{"paper_id":"4a7704e7-110e-4f0d-bddb-9505e1af0a13","body_text":"1 \nDevelopment of a pipeline to evaluate YAP-TEAD inhibitor potential - direct \nTEAD inhibition represses cancerous growth in a NF2-mutant model of \nmesothelioma  \nRichard Cunningham1, Siyang Jia1, Krishna Purohit1, Michaela Noskova Fairley1, Yue Lin1, Ning \nSze Hui 1, Rebecca E. Graham 3, Adriano G. Rossi1, Justyna Cholewa -Waclaw2, Neil O. \nCarragher4, Carsten Gram Hansen1* \n1 University of Edinburgh, Centre for Inflammation Research, Institute for Regeneration and Repair, Edinburgh \nbioQuarter, 5 Little France Drive, Edinburgh EH16 4UU, UK. \n2 High Content Screening Facility, Institute for Regeneration and Repair, The University of Edinburgh, \nEdinburgh, EH16 4UU, UK. \n3 Centre for Clinical Brain Sciences, Anne Rowling Regenerative Neurology Clinic, University of Edinburgh, \nEdinburgh, EH16 4SB, UK. \n4 Cancer Research UK Scotland Centre, Institute of Genetics and Cancer, University of Edinburgh, Edinburgh , \nEH4 2XR, UK \n*Corresponding author.  \n \nABSTRACT \nAs the core, tumorigenic downstream effectors of the Hippo signalling pathway, YAP/TAZ and \nthe TEAD family of transcription factors represent attractive targets for drug discovery efforts \nwithin cancer research. This is particularly true within the context of pleural mesothelioma, \nin which  there are many recent preclinical developments and clinical trials evaluating the \nefficacy of TEAD inhibitors. The range of inhibitors have shown great promise, but \ncomparisons of their performances are so far limited. Here we develop a high content pipeline \nthat enables a comparative analysis of currently developed YAP/TAZ-TEAD inhibitors. We take \nadvantage of isogenic cellular models that enable us to examine inhibitor specificity. We \nidentify genetic compensation of the Hippo pathway transcriptional module, with \nimplications for therapeutic targeting , and implement Cell Painting to develop a detailed \nmorphological profiling pipeline that enables further characterisation, quantification, and \nanalysis of off-target effects. Our pipeline is scalable and allows us to establish specificity and \ncomparative potency within cancer relevant assays in a clinically relevant cellular model.   \n \nINTRODUCTION \nPleural mesothelioma is an asbestos induced cancer of the mesothelial lining of the lung and \nthe deadliest cancer after diagnosis1,2. Patients have a bleak outlook as many are diagnosed \nlate or with aggressive disease, limiting the efficacy of surgical intervention3,4. There are no \ncurative therapies. To this end, therapeutics currently used can be considered palliative care \nand patients only survive on average for 12-18 months after diagnosis. This highlights a clear \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 2 \nand urgent need for the development of targeted therapies and implementation of effective \nchemotherapeutics5. As a relatively uncommon cancer-type, pleural mesothelioma (PM) is \ndefined by a lack of clear, high penetrance driver mutations6–10. Despite this, PM is distinct \nfrom most common cancers in that genomic perturbations within upstream components of \nthe Hippo pathway are overrepresented within its mutational profile. \nThe Hippo pathway comprises an upstream kinase cascade module, which functions to \nregulate the activity of the co-transcriptional activators YAP and TAZ11,12. These, alongside \nthe TEAD family of transcription factors13,14, represent the downstream transcriptional \neffectors of the Hippo pathway. Since the initial discovery of the signalling pathway as a \ncore regulator of a variety of processes in development, differentiation, and \nregeneration15,16, YAP and TAZ are implicated as oncogenic drivers across a range of cancer-\ntypes17–24. Strikingly, PM is defined by a relatively high frequency of loss-of-function \nmutations within the upstream regulatory component of the Hippo pathway relative to \nother cancer types6–10. These mutations include frequent loss-of-function mutations within \nthe upstream Hippo pathway kinase cascade, including most frequently in NF2, with more \noccasional loss-of-function mutations in SAV1, LATS1 and LATS26–10.  \nWhile designing therapeutics specifically against loss of tumour suppressors has been \nchallenging historically, multiple inhibitors of the TEAD family of transcription factors have \nrecently been developed with the hopes to position for clinical use. The realisation of the \ndevelopment of a range of varied direct inhibitors is an exciting advancement in cancer \nresearch, as hyperactive YAP/TAZ-TEAD activity is a widespread phenomenon across many \ncancers25. The majority of the inhibitors developed comprise small molecule disruptors of \nTEAD auto-palmitoylation, a post translational modification required for the interaction \nbetween these transcription factors and YAP/TAZ26,27. Following initial preclinical \ndevelopment, some of these promising inhibitors have now progressed to clinical trials, to \nbe positioned as YAP/TAZ-TEAD inhibitors in PM patients. Due to the genomic evidence and \nfrequent dysregulation of the Hippo pathway within PM, combined with both the \nineffectiveness of surgery28 and general dearth of effective therapeutic interventions10,29, \nthis cancer-type is the primary focus in the development of these therapeutics.  \nBeyond these selective inhibitors that are currently undergoing testing, there are a variety \nof indirect inhibitors of YAP/TAZ-TEAD. These comprise a heterogenous range of chemicals, \nincluding dasatinib, a Src kinase inhibitor commonly used as an anti-cancer agent30,31, \nverteporfin, a photosensitiser historically used as a chemical inhibitor of YAP/TAZ \nactivity32,33, as well as statins34, a family of mevalonate pathway inhibitors reported to \ninhibit YAP/TAZ activity via the disruption of geranylgeranylation19. We focus on these \ndifferent therapeutic classes and prioritize a range of inhibitors with a focus on those that \nhave well-documented, potent anti-YAP/TAZ activity. By focusing on a range of both \nselective and classical, less selective inhibitors, we compare these side-by-side, taking \nadvantage of two separate and complimentary isogenic cellular models to facilitate a direct \ncomparison in the molecular impact of therapeutic families within a clean genetic \nbackground. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 3 \nRESULTS \nIdentifying high potency inhibitors of YAP/TAZ-TEAD \nWith the current focus on YAP-TEAD inhibition in the development of chemotherapeutics \nfor use, especially for the treatment of PM29,35, three recently developed TEAD inhibitors \nwere selected for testing within cell-lines. This includes VT-10736 and K-97537,38, both pan-\nTEAD inhibitors, and IK-930, a TEAD1-specific inhibitor39 that has undergone clinical trials40 \nfor the treatment of PM and other tumours associated with perturbations within the Hippo \npathway (NCT05228015). A multitude of additional inhibitors of YAP-TAZ/TEAD have in the \npast been identified through screening approaches and used primarily as tool compounds to \nprobe pre-clinical impacts of Hippo pathway perturbations, as well as undergoing further \npreclinical development and optimization29,41. Despite widespread use, most of these have \nless well-defined mechanisms of action (MoA) and might have YAP-TAZ/TEAD inhibition as a \nsecondary and potential off-target effect33,42. To compare the inhibitory potential of \nrecently developed, on-target TEAD inhibitors to these more classic YAP inhibitors, we \nincluded verteporfin, lovastatin, and dasatinib as comparators. Dasatinib is prioritized, as its \nprimary mechanism-of-action as an anti-cancer therapeutic is via its function as a Src and \ntyrosine kinase inhibitor30,31, with Dasatinib’s downstream cellular impact on YAP/TAZ-TEAD \nappearing likely through the Src-YAP signalling axis43,44. \nInitial testing was focused on the ability of compounds to inhibit cellular TEAD activity. This \nwas analysed via the use of a luciferase reporter regulated by multiple TEAD binding sites45. \nQuantification of TEAD activity upon 24-hour treatment at 1 µM confirmed the ability of \nalmost all tested compounds to inhibit TEAD activity (figure 1A), apart from lovastatin. \nInterestingly, dasatinib exhibits the most profound inhibition of TEAD activity. In order to \nvalidate this effect, we quantified the expression of YAP/TAZ signature genes, as defined by \nTCGA46, on treatment with a selection of these inhibitors. This reveals a modest, though \nsignificant, decrease in signature expression (figure 1B) under the same treatment \nconditions. Notably, the observed decrease in signature gene expression is TEAD-\ndependent, with no significant effect evident on treatment in TEAD KO cells14 \n(supplementary figure 1A). This suggests that not only TEAD inhibitors, but also off-target \ninhibitors such as verteporfin, modulate the expression of YAP/TAZ target genes in a TEAD-\ndependent manner. Given the difficulties associated in directly targeting YAP/TAZ directly \ndue to their intrinsic disorder29, these findings reinforce the likelihood that YAP/TAZ may be \nprimarily targetable via their interaction with TEADs29. \nIn order to probe the efficacy of candidate compounds within the context of mesothelioma, \nevaluation was then conducted in a preclinical model of PM driver mutations, with CRISPR-\nCas9 mediated knockout47 (KO) of tumour suppressor NF2 in non-malignant, MeT-5A \nmesothelial cells. This unique isogenic cellular model represents the disease well6 and is an \nideal platform for interrogating YAP/TAZ-TEAD dynamics and responses to therapeutic \ninhibition. We have previously shown that the TCGA-defined YAP/TAZ signature gene-set46 \nis upregulated on NF2 loss in the mesothelial context under cancer-related stresses6. Initial \ntesting focused on the effect of treatment on YAP/TAZ-TEAD mediated transcription via \nquantifying the expression of these canonical downstream target genes. Computing the \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 4 \nexpression of this signature reveals that almost all tested inhibitors were sufficient to \nsignificantly downregulate the transcription of the full gene-set (figure 1C) after 24-hour \ntreatment at 1 µM, reinforcing their abilities to suppress YAP/TAZ-TEAD activity. \nInterestingly, IK-930, a TEAD1-specific inhibitor, did not decrease signature expression under \nthese conditions (figure 1C). \nYAP/TAZ-TEAD activity is contingent on nuclear retention of YAP/TAZ, with the upstream \nkinase component of the Hippo pathway known to phosphorylate YAP/TAZ, leading to \nsubsequent cytoplasmic sequestration and therefore inhibition of these co-transcriptional \nactivators48,49. To assess the dynamics of YAP/TAZ alongside the observed reduction in \nYAP/TAZ-TEAD mediated transcription, we next quantified the inhibitory potential of \ncandidate compounds across a range of concentrations in the context of YAP nuclear \nlocalisation. YAP was selected as the sole readout of activity here, with cellular TAZ \ndynamics omitted from study, due to YAP and not TAZ being the prime determinant of \ntumorigenicity within the MeT-5A cellular model6. Additionally, as inhibitors are presumed \nto be cytostatic/cytotoxic and YAP nuclear localisation is reduced under conditions of high \ncell density29,50,51, cells were filtered to those approximating 50% cell-cell contact to limit \neffects from contact inhibition at either extreme6,51. Although lower potency analogues of \nVT-107 and K-975 apparently do not impact nuclear YAP retention in mesothelioma cell-line \nmodels36 and the osteosarcoma U2OS cell-line52 respectively, we identified that the majority \nof selective TEAD inhibitors tested did modulate YAP localisation 24 hours post-treatment. \nThis effect is observed by increased cytoplasmic/nuclear YAP ratio (figure 1D), indicating a \nreduction in the level of transcriptionally active YAP, with the maximal response generally \nobserved at the lowest concentration point included (370 nM; supplementary figure 1B). \nReduction in YAP nuclear localisation broadly matches decreased YAP/TAZ signature gene \nexpression, with a clear effect observed when cells were treated with pan-TEAD inhibitors. \nHowever, in contrast to its effect on YAP-TEAD gene expression, there was no significant \ndecrease in nuclear YAP with verteporfin treatment, with dasatinib being the prime non-\nselective inhibitor that induced cytoplasmic retention of YAP in MeT-5A cells (figure 1E). \nWith the inhibitory potential of these compounds established within MeT-5A cells, we next \nsought to assess this effect within the context of loss of the tumour suppressor NF2. \nNotably, loss of NF2 had little impact on sensitivity to YAP/TAZ-TEAD inhibition in terms of \nexpression of YAP/TAZ signature genes, with just a slight decrease in sensitivity to K-975 \nobserved (supplementary figure 1C), while NF2 KO cells exhibited slightly enhanced \nsensitivity to selective TEAD inhibition on YAP nuclear localisation (supplementary figure \n1D). Though this effect was significant, it was modest, while sensitivity to dasatinib was \nconversely decreased in NF2 KO cells. \nYAP’s principal nuclear/cytoplasmic regulation primarily takes place via LATS1/2 mediated \ninhibitive phosphorylation on five serine residues49,53. In order to interrogate the manner \nwith which this YAP inhibition is mediated, we analysed cellular lysates after 24-hour \ntreatment at 370 nM and 1 µM, performing Phos-Tag western-blotting, a sensitive \ntechnique that allows for the identification of the phosphorylation state of YAP49,53. This \nshows that dasatinib is the sole tested inhibitor that induces YAP phosphorylation at low \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 5 \nconcentrations (figure 1F), while lovastatin treatment results in YAP phosphorylation only at \nhigher concentrations (supplementary figure 1E). This YAP phosphorylation indicates that \ndasatinib’s potent inhibitory effect on YAP/TAZ-TEAD (figure 1F), as well as lovastatin’s more \nmodest effect (supplementary figure 1E), may be mediated via the core kinase module of \nthe Hippo pathway. To explore this, we conducted Phos-Tag based western-blots using \nlysates from similarly treated LATS1/2 double knockout (dKO) HEK293A cells54–56. Cells \nlacking LATS1/2 have, as expected, less phosphorylated YAP, indicating nuclear and \ntherefore hyperactive YAP (supplementary figure 1F). These cells also show no YAP \nphosphorylation response to treatment with lovastatin or dasatinib, while cells with intact \nLATS1/2 exhibit TEAD-independent phosphorylation of YAP on treatment (supplementary \nfigure 1F). Combined, the results highlight that YAP phosphorylation induced by lovastatin \nor dasatinib treatment is dependent on LATS1/2 kinases and therefore likely mediated via \nthe Hippo pathway kinase cascade. \nInterestingly, the increase in phosphorylated YAP observed in NF2 KO MeT-5A cells upon \ndasatinib treatment is diminished relative to WT MeT-5A cells (figures 1F-G). This blunted \nresponse is consistent with the decreased sensitivity of NF2 KO cells in YAP nuclear \nlocalisation to dasatinib (supplementary figure 1D). Considered together, these findings \nsuggest that NF2-loss may result in a resistance to inhibition of YAP/TAZ-TEAD via activation \nof the Hippo pathway’s core kinase cascade, while leaving NF2-deficient cells similarly \nvulnerable to TEAD inhibition as compared to those with intact NF2. \n \nYAP/TAZ-TEAD inhibition disrupts the tumorigenic potential of NF2-deficient mesothelial \ncells \nWith the activity of the candidate compounds established, we next sought to determine the \nin vitro anticancer potential of candidate therapeutics in the context of NF2-deficient \nmesothelioma. Initially, this was assessed by quantifying YAP/TAZ-TEAD inhibition’s impact \non cellular viability. Proliferation assays performed over the course of 72 hours (figure 2A) \nrevealed that selective TEAD inhibitors generally had little effect on cell growth (figure 2B). \nOnly VT-107 showed a consistent, though modest, reduction over the course of 72 hours \ntreatment. This was in stark contrast to non-selective inhibitors of YAP, which caused a \ncomplete cessation of growth at concentrations >1 µM (figure 2B). \nOn loss of NF2, sensitivities across all treatments remain almost equivalent to wildtype (WT) \nMeT-5A cells (figure 2C, supplementary figure 2A). This is somewhat unexpected, as recent \nfindings have indicated that NF2-deficient mesothelioma cells are more sensitive to TEAD \ninhibitors57. Importantly, our studies are conducted on an isogenic background and are \ntherefore not impacted by the complexities of comparing different heterogenous cell lines \nwith varying epigenetic backgrounds and underlying genetic diversity beyond NF2 status. \nCombined, these additional factors might modulate sensitivity to TEAD inhibition, \nobfuscating nuanced molecular dynamics. Interestingly, loss-of-viability induced by non-\nselective YAP inhibitors appears to be YAP-independent, with YAP KO MeT-5A cells \nexhibiting similar sensitivities as WT cells to all compounds tested (figure 2D). YAP-\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 6 \nindependence therefore might explain the initial observation that NF2 KO cells are equally \nsensitive to dasatinib in terms of proliferation (figure 2C, supplementary figure 2A), despite \nthe relative insensitivity in terms of YAP activation (figures 1F-G).  \nGiven the role of NF2 in regulating YAP/TAZ in response to mechanical stimuli, including \ncontact inhibition, substrate stiffness, and the mechanotransduction mediated by the \nYAP/TAZ-TEAD axis, we next focused on the establishment of a functional assay capable of \nquantifying cellular response to these cancer-adjacent mechanical stresses. Previously, we \nidentified NF2 as an upstream mechanoregulator in mesothelial cells, and that mesothelial \nNF2-loss is sufficient to drive YAP-mediated anchorage-independent growth6. As the soft \nagar assay is challenging to scale up for high-throughput assays, we sought to establish an \nanalogue approach. We therefore implemented a spheroid formation assay to analyse the \neffects of the selected inhibitors in a scalable manner. When MeT-5A cells are cultured in an \nultra-low attachment setting, they readily form spheroids, with spheroids exhibiting less \nnuclear YAP than cells cultured in 2D (supplementary figure 2B). Interestingly in spheroids, \nNF2-loss leads to relatively enhanced nuclear YAP signal (supplementary figure 2C), \nsuggesting a resistance to the reduction in nuclear YAP observed in WT spheroids. To assess \nhow these differences in YAP activity may drive distinct spheroid phenotypes, we compared \nthe growth of spheroids with intact NF2 and YAP to those with either KO of NF2 or YAP \n(figure 2E). In addition to a relative increase in nuclear YAP, there is a concurrent increase in \nspheroid size in MeT-5A spheroids on loss of NF2 (figure 2F). In contrast, YAP-deficient \nspheroids exhibit a clear decrease in size (figure 2F). This increased growth observed in NF2-\ndeficient spheroids recapitulates key aspects of the 3D tumour environment and therefore \nreflects tumorigenic potential58, possibly mediated via enhanced YAP activity. \nNext, having established a quantifiable cancer-relevant phenotype robustly driven by NF2 \nloss, we tested spheroid growth in the presence of YAP/TAZ-TEAD inhibitors to quantify the \nimpact of inhibition on the tumorigenic capacity of spheroids. WT MeT-5A cells exhibit \npartial sensitivity selectively to pan-TEAD inhibitors (figure 2G), with a significant reduction \nin spheroid growth observed on treatment with VT-107 and K-975 at high concentrations \n(10 µM). Interestingly, NF2 KO MeT-5A cells were more sensitive to nearly all treatments \n(figure 2H, supplementary figure 2D). Beyond sensitivity, our data highlight that NF2 KO cells \nexhibit a marked decrease in spheroid area, with an enhanced effect observed with on-\ntarget TEAD inhibition as compared to indirect YAP/TAZ-TEAD inhibitors (supplementary \nfigure 2E). This inhibitory effect is sufficient to revert NF2-deficient spheroids to sizes \nequivalent to YAP KO spheroids. \nThese findings collectively point to the spheroid formation assay as effective to evaluate and \nprobe YAP/TAZ-TEAD driven tumorigenic capacity. \n \nDetermining the role of YAP and TEAD isoforms on YAP/TAZ-TEAD inhibitor efficacy \nGiven NF2-mutant in vitro anchorage-independent growth is dependent on YAP6,51,59, we \nnext sought to establish the extent to which YAP-TEAD influence the observed NF2-deficient \nphenotype. To confirm YAP’s role in driving NF2 KO spheroid formation, YAP was \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 7 \noverexpressed in WT MeT-5A cells. We exogenously expressed YAP-5SA, which contains \npoint mutations across five LATS1/2 target residues, rendering YAP non-phosphorylatable by \nLATS, and therefore hyperactive51. Expression of YAP-5SA in MeT-5A cells (figure 3A) reveals \nthat hyperactive YAP drives spheroid growth, closely phenocopying NF2 KO spheroids \n(figure 3B). To resolve whether this is driven by TEAD activity targetable via therapeutic \ninhibition, YAP-5SA spheroids were treated with TEAD-selective inhibitors. Spheroids with \nhyperactive YAP phenocopy the inhibition observed upon loss of NF2 KO, with enhanced \nsensitivity to TEAD inhibition (figure 2H, figure 3C).  \nWe next assessed spheroid growth in NF2/YAP double knockout (dKO) cells (figure 3D). \nWhen YAP is concurrently lost in NF2 KO cells, a reversal of the NF2 KO phenotype is \nobserved, with NF2/YAP dKO spheroids exhibiting a dramatic decrease in spheroid growth \nrelative to NF2 KO (figure 3E). As such, NF2/YAP dKO spheroid sizes were restored roughly \nto those observed in YAP KO spheroids. Taken together, these observations further \nreinforce the likelihood that YAP is the primary effector of NF2 KO spheroid growth and, \ngiven the enhanced sensitivity to TEAD inhibition observed in the presence of hyperactive \nYAP, may be targetable via inhibiting YAP-TEAD transcription.  \nThe TEAD family consists of four proteins in vertebrae, in humans termed TEAD1-413. Recent \nearly-stage clinical evaluation has highlighted that pan-TEAD and TEAD isoform specificity \nmight be a defining clinical feature in toxicity and efficacy60,61. The implications of isoform \nspecific or pan-TEAD targeting in cancer is therefore of importance with pan-TEAD likely \nbeing more effective, but in general higher risk of toxicity. Assessing TEAD isoform \nfunctional importance and drug specificity is therefore critical during drug development to \nguide optimal clinical responses. These implications prompted us to generate NF2/TEAD \nisoform specific dKO cells for TEAD isoforms with antibodies available and where TEAD \nisoforms are robustly expressed and detected in MeT-5A cells. TEAD1 and TEAD4 were \nspecifically recognised and targeted (figure 3D), with both TEAD1 and TEAD4 implicated in \nthe progression and development of various cancers62–64. Notably, TEAD4 loss leads to an \nincrease in protein levels of TEAD1, while loss of either TEAD1 or TEAD4 increases YAP levels \n(figure 3D). This finding has ramifications for mesothelioma, and potential general cancer \ntherapy, given the possibility of compensatory upregulation to counter the inhibition of \nsingle components within the pathway. Loss of TEAD1 or TEAD4, in contrast to YAP loss, \nleads to a modest decrease in spheroid area (figure 3F), reverting spheroid size closer to \nthat of WT spheroids. This effect is particularly pronounced in TEAD4 KO cells. When \ncompared to NF2/YAP dKO, these results suggest complete and partial dependence on YAP \nand individual TEAD isoforms respectively. \nTo assess which of the four TEAD isoforms, if any, coordinate the hypersensitivity of NF2-\ndeficient spheroids to TEAD inhibition, we treated NF2/TEAD1 and NF2/TEAD4 dKO MeT-5A \nspheroids and compared response to parental genotypes. If the observed response to TEAD \ninhibition is mediated via YAP-TEAD, a limited additive effect of spheroid reduction on \ncombination of TEAD inhibition and KO of critical TEAD isoforms would be evident. Resulting \ndKO sensitivities indicate that, although TEAD4 may be more critical for the establishment of \nthe observed NF2-deficient phenotype, with a more complete reversal on concurrent loss of \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 8 \nTEAD4 as compared to TEAD1 (figure 3F), the loss of either isoform is sufficient to offset the \nincreased sensitivity of NF2 KO spheroids to selective TEAD inhibition (figures 3G-H, \nsupplementary figures 3A-B). More specifically, loss of TEAD1 leads to a blunting of NF2 KO \nspheroid sensitivity to selective TEAD inhibition, with all compounds exhibiting reduced \nsensitivity at one or both concentrations tested (figure 3G, supplementary figures 3A). \nSimilar results are seen when TEAD4 is lost (figure 3H, supplementary figures 3B), which is \nparticularly striking given IK-930 is purported to be a TEAD1-selective inhibitor. The \ndecrease in sensitivity observed in NF2/TEAD4 dKO spheroids suggests that TEAD4 may \ncoordinate, at least in part, some response to IK-930 (figure 3H).  \nIn summary, these findings validate the role of YAP and TEADs in coordinating the \ntumorigenic capacity of NF2 loss in vitro, with YAP in particular indispensable to the \nenhanced spheroid forming capability of NF2-deficient cells. Though there is likely some \nredundancy among the TEAD isoforms, the enhanced response to TEAD inhibition in NF2-\ndeficient spheroids is blunted when TEAD isoforms are concurrently eliminated. \n \nMorphometric profiling predicts selectivity of YAP/TAZ-TEAD inhibitors \nDrugs with off-target effects are a major issue for drug development, directly impacting \nmultiple levels of clinical transition, with common issues including toxicity, specificity, and \nefficacy affecting progress65–68. Treatment with inhibitors that exhibit broad off-target \neffects often induce morphological changes beyond their effect on proliferation, stemness, \nand therapeutic resistance. This is exemplified in the disruption of actin organisation in \nMeT-5A cells on treatment with dasatinib (figure 1D), consistent with previous observations \nacross multiple cell-lines30,69. Additionally, a highly disruptive morphology is observed in \nspheroids on treatment with high concentrations of non-selective inhibitors (supplementary \nfigure 3C), pointing to a recurrent phenotype associated with off-target effects of YAP/TAZ-\nTEAD inhibition.  \nAs shifts in cellular morphology are predictive of broad pharmacological effects70 and can be \nleveraged to infer detailed molecular effects of treatment, we next sought to quantify the \nmorphological disruption on YAP/TAZ-TEAD inhibition in our cellular model system. To \ninitially confirm the phenotype observed in MeT-5A spheroids (supplementary figure 3C), \nwe performed a quantitative analysis of the morphological shift of spheroids on treatment \nwith YAP/TAZ-TEAD inhibitors. Brightfield images were segmented, and morphometric \ncharacteristics extracted and processed according to Joint Undertaking in Morphological \nProfiling (JUMP) pipelines71. This revealed a distinct morphological clustering within non-\nspecific YAP inhibitors, while selective TEAD inhibition results in a phenotype that readily \nclustered with vehicle control cells (supplementary figure 3D). \nCell Painting is a newly developed approach to morphological profiling, which has been \nincorporated into high-throughput imaging assay pipelines to impute mechanisms-of-\naction72, bioactivity73, and toxicity74,75 in recent drug discovery efforts. In order to quantify \nthe morphological perturbations induced on inhibition of YAP/TAZ-TEAD in greater \nresolution, we undertook Cell Painting to generate high-dimensionality datasets. Cell \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 9 \nPainting involves the staining of cells with six dyes (supplementary figure 3E) and was \nperformed in accordance with protocols established by the Joint Undertaking in \nMorphological Profiling Cell Painting (JUMP-CP) consortium76. Principal component analysis \n(PCA) of high-dimensional morphometric data reveals a morphological shift on treatment \nwith non-selective inhibitors of YAP (figure 4A). To quantify this shift, the correlation \ndistance across all morphological features was computed between treated wells. The \nresulting correlation distance matrix shows a clustering of non-selective and high \nconcentration inhibitors, which share a high degree of morphological similarity (figure 4B). \nInterestingly, there was little clustering in vehicle control and low concentration TEAD-\nselective inhibitor wells. This suggests that morphological uniformity may be a metric of \nnon-specific inhibition of YAP/TAZ-TEAD. Quantification of uniformity, as defined by the \nmean correlation distance between like-treated cells, reveals significantly enhanced \nmorphological uniformity on high concentration, as well as dasatinib and lovastatin, \ntreatment (figure 4C). To assess whether this observed uniformity is YAP-dependent or -\nindependent, we quantified morphological uniformity on treatment in YAP KO MeT-5A cells \n(figure 4D). This reveals similar trends, with both dasatinib and lovastatin inducing a shift to \nmore morphologically similar cells, thus indicating YAP-independence. Strikingly, high \nconcentration K-975 and verteporfin treatment induced no such effect in YAP-deficient cells \nin contrast to WT, which may suggest a degree of morphological disruption mediated via \nYAP-TEAD inhibition. \nWith the broad impact of treatment on cell morphology established, we next performed in-\ndepth analysis of specific features disrupted on inhibition of YAP/TAZ-TEAD. For this, we \nimplemented a modified robust Z’ scoring, a parameter typically used in quality control of \ndrug screening to extract features for each treatment that resolve distinctly from untreated \ncontrols. The Z’ scoring reveals a disruption of a high percentage of features on treatment \nwith non-specific YAP and high concentrations of TEAD-specific inhibitors (figure 4E), \nconsistent with broad profile analysis. Interestingly, these features are generally equally \ndisrupted in YAP KO cells, particularly in dasatinib treated cells (figure 4E), suggesting that \nthis morphological disruption is YAP-independent. To quantify this YAP specificity in detail, \nwe adjusted modified Z’ statistics in WT to YAP KO cells. This revealed a range of conserved \nfeatures that were consistently disrupted in a YAP independent manner specifically on \ntreatment with the non-specific YAP inhibitors dasatinib and lovastatin (Figure 3F). \nConversely, a smaller proportion of features were YAP dependent, which were disrupted on \ntreatment with both selective TEAD inhibitors, with minimal overlap between the various \nTEAD inhibitors, and non-specific YAP inhibitors. Intriguingly, this suggests that different sets \nof features may be used to quantify YAP-dependent and -independent morphological \nchanges, with verteporfin and both selective TEAD inhibitors showing less off-target YAP \neffects than dasatinib and lovastatin.   \n \nDISCUSSION \nGiven the critical role the Hippo pathway and, more specifically, its downstream effectors \nYAP/TAZ-TEAD, play in the progression of mesothelioma and other cancers, there is a clear \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 10 \nneed to develop pipelines capable of assessing the efficacy and specificity of newly \ndeveloped inhibitors to streamline positioning to clinical use. While in vitro studies assessing \nthe efficacy of TEAD inhibition in NF2-deficient mesothelioma cells have been described57, \nthese are limited by the heterogenous genomic and transcriptomic landscapes associated \nwith cell-lines of mixed origins, potentially obfuscating critical, nuanced molecular dynamics \nof inhibiting of the transcriptional machinery of the Hippo pathway. Here, we present a \npipeline within an isogenic model of loss of upstream and downstream components within \nthe Hippo pathway, allowing the detailed interrogation of the molecular interplay of the \nindividual constituents that comprise this tumorigenic pathway. We additionally describe \nfunctional assays for quantifying the YAP/TAZ-TEAD selective inhibition of tumorigenic \ncapacity, which have shed light into the specifics of efficacy and selectivity of the core \ncompound library used to target the major players within the pathway. \nPrior multiomic and genetic analysis have highlighted the in vivo complexities of pleural \nmesothelioma6–9,77,78. The unfortunate continued lack of curative therapies for this deadly \ndisease is likely due to late-stage diagnosis and the sparsity of complimentary preclinical \nmodels representing the disease required for the predictive evaluation of drug targets \nbefore clinical evaluation in patients. Notably, our observations predict that targeting the \ntranscriptional complex directly via inhibition of TEAD, including via its auto-palmitoylation \nsite, is more specific, and thereby likely more effective, than targeting the upstream kinase \nmodule. Our observations also correlate well with clinical data, including the lack of activity \nand overall unwelcome association with pulmonary toxicities with dasatanib treatment in \nunselected pleural mesothelioma patients79. Our platform combines an extensive, in vitro \nbased isogenic cellular model allowing for specificity evaluation, cancer-relevant assays \nwhere tumour suppressors are functionally active, high content imaging, and Cell Painting. \nCollectively, these allow for direct evaluation of YAP/TAZ-TEAD(1-4) targeting drugs and is \nfeasible to implement alongside synergistic combinatorial drug evaluations. We are hopeful, \nthat our approach provides a milestone within the development of mesothelioma \ntreatments, as we appear to potentially evaluate potency of inhibition80 and clinical \nefficiency. The development of the platform thereby might inform future clinical trials and \noverall complement ongoing efforts to develop advanced cellular models81,82 and ultimately \ncurative therapies.  \nResults point to K-975 as exemplar inhibitor of TEAD auto-palmitoylation, with VT-107 \nexhibiting equivalent potency alongside off-target effects at higher concentrations. \nConversely, while IK-930 treatment is not associated with off-target effects and therefore a \nfavourable safety profile, its potency is markedly less relative to alternative TEAD inhibitors. \nDespite previously reported selectivity for TEAD183, the observation that loss of TEAD4 \nimpacts the sensitivity of spheroids to IK-930 to the same extent of TEAD1 indicates some \ndegree of inhibition of additional TEAD isoforms. Our findings also highlight the apparent \ncompensatory mechanisms in place when targeting individual components (figure 3D). This \ndiscovery may explain the promiscuity observed in IK-930, suggesting a compensatory \nupregulation of, and therefore potential dependence on, TEAD1 on loss of TEAD4. \nInterestingly our results echo preliminary results from clinical studies of IK-930, which \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 11 \nindicate that, despite being well tolerated, IK-930 has not proved effective in patients \nenrolled in clinical trials84. \nIn comparison to selective TEAD inhibitors, upstream inhibitors of YAP/TAZ-TEAD are \nassociated with sweeping off-target effects, with dasatinib inducing cellular changes well \nbeyond inhibition of the YAP-TEAD axis alone. Surprisingly, given the extent of YAP-\nindependent effects previously associated with treatment, verteporfin exhibits a high \ndegree of YAP-specificity across a range of functional assays. The convergence between \nphenotypes in cells treated with verteporfin and selective inhibitors of TEAD reinforces \nprevious findings that verteporfin can act to disrupt YAP-TEAD binding32, though YAP-\nindependent cytotoxicity is validated within our model system. The observation that high \npotency inhibitors of TEAD auto-palmitoylation also increases cytoplasmic retention of YAP \nis novel, though consistent with reports that inhibitors of YAP-TEAD interactions \nconcomitantly reduce levels of nuclear YAP in mesothelioma cell-lines85. This phenomenon \nhas been described as likely a result of displacement of YAP from chromatin-resident TEAD, \nreinforced by the finding that K-975 treatment reduces formation of YAP nuclear \ncondensates, indicative of decreased binding to super-enhancer regions52. \nTo conclude, this work provides a framework for quantifying the efficacy of YAP/TAZ-TEAD \ninhibition for the treatment of mesothelioma. Conventional phenotypic screening \napproaches, which typically gauge success based on the cytotoxic or cytostatic properties of \ntherapeutics, are impeded by their limited scope and are biased to drugs with broad toxicity \nprofiles86. As the Hippo pathway integrates a wide range of stimuli11,12, it might be \nparticularly prone to these perturbations. The pipeline established here benefits from the \nincorporation of a confirmed YAP-TEAD specific phenotypic assay, while the identification of \na feature signature of YAP-specificity allows for the assessment of potential off-target \neffects of putative treatments. We expect this pipeline may help expedite the discovery of \ntherapeutics effective in managing YAP/TAZ-TEAD driven mesothelioma, mitigating some of \nthe risks of moving into in vivo assessment of inhibitor safety and efficacy. \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 12 \nAUTHOR CONTRIBUTIONS \nRC: data curation, conceptualization, formal analysis, validation, investigation, visualization, \nmethodology, and writing—original draft, review, and editing. SJ: data curation, formal \nanalysis, validation, investigation, visualization, methodology, and writing—review and \nediting. KP: data curation, formal analysis, validation, investigation, visualization and \nmethodology. JCW: methodology and resources. MNF: investigation. YL: validation and \nmethodology. NSH: investigation. REG: methodology. AGR: supervision. NOC: resources, \nmethodology, review and editing. CGH: conceptualization, resources, data curation, formal \nanalysis, supervision, investigation, methodology, and writing—original draft, review, and \nediting.  \n \nACKNOWLEDGEMENTS \nOngoing research in the Hansen lab (CGH) was funded by Worldwide Cancer Research (19-\n0238) and CSO-LifeArc. This project was initiated by pump prime funding from ISSF3 and \nJHMRF. MN and KR are funded by MRC Precision Medicine DTP Studentships. Nancy by the \nMartin Lee DTP. SJ is funded by a scholarship from the Chinese Scholarship Council, and the \nEdinburgh Global from University of Edinburgh. We furthermore acknowledge team \nmembers for insightful comments and constructive feedback during this study. We \nacknowledge the technical support and guidance provided by the Institute for Regeneration \nand Repair (IRR) Flow Cytometry and Cell Sorting Facility staff. Single cell sorting necessary \nfor clonal selection of expansion of KO cell populations was conducted with support from \nthe QMRI Flow Cytometry and IRR Flow Cytometry and Cell Sorting Facility, University of \nEdinburgh.  \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 13 \nReferences \n1. Wagner, J. C., Sleggs, C. A. & Marchand, P. Diffuse Pleural Mesothelioma and \nAsbestos Exposure in the North Western Cape Province. Occup. Environ. Med. 17, \n260–271 (1960). \n2. Mirra, L. et al. Therapeutic Strategies to Improve the Treatment of Pleural \nMesothelioma. Curr. Med. Chem. (2024) \ndoi:10.2174/0109298673268206240405084558. \n3. Nowak, A. K., Jackson, A. & Sidhu, C. Management of Advanced Pleural \nMesothelioma—At the Crossroads. JCO Oncol. Pract. 18, 116–124 (2022). \n4. Tsao, A. S., Pass, H. I., Rimner, A. & Mansfield, A. S. New Era for Malignant Pleural \nMesothelioma: Updates on Therapeutic Options. J. Clin. Oncol. 40, 681–692 (2022). \n5. Jin, H., Wang, L. & Bernards, R. Rational combinations of targeted cancer therapies: \nbackground, advances and challenges. Nat. Rev. Drug Discov. 22, 213–234 (2023). \n6. Cunningham, R. et al. YAP/TAZ activation predicts clinical outcomes in mesothelioma \nand is conserved in in vitro model of driver mutations. Clin. Transl. Med. 13, e1190 \n(2023). \n7. Bueno, R. et al. Comprehensive genomic analysis of malignant pleural mesothelioma \nidentifies recurrent mutations, gene fusions and splicing alterations. Nat. Genet. 48, \n407–416 (2016). \n8. Hmeljak, J. et al. Integrative Molecular Characterization of Malignant Pleural \nMesothelioma. Cancer Discov. 8, 1548–1565 (2018). \n9. Mangiante, L. et al. Multiomic analysis of malignant pleural mesothelioma identifies \nmolecular axes and specialized tumor profiles driving intertumor heterogeneity. Nat. \nGenet. 55, 607–618 (2023). \n10. Febres-Aldana, C. A. et al. Diffuse Pleural Mesothelioma: Advances in Molecular \nPathogenesis, Diagnosis, and Treatment. Annu. Rev. Pathol. Mech. Dis. 19, 11–42 \n(2024). \n11. Rausch, V. & Hansen, C. G. The Hippo Pathway, YAP/TAZ, and the Plasma Membrane. \nTrends Cell Biol. 30, 32–48 (2020). \n12. Hansen, C. G., Moroishi, T. & Guan, K.-L. YAP and TAZ: a nexus for Hippo signaling and \nbeyond. Trends Cell Biol. 25, 499–513 (2015). \n13. Currey, L., Thor, S. & Piper, M. TEAD family transcription factors in development and \ndisease. Development 148, (2021). \n14. Lin, K. C. et al. Regulation of Hippo pathway transcription factor TEAD by p38 MAPK-\ninduced cytoplasmic translocation. Nat. Cell Biol. 19, 996–1002 (2017). \n15. Davis, J. R. & Tapon, N. Hippo signalling during development. Development 146, \n(2019). \n16. Zhong, Z., Jiao, Z. & Yu, F.-X. The Hippo signaling pathway in development and \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 14 \nregeneration. Cell Rep. 43, 113926 (2024). \n17. Lamar, J. M. et al. The Hippo pathway target, YAP, promotes metastasis through its \nTEAD-interaction domain. Proc. Natl. Acad. Sci. 109, E2441–E2450 (2012). \n18. Feng, X. et al. Hippo-Independent Activation of YAP by the GNAQ Uveal Melanoma \nOncogene through a Trio-Regulated Rho GTPase Signaling Circuitry. Cancer Cell 25, \n831–845 (2014). \n19. Sorrentino, G. et al. Metabolic control of YAP and TAZ by the mevalonate pathway. \nNat. Cell Biol. 16, 357–366 (2014). \n20. Calvo, F. et al. Mechanotransduction and YAP-dependent matrix remodelling is \nrequired for the generation and maintenance of cancer-associated fibroblasts. Nat. \nCell Biol. 15, 637–46 (2013). \n21. Shao, D. D. et al. KRAS and YAP1 Converge to Regulate EMT and Tumor Survival. Cell \n158, 171–184 (2014). \n22. Lau, A. N. et al. Tumor-propagating cells and Yap/Taz activity contribute to lung \ntumor progression and metastasis. EMBO J. 33, 468–81 (2014). \n23. Seavey, C. N. et al. WWTR1 (TAZ)- CAMTA1 gene fusion is sufficient to dysregulate \nYAP/TAZ signaling and drive epithelioid hemangioendothelioma tumorigenesis. Genes \nDev. 35, 512–527 (2021). \n24. Driskill, J. H. et al. WWTR1(TAZ)-CAMTA1 reprograms endothelial cells to drive \nepithelioid hemangioendothelioma. Genes Dev. 35, 495–511 (2021). \n25. Moroishi, T., Hansen, C. G. & Guan, K.-L. The emerging roles of YAP and TAZ in cancer. \nNat. Rev. Cancer 15, 73–79 (2015). \n26. Noland, C. L. et al. Palmitoylation of TEAD Transcription Factors Is Required for Their \nStability and Function in Hippo Pathway Signaling. Structure 24, 179–186 (2016). \n27. Chan, P. et al. Autopalmitoylation of TEAD proteins regulates transcriptional output of \nthe Hippo pathway. Nat. Chem. Biol. 12, 282–289 (2016). \n28. Lim, E. et al. Extended pleurectomy decortication and chemotherapy versus \nchemotherapy alone for pleural mesothelioma (MARS 2): a phase 3 randomised \ncontrolled trial. Lancet. Respir. Med. 12, 457–466 (2024). \n29. Cunningham, R. & Hansen, C. G. The Hippo pathway in cancer: YAP/TAZ and TEAD as \ntherapeutic targets in cancer. Clin. Sci. 136, 197–222 (2022). \n30. Pichot, C. S. et al. Dasatinib synergizes with doxorubicin to block growth, migration, \nand invasion of breast cancer cells. Br. J. Cancer 101, 38–47 (2009). \n31. Nam, S. et al. Action of the Src Family Kinase Inhibitor, Dasatinib (BMS-354825), on \nHuman Prostate Cancer Cells. Cancer Res. 65, 9185–9189 (2005). \n32. Liu-Chittenden, Y. et al. Genetic and pharmacological disruption of the TEAD-YAP \ncomplex suppresses the oncogenic activity of YAP. Genes Dev. 26, 1300–5 (2012). \n33. Dasari, V. R. et al. Verteporfin exhibits YAP-independent anti-proliferative and \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 15 \ncytotoxic effects in endometrial cancer cells. Oncotarget 8, 28628–28640 (2017). \n34. Xia, Y., Sun, M., Huang, H. & Jin, W.-L. Drug repurposing for cancer therapy. Signal \nTransduct. Target. Ther. 9, 92 (2024). \n35. Franklin, J. M., Wu, Z. & Guan, K.-L. Insights into recent findings and clinical \napplication of YAP and TAZ in cancer. Nat. Rev. Cancer 23, 512–525 (2023). \n36. Tang, T. T. et al. Small Molecule Inhibitors of TEAD Auto-palmitoylation Selectively \nInhibit Proliferation and Tumor Growth of NF2 -deficient Mesothelioma. Mol. Cancer \nTher. 20, 986–998 (2021). \n37. Kaneda, A. et al. The novel potent TEAD inhibitor, K-975, inhibits YAP1/TAZ-TEAD \nprotein-protein interactions and exerts an anti-tumor effect on malignant pleural \nmesothelioma. Am. J. Cancer Res. 10, 4399–4415 (2020). \n38. Kaneda, A. et al. Abstract 3086: Discovery of a first-in-class TEAD inhibitor which \ndirectly inhibits YAP/TAZ-TEAD protein-protein interaction and shows a potent anti-\ntumor effect in malignant pleural mesothelioma. Cancer Res. 79, 3086–3086 (2019). \n39. Amidon, B. S. et al. Abstract 2156: IK-930 is a novel TEAD inhibitor for the treatment \nof cancers harboring mutations in the Hippo signal transduction pathway. Cancer Res. \n82, 2156–2156 (2022). \n40. Tolcher, A. W. et al. A phase 1, first-in-human study of IK-930, an oral TEAD inhibitor \ntargeting the Hippo pathway in subjects with advanced solid tumors. J. Clin. Oncol. \n40, TPS3168–TPS3168 (2022). \n41. Pobbati, A. V & Hong, W. A combat with the YAP/TAZ-TEAD oncoproteins for cancer \ntherapy. Theranostics 10, 3622–3635 (2020). \n42. Wang, C. et al. Verteporfin inhibits YAP function through up-regulating 14-3-3σ \nsequestering YAP in the cytoplasm. Am. J. Cancer Res. 6, 27–37 (2016). \n43. Sun, J. et al. A tightly controlled Src-YAP signaling axis determines therapeutic \nresponse to dasatinib in renal cell carcinoma. Theranostics 8, 3256–3267 (2018). \n44. Hsu, P.-C., Yang, C.-T., Jablons, D. M. & You, L. The Crosstalk between Src and \nHippo/YAP Signaling Pathways in Non-Small Cell Lung Cancer (NSCLC). Cancers \n(Basel). 12, 1361 (2020). \n45. Wang, Z. et al. Interplay of mevalonate and Hippo pathways regulates RHAMM \ntranscription via YAP to modulate breast cancer cell motility. Proc. Natl. Acad. Sci. U. \nS. A. 111, E89-98 (2014). \n46. Wang, Y. et al. Comprehensive Molecular Characterization of the Hippo Signaling \nPathway in Cancer. Cell Rep. 25, 1304-1317.e5 (2018). \n47. Ran, F. A. et al. Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. 8, \n2281–2308 (2013). \n48. Hao, Y., Chun, A., Cheung, K., Rashidi, B. & Yang, X. Tumor Suppressor LATS1 Is a \nNegative Regulator of Oncogene YAP. J. Biol. Chem. 283, 5496–5509 (2008). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 16 \n49. Zhao, B., Li, L., Tumaneng, K., Wang, C.-Y. & Guan, K.-L. A coordinated \nphosphorylation by Lats and CK1 regulates YAP stability through SCF(beta-TRCP). \nGenes Dev. 24, 72–85 (2010). \n50. Kim, N.-G., Koh, E., Chen, X. & Gumbiner, B. M. E-cadherin mediates contact inhibition \nof proliferation through Hippo signaling-pathway components. Proc. Natl. Acad. Sci. \n108, 11930–11935 (2011). \n51. Zhao, B. et al. Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell \ncontact inhibition and tissue growth control. Genes Dev. 21, 2747–61 (2007). \n52. Hao, S. et al. YAP condensates are highly organized hubs. iScience 27, 109927 (2024). \n53. Yu, F.-X. et al. Regulation of the Hippo-YAP Pathway by G-Protein-Coupled Receptor \nSignaling. Cell 150, 780–791 (2012). \n54. Meng, Z. et al. MAP4K family kinases act in parallel to MST1/2 to activate LATS1/2 in \nthe Hippo pathway. Nat. Commun. 6, 8357 (2015). \n55. Park, H. W. et al. Alternative Wnt Signaling Activates YAP/TAZ. Cell 162, 780–94 \n(2015). \n56. Moroishi, T. et al. A YAP/TAZ-induced feedback mechanism regulates Hippo pathway \nhomeostasis. Genes Dev. 29, 1271–84 (2015). \n57. Evsen, L., Morris, P. J., Thomas, C. J. & Ceribelli, M. Comparative Assessment and \nHigh-Throughput Drug-Combination Profiling of TEAD-Palmitoylation Inhibitors in \nHippo Pathway Deficient Mesothelioma. Pharmaceuticals (Basel). 16, (2023). \n58. Ishiguro, T. et al. Tumor‐derived spheroids: Relevance to cancer stem cells and clinical \napplications. Cancer Sci. 108, 283–289 (2017). \n59. Zhao, B. et al. Cell detachment activates the Hippo pathway via cytoskeleton \nreorganization to induce anoikis. Genes Dev. 26, 54–68 (2012). \n60. Muller, F. et al. Abstract 445: A next generation TEAD inhibitor with refined isoform \nspecificity for superior safety &amp; efficacy. Cancer Res. 83, 445–445 (2023). \n61. Muller, F., Kunnimalaiyaan, S., Mangrolia, P. & Olson, J. Abstract 5913: TEAD1/4 \ninhibitors exhibit deeper biological impact and broader activity compared to TEAD1-\nonly inhibitors in both monotherapy and combination without additional kidney \ntoxicity. Cancer Res. 84, 5913–5913 (2024). \n62. Knight, J. F. et al. TEAD1 and c-Cbl are novel prostate basal cell markers that correlate \nwith poor clinical outcome in prostate cancer. Br. J. Cancer 99, 1849–1858 (2008). \n63. Lu, T. et al. Discovery of a subtype-selective, covalent inhibitor against palmitoylation \npocket of TEAD3. Acta Pharm. Sin. B 11, 3206–3219 (2021). \n64. Zhang, W. et al. TEAD4 overexpression promotes epithelial-mesenchymal transition \nand associates with aggressiveness and adverse prognosis in head neck squamous cell \ncarcinoma. Cancer Cell Int. 18, 178 (2018). \n65. Sun, D., Gao, W., Hu, H. & Zhou, S. Why 90% of clinical drug development fails and \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 17 \nhow to improve it? Acta Pharm. Sin. B 12, 3049–3062 (2022). \n66. Seyhan, A. A. Lost in translation: the valley of death across preclinical and clinical \ndivide – identification of problems and overcoming obstacles. Transl. Med. Commun. \n4, 18 (2019). \n67. Lin, A. et al. Off-target toxicity is a common mechanism of action of cancer drugs \nundergoing clinical trials. Sci. Transl. Med. 11, (2019). \n68. Lechner, S. et al. Target deconvolution of HDAC pharmacopoeia reveals MBLAC2 as \ncommon off-target. Nat. Chem. Biol. 18, 812–820 (2022). \n69. Calizo, R. C. et al. Disruption of podocyte cytoskeletal biomechanics by dasatinib leads \nto nephrotoxicity. Nat. Commun. 10, 2061 (2019). \n70. Perlman, Z. E. et al. Multidimensional drug profiling by automated microscopy. \nScience 306, 1194–8 (2004). \n71. Chandrasekaran, S. N. et al. JUMP Cell Painting dataset: morphological impact of \n136,000 chemical and genetic perturbations. bioRxiv 2023.03.23.534023 (2023) \ndoi:10.1101/2023.03.23.534023. \n72. Tian, G., Harrison, P. J., Sreenivasan, A. P., Carreras-Puigvert, J. & Spjuth, O. \nCombining molecular and cell painting image data for mechanism of action \nprediction. Artif. Intell. Life Sci. 3, 100060 (2023). \n73. Fredin Haslum, J. et al. Cell Painting-based bioactivity prediction boosts high-\nthroughput screening hit-rates and compound diversity. Nat. Commun. 15, 3470 \n(2024). \n74. Stossi, F. et al. High throughput microscopy and single cell phenotypic image-based \nanalysis in toxicology and drug discovery. Biochem. Pharmacol. 216, 115770 (2023). \n75. Garcia de Lomana, M., Marin Zapata, P. A. & Montanari, F. Predicting the \nMitochondrial Toxicity of Small Molecules: Insights from Mechanistic Assays and Cell \nPainting Data. Chem. Res. Toxicol. 36, 1107–1120 (2023). \n76. Cimini, B. A. et al. Optimizing the Cell Painting assay for image-based profiling. Nat. \nProtoc. 18, 1981–2013 (2023). \n77. Giotti, B. et al. Single cell view of tumor microenvironment gradients in pleural \nmesothelioma. Cancer Discov. (2024) doi:10.1158/2159-8290.CD-23-0017. \n78. Ma, X. et al. Spatial Landscape of Malignant Pleural and Peritoneal Mesothelioma \nTumor Immune Microenvironments. Cancer Res. Commun. 4, 2133–2146 (2024). \n79. Dudek, A. Z. et al. Phase II Study of Dasatinib in Patients with Previously Treated \nMalignant Mesothelioma (Cancer and Leukemia Group B 30601): A Brief Report. J. \nThorac. Oncol. 7, 755–759 (2012). \n80. Pobbati, A. V, Kumar, R., Rubin, B. P. & Hong, W. Therapeutic targeting of TEAD \ntranscription factors in cancer. Trends Biochem. Sci. 48, 450–462 (2023). \n81. Mazzocchi, A. R., Rajan, S. A. P., Votanopoulos, K. I., Hall, A. R. & Skardal, A. In vitro \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 18 \npatient-derived 3D mesothelioma tumor organoids facilitate patient-centric \ntherapeutic screening. Sci. Rep. 8, 2886 (2018). \n82. Hocking, A. J., Mortimer, L. A., Farrall, A. L., Russell, P. A. & Klebe, S. Establishing \nmesothelioma patient-derived organoid models from malignant pleural effusions. \nLung Cancer 191, 107542 (2024). \n83. Young, N. et al. Abstract 1646: IK-930, a paralog-selective TEAD inhibitor for treating \nYAP/TAZ-TEAD dependent cancers. Cancer Res. 83, 1646–1646 (2023). \n84. Ikena Oncology Announces Strategic Update. (2024). \n85. Chapeau, E. A. et al. Direct and selective pharmacological disruption of the YAP–TEAD \ninterface by IAG933 inhibits Hippo-dependent and RAS–MAPK-altered cancers. Nat. \nCancer 5, 1102–1120 (2024). \n86. Moffat, J. G., Rudolph, J. & Bailey, D. Phenotypic screening in cancer drug discovery — \npast, present and future. Nat. Rev. Drug Discov. 13, 588–602 (2014). \n87. Hansen, C. G., Ng, Y. L. D., Lam, W.-L. M., Plouffe, S. W. & Guan, K.-L. The Hippo \npathway effectors YAP and TAZ promote cell growth by modulating amino acid \nsignaling to mTORC1. Cell Res. 25, 1299–1313 (2015). \n88. Rausch, V. et al. The Hippo Pathway Regulates Caveolae Expression and Mediates \nFlow Response via Caveolae. Curr. Biol. 29, 242-255.e6 (2019). \n89. Spandidos, A., Wang, X., Wang, H. & Seed, B. PrimerBank: a resource of human and \nmouse PCR primer pairs for gene expression detection and quantification. Nucleic \nAcids Res. 38, D792-9 (2010). \n90. Stirling, D. R. et al. CellProfiler 4: improvements in speed, utility and usability. BMC \nBioinformatics 22, 433 (2021). \n91. Stringer, C., Wang, T., Michaelos, M. & Pachitariu, M. Cellpose: a generalist algorithm \nfor cellular segmentation. Nat. Methods 18, 100–106 (2021). \n92. Chandrasekaran, S. N. et al. Three million images and morphological profiles of cells \ntreated with matched chemical and genetic perturbations. Nat. Methods 21, 1114–\n1121 (2024). \n93. Singh, S. et al. Cytominer: Methods for Image-Based Cell Profiling. at https://cran.r-\nproject.org/package=cytominer (2020). \n94. Gu, Z., Gu, L., Eils, R., Schlesner, M. & Brors, B. circlize Implements and enhances \ncircular visualization in R. Bioinformatics 30, 2811–2 (2014). \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 19 \nMATERIALS AND METHODS \nLuciferase assay \nHEK293A cells were seeded in triplicate in 12-well plates before co-transfection with Gal4-\nTEAD, 5×UAS Luciferase reporter, and Renilla luciferase constructs, using GenJet transfection \nreagent ( SL100488, SignaGen). Post-transfection, cells were lysed and luciferase \nquantification was performed using a Dual-Glo® Luciferase Assay kit ( E2920, Promega) \naccording to manufacturer ’s specifications and measured on a Biotek Synergy HT plate \nreader, with luminescence adjusted to Renilla luciferase. \n \nCulture maintenance \nMeT-5A mesothelial cells and CRISPR -Cas9 mediated KO genotypes were generated and \ncultured as described in Cunningham et al 6. HEK293A gene edited cells are described in  \nHansen et al87, Rausch et al88, and Lin et al14 and cultured in high glucose DMEM (21969-035, \nGibco) with 2 mM L -Glutamine (25030-024, Gibco) and 10% FBS (10500 -064, Gibco). Cells \nwere grown in the presence of 100 units/mL of penicillin and 100 µg/mL of streptomycin \n(15140-122, Gibco) and in cubated at 37°C with 20% O 2. Where specified, cells were treated \nwith VT-107 (HY-134957, MedChemExpress), K-975 (HY-138565, MedChemExpress), IK-930, \nverteporfin (SML0534, Sigma Aldrich) , lovastatin (S2061-SEL, Selleckchem) , and dasatinib, \ngifted by Prof Neil Carragher. \n \nGene knockout and ectopic expression of YAP-5SA \nBriefly, CRISPR-Cas9 mediated KO was carried out in MeT-5A mesothelial cells as described in \nCunningham et al 6, with the additional use of the following guide RNA (gRNA):  5’-\nTGGCAGTGGCCGAGACGATC-3’ for TEAD1 and  5’-CTCAAGGATCTCTTCGAACG-3’ for TEAD4. \nValidation of KO via western blotting. Expression of YAP-5SA in MeT-5A cells was performed \nby lentiviral transduction via pQX system under hygromycin B (30-240-CR, Corning) selection. \nLentivirus for transduction was produced in HEK293T cells , harvested 48 and 72 hours after \ntransfection using GenJet transfection reagent ( SL100488, SignaGen), and filtered using low \nbinding 0.45 µm SFCA filters (Corning, 431220). Selection was carried out via hygromycin \ntreatment 24 hours post-transduction to ensure time for the development of resistance. \n \nWestern Blotting \nWestern blots were performed with c ell lysates harvested and run using homecast gels , as \ndescribed in Hansen et al87. PageRuler prestained Protein Ladder (26616, Thermo Scientific) \nwas included in western blots  as a scale for protein size . Separated proteins were \nsubsequently transferred from gels to Immobilon-P PVDF membranes (IPVH00010, Millipore) \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 20 \nand blocked in 5% milk in TBS-T, with subsequent primary and secondary antibody incubation, \nwashing, and development carried out as described in Cunningham et al6. Phos-Tag western \nblots were conducted with  Phos-Tag reagent (304 -93521, Alpha Laboratories) and 10mM \nMnCl2 added to each polyacrylamide resolving gel. Primary antibodies used were as follows: \nYAP (ab52771, Abcam),  CYR61 ( 14479, CST), NF2 (D1D8, CST),  TEAD1 ( 610922, BD \nBiosciences), TEAD4 (ab58310, Abcam), with HSP90 (BD610418, BD Biosciences) used as a \nloading control for samples. \n \nRT-qPCR analysis \nCells were plated and harvested, with cDNA generated as described in Cunningham et al6. RT-\nqPCR assays were carried out using Brilliant III Ultra -Fast SYBR Green (600883, Agilent) and  \nLightCycler 480 SYBR Green I  (04707516001, Roche) RT-qPCR Master Mixes, used according \nto manufacturer’s directions. IDT primers were custom-designed using templates deposited \non PrimerBank89. RT-qPCR was carried out  on a QuantStudio 5 Real -Time PCR System, with  \nresulting data analysed using R statistical software. Primer sequences were as follows: 5 ’-\nGCTCGTTGAGTGAACGGCT-3’ and 5 ’- CATGAGCTAGTACAACATGAGGG -3’ for AMOTL2; 5’-\nAGTAGAGGAACTGGTCACTGG-3’ and 5 ’-TGTTTCTCGCTTTTCCACTGTT-3’ for ANKRD1; 5’-\nCCAAGGTGAGCTTTCCCTCG-3’ and 5 ’-CCTACTAGACCATAGGTCGTCGT-3’ for ARHGEF17; 5’-\nTAGAACAGCCCTTCAGAAAGTGA-3’ and 5 ’-CGGGGTTGTCTCGACTTAAAAA-3’ for ASAP1; 5’-\nGTGGGCAACCCAGGGAATATC-3’ and 5 ’-GTACTGTCCCGTGTCGGAAAG-3’ for AXL; 5’-\nCCCTGTGACGAGTCCAAGTG-3’ and 5 ’-GGTTCCGTAAATCCCGAAGGT-3’ for CRIM1; 5’-\nGAGGCAGAAGTACGGGGTTG-3’ and 5 ’-CAGGAATCACGGTTTCATGCT 3’ for DOCK5; 5’-\nGGCGCTTCAGGCACTACAA-3’ and 5 ’-TTGATTGACGGGTTTGGGTTC-3’ for F3; 5’-\nGCTGGTGGACCTAGTACAATGG-3’ and 5 ’-CTTACGAGCCGGTCGAAGTTG-3’ for FJX1; 5’-\nAATGCCACTCGCCCTACAC-3’ and 5 ’-CGTTCTGGTGCAAGTAGCTCT-3’ for FOXF2; 5’-\nGAGAGCAGAAGACCGAAAGGA-3’ and 5 ’-CACAACACCACGTTATCGGG-3’ for GADD45A; 5’-\nAGAGCACAGATACCCAGAACT-3’ and 5 ’-GGTGATTCAGTGTGTCTTCCATT-3’ for IGFBP3; 5’-\nACTTTTCCTGCCACGACTTATTC-3’ and 5 ’-GATGGCTGTTTTAACCCCTCA-3’ for LATS2; 5’-\nTAATTGGCACGGCGACTGTAG-3’ and 5 ’-GGAGATCAGCTTGTACGGCAG-3’ for MYOF; 5’-\nGCCTGGGAGCTTACGATTTTG-3’ and 5 ’-TAGTGCCCTGGTACTGGTCG-3’ for NT5E; 5’-\nCGCCCAAGCCCCTAATGAAG-3’ and 5 ’-TCCCTCCGTATGTGCATCAGA-3’ for NUAK2; 5’-\nATGCCTTTTGGTCTGAAGCTC-3’ and 5 ’-CCCTGTGCTTTCCACCGAC-3’ for PTPN14; 5’-\nGGGGAACAGTTGAGTAAAACCA-3’ and 5 ’-ACAATTTTTCCATACGGTTGGCA-3’ for RBMS3; and  \n5’-CAGCACACTCGATATGGACCA-3’ and 5’-CCTCGGGCTCAGGATAGTCT-3’ for TGFB2. All gene \nexpression values were normalized to Hypoxanthine Phosphoribosyltransferase 1  (HPRT1) \nexpression. \n \nYAP nuclear localisation high content image-based assays \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 21 \nCells were seeded in 384-well µClear plates ( 781091, Grenier)  at variable cell densities \ndetermined for each genotype to ensure cells equivalent confluency over time-points tested. \nWT MeT-5A cells were seeded at 1,000 cells per well, NF2 KO at 833 cells per well and YAP KO \nat 2,000 cells per well.  Upon attachment, c ells were treated with relevant inhibitor for 24 \nhours before fixing , permeabilising, and staining as described in Cunningham et al 6. \nFluorescent cells were imaged using the Oper a Phenix  Plus high-content imaging system \n(PerkinElmer) at 20x across 3 biological replicates, with 9 fields imaged per well and four wells \nper sample/condition to act as technical controls. A bespoke CellProfiler90 pipeline was \nimplemented to segment cells and quantify relevant intensities and morphological features. \nStatistical analyses were conducted using R, with cell-level data trimmed to ensure detectable \nnuclear and cytoplasmic YAP . Cells were trimmed according to cell -contact, with those \nbetween 45-55% cell-cell contact retained (7.6% cells of 1,381,549 total, with a median of 89 \ncells/well included for downstream analysis ) and v ariance accounted for post-collation by \nremoving outl iers, defined as wells possessing values greater than 2 median absolute \ndeviations (MAD) from the median of biological replicates. Processed data were plotted with \nGraphPad Prism. \n \nProliferation and spheroid formation assays \nFor quantification of proliferation, ce lls were seeded at 2,000, 1,500, or 2,500  cells per well \nfor WT, NF2 KO and YAP KO MeT -5A cells respectively  in 96-well µClear plates (655090, \nGrenier) and treated with relevant inhibitors when attached , 24 -hours post -seeding. Cells \nwere then imaged over the course of 72-hours and confluency calculated using the CELLCYTE \nX (CYTENA). Growth was quantified by normalising slope statistics from  computed logistic \ngrowth curves to DMSO treated wells , with rates of growth < 0 adjusted to 0. For spheroid \nformation assays, c ells were seeded at 500 cells per well in ultra -low attachment 96 -well \nplates (CLS7007, Corning) and cultured for 7 days. For live imaging assays and quantification \nof spheroid size, cells were incubated and imaged throughout this time perio d using the \nCELLCYTE X. To limit the influence of technical artefacts associated with the high variance of \nthis platform, time -points exhibiting spheroid areas greater than 2 MADs from the median  \narea at that time-point were trimmed, with any well with >50% trimmed time-points defined \nas outliers and excluded from analysis . 7 -day spheroid areas were normalised to 24 -hour \ntime-points within each well . For treated spheroids, treatment was initiated ~1 hour  post-\nseeding, with resulting spheroids imaged at day 7 using the EVOS FL Auto 2 Imaging System \n(Invitrogen) at 10x magnification . Quantification of spheroid size was then achieved using \nCellProfiler, with the CellPose91 plugin to enhance spheroid segmentation. Brightfield images \nacquired were then subjected to morphological analysis as described in C ell Painting and \nmorphological analysis, with a pipeline modified for single-channel quantification. \n \nCell Painting and morphological analysis \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 22 \nCell Painting assays were conducted according to established protocols 76, with cells  seeded \nas with high -content imaging in 384-well µClear plates ( 781091, Grenier) and treated with \nappropriate inhibitor for 24 hours before fixing, staining and imaging. Image acquisition using \nthe Opera Phenix Plus high-content imaging system (PerkinElmer)  was performed with the \nsame set-up as described for YAP nuclear localisation high content image -based assays . \nMorphometric datasets were then acquired from resulting images using CellProfiler90 \npipelines esta blished by JUMP -CP protocols 71,92, with the implementation of a modified \nversion of pipelines available at github.com/broadinstitute/imaging-platform-\npipelines/tree/master/JUMP_production. The cytominer93 package was then used to trim low \nvariance or highly correlated features  (leading to retention of 417 of 3,446 input features; \n12.11%) and normalise retained features to vehicle control (DMSO) treated WT MeT-5A cells. \nNormalised high-dimensional morphometric datasets were visualised using circlize94 package \nin R. Morphological uniformity was computed by calculating correlation distances between \nall preserved features in each like -treated well. Feature disruption  was quantified  via a \nmodified robust Z’ scoring, designed to act as a simplified, relative measure  of a feature to \nresolve treated cell morphology from vehicle control (DMSO ). This was calculated by \nremoving the scaling constant of 3  of the traditional Z ’ score and using median/MAD as \nopposed to mean/standard deviation as measures across wells, given no assumption of data \nto conform to a normal distribution. To implement scoring, the following equation was used:  \n \n𝑀𝑜𝑑𝑖𝑓𝑒𝑑 𝑍’ = 1 − 𝑀𝐴𝐷𝑇𝑟𝑒𝑎𝑡 + 𝑀𝐴𝐷𝐶𝑜𝑛𝑡𝑟𝑜𝑙\n𝑀𝑒𝑑𝑖𝑎𝑛𝑇𝑟𝑒𝑎𝑡 − 𝑀𝑒𝑑𝑖𝑎𝑛𝐶𝑜𝑛𝑡𝑟𝑜𝑙\n \n \nModified Z’ scores were adjusted to quantify YAP dependency by subtracting YAP KO from \nWT scores. \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 23 \nFIGURE LEGENDS \nFigure 1| Direct inhibitors of TEAD and indirect inhibitors of YAP differentially effect YAP-\nTEAD activity in vitro.  \na, Bar‐plot shows TEAD luciferase‐based activity in HEK293A cells, treated at 1 µM with \nvarious selective TEADi (red), or non-specific YAP inhibitors (green) for 24 hours. Points \nrepresent individual technical replicates (n=4). All inhibitors, besides lovastatin, decrease \nTEAD based luciferase activity. b, Violin plot shows the expression of YAP/TAZ signature \ngenes, as determined via RT‐qPCR, in HEK293A cells after 24‐hour treatment of select high \npotency inhibitors determined in (a) at 1 µM (n=3). Dots represent the log-transformed fold-\nchanges of individual YAP/TAZ target genes, with a clear downregulation across most \nquantified signature genes. c, Violin plot, as in (b), shows downregulation of YAP/TAZ \nsignature genes in MeT‐5A cells on treatment with YAP/TAZ‐TEAD inhibitors (n=3). d, \nRepresentative confocal images of YAP nuclear localisation in MeT‐5A cells (DMSO, vehicle \ncontrol, left), or with direct TEAD (VT‐107, middle) or non‐specific YAP (dasatinib, right) \ninhibitors at 370 nM for 24 hours. Increased cytoplasmic YAP is seen as compared to vehicle \ncontrol (DMSO). Scale bars = 50 µm. e, Bar‐plot shows quantification of images as depicted \nin (d), with data presented as % nuclear:cytoplasmic YAP ratios adjusted to vehicle controls. \nEach dot represents the median values of individual wells (n=3). f, Phos‐Tag based western \nblot shows YAP phosphorylation in WT MeT‐5A cells in response to 370 nM 24‐hour \ntreatment. YAP phosphorylation is induced solely on treatment with dasatinib. g, Phos‐Tag \nbased western blot conducted as in (f) on cell lysates from NF2 KO MeT‐5A cells shows a \ndecrease in sensitivity to dasatinib, in terms of YAP phosphorylation relative to WT cells. P \nvalues in (a)‐(c) were determined by one‐sample Wilcoxon signed rank test and (e) using \nMann‐Whitney U test, adjusted for multiple comparisons. n.s. = Not significant, *P < 0.05, \n**P < 0.01, ***P < 0.001, and ****P < 0.0001 relative to WT. \n \nFigure 2| Specific TEAD-selective inhibition decreases NF2-deficient cancer-relevant \nphenotypes. a, Representative images of 2D proliferation in WT MeT‐5A cells at 10 µM YAP‐\nTEAD inhibition. Scale bar = 50 µm. b, Bar‐plot of data from images as shown in (a) reveals \ninhibition of growth selectively in all non‐specific YAP inhibitors (green) tested. Direct TEADi \n(red) has modest impact on cell growth in 2D compared to non‐specific YAP inhibitors \n(green). Points represent the growth curve statistics of individual wells (n=3). c, Violin plot \nshows cell proliferative sensitivity of NF2 KO MeT‐5A cells, with decreases in cell growth \nshown relative to those in WT cells (b). Little change in sensitivity is observed upon loss of \nNF2. d, Violin plot as in (c) showing sensitivity to YAP/TAZ‐TEAD inhibition in YAP KO relative \nto WT MeT‐5A cells, in terms of cell growth rates in 2D. e, Representative images of spheroid \nformation in WT (top), NF2 KO (middle) and YAP KO (lower). Scale bar = 50 µm. f, \nQuantification of spheroids imaged at day 7, as in (e), shown as violin plots. WT (grey), two \nNF2 KO clones (red), and YAP KO (green) MeT‐5A cells. Genotypes form spheroids of varying \nsizes, with spheroids containing NF2 loss (larger) and YAP loss (smaller), respectively. #1 and \n#2 NF2 KO are independently generated clones. g, Violin plot shows sensitivity of WT MeT‐\n5A spheroids to YAP‐TEAD inhibition. Each dot represents size of an individual spheroid at \nday 7, relative to size at 24‐hours to account for variable seeding (n=3). Spheroid size is \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 24 \nmodestly decreased upon high concentration TEAD inhibition. h, Violin plot, as in (g) shows \nsensitivity of NF2 KO spheroids to YAP‐TEAD inhibition relative to WT. NF2 KO MeT‐5A \nspheroids have enhanced sensitivity to most tested inhibitors. Inhibitors tested in (b)‐(d) \nwere used at 1.1, 3.3, and 10 µM and in (g)‐(h) at 1 and 10 µM and are compared to vehicle \ncontrol (DMSO). P values in (b)‐(h) were determined by Mann‐Whitney U test, adjusted for \nmultiple comparisons. n.s. = Not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < \n0.0001 relative to WT. \n \nFigure 3| YAP and TEAD differentially orchestrate NF2-deficient spheroid sensitivity. \na, Western blots from MeT‐5A cells stably expressing YAP‐5SA, a hyperactive form of YAP. An \nincrease in protein levels of YAP and CYR61, a downstream transcriptional target of YAP‐\nTEAD, confirm induction of YAP‐5SA expression. b, Violin plot shows the relative spheroid \ngrowth of YAP‐5SA expressing MeT‐5A spheroids, compared to WT control. Gradients show \n% growth relative to parental NF2 single KO (red), WT (grey), and YAP KO (green). c, Violin \nplot shows sensitivity of MeT‐5A spheroids expressing YAP‐5SA to TEAD inhibition, relative to \nWT control spheroids. d, Western blots from gene edited combinatorial KO clones, show KO \nof YAP and TEAD isoforms alongside NF2 in MeT‐5A cells. e-f, Quantification of relative \nspheroid growth shown in violin plot comparing impact of YAP (e) or TEAD (f) KO in NF2‐\ndeficient MeT‐5A spheroids. Gradients as in (b) show % growth relative to single KO \ngenotypes. g-h, Violin plots show sensitivity of NF2/TEAD1 dKO (g) and NF2/TEAD4 dKO (h) \nMeT‐5A spheroids to TEAD inhibition. Loss of either TEAD isoforms results in a decrease in \nsensitivity to TEAD inhibition. Spheroids in (c), (g), and (h) were treated for 24 hours with 1 \nand 10 µM of selective TEAD inhibitors. Each dot represents an individual spheroid across 3 \nbiological replicates. P values in (b), (c), (e)‐(h) were determined by Dunnet’s multiple \ncomparison test. n.s. = Not significant, *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < \n0.0001.  \n \nFigure 4| Morphological disruption predicts specificity of YAP/TAZ-TEAD inhibitors.  \na, PCA plot shows morphological profiles of MeT‐5A cells cultured in 2D, with features \ncollapsed via dimensionality reduction to allow visualisation. Each point represents a single \nwell, with highlighted circles representing the centre of DMSO treated cells coloured \naccording to genotype. Ellipses show the spread of treatments irrespective of genotype and \narrows show trajectories across increasing concentrations of treatment for each genotype. \nb, Heatmap shows correlation distance of computed features between wells. Clustering of \ncells, each representing a single well, was performed via complete linkage, with annotations \n(top) showing the genotype, inhibitor family, name of treatment, and treatment \nconcentration for each well. c, Violin plot portraying morphological uniformity of WT MeT‐\n5A cells on YAP/TAZ‐TEAD inhibition. Each point represents a single well, with uniformity \ncalculated as the mean correlation distance of a single well to all like‐treated wells. d, Violin \nplot as in (c), showing morphological uniformity on YAP/TAZ‐TEAD inhibition in YAP KO MeT‐\n5A cells. e, Heatmap shows the percentage of disrupted features, as defined by modified Z’ \nscores > 0, in WT vs YAP KO MeT‐5A cells in response to various YAP/TAZ‐TEAD inhibitors. As \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 25 \ntreatment concentration increases, a greater percentage of features are disrupted across \nmost inhibitors. f, Heatmap shows modified Z’ scores of features with known discriminatory \npotential, defined as those with modified Z’ scores > 0, adjusted for YAP‐dependency. \nPositive values indicate feature disruption selectively in WT, and not in YAP KO MeT‐5A cells, \nindicating dependence on YAP , while negative values show features disrupted to the same or \ngreater degree in YAP KO cells. Cells in (a), (c), (d), and (f) were treated for 24 hours with 1.1, \n3.3, and 10 µM of compounds. P values in (c) and (d) were determined by Dunnet’s multiple \ncomparison test. n.s. = Not significant, *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < \n0.0001.   \n \n \n \nSupplementary Figure 1| TEAD, but not NF2, mediates effects of YAP/TAZ-TEAD inhibitors.  \na, Violin plot as in figure 1b shows the expression of YAP/TAZ signature genes, as determined \nvia RT‐qPCR, in TEAD KO HEK293A cells, with expression depicted as log‐transformed fold‐\nchange on treatment with select YAP/TAZ‐TEAD inhibitors relative to DMSO. No significant \ndysregulation of signature genes is observed on treatment. b, Dose response quantification \nof decrease in nuclear YAP from images as in figure 1d. Dots represent the median \nnuclear:cytoplasmic YAP ratio across cells in individual wells over 3 biological replicates after \n24‐hour treatment. c, Violin plots comparing sensitivity of NF2 KO relative to WT MeT‐5A \ncells to YAP/TAZ‐TEAD inhibition, in the context of YAP/TAZ signature expression. Dots \nrepresent the individual genes comprising the signature, where sensitivity is the fold‐change \nin up/down‐regulation relative to WT response in two independent NF2 KO clones (left and \nright). d, Violin plots as in (c) show the sensitivity of NF2 KO MeT‐5A cells to inhibition in the \ncontext of YAP nuclear localisation in two independent NF2 KO clones (left and right). e, \nPhos‐tag based western blot, as in figure 1f, shows YAP phosphorylation in WT MeT‐5A cells \nin response to 24‐hour treatment with YAP/TAZ‐TEAD inhibitors at 10 µM. f, Phos‐tag based \nwestern blot (top) showing difference in response between WT (left), LATS1/2 dKO (mid), \nand TEADs KO (right) HEK293A cells to YAP/TAZ‐TEAD inhibition, in terms of difference in YAP \nphosphorylation. The same lysates were analysed on a conventional Western blot and \nprobed for YAP and HSP90 (bottom). Where not stated, cells were treated with 1 µM \ninhibitor for 24‐hours. P values in (a) and (c) were determined by one‐sample Wilcoxon \nsigned rank test and (d) using Mann‐Whitney U test, adjusted for multiple comparisons. n.s. \n= Not significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 relative to WT. \n \nSupplementary Figure 2| NF2 differentially modulates YAP activity in cells cultured in \nspheroids compared to in a 2D monolayer. \na, Violin plot as in figure 2c representing sensitivity of NF2 KO MeT‐5A cells to YAP/TAZ‐TEAD \ninhibition, in terms of 2D proliferation. b, Cross‐section of spheroids at day 7 (left), with \nlabelled actin (phalloidin, green), nuclei (DAPI, blue) and YAP (red). Scale bar = 50 µm. \nQuantification of nuclear:cytoplasmic ratios of YAP in spheroids vs cells cultured in 2D (right) \nreveals a significant decrease in nuclear YAP in MeT‐5A spheroids. Individual points \nrepresent median ratios across well of DMSO treated cells (2D monolayer), as in figure 1d, or \nmedian ratios within an individual spheroid (spheroid). c, YAP nuclear localisation analysed \ndisplayed in violin plots, from confocal images of spheroids as in (b) in WT (grey), NF2 KO \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n 26 \nclone #1 (light red), and clone #2 (dark red) MeT‐5A cells. d, Sensitivity of NF2 KO spheroids \n(decrease in spheroid size, compared to WT spheroids) after treatment with various TEADi \n(red), or non‐specific YAP inhibitors (green), compared to DMSO (grey). e, Spheroid growth \nof NF2 KO clones upon TEADi (red) or non‐specific YAP inhibition (green) normalised to \nDMSO control. Dotted line indicates WT (red) and YAP KO (green) spheroid growth at steady \nstate. Cells in (a) and spheroids in (d), and (e) were treated for 24 hours at 1.1, 3.3, and 10 \nµM and 1 and 10 µM respectively across 3 biological replicates. P values in (a), (d) and (e) \nwere determined by Dunnet’s multiple comparison test, while those in (b) and (c) via Mann‐\nWhitney U test. n.s. = Not significant, *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < \n0.0001. \n \nSupplementary figure 3| Impact of YAP-TAZ/TEAD inhibitors on spheroid morphology.  \na, Violin plot shows spheroid sensitivity to direct TEAD inhibition in NF2/TEAD1 dKO, clone \n#2 relative to parental NF2 KO. Points represent the decrease in spheroid growth, defined as \nspheroid size at day 7 relative to 24‐hours, in individual NF2/TEAD1 dKO spheroids as \ncompared to NF2 KO (n=3). b, Spheroid sensitivity as in (a) to TEAD inhibition in NF2‐TEAD4 \ndKO, clone #2 relative to NF2 KO. c, Representative images of WT MeT‐5A spheroids, at day 7 \nafter treatment with 10 µM of indicated compound. Scale bar = 50 µm. d, PCA plot shows \nthe distinct morphological profiles of MeT‐5A spheroids after YAP/TAZ‐TEAD inhibition. \nSpheroids were imaged and brightfield morphologies quantified, with each point \nrepresenting a single spheroid. Circles represent the centre of DMSO treated spheroids, \ncoloured according to genotype, while ellipses show the morphological spread of treatments \ncombining genotypes. Arrows show the morphometric trajectories along increasing \ntreatment concentrations for each genotype. e, Representative images of Cell Painting in WT \nMeT‐5A cells. Indicated subcellular regions were labelled via fluorescent staining. Scale bar = \n50 µm. f, Violin plots, as in figure 4c, showing the morphological uniformity in NF2 KO MeT‐\n5A cells, as determined via Cell Painting, across 3 biological replicates. Each dot represents a \nsingle well. Spheroids in (a) and (b) and cells in (d) and (f) were treated for 24 hours \nrespectively with either 1 and 10 µM or 1.1, 3.3, and 10 µM of indicated inhibitor. P values in \n(a), (b) and (f) were determined by Dunnet’s multiple comparison test. n.s. = Not significant, \n*P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001. \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\nYAP Phosphorylation – MeT-5A\na\n b\nd\nc\nf\nTEAD Luciferase\nHEK293A\nYAP/TAZ Signature\nHEK293A\nYAP/TAZ Signature\nMeT-5A\nYAP Nuclear Localisation\nMeT-5A\nDMSO VT-107 Dasatinib\nDNA Actin\nYAP DNA\nActin\nYAP\nDNA Actin\nYAP DNA\nActin\nYAP\nDNA Actin\nYAP DNA\nActin\nYAP\ne\nWT\nDMSO\nVT-107\nK-975\nIK-930\nVerteporfin\nLovastatin\nDasatinib\nYAP\nHSP90\nYAP -70kDa\n-110kDa\n- PPPPP- PPPP- PPP- PP- P- –\nNF2 KO\nDMSO\nVT-107\nK-975\nIK-930\nVerteporfin\nLovastatin\nDasatinib\nYAP\nHSP90\nYAP\n- PPPPP- PPPP- PPP- PP- P- –\n-70kDa\n-110kDa\nYAP Nuclear Localisation\nMeT-5A\ng\nFIGURE 1\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\na\nc\ne\nh\nWT\nNF2 KO\n0 0.5 1 3\nDays Post-seeding\n7\nYAP KO\nDMSO\nVT-107\nDasatinib\n0 24 48 72 \nHours Treatment b\nf\ng\nd\nProliferation\nSpheroid Formation\n72 Hour Growth\nYAP/TAZ-TEAD Inhibition\n72 Hour Growth – Sensitivity\nNF2 KO\n72 Hour Growth – Sensitivity\nYAP KO\nSpheroid Growth\nArea Relative to WT\nSpheroid Growth\nYAP/TAZ-TEAD Inhibition\nSpheroid Sensitivity\nNF2 KO\nFIGURE 2\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\nCYR61\nNF2 \nKO TEAD1 TEAD4\nTEAD1\nNF2\nYAP\nHSP90\nTEAD4\n-70kDa\n-55kDa\n-55kDa\n-70kDa\n-110kDa\nWT #1 #2 #1 #2 #1 #2 #1 #2\nYAP\nNF2 dKO\na\ng\nb c\nh\nd NF2 dKO Validation e f\nWT\nYAP-5SA\nYAP -70kDa\n-40kDa\nYAP-5SA Expression\nValidation\nSpheroid Growth\nYAP-5SA Expression\nSpheroid Sensitivity\nYAP-5SA Expression\nSpheroid Sensitivity\nNF2/TEAD1 dKO\nSpheroid Sensitivity\nNF2/TEAD4 dKO\nSpheroid Growth\nNF2/YAP dKO\nSpheroid Growth\nNF2/TEAD dKO\nHSP90 -110kDa\nFIGURE 3\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\ne\nVT-107\nK-975\nIK-930\nVerteporfin\nDasatinib\nLovastatin\nModified Z’\nYAP Specificity\nYAP-Independent features YAP-Dependent features\nVT-107\nK-975\nIK-930\nVerteporfin\nDasatinib\nLovastatin\nWT YAP\nKO\nWT YAP\nKO\nWT YAP\nKO\n1µM 3µM 10µM\n% Disrupted\nFeatures\na\nf\nb\nc d\nGenotype WT\nYAP KO\n Treatment\nVT-107 K-975 IK-930\nVerteporfin Dasatinib Lovastatin\nDMSO\n#1\n#2\n NF2 KO\nCell Painting Morphology\nPCA\nCell Painting Morphology\nCorrelation Distance Heatmap\nMorphological Uniformity\nYAP/TAZ-TEAD Inhibition\nMorphological Uniformity\nYAP KO\nModified Z’ Scoring\nWT vs YAP KO\nModified Z’ Scoring\nAdjusted for YAP-dependency\nFIGURE 4\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\ne\nWT\nDMSO\nVT-107\nK-975\nIK-930\nVerteporfin\nLovastatin\nDasatinib\nYAP\n- PPPPP- PPPP- PPP- PP- P- –\na\nc\nd\nb\nYAP/TAZ Signature\nHEK293A – TEAD KO\nYAP Nuclear Localisation\nMeT-5A\nf\nWTDMSO\nLovastatin\nDasatinib\nDMSO\nLovastatin\nDasatinib\nDMSO\nYAP\nHSP90\nYAP -70kDa\n-110kDa\n- PPPPP\n- PPPP\n- PPP\n- PP\n- P\n- –\nLovastatin\nDasatinib\nLATS dKO TEAD KO\nYAP/TAZ Signature\nNF2 KO, Clone #1\nYAP/TAZ Signature\nNF2 KO, Clone #2\nYAP Nuclear Localisation\nNF2 KO, Clone #1\nYAP Nuclear Localisation\nNF2 KO, Clone #2\nYAP Phosphorylation – MeT-5A YAP Phosphorylation – HEK293A\nSUPPLEMENTARY FIGURE 1\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\n72 Hour Growth – Sensitivity\nNF2 KO, Clone #2\na\ne\nWT\nYAP\nKO\nWT\nYAP\nKO\nDNA Actin\nYAP DNA\nActin\nYAP\nb YAP Nuclear Localisation\nSpheroid Immunofluorescence\nc YAP Nuclear Localisation\nSpheroid, WT vs NF2 KO\nd Spheroid Sensitivity\nNF2 KO, Clone #2\nSpheroid Growth\nNF2 KO, Clone #1\nSpheroid Growth\nNF2 KO, Clone #2\nSUPPLEMENTARY FIGURE 2\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint \n\na\nb d\nDMSO\nVT-107 K-975 IK-930\nVerteporfin Lovastatin Dasatinib\nTEADi\nNon-specific YAP Inhibitors\nc\nTreatment\nVT-107 K-975 IK-930\nVerteporfin Dasatinib Lovastatin\nDMSO\nGenotype WT\nYAP KO\n#1\n#2\n NF2 KO\nf\nSpheroid Sensitivity\nNF2/TEAD1 dKO #2\nSpheroid Sensitivity\nNF2/TEAD4 dKO #2\nSpheroid Morphology\nPCA\nMitochondria\nActin, Golgi &\nPlasma Membrane\nNucleolar & \nCytoplasmic RNA\nEndoplasmic\nReticulum DNA\ne Cell Painting\nRepresentative Images\nMorphological Uniformity\nNF2 KO, Clone #1\nMorphological Uniformity\nNF2 KO, Clone #2\nSUPPLEMENTARY FIGURE 3\n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted November 12, 2024. ; https://doi.org/10.1101/2024.11.12.622964doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}