**Systematic Analysis of Aberrant Biochemical Networks and Potential Drug Vulnerabilities Induced by Tumor Suppressor Loss in Malignant Pleural Mesothelioma**

#### **Haitang Yang 1,2 , Duo Xu <sup>1</sup> , Zhang Yang <sup>1</sup> , Feng Yao <sup>2</sup> , Heng Zhao <sup>2</sup> , Ralph A. Schmid 1,\* and Ren-Wang Peng 1,\***


Received: 13 June 2020; Accepted: 4 August 2020; Published: 17 August 2020

**Abstract:** *Background*: Malignant pleural mesothelioma (MPM) is driven by the inactivation of tumor suppressor genes (TSGs). An unmet need in the field is the translation of the genomic landscape into effective TSG-specific therapies. *Methods*: We correlated genomes against transcriptomes of patients' MPM tumors, by weighted gene co-expression network analysis (WGCNA). The identified aberrant biochemical networks and potential drug targets induced by tumor suppressor loss were validated by integrative data analysis and functional interrogation. *Results*: CDKN2A/2B loss activates G2/M checkpoint and PI3K/AKT, prioritizing a co-targeting strategy for CDKN2A/2B-null MPM. CDKN2A deficiency significantly co-occurs with deletions of anti-viral type I interferon (IFN-I) genes and BAP1 mutations, that enriches the IFN-I signature, stratifying a unique subset, with deficient IFN-I, but proficient BAP1 for oncolytic viral immunotherapies. Aberrant p53 attenuates differentiation and SETD2 loss acquires the dependency on EGFRs, highlighting the potential of differentiation therapy and pan-EGFR inhibitors for these subpopulations, respectively. LATS2 deficiency is linked with dysregulated immunoregulation, suggesting a rationale for immune checkpoint blockade. Finally, multiple lines of evidence support Dasatinib as a promising therapeutic for LATS2-mutant MPM. *Conclusions*: Systematic identification of abnormal cellular processes and potential drug vulnerabilities specified by TSG alterations provide a framework for precision oncology in MPM.

**Keywords:** mesothelioma; tumor suppressor; targeted therapy; immunotherapy

#### **1. Introduction**

Malignant pleural mesothelioma (MPM) is a deadly cancer with incidence and mortality still increasing globally [1]. The leading cause for the poor prognosis of MPM is the extreme dearth of effective treatment options. The great majority of MPM patients present with advanced diseases, for whom a chemotherapy regimen (cisplatin plus pemetrexed) established in 2003 remains the only clinically approved first-line therapy [2].

Comprehensive genomic studies in MPM have revealed a rarity of pharmacologically tractable mutations in oncogenes [3–5], but the prevalence of inactivating alterations in tumor suppressor genes (TSGs), e.g., cyclin-dependent kinase inhibitor 2A/2B (*CDKN2A*/*2B*), BRCA1-associated protein-1 (*BAP1*), neurofibromin 2 (*NF2*), tumor protein p53 (*TP53*), SET domain containing 2 histone lysine

methyltransferase (*SETD2*) and large tumor suppressor kinase 2 (*LATS2*). While the pharmacological inhibition of oncoproteins is successful, targeted therapies that exploit abnormal TSGs have proven far more difficult. Precision oncology, a burgeoning effort aimed at targeting unique molecular alterations of individual patients, has achieved great success in many cancers, but significantly lags behind in MPM. Consequently, clinical trials in MPM without biomarker-directed stratifications have generally failed [6–9].

Although the direct intervention of tumor suppressors is challenging, aberrant TSGs induce the reprogramming of biochemical networks, which creates cancer-specific vulnerabilities and provides an alternative venue for precision oncology in TSG-driven cancer [10]. Systematic correlation analysis is a powerful tool to identify rewired cellular processes, potential therapeutic targets, and associated biomarkers [11]. Here, by implementing weighted gene co-expression network analysis (WGCNA) [12], paralleled by comprehensive data mining and functional interrogation, we systematically delineated the biochemical networks induced by the inactivation of major TSGs (*CDKN2A*/*2B*, *BAP1*, *NF2*, *TP53*, *SETD2*, and *LATS2*) in MPM, and the underlying implications for precision oncology. Identification of molecular traits and the associated drug vulnerabilities co-selected by the functional loss of specific TSGs provides unprecedented insights into MPM pathobiology and may promote personalized treatment of MPM patients with molecularly guided, targeted- and immuno-therapy.

#### **2. Results**

#### *2.1. Systematic Analysis of Rewired Biochemical Networks and Therapeutic Vulnerabilities Enabled by Tumor Suppressor Loss in MPM*

All the major genetic alterations (>10%) occurring in TCGA MPM cohort are TSGs, including *CDKN2A*/*2B* (homozygous deletions (HDs)), *BAP1* (HDs and point mutations), *NF2* (HDs and point mutations), *TP53* (point mutations), *SETD2* (HDs and point mutations), and *LATS2* (HDs and point mutations) (Figure 1A). Notably, there are substantial overlaps of alterations in different TSGs (Figure 1B). For instance, the majority (67.6%) of the MPM tumors that harbor HDs of *CDKN2A*/*2B* have co-occurring alterations in other TSGs, e.g., *BAP1* (40.5%) or *NF2* (37.8%). Importantly, analyses of RPPA data of TCGA MPM cohort (*n* = 61) showed that genetic alterations remarkably decreased the levels of the encoded proteins or downstream effectors (Figure 1C).

To uncover fundamental molecular features associated with the functional loss of TSGs in MPM, we performed WGCNA, based on the transcriptomic data of TCGA MPM cohort (Figure 1D and Figure S1A–D), and delineated a network of multiple modules or clusters, that are significantly positively or negatively correlated with genetic inactivation of the top six TSGs in MPM (Figure S1E). Genes in the positively correlated modules indicate the abundance of the module-specified traits conferred by individual TSG loss, while those in the negatively correlated ones indicate the attenuation. Genes in the gray module are those that cannot be clustered.

**Figure 1.** Major genetic alterations in The Cancer Genome Atlas (TCGA) MPM cohort. (**A**,**B**), Percentage (**A**) and overlap (**B**) of major (>10%) genetic alterations in The Cancer Genome Atlas (TCGA) malignant pleural mesothelioma (MPM) cohort (*N* = 81). (**C**), the association between the major genetic alterations (**A**) and the corresponding protein level in TCGA MPM cohort (*N* = 61). Protein array data were downloaded and reanalyzed from The Cancer Proteome Atlas (TCPA) database (https: //tcpaportal.org/tcpa/). Of note, protein quantification data of LATS2 and SETD2 were not available in the TCPA database. Phospho-YAP (S127) and TAZ are two critical factors, indicating the activity of Hippo pathway. (**D**), Workflow of weighted gene correlation networks analysis (WGCNA).

#### *2.2. CDKN2A*/*2B*

*CDKN2A*/*2B* encodes three tumor suppressors, p16INK4a and p14ARF (by *CDKN2A*) and p15INK4b (by *CDKN2B*), that play critical roles in cell cycle regulation. Moreover, p16INK4a and p15INK4b are functionally redundant by inhibiting cyclin-dependent kinase (CDK) 4/6 and cyclin D, and consequently blocking cell cycle progression from G1 to S [13].

The correlation network showed that *CDKN2A*/*2B* loss in MPM was significantly positively correlated with the green module (508 genes; correlation coefficient Pearson's r = 0.55; *p*-value = 2 × 10−<sup>7</sup> , followed by the yellow (543 genes; *r* = 0.34; *p*-value = 0.002), but negatively with the red (356 genes; *r* = −0.36; *p*-value = 0.001) (Figure S1E). Pathway analyses (GO, KEGG, Reactome) revealed that the green module enriched the genes involved in cell cycle regulation, particularly checkpoints and mitosis (Figure 2A,B and Figure S2A), consistent with the function of *CDKN2A*/*2B* in cell-cycle regulation. The yellow module significantly enriched the genes of extracellular matrix (ECM)-receptor interaction, PI3K/AKT, and focal adhesion pathways (Figure 2C,D and Figure S2B,C). Interrogation of the RPPA data revealed that MPM deficient in *CDKN2A*/*2B* had significantly higher levels of proteins involved in the cell cycle (e.g., Cyclin B1, Cyclin E2, CDK1 (p-Y15), FOXM1) and PI3K (e.g., 4EBP1 and PKC-delta (p-S664)) pathways, but decreased p16INK4a and PTEN (a negative regulator of PI3K) (Figure 2E), further supporting our results.

The red module negatively correlated with *CDKN2A*/*2B* loss enriched genes of anti-viral type I interferon (IFN-I, mainly IFN-α and IFN-β) signaling pathway, suggesting a link between *CDKN2A*/*2B* inactivation and impaired IFN-I pathway (Figure 2E–G and Figure S2D). To explore the underlying mechanisms, we analyzed co-occurring alterations in MPM samples, which revealed that *CDKN2A* and genes of the IFN family were significantly co-deleted (Figure 2H), consistent with a recent study, showing that defects in the IFN-I pathway mainly co-occur with *CDKN2A* loss [14].

We then analyzed intramodular connectivity, given that highly connected genes may serve as the hub with core regulatory roles. The top 20 best-connected genes in the green module are *KIF23*, *KIF4A*, *KIF2C*, *HJURP*, *KIF18B*, *MYBL2*, *BUB1*, *NUF2*, *UBE2C*, *CDCA8*, *CKAP2L*, *PLK1*, *DLGAP5*, *CDC20*, *TOP2A*, *DEPDC1*, *ANLN*, *CENPA*, *CDCA2*, *CEP55*. Most of these genes regulate the mitotic process and predict dismal prognosis in MPM (Figure S2E). Notably, the transcription factor MYBL2 is a central regulator of cell survival, proliferation and differentiation in cancer [15], and PLK1 and TOP2A are druggable by clinically advanced inhibitors. The top 20 best-connected genes in the yellow module are *COL5A1*, *VCAN*, *COL1A2*, *DACT1*, *FN1*, *CTHRC1*, *ITGA11*, *COL5A2*, *FAP*, *PODNL1*, *TGFB1I1*, *COL1A1*, *MMP2*, *COL3A1*, *LTBP1*, *MATN3*, *CHST6*, *POSTN*, *COL16A1*, *SRPX2*. Most of the genes are involved in ECM and associated with the suppression of anticancer immunity [16,17]. Supporting this notion, examining RPPA data revealed significantly decreased LCK, a key molecule in the selection and maturation of developing T-cells [18] (Figure 2E). Moreover, MPM has a high ECM signature compared to other solid tumors (Figure S3A), which predicts poor prognosis in patients (Figure S3B). However, the genetic underpinning for the high ECM of MPM has been unclear. Our data showed that the high ECM might be due to the high percentage (~46%) of MPM tumors with *CDKN2A*/*2B* alterations. The top 20 most connective genes in the red module are *OAS2*, *MX1*, *RSAD2*, *HERC6*, *IFIT3*, *CMPK2*, *IFI6*, *ISG15*, *USP18*, *IFIT2*, *OASL*, *IFI44*, *MX2*, *DDX60*, *IFI44L*, *OAS1*, *LAMP3*, *CYP39A1*, *IFIT1*, *RUFY4*, with the vast majority involved in the IFN-I pathway.

Collectively, these results reveal cellular processes that may represent therapeutic vulnerabilities in *CDKN2A*/*2B* deficient MPM. The enriched green and yellow modules indicate that *CDKN2A*/*2B*-mutant MPM may benefit from the co-targeting of the G2/M checkpoint or mitosis (e.g., PLK1) with PI3K/AKT, but might be associated with suppressive anticancer immunity due to high ECM. Oncolytic viral immunotherapy, a novel anticancer strategy preferentially killing proliferating cancer cells but sparing normal ones, might be particularly effective for the red module-marked subset, in which the IFN-I pathway genes are often co-deleted.

**Figure 2.** Enrichment analyses of genes significantly correlated with MPM tumors harboring HDs in *CDKN2A*/*2B*. (**A**,**B**), Top 10 significantly enriched Reactome pathways based on genes in the green module. In B, genes in the enriched Reactome pathways were listed. (**C**,**D**), Top 10 significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) (**C**) and Reactome (**D**) pathways based on genes in the yellow module. (**E**), Volcano plot showing the significantly (adjusted *p*-value < 0.05) upregulated (red) and downregulated (blue) proteins in malignant pleural mesothelioma (MPM) tumors harboring homozygous deletions (HDs) in CDKN2A/2B (versus wild-type), based on The Cancer Genome Atlas (TCGA) MPM cohort (*N* = 61). Data were downloaded and reanalyzed from The Cancer Proteome Atlas (TCPA) database (https://tcpaportal.org/tcpa/). (**F**,**G**), significantly enriched Reactome pathways based on genes in the red module. In (**G**), genes in the enriched Reactome pathways (**F**) were listed. (**H**), Genes significantly co-deleted with CDKN2A/2B in TCGA MPM samples. Data were downloaded from cBioPortal (https://www.cbioportal.org/). \* *p* < 0.05. 2.3. BAP1.

BAP1 has pleiotropic roles, ranging from the maintenance of genomic stability to the repair of DNA double-strand breaks (DSBs) [19,20]. Our analysis showed that *BAP1* alterations in MPM are positively correlated with the red module only (*r* = 0.41; *p*-value = 2 × 10−<sup>4</sup> ) that enriches the IFN-I pathway (Figure 2F,G), and negatively correlated with *CDKN2A*/*2B* loss (Figure S1E). This finding is supported by our recent study, showing that *BAP1* is negatively correlated with the IFN-I gene signature [21]. Thus, *CDKN2A*/*2B* deficiency plus *BAP1* proficiency defines a unique MPM subset that might particularly be sensitive to oncolytic viral immunotherapy.

#### *2.3. NF2*

NF2 is a plasma membrane protein binding to α-catenin and tight junctions to suppress cell growth. NF2 loss deregulates multiple signal pathways, although a prevalent notion holds that the Hippo pathway is central to the phenotype of *NF2*-mutant MPM.

Akin to *CDKN2A* loss, *NF2* alterations are positively correlated with the green (*r* = 0.34; *p*-value = 0.002) and the yellow (*r* = 0.26; *p*-value = 0.02) modules (Figure S1E), suggesting that NF2 might regulate cell cycle [22,23] and PI3K/AKT/mTORC1 (yellow module) [24], in addition to the canonical Hippo pathway. Supporting the notion, mining the public dataset that elaborates on protein-protein interactions revealed that the proteins involved in the ribosome, tight junction, Hippo and DNA repair are enriched in NF2-binding partners (Figure S4). The similarity between *CDKN2A-* and *NF2-*associated gene expression can alternatively be because *CDKN2A* and *NF2* alterations overlap in MPM (Figure 1B). However, *CDKN2A* and *BAP1* deficiency co-occurs at an even greater extent (Figure 1B) but rewires different gene networks (Figure 2) argues against this possibility.

Thus, like *CDKN2A*/*2B*, the genetic inactivation of *NF2* deregulates cell cycle, ECM and PI3K/AKT pathways, which prioritizes the co-targeting of the G2/M checkpoint/mitosis and PI3K/AKT pathway for *NF2*-altered MPM.

#### *2.4. TP53*

*TP53* mutations are negatively correlated with the purple module (125 genes; *r* = −0.37; *p*-value = 9 × 10−<sup>4</sup> ), to a less extent with the turquoise (1143 genes; r = −0.29; *p*-value = 0.01) and the green-yellow (108 genes; *r* = −0.27; *p*-value = 0.02), but positively with the salmon (57 genes; *r* = 0.23; *p*-value = 0.04), implying that *TP53* mutations deregulate multiple biological processes in MPM (Figure S1E). Notably, the turquoise is also significantly correlated with *LATS2* alterations (Figure S1E); we therefore focused on the purple and green-yellow module in the context of *TP53* mutations.

The purple module enriches genes of adipocyte differentiation/lipid metabolism, suggesting that *TP53*-mutant MPM might have attenuated activity of the processes (Figure 3A,B and Figure S5A) and benefit from differentiation therapy, e.g., peroxisome proliferator-activated receptor (PPAR) activator (Figure S5A). Supporting this notion, PPAR activator has been shown to promote the differentiation of mesenchymal therapy-resistant cancer cells to adipocytes [25]. Furthermore, the green-yellow module negatively correlated with *TP53* mutations enrich genes involved in lung epithelial cell differentiation (Figure 3C,D and Figure S5B), and the positively correlated salmon module enriches for genes of the neuronal system (Figure 3E,F). However, the marginal significance (*p*-value = 0.04) limits the value of this module.

The top 20 best-connected genes within the purple module are AQP7, PLIN1, ADIPOQ, TUSC5, CIDEA, THRSP, PLIN4, CIDEC, C14orf180, AQP7P1, CD300LG, C6, LIPE, LEP, NTRK2, SLC7A10, KCNIP2, GPD1, PDK4, and LPL, among which chemical agonists for PDK4, PRKAR2B and LPL are available. The top 20 best-connected genes of the green-yellow module include PDK4, TUSC5, LIPE, CIDEC, KCNIP2, CTSG, THRSP, CIDEA, AQP7P1, CD300LG, C7, C6, FREM1, THSD7B, MS4A2, TPSB2, C14orf180, FAM107A, TPSAB1, and TNMD.

**Figure 3.** Enrichment analyses of genes significantly correlated with MPM tumors with TP53 alterations. (**A**,**E**), Significantly enriched Reactome pathways based on genes in the purple (**A**,**B**), green-yellow (**C**,**D**) and salmon (**E**,**F**) modules. Cnetplots in (**B**), (**D**) and (**F**) listed genes in the enriched Reactome pathways (**A**, **C** and **E**, respectively).

#### *2.5. SETD2*

*SETD2* is a histone-modifying enzyme responsible for trimethylation of the lysine 36 residue on Histone 3 (H3K36me3) in humans. Impaired H3K36me3 causes aberrant gene regulation and chromosomal instability [26].

MPM with *SETD2* alterations is exclusively abundant (*r* = 0.25; *p*-value = 0.03) in the turquoise module, consisting of 1143 genes, with functions spanning from neuronal biology and receptor tyrosine kinases (particularly EGFR family) to the potassium channel, the Hippo and Wnt (Figure 4A–C). The Hippo and Wnt pathways are tumor-suppressive, precluding the potential as therapeutic targets. However, our results suggest that targeting EGFR might be a novel strategy for *SETD2*-altered MPM (Figure 4A,B).

**Figure 4.** Enrichment analyses of genes significantly correlated with MPM tumors with *SETD2* alterations. (**A**,**C**) Top 10 significantly enriched Reactome (**A**,**B**) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (**C**) pathways based on genes in the turquoise module. Cnetplot in (**B**) listed genes in the enriched Reactome pathways (**A**). (**D**) Volcano plot showing the significantly (adjusted *p*-value < 0.05) upregulated (red) and downregulated (blue) proteins in malignant pleural mesothelioma (MPM) tumors with SETD2 alterations (versus wild-type), based on The Cancer Genome Atlas (TCGA) MPM cohort (*N* = 61). Data were downloaded and reanalyzed from The Cancer Proteome Atlas (TCPA) database (https://tcpaportal.org/tcpa/). (**E**) Box-and-whisker plots show the extent of correlation between cytotoxic effects of each compound and with CDH1 (encoding E-cadherin) mRNA level, across 670 solid cancer cell lines. The y-axis indicates z scored Pearson's correlation coefficients; line, median; box, 25–75th percentile; whiskers, 2.5th and 97.5th percentile expansion; Here, only significantly (*p* < 0.05) correlated inhibitors were shown (in red dots). Labeled dots indicated the most negatively correlated drugs.

Genetic/molecular co-occurrence in tumor samples implies that progression to malignancy is a consequence of cooperative genetic/molecular dysregulations. Indeed, genetic alterations in *EGFR* and *SETD2* frequently co-occur in glioma [27] and TCGA pan-cancer cohort (Figure S6), supporting the notion that co-occurring *EGFR* and *SETD2* alterations cooperate to promote tumor progression, and that *SETD2*-mutant cancer may evolve a dependency on EGFR signaling. To further confirm the link between *SETD2* alterations and sensitivity to EGFR inhibition, we performed integrated analyses of proteomic (RPPA) and drug sensitivity data, which revealed that E-cadherin is significantly upregulated in *SETD2*-altered MPM (Figure 4D) and the expression of *CDH1* (encoding E-cadherin) is most negatively correlated with sensitivity to various EGFR inhibitors (Figure 4E). Of note, the red module, abundant in the IFN-I signature and positively correlated with *BAP1* alterations, is also positively correlated with *SETD2* mutations in MPM. This can be explained by considerably co-occurring *BAP1* and *SETD2* mutations, as 8 of 11 *SETD2*-altered MPM also have aberrant *BAP1* (Figure 1B). RPPA analysis confirmed significantly downregulated BAP1 in *SETD2*-altered MPM (Figure 4D).

The top 20 best-connected genes in the turquoise module are KLK11, CCDC64, CARNS1, CGN, BNC1, CLDN15, COBL, PARD6B, PLLP, PRR15, IGSF9, PRR15L, ANXA9, SELENBP1, PDZK1IP1, TGM1, SOX6, HOOK1, MSLN, NRG4. One of the hub genes in this module is MSLN, encoding mesothelin, a well-characterized biomarker for mesothelial tissue, and commonly overexpressed in epithelial mesotheliomas.

#### *2.6. LATS2*

At the heart of the Hippo pathway stands a core kinase cassette: MST1/2, LATS1/2, and adaptor proteins SAV1, MOB1A/B, which converges at LATS1/2-dependent phosphorylation of Yes-associated protein (YAP) and transcriptional co-activator with TAZ.

*LATS2* alterations show a negative correlation with the turquoise module (Figure 2, *r* = −0.45; *p*-value = 4 × 10−<sup>5</sup> ), which is opposite to *SETD2* alterations (positively correlated with the turquoise), but expected, in that genes involved in the Hippo and tight junction pathways are enriched in the turquoise module. Importantly, *LATS2* alterations in MPM are exclusively positively correlated (*r* = 0.33; *p*-value = 0.004) with the brown module (Figure S1E and Figure 5A), which significantly enriches for genes involved in immunoregulation (Figure 5B,C). These results suggest an immunoregulatory role beyond the canonical Hippo pathway by LATS2 and a rationale of immunotherapy for *LATS2*-altered MPM. Supporting the notion, PD-L1 (encoded by *CD274*) is the most significantly upregulated protein in *LATS2*-mutant MPM (Figure S7A), and LATS1/2 deletion has recently been shown to enhance anti-tumor immune responses [28]. Strikingly, a retrospective analysis of patients after being treated with immune checkpoint blockade showed that mutations of *LATS1*/*2*, rather than of *NF2*, predict significantly better survival (Figure 6A and Figure S7B).

The top 20 best-connected genes in the brown module are LCK, CD3E, IL2RG, SLAMF6, CD2, CD3D, SIT1, SH2D1A, CXCR3, TIGIT, TRAT1, CD6, GZMK, CD247, SIRPG, CD27, ZAP70, TBC1D10C, CD96, CD5. Of these, CD3E, IL2RG, CD2, CD3D, CD6, CD247, CD5, ITK, and CD3G are pharmacologically tractable.

Protein domains are important functional units and crucial for deconvolution of drug targets; we thus explored functional domains of the proteins encoded by the top 20 hub genes. Using SMART and PFAM protein fomains, we found that immunoreceptor tyrosine-based activation motif and Src homology 2 (SH2) domains are significantly enriched (false discovery rate < 0.05) in the hub proteins (Figure S7C). By correlating drug sensitivity with the gene expression of cancer cell lines (*n* = 670), we identified Dasatinib, a potent Abl/Src inhibitor, with the efficacy negatively correlated with several immune biomarkers (*CD274*, *CD47*, *PDCD1LG2*), that are preferentially expressed by cancer cells (Figure 6B). These results suggest that a role by LATS2 in cancer immunity and the potential of Dasatinib to target *LATS2*-altered MPM.

**Figure 5.** Enrichment analyses of genes significantly correlated with MPM tumors with LATS2 alterations. (**A**), Hierarchical clustering dendrogram of module eigengenes (labeled by their colors) and the sample trait (genetic alterations). Heatmap plot of the adjacencies in the eigengene network. In the heatmap, each row and column corresponds to one module eigengene (labeled by colors) or the trait. In the heatmap, green color indicates a negative correlation, while red represents a positive correlation. (**B**,**C**), Top 10 significantly enriched GO (biological process, BP), Kyoto Encyclopedia of Genes and Genomes (KEGG) and Reactome (**C**) pathways based on genes in the brown module. Cnetplot in C listed genes in the enriched Reactome pathways (**B**).

**Figure 6.** Identify Dasatinib as a promising therapeutic drug for MPM with *LATS2* alterations. (**A**) *LATS1*/*2* mutational status is associated with significantly improved overall survival in cancer patients after immune checkpoint blockage. The distribution of sample type (primary vs. metastatic; left panel), cancer type (middle panel) and drug type (anti-CTLA4; anti-PD1/PDL1; right panel) between *LATS1*/*2*-mutant and wild-type cancer. (**B**) Box-and-whisker plots show the extent of correlation between cytotoxic effects of Dasatinib and with several well-characterized immune markers (PDL1, PDL2, CD47), preferentially expressed by cancer cells. The y-axis indicates z scored Pearson's correlation coefficients; line, median; box, 25–75th percentile; whiskers, 2.5th and 97.5th percentile expansion; Here, only significantly (*p* < 0.05) correlated inhibitors were shown (in red dots). Notably, Dasatinib is the most negatively correlated drug. (**C**,**D**) the median inhibitory concentration (IC50) values of a panel of MPM cell lines treated with Dasatinib (72 h). MPM cells seeded in triplicate at 96-well plates were drugged 24 h later, over a 12-point concentration range (two-fold dilution). DMSO-treated cells were used as control. IC50 was determined using GraphPad Prism 7. IC50 values of Dasatinib in three MPM cell lines (H28, MSTO-211H, H2052) cultured in 2D and 3D were compared. \* *p* < 0.05 by Welch's *t*-test. *N* = 3 biological replicates. In D, the genetic annotations of MPM cell lines (**C**) were shown.

As preclinical proof of the concept, we found that *LATS1*/*2*-altered MPM cells exhibited the highest sensitivity to Dasatinib (Figure 6C,D). Importantly, the *LATS1*/*2*-altered MPM cells cultured in 3D retain a high sensitivity to Dasatinib (Figure 6C). Surprisingly, the mutational status of NF2, an upstream factor of LATS1/2 in the Hippo pathway, appeared not to predict the sensitivity to Dasatinib, which may suggest that NF2 and LATS1/2 have distinct and uncoupled functions in MPM. Further supporting our finding, Dasatinib was reported to show durable anticancer effects by promoting anti-tumor T cell responses, besides direct targeting of Abl/Src [29,30].

Finally, by analyzing RPPA data, we identified several antioxidant and anti-ferroptotic proteins, e.g., TFRC, GP6D, and PRDX1, that are significantly enriched in *LATS2*-altered MPM (Figure S7) [31]. In line with this observation, MPM with the aberrant Hippo pathway was reported to be susceptible to ferroptosis induction [32].

These results uncover an unexpected role for LATS2 in modulating immune contexture, suggesting a rationale for Dasatinib to treat *LATS2*-mutant MPM. Our data also argue that LATS2 and NF2 may exert distinct roles in MPM, at odds with the long-held assumption that they act as tumor suppressors through the Hippo pathway.

#### **3. Discussion**

Cancer patients vary in prognosis and response to therapy due to tumor heterogeneity [33,34], highlighting the need for personalized treatment. Unlike many other solid tumors, MPM is characterized by a pharmacologically intractable abnormal tumor genome, mainly TSGs, for which targeted therapy has been poorly established. In this study, we presented, for the first time, a systematic analysis of biochemical networks and associated vulnerabilities induced by the functional loss of TSGs in MPM, which not only sheds light on the mechanisms of MPM biology but also provides a framework of biomarker-guided targeted therapy in MPM (Figure S8).

#### *3.1. CDKN2A*/*2B and NF2*

An important finding of this study is that *CDKN2A* and *NF2* loss leads to similar changes in cellular pathways in MPM. Despite the evidence for targeting PI3K/AKT/mTOR pathway in MPM subsets [3,35–38], whether the deregulation of the pathway is associated with specific genetic events is unclear. Our results reveal the molecular underpinning of *CDKN2A* and *NF2* deficiencies, and further suggest therapeutic options for these MPM subsets. As p16INK4a (product of *CDKN2A*) inhibits CDK4/6 [13], CDK4/6 activation upon *CDKN2A* loss renders *CDKN2A*-deficient MPM particularly vulnerable to CDK4/6 inhibitors [36,39], and co-targeting CDK4/6 and PI3K/AKT/mTOR induce synergistic anti-MPM effects [36]. PI3K/mTOR inhibitors as monotherapy failed in unselected MPM patients [7], highlighting the importance of biomarker-guided stratification in future clinical trials.

Oncolytic viral immunotherapy shows promises in MPM [40], partly due to the special location of the malignancy that facilitates viral administration. We showed that IFN-I pathway genes are often co-deleted with *CDKN2A*, suggesting a rational by oncolytic viral immunotherapy for *CDKN2A*-altered MPM, which is supported by a recent report [14]. As *CDKN2A*/*2B* loss is widely used in pathological diagnosis to distinguish MPM from benign pleural lesions, analyzing the mutations of IFN-I–related genes will improve MPM diagnosis and patient stratification.

MPM has a high ECM signature, which may drive immunotherapy resistance [16,17]. Here, we provided evidence that high ECM in MPM is mainly attributable to *CDKN2A*/*2B* and *NF2* deficiency, that accounts for ~55.6% (45 of 81) of MPM cases (Figure 1B).

#### *3.2. BAP1*

BAP1 loss is frequent in MPM, renal cell carcinoma, peritoneal mesothelioma, and uveal melanoma [41]. Given the role of BAP1 in the maintenance of genomic stability, the association between *BAP1* mutations and sensitivity to PARP1-targeted therapy has been demonstrated in the chicken model of DT40 cells [19]. However, we and others have recently shown that *BAP1* mutations cannot precisely predict the response to PARP1-targeted therapy in MPM [20,42]. In addition, BAP1 status has been shown to determine the sensitivity to Gemcitabine treatment in MPM [43,44]. Here, *BAP1* alterations show significant abundance in IFN-I pathway only, consistent with our finding that *BAP1* is negatively correlated with the IFN-I signature in MPM [21]. Our data suggest that *CDKN2A* deficiency and *BAP1* proficiency should be considered to stratify MPM for oncolytic viral immunotherapy.

#### *3.3. TP53*

Mutant p53 has been proposed to drive metabolic reprogramming, thereby promoting cancer progression [45–48]. Our data reveal a potential role for *TP53* mutation in lipid metabolism, by deregulating the PPAR signaling pathway. Supporting our finding, p53 interacts with PPAR-γ co-activator 1α (PGC-1α) [45–47], and PPAR activator promotes the differentiation of mesenchymal therapy-resistant breast cancer cells [25]. These results warrant further studies to test differentiation therapy for *TP53*-mutant MPM.

Notably, synthetic lethal targets with p53 inactivation have been investigated [49–51]. In particular, MDM2, a nuclear E3 ubiquitin ligase that binds and targets p53 for proteasomal degradation, is detected in 21.3% of clinical MPM samples, and its expression is significantly associated with poor survival [52]. To restore p53 function, several small molecules, such as the Nutlin-like drugs that disrupt MDM2/p53 interaction, have been tested in MPM [53–55]. Moreover, we and others have shown that the inactivation of *CDKN2A*/*2B* and *TP53* is associated with an increased dependence on the G2/M checkpoint, which represents a targetable vulnerability in MPM [56,57].

#### *3.4. SETD2*

We showed that SETD2 might have roles beyond histone modifications. Of note, RTKs, particularly EGFR members (HER1 (EGFR, ERBB1), HER2 (NEU, ERBB2), HER3 (ERBB3), and HER4 (ERBB4)) were exclusively enriched in *SETD2*-altered MPM, suggesting the potential of pan-EGFR inhibitors for this MPM subset. Indeed, co-mutant *EGFR* and *SETD2* are common in glioma and pan-cancer [27], suggesting that *SETD2*-mutant cancer might have evolved a unique dependence on EGFR signaling.

EGFR is not mutated, but overexpressed in MPM [58–60]. A previous study showed that MPM expressed EGFR (79.2%), ErbB4 (49.0%) and HER2 (6.3%), but lacked ErbB3 [61]. In line with this, anti-HER-2 antibody synergizes with cisplatin in a subset of MPM cell lines [62]. However, the first-generation EGFR/ERBB1 inhibitor erlotinib [9] and gefitinib [8] show no clinical benefit, suggesting that pan-EGFR inhibitors might be necessary. To be noted, EGFR and other RTKs (MET, AXL) have been demonstrated to contribute to the activation of the downstream PI3K/AKT/mTOR in MPM [35], and the targeting PI3K/AKT/mTOR pathway, alone or in combination with other agents, have been investigated in MPM [7,36–38]. We showed that E-Cadherin is overexpressed in *SETD2*-altered MPM and predicts the sensitivity to EGFR-targeted therapies. Our finding that E-cadherin is significantly negatively correlated with EGFR inhibitor efficacy prioritizes the need for biomarker-driven selection and pan-EGFR inhibitors that target ERBB2/3/4 as well.

#### *3.5. LATS2*

LATS1/2 are key players of the Hippo pathway, but only LATS2 is frequently mutated in MPM. We identified the significant enrichment of immunoregulatory pathways in *LATS2*-mutant MPM, suggesting an unanticipated role for LATS2 in immunoregulation. Supporting our finding, LATS1/2 can suppress cancer immunity, and their deletion improves tumor immunogenicity by enhancing anti-tumor immune responses [28]. These results support a rationale of immunotherapy to target *LATS2*-altered MPM, although how LATS1/2 modulates the immune response awaits further studies.

Immunotherapy shows promises in MPM, but with low and heterogeneous response rates [63,64], arguing for biomarker-guided stratifications of MPM subsets responsive to immunotherapies. Our data

suggest that *LATS2* mutational status might be a critical factor in selecting MPM patients who can benefit from immunotherapies.

Strikingly, our study identified Dasatinib, a clinically approved RTK inhibitor, as a promising therapeutic for *LATS2*-altered MPM. Dasatinib shows the potential to modulate anticancer immunity (Figure 6B), and selectively impairs *LATS2*-altered MPM cells (Figure 6C), in line with the evidence that Dasatinib enhances anti-PDL1 efficacy in cancer [30]. These data suggest a rationale, by combining Dasatinib with immune checkpoint blockades to treat *LATS2*-altered MPM. Indeed, *LATS2* mutations are associated with beneficial survival in immunotherapy-treated patients (Figure 6A), but Dasatinib as monotherapy failed in unselected MPM patients [6,65], supporting the use of *LATS2* mutational status for patient stratification in clinical trials with Dasatinib.

Finally, we reveal a significant enrichment of proteins regulating ferroptosis in *LATS2*-mutant MPM, but not in those with *NF2* alterations, which is at odds with a recent report, showing that aberrant NF2-Hippo pathway is selectively susceptible to ferroptosis induction [32]. The observation that NF2 and LATS2 likely play different roles in MPM is supported by several lines of evidence. First, *LATS2* rather than *NF2* alterations are associated dysregulated YAP and TAZ (Figure 1C); secondly, *LATS2*- and *NF2-*mutant tumors show strikingly different enrichment of gene and protein signatures (Figures S1E and S7 and Figure 5); thirdly, Dasatinib selectively impairs *LATS2-* but not *NF2*-altered MPM (Figure 6); fourthly, *LATS1*/*2* mutations but not *NF2* alterations predict better survival in patients after immune checkpoint blockade therapy (Figure 6A and Figure S7B). Together, our data suggest that LATS2 and NF2 might have distinct roles in MPM, despite the long-held notion that both function through the Hippo pathway.

#### **4. Materials and Methods**

#### *4.1. WGCNA and Function Enrichment Analyses*

To identify the gene expression profiling associated with the major genetic alterations in MPM, The R package "WGCNA" was applied to the RNA-sequencing data retrieved from TCGA MPM cohort. In WGCNA, genes are clustered based on co-expression patterns to construct a gene co-expression network, which was transformed into the adjacency matrix and then topological overlap matrix (TOM) [12]. According to the TOM-based dissimilarity measure, genes were grouped into different modules (clusters) using the dynamic tree cut algorithm. For each module, the module eigengene (ME) was calculated; the first principal component representative of the module. The ME values were correlated with sample traits defined by specific genetic alterations in MPM samples. Here, we set the soft-thresholding power at 5 (scale-free R2 = 0.86), cut height at 0.25, and minimal module size to 30, to identify key modules. The module significantly correlated with sample traits was selected to explore its biological functions, such as gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and reactome pathway enrichment analyses, using the R package "clusterprofiler" [66]. Hub genes were defined as top 20 intramodular connected genes.

#### *4.2. Cell Viability Assay*

All normal human mesothelial cells Met-5A (MeT-5A, RRID: CVCL\_3749), MPM cell lines H28 (NCI-H28, RRID: CVCL\_1555), H2452 (NCI-H2452, RRID: CVCL\_1553), and H2052 (NCI-H2052, RRID: CVCL\_1518) were obtained from ATCC (American Type Culture Collection, Manassas, VA, USA) [67]. MPM cell lines MESO-1 (ACC-MESO-1, RRID: CVCL\_5113) and MESO-4 (ACC-MESO-4, RRID: CVCL\_5114) were obtained from RIKEN Cell Bank (Ibaraki, Japan). MPM cell lines MSTO-211H (RRID: CVCL\_1430) and JL-1 (RRID: CVCL\_2080) were purchased from DSMZ (German Collection of Microorganisms and Cell Cultures, Brunswick, Germany). A primary MPM cell culture (BE261T) was established from surgically resected tumors of a 67-year-old male patient, using the same protocol as described in [67] and used for short-term studies (up to eight passages in vitro). The human study was performed under the auspices of protocols approved by institutional review board (KEK number: 042/15), and informed consent was obtained from patients. Cells were cultured in RPMI-1640 medium (Cat. #8758; Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10% fetal bovine serum/FBS (Cat. #10270-106; Life Technologies, Grand Island, NY, USA) and 1% penicillin/streptomycin (P/S) solution (Cat. #P0781, Sigma-Aldrich, St. Louis, MO, USA). For 3D culture, cells were cultured in ultra-low attachment plate (Sigma-Aldrich, #CLS3474-24EA) with FBS-free RPMI-1640 medium supplemented with EGF (20 ng/mL; Cat. #PHG0311; Thermo Fisher Scientific (Waltham, MA, USA), bFGF (20 ng/mL; Cat. #PHG6015; Thermo Fisher Scientific), 4µg/mL insulin (Cat. #I9278; Sigma-Aldrich), 1× B-27 (Cat. #17504044; Thermo Fisher Scientific), 1% P/S. All human cell lines have been authenticated using STR profiling within the last three years, and are confirmed free from mycoplasma contamination (Microsynth, Bern, Switzerland).

MPM cells seeded in triplicate at 96-well plates (for 2D: 1000–1500 cells/well in tissue-culture treated plate (Corning, #353072); for 3D: 4000–5000 cells/well in ultra-low attachment plate) were drugged 24 h later, over a 12-point concentration range (two-fold dilution), with DMSO as vehicle. Cell viability was determined 72 h post-treatment by the Acid Phosphatase Assay Kit (ab83367; Abcam) [68]. The median inhibitory concentration (IC50) was calculated using GraphPad Prism 7.

#### *4.3. Public Databases*

RNA-sequencing data of MPM samples (*n* = 87) were downloaded from TCGA (https://portal.gdc. cancer.gov/), in which 81 samples were provided with genetic alterations data. Normalized level 4 data of reverse phase protein array (RPPA) were downloaded from The Cancer Proteome Atlas (TCPA) database (https://tcpaportal.org/tcpa/) [69], which quantified 218 proteins in 61 out of the 87 MPM samples in TCGA. R packages "limma" and "edgeR" were used to normalize the data and identify the differential gene or protein expression, respectively [70]. Protein-interacting data were downloaded from Agile Protein Interactomes DataServer (http://cicblade.dep.usal.es:8080/APID/init.action) [71], and co-occurring analysis data were downloaded from cBioPortal (https://www.cbioportal.org/). Processed drug (*n* = 481) screening and gene expression data across solid cancer cell lines (*n* = 659) were downloaded and reanalyzed from a published study [11]. Fisher's z-transformation was applied to the correlation coefficients to adjust for (normalize) variations in cancer cell line numbers across small molecules and cell lineages. Genetic and survival data of patients after immunotherapies (anti-PD1/PDL1, anti-CTLA4) were from TMB and immunotherapy (MSKCC) cohort in cBioPortal [72].

#### *4.4. Survival Analysis*

Survival analysis was performed using "survminer" and "survival" R packages. Tumor samples within the TCGA MPM cohort were divided into two groups, based on each hub gene's best-separation cut-off value to plot the Kaplan–Meier survival curves.

#### *4.5. ECM Gene Signature*

The extracellular matrix (ECM)/stromal gene signature was scored as the sum of an ECM/stromal gene set (*VCAN*, *FAP*, *POSTN*, *FBLN1*, *COL1A1*, *PDPN*, *THY1*, *CSPG4*, *IL6*, *TGFB1*, *HGF*, *SERPINE1*). The gene list was curated based on previous studies across different cancer lineages [16,17].

#### *4.6. Statistical Analysis*

Data were presented as mean ± SD, with the indicated sample size (*n*) representing biological replicates. Gene expression and survival data derived from the public database, as well as the correlation coefficient, were analyzed using *R* (version 3.6.0). *p* < 0.05 was considered statistically significant.

#### **5. Conclusions**

Overall, we report the systematic identification of biochemical networks and therapeutic potential linked with aberrant TSGs, which provides a framework for biomarker-guided precision oncology for MPM subsets. Our work warrants further studies that verify the drug vulnerabilities and the stratification approaches for future clinical trials.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/8/2310/s1, Figure S1: Weighted gene correlation network analysis (WGCNA) reveal gene modules linked with major genetic alterations in MPM; Figure S2: Pathway enrichment analyses of the genes significantly correlated with CDKN2A/2B loss; Figure S3: MPM has a high extracellular matrix (ECM) gene signature; Figure S4: Pathway enrichment analyses of the genes significantly correlated with NF2 alterations; Figure S5: Pathway enrichment analyses of the genes significantly correlated with TP53 mutations; Figure S6: Mutually exclusive and co-occurring analyses of STED2 and EGFR family genes across TCGA pan-cancer solid tumors; Figure S7: LATS2-altered MPM tumors enrich for the immune-regulatory signature; Figure S8: Tumor suppressor genes (TSGs)-guided precision oncology in MPM.

**Author Contributions:** Conceptualization, H.Y. and R.-W.P. Methodology, H.Y., D.X., Z.Y. Formal Analysis, H.Y. Investigation, H.Y., D.X., Z.Y. Data Curation, F.Y., H.Z., R.A.S., R.-W.P. Writing—Original Draft Preparation, H.Y. Writing—Review and Editing, all authors. Supervision, R.A.S., R.-W.P. Project Administration, R.A.S., R.-W.P. Funding Acquisition, H.Y., Z.Y., R.-W.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from Swiss Cancer League/Swiss Cancer Research Foundation (#KFS-4851-08-2019; to R.-W.P.), Swiss National Science Foundation (SNSF; #310030\_192648; to R.-W.P.) and PhD fellowships from China Scholarship Council (to H.Y. and Z.Y.).

**Acknowledgments:** This study used TCGA Program database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute.

**Conflicts of Interest:** The authors have declared no conflicts of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Evaluation of the Preclinical Efficacy of Lurbinectedin in Malignant Pleural Mesothelioma**

**Dario P. Anobile 1,†, Paolo Bironzo 1,2,†, Francesca Picca 1,3, Marcello F. Lingua <sup>4</sup> , Deborah Morena 1,3 , Luisella Righi 1,5, Francesca Napoli 1,5, Mauro G. Papotti 1,6,7, Alessandra Pittaro 1,6 , Federica Di Nicolantonio 1,8 , Chiara Gigliotti 1,8, Federico Bussolino 1,7,8, Valentina Comunanza 1,8 , Francesco Guerrera 9,10 , Alberto Sandri <sup>11</sup>, Francesco Leo 1,11, Roberta Libener <sup>12</sup>, Pablo Aviles <sup>13</sup> , Silvia Novello 1,2, Riccardo Taulli 1,3, Giorgio V. Scagliotti 1,2,7,\* and Chiara Riganti 1,7,14,\***


**Simple Summary:** The marine drug lurbinectedin revealed an unprecedented efficacy against patientderived malignant pleural mesothelioma cells, regardless of the histological type and the BAP1 mutation status. By inducing strong DNA damages, it dramatically arrested cell cycle progression and induced apoptosis. These results may be translated into the use of lurbinectedin as an effective agent for malignant pleural mesothelioma patients.

**Abstract:** Background: Malignant pleural mesothelioma (MPM) is a highly aggressive cancer generally diagnosed at an advanced stage and characterized by a poor prognosis. The absence of alterations in druggable kinases, together with an immune-suppressive tumor microenvironment, limits the use of molecular targeted therapies, making the treatment of MPM particularly challenging. Here we investigated the in vitro susceptibility of MPM to lurbinectedin (PM01183), a marine-derived drug that recently received accelerated approval by the FDA for the treatment of patients with metastatic small cell lung cancer with disease progression on or after platinum-based chemotherapy. Methods: A panel of primary MPM cultures, resembling the three major MPM histological subtypes (epithelioid, sarcomatoid, and biphasic), was characterized in terms of BAP1 status and histological markers. Subsequently, we explored the effects of lurbinectedin at nanomolar concentration on cell cycle, cell viability, DNA damage, genotoxic stress response, and proliferation. Results: Stabilized

**Citation:** Anobile, D.P.; Bironzo, P.; Picca, F.; Lingua, M.F.; Morena, D.; Righi, L.; Napoli, F.; Papotti, M.G.; Pittaro, A.; Di Nicolantonio, F.; et al. Evaluation of the Preclinical Efficacy of Lurbinectedin in Malignant Pleural Mesothelioma. *Cancers* **2021**, *13*, 2332. https://doi.org/10.3390/ cancers13102332

Academic Editor: Daniel L. Pouliquen

Received: 20 April 2021 Accepted: 10 May 2021 Published: 12 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

MPM cultures exhibited high sensitivity to lurbinectedin independently from the BAP1 mutational status and histological classification. Specifically, we observed that lurbinectedin rapidly promoted a cell cycle arrest in the S-phase and the activation of the DNA damage response, two conditions that invariably resulted in an irreversible DNA fragmentation, together with strong apoptotic cell death. Moreover, the analysis of long-term treatment indicated that lurbinectedin severely impacts MPM transforming abilities in vitro. Conclusion: Overall, our data provide evidence that lurbinectedin exerts a potent antitumoral activity on primary MPM cells, independently from both the histological subtype and BAP1 alteration, suggesting its potential activity in the treatment of MPM patients.

**Keywords:** MPM; lurbinectedin; DNA damage response

#### **1. Introduction**

Malignant pleural mesothelioma (MPM) is a rare but extremely aggressive type of cancer arising from pleural mesothelium and is highly associated with asbestos exposure. The disease is characterized by a long latency between initial exposure to asbestos and the clinical onset of the disease (30–50 years) and, although in Western regions the peak was expected in the 2020s [1], the ongoing use of asbestos in developing countries could lead to a persistence of new cases in the next decades [2]. MPM is classified into three major histological subtypes: epithelioid, sarcomatoid, and biphasic. While the epithelioid subtype occurs more frequently, accounting for approximately 60% of cases, and correlates with a better outcome, the sarcomatoid subgroup represents 10–20% of the cases and is characterized by a worse prognosis [3,4]. Independently from the morphology, the MPM tumor microenvironment is particularly enriched of immunosuppressive cells, which makes this tumor particularly refractory to different therapies [5–10]. Moreover, MPM is generally diagnosed in advanced stage, minimizing the role of curative treatments. For advanced-stage disease, the first-line systemic treatment consists of cisplatin and pemetrexed [11], a combination that prolongs the median survival time of only 3 months. Recently, the combination of immune checkpoint inhibitors directed against programmed death-1 (PD-1) and cytotoxic-T-lymphocyte-associated protein 4 (CTLA-4) showed its superiority over chemotherapy in previously untreated and unresectable MPM, especially in non-epithelioid tumors [12]. Conversely, no second-line standard therapy has been approved, despite the pre-clinical and the clinical evaluation of different therapeutic agents [13,14].

The genomic landscape of MPM reveals a low mutational burden with inactivating alterations mainly on oncosuppressors (BAP1, CDKN2A, NF2, TP53, LATS2, and SETD2) [15–18] thus precluding the use of molecular therapies against activated oncogenes. Among the oncosuppressors, BAP1 (BRCA1-associated protein) alterations account from 30% to 60% of cases [15,17,19,20]. Indeed, BAP1 germline mutations are known to predispose to mesothelioma and other cancer-associated syndromes [21,22] thus indicating a critical role for this deubiquitinase in suppressing tumor development. BAP1 regulates different biological processes among which chromatin modification, cell cycle, apoptosis, ferroptosis, cell metabolism, and differentiation [23]. Notably, BAP1 is involved in DNA synthesis, DNA duplication under stress conditions [24,25], and DNA damage response, by modulating the function of the BRCA1/BARD1 (BRCA1 Associated RING Domain 1) complex and coordinating the recruitment of RAD51 to the damaged DNA loci [26,27].

Lurbinectedin (PM01183) is a marine-derived anticancer drug that exerts a potent antitumor activity in different cancer cell lines and xenografts models and is currently under clinical evaluation in several tumor types [28–35]. Recently, the FDA has released a conditional approval for lurbinectedin for the treatment of second-line metastatic small cell lung cancer patients [36] while promising antitumor activity has been reported in MPM patients in second- and third-line [37]. However, there are no data available on the role of lurbinectedin as monotherapy or in combination in the first-line treatment of MPM. At the molecular level, lurbinectedin covalently binds CG-rich sequences in the DNA minor groove. The presence of the drug on the DNA helix inhibits the transcriptional process and is associated with the generation of DNA breaks [28]. Moreover, the interaction of lurbinectedin with both DNA strand breaks also interferes with the enzymes involved in the DNA damage response [38].

Here, we report about the potential efficacy of lurbinectedin in a panel of primary MPM cultures. Specifically, we demonstrated that lurbinectedin is strongly effective at nanomolar concentration and interferes with the transforming properties of MPM in a way that is independent of the BAP1 status and histological classification. With the caveat that our cell cultures were derived from diagnostic biopsies or surgical resections, our data indicate that lurbinectedin could potentially be explored in the management of patients with advanced MPM as second-line treatment or part of combination treatment in first-line.

#### **2. Results**

#### *2.1. Primary Mesothelioma Cell Cultures Characterization*

Twelve primary MPM cell lines, derived from patients with different histology, were stabilized as 2D cultures (Figure 1A). Flow cytometry for pan-cytokeratin (Figure 1B), immunohistochemical analysis (Figure 1C and Table 1), and immunoblotting for the BAP1 status (Figure 1D) were used to characterize the MPM cell lines. Notably, our panel (6 BAP1+ and 6 BAP1− cultures) was representative of the three major MPM histological subtypes (epithelioid, sarcomatoid, and biphasic) (Table S1).


**Table 1.** Histological characterization of MPM cultures.

Results of the immunohistochemical stainings of MPM samples for BRCA1 associated protein-1 (BAP1), pancytokeratin (pan-CK), Wilms tumor-1 antigen (WT1), calretinin (CALR). POS: positive; NEG: negative.

#### *2.2. Lurbinectedin Exerts Anti-Proliferative Effects in Patient-Derived Mesothelioma Cells*

As shown in Figure 2, lurbinectedin decreased the viability of MPM cells in a dosedependent manner, with an IC<sup>50</sup> in the low nanomolar range for all cell lines (Table 2), independently from the BAP1 status and the histological subtype (Figure 2A–D). Indeed, although the IC<sup>50</sup> was slightly higher in BAP1− vs. BAP1+ cells (Figure 2C) as well as in the sarcomatoid/biphasic vs epithelioid histotype (Figure 2D), the difference was not statistically significant. Notably, UPN6, UPN10, and UPN12 received trabectedin as secondline treatment and their overall survival was <12 months (Table S1). The cell lines derived from these patients had indeed the highest IC<sup>50</sup> in the panel analyzed, but it was below 5 nM for all of them (Table 2).

#### *2.3. Long-Term Lurbinectedin Treatment Impacts on MPM Transforming Abilities*

Since mesothelioma is particularly resistant to conventional chemotherapy, we evaluated the long-term effect of lurbinectedin in terms of inhibiting cell proliferation by

performing a crystal violet viability assay. Also in this setting, nanomolar concentrations of lurbinectedin dramatically reduced cell growth (Figure 3A,B). Furthermore, we extended our analysis by testing lurbinectedin ability to interfere with the anchorage-independent growth of MPM cells. The number of visible colonies was markedly decreased upon treatment, showing long-term anticancer efficacy (Figure 3C,D). Importantly, the consistent reduction in anchorage-independent growth showed no differences between BAP1+ and BAP1− cells, suggesting that lurbinectedin strongly impairs the tumorigenic potential of MPM cells, independently from the BAP1 status.


**Table 2.** IC<sup>50</sup> values of MPM cell lines treated with lurbinectedin.

**Figure 1.** Characterization of patient-derived MPM cell lines. (**A**) Representative images showing different morphology of three BAP1 positive (BAP1+) and three BAP1 negative (BAP1−) MPM cell lines (scale bar = 100 µm). (**B**) Flow cytometry plot representing the percentage of pancytokeratine positive cells in the indicated MPM cell lines. (**C**) Immunohistochemical analysis of BAP1, pan-cytokeratin (pan-CK), Wilms tumor-1 antigen (WT1), and calretinin (CALR) in the indicated MPM cell lines (scale bar = 100 µm). (**D**) Western blot analysis showing BAP1 status of the reported MPM cell lines.

#### *2.4. Lurbinectedin Treatment Interferes with Cell Cycle Progression*

To study the molecular basis of this anti-proliferative activity, we analyzed the effect of lurbinectedin on cell cycle regulation. While we observed variable changes in the percentage of cells in the G2/M-phase, indicating an unlikely strong mitotic arrest, we observed a constant accumulation of cells in the S-phase (Figure 4 and Supplementary Figure S1). This event occurred in both BAP1+ and BAP1− cells, suggesting that lurbinectedin-mediated perturbation of the cell cycle is BAP1-independent.

#### *2.5. Lurbinectedin Induces a Profound DNA Damage Coupled with Strong Apoptosis*

Among the pleiotropic mechanisms of action of lurbinectedin [28,38] the increase of S-phase arrested cells is suggestive of irreversible DNA damage. Indeed, lurbinectedin induced a significant increase in round-shaped and dense cells (Supplementary Figure S2). The presence of irreversible DNA fragmentation was evaluated by the Single Cell Gel Electrophoresis (SCGE). Specifically, in both BAP1+ and BAP1− cells lurbinectedin induced a dose-dependent genomic fragmentation (Figure 5A,B). The presence of genotoxic stress was confirmed by the increase in the phospho (Ser345) Chk1 and phospho (Thr68) Chk2 (Figure 5C,D), two cell cycle checkpoints that block DNA replication after being phosphorylated by the DNA-damaging sensors ATM/ATR kinases [39]. Moreover, in lurbinectedin-treated cells, we observed the accumulation of phospho (Ser15) p53 and phospho (Ser139) H2AX (Figure 5C,D), two additional targets of ATM/ATR kinases that are generally phosphorylated in response to DNA strand breaks and stalled replication [40,41]. This provided additional evidence of the strong DNA damage induced by lurbinectedin, which is also responsible for cell growth arrest (Figure 4 and Supplementary Figure S1). Such mitotic catastrophe is often coupled with apoptosis [40]. Accordingly, lurbinectedin treatment resulted in a strong induction of apoptosis (Figure 6A,B) as also shown by the dose-dependent activation of caspase 3 (Figure 6C,D).

**Figure 2.** Patient-derived MPM cell lines sensitivity to lurbinectedin. (**A**,**B**) Representative dose-response curves and corresponding IC<sup>50</sup> values of the two indicated MPM cell lines treated with lurbinectedin (0.1 nM–100 nM) for 72 h. (**C**) Dot plot of IC<sup>50</sup> values measured in lurbinectedin-treated MPM cell lines positive or negative for BAP1 expression. NS *p* > 0.05. (**D**) Dot plot of IC<sup>50</sup> values measured in lurbinectedin-treated MPM cell lines grouped according to the histological subtype. NS *p* > 0.05.

**Figure 3.** Lurbinectedin impairs long-term proliferation and anchorage-independent growth of MPM cell lines. (**A**,**B**) Representative pictures (lower panels) and quantification (upper panels) of crystal violet staining performed on the indicated MPM cell lines treated or not with lurbinectedin (5-fold the IC50) for 10 days. Data are expressed as means ± SEM; \*\* *p* < 0.01; \*\*\* *p* < 0.001. (**C**,**D**) Soft agar growth assay quantification of the indicated MPM cell lines treated or not with lurbinectedin (5-fold the IC50) for 20 days. The number of colonies obtained from untreated cells was set at 100%. Data are expressed as means ± SEM; \*\*\* *p* < 0.001.

**Figure 4.** Lurbinectedin effects on cell cycle distribution. (**A**,**C**) Representative flow cytometry histogram showing the cell cycle distribution of the indicated MPM cell lines, treated (purple) or not (green) with lurbinectedin (2.5-fold the IC50) for 24 h. (**B**,**D**) Histograms displaying cell number percentage in each cell cycle phase (G0/G1, S and G2/M) of the indicated MPM cell lines, treated or not with lurbinectedin (2.5-fold the IC50) for 24 h. Data are expressed as means ± SEM; NS *p* > 0.05; \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; \*\*\*\* *p* < 0.0001.

**Figure 5.** Lurbinectedin actively induces DNA damage response in MPM cell lines. (**A**) Representative Comet assay images of the indicated BAP1+ and BAP1− MPM cell lines treated or not with increasing lurbinectedin (L) concentrations (2.5-fold and 5-fold the IC50) for 24h (scale bar = 5 µm). (**B**) Histograms showing Comet assay data quantitation by CometScore software. Bars represent a percentage of total DNA in the tail. Data are expressed as means ± SEM; \*\*\* *p* < 0.001. (**C**,**D**) Western blot analysis for the indicated proteins in BAP1+ and BAP1- MPM cell lines treated or not with increasing lurbinectedin (L) concentrations (2.5-fold and 5-fold the IC50) for 24 h. GAPDH was used as a loading control.

**Figure 6.** Lurbinectedin treatment strongly induces apoptosis in MPM cell lines. (**A**,**B**) Histograms representing the percentage of apoptotic MPM cells treated or not with increasing lurbinectedin (L) concentrations (2.5-fold and 5-fold the IC50) for 72 h. The apoptotic rate was measured by TMRM assay. Data are expressed as means ± SEM; \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. (**C**,**D**) Western blot analysis of cleaved caspase 3 in MPM cell lines treated or not with increasing lurbinectedin (L) concentrations (2.5-fold and 5-fold the IC50) for 24 h. GAPDH was used as a loading control.

#### **3. Discussion**

Malignant Pleural Mesothelioma (MPM) is an aggressive tumor marginally impacted by standard chemotherapy regimens. Moreover, the lack of effective molecular therapies as well as the immune-evasive tumor microenvironment makes the treatment of MPM particularly challenging [5–10,42]. Because MPM currently lacks peculiar oncogenic drivers, we have explored the potential therapeutic efficacy of lurbinectedin, an alkylating agent which recently received FDA-conditional approval for the treatment of metastatic small cell lung cancer patients relapsing after chemotherapy [36].

We investigated the antitumor activity of lurbinectedin in a panel of 12 recently established primary MPM cell cultures. Our panel included all three MPM histotypes as well as cultures BAP1 positive and negative. Thus, although limited in terms of absolute number of cell lines, this panel is potentially representative of the different MPM phenotypes. Interestingly, we initially observed that lurbinectedin was effective at nanomolar concentrations and, as reported for other agents, its efficacy was independent of the BAP1 status. These data are particularly encouraging, although we are aware that freshly stabilized cultures could be potentially more sensitive to cytotoxic agents than what is usually observed at the clinical level. It is worthy of note, however, that three patients (UPN6, UPN10, UPN12) subsequently received trabectedin, a previous generation drug binding the minor groove of DNA, as second-line treatment. They did not show a superior clinical benefit compared to patients undergoing other treatments, indicating a limited efficacy of trabectedin. Interestingly, the MPM cells derived from these three patients had the highest IC<sup>50</sup> to lurbinectedin. These data may suggest that the response obtained in our stabilized cultures is a good surrogate of the potential effect of drugs binding the DNA minor groove and targeting the DNA repair observed in vivo.

Our experiments revealed that, as a consequence of the intrinsic ability of lurbinectedin to bind the minor groove of DNA, the drug interferes with the cell cycle, delaying progression through the S-phase. Interestingly, MPM cells immediately responded to genotoxic stress as demonstrated by the phosphorylation of H2AX, an early marker of the cellular response triggered by DNA double-strand breaks. Moreover, we observed the activation of Chk1 and Chk2 as a direct consequence of the stalled replication induced by DNA damage, responsible for the accumulation of MPM cells in the S-phase of the cell cycle. Finally, in our setting, p53 stabilization was not associated with DNA repair but invariably resulted in a massive apoptotic response, as revealed by cleaved caspase 3 activity and irreversible DNA fragmentation detected by Comet assay.

Notably, the efficacy of lurbinectedin against MPM was maintained also upon longterm treatment, as assessed by both crystal violet viability and anchorage-independent growth assays, providing further evidence of its anticancer potential.

As a consequence of DNA damage, replication arrest, and induction of apoptosis, we propose that lurbinectedin impairs the tumorigenic potential of MPM cells, and our results provide support to the clinical data recently reported in a multicentric phase II trial in second- or third-line palliative therapy [37]. Speculatively, considering the high anti-proliferative effect, if the results of the present study will be confirmed in MPM PDXs, lurbinectedin could be potentially investigated in the front line setting, for instance for a short pre-operative treatment in the early stages of MPM. Indeed, the reduction of anchorage-independent growth ability suggests lurbinectedin as a potential cytoreductive agent that, if proven in animal models and at the clinical level, will allow more conservative/less invasive surgery. Finally, the efficacy in all histotypes, independently from the BAP1 status, confers to lurbinectedin a strong advantage compared to other drugs currently used in MPM treatment, since its use could be potentially considered for all patients.

#### **4. Materials and Methods**

#### *4.1. Reagents and Chemicals*

Cell culture plasticware was obtained from Falcon (Glendale, AZ, USA), Biofil (Indore, India), and Costar (Washingtone, DC, USA). Lurbinectedin (PM01183) was kindly provided by PharmaMar (Madrid, Spain).

#### *4.2. Cells*

Primary MPM cells were obtained from biopsies during explorative thoracoscopy or pleurectomy, performed at the Thoracic Surgery Division of AOU Città della Salute e

della Scienza, Torino, Italy; AOU San Luigi Gonzaga, Orbassano, Italy, and AO of Alessandria, Biological Bank of Mesothelioma, Alessandria, Italy. Samples were anonymized by assigning an unknown patient number (UPN). Histological features of the original tumors and clinical features, including the first- and second-line treatment and the overall survival, of the corresponding patients are reported in Table S1. Samples were minced in 1 mm<sup>3</sup> -pieces, enzymatically digested for 1 h at 37 ◦C with 0.2 mg/mL hyaluronidase and 1 mg/mL collagenase [5], centrifuged at 1200× *g* for 5 min and seeded at 1 × 10<sup>6</sup> cells/mL density in DMEM advanced/F12 (Gibco, Dublin, Ireland) until passage #5, when cultures were shifted to DMEM/F12 nutrient mixture medium (Sigma, Saint Louis, MO, USA). All media were supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Sigma), 1% L-glutamine, 1% penicillin/streptomycin. Cells reached a stabilization (i.e., rate of cell subculture ≤1/week) in 2 to 7 months. UPN#3, UPN#4, UPN#5, UPN#6, UPN#10, UPN#11, UPN #12 were directly put in culture. UPN#1, UPN#2, UPN#7, UPN#8, UPN#9 were established from patient-derived xenografts. All cell lines were cultured in a humidified incubator at 37 ◦C in 5% CO<sup>2</sup> and routinely checked for Mycoplasma spp. contamination.

#### *4.3. Patient-Derived Xenograft Generation*

MPM patient-derived xenografts (PDXs) models were established from diagnostic tissue samples obtained at videothoracoscopy or during surgical pleurectomy. Each sample was implanted in the left or right side of the dorsal region of female NOD scid gamma (NSG) mice. A small piece of tumor was implanted subcutaneously and the wound was then stitched by surgical glue (Vetbond, Alcyon Italia, Cherasco, Italy). The tumor growth was monitored until the mass reached 2000 mm<sup>3</sup> . Then the animal was sacrificed by cervical dislocation, after anesthesia. The tumor area was shaved and disinfected with alcohol and the skin around the tumor was cut off. The tumor was divided into smaller pieces for re-implanting and collecting materials for further investigations. In the present work, the PDX platform was used as a tool to generate primary MPM cell cultures, stabilized in a shorter period (i.e., 2–3 months) than cells obtained directly from surgical procedures and used for pharmacological screening. To this aim, 0.2 g of tumors excised from the P1 generation of mice were digested to obtained a single-cell suspension [5] and put in culture as described in paragraph 4.2.

#### *4.4. Immunohistochemical Analysis*

The mesothelial features of cultures were confirmed by immunohistochemical (IHC) staining carried out on cells at passage 1. Specifically, cells were centrifuged at 1200× *g* for 5 min, fixed overnight in 4% *v*/*v* formalin at 4 ◦C, and then paraffin-embedded. The following antibodies were used: BAP-1 (Santa-Cruz Biotechnology, Santa Cruz, CA, USA, sc-28383, 1:100); Pan-cytokeratin AE1/AE3 (Dako, Agilent, Santa Clara, CA, USA, GA053, 1:500); Wilms Tumor-1 antigen (WT1) cl.6FH2 (Thermo Fisher Scientific, Waltham, MA, USA, MA1-46028, 1:10); Calretinin (Thermo Fisher Scientific, RB-9002-R7, 1:100). Mesothelial origin was confirmed if positivity for at least one between calretinin and WT1 was detected, as well as in the case of positivity for pancytokeratin. The histological features are reported in Table 1.

#### *4.5. IC<sup>50</sup> Calculation*

Cells were seeded in 96-well plates at a density of 2 × 103/well and serially diluted lurbinectedin (0.01 nM–100 nM) was added to the medium. After 72 h of treatment, IC<sup>50</sup> was evaluated with CellTiter-Glo (Promega) according to the manufacturer's instructions, using a Cytation 3 Imaging Reader (Bio-Tek Instruments, Winooski, VT, USA).

#### *4.6. Crystal Violet Assay*

For long-term proliferation, cells were seeded at a density of 4 × 103/well in 12 well plates and treated with the indicated concentrations of lurbinectedin for 10 days. Subsequently, cells were fixed and stained with 5% *w*/*v* crystal violet solution in 66% *v*/*v*

methanol and washed. Crystal violet was eluted by adding 10% acetic acid into each well. Quantification was performed by measuring the absorbance (570 nm) with Cytation 3 Imaging Reader (Bio-Tek Instruments).

#### *4.7. Soft-Agar Assay*

For anchorage-independent cell growth assay, cells were suspended in 0.45% type VII low-melting agarose in medium supplemented with 10% FBS at 1 × 10<sup>5</sup> cells/well, plated on a layer of 0.9% agarose in 10% FBS medium in 6-well plates, and cultured for 20–30 days with the indicated concentrations of lurbinectedin.

#### *4.8. Cell Cycle Analysis*

Cells were plated at a density of 1.2 × 105/well in 6-well plates and treated with the indicated concentrations of lurbinectedin for 24 h. Subsequently, cells were washed in PBS, treated with RNAse (167 µg/mL), and stained for 15 min at RT with propidium iodide (33 µg/mL). The cell-cycle distribution in G0/G1, S, and G2/M phases was analyzed by FACSCalibur flow cytometer (Becton Dickinson, Franklin Lanes, NJ, USA) and calculated using the CellQuest program (Becton Dickinson).

#### *4.9. Apoptosis Detection Assay*

MPM cells were plated at a density of 1.2 × 105/well in 6-well plates and treated with the indicated concentrations of lurbinectedin for 72 h. Subsequently, floating and adherent cells were washed with PBS and stained with tetramethylrhodamine methylester perchlorate (TMRM) (200 nM) for 15 min at RT. The percentage of apoptosis was measured by FACSCalibur flow cytometer (Becton Dickinson) and calculated using the CellQuest program (Becton Dickinson).

#### *4.10. Comet Assay*

DNA damage was assessed by Single Cell Gel Electrophoresis assay (Comet assay) [43]. At least 100 nuclei were counted in each condition. The percentage of DNA in the tail was quantified using the CometScore software (TriTek Corp., Sumerduck, VA, USA).

#### *4.11. Western Blot Analysis*

Cells were washed with ice-cold PBS and incubated for 20 min on ice in 0.1% Triton X-100 lysis buffer (20 mM Tris HCl pH 7.4; 150 mM NaCl; 5 mM EDTA; 0.1% Triton X-100; 1 mM Phenylmethanesulfonyl fluoride; 10 mM NaF; 1 mM Na3VO4, supplemented with protease inhibitor cocktail). Cells were then centrifuged at 14,000× *g* for 15 min at 4 ◦C to remove any cellular debris. Protein lysates were subsequently quantified using DC protein assay (Bio-Rad), loaded in 4–12% NuPAGE Bis-Tris Protein Gels (Thermo Fisher Scientific) according to the manufacturer's instructions, and transferred onto Hybond ECL nitrocellulose membranes. Blocking was performed with 5% Nonfat dried milk (PanReac AppliChem, Darmstadt, Germany) for 45 min at RT. Membranes were then incubated O/N at 4◦C with the following antibodies: BAP-1 (Santa Cruz Biotechnology, sc-28383); phospho(Ser345) Chk1 (Cell Signaling, Danvers, MA, USA, 2348); phospho(Thr68) Chk2 (Cell Signaling, 2197); phospho(Ser15) p53 (Cell Signaling, 9286); GAPDH (Cell Signaling, 5174); cleaved Caspase3 (Cell Signaling, 9661); phospho(Ser139)-Histone H2A.X (Cell Signaling, 9718); rabbit IgG, HRP-linked (Cell Signaling, 7074); mouse IgG, HRP-linked (Cell Signaling, 7076). Proteins were detected with horseradish peroxidase-conjugated secondary antibodies and Pierce™ ECL Western Blotting Substrate.

#### *4.12. Image Processing*

Image acquisition was performed with Leica dmire2 microscope and with Olympus BX51. Images were processed with the ImageJ software package (https://imagej.nih.gov/ ij/ accessed on 16 April 2021).

#### *4.13. Statistical Analysis*

All values were expressed as mean ± SEM and derived from at least two independent experiments. Statistical analyses were performed using Microsoft Excel and GraphPad Prism 5. Graphs were generated using Microsoft Excel and GraphPad Prism. Two-tailed Student's t-test was used to evaluate statistical significance: NS *p* > 0.05; \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; \*\*\*\* *p* < 0.0001.

#### **5. Conclusions**

Overall, our work proves the efficacy of lurbinectedin at nanomolar concentration against primary MPM cells. Although obtained in a relatively small cohort, that however is representative of the different MPM phenotypes, our results are particularly encouraging and put the basis for investigating lurbinectedin in different therapeutic settings of MPM.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/cancers13102332/s1, Figure S1: Lurbinectedin effects on cell cycle distribution, Figure S2: Lurbinectedin treatment strongly impairs cell viability in MPM cell lines, Table S1: Histological features of the original tumors and clinical features of the corresponding patients.

**Author Contributions:** Conceptualization, C.R., R.T., G.V.S., F.P., P.B., and D.P.A.; methodology, F.P., D.P.A., M.F.L., D.M., L.R., F.N., M.G.P., A.P., F.D.N., C.G., F.B., V.C., F.G., A.S., F.L., R.L., P.A., and S.N.; writing—review and editing C.R., R.T., G.V.S., and P.B.; supervision, C.R., R.T., P.B., and G.V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research plan has received funding from AIRC under IG 2019—ID. 23760 project to G.V.S and IG 2019 ID. 21408 project to C.R. EX60% Funding 2019 to P.B.; ERA-Net Transcan-2-JTC 2017 (TOPMESO to F.B.). PharmaMar kindly provided the drug for the study without any influence on the conduction of the experiments.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of San Luigi Hospital (#126/2016) and the Biological Bank of Mesothelioma, S. Antonio e Biagio Hospital, Alessandria, Italy (#9/11/2011). All animal procedures were performed in accordance with the national, institutional, and international law, policies, and guidelines (NIH guide for the Care and Use of Laboratory Animals -2011 edition-; European Economic Community (EEC) Council Directive 2010/63/UE; Italian Governing Law D. lg 26/2014). The animal study was approved by the Ethical Committee of the University of Turin and by the Italian Ministry of Health (400/2017-PR).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available in this article (and Supplementary Materials).

**Acknowledgments:** We are grateful to all patients and their families who participated in the study.

**Conflicts of Interest:** G.V.S. received honoraria from AstraZeneca, Eli Lilly, MSD, Pfizer, Roche, Johnson & Johnson, Takeda; consulting or advisory role for Eli Lilly, Beigene, and AstraZeneca, received institutional research funding from Eli Lilly and MSD and received travel, accommodations from Bayer. The Riganti and Taulli laboratories have received research support from PharmaMar. The other authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


#### *Systematic Review*

## **Meta-Analysis of Survival and Development of a Prognostic Nomogram for Malignant Pleural Mesothelioma Treated with Systemic Chemotherapy**

**Rupesh Kotecha 1,2,\* ,†, Raees Tonse 1,†, Muni Rubens <sup>1</sup> , Haley Appel <sup>1</sup> , Federico Albrecht <sup>3</sup> , Paul Kaywin <sup>3</sup> , Evan W. Alley <sup>4</sup> , Martin C. Tom 1,2 and Minesh P. Mehta 1,2**


**Citation:** Kotecha, R.; Tonse, R.; Rubens, M.; Appel, H.; Albrecht, F.; Kaywin, P.; Alley, E.W.; Tom, M.C.; Mehta, M.P. Meta-Analysis of Survival and Development of a Prognostic Nomogram for Malignant Pleural Mesothelioma Treated with Systemic Chemotherapy. *Cancers* **2021**, *13*, 2186. https://doi.org/ 10.3390/cancers13092186

Academic Editors: Daniel L. Pouliquen and Joanna Kopecka

Received: 31 March 2021 Accepted: 30 April 2021 Published: 2 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Simple Summary:** Malignant pleural mesothelioma (MPM) is a rare cancer with an aggressive disease course. For patients who are medically inoperable or surgically unresectable, multi-agent systemic therapy remains an accepted standard-of-care around the world. Given the rare incidence of MPM and the disease's aggressive nature, novel clinical trial designs are required. The purpose of this meta-analysis is to provide baseline summative survival estimates as well as evaluate the influence of prognostic variables to provide comparative estimates for future trial designs. In this study, a nomogram model was created to estimate survival with treatment with platinum-pemetrexed using covariates known to be associated with survival, including median age, gender, ECOG performance status, tumor stage, and tumor pathology subtype. Collaborative efforts can drive the change in the right direction, and appreciable progress has to be facilitated and newer trial designs may need to pave the way for future innovations in this rare disease.

**Abstract:** (1) Purpose: Malignant pleural mesothelioma (MPM) is a rare cancer with an aggressive course. For patients who are medically inoperable or surgically unresectable, multi-agent systemic chemotherapy remains an accepted standard-of-care. The purpose of this meta-analysis is to provide baseline summative survival estimates as well as evaluate the influence of prognostic variables to provide comparative estimates for future trial designs. (2) Methods: Using PRISMA guidelines, a systematic review and meta-analysis was performed of MPM studies published from 2002–2019 obtained from the Medline database evaluating systemic therapy combinations for locally advanced or metastatic disease. Weighted random effects models were used to calculate survival estimates. The influence of proportions of known prognostic factors on overall survival (OS) were evaluated in the creation of a prognostic nomogram to estimate survival. The performance of this model was evaluated against data generated from one positive phase II study and two positive randomized trials. (3) Results: Twenty-four phase II studies and five phase III trials met the eligibility criteria; 2534 patients were treated on the included clinical studies. Ten trials included a platinum-pemetrexed-based treatment regimen, resulting in a pooled estimate of progression-free survival (PFS) of 6.7 months (95% CI: 6.2–7.2 months) and OS of 14.2 months (95% CI: 12.7–15.9 months). Fifteen experimental chemotherapy regimens have been tested in phase II or III studies, with a pooled median survival estimate of 13.5 months (95% CI: 12.6–14.6 months). Meta-regression analysis was used to estimate OS with platinum-pemetrexed using a variety of features, such as pathology (biphasic vs. epithelioid), disease extent (locally advanced vs. metastatic), ECOG performance status, age, and gender. The nomogram-predicted estimates and corresponding 95% CIs performed well when applied to recent randomized studies. (4) Conclusions: Given the rarity of MPM and the aggressive nature of the disease, innovative clinical trial designs with significantly greater randomization to experimental

regimens can be performed using robust survival estimates from prior studies. This study provides baseline comparative values and also allows for accounting for differing proportions of known prognostic variables.

**Keywords:** mesothelioma; first line; meta-analysis; systematic review

#### **1. Introduction**

Malignant pleural mesothelioma (MPM) is a rare cancer with an aggressive disease course associated with poor prognosis [1]. Its incidence in the United States is approximately 3000 new cases diagnosed annually, but is still increasing in the rest of the world, particularly Asia and Europe [2]. Due to its insidious presentation, most patients are diagnosed with locally advanced or metastatic disease-unamenable to radical resection. For patients who are medically inoperable or surgically unresectable, multi-agent systemic chemotherapy remains a current standard-of-care with a median survival of approximately 12 months [3].

Although recent data from the CheckMate 743 trial have demonstrated improved outcomes with first-line immunotherapy [4], the combination of cisplatin and pemetrexed is commonly utilized in the front-line setting worldwide [5,6]. Carboplatin has similar efficacy to cisplatin, with a favorable toxicity profile and ease of administration; therefore, it has often been used in combination with pemetrexed for a large proportion of MPM patients, especially the elderly [7]. The purpose of this meta-analysis is to provide baseline summative survival estimates as well as evaluate the influence of basic prognostic variables to provide comparative estimates for future trial designs.

#### **2. Methods**

#### *2.1. Selection of Articles*

The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) criteria were followed in conducting this systematic review and meta-analysis [8]. The article selection was performed by searching the MEDLINE (PubMed) and Cochrane electronic bibliographic databases for first-line systemic therapy combinations for patients with locally advanced or metastatic MPM. To ensure a comprehensive initial search strategy, generic key words were used in the initial article screen: "mesothelioma" and "locally advanced" and "metastatic" and "first-line" and "systemic therapies" and "platinum/pemetrexed" and "experimental therapies". Full text articles published in the English language were considered and no publishing date restrictions were used through February 2021.

The initial query identified 447 reports that were subsequently screened by thorough review of the article titles and abstracts, as necessary. Inclusion criteria were publication in the English language, phase II and phase III clinical trials with 10 or more patients evaluable and with published outcomes on the efficacy endpoints of interest. Publications that were available in abstract only form and those in languages other than English were excluded. Case reports and limited case series, preclinical trials, studies using locoregional interventions alone, and studies using second-line therapies, were all excluded. A manual review of the references of the articles that were retrieved was performed to identify additional relevant publications. The search strategy used for this meta-analysis and the methodology for study inclusion is illustrated in Figure S1.

The studies were divided by treatment regimen: platinum-pemetrexed-based treatment and other experimental therapies. The demographic data abstracted for this analysis included year of publication, acronyms of the study or study title, duration of the study period, type of study (phase II/III), primary and secondary endpoints, number of patients included, median age, sex (male/female), ECOG Performance status (0,1,2), tumor stage (loco-regional disease; stage I–III and metastatic disease; stage IV), and tumor pathology (epithelioid, biphasic, sarcomatoid). Overall survival (OS), 1-year and 2-year OS rates, progression free survival (PFS), and objective response rate (ORR) were the outcomes evaluated. The radiological response data included patients having complete response (CR), partial response (PR), stable disease (SD), progressive disease (PD), and disease control rate (DCR). The toxicity summary included patients with grade 3–4 toxicities and was subdivided into toxicity category (i.e., general, blood and lymphatic system, cardiac, gastro-intestinal, infections, respiratory, and skin).

#### *2.2. Outcome Measures and Statistical Analysis*

The primary outcomes were OS and PFS; extracted medians of these variables were transferred into a logarithm scale [9]. The random-effects model described by DerSimonian and Laird [10] was used for this analysis. For primary and secondary outcomes, corresponding forest plots were created. Study heterogeneity was assessed using I<sup>2</sup> statistics. Values of 0–30%, 31–60%, 61–75%, and 76–100% indicated low, moderate, substantial, or considerable heterogeneity, respectively [11]. All analyses were performed in R (Version 4.0, R Foundation for Statistical Computing, Vienna, Austria). For identifying publication bias, funnel plots and the Egger test were used. Statistical significance of *p* < 0.05 indicated the presence of bias. To investigate the potential effects of each of the prognostic variables on OS, patient characteristics were also extracted from each study and included as predictors in the meta-regression model. Considered variables include median age, gender, ECOG performance status, tumor stage, and tumor pathology. The extent to which the metaregression model explained heterogeneity of the effect among studies was quantified by the percentage reduction of between-study variability. Plot of residuals was used to check the adequacy of the meta-regression model. Nomograms were used to represent results of the meta-regression model, estimating survival time using the covariates. In developing the nomogram, we used model coefficients to assign points to characteristics and predictions from the model to map cumulative point totals. Finally, the nomogram was used to predict the overall survival outcomes reported in the positive phase II reports (STELLAR [12] study) and phase III studies (MAPS [13] and CheckMate 743 [4]) and compared to the original results to assess the model performance.

#### **3. Results**

Twenty-four phase II studies and five phase III trials were included in this metaanalysis with outcomes data collected on 2534 patients (Figure S1). Key patient characteristics, demographics, and treatment information were not uniformly or consistently reported across the literature. However, there was no publication bias detected (*p* > 0.05) across the included studies regarding the primary outcomes evaluated in this meta-analysis (Figure S2).

#### *3.1. Demographic Data of Platinum-Pemetrexed Regimen*

Ten trials (*n* = 1303 patients) included a platinum-pemetrexed-based treatment regimen with a median of 89 patients in each study (range: 11–302 patients) (Table 1). Across all studies, 81% were male, and the median age was 66 years (range: 59–72 years). The majority of patients (60%) had an ECOG status of 1. The patients diagnosed with loco-regional disease and metastatic disease were 35% and 47%, respectively. The majority of patients across all studies were epithelioid (80%), followed by biphasic (11%), and sarcomatoid (8%).



Abbreviations: OS = overall survival; PFS = progression free survival; ORR = objective response rate; RR = response rate; TTP = time to progression; DCR = disease control rate; QoL = quality of life, ECOG = Eastern Cooperative Oncology Group; NA = not available.

#### *3.2. Treatment Outcomes, Radiological Response, and Toxicity Summary Data of Platinum-Pemetrexed Regimen*

Treatment with a platinum-pemetrexed-based regimen resulted in a pooled PFS of 6.7 months (95% CI: 6.2–7.2 months) and an OS of 14.2 months (95% CI: 12.7–15.9) (Figure 1A,B).

**Figure 1.** Forest plots demonstrating the (**A**) progression-free survival with platinum/pemetrexed; (**B**) overall survival with platinum/pemetrexed; (**C**) progression-free survival with other experimental therapies; and (**D**) overall survival with other experimental therapies. Squares indicate the proportions from individual studies and horizontal lines indicate the 95% confidence interval. The size of the data marker corresponds to the relative weight assigned in the pooled analysis using the random effects model. The diamond indicates the pooled proportion with 95% CI.

> Across all studies, the proportion of ORR was 24% (95% CI: 12–35%) and DCR was 73% (95% CI: 56–90%) (Table 2). Across all patients, the proportion of individual response rates for CR was 1.5% (95% CI: 1–4%), 19% PR (95% CI: 10–27%), 53% SD (95% CI: 37–69%), and 31% PD (95% CI: 14–48%).


**Table 2.**Treatment outcomes, radiological response, and toxicity summary for malignant pleural mesothelioma patients treated with cisplatin/carboplatin and pemetrexed.

Abbreviations: OS = overall survival; PFS = progression free survival; yr. = year; ORR = objective response rate; QoL = quality of life; NA = not available; N = number.

The pooled estimates of the treatment-related toxicity outcomes for patients who received a platinum-pemetrexed regimen (Figure 2A-I) with grade 3–4 blood and lymphatic system toxicities were anemia 10% (95% CI: 8–13%), neutropenia 22% (95% CI: 15–30%), and thrombocytopenia 7% (95% CI: 5–10%). Cardiac toxicity was seen in 1% (95% CI: 0–3%), gastro-intestinal toxicity in 3% (95% CI: 1–5%), fatigue in 6% (95% CI: 3–12%), infections in 5% (95% CI: 3–6%), skin toxicity in 1% (95% CI: 0–3%), and nausea and vomiting in 6% (95% CI: 3–10%).

**Figure 2.** Forest plots demonstrating the toxicity outcomes for platinum/pemetrexed regimen based on toxicity category: (**A**) Anemia; (**B**) Neutropenia; (**C**) Thrombocytopenia; (**D**) Cardiac Toxicity; (**E**) Gastro-intestinal toxicity; (**F**) Fatigue; (**G**) Infections; (**H**) Skin toxicity; and (**I**) Nausea and vomiting. Squares indicate the proportions from individual studies and horizontal lines indicate the 95% confidence interval. The size of the data marker corresponds to the relative weight assigned in the pooled analysis using the random effects model. The diamond indicates the pooled proportion with 95% CI.

#### *3.3. Demographic Data of Experimental Regimens*

Nineteen trials tested 15 experimental chemotherapy regimens (*n* = 1231 patients) in negative phase II or III studies, with a median of 52 patients (range: 20–229 patients) in each study (Table A1). Across these studies, 75% were male, and the median age was 63 years (range: 55–72 years). Patients had an ECOG status of 0 (30%), 1 (60%), and 2 (10%). The patients diagnosed with loco-regional disease and metastatic disease were 39% and 34%, respectively. In these studies, the majority of patients had epithelioid subtype (76%), followed by biphasic (15%), and sarcomatoid (9%).

#### *3.4. Outcomes, Radiological Response, and Toxicity Summary Data of Experimental Regimens*

Treatment with these experimental regimens resulted in a pooled estimate of PFS of 6.6 months (95% CI: 6.2–7.0 months) and OS of 13.5 months (95% CI: 12.6–14.6 months) (Figure 1C,D). Across all studies, the proportion of ORR was 31% (95% CI: 26–36%) and the DCR was 76% (95% CI: 69–84%) (Table A2). Responses using these experimental therapies were low: overall proportions for CR were 0.7% (95% CI: 0.3–1.6%), 29% PR (95% CI: 24–34%), 48% SD (95% CI: 42–55%), 22% PD (95% CI: 13–29%).

The pooled toxicity estimates for patients who received experimental chemotherapy regimens (Figure 3A-I) resulted in blood and lymphatic system grade 3–4 toxicities, with anemia in 4% (95% CI: 2–7%), neutropenia in 21% (95% CI: 12–33%), and thrombocytopenia in 12% (95% CI: 6–24%). Cardiac toxicity was seen in 4% (95% CI: 2–9%), gastro-intestinal toxicity in 4% (95% CI: 2–7%), fatigue in 12% (95% CI: 10–15%), infections in 5% (95% CI: 3–7%), skin toxicity in 1% (95% CI: 0–3%), and nausea and vomiting in 9% (95% CI: 6–15%).

**Figure 3.** Forest plots demonstrating the toxicity outcomes for various experimental regimens: (**A**) Anemia; (**B**) Neutropenia; (**C**) Thrombocytopenia; (**D**) Cardiac Toxicity; (**E**) Gastro-intestinal toxicity; (**F**) Fatigue; (**G**) Infections; (**H**) Skin toxicity; and (**I**) Nausea and vomiting. Squares indicate the proportions from individual studies and horizontal lines indicate the 95% confidence interval. The size of the data marker corresponds to the relative weight assigned in the pooled analysis using the random effects model. The diamond indicates the pooled proportion with 95% CI.

#### *3.5. Development of a Prognostic Nomogram to Estimate Survival*

Meta-regression analysis was used to estimate survival with treatment with platinumpemetrexed using covariates known to be associated with OS, including median age, gender, ECOG performance status, tumor stage, and tumor pathology subtype (Figure 4).

Unlike the aforementioned experimental regimens, two randomized phase III trials and one single-arm phase II trial have demonstrated promising outcomes in this disease entity. The Mesothelioma Avastin Plus Pemetrexed-cisplatin Study (MAPS) [13] evaluated cisplatin/pemetrexed/bevacizumab compared to cisplatin/pemetrexed, the STELLAR trial [12] evaluated the use of tumor-treating fields (TTFields) in addition to cisplatin/pemetrexed, and recently CheckMate 743 [4] evaluated nivolumab plus ipilimumab compared to cisplatin/carboplatin and pemetrexed. To evaluate the prognostic nomogram developed in this study, we compared the estimated outcomes using the patient populations enrolled onto these studies and the proportion of each of the covariates and compared the nomogram estimates with the published results. For the MAPS study, given the patient population in the experimental arm of the phase III study, the OS estimate from the nomogram was 15.76 months (95% CI: 13.96–17.81 months) compared to the reported 18.8 months in the study. Similarly, the OS estimate from the nomogram using the CheckMate 743 trial was 13.65 months (95% CI: 11.41–16.33 months) compared to

18.1 months reported in the experimental arm. Therefore, the results of the experimental arms of these two studies were outside the confidence interval estimate based on historical data and consistent with a positive outcome. For the STELLAR trial, the OS estimate from the nomogram was 16.95 months (95% CI: 10.49–27.38 months) and given the wide confidence interval, potentially could overlap with the 18.2 months reported in the study.

**Figure 4.** Nomogram model developed to predict overall survival (OS) in patients with malignant pleural mesothelioma treated with platinum/pemetrexed therapy. The mean log OS can be calculated by drawing a vertical line connecting the value of each variable with the point score at the top of the nomogram. The point scores for individual variables are then summed to get a total point score. This is then plotted along the total points line at the bottom of the nomogram. This line is projected to the mean log OS of the trial. Then the exponential of mean log OS is calculated to obtain the OS in months.

#### **4. Discussion**

Since 2003, chemotherapy with cisplatin/carboplatin and pemetrexed has been a standard first-line therapy for the majority of newly diagnosed patients who have locally advanced and metastatic MPM [6]. Over the past 15 years, multiple studies have established the outcomes for MPM patients treated with this regimen including single-arm phase II trials [7,14,15], the experimental arms of randomized trials compared to cisplatin alone [6], and the control arms of randomized trials testing novel experimental regimens [4,13,17–20]. In total, 1303 patients have been treated with this regimen across 10 studies, the data of which were abstracted in this systematic review and meta-analysis to determine pooled estimates of a PFS of 6.7 months and an OS of 14.2 months. In fact, a similar number—1231 patients—have been treated with experimental regimens who showed no improved outcomes compared to these historical estimates, underscoring the need for novel therapeutic development in this space. Moreover, despite advances in this field with the addition of bevacizumab and immunotherapy, doublet chemotherapy remains to be commonly used in most parts of the world where mesothelioma incidence continues to rise. Although the addition of bevacizumab to first-line chemotherapy has been added to the national guidelines [13], this regimen has not received FDA approval. Moreover, in CheckMate 743, nivolumab and ipilimumab were compared to pemetrexed-platinum, and although the OS was extended in the experimental arm, subgroup analysis yielded important caveats [4]. For example, for patients with epithelioid histologies (75% of those enrolled), the 12-month OS rates were not as striking (66% vs. 69%). Similarly, for those patients with a PDL-1 < 1%, the Kaplan-Meier curves crossed with longer follow-up, yielding an overall hazard ratio of 0.94. Hence, the role of first-line chemotherapy continues to be evaluated in ongoing trials.

Randomized controlled trials (RCTs) are deemed the gold standard of clinical research [21]. Randomization is often recommended for endpoints with a higher risk of confounding and selection bias, and it has been shown to improve the ability of phase II results to accurately predict phase III success [22,23]. However, modifications to traditional randomized trial designs have been performed to improve their performance in clinical practice. For example, the permuted block randomization has been widely used; however, in this design, there exists a compromise between effective imbalance control with a small block size and accurate allocation target with large block size [24]. Several alternative randomization designs have been proposed, such as the maximal procedure, brick tunnel randomization, and block urn designs [25–27]. However, for cancers such as mesothelioma, there are several logistical constraints for patients with rare diseases, as well as accrual/drop-out issues for those randomized to standard arms with known historically poor outcomes. Therefore, in other similar rare disease entities with robust historical survival estimates, there has been a resurgence in the consideration of alternative clinical trial designs [28]. Bayesian randomized designs and multi-arm multi-stage designs are two different approaches for improving reliability by using patient outcomes [29]. The Bayesian design allocates a greater proportion of prospective patients to well-performing treatments, whereas the multi-arm multi-stage designs use pre-specified stopping boundaries to discontinue novel treatments due to lack of efficacy. Although the Bayesian randomized designs have been shown to be more effective than traditional RCTs in multi-arm studies, their efficiency improvements in two-arm studies have been modest, especially if the rate of accrual outpaces the event rate, since the latter is required to modified the "prior" in a Bayesian concept [29]. Some studies examined the effects of phase II designs for binary endpoints on subsequent phase III trials, and found that randomization is useful when interstudy variability is high or there is a tendency to underestimate the control response [30]. Therefore, there is continued need to develop novel methods of clinical study and pooling historical data may help in future with future trial designs.

Meta-regression, the technique used in this study to develop the nomogram, is often used to assess the relationship between one or more covariates and a dependent variable. Similar approaches can be performed with a meta-analysis alone; however the covariates are at the level of the study rather than the level of the subject [31]. The differences that we need to address as we transition from using primary study data to meta-analysis for regression are similar to those for subgroup analyses. For example, in this meta-analysis, using meta-regression, we identified variables that were associated with OS and developed a nomogram to determine the influence of each of these on survival, including median age, gender, ECOG performance status, tumor stage, and tumor pathology. Using the nomogram, the overall survival was predicted as reported in the positive phase II and III studies and compared to the original result reported in these studies.

In the MAPS study [13], the patient population in the experimental arm of the phase III study showed an OS estimate from the nomogram to be 15.76 months (95% CI: 13.96– 17.81 months), as compared to the 18.8 months that was reported in the original study. Similarly, for another phase III study (CheckMate 743) [4], the OS estimate from the nomogram was 13.65 months (95% CI: 11.41–16.33 months), as compared to 18.1 months reported in the study. Based on the nomogram model developed from historical estimates, the OS reported for the positive phase III trials are outside of the 95% confidence interval range of the historical estimates; however, the predicted OS from the nomogram was also similar to the OS from the control arms in the original studies, indicating good performance. Interestingly, in the single-arm phase II STELLAR study [12], the OS estimate from the nomogram was 16.95 months (95% CI: 10.49–27.38 months), compared to 18.2 months. The OS reported in this study falls within the range of the 95% confidence interval predicted

from the nomogram. This demonstrates the importance of patient numbers in phase II trials, as the effectiveness of a phase II trial cannot be measured due to the wide confidence interval, prompting well-powered confirmatory studies. A well-designed phase II trial with complete reporting of the trial design, patient eligibility, study endpoints, and statistical analyses may be reliable and applicable in rare diseases, such as MPM [32].

There are important limitations to our analysis that should be noted. Formally, any categorical variable should have specific outcome-specific data to optimize the performance of the meta-regression. For example, for gender, male-specific OS and female-specific OS should be calculated. Unfortunately, this was difficult to extract from existing publications, since this level of detail is seldom reported. Similarly, Brims et al. [33] developed a prediction model for MPM using variables like Hb, weight loss, and albumin, which was unable to be extracted from existing publications for this study but would likely improve the performance of the survival estimates. However, in this study, we included percentages as continuous variables in the meta-regression. Furthermore, individual patient-level data can also be used to enhance any created model and should be pursued in subsequent studies. Given this promising approach with study-level data, further projects using individual patient-level data should be performed.

#### **5. Conclusions**

Given the rare incidence of MPM and the aggressive nature of the disease course, innovative clinical trial designs with significantly weighted randomization to experimental regimens can be utilized using robust survival estimates from prior studies. This study provides baseline comparative values and also allows for accounting for differing proportions of known prognostic variables. Collaborative efforts can drive change in the right direction, and appreciable progress has to be facilitated. Newer trial designs may be needed to pave the way for future innovations in this rare disease.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/cancers13092186/s1, Figure S1: PRISMA flow diagram showing the selection of studies for the systematic review of malignant pleural mesothelioma patients treated in the first-line setting with combination chemotherapy regimens, Figure S2: Funnel plots for (A) progression-free survival with platinum/pemetrexed, (B) overall survival with platinum/pemetrexed, (C) progression-free survival with other experimental therapies, and (D) overall survival with other experimental therapies to assess the potential for publication bias.

**Author Contributions:** Conception and Design: R.K., R.T. & M.R. Analysis: M.R. Critical Review of Manuscript: R.K., R.T., M.R., H.A., F.A., P.K., E.W.A., M.C.T., M.P.M. All authors have read and approved the manuscript.

**Funding:** This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** No new data were created or analyzed in this study. Data sharing is not applicable to this article.

**Conflicts of Interest:** R.K.: Honoraria from Accuray Inc., Elekta AB, Viewray Inc., Novocure Inc., Elsevier Inc. Institutional research funding from Medtronic Inc., Blue Earth Diagnostics Ltd., Novocure Inc., GT Medical Technologies, Astrazeneca, Exelixis, Viewray Inc; H.A.: Honoraria from Novocure Inc.; F.A.: Speaker's Bureau for Eli Lilly and Company and Boehringer Ingelheim; M.T.: Institutional research funding from Blue Earth Diagnostics Ltd; M.P.M.: Consulting for Karyopharm, Sapience, Zap, Mevion. Board of Directors: Oncoceutics. Other authors declare no conflict of interest.

#### **Appendix A**


**Table A1.**Demographic data for patients with malignant pleural mesothelioma treated with experimental therapies, negative on phase II/III studies.

Abbreviations: OS = overall survival; PFS = progression free survival; ORR = objective response rate; RR = response rate; TTP = time to progression; DCR = disease control rate; QoL = quality of life,ECOG = Eastern Cooperative Oncology Group; NA = not available.


**Table A2.** Treatment outcomes, radiological response, and toxicity summary for malignant pleural mesothelioma patients treated with experimental therapies, negative on phase II/III studies.

Abbreviations: OS = overall survival; PFS = progression free survival; yr. = year; ORR = objective response rate; QoL = quality of life; NA = not available; N = number.

#### **References**


## *Commentary* **Precision Therapy for Mesothelioma: Feasibility and New Opportunities**

**Sean Dulloo 1,2, Aleksandra Bzura <sup>2</sup> and Dean Anthony Fennell 1,2,\***


**Simple Summary:** Mesothelioma remains a lethal cancer. Personalized treatment is lacking. Emerging insights into the genomic and epigenomic landscape of mesothelioma highlight promising opportunities for precision therapy, where are discussed.

**Abstract:** Malignant pleural mesotheliomas (MPMs) are characterised by their wide variation in natural history, ranging from minimally to highly aggressive, associated with both interpatient and intra-tumour genomic heterogeneity. Recent insights into the nature of this genetic variation, the identification of drivers, and the emergence of novel strategies capable of targeting vulnerabilities that result from the inactivation of key tumour suppressors suggest that new approaches to molecularly strategy therapy for mesothelioma may be feasible.

**Keywords:** mesothelioma; histotype; Hippo pathway; NF2; BAP1; CDKN2A; PTCH1; SETD2; MTAP

#### **1. Introduction**

Over the last decade, multiple landmark next-generation sequencing studies of MPM have shed light on the spectrum of recurrently mutated cancer genes [1–4]. These studies have revealed a preponderance of tumour suppressor gene alterations and dominance of copy number alterations with a relatively low mutation burden of around two mutations per megabase. The absence of a bone fide tyrosine kinase proto-oncogene activating mutations as seen in other cancers (e.g., epidermal growth factor receptor or anaplastic lymphoma kinase, or ROS1 in lung adenocarcinoma), limits the opportunities to target gainof-function somatic alterations directly. However, emerging insights into the biology of MPM highlight opportunities for targeting vulnerabilities that may emerge due to tumour suppressor inactivation, and potentially, oncogenic processes (Figure 1).

**Citation:** Dulloo, S.; Bzura, A.; Fennell, D.A. Precision Therapy for Mesothelioma: Feasibility and New Opportunities. *Cancers* **2021**, *13*, 2347. https://doi.org/10.3390/cancers 13102347

Academic Editors: Daniel L. Pouliquen and Joanna Kopecka

Received: 7 April 2021 Accepted: 3 May 2021 Published: 13 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** Potentially actionable somatic alterations involving common tumour suppressors in pleural or periotoneal mesothelioma, or oncogene (ALK) in peritoneal mesothelioma. Trials shown on the right are evauating these strategies and are denoted by their trials.gov identifier.

#### **2. Histology, Prognosis, and Molecular Stratification of Therapy**

To date, the most commonly used classification of MPM has been histopathological, encompassing prognostically distinct subtypes spanning epithelioid (the most frequent and associated with a better prognosis) to biphasic and sarcomatoid (the latter being the most aggressive). Genomic comparisons of these subtypes do not reveal mutually exclusive somatic alterations, with all harbouring, to some extent, the three most common tumour suppressors at 9p21.3 (CDKN2A), 3p21 (BAP1), or 22q (NF2).

However, phenotypically there is a clear gradient of epithelial–mesenchymal transition or EMT [5–7] which may underpin chemotherapy resistance and the most aggressive behaviour of mesenchymal-like sarcomatoid MPMs. Patients with sarcomatoid MPM tend to have the worst outcomes with median survival ranging between 3.5 to 8 months [8], considerably shorter than for epithelioid subtype [9]. To date, although EMT exhibits plasticity, targeting EMT to revert a mesenchymal-to-epithelial phenotype has proven to be challenging [10].

The MPM histological spectrum may offer opportunities for stratified therapy. One approach has been to target epithelioid MPMs by taking advantage of the differential expression of mesothelin, which is commonly lacking in sarcomatoid MPMs. For example, the antibody-dependent conjugate aneteumab ravtansine has demonstrated clinical activity [11–13] in a molecularly stratified treatment context. Conversely biphasic and sarcomatoid MPMs harbour epigenetic silencing of argininosuccinate synthetase1 (ASS1) which can be therapeutically exploited [9,14–16]. ASS1 catalyses the condensation of citrulline with aspartate to form argininosuccinate. Cells that lacking ASS1 expression exhibit a vulnerability to arginine deprivation owing to a dependency (they are unable to convert endogenous citrulline–known as *auxotrophy*). In vitro, deprivation induces apoptosis which translates to clinical efficacy [9,16]. In the clinical trial called Arginine Deaminase and Mesothelioma (ADAM) study, patients were randomised in a 2:1 ratio to ADI-PEG20 (weekly intramuscular dose) versus best supportive care. The primary endpoint was

progression-free survival. The trial met its primary endpoint with a superior outcome with a Hazard ratio of 0.56 [9] confirming proof of concept.

A correlation between ASS1, and platinum/antifolate sensitivity was investigated in preclinical models [17,18]. The Phase 1 dose-escalation study involving ADI-PEG 20 in combination with pemetrexed and cisplatin has shown seven (out of nine) MPM patients having partial responses (78%) of which three had sarcomatoid/biphasic histology [14]. This phase 1 subsequently led to the development of the randomised phase 2/3 trial called ATOMIC MESO, randomising ADI-PEG20 or placebo with cisplatin and pemetrexed (ADICiSPem) non-epithelioid MPM patients which most commonly lack ASS1 [19].

#### **3. Targeting Hippo Pathway Mutations—Disrupting an Oncogenic Pathway?**

One of the most frequent pathways to be inactivated in MPM involves Hippo signalling, a pathway that regulates organ size. The most common somatic alterations involve the neurofibromatosis 2 gene (NF2 22q12) and the large tumour suppressor gene 2 (LATS2, 13q11-12) [20]. NF2 encodes merlin which recruits LATS1/2 kinases which phosphorylate the downstream effectors of the Hippo pathway, yes-associated protein (YAP), and its paralogue TAZ (WW domain-containing transcription regulator 1, or WWTR1). Inhibition of YAP/TAZ prevents their nuclear entry and ability to activate an oncogenic transcriptional programme in partnership with TEA domain transcription factor (TEAD) [21,22]. Therefore, Hippo pathway mutations de-repress a bone fide oncogenic pathway in MPM that is associated with shorter survival.

Recent analysis exploring the evolution of MPM has revealed that Hippo pathway inactivation involving NF2 almost always occurs as a secondary event during early clonal evolution, preceded by another other driver alteration [23]. Using a deep learning methodology to explore phylogenetic data obtained from multiregional sequencing of MPMs, it was repeated evolution was revealed across the cohort. This suggests that Hippo inactivation is deterministic, highlighting its significance as a potential biomarker for novel therapeutic strategies.

Early preclinical studies demonstrated a correlation between merlin loss and upregulation of focal adhesion kinase (FAK); inhibition of FAK was associated with the selective killing of merlin deficient cell lines, highlighting a potential synthetic lethal relationship [24,25]. This concept was then tested in a merlin-stratified, global randomised phase 2 trial called COMMAND [26–28], comparing maintenance defactinib or placebo. This study was, however, negative. Further preclinical studies revealed a novel function of FAK as an enhancer of regulatory T cell immunosuppression, leading to a phase 1 trial of defactinib and the PD1 inhibitor pembrolizumab, which includes an MPM cohort [29,30].

Alternative approaches to target Hippo-inactivated MPMs are emerging. Preclinical studies have highlighted potential sensitivity to SRC or BCR/Abl inhibition [31]. TEAD inhibitors are currently in development and could directly disable transcriptional oncogenic signalling [32]. Preclinical studies have identified that Hippo inactivation leads to a vulnerability to ferroptosis, a form of iron-dependent cell death. TEAD signalling upregulates ferroptosis modulators ACSL4 and TFRC, leading to enhanced sensitivity to agents such as sorafenib or sulphasalazine that can modify glutamate transport and cellular redox state [33,34].

#### **4. BAP1 Inactivation**

BRCA1 associated protein 1(BAP1) is a frequently inactivated tumour suppressor gene in MPM, which is also rarely associated with germline mutation [35]. Mechanisms through which BAP1 inactivation occurs include mutation, copy number loss, or translocations [4]. BAP1 deubiquitinates histone 2A lysine 119, and BAP1 deletion causing an increase in H3K27me3 associated with repression of enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2) activation. This suggests that small-molecule EZH2 inhibition could be an effective therapy in *BAP1*-mutant cancers. Based on this model, EZH2 inhibition was tested in BAP1 inactivated MPMs [36].

The EZH2 inhibitor tazemetostat was administered to patients with BAP1 inactivated MPM (loss of nuclear expression). The primary endpoint was disease control (i.e., stable disease, complete or partial response) at 12 weeks. The study enrolled 74 patients with the primary endpoint being met at 51%, demonstrating disease control at 12 weeks, with 25% continuing to 24 weeks. Interestingly, 2 of the 61 patients had confirmed partial responses [37]. Based on these data, EZH2 inhibitors might have antitumour activity; however, larger trials would be needed to support these findings.

Nuclear BAP1 regulates homologous recombination (HR) repair via interaction with RAD51 and BRCA1/BARD1 complex [38–40], in contrast to its cytoplasmic function in which it modulates calcium signalling mediated cell death [38]. Recruitment to DNA double-strand break sites is mediated via phosphorylation of BAP1, and the role of BAP in DNA damage response involves its catalytic activity [41].

Synthetic lethality associated with BRCA1/2 and PARP inhibition is well established and widely used for targeting homologous recombination deficient cancers. Cells harbouring HR deficiency switch to base excision repair, which is assisted by PARP, to repair DNA single-strand breaks [42]. PARP inhibitors trap PARP on DNA, resulting in catastrophic accumulation of double-strand breaks due to stalling and collapse of DNA replication forks, triggering cell death. The synthetic lethality interaction is observed in the clinic, in tumours harbouring somatic biallelic inactivation in BRCA1/2, leading to approval of PARP inhibitors in BRCA1/2 cancers [43]. A recent panel sequencing study of MPM reported a 36.9% involvement of HR pathway mutations, and this was deemed the most commonly mutated pathway in MPM [44], warranting evaluation of PARP inhibition in MPM.

Recently a phase 2a trial evaluated the use of rucaparib in patients with BAP1 or BRCA1 deficient MPM-*MPM Stratified Therapy 1* (MiST1) [45]. The primary endpoint for this trial was 12-week disease control which was met with a disease control rate of 58% (95% CI 37–77) with evidence of durable partial responses lasting more than a year, with manageable toxicity. In another study, olaparib (NCT03531840) reported 81% disease control at 6 weeks, with evidence of partial responses (4%) of which one responder harboured an MRE11A mutation [46]. Niraparib is being explored in patients with Trial (NCT03207347) in BAP1 and other DNA damage repair-deficient neoplasms, including MPM.

Evidence to support BAP1 as a bone fide predictor of sensitivity to PARP inhibition is lacking. One study recently identified that the sensitivity of MPM cells is not dependent on BAP1 but is enhanced by temozolomide in cells with high Schlafen 11 and low O6– methylguanine –DNA methyltransferase expression [47]. On the other hand, a novel MPM-specific splice isoform of BAP1 has been identified, lacking a portion of the catalytic domain, and which had decreased deubiquitinating activity compared to its full-length counterpart [48]. Cells expressing more than 20% of BAP1∆ were found to be more sensitive to olaparib than wild-type BAP1 MPM [48]. Coiled-coil domain containing 6 (CCDC6) interacts with BAP1 and has been reported to regulate both homologous recombination and PARP inhibitor sensitivity. Loss of expression of CCDC6 led to increased preclinical sensitivity to PARP inhibitors and is observed in around 30% of MPMs [32].

PARP inhibitors are proinflammatory and activate cytosolic DNA sensing by cyclic GMP–AMP synthase (cGAS) mediated activation of the endoplasmic reticulum-associated stimulator of interferon genes (STING) pathway [49,50]. The cyclic GMP–AMP synthase/stimulator of IFN genes (cGAS/STING) pathway [51] is responsible for sensing of damaged cytosolic DNA leading to activation of innate immune responses via initiation of signalling cascade involving the cytoplasmic DNA sensor cGAS, in concert with STING and TBK1, and transcription factors, such as IRF3 and NF-κB, that collectively induce a type I IFN response [51]. Therefore, the disruption of nuclear DNA integrity, via endogenous or exogenous factors, activates cGAS/STING pathway, leading to immunotherapy response [52]. Combining PARP inhibitors with immune checkpoint inhibitors in MPM is therefore rational and is being explored in the MIST 5 trial.

#### **5. 9p21.3 Deletion**

Homozygous deletion of 9p21, the locus harbouring the *p16ink4a* tumour suppressor is a frequent somatic alteration [53]. This deletion occurs within a cluster of genes that include CDKN2B, CDKN2A, and MTAP in up to 72% of MPMs [54]. CDKN2A regulates two important cell cycle proteins p16ink4a (an inhibitor of cyclin-dependent kinases 4 and 6), and p14ARF, an inhibitor of MDM2 which prevents p53 degradation. Restoring p16ink4a function is feasible with small-molecule CDK4/6 inhibition (to phenocopy p16ink4a). Preclinical studies of CDK4/6 inhibitors have reported evidence of nanomolar potency of palbociclib against MPM xenografts [55]. The MIST2 trial has completed accrual testing abemaciclib in p16ink4a negative MPM (results to be presented at the American Society of Clinical Oncology Conference in 2021).

Adenoviral mediated p14ARF gene transfection has been reported to induce G1 cell cycle arrest and apoptosis which was dependent upon the expression of p53 [56]. The heterodimerisation of MDM2 with its homologue, MDMX protein, enhances p53 ubiquitination and degradation. Phase 1 clinical study has investigated AMG 232, a selective MDM2 inhibitor that restores p53 tumour suppression by blocking the MDM2– p53 interaction with picomolar affinity [57] appears to be safe and could provide a strategy to target CDKN2A deleted MPMs, which harbour wild-type p53.

Methylthioadenosine phosphorylase (MTAP), encoded at the 9p21.3 locus is an enzyme essential in the methionine salvage pathway. MTAP converts methylthioadenosine, a product of polyamine synthesis, to adenine and methylthioribose-1-phosphate. The former is used for AMP and the latter for methionine synthesis [54]. MTAP deficiency leads to a dependency on de novo purine synthesis. The first attempt to target MTAP MPM involved L-alanosine (an inhibitor of de novo purine synthesis); however, there were no reported objective responses [58]. However, recently, it has been shown that loss of MTAP leads to elevation of its substrate methylthioadenosine (MTA). This partially inhibits protein arginine methyltransferase 5 (PRMT5) creating a vulnerability to further inhibition [59]. The old antibiotic quinacrine has been recently reported to silence PRMT5 transcriptionally, phenocopying siRNA-mediated inhibition of cell growth [60]. Inhibition of PRMT5 causes defective mRNA splicing and inactivation of MDM4, leading to p53 activation as a major pathway leading to impaired cell growth [61,62]. Interestingly, this pathway is also used by CDK 4/6 inhibitors [63]. An alternative approach being currently explored is the inhibition of MAT2A, the enzyme involved in the synthesis of the PRMT5 substrate, S-adenosyl methionine. MAT2A inhibition appears to be MTAP dependent and also disrupts mRNA splicing. This is approach is being explored in a phase 1 clinical trial with the agent AG-270.

#### **6. Anaplastic Lymphoma Kinase (ALK)**

ALK rearrangement in NSCLC is well studied and has multiple targeted treatment options with a good prognosis in these patients. In the recent few years, there has been evidence of ALK rearrangements in mesothelioma in the peritoneal subtype. One study that was carried out in pleural MPM identified 25 out of 128 patients (19.5%) with overexpressed ALK transcripts; however, only 10 expressed the ALK protein, and all were negative for ALK rearrangement by fluorescence in situ hybridisation (FISH) [64]. In contrast to the MPM findings, ALK rearrangement tends to be more prevalent in patients with peritoneal MPM, which was confirmed by FISH. ALK positivity was divided into focal weak (no ALK rearrangement) and diffuse strong (ALK rearrangement detected). Sequencing of these samples identified ALK fusion partners STRN, TPM1, ATG16L1 [65]. This has been translated into clinical practice, where the use of ceritinib in a patient with STRN–ALKrearranged malignant peritoneal mesothelioma showed response as early as 6 weeks into treatment [65].

#### **7. PTCH-1**

The hedgehog signalling pathway is involved in embryonic development and is inactivated in the adult mesothelium. Hedgehog ligands (Hh) bind to the transmembrane receptor Patched (PTCH1), which subsequently removes the inhibitory influence of the G-protein-coupled receptor smoothened (SMO). SMO activation leads to the induction of glioma-associated protein (GLI 1) and hedgehog interacting protein (Hhip) [66]. PTCH1 has been shown to be positively selected in the MPM [67] suggesting its role as a relatively rare driver (6%). Targeting the Ptch1 could play an important role in targeting the hedgehog signalling pathway. Vismodegib has recently shown activity in a patient with relapsed malignant MPM harbouring PTCH1F1147fs mutation. This patient had a durable response to Vismodegib [68].

#### **8. Conclusions**

Given the long latency of pleural malignant MPM and ongoing use of asbestos in several non-Western countries, malignant pleural MPM will remain a global health issue during the 21st century. Consequently, there is a pressing need for novel, effective, targeted treatments to improve patient outcomes. Targeted therapy is currently in its infancy for mesothelioma, but emerging developments preclinically and clinically are showing some promise. In summary, targeting altered tumour suppressors in MPM remains a challenge due to the need to identify and action vulnerabilities capable of inducing synthetic lethality; however, promising developments suggest that this may be feasible for the more common somatic alterations in this cancer.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** D.A.F. Research Funding Eli Lilly, GSK, Bergen Bio, MSD, Clovis Oncology, BMS, Boehringer Ingelheim, Astex Therapeutics, Bayer Oncology Honoraria Targovax, Inventiva, BMS, Bayer, Boehringer Ingelheim, Novocure, Lab 21. S.D. and A.B. declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Cancers* Editorial Office E-mail: cancers@mdpi.com www.mdpi.com/journal/cancers

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-2367-5