Research Article

Spatial modeling reveals nuclear phosphorylation and subcellular shuttling of YAP upon drug-induced liver injury

Institute of Pathology, University Hospital Heidelberg, Germany
Department of Modeling of Biological Processes, COS Heidelberg/BioQuant, Heidelberg University, Germany
Institute of Biomedicine (IBIOMED), University of León, Spain
Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Spain
Leibniz Research Centre for Working Environment and Human Factors, Department of Toxicology, Technical University Dortmund, Germany
Department of Forensic Medicine and Toxicology, Faculty of Veterinary Medicine, South Valley University, Egypt

Oct 18, 2022

https://doi.org/10.7554/eLife.78540

Version of Record

Accepted for publication after peer review and revision.

About eLife's process

Download
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Version of Record published: October 18, 2022 (This version)
Accepted: September 21, 2022
Preprint posted: April 1, 2022 (Go to version)
Received: March 10, 2022

1. Of interest
Dynamics of co-substrate pools can constrain and regulate metabolic fluxes

Robert West, Hadrien Delattre ... Orkun S Soyer

Research Article Updated Mar 20, 2023
Further reading

Article
Figures and data
Abstract
Editor's evaluation
Introduction
Results
Discussion
Materials and methods
Appendix 1
Data availability
References
Decision letter
Author response
Article and author information
Metrics

Abstract

The Hippo signaling pathway controls cell proliferation and tissue regeneration via its transcriptional effectors yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ). The canonical pathway topology is characterized by sequential phosphorylation of kinases in the cytoplasm that defines the subcellular localization of YAP and TAZ. However, the molecular mechanisms controlling the nuclear/cytoplasmic shuttling dynamics of both factors under physiological and tissue-damaging conditions are poorly understood. By implementing experimental in vitro data, partial differential equation modeling, as well as automated image analysis, we demonstrate that nuclear phosphorylation contributes to differences between YAP and TAZ localization in the nucleus and cytoplasm. Treatment of hepatocyte-derived cells with hepatotoxic acetaminophen (APAP) induces a biphasic protein phosphorylation eventually leading to nuclear protein enrichment of YAP but not TAZ. APAP-dependent regulation of nuclear/cytoplasmic YAP shuttling is not an unspecific cellular response but relies on the sequential induction of reactive oxygen species (ROS), RAC-alpha serine/threonine-protein kinase (AKT, synonym: protein kinase B), as well as elevated nuclear interaction between YAP and AKT. Mouse experiments confirm this sequence of events illustrated by the expression of ROS-, AKT-, and YAP-specific gene signatures upon APAP administration. In summary, our data illustrate the importance of nuclear processes in the regulation of Hippo pathway activity. YAP and TAZ exhibit different shuttling dynamics, which explains distinct cellular responses of both factors under physiological and tissue-damaging conditions.

Editor's evaluation

The Hippo signaling pathway is essential for multiple physiological processes, most notably the regulation of cell proliferation and survival during wound healing. Wehling et al. provide a molecular framework for an alternative mechanism by which the Hippo effector molecule YAP's sub-cellular localization is regulated by cell compartment-specific phosphorylation. Specifically, the authors demonstrate dynamic regulation of shuttling of YAP both in vitro and in vivo during drug-induced liver injury. Given the importance and developmental conserveness of the Hippo pathway, the work is of broad interest to the field of developmental and regenerative biology.

https://doi.org/10.7554/eLife.78540.sa0

Introduction

The evolutionary conserved Hippo signaling pathway controls tissue homeostasis through the regulation of cell proliferation, apoptosis, differentiation, and cellular fate (Zhao et al., 2011). Activity of this pathway is affected by information derived from extracellular matrix stiffness, actomyosin dynamics, and cell density (Deng et al., 2015; Zhao et al., 2007), which in turn modulate a serine/threonine kinase cassette consisting of serine/threonine kinase 3/4 (STK3/4; MST1/2) and large tumor suppressor 1/2 (LATS1/2). According to the canonical Hippo pathway model, active cytoplasmic LATS1/2 phosphorylate and inactivate two important downstream pathway effectors: the transcriptional co-activators yes-associated protein (YAP) and its paralog transcriptional coactivator with PDZ-binding motif (TAZ). This phosphorylation is associated with nuclear YAP/TAZ exclusion followed by their proteasomal degradation. In contrast, Hippo pathway inactivation in the cytoplasmic compartment causes YAP/TAZ dephosphorylation and nuclear translocation. In the nucleus, YAP/TAZ bind DNA sequence-specific transcription factors such as TEA DNA-binding proteins (TEADs) or forkhead box protein M1 (FOXM1), which control the transcription of genes involved in for example, cell cycle control and paracrine communication (Marti et al., 2015; Thomann et al., 2020; Weiler et al., 2017).

Due to its pivotal role in the regulation of cell proliferation, the Hippo/YAP/TAZ axis is important for tissues maintenance during regenerative processes. Indeed, YAP- or YAP/TAZ deficiency reduce the regenerative capacity of tissues such as liver, skin, and heart (Lee et al., 2014; Lu et al., 2018; Xin et al., 2013). As exemplified in detail for the liver, YAP is activated in hepatocytes under disease conditions that support a continuous regenerative response in different in vivo model systems (Machado et al., 2015; Mooring et al., 2020). Interestingly, the roles of YAP and TAZ are not identical, since TAZ, but not YAP, contributes to fat accumulation in the liver (steatosis) (Wang et al., 2016). In contrast, YAP protects from liver ischemia–reperfusion injury by promoting tissue repair (Liu et al., 2019). These findings strongly suggest a cell-protective role of the Hippo pathway under liver injury conditions; however, the function of YAP and TAZ in this biological process may differ. As YAP and TAZ control the expression of similar target genes in different cell types (Weiler et al., 2020; Zanconato et al., 2015), other regulatory mechanisms must account for these phenotypic differences such as differential nuclear/cytoplasmic shuttling (Reggiani et al., 2021). However, detailed comparative studies regarding subcellular localization dynamics for YAP and TAZ are missing as well as spatially resolved computational framework such as partial differential equation (PDE) models.

Recent data demonstrated a direct impact on YAP by acetaminophen (APAP), which is a widely used analgesic drug and leading cause of acute liver failure (Poudel et al., 2021). Here, APAP overdose in mice caused liver injury, which was associated with a prominent nuclear YAP enrichment in hepatocytes. Interestingly, silencing of YAP did not impair liver regeneration (which would be expected after inactivation of this proproliferative factor), but promoted tissue regeneration upon APAP-induced liver damage (Poudel et al., 2021). In contrast, a supportive role of YAP in tissue repair and regeneration was suggested by chemical blockade of upstream Hippo pathway kinases MST1/2, which reverted APAP-induced liver damage (Fan et al., 2016). These results point to intricate and context-specific roles of YAP and/or TAZ under regenerative conditions.

Our study comprises a quantitative and mechanistic analysis of YAP/TAZ dynamics based on spatially resolved, high-throughput confocal live cell imaging data and PDE modeling. By using this experimental and computational toolbox, we show that nuclear phosphorylation of YAP/TAZ is important for their subcellular shuttling dynamics and that both factors do not equally respond to cell density and APAP-induced hepatocellular damage. We mechanistically show that APAP overdose stimulates a rapid reactive oxygen species (ROS) response, which facilitates the physical nuclear interaction of YAP and AKT as predicted by computational modeling. Thus, our data demonstrate for the first time that APAP-induced YAP activation is a dynamic and specific process that relies on the sequential activation of molecular events.

Results

Establishment of an in vitro model for measuring time-resolved spatial localization of YAP and TAZ in hepatocellular cells

Due to technical limitations, primary human hepatocytes are not suitable for the analysis of dynamic YAP/TAZ shutting in vitro: they rapidly undergo trans-differentiation in culture and different human donors exhibit genomic/epigenetic variability, which diminish reproducibility. We therefore selected the hepatocyte-derived liver cancer cell line Hep3B that was characterized by a prominent YAP/TAZ and cytochrome P450 2E1 (CYP2E1) expression (Figure 1—figure supplement 1A; Weiler et al., 2020). Detectable CYP2E1 levels are required for the enzymatic transformation of APAP to toxic N-acetyl-p-benzoquinone imine (NAPQI) (Lee et al., 1996).

Since we were interested in cell-to-cell variability and spatial localization information of YAP and TAZ, we established a cell line that allowed the quantitative and spatiotemporal investigation of subcellular YAP/TAZ distribution near the single-cell resolution. For this, we generated Hep3B cells that stably express mVenus-tagged YAP (mVenus-YAP) and mCherry-tagged TAZ (mCherry-TAZ) (Figure 1A). In addition, mCerulean-tagged histone H2B (mCerulean-H2B) allowed the spatial allocation of YAP/TAZ. These chosen fluorophores facilitated optimal discrimination between emission spectra by live cell microscopy (Figure 1—figure supplement 1B). Subcellular localization changes of tagged YAP and TAZ under variable cell density conditions illustrated functionality of both proteins regarding nuclear–cytoplasmic shuttling (Figure 1B). However, we observed a high degree of intracellular variability.

Figure 1 with 1 supplement see all

Download asset Open asset

YAP and TAZ show distinct nuclear shuttling in hepatocellular cells upon increasing cell density.

(A) The hepatocyte-derived cell line Hep3B was transduced using three lentiviral vectors coding for mCerulean-tagged H2B (pLentiPGK-mCerulean-H2B), mVenus-tagged YAP (pRRLN-EF1α-mVenus-YAP), and mCherry-tagged TAZ (pRRLN-EF1α-d2mCherry-TAZ). Combined treatment with the antibiotics hygromycin, geneticin, and blasticidin selected for triple-positive cells. Cells were confocally imaged in three channels: 445 nm (CFP, mCerulean), 514 nm (YFP, mVenus), and 561 nm (RFP, mCherry). Scale bar: 250 µm. (B) Exemplary pictures illustrating the subcellular localization of H2B, YAP, and TAZ proteins under low (left) and high (right) cell density conditions. Scale bar: 50 µm. (C) Automatic image analysis workflow depicts analysis of nuclear (left) and cytoplasmic (right) fluorescence intensity in confocal images of living cells. The acquired images were prescreened for imaging artifacts (e.g., out-of-focus images were excluded) and subjected to the image analysis pipeline in Fiji. Object classification (Weka algorithm), thresholding, and object counting were performed. Object masks were overlaid with original data and the nuclear/cytoplasmic ratio (NCR) was calculated by dividing average fluorescent signal intensity of YAP/TAZ in the nucleus with the fluorescence intensity in the cytoplasm. Scale bar: 50 µm. (D) Quantification of YAP and TAZ NCR under increasing cell density conditions. One of four representative experiment is depicted. Left: the NCR of YAP (yellow, n = 310) and TAZ (red, n = 310) was plotted against cell count per visual field (0.4 mm²). A high NCR value indicates predominant nuclear localization. Dashed line shows mean cell count. Right: violin plots summarize shift in NCR between low (below mean cell count) and high (above mean cell count) cell density for YAP and TAZ.

For analysis of variable high-throughput data, we developed an algorithm for the semiquantitative measurement of tagged YAP and TAZ in nuclei and cytoplasm. In brief, spatial image data were classified based on the expression of mVenus-YAP and mCherry-TAZ using the Weka segmentation algorithm (Figure 1C, Figure 1—figure supplement 1C; Arganda-Carreras et al., 2017). The fluorophore intensity measurements were used to calculate the nuclear/cytoplasmic ratio (NCR), which characterized the relative nuclear enrichment of YAP or TAZ. Therefore, this approach allowed us to simultaneously define slightest subcellular changes of YAP and TAZ in small cell populations (field of view, FOV) as well as individual cells under live cell conditions.

Subsequently, the fluorophore-expressing cells were grown under variable cell density conditions and the YAP/TAZ localization was measured by confocal microscopy. The results confirmed that tagged YAP and TAZ dynamically shuttled between cytoplasm and the nucleus depending on abundance of cell–cell contacts (cell density) (Figure 1D). However, TAZ showed a considerably less pronounced dynamic behavior compared to YAP (Figure 1D). YAP was predominantly detectable in cell nuclei under low cell density conditions (NCR > 1), while increasing cell density led to cytoplasmic enrichment of both factors (NCR < 1). Using this algorithm, cell density-dependent protein shuttling was also demonstrated for other liver cancer cells (Tóth et al., 2021).

Together, the confocal imaging and image processing pipeline efficiently detects subtle differences in the dynamic subcellular changes of YAP and TAZ.

Mathematical modeling predicts nuclear phosphorylation of YAP and TAZ

To investigate why Hippo pathway effectors YAP and TAZ differently respond to a low cell density and which Hippo pathway topology can explain the observed localization differences, we mathematically modeled the Hippo pathway using a PDE modeling framework (for detailed description of the PDE modeling approach refer to Materials and methods and Appendix 1).

First, we investigated a canonical Hippo pathway model, where phosphorylation takes place in the cytoplasm and where unphosphorylated YAP/TAZ shuttle to the nucleus (Figure 2A). However, this mathematical model was not able to reproduce the localization patterns which were observed in the experimental data (Figure 2B). For example, the model cannot sufficiently explain the distribution of YAP and TAZ in cell nuclei or at the nuclear membrane. In detail, the cell area around the nuclear envelope on the cytoplasmic side strongly underrepresented the concentration of YAP/TAZ, as indicated by the blue pixels in the residual image (experimental data minus model simulation). Moreover, the simulated YAP/TAZ distribution pattern of the canonical model showed a decreased protein concentration in the center of the nucleus, with increasing gradient toward the nuclear envelope. This indicates that the experimental observation is dominated by a protein that is disseminated from the nucleus and undergoes diffusion and degradation. The canonical Hippo pathway cannot explain this effect, illustrated by residuals (Figure 2B).

Figure 2 with 2 supplements see all

Download asset Open asset

Mathematical modeling predicts that nuclear phosphorylation controls YAP/TAZ subcellular localization.

(A) The model reaction scheme of the canonical Hippo pathway. Only unphosphorylated YAP/TAZ can enter the nucleus, whereas pYAP and pTAZ are exclusively localized to the cytosol. The phosphorylation of YAP/TAZ takes place exclusively outside nuclei. (B) Partial differential equation (PDE) model simulation of the canonical Hippo pathway compared to the experimentally measured YAP/TAZ localization. The residuals (experimental data minus model simulation) indicate low spatial accordance of the model simulation with the experimentally measured subcellular localization of YAP/TAZ. Images were normalized to the maximal value within each image. Nuclear outline is indicated in black. (C) The model reaction scheme of the alternative Hippo pathway model. Phosphorylation and dephosphorylation of YAP/TAZ take place in the nucleus. Unphosphorylated YAP/TAZ is imported in the nucleus and phosphorylated YAP/TAZ is rapidly exported to the cytoplasm. (D) PDE simulation of the alternative model compared to experimental data. Residuals for the subcellular localization (e.g., nuclear distribution) sufficiently reflect results from confocal microscopy. Nuclear outline is indicated in black. (E) Simulated impact of the nuclear phosphorylation to dephosphorylation ratio on subcellular localization of YAP in the alternative Hippo pathway model. Left: model simulation of two phosphorylation/dephosphorylation rates (0.17 and 0.75). Right: summarized residuals with respect to experimental data of YAP and TAZ as a function of phosphorylation to dephosphorylation ratio in the nucleus. (F) Western immunoblot after nuclear and cytoplasmic protein fractionation under low cell density conditions. The central Hippo pathway kinases LATS1/2 were detectable in the nuclear fraction. As expected, high pYAP levels are detectable in the cytoplasm, illustrating that the protein is transported outside the nucleus upon phosphorylation. PARP and Tubulin serve as loading controls for nuclear and cytoplasmic fractions, respectively. Equal amounts of cytoplasmic and nuclear proteins were loaded (n = 4; one representative experiment shown). (G) Proximity ligation assay (PLA) for phosphorylated LATS1/2 (pLATS1/2) and YAP (top row). Red dots indicate physical interaction between pLATS1/2 and YAP. DAPI (4′,6-diamidino-2-phenylindole)-stained nuclei are indicated in cyan. Individual pLATS and YAP antibodies serve as assay controls. Bottom row: computational segmentation of nuclei (empty circles) and dots (black spots). Scale bar: 50 µm. (H) Quantification of the PLA assay. Each dot represents an individual image. Left: quantification of interactions between YAP and pLATS1/2 in the nucleus (n = 18) compared to the negative controls (pLATS1/2 and YAP antibodies alone, n = 16 and n = 17, respectively). Right: the number of PLA dots of pLATS1/2 and YAP interaction in cytoplasm and in nuclei (n = 18). Statistical test: two-tailed paired t-test (p value = 0.01) **p ≤ 0.01.

Figure 2—source data 1 Raw data for Western Blot shown in Figure 2F.: https://cdn.elifesciences.org/articles/78540/elife-78540-fig2-data1-v1.zip
Download elife-78540-fig2-data1-v1.zip

This let us conclude that alternative model topologies must be considered to explain the experimental data. We therefore generated an alternative PDE model, which precisely described YAP/TAZ localization as observed in the image data. In this model, a phosphorylation and dephosphorylation reaction of YAP and TAZ in the nucleus was included (Figure 2C,D, Figure 2—figure supplement 1A, B). Moreover, the alternative model describes that unphosphorylated YAP/TAZ are transported to the nucleus and pYAP/pTAZ are rapidly excluded from the nucleus, which agrees with the general consensus that phosphorylated YAP and TAZ predominantly localize in the cytoplasm (Figure 2—figure supplement 1C). Using well-fitting parametrization of the model, the simulated concentration pattern was dominated by phosphorylated YAP/TAZ disseminating from the nucleus, which now matched the observed experimental data well (Figure 2D).

Since YAP and TAZ are considered biochemically similar molecules, the question arises how the pronounced observed differences in the spatial distribution patterns can be explained. Using the alternative Hippo pathway model, we demonstrated that the different phenotypes could be reproduced with one single PDE model where only the parameters of the YAP/TAZ phosphorylation and dephosphorylation reactions in the nucleus differed, while all other parameters (including nuclear import and export reactions) were kept the same (Figure 2E, Figure 2—figure supplement 1D).

In detail, low nuclear phosphorylation/dephosphorylation ratio (0.17) induced nuclear localization of YAP, while high ratio (0.75) induced nuclear protein exclusion. Upon increased phosphorylation activity (>0.2), the nuclear fraction of TAZ decreased and the summarized residual term for TAZ was reduced. According to residual terms, best accordance between the model and the experimental data was achieved at low phosphorylation/dephosphorylation ratio for YAP and at higher phosphorylation/dephosphorylation ratio for TAZ (Figure 2E, right panel). These findings suggest that differences in the nuclear phosphorylation rates of YAP and TAZ can explain the distinct shuttling patterns of both factors (as observed in Figure 1B). In addition, TAZ may require stronger phosphorylation than YAP to induce its efficient shuttling.

The results of the alternative mathematical model required that major components for YAP/TAZ phosphorylation process were present in the cell nucleus. Indeed, subcellular fractionation experiments illustrated that the Hippo pathway kinases LATS1/2 (total and phosphorylated) were predominantly localized in nuclear protein fraction (Figure 2F). Moreover, we and others reported that nuclear LATS1/2 can control YAP phosphorylation and localization (data not shown) (Ege et al., 2018; Li et al., 2014). The relevance of this nuclear interaction is substantiated by a proximity ligation assay (PLA), which illustrated that pLATS1/2 physically interacted with YAP not only in the cytoplasm but also in the nucleus (Figure 2G). To quantitatively compare the number of YAP/pLATS interactions in cell nuclei and cytoplasm, we utilized a computational algorithm counting signals in nuclear and cytoplasmic area of the cells (Figure 2—figure supplement 2A, B). The signal quantification confirmed that YAP and pLATS interactions were detectable in both compartments with significantly higher interaction count in cell nuclei (Figure 2H).

In summary, our findings suggest that nuclear YAP/TAZ phosphorylation contributes to their inactivation and nuclear exclusion. Moreover, variable phosphorylation rates of YAP and TAZ explain the subcellular shuttling differences between both factors.

APAP regulates YAP protein localization and activity

To test if our findings are of relevance under conditions where YAP and TAZ are actively regulated, we decided to use a drug-induced liver injury (DILI) model. For this, we treated Hep3B cells expressing tagged mVenus-YAP and mCherry-TAZ with APAP (10 nM) (Barbier-Torres et al., 2017), and quantitatively investigated the dynamic shuttling of both factors.

Image quantification revealed that APAP led to a gradual and time-dependent nuclear enrichment of YAP after 48 hr, while TAZ only weakly responded to APAP (Figure 3A,B). This APAP-induced nuclear shuttling effect was clearly detectable under high cell density culture conditions, which is characterized by nuclear YAP/TAZ exclusion. Effects on YAP were less pronounced 24 hr after APAP administration (Figure 3—figure supplement 1A,B). No obvious response was detectable for YAP and TAZ at earlier time points post APAP treatment (data not shown). Severe effects of APAP on cell toxicity and apoptosis in the chosen experimental setup were excluded by measuring cell viability and PARP cleavage (data not shown).

Figure 3 with 2 supplements see all

Download asset Open asset

APAP regulates YAP phosphorylation, localization, and expression of its target genes.

(A) Live cell confocal imaging of H2B-mCerulean, YAP-mVenus, and TAZ-mCherry in Hep3B cells. Upper row: control treatment (phosphate-buffered saline, PBS). Lower row: APAP treatment (10 mM) induces nuclear enrichment of YAP, but not TAZ protein after APAP treatment within 48 hr (n = 4; one representative experiment shown). Scale bar: 50 µm. (B) Live cell confocal microscopy image quantification. Nuclear/cytoplasmic ratio (NCR) of YAP and TAZ with APAP (10 mM for 48 hr; yellow and red circles, n = 180 per channel) and control (PBS; gray circles, n = 180 per channel) treatment under increasing cell density conditions (represented as mean cell count per visual field). Each dot represents a single visual field. Dashed line: mean cell density. (C) Western immunoblot of YAP and pYAP after APAP treatment (10 mM) in Hep3B cells (n = 6; one representative experiment shown). GAPDH served as a loading control. (D) Relative expression of the YAP target genes *ANKRD1* and *CYR61* 24 hr after APAP (10 mM) treatment in HLF cells (n = 2; one out of two biological replicates shown). (E) Rescue experiment: relative expression of YAP and its target genes *CYR61* and *ANKRD1* with and without siRNA-mediated YAP inhibition (for 24 hr) with or without APAP (10 mM) treatment for 24 hr (n = 2; one out of two biological replicates shown).

Figure 3—source data 1 Source data for Western Blot shown in Figure 3C.: https://cdn.elifesciences.org/articles/78540/elife-78540-fig3-data1-v1.zip
Download elife-78540-fig3-data1-v1.zip

The APAP-induced nuclear enrichment of YAP was confirmed with cell population-based Western immunoblotting followed by detection of total and phosphorylated YAP (Figure 3C, for quantification of all western blots see Figure 3—figure supplement 2). Notably, at early time points YAP hyperphosphorylation indicated protein inactivation (up to 3 hr after APAP administration). However, between 24 and 48 hr after APAP treatment, a clear YAP dephosphorylation was detectable (Figure 3C). Similar results were observed for another hepatocyte-derived cell line (Figure 3—figure supplement 1C).

To test if APAP-dependent dephosphorylation/activation of YAP at later time points is leading to its transcriptional activation, we measured the relative expression of known YAP target genes cysteine-rich angiogenic inducer 61 (CYR61) and ankyrin repeat domain 1 (ANKRD1) by real-time PCR. Indeed, we observed induction of these genes already 24 hr after APAP treatment, which reflected the earliest activation of YAP (Figure 3D). To confirm YAP dependence on the target gene induction and to exclude YAP-independent effects on gene expression, we performed additional rescue experiments. For this, the expression of YAP was silenced by RNA interference (RNAi) in APAP-treated cells followed by the measurement of YAP target genes. According to our hypothesis, YAP inhibition partly abolished the APAP-dependent induction of CYR61 and ANKRD1 (Figure 3E).

Together, these data demonstrate that APAP predominantly acts on the Hippo pathway effector YAP in a bimodal manner. Upon APAP treatment an immediate phosphorylation/inactivation of YAP is followed by its late dephosphorylation/activation.

APAP controls YAP phosphorylation via ROS and AKT

Based on previous data, we hypothesized a mechanistic connection between APAP-induced ROS (Barbier-Torres et al., 2017; Shuhendler et al., 2014) and AKT-driven YAP phosphorylation (Basu et al., 2003; Romano et al., 2010). This mechanistic link was investigated at early time points after APAP treatment (up to 3 hr) to exclude unspecific effects caused by APAP at later time points (e.g., due to secondary and/or unspecific APAP effects).

Western immunoblotting showed that both YAP and AKT were phosphorylated early after APAP incubation (Figure 4A), which was in accordance with our previous findings (Figure 3C). In parallel, an ROS detection assay revealed that APAP-induced prominent ROS activity comparable to the known ROS inducers hydrogen peroxide (H₂O₂) and tert-butyl hydroperoxide (TBHP) in hepatocellular cells (Figure 4B). Moreover, treatment of cells with ROS inducers demonstrated a clear (with H₂O₂) or moderate (with TBHP) induction of AKT and YAP phosphorylation after 1 and 2 hr, respectively (Figure 4C,D, quantification of western blots in Figure 3—figure supplement 2). These data indicated that APAP regulated ROS activity, which itself controlled AKT and YAP activity.

Figure 4

Download asset Open asset

APAP controls nuclear YAP enrichment via induction of reactive oxygen species (ROS) and AKT.

(A) Western immunoblot analysis of pYAP, YAP, pAKT, and AKT after APAP (10 mM) administration in Hep3B cells after 1 hr (Hep3B). The concentration of pYAP and pAKT is induced after APAP treatment in comparison to phosphate-buffered saline (PBS)-treated cells (ctrl); n = 2. (B) Spectrophotometric ROS measurement after APAP (10 mM), hydrogen peroxide (H₂O₂, 2 mM), and tert-butyl hydroperoxide (TBHP, 300 µM) treatment in Hep3B cells after 6 hr (n = 8 technical replicates, one out of two biological replicates is shown). APAP as well as H₂O₂ and TBHP induce ROS formation in living cells. Statistical test: one-way analysis of variance (ANOVA) with Geisser–Greenhouse correction (adjusted p value = 0.001) ***p-value ≤ 0.001. Whiskers depict min and max of the dataset. (C) Western immunoblot analysis of pYAP, YAP, pAKT, and AKT after H₂O₂ (2 mM) and APAP (10 mM) treatment for 1 hr. APAP, as well as H₂O₂ induce YAP and AKT phosphorylation compared to untreated cells (ctrl); n = 4. (D) Western immunoblot analysis of pYAP, YAP, pAKT, and AKT dynamics after TBHP (300 µM) and APAP (10 mM) treatment for 2 hr (Hep3B). APAP and TBHP induce AKT and YAP phosphorylation as compared to respective controls (ctrl). (E) Western immunoblot analysis of pAKT, AKT, pYAP, and YAP after hepatocyte growth factor (HGF) treatment (10 ng/µl) for 10 min (Hep3B). The cells were pretreated with an AKT inhibitor VIII (AKTi, 1, 5, and 10 µM) and starved for 3 hr prior HGF administration. Data illustrate that AKT inhibition prevents HGF-induced AKT and YAP phosphorylation. (F) Western immunoblot of pAKT, AKT, pYAP, and YAP after APAP (10 mM) administration for 1 hr (Hep3B). Cells were starved in Fetal calf serum (FCS)-free medium and pretreated with AKTi (5 µM) before APAP treatment for 3 hr. Results demonstrate that AKT inhibition reduces YAP phosphorylation early after APAP treatment. (G) Proximity ligation assay (PLA) of the protein combinations YAP/AKT and pYAP/pAKT. Red dots indicate interactions between YAP/AKT and pYAP/pAKT, respectively. DAPI-stained nuclei are depicted in cyan. Single YAP, AKT, pYAP, and pAKT stains serve as assay controls. Bottom row: segmentation of the nuclei (empty circles) and dots (black spots). One out of two representative experiment is shown. Scale bar: 50 µm. (H) Quantification of the interactions between YAP/AKT and pYAP/pAKT in the nucleus (experiment shown in G). One symbol represents one image (YAP = 17, AKT = 16, YAP-AKT = 22, pYAP = 24, pAKT = 25, pYAP-pAKT = 23). (I) PLA showing YAP–AKT interaction units per nuclei of untreated (ctrl, n = 12) and APAP-treated cells (10 mM, for 2 hr, n = 20). APAP treatment induces nuclear interaction between YAP and AKT. Statistical test: unpaired two-tailed parametric t-test (p value = 0.0155). For A and C–F, actin served as loading control.

Figure 4—source data 1 Source data for Western Blots shown in Figure 4A,C-F.: https://cdn.elifesciences.org/articles/78540/elife-78540-fig4-data1-v1.zip
Download elife-78540-fig4-data1-v1.zip

To confirm the rapid and AKT-dependent phosphorylation of YAP, liver cells were treated with hepatocyte growth factor (HGF), which activates/phosphorylates AKT kinases via binding the receptor c-MET within 10 min (Xiao et al., 2001). HGF administration led to a clear phosphorylation of AKT but not YAP, which might be caused by saturation effects (no further YAP phosphorylation possible under given culture conditions) (Figure 4E). However, simultaneous inhibition of AKT kinase activity by the specific inhibitor AKTi (Adlung et al., 2017; Barnett et al., 2005), not only abolished AKT phosphorylation but also reduced YAP phosphorylation (Figure 4E). As expected, upon AKTi administration concomitantly with APAP, the phosphorylation of both AKT and YAP decreased (Figure 4F).

Since our alternative PDE model predicted that the nuclear YAP phosphorylation is crucial for protein shuttling and its activity (Figure 2C–E), we hypothesized that AKT and YAP not only physically interact in the cytoplasm but also in cell nuclei. Indeed, PLA illustrated that AKT and YAP, as well as the respective phosphorylated isoforms, interacted in the nucleus (Figure 4G,H). Although, interaction between AKT/YAP and pAKT/pYAP was also detectable in the cytoplasm, a prominent protein colocalization took place in the nuclear compartment (Figure 4G). Importantly, image-based quantification revealed that the interaction between nuclear YAP and AKT increased upon 6hr treatment with APAP (Figure 4I), indicating an APAP-dependent shift to nuclear phosphorylation and interaction, as predicted by the PDE model. Indeed, blocking AKT activity by AKTi reduced YAP phosphorylation, moderately elevated nuclear YAP positivity, and increased the physical interaction between AKT and YAP in cell nuclei (Figure 4F and data not shown).

In summary, our data show that APAP affects YAP activity and shuttling behavior by enhancing cellular ROS followed by nuclear AKT/YAP binding and YAP phosphorylation.

Sequential activation of ROS, AKT, and Hippo/YAP in mouse livers after APAP intoxication

The results of our mathematical model (Figure 2) and the in vitro experiments (Figure 3 and Figure 4) demonstrated a distinct sequence of molecular events that control YAP activity after APAP exposure: ROS induction is followed by AKT activation and YAP phosphorylation/inactivation (early events – up to 6 hr). This is followed by phase, where YAP is dephosphorylated and transcriptionally active (late events – 24 to 48 hr).

To confirm this in vivo, mice were injected with a hepatotoxic dose of APAP (300 mg/kg) and liver tissues were collected up to 16 days (in total at nine time points). Subsequently, liver specimens were subjected to expression profiling and results were investigated regarding the presence of gene signatures specific for ROS, AKT, and Hippo/YAP activity (De Marco et al., 2017; Han et al., 2008; Wang et al., 2018). The experimental results showed that gene expression of ROS, AKT, and Hippo/YAP target genes in livers was altered and prominently activated between 6 hr and 2 days after APAP treatment (Figure 5A–C). To compare the temporal dynamic of the gene signatures, z-scores of genes in the signature per time point were summarized and normalized to the number of genes in the signature. As indicated by the results from the cell culture experiments, the expression index illustrated a specific order of signature activation starting with ROS (6–12 hr), followed by AKT (12 hr to 1 day) and Hippo/YAP signature genes (1–2 days) (Figure 5D).

Figure 5 with 1 supplement see all

Download asset Open asset

APAP stimulates the sequential activation of reactive oxygen species (ROS), AKT, and YAP in vivo.

Gene expression profiling over time of mouse livers after APAP treatment (**A–C**) was performed for up to 16 days (3–5 animals/group, 300 mg/kg). Abundance of gene signatures characteristic for the activity of ROS consisting of 23 genes (Han et al., 2008) (A), AKT (B) with 28 genes (De Marco et al., 2017) and YAP (C) consisting of 23 genes (Wang et al., 2018) were analyzed. In A–C, gene expression values are z-score normalized. (D) Summarized gene expression scores (expression index) over time illustrate the timely order of ROS (max values between 6 and 12 hr), AKT (max values between 12 hr and 1 day) and YAP (max values between 1 and 2 days) target gene signatures after APAP treatment. (E) Mouse liver tissue sections stained for YAP (top tow) and total AKT (bottom row) under control condition and at 6 hr as well as 1, 2, 6, and 16 days after APAP treatment (300 mg/kg). Arrows indicate high YAP and AKT positivity in nuclei. Scale bar: 50 µm. (F) Automatic quantification of nuclei with YAP and AKT positivity in mouse liver tissues. Each dot represents the count of stained nuclei in one image (a tile, 1 mm²) for YAP (n = 733) and AKT (n = 520). (G) Quantification of YAP and AKT nuclear positivity 6 hr and 2 days after APAP treatment. One dot represents the average count of positive nuclei in one animal (n = 4).

To confirm the findings from the gene expression analysis, we performed immunohistochemical staining of liver tissues isolated at five time points after APAP treatment. Staining for total YAP illustrated not only a general increase of YAP positivity in the cytoplasm of surviving hepatocytes, but also its prominent nuclear accumulation after 1–2 days (Figure 5E, arrows). Moreover, nuclear AKT enrichment was already detectable 6 hr after APAP treatment. Interestingly, nuclear AKT accumulation in hepatocytes remained high for the rest of the experiment compared to untreated mice (Figure 5E, arrows).

Because of a high mouse-to-mouse variability, we decided to objectify these results. For this, the immunohistochemical staining of YAP and AKT was quantitatively analyzed using machine learning (random forest) and image analysis methods (Figure 5F). For this, 733 (for YAP) and 516 (for AKT) images were unbiasedly selected from all investigated animals and analyzed with an algorithm, which was trained to detect positively stained nuclei outside of the necrotic areas and tissue borders (Figure 5—figure supplement 1; for detailed information see Materials and methods). The quantification results confirmed the observed nuclear YAP positivity 1–2 days after APAP administration, while nuclear AKT was already detectable at 6 hr and maintained during the experiment (Figure 5F). The shift from exclusive nuclear AKT to nuclear YAP was illustrated by direct comparison of the time points 6 hr and 2 days after APAP injection (Figure 5G). The mechanistic connection between AKT activity and YAP induction in murine hepatocytes was confirmed in independent experiments. Here, hydrodynamic gene delivery of myristoylated AKT led to nuclear enrichment of YAP expression in hepatocytes and subsequent induction of typical YAP target genes (data not shown).

To sum up, we confirm the sequential order of APAP-induced events that cause YAP activation in mouse liver tissue.

Discussion

In this study, we aimed to investigate the different properties of the Hippo pathway effectors YAP and TAZ under in vitro and in vivo conditions that resembled the situation in DILI. For this, we developed a cell-based model system that allowed the systematic and comparative tracing of spatial YAP/TAZ dynamics upon stimulation with APAP, which is one of the most widely used analgesic with high potential of liver toxicity. By integrating time-resolved experimental data, computational modeling, image analysis tools as well as confirmatory in vivo results, we could draw several important conclusions on the role of YAP and TAZ under physiological conditions and in liver cells upon DILI.

Several mathematical modeling approaches have been previously applied to explain the dynamic behavior of the Hippo pathway under physiological or pathological conditions in different organisms. For example, computational modeling was used to explain how Hippo signaling cross-talks with other signaling pathways via protein interactions (Labibi et al., 2020; Romano et al., 2014). Other studies exclusively focused on biophysical/biochemical regulation of the Hippo pathway to investigate how YAP/TAZ activity relates to mechanical input information (Ege et al., 2018; Eroumé et al., 2021; Gou et al., 2018; Scott et al., 2021; Sun et al., 2016).

The goal of our approach was to determine whether nuclear phosphorylation of YAP plays a role in Hippo signaling, specifically in governing YAP localization. A model not including nuclear phosphorylation of YAP could not sufficiently explain the experimentally observed localization pattern and was excluded from further analyses. However, the alternative model proposed in this work, can explain the experimental data and therefore establishes a possible mechanism for YAP localization. However, more complex mechanisms regulating subcellular localization and dynamics of YAP and TAZ cannot be excluded. Therefore, our model partly explains how the Hippo pathway is organized, however, it does not aim to comprehensively describe it.

Here, we decided to use PDE modeling, which allowed us to discover spatial aspects of signaling dynamics and to use live cell image data for parametrizing the Hippo signaling steady-state models. By doing so, we compared PDE model simulations of different model topologies with the experimentally acquired data for YAP as well as TAZ and selected a model topology, which best corresponded to data generated in vitro. The selected alternative Hippo pathway model is based on experimental observations and parametrized within the considered boundaries of experimental evidence (see Appendix 1). The obtained model parameters are results of the best model fit obtained, and allow reproducing the described model behavior. However, since the model fitting is based on steady-state data only and concentration observations are in arbitrary units, the parameter values are subject to structural non-identifiability. Nevertheless, this does not affect the qualitative conclusions drawn from the models for the investigated biological process.

The results of our PDE modeling approach strongly suggested that the nuclear phosphorylation reaction of YAP/TAZ was necessary and sufficient for the reproduction of their subcellular distribution patterns in vitro (i.e., the halo-like nuclear distribution of YAP and TAZ). This observation is of importance since the conventional view on the Hippo pathway topology represents sequential phosphorylation steps in the cytoplasm. Our PDE modeling-based finding would therefore extend the current understanding of Hippo pathway regulation since it adds the cell nucleus as a pivotal spatial compartment in the modulation and adjustment of Hippo/YAP/TAZ signaling.

This conclusion is supported by our experimental findings illustrating that LATS1/2 are expressed in cell nuclei and that the interaction between phosphorylated LATS1/2 and YAP as well as between AKT and YAP is also detectable in this compartment. Importantly, previously published experimental and computational approaches support our findings. For example, nuclear LATS1/2 controls YAP phosphorylation in the context of NF2/Merlin deficiency (Li et al., 2014). In addition, the combination of quantitative photobleaching experiments and modeling illustrated that nuclear export is a crucial reaction for the subcellular YAP distribution (Ege et al., 2018). Combining these insights on nuclear export mechanisms with our findings on nuclear YAP phosphorylation strongly suggests that both processes closely cooperate in the efficient regulation of the Hippo pathway effectors. Lastly, our model does not explicitly exclude phosphorylation reaction in the cytoplasm; however, it suggests the necessity to reevaluate the canonical Hippo pathway scheme. The appeal for the reevaluation and extension of the canonical Hippo pathway, for example, by combining models on nuclear phosphorylation and nuclear transport, would further broaden our understanding on the dynamic spatial behavior of complex signaling pathways (Ege et al., 2018; Shreberk-Shaked and Oren, 2019).

The extension of the canonical Hippo pathway model with processes that control active nuclear YAP phosphorylation illustrates that the regulatory YAP/TAZ phosphorylation must be considered as continuum process (Shreberk-Shaked and Oren, 2019). For instance, the existence of nuclear phosphorylation possibilities by LATS-dependent and -independent (e.g., by AKT) mechanisms increases the complexity but also flexibility and redundancy of the signaling pathway (Ege et al., 2018; Gao et al., 2017; Low et al., 2014). These cellular aspects are part of a fine-tuned regulatory network that allows a rapid and adjustable proliferative response of YAP and TAZ under diverse physiological and pathological cellular conditions (e.g., upon induction of tissue regeneration).

To our knowledge, a direct mathematical comparison of YAP and TAZ regarding their spatial distribution has not been performed, yet. Using the PDE model, we comparatively investigated YAP and TAZ distribution and identified kinetic parameters, which discriminate between both species. Specifically, the observed localization differences (nuclear exclusion vs. nuclear localization) between YAP and TAZ might be caused by distinct phosphorylation rates of YAP and TAZ. Nuclear/cytoplasmic shuttling dynamics for YAP and/or TAZ have been investigated previously (Ege et al., 2018; Plouffe et al., 2018). However, our model predicts that the phosphorylation efficiency for TAZ must be higher than for YAP to achieve the observed differential localization of YAP and TAZ.

By utilizing our image analysis approach for the detection of minor changes in the subcellular distribution of YAP and TAZ, we showed a dynamic shuttling response of YAP (and to a much lesser extent of TAZ) upon APAP treatment. Although, APAP-dependent activation of YAP was described in the literature (Poudel et al., 2021), the underlying molecular mechanisms were poorly understood. Our results demonstrated that APAP-induced YAP shuttling is not the result of an unspecific cellular response but is due to a sequential order of molecular events that include ROS induction (Ghallab et al., 2016; Shuhendler et al., 2014), followed by AKT (in)activation (Koundouros NPoulogiannis, 2018) and nuclear YAP (de)phosphorylation. Time-dependent APAP-induced ROS activation was recently supported by a study by Ghallab et al., which showed transient induction of oxidative stress in mice 8 hr after APAP overdose with a peak at 2 hr (Ghallab et al., 2022). Interestingly, we observed a biphasic YAP regulation: YAP phosphorylation/inactivation at early time points (up to 3 hr) and its dephosphorylated/activation at later time points (>24 hr). Especially, the functionally important activation of YAP, which is associated with the induction of cell proliferation, is well reflected by the temporal activation of YAP in murine tissues upon APAP administration after 24–48 hr.

Importantly, several cellular mechanisms might simultaneously contribute to the APAP-dependent activation of YAP. For example, the direct impact of APAP on MST1/2 kinases, which are central Hippo pathway constituents, represents one additional process that may contribute to the observed effects on YAP (Fan et al., 2016). Moreover, the serine-/threonine-kinase AKT could control the Hippo/YAP axis at different levels. For example, a physical interaction of AKT and YAP in the nucleus (shown here) or the cytoplasm might directly control YAP phosphorylation (Basu et al., 2003). Alternatively, AKT can phosphorylate MST2 and therefore indirectly contribute to YAP activity (Romano et al., 2014). Thus, APAP probably controls YAP in a multimodal manner to achieve a cellular response and it is therefore likely that this multimodal process is part of cell-protective mechanism that counteracts the massive loss of hepatocytes upon APAP-induced DILI (Holland et al., 2022; Shuhendler et al., 2014).

In our manuscript, we show that APAP administration induces a chain of molecular events through the APAP/ROS/AKT axis, which is leading to YAP inactivation (early) followed by YAP activation (late). The late induction of YAP activity might be considered as cell-protective cellular response upon long-term tissue damage; however, YAP also acts as a potent oncogene in different cell types, including hepatocytes, and its overexpression is associated with tumor initiation (Dong et al., 2007; Weiler et al., 2017). Thus, it is tempting to speculate that APAP overdose may not only cause DILI but also increases the risk for liver tumor development. Although controversially discussed in the literature, a recent evaluation of 139 published epidemiologic studies strongly argues against a relationship between APAP exposure and cancer (Weinstein et al., 2021). Actually, APAP has been discussed as therapeutic option for patients with YAP-induced cancer (Poudel et al., 2021). Although, our and other studies clearly demonstrate a mechanistic link between APAP uptake and Hippo/YAP pathway activity, these partly controversial findings illustrate the necessity to decipher this connection under distinct disease conditions in the future.

Materials and methods

Key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Gene (Homo sapiens)	YAP	Prof. Loewer (TU Darmstadt)	ENSG00000137693	–
Gene (Homo sapiens)	TAZ	Cloned	ENSG00000018408	–
Cell line (Homo sapiens)	Hep3B	DSMZ	#ACC93	–
Cell line (Homo sapiens)	HepG2	LGC	ATCC-HB-8065	–
Cell line (Homo sapiens)	HLF	JCRB	JCRB0405	–
Cell line (Homo sapiens)	SNU182	ATCC	CRL-2235	–
Transfected construct	pRRLN-EF1α-mVenus-YAP	Prof. Loewer (TU Darmstadt)	End-to-end sequencing	–
Transfected construct	pRRLN-EF1α-d2mCherry-TAZ	This paper	End-to-end sequencing	–
Transfected construct	pLentiPGK-mCerulean-H2B	Addgene	90234	–
Chemical compound, drug	AKT inhibitor VIII	Merck Millipore	124018	–
Sequence-based reagent	siRNA YAP#1	This paper		CCACCAAGCUAG AUAAAGA-dT-dT
Sequence-based reagent	siRNA YAP#2	This paper		GGUCAGAGAUAC UUCUUAA-dT-dT
Sequence-based reagent	control siRNA	This paper		UGGUUUACAUG UCGACUAA
Antibody	YAP (rabbit monoclonal)	Cell Signaling Technology	RRID:AB_2650491	WB (1:1000)
Antibody	pYAP (rabbit polyclonal)	Cell Signaling Technology	RRID:AB_2218913	WB (1:400) PLA (1:200)
Antibody	pAKT (rabbit monoclonal)	Cell Signaling Technology	RRID:AB_2315049	WB (1:1000)
Antibody	AKT (rabbit polyclonal)	Cell Signaling Technology	RRID:AB_329827	WB (1:1000) PLA (1:200)
Antibody	β-Actin (mouse monoclonal)	MP Biomedicals, Solon	RRID:AB_2335127	WB (1:10,000)
Antibody	LATS1 (mouse monoclonal)	Santa Cruz Biotechnology	sc-398560	WB (1:200)
Antibody	LATS2 (mouse monoclonal)	Santa Cruz Biotechnology	sc-515579	WB (1:200)
Antibody	pLATS1/2 (rabbit monoclonal)	Cell Signaling Technology	RRID:AB_10971635	WB (1:500) PLA (1:200)
Antibody	CYP2E1 (rabbit unknown)	Novus Biologicals	RRID:AB_11021447	WB (1:500)
Antibody	PARP (rabbit polyclonal)	Cell Signaling Technology	RRID:AB_2160739	WB (1:10,000)
Antibody	β-Tubulin (mouse monoclonal)	Santa Cruz Biotechnology	RRID:AB_2288090	WB (1:200)
Antibody	GAPDH (chicken polyclonal)	Merck Millipore	RRID:AB_10615768	WB (1:10,000)
Antibody	YAP (mouse monoclonal)	Santa Cruz Biotechnology	RRID:AB_10612397	PLA (1:25)
Antibody	pAKT (mouse monoclonal)	Cell Signaling Technology	RRID:AB_331158	PLA (1:200)
Antibody	YAP (rabbit monoclonal)	Cell Signaling Technology	RRID:AB_2650491	IHC (1:50)
Antibody	AKT (rabbit polyclonal)	Cell Signaling Technology	RRID:AB_329827	IHC (1:50)
Commercial assay or kit	ROS assay kit	Abcam	ab287839	–
Sequence- based reagent	YAP (human) – forward	This paper	NM_006106	CCTGCGTAGCCAGTTACCAA
Sequence- based reagent	YAP (human) – reverse	This paper	NM_006106	CCATCTCATCCACACTGTTC
Sequence- based reagent	ANKRD1 (human) – for	This paper	NM_014391.3	AGTAGAGGAACTGGTCACTGG
Sequence- based reagent	ANKTD1 (human) – rev	This paper	NM_014391.3	TGGGCTAGAAGT GTCTTCAGA T
Sequence- based reagent	CYR61 (human) – for	This paper	NM_001554.5	AGCCTCGCATCCTATACAACC
Sequence- based reagent	CYR61 (human) – rev	This paper	NM_001554.5	TTCTTTCACAAGGCGGCACTC
Sequence- based reagent	GAPDH (human) – for	This paper	NM_002046.7	CTGGTAAAGTGGATATTGTTGCCAT
Sequence- based reagent	GAPDH (human) – rev	This paper	NM_002046.7	TGGAATCATATTGGAACATGTAAACC
Sequence- based reagent	RPL41 (human) – for	This paper	NM_001035267	AAACCTCTGCGCCATGAGAG
Sequence- based reagent	RPL41 (human) – rev	This paper	NM_001035267	AGCGTCTGGCATTCCATGTT
Sequence- based reagent	SRDF4 (human) – for	This paper	NM_005626	TGCAGCTGGCAAGACCTAAA
Sequence- based reagent	SRSF4 (human) – rev	This paper	NM_005626	TTTTTGCGTCCCTTGTGAGC
Sequence- based reagent	B2M (human) – for	This paper	NM_004048	CACGTCATCCAGCAGAGAAT
Sequence- based reagent	B2M (human) – rev	This paper	NM_004048	TGCTGCTTACATGTCTCGAT
Software, algorithm	ASAP software	https://github.com/computationalpathologygroup/ASAP	v1.6	–
Software, algorithm	ImageJ	Rueden et al., 2017	v1.53f51	–
Software, algorithm	Weka segmentation	Arganda-Carreras et al., 2017	v3.3.1	–
Software, algorithm	Ilastik	Berg et al., 2019	v1.3.3	–
Software, algorithm	Spatial Model Editor	https://spatial-model-editor.github.io	v1.2.1	–
Software, algorithm	sme-contrib	https://spatial-model-editor.github.io	v0.014	–

In vitro experiments

Cell culture and the establishment of genetically modified cells

Request a detailed protocol

The hepatocyte-derived cell lines Hep3B (#ACC93, DSMZ, Braunschweig, Germany), HLF (JCRB, Japan), HepG2, and SNU182 (ATCC/LGC Standards, Wesel, Germany) were cultured in Minimum Essential Medium (MEM), Dulbeccos Modified Eagle Medium (DMEM), and Roswell Park Memorial Institute (RPMI) medium, respectively. The media were supplemented with 10% FCS and 1% penicillin–streptomycin (Sigma-Aldrich, Taufkirchen, Germany). Cells were maintained at 37°C in a 5% CO₂ atmosphere. Hep3B cells were transduced with vectors carrying cDNAs for human histone H2B, YAP, and TAZ/WWTR1 genes, fused with mCerulean, mVenus, and mCherry genes, respectively. The plasmid pRRLN-EF1α-mVenus-YAP was kindly provided by Prof. Dr. Alexander Loewer (TU Darmstadt), pLentiPGK-mCerulean-H2B was purchased (Addgene, Watertown, USA; No. 90234). The pRRLN-EF1α-d2mCherry-TAZ was cloned. Correctness of vectors was verified by end-to-end sequencing.

Vectors were stably integrated into the Hep3B cells using lentiviral particles. For this, HEK293 cells were transfected with lentiviral vectors, packaging (psPAX2, Addgene, Watertown, USA; No. 12260), and envelope plasmids (pMD2.G, Addgene, Watertown, USA; No. 12259) using polyethylenimine and incubated for 40 hr. Subsequently, virus particles in the cell culture medium were collected and used for infections. The transduced cells were selected for stable vector integration using geneticin (0.6 mg/ml, pRRLN-EF1α-mVenus-YAP), hygromycin (0.6 mg/ml, pLentiPGK-mCerulean-H2B), and blasticidin (2 µg/ml, pRRLN-EF1α-d2mCherry-TAZ). Cells were frequently checked for mycoplasma contamination and authenticated by short tandem repeat analysis (DSMZ, Braunschweig, Germany).

The APAP stock solution (100 mM) was freshly prepared by dissolving 0.3 g APAP crystalline powder in 20 ml phosphate-buffered saline (PBS, GE Healthcare, Solingen, Germany) under heating (42°C) and stirring (Sigma-Aldrich, St. Louis, USA). The APAP solution was filtered (Millex-GS filters, 0.22 µm; Merck Millipore Ltd, Carrigtwohill, Ireland) and immediately administered to cells to a final concentration of 10 mM. TBHP (Luperax, Sigma-Aldrich) with 70 wt% in H₂O was applied to cells to a final concentration of 300 µM for 2 hr prior protein extraction. Hydrogen peroxide (H₂O₂) 30% (Rotipuran, Carl Roth GmbH, Karlsruhe, Germany) was applied to a final concentration of 2 mM for 1 hr.

AKT1/2 phosphorylation was inhibited with InSolution Akt Inhibitor VIII (1–10 µM, Merck KGaA, Darmstadt, Germany) in FCS-free medium for 3 hr prior to treatment with HGF (10 ng/µl).

Gene silencing by RNAi

Request a detailed protocol

RNAi was performed to inhibit YAP gene expression. For this, 2 × 10⁵ cells were seeded per well in a 6-cm-well plate and incubated overnight. Oligofectamine (Invitrogen, Carlsbad, CA, USA) and Opti-MEM (Gibco Life Technologies Ltd, Paisley, UK) were used for transfection of the siRNA (final concentration: 5 nM) according to the manufacturer’s instructions. A random sequence oligonucleotide was used as a control treatment (control siRNA: ctrl siRNA). The following siRNA sequences were used: YAP siRNA #1 (5′–3′): CCA CCA AGC UAG AUA AAG A-dT-dT and YAP siRNA #2 (5′–3′): GG UCA GAG AUA CUU CUU AA-dT-dT, and ctrl siRNA (5′–3′): UGG UUU ACA UGU CGA CUA A. YAP siRNA #1 and #2 were pooled to achieve an optimal gene knockdown. Twenty-four hours after siRNA transfection, cells were treated with APAP (10 mM) for 24 hr.

Live cell imaging

Request a detailed protocol

Stably transduced Hep3B cells were seeded on 24-well glass-bottom black wall plates (MoBiTec GmbH, Goettingen, Germany) in a phenol red-free RPMI 1640 medium (Gibco/Life Technologies Corporation, Paisley, UK). Dynamic protein localization in living cells was measured using confocal laser scanning microscope (Nikon A1R on an inverted Nikon Ti2 microscope), which was supplemented with an on-stage incubator (TokaiHit) to maintain 37°C and 5% CO₂ at all stages of the experiments. Selected images were acquired with additional Nikon AxR microscope system.

Three lasers were employed: 445 nm (mCerulean, PMT detector), 514 nm (mVenus, GaAsP detector), and 561 nm (mCherry, GaAsP detector) to obtain highly resolved live cell images. Two channels – 445 and 514 nm – were acquired simultaneously. The employed pinhole size was 17.9 µm (445 nm: 1.5 a.u., 514 nm: 1.3 a.u., 561 nm: 1.2 a.u.). Images were acquired using a galvano-scanner of size 512 × 512 px (FOV 0.64 × 0.64 mm, pixel resolution of 1.24 µm). The objectives Plan Apo λ 20× (NA 0.75, working distance 1 mm, FOV 0.64 × 0.64 mm) or Plan Apo λ 40× (NA 0.95, working distance 0.21 mm, FOV 0.435 × 0.435 mm) were used.

Western immunoblotting

Request a detailed protocol

Western immunoblotting was performed as previously described (Weiler et al., 2017). In brief, for total protein fraction preparation, cultured cells were harvested using 1× Cell Lysis Buffer (Cell Signaling Technology, Danvers, USA) supplemented with 1× Protease Inhibitor Mix G (Serva, Heidelberg, Germany) and 1× PhosSTOP (Roche Diagnostics Deutschland GmbH, Mannheim, Germany). For protein fractionation, the NE-PER Nuclear and Cytoplasmic Extraction kit was used (Thermo Scientific, Rockford, USA). The following antibodies were used for western immunoblotting experiments: YAP XP (1:1000, Cell Signaling Technology, Danvers, USA, #14074, RRID:AB_2650491), pYAP (1:400, Cell Signaling Technology, #4911, RRID:AB_2218913), pAKT (1:1’000, Cell Signaling Technology, #4060, RRID:AB_2315049), AKT (1:1’000, Cell Signaling Technology, #9272, RRID:AB_329827), β-actin (1:10,000, MP Biomedicals, Solon, USA, #08691001, RRID:AB_2335127), LATS1 (1:200, Santa Cruz Biotechnology, Dallas, USA, sc-398560), LATS2 (1:200, Santa Cruz Biotechnology, sc-515579), and pLATS1/2 (1:500, Cell Signaling Technology, #8654, RRID:AB_10971635) and Cytochrome P450 2E1 (1:500, Novus Biologicals, Centennial, USA, NVP1-85367, RRID:AB_11021447). PARP (1:10,000, Cell Signaling Technology, #9542, RRID:AB_2160739), β-tubulin (1:200, Santa Cruz Biotechnology, sc-5274, RRID:AB_2288090), and GAPDH (1:10,000, Merck Millipore, Darmstadt, Germany, ab2302, RRID:AB_10615768) were used as loading controls.

Western blot detection and quantification were performed using the Odyssey-CLx Infrared Imaging system with the ImageStudio software (LI-COR Biosciences, Bad Homburg, Germany). Phosphorylated protein band intensity was measured and normalized to total protein concentrations. Equal amount of protein was loaded for each line, as measured by the Bradford reagent (Millipore Sigma, Saint Louis, USA). Raw unedited image files and uncropped blots are available as source data files of this manuscript.

RNA isolation, reverse transcription, and quantitative real-time PCR (qPCR)

Request a detailed protocol

RNA was isolated using Extractme kit (Blirt, Gdańsk, Poland) according to the manufacturer’s protocol. Reverse transcription was performed using the RevertAid kit (Thermo Scientific). Semiquantitative real-time PCR reactions were set up using the primaQuant 2× qPCR-SYBR-Green-Mastermix (Steinbrenner Laborsysteme, Wiesenbach, Germany) and analyzed with the QuantStudio 3 real-time PCR system (Applied Biosystems, Thermo Fisher Scientific, Singapore). The following cycling conditions were applied: 95°C for 15 min, followed by 40 cycles of 95°C for 15 s, and 60°C for 60 s. Product specificity was confirmed by melting curve analysis (95°C for 15 s, 60°C for 30 s, 60–95°C with 0.5°C/s).

The mRNA levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 60S ribosomal protein L41 (RPL41), serine/arginine-rich splicing factor (SRSF4), and β2-microglobulin (B2M). The following primers for human cDNAs were used: YAP-for: 5′-CCT GCG TAG CCA GTT ACC AA-3′; YAP-rev: 5′-CCA TCT CAT CCA CAC TGT TC-3′; ANKRD1-for: 5′-AGT AGA GGA ACT GGT CAC TGG-3′; ANKRD1-rev: 5′-TGG GCT AGA AGT GTC TTC AGA T-3′; CYR61-for: 5′-AGC CTC GCA TCC TAT ACA ACC-3′; CYR61-rev: 5′-TTC TTT CAC AAG GCG GCA CTC-3′; GADPH-for: 5′-CTG GTA AAG TGG ATA TTG TTG CCA T-3′; GAPDH-rev: 5′-TGG AAT CAT ATT GGA ACA TGT AAA CC-3′; RPL41-for: 5′-AAA CCT CTG CGC CAT GAG AG-3′; RPL41-rev: 5′-AGC GTC TGG CAT TCC ATG TT-3′; SRSF4-for: 5′-TGC AGC TGG CAA GAC CTA AA-3′; SRSF4-rev: 5′-TTT TTG CGT CCC TTG TGA GC-3′; B2M-for: 5′-CAC GTC ATC CAG CAG AGA AT-3′; B2M-rev: 5′-TGC TGC TTA CAT GTC TCG AT-3′. For the analysis of gene expression in tissue samples, a panel of housekeeping genes was analyzed using the geNorm algorithm to find the most stable reference gene (Vandesompele et al., 2002).

In situ PLA

Request a detailed protocol

The DuoLink in situ PLA was performed according to the manufacturer’s instructions (Sigma-Aldrich). Briefly, cells were seeded on glass coverslips and prior to APAP administration cells were grown under FCS-free conditions for 1 day, then incubated with APAP (10 mM) or PBS. After treatment, cells were washed three times with 2 mM MgCl₂ in PBS and fixed with 4% paraformaldehyde for 10 min at room temperature. Fixed cells were washed four times with PBS for 5 min, permeabilized with 0.2% Triton X-100 in PBS for 5 min at room temperature, and washed again twice with PBS for 5 min. Subsequently, cells were blocked with Blocking solution (Sigma-Aldrich) for 30 min at room temperature and incubated with the following primary antibodies diluted in Antibody Diluent (Sigma-Aldrich) overnight at 4°C: anti-YAP (1:25, Santa Cruz Biotechnology, sc-271134, RRID:AB_10612397), anti-AKT (1:200, Cell Signaling Technology, #9272, RRID:AB_329827), anti-pYAP (1:200, Cell Signaling Technology, #4911, RRID:AB_2218913), anti-pAKT (1:200, Cell Signaling Technology, #4051, AB_331158), and anti-pLATS1/2 (1:200, Cell Signaling Technology, #8654, RRID:AB_10971635). Subsequently, cells were washed twice with Wash Buffer A (Sigma-Aldrich) and incubated with prediluted rabbit PLUS and mouse MINUS probes (Sigma-Aldrich) in Antibody Diluent for 1 hr at 37°C. After incubation, cells were washed twice with Wash Buffer A and then incubated with ligation solution (Sigma-Aldrich) for 30 min at 37°C. After ligation, samples were again washed twice with Wash Buffer A for 2 min at room temperature and incubated with amplification solution, and detection reagent Orange (Sigma-Aldrich) for 100 min at 37 °C. Finally, cells were washed twice with Wash Buffer B (Sigma-Aldrich) for 10 min at room temperature, once with 0.01× Wash Buffer B for 1 min at room temperature and coverslips were mounted on the slide with DAPI Fluoromount-G mounting medium (SouthernBiotech, Birmingham, USA).

Fluorescence images were captured using an inverted Nikon Ti2 microscope with Nikon S Plan Fluor ELWD ×40 NA 0.60 objective in a widefield fluorescence mode using Lumencor Sola SE II lamp. Images were captured in DAPI and TRITC channels (460 and 580 nm) with Nikon DS-Qi2 monochrome camera (image size 2404 × 2404 px, pixel resolution of 0.18 µm/px).

ROS activity measurement

Request a detailed protocol

Measurements of the intracellular ROS levels were performed using the DCFDA/H2DCFDA cellular ROS assay kit (Abcam, Amsterdam, Netherlands) according to the manufacturer’s protocol. In brief, cells were seeded on white clear-bottom 96-well plate (Corning, Corning, USA) and incubated overnight. Cells were treated with 10 mM APAP, 2 mM H₂O₂, or 300 µM TBHP (H₂O₂ and TBHP served as positive controls for ROS induction) for 6 hr in FCS-free cell culture medium. Fluorescence was measured using a microplate reader (FluoStar Omega, BMG Labtech GmbH, Ortenberg, Germany). Buffer solution without cells served as a background control.

In vivo experiments and sample analyses

Housing and treatment of mice and induction of acute liver injury by acetaminophen

Request a detailed protocol

Male C57BL/6N mice (8- to 10-week-old) were bought from Janvier Labs (Janvier Labs, Le Genest-Saint-Isle, France). Animals were housed under 12 hr light/dark cycles at controlled ambient temperature of 25°C with free access to water and were fed ad libitum with a standard diet (Ssniff, Soest, Germany) before starting the experiments. Induction of acute liver injury with APAP was done as previously described (Schneider et al., 2021). Briefly, the mice were fasted overnight, then challenged with a dose of 300 mg/kg APAP intraperitoneally. APAP was dissolved in warm PBS with an application volume of 30 ml/kg. Control group was treated with PBS only. The mice were fed ad libitum after APAP administration. All animals were included for further analyses. All experiments were approved by the local animal welfare committee (LANUV, North Rhine-Westphalia, Germany, application number: 84-02.04.2016.A279). Hydrodynamic gene delivery experiments in mice were performed as recently described (Luiken et al., 2020).

Liver tissue sample collection, processing, and staining

Request a detailed protocol

Tissues were collected time dependently after APAP injection from the left liver lobe. The tissues were fixed and embedded in paraffin as previously described (Ghallab et al., 2016). YAP and AKT immunostaining were performed using 4-µm-thick paraffin-embedded tissue sections. For immunohistochemistry, an anti-YAP antibody (1:50, Cell Signaling Technology, #14074, RRID:AB_2650491) and anti-AKT (1:50, Cell Signaling Technology, #9272, RRID:AB_329827) were used. Embedded tissue sections were pretreated with a heat-induced epitope retrieval method (pH 6, DAKO, Hamburg, Germany). As secondary antibody anti-rabbit Polymer-AP (Enzo Life Sciences, Farmingdale, USA, ENZ-ACC110-0150) was used. Detection was performed with Permanent AP (Zytomed Systems GmbH, Berlin, Germany). Following staining, the whole slides were digitally documented using a slide scanner (Aperio AT2, Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany).

Expression profiling and bioinformatics

Request a detailed protocol

RNA isolation and gene array analysis were performed as published before (Campos et al., 2020) and bioinformatic analysis was done as described (Holland et al., 2022). The gene expression data are available under ArrayExpress accession number GSE167032. The expression data were applied to three known signatures that are informative for ROS activity (Han et al., 2008), AKT activity (De Marco et al., 2017), and YAP/TAZ activity (Wang et al., 2018). Due to the high number of AKT signature genes, only genes whose response to APAP treatment was larger than fold change of 2 (compared to control animals) were considered. The expression data were z-score normalized and clustered with seborn clustermap python module (v0.11.0). The signature score was obtained by summarizing z-scored expression values at the given time point if the z-scored value was greater than 0.5. The summarized expression values were normalized to the number of genes in the signature.

Computational methods

Analysis of the live cell images

Request a detailed protocol

The confocal images of the living cells were analyzed in a high-throughput manner using ImageJ (v1.53f51) platform (Rueden et al., 2017). Images were first manually selected with respect to the quality criteria: images with insufficient sharpness or with artifacts were discarded from the analysis. The image processing pipeline was based on Weka segmentation (v3.3.1) (Arganda-Carreras et al., 2017) of foreground and background areas, and subsequent thresholding, object detection, and counting. After object detection and counting, the respective masks were overlaid on the initial images to acquire YAP and TAZ intensity values for nuclei and cytoplasm. Mean pixel intensity from nuclear areas of the cells was divided by the mean pixel intensity of cytoplasmic regions, thus obtaining an NCR.

Analysis of PLA images

Request a detailed protocol

PLA slides were analyzed using ImageJ (v1.53f51). First, nuclei and dots were classified using the Weka segmentation algorithm (v3.3.1) (Arganda-Carreras et al., 2017), thresholded, and counted. The pseudo-cytoplasmic (ring-shaped) area was created using the ImageJ’s binary mask option dilate for 30 iterations on nuclear masks to obtain nuclear and cytoplasmic area for a comparative analysis. The thresholded nuclei or cytoplasmic masks were overlaid with detected dots to obtain the information on the subcellular localization of the protein interaction.

Analysis of IHC images

Request a detailed protocol

Stained tissue samples were digitalized using Aperio slide scanner with ×40 magnification and pixel resolution of 0.253 µm/px (Aperio AT2, Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany). The selected time points (6 hr and 1, 2, 6, 16 days) and control treatment were quantified using a pipeline, which consisted of python scripts (modules PIL v5.3.0, matplotlib v2.2.3), ASAP software (v1.6, https://github.com/computationalpathologygroup/ASAP) and Ilastik software (v1.3.3) (Berg et al., 2019).

First, digital images of the tissue sections were divided in tiles (1 mm², 400 × 400 pixels), which were binned (to reduce processing time and storage load) using ASAP and python modules. Some images were excluded due to staining and/or scanning artifacts. Tiles, which displayed tissue for at least 50% of their area, were kept for further processing. Machine-learning model, which was based on a random forest algorithm, was trained on a selected set of training tiles for YAP or AKT using Ilastik software. The algorithm was trained to detect positively stained nuclei, excluding necrotic areas and staining artifacts. Ilastik software was further used to export probability maps, which were possessed with ImageJ. In ImageJ, probability maps were thresholded and positive nuclei were counted and normalized to the area of the tile, which was occupied by the tissue.

PDE modeling

Request a detailed protocol

Spatial modeling aimed at describing the variations in space of fluorescently labeled YAP and TAZ distribution within living Hep3B cells. The mathematical model is based on a system of PDEs. PDE modeling and parameter estimation were performed with the Spatial Model Editor (SME) software (v1.2.1) and sme-contrib (v0.0.14) python module (https://spatial-model-editor.github.io/). SME is graphical user interface-based model editing and simulation software compatible with systems biology markup language (SBML) standards. Models were simulated using simple Forward Time Centered Space (FTCS) solver.

In the demonstrated models, all reactions were defined by first-order kinetics (for PDEs see Appendix 1). For the parameter estimation, the model behavior was evaluated at steady state, that is the model was simulated until the concentration distribution did not change over time. Parameter estimation was performed using the particle swarm algorithm (20 particles, 200 iterations) to minimize a cost function consisting of the weighted sum of two terms: the squared per pixel differences between model and the target image, and the sum of squares of species concentration rates of change. In total 200 fitted parametrizations for each tested model were generated. The parameter space for the optimization algorithm was defined based on published data (Appendix 1). The parameter ranges of our canonical and alternative models were selected in a way to describe biologically meaningful value ranges and to avoid numeric instability during PDE simulations of the canonical model. The mathematical models were uploaded to the BioModels repository under the model identifier number MODEL2202080001. PDE model equations and parameters can be found in Supplementary Information (Appendix 1—Tables 1–6; Ege et al., 2018; Jack et al., 1990). Tables showing all fitted parameters for YAP and TAZ are provided (Supplementary file 1, Supplementary file 2).

Statistics

Statistical analysis was performed using GraphPad Prism 9.2.0. Statistical tests are indicated in figure legends. Error bars depict standard deviation. Significance levels are as follows: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Appendix 1

PDE model equations

In the computational model, the equations which are simulated in cytoplasm or nucleus follow two-dimensional reaction diffusion equation. Initial conditions for the PDE are arbitrary since only steady-state solutions are considered. The equations in TAZ model are equivalent to YAP model equations.

Here, described in general terms as follows:

\frac{\partial C_{s}}{\partial t} = R_{s} + D_{s} \nabla^{2} C_{s}

where

· $C_{s}$ is the concentration of species $S$ at position $(x, y)$ and time $t$

· $R_{s}$ is the reaction term for species $S$

· $D_{s}$ is the diffusion constant for species $S$

· $\nabla^{2} C_{s}$ is the Laplacian of species $S$ , which can be rewritten as

\nabla^{2} C_{s} = \nabla \cdot \nabla C_{s} = \frac{\partial^{2} C_{s}}{\partial x^{2}} + \frac{\partial^{2} C_{s}}{\partial y^{2}}

The PDEs for the dynamics of YAPn/pYAPn (nuclear fraction) and YAP/pYAP (cytosolic fraction) proteins in the alternative model in the nucleus or cytoplasm are presented as follows.

\frac{\partial [Y A P]}{\partial t} = R_{1} + D_{Y A P} \nabla^{2} [Y A P]

\frac{\partial [p Y A P]}{\partial t} = - R_{2} + D_{p Y A P} \nabla^{2} [p Y A P]

\frac{\partial [Y A P n]}{\partial t} = - R_{5} + R_{6} + D_{Y A P n} \nabla^{2} [Y A P n]

\frac{\partial [p Y A P n]}{\partial t} = R_{3} - R_{4} + D_{p Y A P n} \nabla^{2} [p Y A P n]

where reaction rates $R$ are defined as follows:

R₁ as translation:

R_{1} = K_{t r a n s l a t i o n}

R₂ as degradation:

R_{2} = K_{d e g r a d a t i o n} [p Y A P]

R₅ as phosphorylation:

R_{3} = K_{p h o s p h o} [Y A P n]

R₆ as dephosphorylation:

R_{4} = K_{d e p h o s p h o} [p Y A P n]

The transport reactions for YAP and pYAP are defined as flux densities across the nuclear membrane that depends on the respective concentrations; these are translated into Neumann-type interface conditions for the inner boundary of the cytoplasm and the outer boundary of the nucleus by the simulator software.

Flux density of YAP import
- $F_{i m p o r t} = k_{i m p o r t} [Y A P]$
Flux density of YAP export
- $F_{e x p o r t} = k_{e x p o r t} [p Y A P n]$

Boundary conditions on the outer membrane of the cytoplasm are ‘zero-flux’ Neumann type for all variables.

Model parameters

Appendix 1—table 1

Canonical model parameters.

Parameter	Symbol	YAP model	TAZ model	Unit
Diffusion	$D_{Y A P}$ $D_{Y A P n}$	0.0391776	0.00812799	µm²/s
Diffusion phosphorylated	$D_{p Y A P n}$ $D_{p Y A P}$	0.433785	0.501116	µm²/s
Phosphorylation	$K_{p h o s p h o}$	3.2462731384467	9.2660885020866	s^–1
Dephosphorylation	$K_{d e p h o s p h o}$	2.4380091829261	2.9721728333365	s^–1
Translation	$K_{t r a n s l a t i o n}$	0.013295743091699	0.02329254213393	a.u./s
Degradation	$K_{d e g r a d a t i o n}$	0.0081058910608829	0.0090941179957213	s^–1
Nuclear import	$K_{i m p o r t}$	9.7229134833679e−16	8.1106772583928e−16	µm/s
Nuclear export	$K_{e x p o r t}$	9.5119620861314e−17	6.1050891990604e−17	µm/s

Appendix 1—table 2

Alternative model parameters.

Parameter	Symbol	YAP model	TAZ model	Unit
Diffusion	$D_{Y A P}$ $D_{Y A P n}$	9.84707	9.14393	µm²/s
Diffusion phosphorylated	$D_{p Y A P n}$ $D_{p Y A P}$	1.04518	1.00732	µm²/s
Phosphorylation	$K_{p h o s p h o}$	0.028839815689838	0.10286291836829	s^–1
Dephosphorylation	$K_{d e p h o s p h o}$	0.09121073484801	0.10517790271802	s^–1
Translation	$K_{t r a n s l a t i o n}$	0.009403085138566	0.0030932046019853	a.u./s
Degradation	$K_{d e g r a d a t i o n}$	0.0094299166245222	0.0097940321843269	s^–1
Nuclear import	$K_{i m p o r t}$	6.5845623444598e−15	9.5850453479782e−15	µm/s
Nuclear export	$K_{e x p o r t}$	7.8819897236132e−15	4.5876638224922e−15	µm/s

Appendix 1—table 3

Parameters of the YAP-to-TAZ transition model.

Parameter	Symbol	Value	Unit
Diffusion	$D_{Y A P}$ $D_{Y A P n}$	10	µm²/s
Diffusion phosphorylated	$D_{p Y A P n}$ $D_{p Y A P}$	1	µm²/s
Phosphorylation	$K_{p h o s p h o}$	–	s^–1
Dephosphorylation	$K_{d e p h o s p h o}$	–	s^–1
Translation	$K_{t r a n s l a t i o n}$	0.01	a.u./s
Degradation	$K_{d e g r a d a t i o n}$	0.004	s^–1
Nuclear import	$K_{i m p o r t}$	7.5e−15	µm/s
Nuclear export	$K_{e x p o r t}$	1.6e−15	µm/s

Appendix 1—table 4

Parameter range for particle swarm algorithm of the canonical YAP/TAZ model.

Parameter	Symbol	Range (min, max)	Unit
Diffusion	$D_{Y A P}$ $D_{Y A P n}$	(0.001, 5)	µm²/s
Diffusion phosphorylated	$D_{p Y A P n}$ $D_{p Y A P}$	(0.001, 1)	µm²/s
Phosphorylation	$K_{p h o s p h o}$	(0.1, 10)	s^–1
Dephosphorylation	$K_{d e p h o s p h o}$	(0.1, 10)	s^–1
Translation	$K_{t r a n s l a t i o n}$	(1e−3, 0.1)	a.u./s
Degradation	$K_{d e g r a d a t i o n}$	(1e−3, 1e−2)	s^–1
Nuclear import	$K_{i m p o r t}$	(1e−16, 1e−15)	µm/s
Nuclear export	$K_{e x p o r t}$	(1e−17, 1e−15)	µm/s

Appendix 1—table 5

Parameter range for particle swarm algorithm of the alternative YAP/TAZ model.

Parameter	Symbol	Range (min, max)	Unit
Diffusion	$D_{Y A P}$ $D_{Y A P n}$	(1, 10)	µm²/s
Diffusion phosphorylated	$D_{p Y A P n}$ $D_{p Y A P}$	(1, 1.5)	µm²/s
Phosphorylation	$K_{p h o s p h o}$	(0.01, 0.5)	s^–1
Dephosphorylation	$K_{d e p h o s p h o}$	(0.01, 0.5)	s^–1
Translation	$K_{t r a n s l a t i o n}$	(1e−3, 0.01)	a.u./s
Degradation	$K_{d e g r a d a t i o n}$	(1e−3, 0.01)	s^–1
Nuclear import	$K_{i m p o r t}$	(1e−16, 1e−14)	µm/s
Nuclear export	$K_{e x p o r t}$	(1e−16, 1e−14)	µm/s

Appendix 1—table 6

Parameter values from the literature.

Parameter	Value	Source
Hepatocyte cell volume	10–11 l	Jack et al., 1990
Hepatocyte nucleus volume	5 ×10^–13	Jack et al., 1990
Nuclear export	10–100 s	Ege et al., 2018
Nuclear import	50 s	Ege et al., 2018
Diffusion rate YAP/TAZ	19 μm²/s	Ege et al., 2018

Data availability

The mathematical models generated and analyzed during the current study are available in the BioModels repository (https://www.ebi.ac.uk/biomodels/) under accession number MODEL2202080001. The PDE simulation software is available over Github (https://spatial-model-editor.github.io/). The gene expression data analyzed during the current study are available in the Gene Expression Omnibus (GEO) repository under accession number GSE167032.

The following data sets were generated

1. Wehling L
2. Sahle S
(2022) EMBL-EBI BioModels
ID MODEL2202080001. MODEL2202080001.

https://www.ebi.ac.uk/biomodels/MODEL2202080001

The following previously published data sets were used

(2021) NCBI Gene Expression Omnibus
ID GSE167032. Expression data of the livers of male C57Bl6/N mice after i.p. injection of paracetamol (APAP).

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE167032

References

1. Adlung L
2. Kar S
3. Wagner M-C
4. She B
5. Chakraborty S
6. Bao J
7. Lattermann S
8. Boerries M
9. Busch H
10. Wuchter P
11. Ho AD
12. Timmer J
13. Schilling M
14. Höfer T
15. Klingmüller U
(2017) Protein abundance of Akt and ERK pathway components governs cell type-specific regulation of proliferation
Molecular Systems Biology 13:904.
https://doi.org/10.15252/msb.20167258
- PubMed
- Google Scholar
(2017) Trainable weka segmentation: a machine learning tool for microscopy pixel classification
Bioinformatics 33:2424–2426.
https://doi.org/10.1093/bioinformatics/btx180
- PubMed
- Google Scholar
(2017) The mitochondrial negative regulator MCJ is a therapeutic target for acetaminophen-induced liver injury
Nature Communications 8:2068.
https://doi.org/10.1038/s41467-017-01970-x
- PubMed
- Google Scholar
1. Barnett SF
2. Defeo-Jones D
3. Fu S
4. Hancock PJ
5. Haskell KM
6. Jones RE
7. Kahana JA
8. Kral AM
9. Leander K
10. Lee LL
11. Malinowski J
12. McAvoy EM
13. Nahas DD
14. Robinson RG
15. Huber HE
(2005) Identification and characterization of pleckstrin-homology-domain-dependent and isoenzyme-specific Akt inhibitors
The Biochemical Journal 385:399–408.
https://doi.org/10.1042/BJ20041140
- PubMed
- Google Scholar
1. Basu S
2. Totty NF
3. Irwin MS
4. Sudol M
5. Downward J
(2003) Akt phosphorylates the Yes-associated protein, YAP, to induce interaction with 14-3-3 and attenuation of p73-mediated apoptosis
Molecular Cell 11:11–23.
https://doi.org/10.1016/s1097-2765(02)00776-1
- PubMed
- Google Scholar
1. Berg S
2. Kutra D
3. Kroeger T
4. Straehle CN
5. Kausler BX
6. Haubold C
7. Schiegg M
8. Ales J
9. Beier T
10. Rudy M
11. Eren K
12. Cervantes JI
13. Xu B
14. Beuttenmueller F
15. Wolny A
16. Zhang C
17. Koethe U
18. Hamprecht FA
19. Kreshuk A
(2019) Ilastik: interactive machine learning for (BIO) image analysis
Nature Methods 16:1226–1232.
https://doi.org/10.1038/s41592-019-0582-9
- PubMed
- Google Scholar
1. Campos G
2. Schmidt-Heck W
3. De Smedt J
4. Widera A
5. Ghallab A
6. Pütter L
7. González D
8. Edlund K
9. Cadenas C
10. Marchan R
11. Guthke R
12. Verfaillie C
13. Hetz C
14. Sachinidis A
15. Braeuning A
16. Schwarz M
17. Weiß TS
18. Banhart BK
19. Hoek J
20. Vadigepalli R
21. Willy J
22. Stevens JL
23. Hay DC
24. Hengstler JG
25. Godoy P
(2020) Inflammation-Associated suppression of metabolic gene networks in acute and chronic liver disease
Archives of Toxicology 94:205–217.
https://doi.org/10.1007/s00204-019-02630-3
- PubMed
- Google Scholar
1. De Marco C
2. Laudanna C
3. Rinaldo N
4. Oliveira DM
5. Ravo M
6. Weisz A
7. Ceccarelli M
8. Caira E
9. Rizzuto A
10. Zoppoli P
11. Malanga D
12. Viglietto G
(2017) Specific gene expression signatures induced by the multiple oncogenic alterations that occur within the PTEN/PI3K/Akt pathway in lung cancer
PLOS ONE 12:e0178865.
https://doi.org/10.1371/journal.pone.0178865
- PubMed
- Google Scholar
1. Deng H
2. Wang W
3. Yu J
4. Zheng Y
5. Qing Y
6. Pan D
(2015) Spectrin regulates Hippo signaling by modulating cortical actomyosin activity
eLife 4:e06567.
https://doi.org/10.7554/eLife.06567
- PubMed
- Google Scholar
1. Dong J
2. Feldmann G
3. Huang J
4. Wu S
5. Zhang N
6. Comerford SA
7. Gayyed MF
8. Anders RA
9. Maitra A
10. Pan D
(2007) Elucidation of a universal size-control mechanism in Drosophila and mammals
Cell 130:1120–1133.
https://doi.org/10.1016/j.cell.2007.07.019
- PubMed
- Google Scholar
1. Ege N
2. Dowbaj AM
3. Jiang M
4. Howell M
5. Hooper S
6. Foster C
7. Jenkins RP
8. Sahai E
(2018) Quantitative analysis reveals that actin and Src-family kinases regulate nuclear Yap1 and its export
Cell Systems 6:692–708.
https://doi.org/10.1016/j.cels.2018.05.006
- PubMed
- Google Scholar
(2021) Exploring the influence of cytosolic and membrane FAK activation on YAP/TAZ nuclear translocation
Biophysical Journal 120:4360–4377.
https://doi.org/10.1016/j.bpj.2021.09.009
- PubMed
- Google Scholar
1. Fan F
2. He Z
3. Kong LL
4. Chen Q
5. Yuan Q
6. Zhang S
7. Ye J
8. Liu H
9. Sun X
10. Geng J
11. Yuan L
12. Hong L
13. Xiao C
14. Zhang W
15. Sun X
16. Li Y
17. Wang P
18. Huang L
19. Wu X
20. Ji Z
21. Wu Q
22. Xia NS
23. Gray NS
24. Chen L
25. Yun CH
26. Deng XZhou D
(2016) Pharmacological targeting of kinases MST1 and MST2 augments tissue repair and regeneration
Science Translational Medicine 8:352ra108.
https://doi.org/10.1126/scitranslmed.aaf2304
- Google Scholar
1. Gao J
2. He L
3. Shi Y
4. Cai M
5. Xu H
6. Jiang J
7. Zhang L
8. Wang H
(2017) Cell contact and pressure control of YAP localization and clustering revealed by super-resolution imaging
Nanoscale 9:16993–17003.
https://doi.org/10.1039/c7nr05818g
- PubMed
- Google Scholar
1. Ghallab A
2. Cellière G
3. Henkel SG
4. Driesch D
5. Hoehme S
6. Hofmann U
7. Zellmer S
8. Godoy P
9. Sachinidis A
10. Blaszkewicz M
11. Reif R
12. Marchan R
13. Kuepfer L
14. Häussinger D
15. Drasdo D
16. Gebhardt R
17. Hengstler JG
(2016) Model-Guided identification of a therapeutic strategy to reduce hyperammonemia in liver diseases
Journal of Hepatology 64:860–871.
https://doi.org/10.1016/j.jhep.2015.11.018
- PubMed
- Google Scholar
1. Ghallab A
2. Hassan R
3. Hofmann U
4. Friebel A
5. Hobloss Z
6. Brackhagen L
7. Begher-Tibbe B
8. Myllys M
9. Reinders J
10. Overbeck N
11. Sezgin S
12. Zühlke S
13. Seddek A-L
14. Murad W
15. Brecklinghaus T
16. Kappenberg F
17. Rahnenführer J
18. González D
19. Goldring C
20. Copple IM
21. Marchan R
22. Longerich T
23. Vucur M
24. Luedde T
25. Urban S
26. Canbay A
27. Schreiter T
28. Trauner M
29. Akakpo JY
30. Olyaee M
31. Curry SC
32. Sowa J-P
33. Jaeschke H
34. Hoehme S
35. Hengstler JG
(2022) Interruption of bile acid uptake by hepatocytes after acetaminophen overdose ameliorates hepatotoxicity
Journal of Hepatology 77:71–83.
https://doi.org/10.1016/j.jhep.2022.01.020
- PubMed
- Google Scholar
1. Gou J
2. Lin L
3. Othmer HG
(2018) A model for the Hippo pathway in the Drosophila wing disc
Biophysical Journal 115:737–747.
https://doi.org/10.1016/j.bpj.2018.07.002
- PubMed
- Google Scholar
1. Han E-S
2. Muller FL
3. Pérez VI
4. Qi W
5. Liang H
6. Xi L
7. Fu C
8. Doyle E
9. Hickey M
10. Cornell J
11. Epstein CJ
12. Roberts LJ
13. Van Remmen H
14. Richardson A
(2008) The in vivo gene expression signature of oxidative stress
Physiological Genomics 34:112–126.
https://doi.org/10.1152/physiolgenomics.00239.2007
- PubMed
- Google Scholar
1. Holland CH
2. Ramirez Flores RO
3. Myllys M
4. Hassan R
5. Edlund K
6. Hofmann U
7. Marchan R
8. Cadenas C
9. Reinders J
10. Hoehme S
11. Seddek A-L
12. Dooley S
13. Keitel V
14. Godoy P
15. Begher-Tibbe B
16. Trautwein C
17. Rupp C
18. Mueller S
19. Longerich T
20. Hengstler JG
21. Saez-Rodriguez J
22. Ghallab A
(2022) Transcriptomic cross-species analysis of chronic liver disease reveals consistent regulation between humans and mice
Hepatology Communications 6:161–177.
https://doi.org/10.1002/hep4.1797
- PubMed
- Google Scholar
(1990) Increase in hepatocyte and nuclear volume and decrease in the population of binucleated cells in preneoplastic foci of rat liver: a stereological study using the nucleator method
Hepatology 11:286–297.
https://doi.org/10.1002/hep.1840110220
- PubMed
- Google Scholar
1. Koundouros NPoulogiannis G
(2018) Phosphoinositide 3-kinase/Akt signaling and redox metabolism in cancer
Frontiers in Oncology 8:160.
https://doi.org/10.3389/fonc.2018.00160
- Google Scholar
(2020) Modeling the control of TGF-beta/smad nuclear accumulation by the Hippo pathway effectors
Taz/Yap. IScience 23:101416.
https://doi.org/10.1016/j.isci.2020.101416
- Google Scholar
(1996) Role of CYP2E1 in the hepatotoxicity of acetaminophen
The Journal of Biological Chemistry 271:12063–12067.
https://doi.org/10.1074/jbc.271.20.12063
- PubMed
- Google Scholar
(2014) Yap and TAZ regulate skin wound healing
The Journal of Investigative Dermatology 134:518–525.
https://doi.org/10.1038/jid.2013.339
- PubMed
- Google Scholar
1. Li W
2. Cooper J
3. Zhou L
4. Yang C
5. Erdjument-Bromage H
6. Zagzag D
7. Snuderl M
8. Ladanyi M
9. Hanemann CO
10. Zhou P
11. Karajannis MA
12. Giancotti FG
(2014) Merlin/Nf2 loss-driven tumorigenesis linked to CRL4 (DCAF1) -mediated inhibition of the Hippo pathway kinases LATS1 and 2 in the nucleus
Cancer Cell 26:48–60.
https://doi.org/10.1016/j.ccr.2014.05.001
- PubMed
- Google Scholar
1. Liu Y
2. Lu T
3. Zhang C
4. Xu J
5. Xue Z
6. Busuttil RW
7. Xu N
8. Xia Q
9. Kupiec-Weglinski JW
10. Ji H
(2019) Activation of YAP attenuates hepatic damage and fibrosis in liver ischemia-reperfusion injury
Journal of Hepatology 71:719–730.
https://doi.org/10.1016/j.jhep.2019.05.029
- PubMed
- Google Scholar
1. Low BC
2. Pan CQ
3. Shivashankar GV
4. Bershadsky A
5. Sudol M
6. Sheetz M
(2014) Yap/Taz as mechanosensors and mechanotransducers in regulating organ size and tumor growth
FEBS Letters 588:2663–2670.
https://doi.org/10.1016/j.febslet.2014.04.012
- PubMed
- Google Scholar
(2018) Hippo pathway coactivators YAP and TAZ are required to coordinate mammalian liver regeneration
Experimental & Molecular Medicine 50:e423.
https://doi.org/10.1038/emm.2017.205
- PubMed
- Google Scholar
1. Luiken S
2. Fraas A
3. Bieg M
4. Sugiyanto R
5. Goeppert B
6. Singer S
7. Ploeger C
8. Warsow G
9. Marquardt JU
10. Sticht C
11. De La Torre C
12. Pusch S
13. Mehrabi A
14. Gretz N
15. Schlesner M
16. Eils R
17. Schirmacher P
18. Longerich T
19. Roessler S
(2020) Notch target gene Hes5 mediates oncogenic and tumor suppressive functions in hepatocarcinogenesis
Oncogene 39:3128–3144.
https://doi.org/10.1038/s41388-020-1198-3
- PubMed
- Google Scholar
(2015) Accumulation of duct cells with activated YAP parallels fibrosis progression in non-alcoholic fatty liver disease
Journal of Hepatology 63:962–970.
https://doi.org/10.1016/j.jhep.2015.05.031
- PubMed
- Google Scholar
1. Marti P
2. Stein C
3. Blumer T
4. Abraham Y
5. Dill MT
6. Pikiolek M
7. Orsini V
8. Jurisic G
9. Megel P
10. Makowska Z
11. Agarinis C
12. Tornillo L
13. Bouwmeester T
14. Ruffner H
15. Bauer A
16. Parker CN
17. Schmelzle T
18. Terracciano LM
19. Heim MH
20. Tchorz JS
(2015) Yap promotes proliferation, chemoresistance, and angiogenesis in human cholangiocarcinoma through TeaD transcription factors
Hepatology 62:1497–1510.
https://doi.org/10.1002/hep.27992
- PubMed
- Google Scholar
1. Mooring M
2. Fowl BH
3. Lum SZC
4. Liu Y
5. Yao K
6. Softic S
7. Kirchner R
8. Bernstein A
9. Singhi AD
10. Jay DG
11. Kahn CR
12. Camargo FD
13. Yimlamai D
(2020) Hepatocyte stress increases expression of Yes-associated protein and transcriptional coactivator with PDZ-binding motif in hepatocytes to promote parenchymal inflammation and fibrosis
Hepatology 71:1813–1830.
https://doi.org/10.1002/hep.30928
- PubMed
- Google Scholar
1. Plouffe SW
2. Lin KC
3. Moore JL
4. Tan FE
5. Ma S
6. Ye Z
7. Qiu Y
8. Ren B
9. Guan KL
(2018) The Hippo pathway effector proteins YAP and TAZ have both distinct and overlapping functions in the cell
The Journal of Biological Chemistry 293:11230–11240.
https://doi.org/10.1074/jbc.RA118.002715
- PubMed
- Google Scholar
1. Poudel S
2. Cabrera DP
3. Bhushan B
4. Manley MW
5. Gunewardena S
6. Jaeschke H
7. Apte U
(2021) Hepatocyte-Specific deletion of Yes-associated protein improves recovery from acetaminophen-induced acute liver injury
Toxicological Sciences 184:276–285.
https://doi.org/10.1093/toxsci/kfab115
- PubMed
- Google Scholar
(2021) Yap and TAZ are not identical twins
Trends in Biochemical Sciences 46:154–168.
https://doi.org/10.1016/j.tibs.2020.08.012
- PubMed
- Google Scholar
1. Romano D
2. Matallanas D
3. Weitsman G
4. Preisinger C
5. Ng T
6. Kolch W
(2010) Proapoptotic kinase MST2 coordinates signaling crosstalk between RASSF1A, Raf-1, and Akt
Cancer Research 70:1195–1203.
https://doi.org/10.1158/0008-5472.CAN-09-3147
- PubMed
- Google Scholar
1. Romano D
2. Nguyen LK
3. Matallanas D
4. Halasz M
5. Doherty C
6. Kholodenko BN
7. Kolch W
(2014) Protein interaction switches coordinate Raf-1 and MST2/hippo signalling
Nature Cell Biology 16:673–684.
https://doi.org/10.1038/ncb2986
- PubMed
- Google Scholar
1. Rueden CT
2. Schindelin J
3. Hiner MC
4. DeZonia BE
5. Walter AE
6. Arena ET
7. Eliceiri KW
(2017) ImageJ2: imagej for the next generation of scientific image data
BMC Bioinformatics 18:529.
https://doi.org/10.1186/s12859-017-1934-z
- PubMed
- Google Scholar
1. Schneider KM
2. Elfers C
3. Ghallab A
4. Schneider CV
5. Galvez EJC
6. Mohs A
7. Gui W
8. Candels LS
9. Wirtz TH
10. Zuehlke S
11. Spiteller M
12. Myllys M
13. Roulet A
14. Ouzerdine A
15. Lelouvier B
16. Kilic K
17. Liao L
18. Nier A
19. Latz E
20. Bergheim I
21. Thaiss CA
22. Hengstler JG
23. Strowig T
24. Trautwein C
(2021) Intestinal dysbiosis amplifies acetaminophen-induced acute liver injury
Cellular and Molecular Gastroenterology and Hepatology 11:909–933.
https://doi.org/10.1016/j.jcmgh.2020.11.002
- PubMed
- Google Scholar
(2021) A spatial model of YAP/TAZ signaling reveals how stiffness, dimensionality, and shape contribute to emergent outcomes
PNAS 118:e20215–e27111.
https://doi.org/10.1073/pnas.2021571118
- PubMed
- Google Scholar
1. Shreberk-Shaked M
2. Oren M
(2019) New insights into YAP/TAZ nucleo-cytoplasmic shuttling: new cancer therapeutic opportunities?
Molecular Oncology 13:1335–1341.
https://doi.org/10.1002/1878-0261.12498
- PubMed
- Google Scholar
1. Shuhendler AJ
2. Pu K
3. Cui L
4. Uetrecht JP
5. Rao J
(2014) Real-Time imaging of oxidative and nitrosative stress in the liver of live animals for drug-toxicity testing
Nature Biotechnology 32:373–380.
https://doi.org/10.1038/nbt.2838
- PubMed
- Google Scholar
1. Sun M
2. Spill F
3. Zaman MH
(2016) A computational model of YAP/TAZ mechanosensing
Biophysical Journal 110:2540–2550.
https://doi.org/10.1016/j.bpj.2016.04.040
- PubMed
- Google Scholar
1. Thomann S
2. Weiler SME
3. Marquard S
4. Rose F
5. Ball CR
6. Tóth M
7. Wei T
8. Sticht C
9. Fritzsche S
10. Roessler S
11. De La Torre C
12. Ryschich E
13. Ermakova O
14. Mogler C
15. Kazdal D
16. Gretz N
17. Glimm H
18. Rempel E
19. Schirmacher P
20. Breuhahn K
(2020) Yap orchestrates heterotypic endothelial cell communication via HGF/c-Met signaling in liver tumorigenesis
Cancer Research 80:5502–5514.
https://doi.org/10.1158/0008-5472.CAN-20-0242
- PubMed
- Google Scholar
1. Tóth M
2. Wehling L
3. Thiess L
4. Rose F
5. Schmitt J
6. Weiler SME
7. Sticht C
8. De La Torre C
9. Rausch M
10. Albrecht T
11. Grabe N
12. Duwe L
13. Andersen JB
14. Köhler BC
15. Springfeld C
16. Mehrabi A
17. Kulu Y
18. Schirmacher P
19. Roessler S
20. Goeppert B
21. Breuhahn K
(2021) Co-Expression of YAP and TAZ associates with chromosomal instability in human cholangiocarcinoma
BMC Cancer 21:1079.
https://doi.org/10.1186/s12885-021-08794-5
- PubMed
- Google Scholar
(2002) Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes
Genome Biology 3:RESEARCH0034.
https://doi.org/10.1186/gb-2002-3-7-research0034
- PubMed
- Google Scholar
1. Wang X
2. Zheng Z
3. Caviglia JM
4. Corey KE
5. Herfel TM
6. Cai B
7. Masia R
8. Chung RT
9. Lefkowitch JH
10. Schwabe RF
11. Tabas I
(2016) Hepatocyte TAZ/WWTR1 promotes inflammation and fibrosis in nonalcoholic steatohepatitis
Cell Metabolism 24:848–862.
https://doi.org/10.1016/j.cmet.2016.09.016
- PubMed
- Google Scholar
1. Wang Y
2. Xu X
3. Maglic D
4. Dill MT
5. Mojumdar K
6. Ng PK-S
7. Jeong KJ
8. Tsang YH
9. Moreno D
10. Bhavana VH
11. Peng X
12. Ge Z
13. Chen H
14. Li J
15. Chen Z
16. Zhang H
17. Han L
18. Du D
19. Creighton CJ
20. Mills GB
21. Cancer Genome Atlas Research Network
22. Camargo F
23. Liang H
(2018) Comprehensive molecular characterization of the Hippo signaling pathway in cancer
Cell Reports 25:1304–1317.
https://doi.org/10.1016/j.celrep.2018.10.001
- PubMed
- Google Scholar
1. Weiler SME
2. Pinna F
3. Wolf T
4. Lutz T
5. Geldiyev A
6. Sticht C
7. Knaub M
8. Thomann S
9. Bissinger M
10. Wan S
11. Rössler S
12. Becker D
13. Gretz N
14. Lang H
15. Bergmann F
16. Ustiyan V
17. Kalin TV
18. Singer S
19. Lee J-S
20. Marquardt JU
21. Schirmacher P
22. Kalinichenko VV
23. Breuhahn K
(2017) Induction of chromosome instability by activation of Yes-associated protein and forkhead box M1 in liver cancer
Gastroenterology 152:2037–2051.
https://doi.org/10.1053/j.gastro.2017.02.018
- PubMed
- Google Scholar
1. Weiler SME
2. Lutz T
3. Bissinger M
4. Sticht C
5. Knaub M
6. Gretz N
7. Schirmacher P
8. Breuhahn K
(2020) Taz target gene ITGAV regulates invasion and feeds back positively on YAP and TAZ in liver cancer cells
Cancer Letters 473:164–175.
https://doi.org/10.1016/j.canlet.2019.12.044
- PubMed
- Google Scholar
(2021) A systematic assessment of the epidemiologic literature regarding an association between acetaminophen exposure and cancer
Regulatory Toxicology and Pharmacology 127:105043.
https://doi.org/10.1016/j.yrtph.2021.105043
- PubMed
- Google Scholar
(2001) Anti-Apoptotic signaling by hepatocyte growth factor/Met via the phosphatidylinositol 3-kinase/Akt and mitogen-activated protein kinase pathways
PNAS 98:247–252.
https://doi.org/10.1073/pnas.98.1.247
- PubMed
- Google Scholar
1. Xin M
2. Kim Y
3. Sutherland LB
4. Murakami M
5. Qi X
6. McAnally J
7. Porrello ER
8. Mahmoud AI
9. Tan W
10. Shelton JM
11. Richardson JA
12. Sadek HA
13. Bassel-Duby R
14. Olson EN
(2013) Hippo pathway effector YAP promotes cardiac regeneration
PNAS 110:13839–13844.
https://doi.org/10.1073/pnas.1313192110
- PubMed
- Google Scholar
1. Zanconato F
2. Forcato M
3. Battilana G
4. Azzolin L
5. Quaranta E
6. Bodega B
7. Rosato A
8. Bicciato S
9. Cordenonsi M
10. Piccolo S
(2015) Genome-Wide association between YAP/TAZ/TEAD and AP-1 at enhancers drives oncogenic growth
Nature Cell Biology 17:1218–1227.
https://doi.org/10.1038/ncb3216
- PubMed
- Google Scholar
1. Zhao B
2. Wei X
3. Li W
4. Udan RS
5. Yang Q
6. Kim J
7. Xie J
8. Ikenoue T
9. Yu J
10. Li L
11. Zheng P
12. Ye K
13. Chinnaiyan A
14. Halder G
15. Lai Z-C
16. Guan K-L
(2007) Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control
Genes & Development 21:2747–2761.
https://doi.org/10.1101/gad.1602907
- PubMed
- Google Scholar
(2011) The Hippo pathway in organ size control, tissue regeneration and stem cell self-renewal
Nature Cell Biology 13:877–883.
https://doi.org/10.1038/ncb2303
- PubMed
- Google Scholar

Decision letter

Matthew A Quinn

Reviewing Editor; Wake Forest School of Medicine, United States
Mone Zaidi

Senior Editor; Icahn School of Medicine at Mount Sinai, United States
Matthew A Quinn

Reviewer; Wake Forest School of Medicine, United States
Dirk Fey

Reviewer; University College Dublin, Ireland

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Spatial modeling reveals nuclear phosphorylation and subcellular shuttling of YAP upon drug-induced liver injury" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Matthew A Quinn as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Reviewing Editor and Mone Zaidi as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Dirk Fey (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) The model features very different diffusion rates for the canonical and alternative model, but is not explained why. For the estimation, different parameter ranges were specified for the canonical and alternative model, also for other parameters to be estimated, e.g. phosphorylation and dephosphorylation rates. This also lacks an explanation. What is the rationale behind this? Please update the discussion and/or methods to clarify this issue.

2) 200 parametrizations were obtained. Is the simulation result similar? Maybe include a panel in figure S2 in the supplement about this. Also, include a table with all fitted parameters.

3) Although of a dynamic nature, the model was not analyzed for the time-dependency of its simulations nor compared to time-course data, and the manuscript would benefit from a brief discussion, in particular with respect to the time-dependent biphasic response.

4) In Figures 2B and 2D, please highlight the boundary between nucleus and cytoplasm in the pictures, as the main difference between the two modeling results is the direction of gradient around the nucleus-cytoplasm boundary. In addition to the modeling result confined by the parameter space, please also provide an intuitive explanation of why the canonical model cannot account for the discrepancies between experiment and simulation.

5) Figure 3 on APAP regulation of YAP phosphorylation: As APAP generally induces toxicity, it makes sense to quantify cell death in the culture and limit the measurement to live cells. If possible, it will be nice to monitor the same cell population over the course of APAP or control treatment, and see how the nuclear/cytoplasmic ratio of YAP and TAZ change as a function of a cell's local environment. Also, did the proposed YAP phosphorylation in early APAP treatment (<6h) lead to YAP shuttling inside the nuclei?

6) The western immunoblot bands are hard to be interpreted by eye except when showing a binary result. Please add quantification for the plots in Figure 3C and Figure 4A, C, D, E, and F.

Reviewer #1 (Recommendations for the authors):

This is an interesting and thorough investigation of the regulation of Yap localization via its nuclear phosphorylation. While this is a novel concept and well-supported by the data provided, there are several unanswered questions that need to be resolved either by additional experiments or by an enhanced discussion as detailed below:

1) In figure 2, the authors demonstrate kinases able to phosphorylate Yap (LATS1/2) display a nuclear localization under low cell seeding conditions. However, is this localization of LATS1/2 dependent on cell density? Does a LATS1/2 knockdown or overexpression affect Yap phosphorylation and shuttling? Does APAP affect LATS1/2 localization? Answering these questions experimentally will give more credence to the claim that Yap regulatory machinery is located in the nucleus and affords functional consequences to those regulatory proteins.

2) Figure 4 shows the quantification of duolink particles for various proteins under control and APAP groups. However, the APAP treated group is not displayed. A representative image of the APAP duolink is needed.

3) From a mechanistic perspective the authors utilize an Akt inhibitor and show dephosphorylation of YAP. They also show that Akt inhibitors block APAP-induced phosphorylation. However, the authors do not determine the effects of the downstream sub-cellular localization of Yap. Determining if Akt inhibitors affect Yap localization during APAP treatment is needed to make a mechanistic link between APAP/Akt/Yap translocation.

4) The authors show APAP induces Yap translocation in vivo during DILI. While this demonstrates in vivo relevance to their in vitro findings, it does not convey any functional relevance in the modulation of the regenerative response. Does Akt inhibition affect hepatic Yap translocation and downstream survival or regeneration? Performing additional in vivo experiments modulating the Akt/ROS/Yap pathway is essential in order to draw conclusions on whether this is a viable therapeutic target in vivo during DILI.

Reviewer #2 (Recommendations for the authors):

2) 200 parametrizations were obtained. Is the simulation result similar? Maybe include a panel in figure S2 in the supplement about this. Also, include a table with all fitted parameters.

Reviewer #3 (Recommendations for the authors):

I believe the conclusions are mostly supported by the result. My comments are mostly on data presentation and interpretation:

1. In Figures 2B and 2D, please highlight the boundary between nucleus and cytoplasm in the pictures, as the main difference between the two modeling results is the direction of gradient around the nucleus-cytoplasm boundary. In addition to the modeling result confined by the parameter space, please also provide an intuitive explanation of why the canonical model cannot account for the discrepancies between experiment and simulation.

2. Figure 3 on APAP regulation of YAP phosphorylation: As APAP generally induces toxicity, it makes sense to quantify cell death in the culture and limit the measurement to live cells. If possible, it will be nice to monitor the same cell population over the course of APAP or control treatment, and see how the nuclear/cytoplasmic ratio of YAP and TAZ change as a function of a cell's local environment. Also, did the proposed YAP phosphorylation in early APAP treatment (<6h) lead to YAP shuttling inside the nuclei?

3. The western immunoblot bands are hard to be interpreted by eye except when showing a binary result. Please add quantification for the plots in Figure 3C and Figure 4A, C, D, E, and F.

Finally, it would be nice if the authors could discuss (1) the role of nuclear phosphorylation in cell proliferation, tissue damage, and regeneration – i.e. whether the YAP/TAZ localization is the cause or consequence of cell density change, and (2) the biological implication of context-dependent effect of APAP in early and late treatment.

https://doi.org/10.7554/eLife.78540.sa1

Author response

Essential revisions:

1) The model features very different diffusion rates for the canonical and alternative model, but is not explained why. For the estimation, different parameter ranges were specified for the canonical and alternative model, also for other parameters to be estimated, e.g. phosphorylation and dephosphorylation rates. This also lacks an explanation. What is the rationale behind this? Please update the discussion and/or methods to clarify this issue.

This question is answered in detail below (see Reviewer 2, comment #1).

2) 200 parametrizations were obtained. Is the simulation result similar? Maybe include a panel in figure S2 in the supplement about this. Also, include a table with all fitted parameters.

This question is answered in detail below (see Reviewer 2, comment #2).

3) Although of a dynamic nature, the model was not analyzed for the time-dependency of its simulations nor compared to time-course data, and the manuscript would benefit from a brief discussion, in particular with respect to the time-dependent biphasic response.

This question is answered in detail below (see Reviewer 2, our reply to the general comment as well as comment #3).

4) In Figures 2B and 2D, please highlight the boundary between nucleus and cytoplasm in the pictures, as the main difference between the two modeling results is the direction of gradient around the nucleus-cytoplasm boundary. In addition to the modeling result confined by the parameter space, please also provide an intuitive explanation of why the canonical model cannot account for the discrepancies between experiment and simulation.

This question is answered in detail below (see Reviewer 3, comment #1).

5) Figure 3 on APAP regulation of YAP phosphorylation: As APAP generally induces toxicity, it makes sense to quantify cell death in the culture and limit the measurement to live cells. If possible, it will be nice to monitor the same cell population over the course of APAP or control treatment, and see how the nuclear/cytoplasmic ratio of YAP and TAZ change as a function of a cell's local environment. Also, did the proposed YAP phosphorylation in early APAP treatment (<6h) lead to YAP shuttling inside the nuclei?

This question is answered in detail below (see Reviewer 3, comment #2).

6) The western immunoblot bands are hard to be interpreted by eye except when showing a binary result. Please add quantification for the plots in Figure 3C and Figure 4A, C, D, E, and F.

This question is answered in detail below (see Reviewer 3, comment #3).

Reviewer #1 (Recommendations for the authors):

This is an interesting and thorough investigation of the regulation of Yap localization via its nuclear phosphorylation. While this is a novel concept and well-supported by the data provided, there are several unanswered questions that need to be resolved either by additional experiments or by an enhanced discussion as detailed below:

1) In figure 2, the authors demonstrate kinases able to phosphorylate Yap (LATS1/2) display a nuclear localization under low cell seeding conditions. However, is this localization of LATS1/2 dependent on cell density? Does a LATS1/2 knockdown or overexpression affect Yap phosphorylation and shuttling? Does APAP affect LATS1/2 localization? Answering these questions experimentally will give more credence to the claim that Yap regulatory machinery is located in the nucleus and affords functional consequences to those regulatory proteins.

We thank reviewer for the suggestions. The mechanistic connection between LATS1/2 kinases and the Hippo pathway effectors has been intensively investigated in different species and cell types. To further emphasize the relevance of the LATS1/2-YAP connection in nuclei several experiments were initiated.

(A) To answer the question whether LATS1/2 localization is dependent on cell density, we performed nuclear-cytoplasmic fractionation experiments for low and high cell density conditions (Author response image 1A). Indeed, cell density conditions affected LATS1 and LATS2 localization differently. The amount of nuclear LATS1 protein was increasing at high cell density (NCR 1.4 at low cell density to 4.1 at high cell density). In contrast, LATS2, which was predominantly localized in the cytoplasm, did not shuttle in a comparable manner. Phosphorylated LATS1/2 (pLATS1/2) increased in the nucleus at high cell density conditions (NCR from 0.7 to 1.1), indicating that the LATS-dependent regulatory processes take place at high cell density in cell nuclei. To sum up, these experiments demonstrate that nuclear LATS abundancy is indeed regulated by cell density.

(B) To answer the question whether LATS knockdown affects YAP phosphorylation and shuttling, we inhibited LATS1/2 expression using gene-specific siRNAs targeting LATS1 and LATS2 (Author response image 1B). As illustrated by Western blot quantification, LATS2 silencing reduced phosphorylation of YAP about 30%, whereas LATS1 knockdown did not significantly alter pYAP abundance. Thus, both LATS proteins contribute YAP in our experimental setup differentially.

(C) We further quantitatively explored the localization of YAP after siLATS treatment using live-cell imaging (Author response image 1C). Data illustrated significantly increased NCR for YAP in siLATS treated cells versus controls cells at low and high cell density conditions (see quantification).

Together with the data we showed in the first version of the manuscript (nuclear localization of LATS1/2 and physical interaction between active LATS1/2 and YAP in nuclei), we conclude that nuclear LATS is indeed an important player in regulating Hippo pathway.

Author response image 1

Download asset Open asset

Western blot of nuclear and cytoplasmic protein fractions derived from Hep3B cells under low and high cell density conditions (5x10⁵ for low cell density and 3x10⁶ cells for high cell density per 10 cm dish).

Quantification of the NCR for LATS1, LATS2, and phosphorylated LATS (pLATS1/2) under low and high cell density is depicted on the right. C: cytoplasmic fraction; N: nuclear fraction. (B): Inhibition of LATS proteins in Hep3B cells using siRNAs at different concentrations for 4 hr (20 and 40 nM; (siLATS1, 5’-CCA GAA GGA UAU AGA CAA AdTdT-3’; siLATS2, 5’-CAU GAA GAC CCU AAG GAA AdTdT-3’). Quantification of LATS knockdown efficiency and effect on YAP phosphorylation is shown on the right. Ctrl: untreated control, NTC: no template control. Cytoplasmic (C) and nuclear (N) protein fraction. (C): Live-cell imaging of Hep3B cells after LATS inhibition (siLATS1/2) and NTC (no template control) treatment (left). Scale bar 50 µm. Quantification of the NCR between NTC (n=60) and siLATS (n=55) with respect to cell count per image (representing cell density). Shaded area around the green and yellow curves represent 96% confidence interval of the fitted log(x) function. Statistics: two-tailed Mann-Whitney test with p-value<0.0001.

(D) Finally, we aimed to answer the question whether APAP affects LATS localization in our model system (Author response image 2). In cytoplasmic-nuclear fractionation experiments we observed that APAP treatment causes a moderate decrease in nuclear fraction of LATS1. Here, LATS2 localization was strongly influenced by APAP treatment, especially after 48 hours of treatment (NCR reduced about 60% in comparison to control). In addition, the localization of the phosphorylated LATS isoform (pLATS1/2) was also regulated by APAP. In general, APAP treatment led to reduced pLATS in nuclei and cytoplasm. Taken together, this data illustrates that APAP treatment reduced total amount of LATS1, LATS2 and pLATS1/2 proteins and also induced the cytoplasmic shift for those proteins. Thus, LATS proteins are less active after APAP administration and may contribute to the observed less efficient YAP phosphorylation.

Author response image 2

Download asset Open asset

Western blotting of cytoplasmic (C) and nuclear (N) fraction 2 and 48 hours after APAP (10 mM) and control (ctrl) treatment in Hep3B cells.

The localization of LATS1, LATS2 and pLATS1/2 proteins was detected. PARP and Tubulin served as loading controls for nucleus and cytoplasm, respectively. The quantification of the NCR of LATS1, LATS2 and pLATS1/2 shown on the right. Western blots were obtained from one experiment.

In summary, the performed experiments confirming that nuclear LATS1/2 protein are of importance for YAP shuttling. In addition, APAP (at least partly) may contribute to the subcellular localization of YAP via the regulation/exclusion of LATS1/2 from hepatocellular cell nuclei.

Because this information is of relevance for the readership, a clarifying statement was added in the revised version of the manuscript: " Indeed, we and others showed that nuclear LATS1/2 can control YAP phosphorylation and localization (data not shown) (Ege et al., 2018; Li et al., 2014). The relevance of this nuclear interaction is substantiated by a proximity ligation assay (PLA), which illustrated that pLATS1/2 physically interacted with YAP not only in the cytoplasm but also in the nucleus (Figure 2G)." Page 7, lines 251-256

2) Figure 4 shows the quantification of duolink particles for various proteins under control and APAP groups. However, the APAP treated group is not displayed. A representative image of the APAP duolink is needed.

We thank the reviewer for the suggestion. According to the comment, we now added representative images for the experiment in Figure 4I to the manuscript.

3) From a mechanistic perspective the authors utilize an Akt inhibitor and show dephosphorylation of YAP. They also show that Akt inhibitors block APAP-induced phosphorylation. However, the authors do not determine the effects of the downstream sub-cellular localization of Yap. Determining if Akt inhibitors affect Yap localization during APAP treatment is needed to make a mechanistic link between APAP/Akt/Yap translocation.

We thank the reviewer for the suggestion. Indeed, the mechanistic link between AKT and YAP was not intensively investigated in the first manuscript draft.

Before we present/discuss our new results, we want to emphasize that long-term experiments with AKTi (24 or 48 hours) were difficult to perform due to (unspecific?) cell toxic effects of the inhibitor. Instead, all experiments were performed at early time points (up to 3 hours), which is of special importance when interpreting those experiments where APAP is equally given (early APAP is leading to YAP phosphorylation, see figure 3C of main manuscript).

Since the reviewer asked for the effects of AKTi on YAP localization, we also performed short- and long-term life cell imaging experiments using fluorescently-tagged YAP after APAP and AKTi treatment. However, it turned out that the results couldn’t be analyzed for technical reasons. In detail, combined APAP/AKTi treatment disintegrated the signal of the nuclear marker (H2B-tagged cerulean), which prevented the quantitative analysis of nuclear and cytoplasmic YAP (Author response image 3C). Indeed, the AKT-dependent regulation of DNA fragmentation and histone deacetylase proteins has been described before (Ahn et al., 2006; Ozaki et al., 2010). Because the impaired nuclear signal quality led to unreliable interpretation of the live-cell imaging results, these results cannot be used to substantiate our findings.

Based on the data derived from experiments shown in Author response image 3A/B we conclude that AKTi leads to a (moderate) nuclear localization of YAP early (!) after APAP treatment. Thus, blocking AKT at early time points after APAP treatment prevented YAP phosphorylation (which is observed at early time-points) and therefore reduced nuclear exclusion.

We adjusted a statement in the revised manuscript: "Indeed, blocking AKT activity by AKTi reduced YAP phosphorylation, moderately elevated nuclear YAP positivity, and increased the physical interaction between AKT and YAP in cell nuclei (Figure 4F and data not shown)." Page 9, line 351 – page 10, line 354

Author response image 3

Download asset Open asset

(A): Left: nuclear and cytoplasmic protein fractionation after APAP (10 mM, 1h in FCS-free medium) and combined APAP and AKTi (AKT inhibitor VIII, 10 µM 3h in FCS-free medium) treatment of Hep3B cells.

PARP and Tubulin serve as loading controls for nuclear and cytoplasmic fraction, respectively. Right: quantification of the NCR of YAP after APAP and combined APAP and AKTi treatment. (B): DuoLink PLA between YAP and AKT proteins under APAP (10 mM, 2h) and combined APAP and AKTi (10 µM, 3h) treatment. Top row: nuclei (blue) and YAP-AKT interaction (red dots); bottom row: segmentation of the nuclei and PLA dots. Right: quantification of the PLA dots in nuclei for APAP (27 images) and combined APAP/AKTi treatment (31 images). Statistics: two tailed Mann-Whitney test, p-value<0.0001. Scale bar 50 µm. (C): Live-cell imaging of Hep3B cells under APAP (10 mM 2h, top row) and combined APAP and AKTi (10µM 3h, bottom row) treatment.

4) The authors show APAP induces Yap translocation in vivo during DILI. While this demonstrates in vivo relevance to their in vitro findings, it does not convey any functional relevance in the modulation of the regenerative response. Does Akt inhibition affect hepatic Yap translocation and downstream survival or regeneration? Performing additional in vivo experiments modulating the Akt/ROS/Yap pathway is essential in order to draw conclusions on whether this is a viable therapeutic target in vivo during DILI.

We thank the reviewer for the suggestion to further investigate the therapeutic relevance of AKT and YAP inhibition during APAP-induced DILI. We discussed this aspect with our collaboration partners (A. Ghallab and J. Hengstler, Department of Toxicology, Dortmund). However, we decided not to perform e.g., AKT inhibition experiments in vivo for ethical and experimental reasons:

First, the inhibitor used in our study is very potent and efficiently abrogates AKT_1/2 activity (see main Figure 4E). Although the genetic inactivation of AKT_1/2 in specific cell types such as cardiomyocytes can be investigated in vivo, the general blockade of AKT_1/2 can cause rapid lethality as illustrated in newborn and adult mice with AKT1- and AKT2-deficiency (Liu et al., 2019; Peng et al., 2003; Wang et al., 2016). Thus, it is likely that the complete inhibition of both AKT isoforms by using an efficient chemical inhibitor may cause unpredictable side effects or is not compatible with life.
Second, as seen for APAP administration, the genetic inhibition of AKT isoforms in liver tissue cause severe hepatotoxicity and liver injury (Black, 1984; Wang et al., 2016). This strongly suggests that the treatment with APAP and concomitant silencing of AKT isoforms would further intensify hepatotoxicity leading to the unpredictable behavior of YAP expression and localization.

Due to these obstacles, we believe that these experiments are ethically problematic and very likely do not provide generate meaningful results.

However, we decided to perform additional mouse experiments to substantiate our experimental conclusions that AKT supports YAP activity in hepatocytes. For this, we did hydrodynamic gene delivery experiments in vivo with vectors coding for constitutively active myristoylated AKT (myrAKT) (Luiken et al., 2020). According to our hypothesis, permanent myrAKT activity should support nuclear YAP enrichment followed by the induction of typical transcriptional target genes (e.g., Ctgf, Cyr61). Indeed, immunohistochemical (IHC) tissue staining of extracted mice liver illustrated not only illustrated elevated AKT expression followed by hepatocellular steatosis (Wang et al., 2015) but also nuclear YAP positivity. Moreover, both investigated target genes were induced at the transcript level in mice after myrAKT injection (Author response image 4A/B).

Author response image 4

Download asset Open asset

(A): Immuno-histochemical staining of AKT (left) and YAP (right) of tissues derived from FVB mice after hydrodynamic injection of expression vectors coding myrAKT.

Experiments and ethical aspects have been published previously (Luiken *et al.*, 2020). Scale bar 50 µm. (B): Quantitative polymerase chain reaction (qPCR) measurements of murine Hippo target genes Ctgf and Cyr61 (biological replicates: n_control = 3, n_myrAKT = 4). Statistics: unpaired two tailed t test; p-value_Ctgf = 0.049, p-value_Cyr61 = 0.026.

The results confirm that AKT support nuclear YAP enrichment. To keep the manuscript concise, we suggest to include the following clarifying statement in the manuscript: "The mechanistic connection between AKT activity and YAP induction in murine hepatocytes was confirmed in independent experiments. Here hydrodynamic gene delivery of myristoylated AKT led to nuclear enrichment of YAP expression in hepatocytes and expression of typical target genes (data not shown)." in paragraph ' Sequential activation of ROS, AKT, and Hippo/YAP in mouse livers after APAP intoxication'. Page 11, lines 402-407

Author response table 1

Murine primers used for qPCR.

Expression of Cyr61 and Ctgf were normalized to a set of house-keeping genes: Actin, Gapdh, Tbp, Tubulin, Hprt, Ppia (software: GeNorm).

Gene	Forward	Reverse
Actin	GCTTCTTTGCAGCTCCTTCGT	ACCAGCGCAGCGATATCG
Gapdh	TGTCCGTCGTGGATCTGAC	CCTGCTTCACCACCTTCTTG
Tbp	TTGTCTGCCATGTTCTCCTG	CAGGGTGATTTCAGTGCAGA
Tubulin	TCACTGTGCCTGAACTTACC	GGAACATAGCCGTAAACTGC
Ctgf	GGAGAACTGTGTACGGAGCG	CCAGGCAAGTGCATTGGTA
Hprt	TCCTCCTCAGACCGCTTTT	CCTGGTTCATCATCGCTAATC
Ppia	GCATACAGGTCCTGGCATCT	AGCTGTCCACAGTCGGAAAT
Cyr61	GATCTGTGAAGTGCGTCCTTGTGG	GACACTGGAGCATCCTGCATAAG

Reviewer #2 (Recommendations for the authors):

1) The model features very different diffusion rates for the canonical and alternative model, but is not explained why. For the estimation, different parameter ranges were specified for the canonical and alternative model, also for other parameters to be estimated, e.g. phosphorylation and dephosphorylation rates. This also lacks an explanation. What is the rationale behind this?

This comment/question involves two aspects: 1) Why are parameters that should be the same physical constant in both models different (e.g. diffusion coefficients)? 2) Why did we choose different value ranges for the fitting procedure in the different models?

In this work it was not our intention to exactly determine the parameter values. Indeed, all parameters are known to be structurally unidentifyable since only stationary data with arbitrary concentration units is used for the fitting. Choosing a different time scale for the model would change all parameter values while still resulting in the same steady state concentration pattern. This means the actual parameter values are de facto arbitrary. Instead, we focus on the qualitative result of our modeling: one version of the model cannot fit the data (given a wide range of allowed parameter values) while the other version can. Identical diffusion constants in both model versions would not strengthen our conclusion. Specifically, we have shown that the simple model without the same diffusion constants cannot reproduce the observations. Thus, adding an additional restriction on the diffusion constants will definitely not lead to a better fitting model.

The reason for different parameter ranges in both models is the numeric stability of the PDE solver for the canonical model. Extreme numeric values for the diffusion constants lead to very stiff models, which complicate PDE simulations. Therefore, parameter values for the canonical model were chosen so that the complete range could be reliably simulated for parameter fitting. As explained above, the actual parameter values are not identifyable and not the scope of this study. It is important that the range is big enough that a good fit for the simple model was not arbitrarily excluded. From our experience we are convinced that this condition is satisfied. We agree that identical ranges for both models would be 'nicer' and future versions of our simulator software may offer this possibility. However, for the reasons mentioned above we are convinced that our results are sound and can explain the biological observations sufficiently.

We want to emphasize that even in the absence of non-identifyability the parameter values resulting from a 'failed' fitting procedure are not physically meaningful. We have shown with extensive parameter estimation runs that we were not able to generate at least one single satisfying fit for the canonical model. Any specific parameter values from these unsuccessful fitting runs, even if not affected by structural non-identifyability, are inherently meaningless.

In order to clarify the discrepancies between parameters of canonical and alternative models, we added the following paragraph to our methods sections:

“The parameter ranges of our canonical and alternative models were selected in a way to describe biologically meaningful value ranges and to avoid numeric instability during PDE simulations of the canonical model.” Page 25, lines 842-844

2) 200 parametrizations were obtained. Is the simulation result similar? Maybe include a panel in figure S2 in the supplement about this. Also, include a table with all fitted parameters.

According to the reviewer's comment, we now provide a csv file with all obtained parameter values for each alternative model and the objective functions of the models. See Supplementary File 1 (for YAP) and 2 (for TAZ).

In addition, we provide additional information on distribution of the objective function across models (Figure 2—figure supplement 1A) and parameter values of the 30 best models (Figure 2—figure supplement 1B). Here, we overlaid the values of the 30 best models (based on the objective function minimum value after 200 iterations) with the parameter values form the rest of the models. In general, this illustrates the structural non-identifyability of all parameters in the models (as discussed in the manuscript), which is not a problem since we focus on the ability of the various model variants to fit the data rather than the values of the specific parameters. Interestingly, some specific parameters, such as, phosphorylation reaction in the nucleus (both models YAP and TAZ) and degradation parameter for TAZ models, seem to cluster in the 'good fit', possibly indicating the importance of these parameter values for the fit. Lastly, we show exemplarily how a “bad fit” looks in comparison to a 'good fit' (Author response image 5)

Author response image 5

Download asset Open asset

A visual representation of one of the best models for YAP and TAZ, and the worst models.

Parameter value distribution of YAP.

3) Although of a dynamic nature, the model was not analyzed for the time-dependency of its simulations nor compared to time-course data, and the manuscript would benefit from a brief discussion, in particular with respect to the time-dependent biphasic response.

We thank the reviewer for pointing out the time-dependency as an important player in Hippo signaling dynamics. This suggestion could indeed initiate an interesting and challenging follow up study. Model fitting and parameter estimation for spatial models is challenging even for not time-dependent data (indeed not many examples exist in the published literature) and using time-dependent data for fitting is not yet implemented in our software platform.

The reviewer is correct when he/she emphasizes that simulation of the time dependency could answer complex questions of dynamic nature. However, the initial question of our study (which pathway topology explains the data) was tailored to steady state models. A more complex model with further assumptions would indeed expand the view on the Hippo pathway, however, this would require further intensive work which would not significantly affect the outcome of our study and therefore could not be in the scope of the manuscript.

To clarify our aims regarding steady state PDE spatial modelling, we added the following passage to the discussion paragraph:

“The goal of our approach was to determine whether nuclear phosphorylation of YAP plays a role in Hippo signaling, specifically in governing YAP localization. A model not including nuclear phosphorylation of YAP could not sufficiently explain the experimentally observed localization pattern and was excluded from further analyses. However, the alternative model proposed in this work, can explain the experimental data and therefore establishes a possible mechanism for YAP localization. However, more complex mechanisms regulating subcellular localization and dynamics of YAP and TAZ cannot be excluded. Therefore, our model partly explains how the Hippo pathway is organized, however, it does not aim to comprehensively describe it." Page 12, lines 430-440

Reviewer #3 (Recommendations for the authors):

I believe the conclusions are mostly supported by the result. My comments are mostly on data presentation and interpretation:

1. In Figures 2B and 2D, please highlight the boundary between nucleus and cytoplasm in the pictures, as the main difference between the two modeling results is the direction of gradient around the nucleus-cytoplasm boundary. In addition to the modeling result confined by the parameter space, please also provide an intuitive explanation of why the canonical model cannot account for the discrepancies between experiment and simulation.

We thank the reviewer for the suggesting and now added a thin contour line around the nuclear border to highlight boundaries between cytoplasm and nuclei in main Figure 2B and 2D.

In addition, we agree with the reviewer that the inappropriateness of the canonical model (and why the extended model is more appropriate) can be explained better. In detail, the measured distribution of YAP and TAZ would imply diffusion out of the nucleus due to a gradient formed from nucleus to cytosol. Since YAP/TAZ are not synthesized in the nucleus, a second chemical species with different diffusion coefficients has to be exported out of the nucleus (the phosphorylated version of the protein). This prerequisite cannot be fulfilled by the canonical model.

This is obvious considering the residuals of the canonical model that clearly show discrepancy between simulation and the experimental results. Especially in areas close to the nuclear envelope in figure 2B/D, the canonical model is underrepresenting YAP/TAZ concentrations as depicted by deep blue pixels around the nucleus. Furthermore, the canonical model does not correctly represent the nuclear distribution of YAP and TAZ. Here, nuclear YAP and TAZ are underrepresented in the nuclear center (blue pixels), which is an inherent property of the canonical model topology.

To clarify the process behind the model selection, we added/adjusted the following passages in the results chapter of our manuscript:

“For example, the model cannot sufficiently explain the distribution of YAP and TAZ in cell nuclei or at the nuclear membrane. In detail, the cell area around the nuclear envelope on the cytoplasmic side strongly underrepresented the concentration of YAP/TAZ, as indicated by the blue pixels in the residual image. Moreover, the simulated YAP/TAZ distribution pattern of the canonical model showed a decreased protein concentration in the center of the nucleus, with increasing gradient towards the nuclear envelope. This indicates that the observation is dominated by a protein that is disseminated from the nucleus and undergoes diffusion and degradation. The canonical Hippo pathway cannot explain this effect, illustrated by residuals (experimental data minus model simulation) (Figure 2B)." Page 6, lines 203-214

2. Figure 3 on APAP regulation of YAP phosphorylation: As APAP generally induces toxicity, it makes sense to quantify cell death in the culture and limit the measurement to live cells. If possible, it will be nice to monitor the same cell population over the course of APAP or control treatment, and see how the nuclear/cytoplasmic ratio of YAP and TAZ change as a function of a cell's local environment. Also, did the proposed YAP phosphorylation in early APAP treatment (<6h) lead to YAP shuttling inside the nuclei?

Topic: Toxicity

The reviewer is right when he/she underlines the cell-toxic properties of APAP, which may differ in cell culture models. To investigate this aspect in more detail, we performed three different experimentals: CellTox assay, Western blotting of cleaved PARP fragment, and live-cell imaging.

We measured time-dependent cell-toxicity after APAP (10 mM) treatment with a CellTox Green Cytotoxicity Assay (Promega, # G8742, Author response image 6A). Here, APAP treatment for 48 hours moderately increased cell toxicity about 10% on average, whereas at 24 hours post treatment slightly decreased toxicity was observed. Although, mild toxicity is detectable in our experimental setup, we can conclude that the majority of cells is still viable at all time points of our experiments.
We quantified cleaved cytoplasmic PARP 89 kDa fragments whose presence indicate cell death (Author response image 6B). This experiment confirm that the amount of cleaved PARP slightly increased after APAP treatment for 48 hours.
More important, for all live-cell imaging experiments, which are the basis for our study and mathematical modeling, only viable (attached) cell populations with clear nuclear/cytoplasmic structure were used (see main Figures 3 and Figure 3—figure supplement 1A/B). Thus, quantification of the nuclear to cytoplasmic ratio was exclusively performed on viable cells, since rounded and floating (dead) cells were excluded from the classification algorithm.

These results illustrate a weak and neglectable role of APAP on cell toxicity and cell death in our experimental setup. Because the chosen live cell approach excludes damaged and/or apoptotic cells, we conclude that our experimental setup allows an unbiased view on APAP-induced effects on cell signaling.

An explanatory statement summarizing the results is now included in the revised manuscript: "Severe effects of APAP on cell toxicity and apoptosis in the chosen experimental setup were excluded by measuring cell viability and PARP cleavage (data not shown)." in paragraph 'APAP regulates YAP protein localization and activity'. Page 8, lines 281-283

Topic: early response

As to early APAP treatment (<6 hours), we indeed did not explicitly mention YAP localization in our previous manuscript, since we believe that 'later' time points better describe the biologically relevant APAP responses of a cell.

To answer the reviewer's question, we approached this aspect with two different experimental techniques: live-cell imaging and Western immunoblotting (Author response image 6B and 6C, respectively).

Live-cell imaging illustrates no significant nuclear/cytoplasmic shift of YAP after APAP treatment for 2 hours (Author response image 6C). In this experimental setup, we also investigated different cell density conditions to exclude this aspect as important parameter.
(B)These results indicate no or very weak cytoplasmic translocation of YAP. A possible explanation for this mild effect after short-term treatment of APAP might be the existence of molecular mechanisms (e.g., phosphorylation of alternative serine residues of YAP), which are not part of the current model. Due to the weak character of the results and the fact that we cannot offer a reasonable explanation for this observation, we would appreciate not to include these results in the revised version of the manuscript.

However, this aspect was already mentioned in the first submitted version of the manuscript. We wrote: " No obvious response was detectable for YAP and TAZ at earlier time points post APAP treatment (data not shown)." Page 8, lines 298-281

Author response image 6

Download asset Open asset

(A): CellTox assay quantifying non-viable cells in Hep3B cells after APAP (10 mM) treatment for 24 and 48 hours (n=10, technical replicates, independent biological replicates are not shown).

Statistics: two-tailed unpaired t test, p-value<0.0001. (B): Nuclear/cytoplasmic protein fractioning of Hep3B cells after APAP (10mM) treatment for 2 and 48 hours. Total (116 kDa) and cleaved PARP (89 kDa) were detected. Tubulin served as fractionation control. Quantification of cleaved PARP fragments is shown below. C: cytoplasmic; N: nuclear fraction. (C): Live-cell imaging of Hep3B cells expressing Venus-tagged YAP 2 hours after APAP (10 mM) or control treatment (PBS). Exemplary fluorescent images (left) and data quantification of NCR as a function of cell count per image (right) are shown. One (out of three) representative experiment is shown (n_ctrl = 68; n_APAP = 68). Scale bar 50 µm. (D): Nuclear/cytoplasmic protein fractionation after APAP treatment (10 mM, 2 hours). Detection of YAP and pYAP is shown. PARP and tubulin serve as loading controls. Quantification of NCR between YAP and pYAP is shown (right). C: cytoplasmic; N: nuclear fraction. Western immunoblots in (B) and (D) were obtained from one experiment.

3. The western immunoblot bands are hard to be interpreted by eye except when showing a binary result. Please add quantification for the plots in Figure 3C and Figure 4A, C, D, E, and F.

According to the reviewer's suggestion, we now included quantification of all blots of the revised version of the manuscript (Figure 3—figure supplement 2).

Finally, it would be nice if the authors could discuss (1) the role of nuclear phosphorylation in cell proliferation, tissue damage, and regeneration – i.e. whether the YAP/TAZ localization is the cause or consequence of cell density change, and (2) the biological implication of context-dependent effect of APAP in early and late treatment.

According to the reviewer's suggestion, two additional paragraphs were included in the Discussion chapter of the revised manuscript. The first paragraph focuses on the aspect how nuclear YAP phosphorylation can extent our current thinking about Hippo pathway biology. Whereas the second paragraph discusses our findings in the context of biological/medical implications after long-term APAP ingestion. For both paragraphs additional references were included in the manuscript.

“The extension of the canonical Hippo pathway model with processes that control active nuclear YAP phosphorylation illustrates that the regulatory YAP/TAZ phosphorylation must be considered as continuum process (Shreberk-Shaked and Oren, 2019). For instance, the existence of nuclear phosphorylation possibilities by LATS-dependent and independent (e.g., by AKT) mechanisms increases the complexity but also flexibility and redundancy of the signaling pathway (Ege et al. 2018, Gao et al. 2017, Low et al. 2017). These cellular aspects are part of a fine-tuned regulatory network that allows a rapid and adjustable proliferative response of YAP and TAZ under diverse physiological and pathological cellular conditions (e.g., upon induction of tissue regeneration).” Page 13, lines 483-493

“In our manuscript we show that APAP administration induces a chain of molecular events through the APAP/ROS/AKT axis, which is leading to YAP inactivation (early) followed by YAP activation (late). The late induction of YAP activity might be considered as cell protective cellular response upon long-term tissue damage; however, YAP also acts as a potent oncogene in different cell types, including hepatocytes, and its overexpression is associated with tumor initiation (Dong et al. 2007; Weiler et al. 2017). Thus, it is tempting to speculate that APAP overdose may not only cause DILI but also increases the risk for liver tumor development. Although controversially discussed in the literature, a recent evaluation of 139 published epidemiologic studies strongly argues against a relationship between APAP exposure and cancer (Weinstein et al. 2021). Actually, APAP has been discussed as therapeutic option for patients with YAP-induced cancer (Poudel et al. 2021). Although, our and other studies clearly demonstrate a mechanistic link between APAP update and Hippo/YAP pathway activity, these partly controversial findings illustrate the necessity to decipher this connection under distinct disease conditions in the future." Page 15, lines 538-554

References

Ahn JY, Liu X, Liu Z, Pereira L, Cheng D, Peng J, Wade PA, Hamburger AWYe K. 2006. Nuclear Akt associates with PKC-phosphorylated Ebp1, preventing DNA fragmentation by inhibition of caspase-activated DNase. EMBO J 25:2083-2095. doi:10.1038/sj.emboj.7601111.

Black M. 1984. Acetaminophen hepatotoxicity. Annu Rev Med 35:577-593. doi:10.1146/annurev.me.35.020184.003045.

Ege N, Dowbaj AM, Jiang M, Howell M, Hooper S, Foster C, Jenkins RPSahai E. 2018. Quantitative Analysis Reveals that Actin and Src-Family Kinases Regulate Nuclear YAP1 and Its Export. Cell Syst 6:692-708 e613. doi:10.1016/j.cels.2018.05.006.

Li W, Cooper J, Zhou L, Yang C, Erdjument-Bromage H, Zagzag D, Snuderl M, Ladanyi M, Hanemann CO, Zhou P, Karajannis MAGiancotti FG. 2014. Merlin/NF2 loss-driven tumorigenesis linked to CRL4(DCAF1)-mediated inhibition of the hippo pathway kinases Lats1 and 2 in the nucleus. Cancer Cell 26:48-60. doi:10.1016/j.ccr.2014.05.001.

Liu W, Jing ZT, Xue CR, Wu SX, Chen WN, Lin XJLin X. 2019. PI3K/AKT inhibitors aggravate death receptor-mediated hepatocyte apoptosis and liver injury. Toxicol Appl Pharmacol 381:114729. doi:10.1016/j.taap.2019.114729.

Luiken S, Fraas A, Bieg M, Sugiyanto R, Goeppert B, Singer S, Ploeger C, Warsow G, Marquardt JU, Sticht C, De La Torre C, Pusch S, Mehrabi A, Gretz N, Schlesner M, Eils R, Schirmacher P, Longerich TRoessler S. 2020. NOTCH target gene HES5 mediates oncogenic and tumor suppressive functions in hepatocarcinogenesis. Oncogene 39:3128-3144. doi:10.1038/s41388-020-1198-3.

Ozaki K, Kosugi M, Baba N, Fujio K, Sakamoto T, Kimura S, Tanimura SKohno M. 2010. Blockade of the ERK or PI3K-Akt signaling pathway enhances the cytotoxicity of histone deacetylase inhibitors in tumor cells resistant to gefitinib or imatinib. Biochem Biophys Res Commun 391:1610-1615. doi:10.1016/j.bbrc.2009.12.086.

Peng XD, Xu PZ, Chen ML, Hahn-Windgassen A, Skeen J, Jacobs J, Sundararajan D, Chen WS, Crawford SE, Coleman KGHay N. 2003. Dwarfism, impaired skin development, skeletal muscle atrophy, delayed bone development, and impeded adipogenesis in mice lacking Akt1 and Akt2. Genes Dev 17:1352-1365. doi:10.1101/gad.1089403.

Shreberk-Shaked MOren M. 2019. New insights into YAP/TAZ nucleo-cytoplasmic shuttling: new cancer therapeutic opportunities? Mol Oncol 13:1335-1341. doi:10.1002/1878-0261.12498.

Wang Q, Yu WN, Chen X, Peng XD, Jeon SM, Birnbaum MJ, Guzman GHay N. 2016. Spontaneous Hepatocellular Carcinoma after the Combined Deletion of Akt Isoforms. Cancer Cell 29:523-535. doi:10.1016/j.ccell.2016.02.008.

Wang YH, Viscarra J, Kim SJSul HS. 2015. Transcriptional regulation of hepatic lipogenesis. Nat Rev Mol Cell Bio 16:678-689. doi:10.1038/nrm4074

https://doi.org/10.7554/eLife.78540.sa2

Article and author information

Author details

Lilija Wehling
1. Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany
2. Department of Modeling of Biological Processes, COS Heidelberg/BioQuant, Heidelberg University, Heidelberg, Germany
Contribution
Conceptualization, Data curation, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-8697-5348
Liam Keegan

Department of Modeling of Biological Processes, COS Heidelberg/BioQuant, Heidelberg University, Heidelberg, Germany

Contribution
Software, Writing – review and editing

Competing interests
No competing interests declared
Paula Fernández-Palanca
1. Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany
2. Institute of Biomedicine (IBIOMED), University of León, León, Spain
3. Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Contribution
Validation, Investigation, Methodology, Writing – review and editing

Competing interests
No competing interests declared
Reham Hassan
1. Leibniz Research Centre for Working Environment and Human Factors, Department of Toxicology, Technical University Dortmund, Dortmund, Germany
2. Department of Forensic Medicine and Toxicology, Faculty of Veterinary Medicine, South Valley University, Qena, Egypt
Contribution
Resources, Investigation, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-6569-7676
Ahmed Ghallab
1. Leibniz Research Centre for Working Environment and Human Factors, Department of Toxicology, Technical University Dortmund, Dortmund, Germany
2. Department of Forensic Medicine and Toxicology, Faculty of Veterinary Medicine, South Valley University, Qena, Egypt
Contribution
Resources, Investigation, Project administration, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-0695-3403
Jennifer Schmitt

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared
Yingyue Tang

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared
Maxime Le Marois

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Investigation, Writing – review and editing

Competing interests
No competing interests declared
Stephanie Roessler

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Formal analysis, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Peter Schirmacher

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Conceptualization, Resources, Funding acquisition, Writing – review and editing

Competing interests
No competing interests declared
Ursula Kummer

Department of Modeling of Biological Processes, COS Heidelberg/BioQuant, Heidelberg University, Heidelberg, Germany

Contribution
Conceptualization, Resources, Project administration, Writing – review and editing

Competing interests
No competing interests declared
Jan G Hengstler

Leibniz Research Centre for Working Environment and Human Factors, Department of Toxicology, Technical University Dortmund, Dortmund, Germany

Contribution
Resources, Project administration, Writing – review and editing

Competing interests
No competing interests declared
Sven Sahle

Department of Modeling of Biological Processes, COS Heidelberg/BioQuant, Heidelberg University, Heidelberg, Germany

Contribution
Resources, Software, Supervision, Funding acquisition, Writing – original draft, Project administration, Writing – review and editing

Contributed equally with
Kai Breuhahn

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-5458-7404
Kai Breuhahn

Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany

Contribution
Conceptualization, Resources, Supervision, Funding acquisition, Methodology, Writing – original draft, Project administration, Writing – review and editing

Contributed equally with
Sven Sahle

For correspondence
Kai.Breuhahn@med.uni-heidelberg.de

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-2462-1229

Funding

German Federal Ministry of Education and Research (031L0074H)

Kai Breuhahn

Deutsche Forschungsgemeinschaft (505755359)

Kai Breuhahn

Deutsche Forschungsgemeinschaft (314905040)

Kai Breuhahn

German Federal Ministry of Education and Research (031L0158)

Liam Keegan
Ursula Kummer
Sven Sahle

Heidelberg University (Research training group "Mathematical Modeling for the Quantitative Biosciences (MMQB)")

Lilija Wehling

Deutsche Forschungsgemeinschaft (GH 276)

Ahmed Ghallab

Ministerio de Ciencia e Innovación (Programa Estatal de Promoción del Talento y su Empleabilidad (FPU17/01995))

Paula Fernández-Palanca

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication.

Acknowledgements

We want to thank Patrizia Birner and Michaela Bissinger for their excellent technical support. In addition, we thank the Nikon Imaging Center at Heidelberg University (BioQuant), especially Dr. Ulrike Engel and Dr. Christian Ackermann, for supporting confocal imaging experiments. We also acknowledge the LSDF2 (URZ) for data storage. We thank Prof. Dr. Alexander Loewer (TU Darmstadt) for providing the YAP-mVenus vector and the Center for Model System and Comparative Pathology (CMCP, Heidelberg) for staining mouse tissue samples, especially Heike Conrad, Christine Schmitt, and Sarah Lammer. Tissue scanning was supported by the Tissue Bank of the National Center of Tumor Diseases (NCT, Heidelberg).

Ethics

This study was performed according to animal welfare committee of The Ministry for Environment, Agriculture, Conservation and Consumer Protection of the State of North Rhine-Westphalia (LANUV, North Rhine-Westphalia, Germany, application number: 84-02.04.2016.A279).

Senior Editor

Mone Zaidi, Icahn School of Medicine at Mount Sinai, United States

Reviewing Editor

Matthew A Quinn, Wake Forest School of Medicine, United States

Reviewers

Matthew A Quinn, Wake Forest School of Medicine, United States
Dirk Fey, University College Dublin, Ireland

Publication history

Received: March 10, 2022
Preprint posted: April 1, 2022 (view preprint)
Accepted: September 21, 2022
Version of Record published: October 18, 2022 (version 1)

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

380

Page views
83

Downloads
1

Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Mendeley

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Lilija Wehling
Liam Keegan
Paula Fernández-Palanca
Reham Hassan
Ahmed Ghallab
Jennifer Schmitt
Yingyue Tang
Maxime Le Marois
Stephanie Roessler
Peter Schirmacher
Ursula Kummer
Jan G Hengstler
Sven Sahle
Kai Breuhahn

(2022)

Spatial modeling reveals nuclear phosphorylation and subcellular shuttling of YAP upon drug-induced liver injury

eLife 11:e78540.

https://doi.org/10.7554/eLife.78540

Categories and tags

Research organisms

Human
Mouse

Share this article

Cite this article

YAP and TAZ show distinct nuclear shuttling in hepatocellular cells upon increasing cell density.

Mathematical modeling predicts that nuclear phosphorylation controls YAP/TAZ subcellular localization.

Figure 2—source data 1

APAP regulates YAP phosphorylation, localization, and expression of its target genes.

Figure 3—source data 1

APAP controls nuclear YAP enrichment via induction of reactive oxygen species (ROS) and AKT.

Figure 4—source data 1

APAP stimulates the sequential activation of reactive oxygen species (ROS), AKT, and YAP in vivo.

Canonical model parameters.

Alternative model parameters.

Parameters of the YAP-to-TAZ transition model.

Parameter range for particle swarm algorithm of the canonical YAP/TAZ model.

Parameter range for particle swarm algorithm of the alternative YAP/TAZ model.

Parameter values from the literature.

Western blot of nuclear and cytoplasmic protein fractions derived from Hep3B cells under low and high cell density conditions (5x105 for low cell density and 3x106 cells for high cell density per 10 cm dish).

Western blotting of cytoplasmic (C) and nuclear (N) fraction 2 and 48 hours after APAP (10 mM) and control (ctrl) treatment in Hep3B cells.

(A): Left: nuclear and cytoplasmic protein fractionation after APAP (10 mM, 1h in FCS-free medium) and combined APAP and AKTi (AKT inhibitor VIII, 10 µM 3h in FCS-free medium) treatment of Hep3B cells.

(A): Immuno-histochemical staining of AKT (left) and YAP (right) of tissues derived from FVB mice after hydrodynamic injection of expression vectors coding myrAKT.

Murine primers used for qPCR.

A visual representation of one of the best models for YAP and TAZ, and the worst models.

(A): CellTox assay quantifying non-viable cells in Hep3B cells after APAP (10 mM) treatment for 24 and 48 hours (n=10, technical replicates, independent biological replicates are not shown).

Author details

Lilija Wehling

Contribution

Competing interests

Liam Keegan

Contribution

Competing interests

Paula Fernández-Palanca

Contribution

Competing interests

Reham Hassan

Contribution

Competing interests

Ahmed Ghallab

Contribution

Competing interests

Jennifer Schmitt

Contribution

Competing interests

Yingyue Tang

Contribution

Competing interests

Maxime Le Marois

Contribution

Competing interests

Stephanie Roessler

Contribution

Competing interests

Peter Schirmacher

Contribution

Competing interests

Ursula Kummer

Contribution

Competing interests

Jan G Hengstler

Contribution

Competing interests

Sven Sahle

Contribution

Contributed equally with

Competing interests

Kai Breuhahn

Contribution

Contributed equally with

For correspondence

Competing interests

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Categories and tags

Research organisms

Further reading

Western blot of nuclear and cytoplasmic protein fractions derived from Hep3B cells under low and high cell density conditions (5x10⁵ for low cell density and 3x10⁶ cells for high cell density per 10 cm dish).