A Novel Irreversible TEAD Inhibitor, SWTX-143, Blocks Hippo Pathway Transcriptional Output and Causes Tumor Regression in Preclinical Mesothelioma Models

This article is featured in Selected Articles from This Issue, p. 1

The Hippo pathway has emerged as a promising target for cancer therapy due to the prominent hyperactivation of the Hippo pathway effectors YAP and TAZ in many human solid cancer types (1–6). YAP and TAZ are transcriptional coactivators that form a complex with partner transcription factors, primarily with the TEAD transcription factors (1–6). The TEAD transcription factors have four paralogs, TEAD1–4, that share highly conserved DNA-binding and YAP/TAZ interaction domains. The binding of YAP and TAZ to TEADs stimulates the expression of target genes such as CYR61, CTGF, CCND1, and MYC that promote cell proliferation, survival, and plasticity (1–6). In normal homeostasis, the Hippo pathway core kinases LATS1/2 phosphorylate and regulate the stability and nuclear localization of YAP and TAZ thereby preventing their inductive effect on target gene expression. Indeed, YAP/TAZ are largely dispensable for normal homeostasis of adult mice, showing no overt phenotypes upon genetic codeletion of Yap and Taz in many different adult organs (2, 5, 7–9). However, when YAP/TAZ are hyperactivated, for example, by loss-of-function mutations in LATS1/2, unphosphorylated YAP/TAZ can translocate to nucleus and activate transcription of downstream target genes thereby driving tumor progression, metastasis, and therapy resistance (10, 11). Thus, the broad activation of YAP/TAZ in solid tumors, their importance for cancer cell phenotypes but their irrelevance for homeostasis makes them promising targets for cancer therapy.

Mesothelioma is the primary indication for treatment with YAP/TAZ-TEAD inhibitors due to the high frequency of mutations in Hippo pathway components (12–15). Mesothelioma is a highly aggressive cancer that originates from the mesothelium, an epithelial cell layer that covers the lungs (pleura), heart (pericardium), and abdominal cavity (peritoneum) and that provides a nonadhesive surface to allow organ movement (16). However, most mesothelioma develop in the pleura and are called malignant pleural mesothelioma (MPM). Around 80% of all MPM cases are linked to asbestos exposure and the disease is often diagnosed at advanced stages after a long latency resulting in a poor prognosis (16). Currently, no effective therapies are available (17). Genetic and transcriptional analyses showed that mutations in genes encoding Hippo pathway components are very frequent (up to 75%) with NF2 or LATS2 being affected in up to 50% of MPMs (13–15). Also, knockdown of YAP in mesothelioma cells impaired cell growth, invasion, and tumor formation (12, 18). Therefore, targeting the Hippo pathway is predicted to be an effective treatment for at least a subgroup of mesothelioma.

Several compounds have been identified that inhibit the effects of YAP/TAZ (19–23). On the basis of their mechanism of action, they can be classified into (i) compounds that stimulate the upstream kinases of the Hippo pathway thereby causing the inhibition of YAP/TAZ activity, (ii) compounds that target YAP/TAZ or TEAD directly and interfere with their binding, and (iii) compounds that act on target genes downstream of YAP/TAZ-TEAD (19, 24). Recent reports describe a novel class of small-molecule YAP/TAZ-TEAD inhibitors including K-975 (20) and VT104 (21), that bind (covalently or noncovalently) into a hydrophobic palmitoylation pocket (so-called P-pocket) of the TEADs, where they interact with a cysteine residue located at the entrance of the P-pocket (25, 26). Palmitoylation of TEAD proteins is essential for their stability and activity, and compound binding blocked autopalmitoylation and YAP/TAZ binding (25, 27–30). These P-pocket inhibitors thus blocked expression of canonical Hippo pathway target genes such as CYR61, CTGF, IGFBP3, and NPPB in different mesothelioma model systems and suppressed the growth of xenograft tumors from Hippo-mutant mesothelioma cell lines (19–23). These results show that P-pocket binders inhibit YAP/TAZ-TEAD transcriptional activity and suppress mesothelioma tumor growth in vitro and in vivo.

While mutations in Hippo pathway genes are common in MPM, genetic hyperactivation of YAP/TAZ-TEAD occurs in a variety of other solid tumor types (1–6, 31). In particular, biallelic inactivation of NF2 commonly occurs in aggressive kidney cancer subtypes including clear-cell renal cell carcinoma with sarcomatoid dedifferentiation (19%; refs. 32, 33), type II papillary RCC (13%; refs. 34–36) and unclassified RCC (18%; ref. 37), for which there are no good therapies. Using inducible knockout of Yap and Taz, White and colleagues found that YAP/TAZ are required for the maintenance of NF2-deficient kidney cancer tumors in mice (38). However, their sensitivity to pharmacologic YAP/TAZ-TEAD inhibitors has not yet been described.

Here, we report the identification and characterization of a novel covalent and irreversible YAP/TAZ-TEAD inhibitor, SWTX-143. We describe the direct binding of SWTX-143 to all TEAD isoforms at atomic resolution, as well as its selective transcriptional and phenotypic impact on various tumor models characterized by Hippo pathway alterations. We demonstrate a significant reversion of mesothelioma tumorigenesis and rescue of survival in a genetically induced orthotopic mesothelioma mouse model. Finally, SWTX-143 also potently reduced growth of NF2-mutant kidney cancer cell lines, indicating that the sensitivity of mesothelioma models to YAP/TAZ-TEAD inhibitors may be applicable to other Hippo-mutant tumor subsets.

NCI-H226 (RRID:CVCL_1544), NCI-H2052 (RRID:CVCL_1518), NCI-H28 (RRID:CVCL_1555), HEK293 (RRID:CVCL_0045), and ACHN (RRID:CVCL_1067) were purchased from ATCC and UO31 (RRID:CVCL_1911) and SN12C (RRID:CVCL_1705) were acquired from NIH at the DCTD TUMOR REPOSITORY. Mutational data for each cell line was obtained from cBioportal.org. Additional cell authentication was not performed. Presence of Mycoplasma was tested by using the Lonza kit (MycoAlert). Cell passages were below 25 (starting from the original vials from the suppliers).

Palbociclib and cisplatin were acquired at Tocris. Chemical synthesis and preparative chiral SFC (supercritical fluid chromatography) of SWTX-143 is described in WO2022072741 (Cpd. no. 075, p. 224–225; analytic data: p. 291; ref. 39). Briefly, we used a Waters Thar SFC-80 instrument with a Chiralpak AD-H (5 μm, 30 × 250 mm) column, CO₂ as the mobile phase, MeOH as the cosolvent at 30°C, and detection at 214 nm. Elution was with 80% CO₂ and 20% EtOH, a flow of 60 g/minute, and 1,500 psi ABPR at 30°C.

Human TEAD1–4 YAP-binding domains were expressed as His-tagged proteins in BL21 (DE3) cells. After sonication in lysis buffer (50 mmol/L Tris HCl pH 7.5, 250 mmol/L NaCl, 10% glycerol, 10 mmol/L imidazole) TEADs were purified by affinity chromatography (IMAC-5 mL FF). Proteins were eluted by imidazole, His-tags removed by thrombin cleavage, and further purified by size exclusion chromatography (Superdex S75 16/60) in 20 mmol/L TrisHCl pH7.0; 150 mmol/L NaCl; 5% Glycerol; 2 mmol/L dithiothreitol (DTT). For crystallization, TEAD1 was concentrated to 10 mg/mL. A 3-molar excess of compound SWTX-143 was added to the protein prior to crystallization trials.

TEAD1 crystals grew in drops (Mosquito robot SPT Labtech) at 22°C in 2 mol/L Ammonium sulfate. Data were collected on the PROXIMA-2 beamline at SOLEIL synchrotron. Data were processed with autoPROC/stararaniso (GlobalPhasing), structures solved by molecular replacement with Phaser using an in-house structure of TEAD1 as search model, and refined by iterative cycles using Buster, and manually fitted with Coot (RRID:SCR_014222). Model quality was monitored with Rampage.

TEAD stock solutions were desalted on Zeba columns against AcONH₄ 200 mmol/L pH7.5. Protein concentration was adjusted to 40 μmol/L in AcONH₄ 200 mmol/L pH7.5. When desired, Palmitoyl CoA was added (5eq, room temperature, 30 minutes), followed by desalting on Zeba columns. For native mass spectrometry (MS) analysis, a final protein concentration of 10 μmol/L in 50 mmol/L AcONH₄ was used. Denaturing MS analysis was done at 1 μmol/L protein in ACN/H₂O/FA, 50/50/0.5 solution.

Data interpretation was performed using MassLynx v4.1 software (Waters, RRID:SCR_014271) and masses were calculated using the peak picking method. Because the molecular isotopic cluster was not fully resolved, average molecular masses were measured.

Thermal shift assays were performed using the StepOnePlus Real-Time PCR System (Applied Biosystems). Compound predilutions were made to maintain an equal percentage of DMSO. The four TEAD proteins were incubated with compound for 30 minutes at room temperature in 25 mmol/L HEPES, 150 mmol/L NaCl, pH 7.5 buffer and analyzed according to the Applied Biosystems protocol.

HEK293 cells were transiently transfected with the 8xGTIIC TEAD luciferase reporter (RRID:Addgene_34615) using JetPEI (Polyplus) and plated with prediluted compounds. Luciferase levels were measured after 24 hours using Perkin Elmer SteadyLite substrate on an Envision reader (PerkinElmer). The ability of a test compound to inhibit this activity was determined as: Percentage inhibition = [1 − ((RLU with vehicle − RLU with test compound) divided by (RLU with vehicle − RLU with positive control inhibitor))] * 100.

Cell lines were seeded, medium with compound was added the next day (day 1), and cell numbers were determined on day 6 by using ATPLite (Perkin Elmer 6016943) or with CellTiterGlo. The Click-iT EdU proliferation assay (Thermo Fisher scientific, # C10499) was applied to NCI-H226 cells that were seeded at 1,500 cells/well in a 96-well plate and incubated overnight. Active Caspase3/7 were detected using the Caspase-Glo assay (Promega, G8091) on NCI-H226 cells.

NCI-H226 and ACHN cells were seeded at 600,000 cells per well and incubated with SWTX-143 overnight (100 nmol/L). RNA was extracted using RNeasy Kit (Qiagen, #74004), treated with DNase (Qiagen, #79254), and stored at −80°C. Samples were sequenced at BGI. mRNA fragments were purified, used for cDNA library preparation, and analyzed through the BGISEQ-500 platform (RRID:SCR_017979). Quality control and sequence trimming was done with FastQC and Cutadapt followed by mapping to the human genome (GRCh37, hg19) using STAR (RRID:SCR_004463). After filtering lowly expressed genes (<1 cpm), differential gene expression analysis was performed using the DESeq2 R-package (RRID:SCR_000154). Genes were considered significantly downregulated, when adjusted P values were < 0.05 and log₂ fold changes were <−0.6. Heat maps show row-normalized log₂ scale expressions as obtained by “regularized log” transformation from DESeq2. Gene set enrichment analysis (GSEA) was performed by the GSEA algorithm using clusterProfiler and visualized using enrichplot (RRID:SCR_003199). The motif enrichment analysis was done using i-CisTarget (40) with default settings. The YAP signature was from (31).

SWTX-143 was formulated in PEG-400 40%, 1,2-Propanediol 30%, and H₂O 30% and dosed orally at the indicated concentrations. Bioanalysis of SWTX-143 in plasma was determined using standard LC/MS-MS procedures.

A total of 10 × 10⁶ NCI-H226 cells in 0.2 mL PBS mixed with 1:1 Matrigel (Corning Inc) were subcutaneously inoculated into the flank of female BALB/c nude mice (RRID:IMSR_ORNL:IE-BALB/c). Mice were randomized when the mean tumor volume reached approximately 170 mm³ based on “matched distribution” method using tumor volumes (StudyDirector software, version 3.1.399.19). Tumor volumes were measured twice per week and calculated using the formula: “V = (L × W × W)/2, with L = tumor length (mm) and W = tumor width (longest dimension perpendicular to L). Mice were monitored daily and sacrificed when the humane endpoint was reached according to the national guidelines (tumor volume >3,000 mm³ or >20% body weight loss relative to the weight at the start).

Lats1^fl/fl; Lats2^fl/fl (41) were crossed to p53^fl/fl mice (42) to produce p53^fl/fl;Lats1^fl/fl;Lats2^fl/fl mice. SWTX-143 was administered at 30 mg/kg via gavage twice daily for 2 weeks. Adult mice were sedated with Ketamine:Xylazine solution and 10⁸ Adeno-Cre virus particles (VectorBiolabs dE1/E3, #1700) were injected intrapleurally by fixing mice on their back and inserting the needle between the ribs through the chest wall (43). All mouse experiments were approved by the institutional ethical commission at KU Leuven and performed in accordance with relevant institutional and national guidelines and regulations.

Mouse lungs were fixed in 4% paraformaldehyde (VWR) in 1× PBS for 48 hours at 4°C, embedded in paraffin and cut (5 μm) using a microtome. Paraffin sections were stained with hematoxylin and eosin (H&E) and imaged on a Axio Scan.Z1 slide scanner. For immunofluorescence staining, samples were dewaxed, rehydrated in several steps, permeabilized, blocked for 2 hours in 3% BSA at room temperature, and incubated overnight with primary antibodies [WT1 (RRID:AB_2043201); Mesothelin (RRID:AB_1279130); Vimentin (RRID:AB_10695459)]. The following day, samples were washed in PBS, incubated in secondary antibody solution for 1 hour at room temperature [Cy3-conjugated donkey anti-rabbit IgG, Jackson, (RRID:AB_2313568), Cy3-conjugated donkey anti-rat IgG, (RRID:AB_2340669), and 488-conjugated donkey anti-rabbit, (RRID:AB_2313584)] and mounted in Mowiol with the nuclear marker DAPI (4′,6-diamidino-2-phenylindole) RRID:AB_2307445 (Sigma), and scanned on an Olympus FV1200 confocal microscope. The immunofluorescent stainings were analyzed to determine mesothelial layer thickness and to confirm cell fate using the mesothelin stainings at high magnification (40× and 60× objectives).

ImageJ software was used for quantifications (RRID:SCR_003070). Mesothelioma cell layer thickness was quantified on H&E stainings by measuring the distance from the basal membrane to the edge of the mesothelium at multiple points regularly distributed around the perimeter of the section. These data are presented as the mean ± SEM and groups were tested for significant differences using unpaired t tests in which a P value of 0,05 was considered statistically significant.

The RNA-seq data can be accessed by GSE222963. The X-ray structure is deposited in the PBD (8Q68).

To identify compounds that bind to the P-pocket of TEADs, we initiated a structure-based drug design approach using the publicly available structural information at the time [Apo-TEAD2 and TEAD2 in complex with ligands such as flufenamic acid (25) and myristic acid (28)]. A native MS assay was established to determine binding of compounds to the P-pocket of TEAD1. Upon purification from bacteria, the YAP-binding domain of TEAD1 (TEAD1^R209-D426) is typically moderately palmitoylated (up to 15%) but can be fully covalently palmitoylated using palmitoyl CoA (detected by MS shift of 236 Da). We thus determined the extent and specificity of compound binding to the P-pocket by comparing the percentage of binding to the native (poorly palmitoylated) versus the fully palmitoylated protein using native MS. Several iterative cycles of medicinal chemistry to optimize selective TEAD P-pocket binders led to the identification of SWTX-143 (Fig. 1A; medicinal chemistry optimization to be described in detail elsewhere; ref. 39). SWTX-143 binds to native TEAD1 (67% and 92% binding at 5 and 20 μmol/L, respectively) but did not bind to the fully palmitoylated TEAD1 protein (0% and 9% binding at 5 and 20 μmol/L, respectively), indicating that SWTX-143 exclusively binds into the P-pocket of TEAD1 (Fig. 1B and C). An MS-based assay under denaturing conditions further demonstrated that SWTX-143 was covalently bound to TEAD1 protein at 20 μmol/L, consistent with the presence of an acrylamide electrophile in the molecule (Fig. 1D).

To elucidate the binding mode of SWTX-143 to TEAD1 at atomic resolution, we soaked SWTX-143 into crystals of TEAD1^R209-D426. X-Ray data collection of the complex to a resolution of 1.6 Å and crystallographic structure refinement showed that SWTX-143 occupied up to 80% of the volume of the P-pocket and that it binds in an orientation suitable to form a covalent bond with Cys359 via its acrylamide function (Fig. 1E; Supplementary Materials and Methods).

To determine whether SWTX-143 can bind to all four members of the TEAD family, thermal shift assays were performed on the YAP-binding domains of the four TEAD isoforms (TEAD1^R209-D426, TEAD2^A217-D447, TEAD3^R218-D435, TEAD4^S217-E434). When coincubating SWTX-143 (10 μmol/L) with the TEAD protein domains, a significant thermal stabilization was observed for all four TEAD isoforms (ranging from 3°C to 9°C), demonstrating that SWTX-143 is a pan-TEAD binder (Fig. 1F). Interestingly, the modest binding of the SWTX-143 racemic mixture to TEAD3 was due to the lack of TEAD3 binding to one of the enantiomers, SWTX-143-A, which did not stabilize TEAD3 under the conditions tested, whereas the alternate enantiomer, SWTX-143-B, clearly bound to TEAD3 (stabilization of 3.78°C; Fig. 1F). These data show that subtle conformational differences between compounds can impact TEAD isoform binding preference, despite the highly conserved nature of the P-pocket across all isoforms.

Autopalmitoylation of TEAD proteins impacts YAP/TAZ-TEAD complex formation and stability (27, 28). To investigate the impact of our TEAD P-pocket binders on TEAD transcriptional activity, we tested SWTX-143, along with the pure enantiomers, SWTX-143-A and SWTX-143-B, in HEK293 cells transiently transfected with a TEAD-reporter construct that expressed firefly luciferase under the control of eight repeats of a TEAD responsive element (44). SWTX-143 potently repressed luciferase expression from the TEAD reporter (IC₅₀ = 12 nmol/L) while not affecting the viability of HEK293 cells (Fig. 2A). No significant differences between the enantiomers of SWTX-143 were observed indicating that TEAD3 does not significantly contribute to YAP/TAZ-TEAD transcriptional output in this assay consistent with its low expression levels (Fig. 2B). Moreover, the action of SWTX-143 was selective as it did not affect luciferase expression driven by a HIF responsive element (HRE; Fig. 2B).

Because we demonstrated covalent binding of SWTX-143 to TEAD1 protein, we examined the reversibility of TEAD transcriptional inhibition by SWTX-143 in this cellular reporter system. HEK293 cells were transiently transfected with the TEAD-reporter construct and incubated for 3, 6, 12, or 24 hours with SWTX-143, after which the compound-containing medium was removed and replaced with medium for a total period of 24 hours when reporter activity was measured (Fig. 2C). Incubation of the transfected cells for as short as 3 hours with SWTX-143, followed by a washout period of 21 hours, effectively reduced transcriptional reporter activity (IC₅₀ = 95 nmol/L), consistent with a fast and irreversible mechanism of action of SWTX-143. Longer incubation for 6, 12, or 24 hours still moderately increased the potency of transcriptional repression (IC₅₀ = 58, 34, and 16 nmol/L, respectively), as expected for a reactive covalent inhibitor. Thus, a short exposure of tumor cells to SWTX-143 may be sufficient to evoke a long-lasting transcriptional impact.

To determine the specificity of SWTX-143 on gene expression in more detail, we turned to the Hippo pathway–inactivated (NF2-deficient) mesothelioma cell line, NCI-H226. We measured the impact of 100 nmol/L SWTX-143 on the whole transcriptome of this cell line by RNA-seq after 6 hours of compound incubation. Differential expression analysis identified 327 downregulated and 380 upregulated genes after SWTX-143 treatment compared with vehicle treatment (adjusted P value < 0.05; log₂FC > 0.6). A YAP signature was strongly repressed by SWTX-143 (Fig. 2D and E; ref. 31), but the transcriptional output of most other signaling pathways was not strongly affected (Fig. 2F). To further investigate the regulation of the downregulated genes, we employed iCis-target to identify transcription factor binding motifs that were enriched in the 20 kb genomic region surrounding their transcription start site (40). This identified binding sites for TEAD factors as the top hit with a high enrichment score (NES = 8.3) and binding sites for E2F family members as the only other enriched motifs (NES = 6.54; Fig. 2G). Strikingly, 80% of the downregulated genes (262/327 genes) had TEAD binding sites, indicating they could be direct targets of TEAD factors (Fig. 2H and I). Altogether, these data demonstrate that SWTX-143 represses YAP/TAZ-TEAD transcriptional activity in a selective manner.

MPM exhibits recurrent genetic alterations in NF2 and LATS2, resulting in constitutive YAP/TAZ activity (13–15). Here, we tested the impact of SWTX-143 on the proliferation of three distinct mesothelioma cancer cell lines that are characterized by and depend on loss-of-function alterations in NF2 (Mero-14, NCI-H226, and NCI-H2052; Supplementary Fig. S1A) and one that is deficient in LATS1 and LATS2 (MSTO-211H). Strikingly, proliferation of all four cell lines was highly sensitive to SWTX-143 (IC₅₀ ranging between 5 and 207 nmol/L; Fig. 3A and B). On the other hand, SWTX-143 treatment only modestly impacted the proliferation of the NCI-H28 and NCI-H2452 mesothelioma, HeLa, SiHa, and CaSki cancer cell lines which lack genetic alterations in known Hippo pathway components [Hippo-WT (wild-type); Fig. 3A and B; Supplementary Fig. S1B].

To decipher the mechanism responsible for the observed decrease in cell number over time in Hippo-mutant mesothelioma cell lines, we distinguished whether SWTX-143 was cytotoxic (inducing apoptosis) or cytostatic (promoting cell-cycle arrest). SWTX-143 did not promote apoptosis in the NF2-deficient mesothelioma cell line NCI-H226, in contrast to cisplatin (Fig. 3C). Rather, SWTX-143 induced cell-cycle arrest with similar kinetics as the CDK4/6 inhibitor palbociclib, as shown by a strong reduction of 5-ethynyl-2′-deoxyuridine (EdU)-positive cells over time (Fig. 3D). Thus, SWTX-143 acts by inhibiting the proliferation of Hippo-mutant mesothelioma cells in vitro.

In addition to mesothelioma, a subset of kidney cancers exhibits frequent mutations in Hippo pathway genes. We thus investigated the response of a small panel of kidney cancer cell lines to SWTX-143 and its enantiomers (Fig. 3E and F). Interestingly, cell lines carrying mutations in Hippo pathway components (SN12C NF2^LOF, UO31 LATS2^A324V, and ACHN NF2^R57*/LATS2^LOF) were particularly sensitive to SWTX-143 incubation, whereas the proliferation of HEK293 kidney cells (Hippo-WT) was not affected by SWTX-143. Like the specificity of SWTX-143 action in NCI-H226 mesothelioma cells, we observed robust downregulation of canonical YAP/TAZ-TEAD target genes and the YAP signature by SWTX-143 treatment of ACHN kidney cancer cells (NES = −2.19, P = 8.14e-08; Supplementary Fig. S2A and S2B).

Finally, we compared the efficacy of SWTX-143 with the efficacy of the previously identified TEAD P-pocket binders K-975 and VT104 in the TEAD and HIF reporter gene assays and the mesothelioma proliferation assays (Fig. 3G). This showed that SWTX-143 was equally effective in suppressing TEAD transcriptional activity but had higher efficacy in inhibiting the proliferation of mesothelioma cells with lower toxicity in HEK293 and H28 cells compared with K-975 and VT104.

SWTX-143 was optimized to be suitable for oral administration and to have sufficient plasma exposure allowing animal testing (oral bioavailability ≥75% and half-life of about 2 hours). We thus evaluated the efficacy of SWTX-143 in a human mesothelioma xenograft mouse model. We subcutaneously injected NCI-H226 cells and when tumors reached about 170 mm³, mice were randomized into groups and treated once daily with SWTX-143 (10, 25, and 50 mg/kg) or vehicle for 19 days. Although administration of SWTX-143 had little effect on tumor volume during the first week of dosing, a clear dose-dependent tumor regression was observed during the 2 subsequent weeks of dosing (Fig. 4A). Significant antitumor efficacy was observed at the end of the 18 days of compound administration for all dosing regimens (Fig. 4B). Notably, none of the treatments resulted in overt toxicity or loss of body weight of the mice (Fig. 4C). Thus, SWTX-143 treatment causes strong regression of mesothelioma tumors in vivo without significant secondary adverse effects.

To corroborate that these antitumor effects were driven by inhibition of YAP/TAZ-TEAD-driven transcription in the tumor, we analyzed tumor gene expression profiles by RNA-seq. Indeed, increasing doses of SWTX-143 (resulting in dose-proportional increases of plasma exposure) correlated with stronger repression of YAP/TAZ-TEAD target genes in tumors, measured at the endpoint of the study (Fig. 4D; Supplementary Fig. S3). Notably, although SWTX-143 plasma exposure was transient (Fig. 4E), it was sufficient to trigger a sustained repression of YAP/TAZ-TEAD target genes after 4 and 24 hours after the last dose of SWTX-143 (Fig. 4D), probably due to irreversible covalent binding of the compound to the TEAD complex. These results show that SWTX-143 treatment causes a strong and specific inhibition of YAP/TAZ-TEAD target gene expression in vivo.

Next, we investigated the efficacy of SWTX-143 on an orthotopic mesothelioma mouse model that was generated by conditionally deleting p53, Lats1, and Lats2 in mesothelial cells of adult mice. An adenovirus that constitutively expressed Cre (adeno-Cre) was injected into the pleural cavity of p53^fl/fl; Lats1^fl/fl; Lats2^fl/fl mice (Fig. 5A). This virus efficiently infected mesothelial cells, which normally form a single-cell layer covering the pleural cavity and triggered massive mesothelial cell proliferation that developed into a thick layer of ectopic mesothelial cells after 2 weeks (Fig. 5B). These cells expressed the mesothelioma markers WT1, Mesothelin, and Vimentin, which are used in the clinic to diagnose mesothelioma (Fig. 5B; Supplementary Fig. S4A; ref. 45). Soon after these 2 weeks, mutant mice needed to be sacrificed because of labored breathing (Fig. 5C). Mutant mice also had ectopic proliferation of pericardial cells surrounding the heart, while other organs that do not reside in the pleural cavity such as the liver were not affected. Mutant mice were then treated with SWTX-143 or vehicle starting 2 weeks after adeno-Cre injection when they already had advanced mesothelioma and were just a few days before they would normally have had to be sacrificed. Vehicle treatment did not affect mesothelioma progression and mice had to be sacrificed 2–4 days after the start of treatment (Fig. 5D). In contrast, SWTX-143–treated mice had normal breathing within a few days after the start of treatment and survived until the termination of the experiment after 2 weeks of treatment (Fig. 5D).

We then quantified the thickness of the mesothelial cell layer on H&E-stained lung sections and measured the distance from the basal membrane to the edge of the mesothelium at multiple points around the perimeter of the section. At the start of treatment and after a few days of vehicle treatment, the lungs of p53;Lats1;Lats2 triple mutant mice showed a massive expansion of the mesothelial layer from a 30-μm-thick to 50-μm-thick single-cell layer, to a large, over 400-μm-thick layer of about 10 cells stacked on top of each other (Fig. 5B and D). This mesothelial overgrowth dramatically regressed in SWTX-143–treated mice to an average thickness of 80 μm that appeared to be normal with a single-cell layer around most of the lung surface (Fig. 5D). A few areas still had multiple layers of cells; however, these regions displayed many bloated and necrotic cells, and cells that were detaching from the tissue (Supplementary Fig. S4B). These observations were confirmed by specifically detecting mesothelial cells by WT1 antibody staining (Fig. 5B, right). We conclude that SWTX-143 treatment potently caused mesothelioma regression in this orthotopic in vivo mesothelioma model.

The Hippo pathway is an important target for drug discovery because constitutive activation of YAP/TAZ by genetic and nongenetic dysregulation occurs across multiple hard-to-treat solid tumor types and is associated with treatment resistance (1–6). As the TEAD proteins are the main partner transcription factors of YAP/TAZ, targeting the TEAD transcription factors is emerging as a promising therapeutic strategy to treat Hippo-inactivated tumors (3–5). Here, we provide preclinical evidence that direct covalent TEAD P-pocket engagement is remarkably effective to treat mesothelioma, a prime example of a Hippo-mutant cancer. Using a novel, potent and selective pan-TEAD inhibitor, SWTX-143, we demonstrate tumor regression of subcutaneously implanted human mesothelioma cells, and more importantly, near complete normalization of orthotopic mouse mesothelioma and full rescue of survival in a newly developed genetic model. In these mouse studies, we did not observe any overt signs of tolerability alerts, even though YAP/TAZ-TEAD transcription in tumors was profoundly suppressed. More extensive tolerability studies are needed, but this analysis suggests that a therapeutic index can be achieved using this mechanism of YAP/TAZ-TEAD inhibition. Of note, a potential advantage of a covalent irreversible TEAD binder, such as SWTX-143, is the limited systemic exposure needed to achieve long-term sustained target engagement in tumors and efficacy. On the other hand, covalent Cys binders may have the downside of being susceptible to resistance by mutation of this residue, as has been clinically observed for covalent kinase inhibitors such as ibrutinib for example (46). Importantly however, the Cys residue targeted by SWTX-143 (Cys 359 in TEAD1) is autopalmitoylated and important for TEAD functionality which decreases the risk of resistance through mutation of this Cys as this would directly impair the activity of the mutant TEAD protein.

Despite the impressive tumor regression in distinct mouse cancer models in vivo, we did not observe killing of mesothelioma cancer cells by SWTX-143 in vitro. We speculate that YAP/TAZ-TEAD inhibition leads to more impressive tumor cell killing in vivo due to its impact on cell–cell and/or cell–matrix interactions, processes that are abnormal in cells cultured on artificial surfaces, which may make them less vulnerable to YAP/TAZ-TEAD inhibition. A better understanding of the precise mechanism of action of YAP/TAZ-TEAD inhibitors may also lead to rationally selected combination regimens. Recently, a detailed analysis of clinical pleural mesothelioma datasets demonstrated that YAP/TAZ transcriptional markers track with worse outcome (47). Furthermore, loss of NF2 increases tumorigenicity in mesothelioma cells. The data reported here support a clinical stratification strategy for TEAD inhibitors based on YAP/TAZ transcriptional activation and/or loss of NF2. However, further (clinical) research is required to determine what the most optimal predictive biomarker would be to guide patient stratification. Finally, we provided evidence that the efficacy observed with YAP/TAZ-TEAD inhibitors is not limited to mesothelioma models but may be extended to other tumor types with Hippo pathway alterations such as kidney cancer.

A. Candi reports other support from Springworks Therapeutics during the conduct of the study; in addition, A. Candi has a patent for WO2022072741 issued to Springworks Therapeutics. W. Kowalczyk reports grants from FWO during the conduct of the study. M. Nijs reports other support from KU Leuven and SpringWorks during the conduct of the study. E. Soons reports other support from KU Leuven and SpringWorks during the conduct of the study. W. Haeck reports other support from KU Leuven and SpringWorks during the conduct of the study. H. Klaassen reports other support from KUleuven during the conduct of the study. S.A. Spieser reports other support from SpringWorks Therapeutics and KU Leuven during the conduct of the study. A. Marchand reports other support from Springworks Therapeutics and KU Leuven during the conduct of the study; in addition, A. Marchand has a patent for WO2022072741 pending, licensed, and with royalties paid. P. Chaltin reports other support from SpringWorks Therapeutics during the conduct of the study. A. Kamal reports other support from SpringWorks Therpauetics outside the submitted work; and employee of SpringWorks Therapeutics. S.L. Gwaltney reports personal fees from SpringWorks Therapeutics outside the submitted work; in addition, S.L. Gwaltney has a patent for TEAD inhibitors pending to SpringWorks Therapeutics. M. Versele reports other support from SpringWorks Therapeutics during the conduct of the study; in addition, M. Versele has a patent for WO2022072741 pending and licensed to SpringWorks Therapeutics. G.A. Halder reports grants from Stichting tegen kanker and other support from Springworks Therapeutics during the conduct of the study; in addition, G.A. Halder has a patent for WO2022072741 issued to The Katholieke Universiteit Leuven, SpringWorks Therapeutics Inc., Vib Vzw. No disclosures were reported by the other authors.

H. Hillen: Data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. A. Candi: Resources, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. B. Vanderhoydonck: Formal analysis, validation, investigation, methodology, writing–review and editing. W. Kowalczyk: Data curation, software, formal analysis, investigation, visualization. L. Sansores-Garcia: Formal analysis, validation, investigation, methodology, writing–review and editing. E.C. Kesikiadou: Formal analysis, investigation, methodology. L. Van Huffel: Investigation, methodology, project administration. L. Spiessens: Investigation, methodology, project administration. M. Nijs: Validation, investigation, methodology, writing–review and editing. E. Soons: Validation, investigation, methodology, writing–review and editing. W. Haeck: Validation, investigation, methodology, writing–review and editing. H. Klaassen: Validation, investigation, methodology, writing–review and editing. W. Smets: Validation, investigation, visualization, methodology, writing–review and editing. S.A. Spieser: Validation, investigation, writing–review and editing. A. Marchand: Data curation, validation, investigation, visualization, methodology, writing–review and editing. P. Chaltin: Conceptualization, resources, supervision, funding acquisition, project administration, writing–review and editing. F. Ciesielski: Formal analysis, validation, investigation, writing–review and editing. F. Debaene: Formal analysis, validation, investigation, visualization, methodology, writing–review and editing. L. Chen: Validation, investigation, visualization, methodology, writing–review and editing. A. Kamal: Formal analysis, investigation, visualization, methodology, writing–review and editing. S.L. Gwaltney: Conceptualization, resources, supervision, funding acquisition, project administration, writing–review and editing. M. Versele: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration. G.A. Halder: Conceptualization, data curation, formal analysis, supervision, funding acquisition, writing–original draft, project administration.

We are grateful for excellent technical support from Marianne Croonenborghs (Cistim Leuven), Merja Bhailal Chakubhai (Aragen), and Tao Yang (Crownbio). We thank Randy Johnson and Chris Marine for sending mouse strains. This work was supported by grants from Stichting tegen Kanker to G.A. Halder.

This work was supported by SpringWorks Therapeutics internal funding and a Stichting Tegen Kanker grant (ZKE0232-02-W01).