Insights Into Genomic Drivers, Transcriptomic Heterogeneity, and Therapeutic Vulnerabilities From Novel Mesothelioma Cell Lines

ABSTRACT

Background

Pleural mesothelioma (PM) is an aggressive cancer with limited therapeutic options and poor prognosis, necessitating comprehensive model systems for mechanistic studies and drug discovery.

Methods

Tumor cells were isolated from malignant pleural effusion of PM patients and subjected to primary culture and continuous passaging. Cell lines were successfully established after more than 40 continuous passages in vitro. Cell identification included morphological analysis, species identification, short tandem repeat (STR) profiling, mycoplasma detection, in vitro proliferation assays, in vivo tumorigenicity testing in NOD/SCID mice, and immunohistochemical characterization. Genomic landscapes and transcriptomic profiling were defined by whole‐exome, whole‐genome sequencing, and RNA‐sequencing.

Results

Three Chinese‐derived pleural mesothelioma cell lines (PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4) were successfully established. All three cell lines exhibited epithelioid morphology with adherent growth patterns. Upon subcutaneous transplantation into NOD/SCID mice, PUMC‐MESO1 and PUMC‐MESO4 demonstrated tumorigenicity, while PUMC‐MESO3 remained non‐tumorigenic after 3 months of observation. Histopathological examination confirmed their mesothelioma origin through positive staining for mesothelial markers including calretinin, WT‐1, and D2‐40. Genomic analysis revealed characteristic PM genomic alterations including mutations in BAP1, NF2, and TP53, along with CDKN2A deletions. Transcriptomic analyses revealed heterogeneous molecular features among the three cell lines. Although the PUMC‐MESO cell lines displayed distinct clustering patterns relative to the CCLE mesothelioma cohort, each retained transcriptomic similarities to specific established mesothelioma models. Drug sensitivity assays further demonstrated heterogeneous responses to standard therapeutic agents.

Conclusion

This study reports the establishment and characterization of three novel PM cell lines. These models recapitulate key aspects of PM biology, exhibit diverse therapeutic responses, and provide a valuable new resource for investigating disease mechanisms and advancing precision oncology research.

Three novel malignant pleural mesothelioma (MPM) cell lines (PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4) were established from patient‐derived pleural effusions. All three cell lines expressed mesothelioma markers (calretinin, WT1, and D2‐40) and showed BAP1 loss of expression. Integrated genomic and transcriptomic analyses revealed distinct mutational profiles and heterogeneous molecular features. Drug sensitivity profiling identified differential responses to pemetrexed and cisplatin, providing new tools for MPM preclinical research.

1. Introduction

Pleural mesothelioma (PM) is an aggressive malignancy originating from the pleural mesothelial cells, characterized by significant molecular heterogeneity that influences prognosis and therapeutic response [1]. Despite advances in multi‐omics profiling of clinical cohorts, a critical gap persists in the availability of well‐characterized in vitro models that reflect this diversity. A substantial portion of existing PM cell lines originates from Caucasian patients, potentially limiting their representation of the full spectrum of biological and genomic subtypes observed globally. This limitation hinders preclinical research and the development of broadly applicable therapeutic strategies.

The establishment of novel, patient‐derived cell lines is fundamental for advancing cancer research. Such models serve as renewable resources for investigating disease mechanisms, exploring drug sensitivity, and identifying novel biomarkers [2]. For PM, key tumor suppressor genes such as BAP1, TP53, CDKN2A, and NF2 are frequently altered, yet the interplay of these mutations with other genomic events and their impact on the tumor phenotype is not fully understood. Furthermore, while immune checkpoint inhibitors have shown promise, their efficacy in PM is limited, and reliable biomarkers beyond PD‐L1 expression are urgently needed [3, 4]. Characterizing the genomic and transcriptomic landscapes of new cell lines is essential to position them within the known context of PM and to uncover unique features relevant to therapy response and immune modulation.

This study reports the establishment and comprehensive characterization of three novel PM cell lines derived from Chinese patients. Through an integrated multi‐omics approach, these cell lines were systematically analyzed to define their phenotypic, genomic, and transcriptomic features. The primary objectives were to validate their mesothelial origin, to map their genomic landscapes in the context of known PM driver genes, to determine their drug sensitivity profiles to standard‐of‐care chemotherapies, and to position their transcriptomic signatures relative to the primary tumors and existing cell line models. This work provides a valuable new resource for the global cancer research community, offering unique models for investigating PM biology and developing novel therapeutic interventions.

2. Material and Methods

2.1. Main Reagents

DMEM/F12 medium, 0.05% trypsin, and fetal bovine serum (FBS) were obtained from the National Science and Technology Infrastructure (NSTI‐BMCR). The following reagents were also used for cell culture: penicillin–streptomycin (PS, #15140122, Gibco, USA), insulin‐transferrin‐selenium (ITS, #41400‐045, Gibco, USA), human epidermal growth factor (hEGF, #E9644, Sigma‐Aldrich, USA), and Y‐27632 (#M1817, AbMole, USA). The Cell Counting Kit‐8 (CCK8, #HD666, Tokyo, Japan) was from Dojindo Molecular Technologies. Working solutions for immunohistochemistry (IHC) staining included primary antibodies against calretinin (POLY, ZSGB‐BIO), WT‐1 (OTIR1F1, ZSGB‐BIO), D2‐40 (D2‐40, GeneTech), BAP1 (C‐4, GeneTech), P40 (GR006, GeneTech), Ber‐EP4 (Ber‐EP4, GeneTech), vimentin (OTIR5D6, ZSGB‐BIO), cytokeratin (AE1/AE3, ZSGB‐BIO), and CK5/6 (OTI1F8, ZSGB‐BIO). Pemetrexed (S5971), gemcitabine (S1714), and cisplatin (S1166) were from Selleckchem (Houston, USA).

2.2. Clinical Samples and Cell Line Establishment

The study was approved by the Ethics Committee of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Approval No. 2022129). Pleural effusions were collected from eight patients diagnosed with PM, from whom written informed consent had been obtained. Cells were isolated by centrifugation and cultured for primary and subsequent passages. A cell line was considered successfully established after more than 20 continuous passages in vitro.

Cells were cultured in DMEM/F12 (1:1) mixed medium supplemented with 10% FBS, 1 × ITS, 40 ng/mL EGF, and 20 μM Y27632. The PM cells were seeded in culture flasks and maintained in a humidified incubator at 37°C with 5% CO₂. The medium was changed every 48 h. Upon reaching approximately 90% confluency, the cells were passaged. Detachment was achieved using 0.05% trypsin, and the digestion was stopped with complete medium to create a single‐cell suspension. The cells were then split at a 1:2 ratio for subculture. Throughout the culture process, repeated sample collection was performed for cell species identification, short tandem repeat (STR) analysis, and mycoplasma detection, ensuring no contamination from other cell lines or mycoplasma.

2.3. Cell Proliferation Assay

PM cells were seeded in six‐well plates at a density of 1.5 × 10⁵ cells or 2.5 × 10⁵ cells per well, with four plates prepared in parallel. Each day for 7 consecutive days, cells from three replicate wells were harvested and counted. A growth curve was plotted using Prism 10 software. The cell doubling time was calculated using an online tool (http://www.doubling‐time.com/compute.php) with the formula:

$$\begin{array}{c}\text{Doubling Time}=\left(\text{Duration}\times \log \left(2\right)\right)\hfill \\ /\left(\log \left(\text{Final Concentration}\right)-\log \left(\text{Initial Concentration}\right)\right).\hfill \end{array}$$

2.4. Species Identification and STR Profiling

Approximately 2.0 × 10⁶ cells were collected, washed twice with PBS, and pelleted. Genomic DNA was extracted from the cells. PCR with species‐specific primers was performed to confirm the species of origin [5]. STR profiling was outsourced to the BGI‐write Co. Ltd. (Beijing, China).

2.5. In Vivo Tumorigenicity and Histology

Five 6‐week‐old male NOD/SCID mice (the Vital RiverLaboratory Animal Technology Co. Ltd., Beijing, China) were maintained under specific pathogen‐free (SPF) conditions. Each mouse was injected subcutaneously in the right axilla with 1.0 × 10⁷ PM cells. Mice were euthanized when tumor diameter reached approximately 1.5 cm. The tumor tissue was collected, fixed in 4% paraformaldehyde, and processed for dehydration, clearing, and wax embedding. Sections were then prepared for hematoxylin and eosin (H&E) staining and IHC analysis. All procedures were conducted in accordance with institutional and national ethical guidelines. The animal study protocol was approved by the Animal Ethics Committee of the Institute of Basic Medical Sciences, CAMS, and performed according to the China Regulations on the Administration of Laboratory Animals (Animal Ethics No. ACUC‐A02‐2021‐041).

IHC staining was performed automatically using either the Ventana Benchmark Ultra (Roche Diagnostics) or the BOND‐III (Leica Biosystems) platform. Staining results were independently evaluated by two pathologists. Representative images were acquired using the P250 FLASH scanning system and analyzed using the CaseViewer software (3DHistech Ltd., Beijing, China).

2.6. Cell Block Preparation

For PM cells that did not form subcutaneous tumors, 1.0 × 10⁷ cells were collected by centrifugation at 1000 rpm. After washing with PBS, the cells were fixed in 95% ethanol for 1 h. The suspension was centrifuged at 2500 rpm for 10 min. The supernatant was discarded, leaving 1–2 mL of liquid containing the cell pellet. This cell suspension was transferred to a special embedding cassette tube and placed in a preheated micro‐specimen/cell block centrifuge embedding machine (Beijing Saipu Jiuzhou Technology Development Co. Ltd.) at 70°C [6]. The sample was centrifuged at 2000 rpm for 5 min. Following centrifugation, the supernatant was removed, and 0.2–0.4 mL of molten fixing glue was added and mixed thoroughly with the cells. The mixture was then centrifuged for embedding (2000 rpm, 5 min, 60°C). The embedding cassette was removed and cooled at room temperature for 10 min. Subsequent wax immersion, sectioning, and staining were performed as described previously.

2.7. Drug Sensitivity Assay and Synergy Analysis

Cells were seeded in 96‐well plates at 3000 cells per well and cultured as described above. After a 24‐h incubation at 37°C and 5% CO₂ to allow for cell adherence, the medium was replaced with complete culture medium containing gradient concentrations of pemetrexed, cisplatin, or gemcitabine (0, 1, 100, 1000, 10 000, 100 000, and 1 000 000 nM). After 72 h of incubation, the medium was removed. A mixture of 10 μL of CCK‐8 reagent and 100 μL of culture medium was added to each well, followed by a 1‐h incubation in the dark. The optical density at 450 nm was measured using a microplate reader. Survival rates and IC₅₀ values were calculated using Prism 10 software.

To evaluate the interaction between pemetrexed and cisplatin, a drug combination concentration matrix was designed. The concentration gradients were 0, 2500, 5000, 10 000, and 20 000 nM for pemetrexed and 0, 500, 1000, 2000, and 4000 nM for cisplatin. Cells were seeded at the same density in 96‐well plates. After 24 h, the medium was replaced with medium containing the different drug concentration combinations. The matrix covered the effective concentration ranges of both drugs and included single‐drug gradients and blank controls. After a 72‐h culture period, the cell survival rate of each well was determined as described above. The expected drug combination responses were calculated based on zero interaction potency (ZIP) reference model using SynergyFinder [7]. Deviations between observed and expected responses with positive and negative values denote synergy and antagonism, respectively.

2.8. Multi‐Omics Data Generation and Processing

Multi‐omics data generation was performed by Novogene Co. Ltd. (Beijing, China) and JMDNA Co. Ltd. (Shanghai, China).

Whole‐exome sequencing (WES) and whole genome sequencing (WGS): Raw sequencing data were quality‐controlled with FastQC (v0.23.1). After filtering, reads were aligned to the human reference genome hg38 using BWA (v0.7.17). Somatic SNVs, InDels, CNVs, and SVs were detected using Mutect, Strelka, Control‐FREEC, and Lumpy, respectively.

RNA‐sequencing (RNA‐seq): Gene expression was quantified using featureCounts (v2.0.6) based on GENCODE v36 annotation to obtain a raw count matrix [8].

2.9. Bioinformatics Analysis

2.9.1. Reference Cohort Acquisition and Processing

TCGA‐MESO Cohort [9]: Public CNV data (hg19) from the TCGA mesothelioma cohort were downloaded. Tools like rtracklayer (v1.68.0) were utilized to align coordinates and data formats with the current study (hg38) for background frequency comparison.

CCLE‐PM Cohorts [10]: PM cell line data from the CCLE served as benchmarks for in vitro model analysis.

2.9.2. Genomic Event Analysis

Somatic mutation analysis: High‐confidence mutation annotation format (MAF) files were generated by strictly filtering high‐frequency variants from public databases (gnomAD, ExAC, 1000 Genomes). The maftools package (v2.24.0) was employed for OncoPlot visualization [11].

CNV analysis: FREEC software (v11.6) was used to analyze WGS data to identify copy number gains and losses. By comparing these with CNV frequencies in the TCGA‐MESO cohort, the typicality and specificity of CNV events in the new cell lines were assessed on a genome‐wide scale.

2.9.3. Transcriptome Subtyping and Positioning

Batch effect correction: Raw read counts from the newly established cell lines were converted to TPM values using gene lengths from GENCODE v36 annotation and subsequently log2‐transformed (log2(TPM + 1)) to match the CCLE reference dataset. The two expression matrices were then merged, and the ComBat function from the sva R package (v3.56.0) was applied to remove batch effects while preserving biological variability [12, 13]. All subsequent transcriptome analyses were performed on this batch‐corrected expression matrix.

PCA projection: The top 1000 most variable genes were selected from the batch‐corrected expression matrix based on across‐sample variance. Principal component analysis (PCA) was performed on all samples jointly for dimensionality reduction and visualization to determine the transcriptomic positioning of the new cell lines relative to the CCLE cohort.

Hierarchical clustering: Hierarchical clustering was performed on the top 1000 most variable genes using Ward's minimum variance method and visualized as a heatmap with row‐wise Z‐score normalization.

3. Results

3.1. Establishment and Phenotypic Characterization of Three Novel PM Cell Lines

Three stable cell lines, designated PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4, were successfully established from pleural effusion specimens of eight individual PM patients, with a success rate of 37.5% (Table 1). Each cell line has been subcultured in vitro for more than 40 passages and exhibited adherent growth with distinct morphologies in culture (Figure 1A). PUMC‐MESO1 and PUMC‐MESO4 cells displayed a more polygonal, epithelioid shape, while PUMC‐MESO3 cells appeared more elongated. All three cell lines exhibited rapid proliferation, with population doubling times of 38.56, 56.55, and 38.23 h, respectively (Table 1; Figure S1A). The human origin of the cell line was verified through PCR amplification, and STR profiling further confirmed that the genetic signature of the cells was highly consistent with that of the original tumor tissue.

TABLE 1.

Characteristics of the pleural mesothelioma cell lines.

Cell line	Clinical data				Sample type	Cell line characteristics
	Age	Sexuality	Asbestos exposure	Histopathological diagnosis		Population doubling time	Tumorigenicity
PUMC‐MESO1	69	Female	Unknown	Pleural mesothelioma, epithelioid	Pleural effusion	38.56 h	Yes (5/5)
PUMC‐MESO3	85	Female	Unknown	Pleural mesothelioma	Pleural effusion	56.55 h	No (0/5)
PUMC‐MESO4	56	Female	Non exposed	Pleural mesothelioma, epithelioid	Pleural effusion	38.23 h	Yes (4/5)

FIGURE 1.

Phenotypic and Histological characterization of three novel human PM cell lines. (A) Phase‐contrast microscopy images showing the morphology of PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4 cells in vitro. Scale bar = 50 μm. (B) Histological and IHC analysis. H&E staining shows the tissue architecture of xenograft tumors (PUMC‐MESO1, PUMC‐MESO4) and a cell block (PUMC‐MESO3). IHC staining for BAP1, and the mesothelial markers WT1, D2‐40, and calretinin confirm the cell lineage and protein expression status. Scale bar = 100 μm.

To evaluate their in vivo tumorigenicity and confirm their mesothelial origin, the cell lines were subcutaneously injected into NOD/SCID mice. PUMC‐MESO1 exhibited strong tumorigenicity: tumors developed in all five mice. Palpable subcutaneous tumors were observed approximately 1.5 months after inoculation, and the tumor diameter reached nearly 1.5 cm at 2.5 months postinoculation. In the five mice injected with PUMC‐MESO4, palpable subcutaneous tumors were detected in all animals at 2 months after inoculation. Among them, the tumor in one mouse began to regress at 3 months and eventually disappeared completely, while the tumor diameter in the remaining 4 mice approached 1.5 cm at 4 months postinoculation (Table 1; Figure S1B). In contrast, no tumor formation was observed in mice inoculated with PUMC‐MESO3 for up to 4 months in vivo (Table 1).

Consequently, IHC staining was performed on xenograft tumor tissues from PUMC‐MESO1 and PUMC‐MESO4, and on a cell block prepared from the non‐tumorigenic PUMC‐MESO3 cells (Figure 1B). All three cell lines were stained positive for the mesothelial markers calretinin, WT1, and D2‐40. Conversely, they were negative for the adenocarcinoma marker Ber‐EP4, squamous cell carcinoma marker P40, confirming their mesothelial lineage (Figure S2A). Histopathological examination of the xenografts revealed that PUMC‐MESO1 formed tumors with a solid growth pattern, whereas PUMC‐MESO4 exhibited a tubulopapillary architecture. IHC analysis showed a complete loss of BAP1 protein expression in PUMC‐MESO1 and PUMC‐MESO4 [14]. In contrast, PUMC‐MESO3 exhibited a partial loss of BAP1 expression, with focal retention of nuclear staining in a subset of tumor cells. These findings were consistent with the pathological features characteristic of PM.

3.2. Chemosensitivity Profiles Reveal Heterogeneous Responses to Standard Therapies

To evaluate the potential of these cell lines as models for therapeutic testing, their sensitivity to three standard chemotherapeutic agents used in PM treatment—pemetrexed, cisplatin, and gemcitabine—was assessed. The cell lines displayed marked heterogeneity in their response profiles (Figure 2A–C). PUMC‐MESO4 was the most sensitive to both pemetrexed (IC₅₀ = 0.23 μM) and gemcitabine (IC₅₀ = 0.014 μM). In contrast, PUMC‐MESO3 demonstrated significant resistance, particularly to pemetrexed, with an IC₅₀ value of 11.15 μM. PUMC‐MESO1 showed an intermediate sensitivity to these agents. All three cell lines exhibited comparable sensitivity to cisplatin, with IC₅₀ values ranging from 2.8 to 4.7 μM.

FIGURE 2.

Chemosensitivity profiles and drug combination effects in PM cell lines. (A–C) Dose–response curves and corresponding IC₅₀ values for PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4 cells treated with (A) pemetrexed, (B) gemcitabine, and (C) cisplatin for 72 h. Cell viability was determined by CCK‐8 assay. Data are presented as mean ± SD. (D–F) Two‐dimensional synergy maps for the combination of pemetrexed and cisplatin in (D) PUMC‐MESO1, (E) PUMC‐MESO3, and (F) PUMC‐MESO4. The interaction landscape was quantified using the ZIP model, with red areas indicating synergy and green areas indicating antagonism. The overall ZIP synergy score is displayed for each cell line.

Given that the combination of pemetrexed and cisplatin is a cornerstone of PM therapy [15], the interaction between these two drugs was investigated using the ZIP model [16]. This analysis revealed distinct interaction patterns among the cell lines (Figure 2D–F). A positive ZIP score indicates synergy, while a negative score indicates antagonism. In PUMC‐MESO1, the combination resulted in a synergistic effect, with an overall synergy score of 6.12. The synergy map for this cell line showed broad areas of synergy, indicating that the combination was more effective than expected across a wide range of concentrations. In contrast, the same drug combination produced an antagonistic effect in PUMC‐MESO3 and PUMC‐MESO4, with synergy scores of −12.42 and −5.03, respectively. These findings highlight that the efficacy and nature of this standard combination therapy are highly genomic context‐dependent.

3.3. Genomic Analysis Reveals Classic PM Alterations and Unique Mutational Signatures

To delineate the genomic foundations of the observed phenotypes, whole‐exome and whole‐genome sequencing were performed. The analysis focused on somatic mutations and CNVs in key cancer‐related genes (Table S1; Figure 3A–D). The oncoplot displays somatic mutation patterns for 10 frequently altered genes (BAP1, NF2, CDKN2A, TP53, LATS2, SETD2, MTAP, NF1, MGA, and RBFOX1 [9]) across PUMC‐MESO1, PUMC‐MESO3, PUMC‐MESO4, and 22 PM cell lines from the CCLE(Figure 3A). The three cell lines harbored mutations in core PM driver genes. PUMC‐MESO1 featured concurrent mutations in BAP1 and NF2. PUMC‐MESO4 also carried a BAP1 mutation. In contrast, PUMC‐MESO3 was characterized by a TP53 mutation and homozygous deletion of the CDKN2A locus. These alterations are consistent with the loss of BAP1 protein observed by IHC in PUMC‐MESO1 and PUMC‐MESO4.

FIGURE 3.

Genomic landscape of the three PM cell lines. (A) OncoPlot visualizing key somatic mutations and copy number variation (CNV) status in selected cancer‐related genes across the three cell lines and 22 PM cell lines from the CCLE. Different colors denote mutation types and subtypes. Triangle markers indicate genes without CNV data. (B–D) Genome‐wide CNV profiles for (B) PUMC‐MESO1, (C) PUMC‐MESO3, and (D) PUMC‐MESO4. The plots show log2 copy number ratios across all chromosomes. Key PM‐related genes located in altered regions are highlighted.

Genome‐wide CNV analysis provided further insight into the genomic architecture of each cell line (Table S2; Figure 3B–D). The profiles revealed large‐scale chromosomal gains and losses consistent with the genomic instability characteristic of PM. For instance, PUMC‐MESO3 exhibited a clear homozygous deletion at the 9p21.3 locus, which contains the CDKN2A gene, confirming the mutation analysis. PUMC‐MESO1 and PUMC‐MESO4, both harboring BAP1 mutations, displayed distinct CNV landscapes. These comprehensive genomic data connect the cell lines to established molecular subtypes of PM [17] and provide a basis for mechanistic studies linking genotype to phenotype.

3.4. Transcriptomic Profiling Positions the Cell Lines Within the PM Landscape

To understand the functional output of their genomic alterations, RNA sequencing was performed, and the transcriptomic profiles of the three cell lines were compared with those of 22 PM cell lines from the CCLE. PCA was used to project the new cell lines onto the transcriptomic space of the CCLE cohort (Table S3; Figure 4). This analysis revealed that PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4 clustered together and were separated from the majority of existing CCLE PM cell lines.

FIGURE 4.

Transcriptomic positioning of the three PM cell lines relative to the CCLE reference cohort. Principal component analysis (PCA) of the top 1000 most variable genes from the batch‐corrected expression matrix. Each dot represents one cell line, colored by histological subtype. The three newly established PUMC cell lines are indicated by triangles. Shaded ellipses represent 95% confidence intervals for each subtype.

To further assess the relationship between the PUMC‐MESO models and existing CCLE PM cell lines, we additionally analyzed the 1000 most variably expressed genes across all cell lines. This analysis revealed that PUMC‐MESO1 was most closely related to Mero‐48a, PUMC‐MESO3 to IST‐MES2, and PUMC‐MESO4 to JL‐1(Figure 5). These findings suggest that, although the PUMC‐MESO cell lines formed a separate cluster in the global PCA analysis, each line still shares transcriptomic similarities with specific established PM models in the CCLE models.

FIGURE 5.

Hierarchical clustering of the top 1000 most variable genes across PM cell lines. Heatmap displaying the expression of the top 1000 most variable genes across the three newly established PUMC cell lines and 22 CCLE PM cell lines. Expression values are row‐wise Z‐score normalized from the batch‐corrected matrix.

4. Discussion

This study successfully established and characterized three novel human PM cell lines, PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4, derived from Chinese patients. These cell lines represent a significant contribution to the existing repertoire of in vitro PM models, as they exhibit unique biological, genomic, and transcriptomic features that capture the heterogeneity of the disease. As a resource, they provide new opportunities to investigate PM pathogenesis and to explore novel therapeutic strategies.

The genomic characterization confirmed that the cell lines recapitulate the hallmark genetic alterations of PM. The presence of inactivating mutations in key tumor suppressor genes—BAP1 and NF2 in PUMC‐MESO1, TP53 and CDKN2A in PUMC‐MESO3, and BAP1 in PUMC‐MESO4—anchors these models directly to the primary molecular subtypes of PM. This clear genotype–phenotype linkage makes them ideal tools for dissecting the functional consequences of these specific alterations. For example, the profound pemetrexed resistance observed in the TP53‐mutant/CDKN2A‐deleted PUMC‐MESO3 line, compared to the sensitivity of the BAP1‐mutant PUMC‐MESO4 line, provides a robust system for studying the molecular mechanisms of chemotherapy response and resistance driven by these distinct genetic backgrounds.

When compared to the extensive CCLE database, the PUMC‐MESO lines form a discrete cluster. Their shared transcriptomic signature, which differs from most existing models, underscores their value in expanding the diversity of available research tools [18]. Interestingly, despite forming a separate cluster in the global PCA analysis, each PUMC‐MESO cell line showed transcriptomic similarity to a specific established PM model in the CCLE cohort. PUMC‐MESO1 was most closely related to Mero‐48a, an epithelial mesothelioma cell line characterized by rapid proliferation and mixed epithelial/spindle morphology [19]. PUMC‐MESO3 showed the highest similarity to IST‐MES2, a biphasic mesothelioma model with mesenchymal features and reported stem cell–associated characteristics [20]. PUMC‐MESO4 was most closely related to JL‐1, an epithelioid mesothelioma cell line widely used in studies of mesothelioma invasion, inflammatory signaling, and therapeutic response [20, 21].

The drug sensitivity profiling not only highlighted the heterogeneous nature of PM but also yielded clinically relevant insights. The observation of both synergistic and antagonistic interactions between pemetrexed and cisplatin challenges the one‐size‐fits‐all application of this standard‐of‐care combination [22]. The synergistic effect in PUMC‐MESO1 validates the rationale for this combination in certain molecular contexts. Conversely, the strong antagonism observed in PUMC‐MESO3 and PUMC‐MESO4 suggests that for other molecular subtypes, this combination may be suboptimal or even detrimental, with the drugs potentially counteracting each other's effects. These findings emphasize the critical need for predictive biomarkers to guide combination therapy [23] and highlight the utility of these cell lines for systematic drug screening and the discovery of novel, effective drug pairings tailored to specific genomic contexts.

In conclusion, the three PM cell lines established in this study represent well‐characterized preclinical models that capture diverse molecular and transcriptional features of PM. These models may provide useful platforms for investigating tumor suppressor alterations, transcriptional heterogeneity, and differential therapeutic responses in PM. Future studies using these cell lines may further facilitate the identification of candidate therapeutic targets, mechanisms of treatment resistance, and biomarkers associated with therapeutic response in malignant mesothelioma.

Author Contributions

Yani Wu: investigation, data curation, formal analysis, validation, software, writing – original draft. Zhenli Yang: investigation. Dongmei Lin: resources. Jiayin Dai: investigation, data curation. Yanli Zhu: resources. Yuqin Liu: conceptualization, supervision, funding acquisition, writing – review and editing. Xiaocui Bian: conceptualization, methodology, writing – review and editing, supervision, investigation, validation. Fang Cao: conceptualization, methodology, resources, formal analysis, investigation, writing – review and editing, validation.

Funding

This work was supported by “CAMS Innovation Fund for Medical Sciences (CIFMS, 2021‐1‐I2M‐053 to Yuqin Liu)” and “Science Foundation of Peking University Cancer Hospital (Grant/Award Number: JC202607)”.

Ethics Statement

Clinical sample was approved by the Ethics Committees of Perking University Cancer Hospital with informed consent obtained from the patient (approval number: 2024YJZ81‐GZ01).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Phenotypic characterization of three human MPM cell lines in vivo and in vitro. (A) In vitro proliferation curves of the three cell lines over a 6‐day period. Cell numbers were counted daily. Data are presented as mean ± SD. (B) In vivo tumorigenicity of PUMC‐MESO1 and PUMC‐MESO4 cells. Images show tumor formation in NOD/SCID mice (left) and the corresponding dissected tumors (right) after subcutaneous injection.

Figure S2: Histological characterization of three novel human MPM cell lines. (A) Histological and IHC analysis. H&E staining shows the tissue architecture of xenograft tumors (PUMC‐MESO1, PUMC‐MESO4) and a cell block (PUMC‐MESO3). IHC staining for cytokeratin(CK), P40, Ber‐EP4, CK5/6 and vimentin. Scale bar = 100 μm.

Table S1: Detailed somatic mutation data for PUMC‐MESO1, PUMC‐MESO3, and PUMC‐MESO4.

Table S2: Gene‐level copy number variations for the three cell lines.

Table S3: Gene expression matrix for the three cell lines.

Data Availability Statement

Detailed mutation, copy number, and expression data for the three cell lines are available in the Supporting Information. All raw sequencing data supporting the findings of this study can be obtained from the corresponding author upon a reasonable request.