The YAP1 Signaling Inhibitors, Verteporfin and CA3, Suppress the Mesothelioma Cancer Stem Cell Phenotype

Mesothelioma is an aggressive/poor-prognosis cancer arising from the mesothelial lining of the pleura, peritoneum, and pericardium in response to asbestos exposure (1, 2). The first-line treatment is surgical resection coupled with chemotherapy (3). This treatment initially works, but the cancer cells eventually become resistant (2) leading to tumor recurrence and poor posttherapy survival (4). This poor clinical experience speaks to the need for improved therapeutic approaches. Cancer stem cells maintain tumors and mediate drug resistance (5–7). We have characterized mesothelioma cancer stem (MSC) cells (8, 9), which comprise 0.15% of tumor cell population (8). MCS cells are highly aggressive compared with nonstem cancer cells and display vigorous spheroid formation, Matrigel invasion, and migration. They also form highly aggressive and rapidly growing tumors relative to nonstem cancer cells (8, 9).

YAP1/TAZ/TEAD signaling is a centrally important procancer signaling pathway (10) and is often highly activated in cancer stem cells (11–13). YAP1 and TAZ factors interact with nuclear TEAD transcription factors to drive cancer cell survival–related gene expression and tumor growth (14) and are considered important cancer therapy targets (15). The central importance of this signaling cascade has engendered a search for inhibitors (16, 17). Verteporfin, a photosensitizer that is approved for the treatment of macular degeneration (18), was recently shown to inhibit YAP1/TEAD interaction to reduce YAP1 signaling (16). In addition, CA3 was identified in a TEAD reporter transcription screen as suppressing YAP1/TEAD-dependent transcription and tumor formation in mice (17). YAP1/TAZ and TEAD have a role in mesothelioma where they are overexpressed and drive the malignant phenotype (19–21).

In this study, we examine the impact of suppressing YAP1/TAZ/TEAD signaling on the MCS cell phenotype using cell culture and tumor models. We show that YAP1 function is required to maintain the MCS cell phenotype, and that treatment with CA3 or verteporfin suppresses MCS cell spheroid formation, Matrigel invasion, and migration. Verteporfin and CA3 also suppress xenograft tumor formation without evidence of adverse side effects. In each case, these agents inhibit YAP1/TAZ signaling and activate apoptosis. Thus, these agents target YAP1 signaling via different mechanisms to reduce the survival of pleural- and parietal-derived mesothelioma cancer cells.

RPMI1640 with l-glutamine (10-040-CV) was purchased from Corning and 0.25% trypsin-EDTA (T4049) was purchased from Sigma-Aldrich. Anti-PARP (556494) was obtained from BD Biosciences. Antibodies for Slug (ab27568) and Twist (ab49254) were obtained from Abcam. β-actin (A5441) antibody, heat-inactivated FCS (F4135) and DAPI (D9542) were purchased from Sigma. Antibodies for caspase-3 (9665), caspase-8 (9746), caspase-9 (9502), LATS1-P (9157), LATS1 (9153), YAP1-P (13008), YAP1 (4912), TAZ (4883), TEAD (13295), and Snail (3895) were purchased from Cell Signaling Technology. Anti-TAZ-P (sc-17610) was obtained from Santa Cruz Biotechnology. Peroxidase-conjugated anti-mouse IgG (NA931V) and anti-rabbit IgG (NA934V) were obtained from GE Healthcare, and anti-goat IgG (PA1-28664) was obtained from Invitrogen. These secondary antibodies were used as a 1:5,000 dilution. Matrigel (354234) and BD BioCoat Millicell inserts (353097) were purchased from BD Biosciences. The YAP1 signaling inhibitors, CA3 (CIL56) (17) and verteporfin (16) were purchased, respectively, from Selleckchem and Tocris. For use in cell culture, the compounds were dissolved in DMSO at a 1,000-fold stock. YAP(S127A) (Addgene plasmid # 27370) is a plasmid encoding a constitutively active form of YAP1.

Meso-1 and NCI-Meso-17 are immortal mesothelioma cell lines generated, respectively, from malignant epithelioid peritoneal and malignant sarcomatoid pleural mesothelioma, were maintained in RPMI1640 medium supplemented with 10% heat-inactivated FCS (9, 22). Meso-1 cells have been in culture for several years. NCI-Meso-17 cells are short-term passage cells derived from a pleural mesothelioma tumor that has been used for patient-derived xenograft studies (22). The NCI-Meso-17 cell line was kindly provided by Dr. Raffit Hassan (NIH, Bethesda, MD). Monolayer cells were plated at 2 × 10⁵ cells per 9.6 cm² plate (#353001, Corning) in RPMI1640 medium supplemented with 10% heat-inactivated FCS. After attachment, plates were treated with 0–5 μmol/L of verteporfin and cell number was counted on day 2 and expressed as mean ± SEM. The cells lines are periodically confirmed as authentic using short tandem repeat profiling. Cell lines were also tested for Mycoplasma contamination at the time of initiation of the studies, and are generally retested yearly.

For spheroid formation assay, Meso-1 or NCI-Meso-17 cell cultures (near-confluent) were dissociated with trypsin, collected by centrifugation, and resuspended in RPMI1640 medium containing 10% FBS. Single-cell suspension was plated for spheroid growth at 40,000 cells per 9.5 cm² ultra-low attachment dishes (#4371, Corning) and allowed to form spheroids for 0–5 days. We count the total number of spheroids and/or average spheroid diameter. We define a spheroid as a clonal collection of cells achieving a diameter greater than or equal to 25 μm. Verteporfin and CA3 impact on spheroid growth/survival was monitored using growing and/or preformed spheroids. To monitor impact on nascent spheroid growth, cells were seeded on ultra-low attachment plates, the compound was added the next day, and spheroid expansion was monitored thereafter. Alternately, spheroids were permitted to pre-form for 3–5 days prior to initiation of drug treatment.

To measure invasion, BioCoat Millicell inserts (d = 1 cm, 8 μm pore size, #353097) were coated with 120 μL of 250 μg/mL Matrigel. Cells (20,000) were seeded atop the Matrigel in 500 μL of RPMI1640 containing 1% FCS. The bottom chamber contained the identical medium containing 10% FCS. When appropriate, pharmacologic agents were added to the bottom chamber. After 18 hours the membrane was washed, fixed with 4% paraformaldehyde, and cells on the membrane inner surface were visualized by staining with 4′, 6-diamidino-2-phenylindole (DAPI) for fluorescence detection of nuclei. For migration assay, confluent monolayer cultures were wounded by scraping with a 10 μL pipette tip and the released cells were removed. Migration of cells to close the wound was monitored at 0–24 hours.

Cell or tumor samples were lysed in Laemmli Buffer (0.063 mol/L Tris-HCl, pH 7.5, 10% glycerol, 5% SDS, and 5% β-mercaptoethanol) and equivalent amounts of protein were electrophoresed on denaturing and reducing 10% polyacrylamide gels and transferred to nitrocellulose. The membrane was blocked by 5% nonfat dry milk and then incubated with the appropriate primary (1:1,000) and secondary antibodies (1:5,000). Secondary antibody binding was visualized using chemiluminescence detection technology.

Cells (1 × 10⁶) were collected and electroporated with nucleofection reagent VPD-1002 (Lonza, Germany; ref. 11). The mixture, containing 3 μg siRNA or plasmid as indicated, was mixed and electroporated using the T-018 setting on the AMAXA Electroporator. Prewarmed spheroid media were added and the suspension was transferred to monolayer culture. At 24 hours, the cells were used for spheroid formation, migration, and invasion assays. For siRNA experiments, the cells were harvested and electroporated a second time, following the same protocol at 48 hours after the initial electroporation (11).

Cells were transfected using Fugene 6 with 1 μg of pGL3B-Luc (empty vector) or pGL3B-8xGTIIC-Luc vector and permitted to recover for 24 hours before treatment for 24 hours with 2 μmol/L verteporfin, an inhibitor of YAP1/TEAD interaction (16) or 1 μmol/L CA3, which inhibits YAP1 signaling (17). After 24 hours the cells were harvested to assay for luciferase activity (23).

MCS cells (3 × 10⁶, derived from Meso-1 spheroids) were suspended in 100 μL PBS containing 30% Matrigel followed by subcutaneous injection into each front flank of 5 female NOD scid IL2 receptor gamma chain–knockout mice (NSG mice) per treatment group using a 26.5 gauge needle. Verteporfin was prepared in DMSO and the diluted in PBS to prepare an injection solution containing 10% DMSO. CA3 was prepared in DMSO and diluted in PBS to prepare an injection solution containing 0.6% DMSO. Inhibitor treatment was initiated at 8 weeks, the time of first tumor detection, by intraperitoneal injection of 100 μL delivering final concentrations of 0, 50, and 100 mg/kg verteporfin or 1.5 mg/kg CA3 delivered three times per week (Monday/Wednesday/Friday). Tumor growth was monitored by measuring tumor diameter and calculating tumor volume = 4/3π × (diameter/2)³ (8). Tumors were harvested at 11 weeks to acquire images and prepare extracts for immunoblot analysis. These studies were approved by the institutional board and followed accepted international practices for the treatment and welfare of animals. Statistics are assessed using the Student t test.

We first monitored the effect of verteporfin on the Meso-1 peritoneal-derived mesothelioma cancer cells (9). As shown in Fig. 1A, verteporfin causes a dose-dependent reduction in Meso-1 monolayer cell growth that is associated with abnormal cell morphology suggestive of cell death. MCS cells, which comprise 0.15% of the cells in monolayer cultures (8), are enriched in spheroid culture and display vigorous spheroid formation, Matrigel invasion, and migration (ability to close a scratch wound; ref. 8). To assess the impact on the MCS cell phenotype, we examined the impact of verteporfin on spheroid formation, invasion, and migration (8). We grew 5 day preformed spheroids on low-attachment plates and then treated with verteporfin for 48 hours. Verteporfin treatment promotes spheroid destruction and reduces spheroid number (Fig. 1B). Verteporfin treatment also suppresses MCS cell migration (Fig. 1C) and Matrigel invasion (Fig. 1D). Verteporfin has been shown to inhibit YAP1/TEAD signaling (16), and so we assessed the impact of verteporfin treatment on YAP1, TAZ, and TEAD signaling in MCS cell spheroids. Figure 1E shows that verteporfin treatment reduces LATS1, LATS1-P, and YAP1 levels. In contrast, TAZ and pan-TEAD levels are slightly increased. Moreover, verteporfin treatment is associated with enhanced apoptosis as measured by loss of procaspase 8, increased levels of active (cleaved) caspase 3 and 9, and PARP (Fig. 1F).

To assure that this effect is generalizable, we monitored the verteporfin impact on NCI-Meso-17 cell spheroid formation, invasion, and migration. NCI-Meso-17 cells are low passage patient-derived pleural mesothelioma cells (22). Cells were seeded on ultra-low attachment plates to assay spheroid formation in the presence of verteporfin. Verteporfin reduces spheroid size and promotes accumulation of cell debris in the cultures (Fig. 1G). Verteporfin also reduces cell invasion and migration (Fig. 1H and I), and this is associated with reduced LATS1, LATS1-P, YAP1, YAP1-P, and procaspase-8 and -9 (Fig. 1J). Cleaved PARP is present in treated and untreated cells. No cleaved caspase is detected, but loss of procaspases without appearance of the cleaved products is not uncommon in apoptosing cells (24). Thus, verteporfin decreases NCI-Meso-17 MCS cell YAP1 signaling and increases apoptosis.

Because–verteporfin may produce photosensitivity in patients, we tested another recently identified YAP1 signaling inhibitor, CA3 (Fig. 2A; ref. 17). Figure 2B–D show that CA3 treatment reduces the size and number of preformed Meso-1 cell spheroids, and suppresses Matrigel invasion and migration. Moreover, CA3 reduces LATS1, LATS1-P, YAP1, YAP1-P, TAZ, and pan-TEAD signaling (Fig. 2E). In addition, CA3 reduces procaspase-3, -8, and -9, and PARP level, and increases the level of cleaved (active) caspase-8 and -9 and PARP (Fig. 2F). Figure 2G–I show that CA3 treatment of NCI-Meso-17 cells causes destruction of preformed spheroids, and suppresses Matrigel invasion and migration. Moreover, these changes are associated with reduced YAP1/TAZ signaling and increased apoptosis (Fig. 2J and K). We also examined the impact of inhibitor treatment on YAP1-dependent transcription. Cells were transfected with empty vector luciferase or 8xTGIIC-Luc, which encodes repeats of the TEAD response element (25, 26). Figure 2L shows that CA3 and verteporfin treatment reduce YAP1/TEAD response element–dependent transcription in Meso-1 and NCI-Meso-17 cells. The reduction in transcription activity following inhibitor treatment is consistent with the treatment-dependent reduction in YAP1 level (Fig. 1E and 2E and J).

To confirm that the regulation we observe in cells is observed in tumors, we examined the impact of verteporfin and CA3 treatment on MCS cell tumor formation, and YAP1/TAZ signaling. Meso-1 spheroid-derived MCS cells were injected into NSG mice and tumor formation was monitored. When tumors were first palpated at 7–9 weeks, inhibitor treatment was imitated. Figure 3A and B shows that verteporfin produces a dose-dependent reduction in tumor formation. Figure 3C shows immunoblot profiles from two separate tumors, which indicate that YAP1/TAZ signaling is suppressed in verteporfin-treated tumors. These findings show that verteporfin suppresses TAZ and pan-TEAD level, but not YAP1 level. We also observe a reduction in Slug and Snail, which are regulators that drive epithelial-to-mesenchymal transition (EMT), suggesting that verteporfin treatment suppresses EMT. Another EMT factor, Twist, was not regulated. Figure 3D shows verteporfin activates tumor cell apoptosis, as evidenced by reduced procaspase-3, -8, and -9, and PARP, and increased cleaved/active caspase-3 and -9 levels. CA3 also reduces tumor formation (Fig. 3E and F). Moreover, as with verteporfin, there is a substantial reduction in YAP1, TAZ, and pan-TEAD. Slug and Snail (EMT markers) levels are also reduced (Fig. 3G). Apoptosis is increased, as evidenced by loss of procaspase 9 and accumulation of cleaved caspase 9, and reduced procaspase 8 and PARP (Fig. 3H).

If YAP1 has a key role in maintaining mesothelioma cancer cell function, YAP1 knockdown should reduce MCS cell survival and overexpression of constitutively active YAP1, YAP(S127A), and should antagonize verteporfin and CA3 action. To assess these concepts, we treated Meso-1 and NCI-Meso-17 cells with control- or YAP1-siRNA and monitored the impact on Matrigel invasion. Figure 4A and B show that YAP1 knockdown reduces Meso-1 and NCI-Meso-17 cell Matrigel invasion. Moreover, Fig. 4C shows that forced expression of YAP(S127A) enhances Meso-1 MCS cell Matrigel invasion and partially reverses the suppression of migration observed in verteporfin-treated cells. Forced YAP(S127A) expression also antagonized CA3 suppression of Matrigel invasion (Fig. 4D). Note that YAP1 levels in verteporfin- and CA3-treated cells is only marginally elevated (Fig. 4C and D), but this is not surprising because inhibitor treatment suppresses YAP1 level. The above findings are consistent with recent reports suggesting that YAP1/TAZ/TEAD signaling is required for mesothelioma cancer cell survival (21, 27, 28).

We next studied the relationship between YAP1, TAZ, and pan-TEAD level and proteasome activity. Preformed spheroids were treated with CA3 in the presence or absence of lactacystin, a proteasome inhibitor. Figure 4E shows CA3 reduces YAP1, TAZ, pan-TEAD, procaspase-9, and PARP levels, and that cotreatment with lactacystin reverses these changes suggesting that proteosome degradation reduces YAP1, TAZ, and pan-TEAD signaling.

Mesothelioma is an aggressive and unforgiving asbestos and nanotube-induced cancer (1, 2). Current treatment is surgical resection with chemotherapy, but this is only marginally successful (3). This fact has led to the search for new therapeutic options. The Hippo signaling cascade is a centrally important cancer cell survival cascade (10). Mst1/2 kinase phosphorylates LATS1/2 kinase, which phosphorylates/inactivates YAP1 and TAZ (10). This cascade controls many cellular processes, including organ development (10, 29). In growing cells Mst1/2 and LATS1/2 kinases are silent, which permits unphosphorylated YAP1 and TAZ transcription adaptor proteins to accumulate in the nucleus where they bind to TEAD transcription factors and activate expression of pro-proliferation genes (10). In contrast, in senescent cells, Mst1/2 and LATS1/2 kinases are active, and phosphorylate YAP1 and TAZ, which are then released to the cytoplasm where they are degraded by the proteasome, thereby relieving the growth stimulus (10). This cascade is extremely important in cancer where LATS1/2 signaling is reduced leading to enhance accumulation of active YAP1 in the nucleus. For example, highly aggressive hepatocellular carcinoma develops in mice in response to constitutive YAP1/TAZ activity (30). YAP1/TAZ/TEAD are also important in human cancers (10, 31) including mesothelioma where they drive the malignant phenotype (19–21). We confirm that YAP1 maintains the MCS cell phenotype, such that loss of YAP1 reduces MCS cell spheroid formation, Matrigel invasion, and migration. Moreover, expression of constitutively active YAP1 antagonizes verteporfin and CA3 suppression of Matrigel invasion, suggesting that loss of YAP1 is a requirement for action of these agents.

Because of the procancer activity of this pathway (19–21), YAP1/TAZ signaling is an important cancer therapy target. Verteporfin is a photosensitizer that is used to treat excess vascularization in eye disease (32, 33), but has been shown to inhibit YAP1/TEAD interaction to reduce YAP1 signaling (12, 16, 34–36). In this study we show that verteporfin suppresses cell proliferation, spheroid formation, Matrigel invasion, and migration of MCS cells derived from Meso-1 (peritoneal mesothelioma) and NCI-Meso-17 (pleural mesothelioma) cells. A common theme is suppression of LATS1, LATS1-P, YAP1, YAP1, TAZ, and pan-TEAD levels. It is interesting that verteporfin, which inhibits YAP1/TEAD interaction (16), reduces the level of both LATS1 and LATS1-P. This is surprising as LATS1/LATS1-P is upstream of YAP1 in the signaling cascade (16). We are not certain why inhibition of YAP1/TEAD function reduces LATS1/LATS1-P; however, LATS1/LATS1-P loss is not likely to be important because verteporfin treatment reduces the level of the penultimate regulators, YAP1, TAZ, and TEAD.

Photosensitizers, like verteporfin, have been used to treat mesothelioma (37, 38). In this procedure, the tumor is surgically removed and the body cavity is bathed in photosensitizer which is then activated by laser irradiation (39–42). The goal is use of the photosensitizer to kill residual tumor cells that could not be removed by surgery. The drawback is that these photoactive agents cause photosensitivity in light-exposed patients. This feature makes it difficult to use such agents as therapeutics. Thus, other YAP1-targeting agents must be considered. CA3 was identified as reducing YAP1 level and YAP1/TEAD-dependent transcription as measured by suppression of TEAD-dependent transcription (17). The mechanism whereby CA3 inhibits YAP1 function is not clear, and may include interfering with signaling events upstream of YAP1/TEAD. Our findings show that CA3 suppresses YAP1/TAZ/TEAD signaling by reducing LATS1, LATS1-P, YAP1, YAP1-P, TAZ, TAZ-P, and pan-TEAD levels.

The fact that forced expression of constitutively active YAP1 reverses CA3 (and verteporfin) suppression of the MCS cell phenotype suggests that YAP1/TAZ/TEAD signaling is a biologically important target of these agents. In addition, treatment of cells with lactacystin, a proteasome inhibitor, partially reverses YAP1 inhibitor–associated loss of YAP1, TAZ, and pan-TEAD, suggesting that loss of these proteins is via apoptotic degradation.

To have clinical importance, any candidate therapeutic agent must suppress in vivo tumor formation. We show that verteporfin and CA3 suppress tumor formation (Fig. 3). Verteporfin treatment is associated with reduced YAP1/TAZ/TEAD signaling as evidence by reduced TAZ and pan-TEAD levels in verteporfin-treated tumors. However, in contrast to the studies in cultured cells, YAP1 level is not altered. In addition, we observe enhanced tumor cell apoptosis as shown by reduced procaspase-3, -8, and -9, and PARP, and increased active caspase-3 and -9. CA3 also suppresses tumor formation and this is associated with reduced YAP1, TAZ, and pan-TEAD levels. Apoptosis is increased as indicated by loss of procaspase 9 and 8 and PARP and enhance levels of active caspase 3.

These findings are consistent with the observations in cultured cells in that verteporfin suppresses YAP1/TAZ/TEAD signaling and enhances apoptosis. The subtle differences in YAP1/TAZ/TEAD signaling protein modulation in cultured cells versus tumors is likely due to differences in cellular environment. Verteporfin and CA3 also reduce Slug and Snail, which are positive regulators of EMT, suggesting that verteporfin suppresses tumor cell EMT.

We have shown that MCS cells comprise 0.15% of the mesothelioma cell population and that these cells can be enriched by growth as unattached spheroids (8). MCS cells display a highly aggressive phenotype including enhanced spheroid growth, Matrigel invasion, migration, and tumor formation (8). Our current studies show that verteporfin and CA3 suppress the MCS cell phenotype. Moreover, verteporfin and CA3 suppression of the MCS cell phenotype requires reduced YAP1/TAZ/TEAD signaling, as demonstrated by the fact that overexpression of constitutively active YAP1 antagonizes the action of these agents. It is encouraging that the highly aggressive MCS cell subpopulation is responsive to these agents. On the basis of these studies, we suggest that verteporfin and CA3 suppresses YAP1/TAZ/TEAD signaling leading to induction of apoptosis, which further compromises YAP1/TAZ/TEAD signaling to suppress the highly aggressive MCS cell phenotype (Fig. 4F). These findings suggest that use of YAP1/TEAD targeting agents may be useful in the treatment of mesothelioma. Additional experiments should study combination therapy of chemotherapeutic agents and YAP1/TEAD inhibitors.

No potential conflicts of interest were disclosed.

Conception and design: S. Kandasamy, J.S. Friedberg, H.R. Alexander, R.L. Eckert

Development of methodology: S. Kandasamy, E.A. Rorke

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Kandasamy, G. Adhikary, M.B. Mickle, H.R. Alexander

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Kandasamy, G. Adhikary, R.L. Eckert

Writing, review, and/or revision of the manuscript: S. Kandasamy, J.S. Friedberg, R.L. Eckert

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H.R. Alexander, R.L. Eckert

Study supervision: R.L. Eckert

This work was supported by the NIH (CA211909 and CA184027 to R.L. Eckert).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.