YAP-TEAD interaction inhibitors decrease YAP nuclear localization and cell velocity.
With the objective to investigate the biological effects of different Inventiva™-supplied YAP/TEAD interaction inhibitors compounds, immunofluorescence assays were performed to calculate the intensity ratio between YAP nucleus and cytoplasm staining. These tests were carried out using A549, H2052 and Met5A cell lines, grown at low to moderate density, to avoid any mechano-transduction effect potentially leading to YAP nucleus exit in the event of high cell density (Fig. 1). Following treatment, this ratio was significantly lowered in the A549 and H2052 cells, namely two cell lines previously reported to have increased basal YAP activity, owing to STK11 mutation and RASSF1A methylation for the former (15, 16), and NF2 and LATS2 mutation for the latter (20). The effect was particularly marked following IV#6 (p < 0.0001, n = 3, Student t-test) (Fig. 1, upper and middle panel). Conversely, the ratios observed in the Met5A cell line did not support any IV#6 effect in such immortalized, non-tumorigenic cells, without known alteration in the Hippo/YAP pathway (Fig. 1, lower panel).
Thereafter, we analyzed the migration speed of A549 and H2052 cells in the presence of two YAP/TEAD interaction inhibitors IV#5 and IV#6 at increasing concentrations (Fig. 2). These treatments significantly and dose-dependently decreased the 2D cell velocity. Representative images of the wound healing assay were taken at 0 and 48 hours after cell- free gap creating (Fig. 2), and the velocity (µm/h) was calculated using Avemap™ software, which determines cell velocity, using an image correlation method, without being influenced by cell proliferation (23). The 20µM dose was likely the most efficient dose, resulting in significantly decreased cell migration (p < 0.0001, n = 3, Student t-test), without any cell toxicity. Accordingly, all other Inventiva™ YAP-TEAD interaction inhibitors reduced cell velocity of both cell lines at 20µM either (Suppl. Figure 1), although their effect was more pronounced in H2052 cells, suggesting a higher YAP inhibition susceptibility in these cells, as compared with A549 cells.
YAP-TEAD interaction inhibitors partly revert epithelial to mesenchymal (EMT) transition induced by TGF-β in A549 lung cancer cells.
As previously reported by us (15), along with other authors, YAP signaling activation, by RASSF1A depletion for instance, could induce EMT in epithelial lung cancer cells. A549 cells are already engaged towards a partial EMT, as reflected by low basal N-Cadherin expression levels (Figure Suppl.2, panel A). This could be further stimulated by TGFβ treatment as shown by the increased N-cadherin expression, with concurrent decrease of E-cadherin (26). Upon YAP pharmacological inhibition with the two most active compounds at 20µM (IV#5 and IV#6), we were able to inhibit N-cadherin expression induced by TGFβ treatment (Figure Suppl.2, line 1 panel B). Concurrently E-cadherin failed to increase, suggesting that YAP pharmacological inhibition only partly reverted EMT in this lung cancer cell model (Figure Suppl. 2, panel C for quantification of four independent experiments). Such results clearly support an actual effect of tested YAP/TEAD inhibitors in disrupting YAP/TEAD interaction, with partial reversibility of TGFβ-induced EMT as a functional consequence.
Anti YAP/TEAD Compounds sensitizes A549 cells to MEK inhibitor
Trevor Bivona's team identified YAP pathway activation as a novel resistance mechanism to kinase inhibitors targeting RAF and MEK therapy, in the presence of B-RAF V600E or K-RAS activating mutations. These authors revealed a synthetic lethality effect induced by RNA interference-mediated depletion of YAP, with simultaneous inhibition of MEK by trametinib, specifically in K-Ras or B-Raf mutated cell lines. Of note such effect was not observed in cells without MAP-K signaling constitutive activation (27). We thus sought to recapitulate this effect using pharmacological compounds targeting YAP/TEAD interaction, in combination with trametinib. This was meant to support their specific effect on YAP downstream signaling in the A549 cell line exhibiting a K-Ras G12S oncogenic mutation, with predominant nuclear YAP staining. For this, we calculated the trametinib IC50 in our cancer cells, in the presence of Inventiva™-supplied compounds over 7 days (Fig. 3A). Using pharmacological agents, we recapitulated the synthetic lethality induced in such cells by YAP genetic depletion while combining trametinib with the IV#6 compound. Indeed, the trametinib IC50 was significantly reduced by IV#6 at 20µM for 7 days (p = 0.0432, Student t-test, n = 3). The compounds without trametinib did not exhibit significant toxicity (Fig. 3B). Thereafter, we calculated the trametinib IC50 in the presence of YAP SiRNA for three days, prior to the YAP transient knockdown’s fading (Fig. 3C). YAP knockdown did not decrease cell viability per se (Fig. 3D). YAP knockdown was maximal over 90% at 3 days (Fig. 3E). Trametinib’s pharmacological efficacy was verified in cell extracts by the decreased phospho-ERK band on Western blot (Fig. 3F). As observed with pharmacological inhibition, YAP knockdown actually enhanced trametinib’s efficacy on A549 cells (p < 0.05, Student's t-test, n = 3), as did YAP/TEAD pharmacological inhibitor (Fig. 3C).
Chemo-resistance of paclitaxel-resistant lung cancer cell line correlates with an increase of YAP activity
To compare basal activity of YAP in parental cell lines and their paclitaxel-resistant counterparts, we performed a functional transactivation luciferase reporter assay as described by the S. Piccolo's team (28). The protocol involves two plasmids co-transfection including a reporter plasmid containing TEAD minimal promoter sequences upstream the gene encoding firefly luciferase (Firefly plasmid - pGL3b 8xGTIIC TEAD), as well as a plasmid constitutively expressing renilla luciferase (Renilla lpRL-TK plasmid), to normalize firefly luciferase’s enzymatic activity, for overcoming the variation of plasmid’s transfection efficiency and transfection-induced cell death. Thus, the bioluminescence is expected to be proportional to the nuclear content of active YAP interacting with nuclear TEAD. YAP activity, measured by the luciferase reporter assay, was revealed to be clearly higher in resistant A549 and HCC827 cells than in their parental counterparts (p = 0.0082 for A549 and p = 0.0014 for HCC827; n = 10 for each cell line; Student-t test) (Fig. 4A).
To further confirm these results, we performed YAP immunofluorescence to compare YAP’s localizations among different cell lines. To this end, we calculated the YAP nuclear to cytoplasmic labelling ratio in 150 arbitrarily chosen cells for each condition, ensuring the cell density was low-to-moderate and being comparable. The nuclear to cytoplasm YAP fluorescent signal ratio was significantly higher in the resistant A549 cells (p < 0.0001; n = 150 cells, manual cell quantification, Student-t test) than in A549 parental cells (Fig. 4B). Yet, the main localization of YAP did not significantly differ between parental and resistant HCC827 cancer cells (p = 0,6; n = 150 cells; paired t-test) (Fig. 4B).
YAP/TEAD transcription activity is inhibited by IV#6
To characterize IV#6's effect, we performed luciferase reporter assays and analyzed YAP nuclear immunofluorescence staining in in the compound’s presence or absence. YAP activity, was significantly decreased in the presence of IV#6 in A549 cells (p = 0.0156 for parental and resistant A549 cell line, n = 7; Wilcoxon test) (Fig. 5A). Conversely, despite a clear trend as (shown in Fig. 5A), YAP activity was not significantly altered by the compound's presence in HCC827 cells (p = 0.1250 for parental HC827 and p = 0.3125 for resistant HCC827; n = 7, Wilcoxon test), contrasting with IV#6's effect in cells exhibiting molecular alterations of Hippo pathway genes, which show exquisite sensitivity to YAP inhibition, possibly related to the level of YAP activation (21). The IV#6 dose used was 10µM in A549 cells and 3µM in HCC827, since 10 µM was shown to be excessively toxic in the latter (data not shown), but which also could explain why the effects were weaker in HCC827 cells than in A549 cell line.
YAP localization was studied in the presence or absence of the IV#6 YAP/TEAD interaction inhibitor compound, using three independent immunofluorescence experiments, with a total of 150 nucleocytoplasmic ratio measurements in each A549 and HCC827 cell sub-clone. A significant decrease in YAP nuclear localization was observed for cells treated with the inhibitor, 24 hours prior to labeling, in parental as in resistant cells versus untreated cells (p < 0.0001 for A549 and A549R cells; p = 0.002 for parental HCC827 and p = 0.0007 for resistant HCC827 cells, Student's t-test) (Fig. 5B).
The YAP activity was then further measured by luciferase assay on three other cell types (Fig. 5C). PC-9 and H2052 lines were reported to display high YAP activity (20, 21). In both cell types, a significant decrease in the YAP activity was observed in the presence of relatively low IV#6 doses (for H2052 cell line: p = 0.0236, IV#6 3µM, n = 5, Student's t-test and for PC9 cell line: p = 0.0156; IV#6 3µM, n = 7, Wilcoxon test). PC-9 cell line, with oncogenic EGFR mutation, was recently reported to exhibit YAP/TEAD signaling activation in line with a TP53 gain of function mutation, which results in mevalonate pathway activation, inducing cytoskeleton rearrangements via Rho activation, and ultimately YAP nuclear translocation (21). H2052 exhibits two mutations in genes of the Hippo pathway, leading to YAP nuclear accumulation and activation (20). This YAP pharmacological inhibition was not found in a third EGFR-mutated cell model, the HCC4006 cell line (p = 0.1250; IV#6 1.5µM, n = 4, Wilcoxon test) (Fig. 5C).
YAP/TEAD transcriptional signature in parental and paclitaxel-resistant cells
Differential transcriptomic analysis between parental and resistant A549 cells revealed resistant cells to display an overexpression of YAP/TEAD target mRNAs as compared to the parental cell line (analysis using Metascape and EnrichR) (Fig. 6A). Similar pathways were highly represented in both cell lines, including cell cycle, MYC target, oxidative phosphorylation, and RNA metabolism. We assessed the expression of YAP/TEAD target genes by constructing a 42-genes YAP/TEAD signature (Fig. 6B left), based on our data mining from published literature, only selecting YAP/TEAD target genes by CHIPseq defined by at least two different teams, and two different cell types (7, 37, 38). Such list comprised well-known targets, some of which being Hippo pathway members (AJUBA, FAT1, FRMD6, TEAD1/4, STK3/MST2, STK4/MST1 (Hippo), SCRIB, LATS2, WWTR1/TAZ, WWC1/Kibra), while others were classical targets responsible the YAP/TEAD’s signaling pleïotropic effects, including AXL, AREG, CCND1, BIRC5/survivin. We compared mean expression and mean z-score of individual genes as well as the whole signature in our two cell lines A549 and A549R, either treated or not, with the YAP/TEAD inhibitor (Fig. 6B right). GSEA analysis, based on the Broad Hallmark geneSet, confirmed the significant down-regulation of these YAP/TEAD target genes in resistant A549 cells, after treatment with the compound (p = 0.05 mean expression; p = 0.037 Z-score). Despite an objective trend towards reduced YAP/TEAD targets expression in parental cells, following treatment with the compound, the results did not reach statistical significance (p = 0,7 for mean expression; p = 0,5 for Z-score). This suggested that pharmacological YAP/TEAD inhibition displayed specifically dramatic effects only on cells with basal high-level YAP activation, including the paclitaxel-resistant A549. Individual classical transcriptional YAP target genes, namely CTGF and CYR61, were also up-regulated in resistant cells as compared with parental cells, with a decreased expression upon IV#6 in both resistant and parental A549 cells (p-value < 0.05) (Fig. 6C).
IV#6 compound partly restores chemo-sensitivity to Paclitaxel
Figure 7A displays the IC50 curves of paclitaxel from the four independent experiments combined, illustrating the pharmacological inhibitor effect in restoring chemotherapy sensitivity in resistant cells (purple curve). Calculation of the areas under the curve (Fig. 7B) confirmed that adding IV#6 (10µM) significantly reduced the area under the curve in resistant A549 cells (p = 0.02; n = 4; Mann Whitney test), thereby increasing the sensitivity of these cells cultured in 2D, to paclitaxel. This difference was not observed in parental A549 cells, which remained sensitive to paclitaxel (Fig. 7C).
To further demonstrate the Inventiva™-supplied anti-YAP compound ability to restore sensitivity to the cytotoxic agent paclitaxel, we moved to a "tumor on chip" 3D micro-fluidic culture experimental system (Fig. 8A). We started with the A549 lung cancer cell line model, which revealed some clear YAP signaling activation, shown in our transcriptomic data. We therefore incubated A549 cells embedded in collagen with a medium containing different drugs at different concentrations, as well as a green apoptosis reporter (Cell Event Caspase 3/7). The microfluidic chips were imaged under the microscope for 72h. We used STAMP software to quantify apoptosis events over time (24). Parental and resistant A549 cells were treated either with 100 nM paclitaxel alone, IV#6 compound at 20 µM, or 100nM paclitaxel and 20µM IV#6 combined. DMSO and 20 nM paclitaxel constituted the control condition for A549 cells and paclitaxel-resistant A549R cells respectively, since the latter are routinely grown in presence of low-dose 20 nM paclitaxel.
In parental cells, combining IV#6 and paclitaxel, resulted in a slightly higher effect on cell death than the control (p = 0.04, 2-way Anova, n = 3) (Fig. 8B left). In resistant A549 cells, combining 20µM IV#6 and 100nM paclitaxel, significantly and substantially increased cell death, versus the control conditions (p = 0.02, 2way Anova, n = 3) and the 100nM paclitaxel single-therapy (p = 0.04, 2-way Anova, n = 3) (Fig. 8B right). This result indicated that inhibition of YAP by the IV#6compound under 3D chip culture conditions, at least partly restored paclitaxel chemo-sensitivity in resistant cells, supporting the YAP-TEAD pathway's role in paclitaxel resistance of A549 lung cancer cells.
We additionally studied the compound effect in combination with paclitaxel on cancer cells from fresh tumors, from lung cancer patients operated on at Bichat Hospital, providing herein the example of a 70-year-old smoker, who underwent right upper lobectomy for a 10% PD-L1-expressing lung adenocarcinoma, without molecular addictive driver alteration. For this experiment, we used DRAQ7 as a fluorescence general death-induced marker to monitor all-cause cell death. To this end we manually counted dying cells as a function of time. Figure 8C illustrates the cell death rate per 10 hours (left panel), and the cumulative cell death rate (right panel), over 48 hours. The anti-YAP/TEAD compound alone was not toxic at 5 and 10µM in this experiment (red and magenta curves), while adding paclitaxel alone (200 nM, green curve) led to cell death, as expected. Combining paclitaxel and 10µM IV#6 (but not 5µM), thus likely induced a higher effect with a significantly greater increase in cell death versus paclitaxel alone. Such preliminary data actually indicated that patient-derived tumor-on chips consisted of an efficient strategy to investigate chemotherapy combinations with YAP inhibitors, and to address the complexity of patient response heterogeneity.