (F) Binding of h3B7A and MF-T to natural MSLN antigen on AsPC-1 cells was evaluated via flow cytometry. (G) The endocytosis rates of h3B7A and MF-T were detected in AsPC-1 cells by flow cytometry. Data represent mean ± SD (n = 3). (H) Binding affinity of h3B7A to murine MSLN was measured by enzyme-linked immunosorbent assay (ELISA). ^∗∗∗P < 0.001 vs. the control. GAPDH: glyceraldehyde-3-phosphate dehydrogenase; FITC: fluorescein isothiocyanate; DAPI: 4′,6-diamidino-2-phenylindole; LAMP1: lysosomal associated membrane protein 1; OD: optical density.

The binding specificity of mAb for human tumor tissues is critical for the efficacy and safety of antibody-based therapeutics. Accordingly, immunohistochemical staining was used to evaluate the binding of 3B7A to human tumor tissues. We found that in ovarian and pancreatic cancer, compared with their corresponding adjacent tissues, 3B7A specifically binds to tumor tissue (Fig. 2C). Furthermore, to assess the potential of 3B7A as a payload carrier for ADC development, we evaluated its cellular internalization capacity in MSLN-expressing cancer cells using confocal microscopy. As shown in Fig. 2D, in OVCAR3 cells at 4 °C, green fluorescence (3B7A) was localized on the cell surface, and red fluorescence (lysosome) was present in the cytosol after labeling lysosomes with anti-LAMP1 (lysosome-associated membrane protein 1) antibody (red). This demonstrates that 3B7A can directly bind to MSLN-positive cancer cells. Following incubation at 37 °C, green staining of 3B7A appeared throughout the cytoplasm, with partial colocalization with LAMP1 (yellow), indicating that 3B7A is internalized by antigen-positive cancer cells and trafficked to the lysosome (Fig. 2D). However, in SKOV3 cells, no green fluorescence was observed at either 4 or 37 °C, confirming that 3B7A does not bind to antigen-negative cancer cells.

The in vivo tumor targeting and biodistribution of 3B7A were evaluated in tumor-bearing mice using an in vivo fluorescence imaging system. DyLight 680-labeled 3B7A was injected into OVCAR3-, AsPC-1-, and SKOV3-bearing mice. An intense red/yellow fluorescence signal was observed within 60 h (or 12 h) and sustained for approximately 15 days in the tumor region of MSLN-positive OVCAR3 and AsPC-1 xenograft model mice (Figs. 2E and S2A). In contrast, the MSLN-negative SKOV3 model exhibited significantly reduced fluorescence signals at the tumor site (Fig. 2E). At the end of the experiments, major organs and tumor tissues were harvested. Fluorescence signals were detected only in tumor tissue, whereas the heart, liver, spleen, lungs, kidneys, and other tissues showed no fluorescence, indicating the specific accumulation of 3B7A in tumors (Fig. S2A).

3.4. Humanization of the murine 3B7A antibody

To humanize the mouse antibody 3B7A, we designed 27 humanized antibodies sequences through sequence alignment and structural modeling based on CDR grafting. High-affinity antibodies T2H1/T2L1 and T2H2/T2L1 were subsequently obtained after the transient transfection of the antibody expression plasmids and screening via ELISA (Fig. S2B). In theory, a successful humanized antibody should have a humanization ratio exceeding 80% [35]. The IMGT Domain GapAlign Tool (www.imgt.org) was used to evaluate the humanization ratios of T2H1/T2L1 and T2H2/T2L1, which confirmed that both antibodies met the criteria (Fig. S2C). ELISA results further indicated that the affinity of T2H1/T2L1 was higher than that of T2H2/T2L1 after purification (Fig. S2D). Additionally, the SPR analysis demonstrated that the K_D value of T2H1/T2L1 was 2.757 nM (Fig. S2E), slightly lower than that of the parental mouse antibody 3B7A (3.96 nM) (Fig. 1G). Therefore, the humanized antibody T2H1/T2L1 was designated as h3B7A and selected for further research.

3.5. Activity comparison of h3B7A and MF-T

Anetumab ravtansine is an anti-MSLN ADC composed of the fully human mAb MF-T linked to the cytotoxic maytansinoid DM4, a potent tubulin inhibitor. It is currently under clinical evaluation [36]. To further characterize the h3B7A antibody, we compared its functional properties with those of MF-T. Flow cytometry analysis revealed that h3B7A and MF-T exhibited comparable binding affinities to native MSLN antigens expressed on AsPC-1 cells at an equivalent concentration of 10 nM (Fig. 2F). Furthermore, we detected the endocytosis of h3B7A and MF-T in AsPC-1 and SW1990 cells using flow cytometry. The results demonstrated that h3B7A had a significantly higher endocytosis rate than MF-T after incubation at 37 °C for 1, 2, 4, and 6 h (Figs. 2G and S2F). Notably, h3B7A exhibited cross-species reactivity, showing detectable binding to murine MSLN (Fig. 2H), whereas MF-T did not [37]. This suggests that h3B7A can undergo preliminary safety evaluation in mice. These findings indicate that h3B7A has advantages over MF-T, particularly in terms of internalization efficiency and species cross-reactivity.

3.6. Construction and expression of h3B7A-LDP

To prepare a novel ADC composed of the humanized antibody h3B7A and LDM, we first designed and constructed an expression vector of the h3B7A-LDP fusion protein using genetic recombination. Briefly, the LDP sequence of LDM was linked to the N-terminus of the V_L domain of the h3B7A antibody via a non-cleavable peptide linker (SGGPEGGS), generating the fusion gene sequence LDP-V_L. Subsequently, the LDP-V_L and V_H sequences of h3B7A were inserted into the expression plasmid pIZDHL, yielding the pIZDHL-h3B7A-LDP vector (Fig. S3A). Following plasmid transfection, positive clone screening, antibody expression, and purification, we successfully obtained the h3B7A-LDP fusion protein. As a parallel control, the unconjugated h3B7A antibody was also expressed. SDS-PAGE analysis revealed that the heavy chains of both h3B7A and h3B7A-LDP had molecular weights (MWs) of approximately 50 kDa, whereas their light chains exhibited MWs of 25 kDa (h3B7A) and 35 kDa (h3B7A-LDP), respectively (Fig. 3A). The increased MW of the light chain in h3B7A-LDP was consistent with the expected size, accounting for the additional LDP protein (MW: approximately 10 kDa) fused to it. These results confirmed the correct expression of h3B7A-LDP.

Fig. 3.

Specificity, internalization and in vivo tumor imaging analysis. (A) Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis of h3B7A-apoprotein of lidamycin (LDP) and h3B7A antibodies under non-reducing (left) and reducing (right) conditions. (B) Binding affinity of h3B7A-LDP and h3B7A to the human mesothelin (MSLN) antigen was measured by enzyme-linked immunosorbent assay (ELISA). The data are represented as mean ± standard deviation (SD) (n = 3). (C) Binding kinetics of h3B7A-LDP to the MSLN antigen were measured via biolayer interferometry (BLI). (D) Binding of h3B7A-LDP and h3B7A to the natural MSLN antigen on tumor cells was evaluated via flow cytometry. ‌‌(E) Binding of different concentrations of h3B7A-LDP to MSLN-positive cancer cells was assessed based on the flow cytometry analysis. (F) Internalization and lysosomal trafficking of h3B7A-LDP in H460 and SW1990 cells were observed using confocal microscopy. (G) In vivo fluorescence imaging of DyLight 680-labeled hIMB1636-LDP in H460, SW1990, and AsPC-1 xenograft mouse models. The color scale represents photon flux (photons/s/cm²/steradian). V_H: variable heavy chain; V_L: variable light chain; OD: optical density; K_D: equilibrium dissociation constant; FITC: fluorescein isothiocyanate; DAPI: 4′,6-diamidino-2-phenylindole; LAMP1: lysosomal associated membrane protein 1.

3.7. Characterization and biological activity of h3B7A-LDP

The binding of fusion protein h3B7A-LDP to MSLN antigen and cancer cells was evaluated by the methods of ELISA, BLI, and flow cytometry. As shown in Fig. 3B, a saturable dose-response curve was observed for the binding of h3B7A-LDP to MSLN antigen, which was consistent with that of the parental h3B7A in ELISA. The BLI results show that h3B7A-LDP has a strong binding affinity to MSLN antigen with a K_D value of 1.19 nM (Fig. 3C), which closely matched that of h3B7A (K_D = 1.18 nM) (Fig. S3B). Western blot analysis shows that SW1990, OVCAR3, H460, and AsPC-1 cells express high levels of MSLN, whereas H1975, Panc-1, and Panc-28 cells express low levels of MSLN (Fig. S3C). Therefore, we selected OVCAR3, H460, and SW1990 as MSLN-positive cells and H1975 as negative control cells for the subsequent flow cytometry analysis. The results demonstrated that the shifts in fluorescence values of h3B7A-LDP and h3B7A in these cells were similar (Fig. 3D), which indicates that fusion of LDP moiety did not adversely affect the binding activity of the parent antibody h3B7A. Then, we evaluated the binding specificity of h3B7A-LDP to MSLN-positive cells. As expected, a concentration-dependent shift in fluorescence intensity was observed for all cells, and the fluorescence shift at 10 μg/mL h3B7A-LDP was significantly greater than at 0.1 μg/mL (Fig. 3E).

Furthermore, we detected the internalization ability of h3B7A-LDP fusion protein in MSLN⁺ cancer cells by confocal microscopy. After incubation at 4 °C, a strong green fluorescence signal (h3B7A-LDP) was observed on the surface of H460 and SW1990 cells. However, after transfer to 37 °C and 2 h incubation, the membrane signal decreased markedly, while cytoplasmic fluorescence intensified. Some signals colocalized with lysosomes (yellow fluorescence from merged green h3B7A and red LAMP-1 staining) (Fig. 3F). These results indicate that h3B7A-LDP can be internalized and trafficked to lysosomes by MSLN-positive cancer cells.

The in vivo tumor targeting of h3B7A-LDP was assessed in an H460 xenograft model. An obvious fluorescence signal was detected at the tumor site in mice 6 h after drug injection, and the tumor retention time of h3B7A-LDP exceeded 11 days (Fig. 3G). Similar results were observed in SW1990 and AsPC-1 xenograft models (Fig. 3G). At the end of the experiments, fluorescence signals were detectable only in tumor tissue, with no visible signals in heart, liver, lungs, spleen, or kidneys (Fig. S3D).

3.8. Assembly of h3B7A-LDM

We have demonstrated the potent affinity, specificity, internalization, and in vivo tumor targeting of h3B7A-LDP, which are essential for its function in cancer therapy. However, h3B7A-LDP requires a potent cytotoxic payload to directly destroy cancer cells. As shown in Fig. 4A, the fusion protein h3B7A-LDM, composed of one antibody and two intact LDM molecules with a uniform drug-to-antibody ratio (DAR) value of 2, was prepared. First, we separated the LDM molecule into LDP protein and AE via HPLC and collected the AE (Fig. S4A). Next, AE and h3B7A-LDP were assembled at 4 °C overnight to obtain h3B7A-LDM, followed by purification and subsequent analysis. The C4-300A chromatogram results revealed a characteristic absorption peak for AE at 340 nm (Fig. 4B), indicating that AE properly fits into the hydrophobic pocket formed by the LDP protein in the h3B7A-LDP molecule. Furthermore, the assembly operation did not compromise antibody activity, as confirmed by flow cytometry, since h3B7A-LDM exhibited nearly identical binding activity to h3B7A-LDP in MSLN-positive tumor cell lines (Fig. 4C).

Fig. 4.

In vitro antitumor effects of h3B7A-lidamycin (LDM). (A) Assembly diagram of h3B7A-LDM. (B) ‌High performance liquid chromatography (HPLC) analysis of h3B7A-LDM using a Delta-Pak C4-300 Å column at 340 nm. (C) Binding ability of h3B7A-LDM and h3B7A-apoprotein of LDM (LDP) to mesothelin (MSLN)-positive tumor cells was analyzed by flow cytometry. Data represent mean ± standard deviation (SD) (n = 3). ‌‌(D) Cytotoxicity of h3B7A-LDM and LDM against cancer cell lines was determined by Cell Counting Kit-8 (CCK-8) assay. Data represent mean ± SD (n = 3). (E, F) Proliferation (E) and migration (F) of OVCAR3, H460, SW1990, and AsPC-1 cells treated with h3B7A-LDM (indicated concentrations) were monitored in real time using the xCELLigence system. AE: active enediyne chromophore of LDM; V_L: variable light chain; V_H: variable heavy chain; UV-Vis-WVL: ultraviolet-visible spectroscopy-wavelength; FITC: fluorescein isothiocyanate.

3.9. In vitro antitumor activity of h3B7A-LDM

We evaluated the antitumor efficacy of h3B7A-LDM in vitro using MSLN-positive cancer cells, with naked LDM as the positive control. Both compounds displayed potent cytotoxic activity against OVCAR3, H460, SW1990, and AsPC-1 cells. The half maximal inhibitory concentration (IC₅₀) values of h3B7A-LDM were 0.05 ± 0.01, 0.43 ± 0.01, 0.51 ± 0.04, and 0.16 ± 0.01 nM, while those of LDM were 0.11 ± 0.02, 0.24 ± 0.03, 0.26 ± 0.09, and 0.10 ± 0.02 nM, respectively (Fig. 4D). The cytotoxicity of unintegrated fusion protein h3B7A-LDP was also assessed, but no significant inhibitory effects were observed in these cells (Fig. S4B). This suggests that the cytotoxicity of h3B7A-LDM is primarily mediated by the AE moiety. Subsequently, we investigated the effects of h3B7A-LDM on the proliferation and migration of MSLN-positive tumor cell lines. As shown in Figs. 4E and F, compared to the control, 1 nM h3B7A-LDM inhibited cell proliferation by 87.7% ± 1.0%, 78.4% ± 2.6%, 60.2% ± 2.0%, and 84.6% ± 0.2% in OVCAR3, H460, SW1990, and AsPC-1 cells, respectively, at 72 h (Fig. 4E). Moreover, 0.3 nM h3B7A-LDM suppressed cell migration by 47.0% ± 1.3%, 37.0% ± 0.1%, 31.7% ± 1.3%, and 32.0% ± 1.3%, respectively, at 36 or 48 h (Fig. 4F). In summary, h3B7A-LDM demonstrated significant, dose-dependent inhibition of cancer cell proliferation and migration.

3.10. h3B7A-LDM induces cell cycle arrest and apoptosis

Cell cycle arrest and apoptosis induction represent two pivotal mechanisms contributing to regulated cell death. First, we performed flow cytometry to detect the effects of h3B7A-LDM on cell cycle kinetics and apoptosis in MSLN-positive cells. After the treatment with different concentrations of h3B7A-LDM, a significant decrease in G1-or S-phase DNA content and an increase in G2 DNA content were observed in H460, AsPC-1, and SW1990 cells (Figs. 5A and S5A). Furthermore, Western blot results revealed that the expression levels of cyclin D1 and CDK6 decreased, whereas those of P21 increased (Fig. 5B). In addition, compared with the control, h3B7A-LDM induced early and late apoptotic events in a dose-dependent manner. Treatment with 30 nM h3B7A-LDM increased the percentage of apoptotic H460 and AsPC-1 cells to 27.02% ± 2.02% and 47.70% ± 3.27%, respectively (Fig. 5C). h3B7A-LDM also enhanced the apoptosis rate of SW1990 cell in a concentration-dependent manner (Fig. S5B). These results were further confirmed by Western blot analysis, which showed increased levels of proapoptotic proteins, including cleaved caspase-3, -7, and -9 (Fig. 5D). Collectively, these results indicate that h3B7A-LDM mediates its antitumor activity through the induction of cell cycle arrest and caspase-dependent apoptotic pathways.

Fig. 5.

h3B7A-lidamycin (LDM) induces cell cycle arrest and apoptosis. (A) Representative cell cycle histograms and quantitative measurement of the cell cycle phase distribution in H460 and AsPC-1 cells after the treatment with the indicated concentrations of h3B7A-LDM for 48 h. Data are presented as mean ± standard deviation (SD) (n = 3). (B) Western blot analysis of cell cycle-related protein expression in H460 and AsPC-1 cells. Protein levels were quantified and normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Data represent mean ± SD (n = 3). (C) Representative flow cytometry and quantitative analysis of apoptosis in H460 and AsPC-1 cells after treatment with the indicated concentrations of h3B7A-LDM for 48 h. ‌‌(D) Western blot analysis of apoptosis-associated protein expression in H460 and AsPC-1 cells. Protein levels were quantified and normalized to GAPDH. Data represent mean ± SD (n = 3). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001 vs. the control. PI: propidium iodide; CDK6: cyclin-dependent kinase 6; CyclinD1: G1/S-specific cyclin-D1; FITC: fluorescein isothiocyanate.

3.11. h3B7A-LDM and LDM induced the ICD response and activated immune cells

Some chemotherapy drugs, such as anthracyclines and oxaliplatin, can induce ICD. In other words, they can convert dying tumor cells into a therapeutic vaccine, activate the immune system by blocking various immuno-suppressive tumor networks, and promote an antitumor immune response in the body [38]. ICD is distinguished by the release of damage-associated molecular patterns (DAMPs), such as the translocation of CRT to the cell surface, the secretion of ATP, and released HMGB1. To explore whether h3B7A-LDM can induce ICD in the cancer cells, we detected the emission of key DAMPs in H460, SW1990, and AsPC-1 cells. Flow cytometry results show that, compared with the control group, cell-surface CRT expression increased in a time-dependent manner in these cells after h3B7A-LDM or LDM treatment (Figs. 6A, 6B, and S6A). Next, we found that h3B7A-LDM or LDM treatment upregulated the extracellular level of ATP in cancer cells, indicating that both compounds induced ATP secretion (Figs. 6C and S6B). Finally, extracellular HMGB1 was increased in the cell-free supernatant of cells treated with h3B7A-LDM or free LDM (Figs. 6D and S6C). Overall, these results demonstrated that h3B7A-LDM or LDM treatment triggers the presentation of three hallmarks of ICD, suggesting their potential to induce ICD.

Fig. 6.

h3B7A-lidamycin (LDM) and LDM induce immunogenic cell death (ICD) in vitro and activate host immune cells in vivo. (A, B) Cell-surface expression of calreticulin‌‌ (CRT) was measured via flow cytometry in H460 (A) and SW1990 (B) cells after 48 h of treatment with h3B7A-LDM or LDM. (C, D) Extracellular adenosine triphosphate (ATP) (C) and recombinant high mobility group protein 1 (HMGB1) (D) levels were measured in H460 or SW1990 cell-free supernatants following 48 h of treatment with h3B7A-LDM or LDM. Data are presented as mean ± standard deviation (SD) (n = 3). (E) Proportions of CD11c⁺ dendritic cells (DCs) and Ly6G⁺ cells in H460 xenografts after the administration of h3B7A-LDM or LDM. Immune cell levels were detected by flow cytometry. Data represent mean ± SD (n = 5). (F) Intratumoral proinflammatory cytokine levels in H460 xenograft-bearing mice after injection of h3B7A-LDM or LDM. Cytokines were measured using enzyme-linked immunosorbent assay (ELISA). Data represent mean ± SD (n = 5). (G) Increased levels of murine serum chemokines and cytokines in H460 xenograft-bearing mice after administration of h3B7A-LDM or LDM. Cytokines were measured via ELISA. Data represent mean ± SD (n = 5). (H) Schematic representation of the in vivo vaccination assay. (I) Tumor formation in mice after vaccine intervention by subcutaneous rechallenge with growing Lewis lung carcinoma (LLC) cells (n = 6). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001 vs. control. ns: not significant. FITC: fluorescein isothiocyanate; CD11c: integrin alpha-X; Ly6g: lymphocyte antigen 6G. IP10/CXCL10: C−X−C motif chemokine ligand 10; MIP-1α: macrophage inflammatory protein-1 alpha; TNF-α: tumor necrosis factor-alpha.

Furthermore, we investigated whether the dying cells could activate innate and adaptive immune responses in vivo. To this end, the host innate immune responses to h3B7A-LDM- or LDM-treated tumors were explored in vivo using an H460 cell-derived xenograft model in severe combined immune deficiency (SCID) mice. SCID mice lack T and B cells but retain innate immune cells such as dendritic cells (DCs) and macrophages, making them suitable for evaluating the activation and recruitment of antigen-presenting cells. ELISA results showed that in tumors treated with h3B7A-LDM or LDM, the proportion of antigen-presenting CD11c⁺ DCs increased, while that of Ly6G⁺ neutrophils decreased (Fig. 6E). This suggested that the treatment reshaped the immune system by preferentially promoting the differentiation of innate immune cells toward mature DCs (CD11c⁺ DCs). Consequently, DCs were either recruited into the tumor or infiltrating monocytes differentiated into DCs. In addition, h3B7A-LDM or LDM significantly or trend-wise increased proinflammatory cytokines, including IP10, MIP-1α, and TNF-α in both tumors and serum of mice (Figs. 6F and G). Taken together, the in vitro release of key DAMPs from cancer cells, the in vivo recruitment of murine DCs, and the release of proinflammatory cytokines following h3B7A-LDM or LDM treatment were consistent with the ICD induction. These findings suggest that LDM-based drugs have the potential to induce an ICD response and activate immune cells to further kill cancer cells.

Vaccination trials are the gold standard for detecting ICD. Given that the cytotoxicity of h3B7A-LDM primarily originates from LDM component, we performed a vaccination assay using LDM directly to assess its effects on ICD induction (Fig. 6H). The results demonstrated that vaccination with LDM-treated dying cells significantly reduced the tumor formation rate (1/6 vs. 6/6 in controls) and suppressed tumor growth (Figs. 6I and S6D), confirming that LDM treatment elicits ICD in vivo and inhibits tumorigenesis. These findings further support the conclusion that h3B7A-LDM induces an ICD response, contributing to its anti-tumor effects.

3.12. h3B7A-LDM inhibits tumor growth in vivo

Finally, we evaluated the in vivo antitumor effects of h3B7A-LDM using subcutaneous xenograft models. In the H460 model, h3B7A-LDM displayed a concentration-dependent inhibition of tumor growth. At 0.8 mg/kg, h3B7A-LDM achieved a 69.8% ± 14.4% tumor inhibition rate, significantly higher than the inhibition rate of free LDM (55.4% ± 12.1%) at its maximally tolerated dose (0.045 mg/kg); free antibody h3B7A at 0.8 mg/kg had no antitumor effects, and the tumor volumes were similar to those of the control group (Figs. 7A and S7A). Except for significant weight loss in the LDM-treated group (P < 0.01 vs. h3B7A-LDM at 0.8 mg/kg, Fig. 7A), no deaths or other significant adverse effects were observed in all groups. Further, we explored the therapeutic potential of 0.8 mg/kg h3B7A-LDM in other MSLN⁺ xenograft models. As expected, h3B7A-LDM treatment significantly suppressed tumor growth and reduced tumor weight/size compared with the control, with tumor inhibition rates of 66.3% ± 11.1% and 62.0% ± 5.3% in the SW1990 and AsPC-1 xenograft models, respectively (Figs. 7B, 7C, and S7A). Similarly, the LDM treatment induced weight loss in the SW1990 xenograft model mice (Fig. 7B). However, no significant toxico-pathological changes were observed in major organs (heart, liver, spleen, lungs, kidneys, bone marrow, and small intestine) of H460 or SW1990 xenograft mice after treatment (Figs. 7D and S7B), indicating minimal toxicity at the administered doses. In the H460 model, 0.8 mg/kg h3B7A-LDM did not severely affect hematological parameters (hemoglobin, leukocytes, platelets, or neutrophils), suggesting minimal myelotoxicity (Fig. 7E).

Fig. 7.

Therapeutic efficacy of h3B7A-lidamycin (LDM) against mesothelin (MSLN)⁺ tumor cells in vivo. (A–C) Anti tumor activity in stenography models: tumor growth curves, final tumor weights, and body weight changes in mice bearing H460 (A), SW1990 (B), or AsPC-1 (C) stenographer following h3B7A-LDM treatment. (D) Hematoxylin and eosin (H&E) staining of multiple organs and tumor tissues from H460 stenography mice following treatment with h3B7A-LDM (0.8 mg/kg). (E) Hematological profiles of mice following various treatment regimens. Data are represented as mean ± standard deviation (SD) (n = 6). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001 vs. control.

Finally, considering that H460 cells are MSLN-positive but Trop2-negative, we used anti-Trop2 ADC hIMB1636-LDM as a non-targeting IgG-LDM control in H460 xenograft model to verify the MSLN-dependent efficacy of h3B7A-LDM and rule out its off-target cytotoxicity. The results revealed that h3B7A-LDM at 0.8 mg/kg significantly inhibited tumor growth with an inhibition rate of 72.7% (P < 0.001 vs. control), while hIMB1636-LDM at the same dose exhibited virtually no inhibitory effect (P > 0.05 vs. control) (Fig. S8A). No deaths or other significant adverse effects were observed in all groups (Fig. S8B).

4. Discussion

In this study, we first developed a novel MSLN-targeting IgG1 mAb, named 3B7A, using the traditional hybridoma technology. 3B7A binds to the recombinant MSLN antigen with a K_D value of 3.96 nM (Fig. 1G), and specifically binds to natural MSLN in cancer cell lines and tumor tissues, with little or no binding to normal tissues (Fig. 2C). As an essential step for ADCs to enter tumor cells, endocytosis facilitates the delivery of cytotoxic payloads to their intracellular targets and plays a critical role in the antitumor effects of ADCs [39]. We found that 3B7A binds to cancer cells and is rapidly internalized into lysosomes (Fig. 2D). Furthermore, in in vivo imaging experiments, after intravenous injection via the tail vein, 3B7A rapidly accumulates at the tumor site in mice, persisting for more than 11 days (Fig. 2E). These attributes of 3B7A, including specific and high affinity toward antigen-positive cancer cells, efficient internalization, specific targeting, and prolonged retention within tumors in vivo, make it a promising antibody carrier for ADC design. Next, we successfully humanized the 3B7A antibody by CDR grafting, and the resulting antibody, h3B7A, has a K_D value of 2.757 nM (Fig. S2E). Notably, our findings demonstrate that, compared to MF-T, h3B7A exhibits not only comparable binding activity to cancer cells but also superior functional properties, including more potent endocytosis activity and the ability to directly bind to mouse MSLN (Figs. 2G and H).

MSLN-targeted ADCs have emerged as a promising strategy for treating MSLN-positive cancers, though no related product has been approved for clinical use. As mentioned earlier, most anti-MSLN ADCs in clinical trials employ tubulin inhibitors as cytotoxic payloads. However, these inhibitors exhibit limited efficacy against quiescent cancer cells, often leading to tumor recurrence and metastasis. To address this limitation, novel ADC payloads with improved safety and efficacy, particularly those leveraging mechanisms of action distinct from current clinical ADCs, must be developed. LDM, a DNA-damaging enediyne antibiotic, demonstrates potent antitumor activity with unique characteristics: the special mechanism of action of binding to the minor groove of the DNA double helix, preferentially targeting of hypoxic tumor cells, ability to overcome radioresistance in malignant cells, and efficacy against multidrug-resistant cancer cells [40]. Moreover, the approval of besponsa and mylotarg, which use enediyne-containing calicheamicin as payload, also highlighted the potential clinical value of enediyne antibiotics. In addition, an ideal ADC payload should combine high cytotoxicity with low immunogenicity [25], notably, phase-I trials of LDM reported no immunogenicity concerns. Collectively, these data support LDM as a compelling candidate for ADC development.

Using h3B7A as a tumor-targeting carrier and LDM as the payload, we generated the MSLN-directed ADC h3B7A-LDM via a two-step method based on the distinctive attribute of LDM, which can be disassembled and reassembled in vitro [41]. First, the fusion protein h3B7A-LDP was generated by fusing the antibody h3B7A with the LDP moiety of LDM via gene recombination. BLI analysis showed that the K_D value of h3B7A-LDP was 1.19 nM, similar to that of the naked antibody h3B7A (1.18 nM) (Figs. 3C and S3B). Notably, h3B7A-LDP and h3B7A exhibited nearly identical binding affinities toward various cancer cell lines (Fig. 3D). Confocal microscopy further confirmed that h3B7A-LDP was effectively internalized by MSLN-positive cancer cells and subsequently trafficked to lysosomes (Fig. 3F). Consistent with these findings, in vivo imaging demonstrated specific and prolonged accumulation of h3B7A-LDP at tumor sites, with detectable signals persisting for over 11 days (Fig. 3G). Collectively, these data indicate that fusion of LDP with h3B7A had almost no adverse effect on the biological activities of the parent antibody. Second, we obtained the ADC h3B7A-LDM through molecular assembly. Unlike the traditional nonspecific conjugation, which produces mixed ADC molecules containing antibodies conjugated with varying numbers of payloads, potentially leading to reduced efficacy and unexpected side effects, the h3B7A-LDM generated via this two-step process is homogeneous, well-defined, and has a DAR value of 2. Additionally, we fused LDP with the N-terminus of the V_L rather than C-terminus of antibody Fc domain, to avoid compromising the ADCC effect of the h3B7A antibody. This site-specific construct preserved the parental antibody's antigen affinity while ensuring ADC stability and enhanced binding efficiency.

Both h3B7A-LDM and LDM displayed potent cytotoxic effects against cancer cells with IC₅₀ values in the sub-nanomolar range, whereas h3B7A-LDP had no effect on tumor cell survival. This suggests that the enediyne chromophore AE is the main cytotoxic moiety responsible for killing tumor cells in h3B7A-LDM (Figs. 4D and S4B). In addition, h3B7A-LDM inhibited the proliferation and migration of cancer cells (Figs. 4E and F). Moreover, we observed that h3B7A-LDM induced G2/M phase cell cycle arrest and significantly altered the expression levels of key cyclins, including the downregulation of cyclin D1 and CDK6 and the upregulation of P21 (Figs. 5A and B). Furthermore, h3B7A-LDM triggered apoptosis in cancer cells through a caspase-dependent pathway, as confirmed by the increased expression of the proapoptotic proteins such as cleaved caspase-3, -7, and -9 in Western blot assays (Figs. 5C and D).

ICD is a distinct form of apoptotic cell death characterized by the expression and release of DAMPs and proinflammatory cytokines. ICD can elicit an immune response at the tumor site, turn immunosuppressive “cold tumors” into “hot tumors”, and is considered an effective strategy to restore anti-tumor immune responses [42]. It facilitates the tumor antigen presentation in dying tumor cells via the enhanced expression of DAMPs, such as CRT, ATP, and HMGB-1, which promotes the activation of DCs to engulf dying tumor cells. This process further recruits activated cytotoxic T lymphocytes (CTLs) to the tumor microenvironment, where they destroy cancer cells by releasing immune-promoting cytokines. Consequently, ICD enhances the long-term efficacy of anticancer agents by directly killing cancer cells and activating antitumor immunity [43]. Some ADCs have been reported as ICD inducers, such as brentuximab vedotin [44] and D3-GPC2-PBD [45]. Therefore, we evaluated whether h3B7A-LDM treatment of MSLN⁺ tumor cells could induce ICD. We found that h3B7A-LDM induced several hallmarks of ICD, including cell-surface expression of CRT, secretion of ATP, and extracellular release of HMGB1 in MSLN-positive cells (Figs. 6A–D). In the H460 xenograft model, h3B7A-LDM treatment led to the recruitment of DCs into tumors (Fig. 6E) and increased levels of proinflammatory cytokines such as MIP-1α, IP10/CXCL10, and TNF-α in both serum and tumor tissues (Figs. 6F and G). These findings align with ICD-driven activation of innate immunity, effectively stimulating the host immune system to mount a long-lasting antitumor adaptive immune response. Furthermore, the in vivo vaccination experiment confirmed that LDM-based ADCs can induce an ICD response (Fig. 6I). Our findings reveal, for the first time, that LDM-based ADCs may act as potential ICD inducers and promote antitumor immunity. Taken together, these findings reveal that h3B7A-LDM exerts multiple anticancer mechanisms of action beyond general proliferation inhibition, migration suppression, and the induction of apoptosis and cell cycle arrest.

The results of animal experiments further verified that h3B7A-LDM significantly enhanced antitumor efficacy (inhibition rate: 69.8% ± 14.4% vs. 55.4% ± 12.1%) compared with free LDM alone at the maximal tolerated dose of 0.045 mg/kg (Fig. 7A). Additionally, h3B7A-LDM exhibited a favorable safety profile in histopathological examinations, as no organ lesions were observed in mice treated with therapeutic doses (Fig. 7D). Although myelosuppression is the most common and serious adverse effect for most of the ADCs currently used in clinics due to the cytotoxic payload [46,47], no obvious myelotoxicity was detected in h3B7A-LDM-treated mice (Fig. 7E). Notably, fusing LDM with the h3B7A mAb improved the tolerability and tumor-therapeutic efficacy of LDM in nude mice, suggesting that our strategy of conjugating h3B7A with LDM is feasible and successful. Thus, h3B7A-LDM may be a potential therapeutic candidate for MSLN⁺ tumors. However, despite demonstrating superior antitumor efficacy compared to the maximal tolerated dose of LDM, h3B7A-LDM still exhibited relatively modest performance when compared to most of other ADCs reported in the literature [48,49]. This limitation highlights the need for further optimization of this approach. Future studies should focus on two potential strategies to enhance therapeutic efficacy: i) dose escalation studies to determine whether higher doses of h3B7A-LDM can improve efficacy while maintaining an acceptable safety profile, and ii) exploration of combination regimens with established clinical antitumor agents to potentially achieve synergistic effects. On the other hand, to establish the relative merits of LDM over clinically used ADC payloads such as MMAE, a direct head-to-head comparison between h3B7A-LDM and h3B7A-MMAE is necessary. Unfortunately, due to the presence of aggregates in the purified h3B7A, our attempts to prepare h3B7A-MMAE have thus far been unsuccessful. We will therefore focus on optimizing the h3B7A purification process to enable the synthesis of h3B7A-MMAE, which will allow a direct comparative evaluation and ultimately demonstrate the advantages of LDM as a payload.

5. Conclusion

In this study, we generated a novel humanized anti-MSLN mAb, h3B7A, with properties suitable for ADC development, including high and specific binding to cancer cells and tissues, efficient internalization by cancer cells, and effective in vivo tumor targeting with long-term persistence. We then designed and developed a novel ADC, h3B7A-LDM, by conjugating LDM to h3B7A via a peptide linker using genetic engineering and molecular assembly techniques. h3B7A-LDM exhibited potent antitumor activity both in vitro and in vivo through multiple mechanisms, including the proliferation and migration inhibition, induction of apoptosis and cell cycle arrest, and activation of the host immune response. The preclinical data from this study suggest that h3B7A-LDM could be a promising therapeutic candidate for treating MSLN-positive solid tumors.

CRediT authorship contribution statement

Dan-Dan Zhou: Writing – original draft, Validation, Software, Resources, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Zi-Hui Xie: Resources, Methodology, Formal analysis, Data curation. Ai-Jun Duan: Methodology, Formal analysis, Data curation. Shi-Yu Zhu: Methodology, Formal analysis. Ying Wang: Resources, Methodology. Yong-Su Zhen: Supervision, Conceptualization. Rui-Juan Gao: Supervision, Funding acquisition. Qing-Fang Miao: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is the Supplementary data to this article: