Impaired degradation of YAP1 and IL6ST by chaperone-mediated autophagy promotes proliferation and migration of normal and hepatocellular carcinoma cells

ABSTRACT Impaired degradation of the transcriptional coactivator YAP1 and IL6ST (interleukin 6 cytokine family signal transducer), two proteins deregulated in liver cancer, has been shown to promote tumor growth. Here, we demonstrate that YAP1 and IL6ST are novel substrates of chaperone-mediated autophagy (CMA) in human hepatocellular carcinoma (HCC) and hepatocyte cell lines. Knockdown of the lysosomal CMA receptor LAMP2A increases protein levels of YAP1 and IL6ST, without changes in mRNA expression. Additionally, both proteins show KFERQ-dependent binding to the CMA chaperone HSPA8 and accumulate into isolated lysosomes after stimulation of CMA by prolonged starvation. We further show that LAMP2A downregulation promotes the proliferation and migration in HCC cells and a human hepatocyte cell line, and that it does so in a YAP1- and IL6ST-dependent manner. Finally, LAMP2A expression is downregulated, and YAP1 and IL6ST expression is upregulated, in human HCC biopsies. Taken together, our work reveals a novel mechanism that controls the turnover of two cancer-relevant proteins and suggests a tumor suppressor function of CMA in the liver, advocating for the exploitation of CMA activity for diagnostic and therapeutic purposes. Abbreviations: ACTB: actin beta; ATG5: autophagy related 5; ATG7: autophagy related 7; CMA: chaperone-mediated autophagy; eMI: endosomal microautophagy; HCC: hepatocellular carcinoma; HSPA8: heat shock protein family A (Hsp70) member 8; IL6ST: interleukin 6 cytokine family signal transducer; JAK: Janus kinase; LAMP1: lysosomal associated membrane protein 1; LAMP2A: lysosomal associated membrane protein 2A; MAPK8: mitogen-activated protein kinase 8; P6: pyridine 6; SQSTM1: sequestosome 1; TUBA: tubulin alpha; VDAC1: voltage dependent anion channel 1; VP: verteporfin; YAP1: Yes1 associated transcriptional regulator.


Introduction
Autophagy is a catabolic mechanism responsible for the degradation of cellular components via the lysosomal pathway and plays a pivotal role in maintaining protein homeostasis and protecting cells in stress conditions such as nutrient shortage or oxidative stress [1]. In mammalian cells, three primary forms of autophagy exist, each with its distinctive features: macroautophagy, in which double-membrane structures called phagophores engulf cell components and mature into autophagosomes, which then fuse with lysosomes to degrade their cargo; microautophagy, including endosomal microautophagy (eMI), in which proteins are imported into late endosomes or lysosomes; and chaperone-mediated autophagy (CMA), in which proteins translocate into the lysosomes through LAMP2A (lysosomal associated membrane protein 2A) [2]. Together with eMI, CMA is responsible for the degradation of proteins containing a recognition sequence named the KFERQ-like motif [3]. During CMA, the KFERQ motif is recognized by the chaperone HSPA8 (heat shock protein family A (Hsp70) member 8). The substratechaperone complex reaches the lysosome where it interacts with the CMA receptor LAMP2A and translocates into the lysosomal lumen [4]. Basal levels of CMA are detectable in unstimulated cells, but higher CMA activity is observed in response to stresses such as starvation, oxidative or genotoxic stress [5]. About 30% of cytosolic proteins contain KFERQlike motifs and are therefore potential eMI or CMA substrates [6]. Aberrant degradation of these substrates can profoundly impact the cellular proteome, and reduced CMA activity has already been implicated in many pathological conditions, such as neurodegeneration, lysosomal storage disorders, and aging [5]. In cancer, the role of CMA is still controversial. LAMP2A has been observed elevated in many human cancer tissues, and CMA downregulation decreases the proliferation of lung and gastric cancer cells [7]. However, CMA also prevents fibroblast transformation by promoting the degradation of the MYC proto-oncogene [8] and many other cancer-related proteins, such as mutant TP53/p53 (tumor protein p53), LDHA (lactate dehydrogenase A), PKM (pyruvate kinase M1/2), and HK2 (hexokinase 2) are CMA substrates [9]. In the liver, CMA deficiency leads to metabolic dysregulation, hepatic glycogen depletion and hepatosteatosis [10], and restoration of CMA activity in aged mice reduces the accumulation of damaged proteins and improves liver function, highlighting hepatoprotective and possibly anti-tumor functions of CMA [11].
Hepatocellular carcinoma (HCC) is the primary liver malignancy and the third cause of cancer-related death worldwide. HCC incidence is high in Asia, Africa and Southern Europe, and it is one of the few cancer types whose mortality increased in the last decades [12]. Major risk factors for the development of HCC are persistent infections with hepatitis B and C viruses, alcohol abuse, obesity, exposure to toxins (e.g., aflatoxin B), and nonalcoholic fatty liver disease [12]. HCC is a highly heterogeneous cancer, in which aberrant activation of many oncogenic signaling pathways is observed with variable frequencies [13]. The transcriptional coactivator YAP1 (Yes1 associated transcriptional regulator) and IL6ST (interleukin 6 cytokine family signal transducer)-JAK (Janus kinase) signaling axis have a prominent role in tumor formation and growth, promoting cancer cell survival, proliferation, migration and metastasis [14,15]. They are coordinately upregulated at the post-translational level in mouse and human RAF1 knock-out HCC models, and responsible for the increased tumorigenesis induced by RAF1 ablation [16]. Impaired degradation of YAP1 and IL6ST by either the ubiquitinproteasome system or autophagy was shown to favor the growth of different tumor types [17][18][19][20][21]. In this work, we provide evidence that CMA controls protein levels of YAP1 and IL6ST, which we validate as novel CMA substrates. Consistent with the pro-tumorigenic function of YAP1 and IL6ST in HCC, we also show that inhibition of CMA enhances proliferation and migration of a hepatocyte cell line and HCC cells in a YAP1-and IL6ST-dependent manner. Our work advances the current knowledge of mechanisms regulating the expression of cancer-related proteins, laying the groundwork for developing novel and specific strategies to be employed for therapeutic purposes.

YAP1 and IL6ST are preferentially degraded by a lysosomal mechanism independent of macroautophagy in HCC cells
To identify the mechanism through which YAP1 and IL6ST are preferentially degraded in HCC cells, we treated Hep3B cells either with lysosomal inhibitors (leupeptin, NH 4 Cl, or a combination of the two) or with the proteasome inhibitor MG132. Inhibition of lysosomal activity caused a much stronger increase in IL6ST and YAP1 expression (Figures 1A and 1B) than inhibition of the proteasome (Figures 1C, 1D; the efficacy of MG132 and absence of cell death are shown in Figure S1A). Similar results were obtained in the HCC cell line HepG2 and in the human hepatocyte cell line HuS (Figures S1B and S1C). In the latter cell line, the effect of lysosomal inhibition on IL6ST was stronger than proteasome inhibition, while the inhibition of either mechanism had similar effects on YAP1, possibly reflecting a higher proteasomal turnover of YAP1 in normal than in HCC cells. In any case, these data demonstrate that lysosomes control the turnover of YAP1 and IL6ST in cells of liver origin. To verify whether YAP1 and IL6ST were degraded by macroautophagy, this process was either inhibited by siRNA against ATG5 (autophagy related 5) or ATG7 (autophagy related 7) or stimulated by rapamycin ( Figure 1E and 1F). Inhibition of macroautophagy, clearly reflected in the accumulation of SQSTM1/p62 (sequestosome 1), did not increase YAP1 and IL6ST protein levels. Conversely, induction of macroautophagy, shown by SQSTM1 degradation, did not reduce YAP1 or IL6ST levels ( Figure 1G). In contrast, inhibition of macroautophagy decreased YAP1 levels, and rapamycin treatment increased IL6ST levels. This phenomenon could be explained by a previously reported crosstalk between macroautophagy and other degradation pathways, such as CMA [22,23]. In Hep3B cells, modulation of macroautophagy had a negative impact on LAMP2A expression (Figures S1D and S1E), while LAMP2A silencing had a moderately positive effect on macroautophagy ( Figure S1F). Be that as it may, the results are consistent with the idea that the lysosomal degradation of YAP1 and IL6ST is independent of macroautophagy.

Modulation of CMA affects YAP1 and IL6ST levels
In addition to macroautophagy, lysosomes are the terminal station for other autophagic processes, such as CMA. Prolonged starvation, a condition that triggers CMA activation, led to a decrement of YAP1 and IL6ST levels in Hep3B cells and mouse livers (Figures 2A and B). Similarly, chemical stimulation of CMA by AKT1/2 (AKT serine/threonine kinase 1/2) inhibition [24] and the atypical RARA/RARα receptor (retinoic acid receptor alpha) antagonist, AR7 [25], caused a decrease of YAP1 and IL6ST ( Figures 2C and 2D). Silencing of LAMP2A, the rate-limiting factor in CMA, lead to the accumulation of YAP1 and IL6ST proteins in Hep3B cells and in the human hepatocyte cell line HuS ( Figures 2E and  S2A) without significant changes in mRNA expression in either cell line ( Figures 2F and S2B), and rescued starvationinduced decrease of both proteins (siLAMP2A #1 was used in this and all the following experiments) ( Figure 2G). To clarify the contribution of CMA in the lysosomal degradation of YAP1 and IL6ST, we inhibited lysosomal activity in LAMP2Asilenced cells and measured the expression of the two proteins. The results showed that lysosomal inhibitors further increased IL6ST but not YAP1 levels ( Figure S2C), indicating that CMA accounted for most of the lysosomal degradation of YAP1, while other lysosomal-dependent pathways appear to cooperate with CMA to regulate the turnover of IL6ST. Consistent with previous results, inhibition of the proteasome did not have any additive effect to CMA inhibition on either protein. Inhibition of lysosomal activity was also able to increase the expression of a YAP1 S127A mutant ( Figure  S2D), which is resistant to repression by the Hippo pathway [20,26]. LAMP2A knockdown had a similar effect ( Figure  S2E), while ATG7 silencing did not promote the accumulation of either construct (Figure S2F), further confirming that lysosomal-dependent degradation of YAP1 relies on CMA. CMA downregulation has a milder effect on the YAP1 S127A mutant than on the WT YAP1. Serine 127 phosphorylation is important for the retention of YAP1 into the cytoplasm [20,26]. Therefore, this result may reflect a reduced accessibility of the YAP1 S127A mutant to the CMA machinery. Although CMA has already been shown to degrade nuclear proteins (e.g. CHEK1/Chk1) [27], it may target them less efficiently than cytosolic substrates. Importantly, LAMP2A silencing in Hep3B and HuS cells promoted YAP1 nuclear translocation and increased the expression of CCN2/CTGF (cellular communication network factor 2) (YAP1 target gene) and BIRC5 (baculoviral IAP repeat containing 5) (common YAP1-IL6ST pathway target) (Figures S2G-S2J). Together, these results demonstrate that CMA modulation impacts the levels of YAP1 and IL6ST as well as their output.

YAP1 and IL6ST are bona fide substrates of CMA
Next, we investigated whether YAP1 and IL6ST were substrates of CMA. To be validated as a CMA substrate, a protein must satisfy established criteria, described elsewhere [28]. We have already shown that protein levels of YAP1 and IL6ST were increased by LAMP2A silencing and decreased by conditions activating CMA. In addition, all CMA substrates must contain a KFERQ-motif, recognized by HSPA8 and necessary for the lysosomal translocation of the substrates. Sequence analysis revealed that both YAP1 and IL6ST featured a putative KFERQ motif In all panels, the intensity of the bands of interest was normalized to TUBA prior to fold-change calculations. Data are plotted as the mean ± SEM of n = 3 biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001 vs control.
( Figures 3A and B). Mutation of the KFERQ motif reduced the interaction of YAP1 and IL6ST with HSPA8 ( Figures  3C and D). The basal expression levels of the KFERQ mutants were higher than those of the corresponding WT, although not significantly so (compare lanes 1 with lane 3, Figure 3E and 3F; see also left bottom panels). More importantly, mutation of the KFERQ motifs dramatically reduced the accumulation of YAP1 and IL6ST in response to treatment with inhibitors of lysosomal activity (NL: NH 4 Cl and leupeptin) (Figures 3E and F; compare the fold change in the expression of the WT and KFERQ mutant in cells untreated or treated with NL, right bottom panel). These data imply that CMA is the main lysosomal degradation mechanism for these two proteins. In line with this, stimulation of CMA by prolonged starvation also promoted the interaction of both proteins with LAMP2A ( Figures 3G and H) and their accumulation into the lysosomes ( Figure 3I; purity of lysosomal fraction is shown in Figure S3). Knockdown of LAMP2A, but not of ATG7, reduced starvation-induced lysosomal accumulation of YAP1 and IL6ST ( Figure 3J), further confirming that macroautophagy was not involved in this process. Overall, these data indicate that YAP1 and IL6ST are novel substrates of CMA.

LAMP2A silencing promotes proliferation and migration of normal and HCC cells in a YAP1 and IL6ST-dependent manner
Both IL6ST and YAP1 have been shown to promote tumorigenic features of tumor cells, such as proliferation and migration [15,29]. If, as indicated above, these proteins are CMA substrates, inhibition of CMA might also affect these processes. To test this hypothesis, we employed one human hepatocyte cell line, HuS, and two HCC cell lines, Hep3B and HepG2, in which CMA was downregulated by LAMP2A knockdown (Figures S4A-S4C). Live cell counting ( Figure 4A) and crystal violet staining ( Figure 4B) showed that LAMP2A downregulation significantly increased cell proliferation in HuS cells.  (Figures S4E and S4F). Thus, reducing the levels of LAMP2A promoted proliferation and migration, two hallmark features of tumor cells, suggesting an anti-tumorigenic function of LAMP2A and CMA. In support of this hypothesis, immunoblot analysis of biopsies of HCC patients clearly showed lower expression of LAMP2A in tumor tissues compared to adjacent non-tumor tissues ( Figures 4I, 4J and Figure S4G). In the same samples, the expression of YAP1 and IL6ST was instead increased, consistent with the results obtained in cultured liver cell lines.
To address whether YAP1 and IL6ST contributed to the proliferative phenotype of CMA impaired cells we tested whether silencing these genes would rescue the phenotypes of LAMP2A knockdown cells (Figures S5A and  S5B). In HuS and Hep3B cells, silencing of YAP1 or IL6ST efficiently reverted the increased proliferation of LAMP2A-knockdown cells (Figures 5A and 5C). Similar results were obtained using verteporfin (VP) and the pan JAK inhibitor pyridone 6 (P6), inhibitors of YAP1 and IL6ST-JAK signaling, respectively (Figures S5C and  S5E). Analysis of migration by the wound healing assay also showed that silencing of either YAP1 or IL6ST reverted the increased migration of LAMP2A knockdown HuS and Hep3B cells (Figures 5B and 5D). Again, similar results were obtained using VP and P6 (Figures S5D and  S5F). Collectively, these data demonstrated that YAP1 and IL6ST contributed to the positive effect of CMA downregulation on cell proliferation and migration.

Discussion
Many cellular functions, such as cell cycle progression and signal transduction, are characterized by a finely regulated fluctuation of protein levels, determined by a balance between gene transcription, protein synthesis, and protein degradation. In eukaryotic cells, the activity of the ubiquitin-proteasome system and autophagy determine protein turnover, removing defective or misfolded proteins and providing amino acids for de novo protein synthesis, thus maintaining homeostasis [30]. Loss of protein homeostasis can lead to abnormal protein expression, altering cell functions, and it is linked to a wide range of diseases, including cancer [31][32][33].
In this study, we demonstrate that CMA controls the levels of YAP1 and IL6ST, expanding the range of cancer-relevant proteins whose turnover is controlled by CMA [9]. Silencing of the CMA receptor LAMP2A increased the levels of both proteins without changes in mRNA expression, and further analysis demonstrated that YAP1 and IL6ST were direct CMA substrates, revealing a novel mechanism that controls the expression of two cancer-relevant proteins. In fact, YAP1 and IL6ST signaling pathways have a role in promoting liver tumors [29,34], and the impaired degradation of the two proteins by either the proteasome or macroautophagy has already been shown to promote tumor growth [17][18][19]21,35]. In line with the identification of YAP1 and IL6ST as CMA substrates in liver cells, we could show that CMA downregulation promoted proliferation and migration in a hepatocyte cell line and HCC cell lines. Genetic or pharmacological inhibition of YAP1 and IL6ST restrained the growth advantage of LAMP2A-knockdown cells, demonstrating that the impaired degradation of the two proteins by CMA has a significant role in the phenotype imposed by LAMP2A silencing. Interestingly, inhibition of YAP1 and IL6ST had a cell line-specific effect on the proliferation of WT cells, with the HuS cell showing little sensitivity. When LAMP2A was silenced, however, both cell lines tested responded in a similar manner to inhibition of either pathway. In light of this, LAMP2A expression may represent a predictor of cell sensitivity to specific inhibitors and eventually find application in biomarker-driven anticancer therapy. This aspect is particularly important in HCC, in which the high heterogeneity and the high number of oncogenic signaling pathways potentially involved in tumor growth make the identification of feasible therapeutic strategies challenging [36].
The demonstration that CMA modulation controls the turnover of two cancer proteins and, in turn, influences cell proliferation and migration argues for a tumor suppressor role of CMA in the liver. The role of CMA in cancer is still unclear, with numerous lines of evidence arguing for a pro-or an anti-tumorigenic function of this mechanism, depending on the cell and tumor type [7,[37][38][39]. In support of the hypothesis that CMA acts as a tumor suppressor in the liver, LAMP2A expression was reduced in HCC biopsies compared to their matched adjacent normal tissues, and an anticorrelation with YAP1 and IL6ST could be observed. A tumor suppressor function of CMA in the liver can also be envisaged from the work of the Cuervo´s group, showing that deficient CMA leads to liver metabolic dysregulation and hepatosteatosis [10], emerging as risk factors for HCC in developed countries [40]; and that restoration of CMA activity in aged mice reduces the accumulation of damaged proteins and improves liver function [11]. Stimulation of CMA could be deployed in cancer prevention to prevent aberrant protein accumulation and maintain protein homeostasis, for instance in people with a higher than average risk of developing liver cancer. In this context, activators of CMA have already been characterized in vitro and in vivo [25,41]. In particular, the mitochondrial-derived peptide humanin and its analogs were shown to protect cells from a multitude of stress conditions and exert cardioprotective and neuroprotective functions in vitro and in vivo [41,42]. CMA mediates the cytoprotective effects of humanin in vitro [41], but whether and to which extent this also occurs in vivo is still to be clarified. In conclusion, our work expands the list of cancer-related CMA substrates, advocating for the exploitation of CMA activity for diagnostic and therapeutic purposes.

Treatments
Lysosomal inhibitors NH 4 Cl (Sigma-Aldrich, A9434) and leupeptin (Sigma-Aldrich, SAE0153) were used at the final concentration of 20 mM and 50 μM, respectively. Proteasome inhibitor MG132 was purchased from Sigma-Aldrich (474791) and used at the concentrations indicated in the text. Rapamycin was purchased from Enzo Lifesciences (BML-A275-0005) and used at the concentrations indicated in the text. AKT inhibitor VIII was purchased from Sigma-Aldrich (124017) and used at the final concentration of 10 μM. The atypical RARA receptor antagonist AR7 was purchased from Sigma-Aldrich (SML0921) and used at the indicated concentrations. YAP1 inhibitor verteporfin (VP) was purchased from Sigma-Aldrich (SML0534) and used at the final concentration of 5 μM. Pan JAK inhibitor pyridone 6 (P6) was purchased by Sigma-Aldrich (420097) and used at the final concentration of 1 μM.

Patient samples
Human HCC biopsies were provided by Prof. Grazi from the Hepato-Pancreato-Biliary Surgery Unit, Department of Clinical and Experimental Oncology, Regina Elena National Cancer Institute, Rome, Italy. Analysis was performed after approval from the Regina Elena Cancer Institute ethics committee.

Animal experiments
Male mice on an Sv/129 background were food-deprived overnight, euthanized with CO 2, and livers collected. Animal experiments were authorized by the Austrian Ministry of Science, Research and Economy.

Immunoblotting and co-immunoprecipitation
Immunoblotting was carried out as previously described [16]. ) supplemented with 10 mM NaF (Sigma-Aldrich, S7920), 1 mM Na 3 VO 4 (Sigma-Aldrich, 567540) and protease inhibitor cocktail (Roche, P8849) and centrifuged at 20,000 × g for 15 min. Five hundred μg of protein lysate were incubated overnight at 4°C with 2 μg of primary antibody, followed by 4 h of incubation with 50 μl of Rec-Protein G -Sepharose® 4B beads (Thermo Scientific, 101241). An isotype control antibody (Cell Signaling Technology, 3900) was used as a negative control. Beads were washed three times in lysis buffer, incubated for 6 min at 95°C in Laemmli sample buffer (Bio-Rad Laboratories, 1610737) and subjected to SDS-PAGE. Immunoblots (representative of at least 2 experiments) were acquired using ChemiDoc Imaging Systems and quantified using the Image Lab software (Bio-Rad Laboratories, Hercules, USA).

Site-directed mutagenesis
Mutation of the KFERQ sequence of YAP1 and IL6ST was performed using the QuikChange site-directed mutagenesis kit (Stratagene, 200522), using the following primers: YAP1: 5'-CTGCGGCTGAAACAGGCCGCCCTGCTTCGG CAGGAGTTAG-3'; IL6ST: 5'-ACTGTTTCAAATCTTTCTGCTGCTCCCTCA GTACCTGGACCAAAAGC-3'. The presence of the desired mutations and the absence of additional mutations within the coding sequence was confirmed by DNA sequencing.

Live cell counting
Cells were transfected as required and re-seeded after 12 h. After 12 h (T 0h ), 36 h (T 24h ) and 60 h (T 48h ), cells were collected and live cells counted after staining with 0.08% Trypan Blue (Sigma-Aldrich, T8154). Where required, inhibitors were added at T 24 and cells counted after an additional 24 h.

Crystal violet staining
Cells transfected with either a siScr or siLAMP2A were plated in 35-mm dishes and let grow for 72 h. At the end of the experimental time, cells were fixed for 15 min with ice-cold methanol, washed with PBS (137 mM NaCl [Sigma-Aldrich, S9625], 2.7 mM KCl [Sigma-Aldrich, 104936], 10 mM Na 2 HPO 4 [Sigma-Aldrich, S0876], and 1.8 mM KH 2 PO 4 [Sigma-Aldrich, P0662], pH 7.4) and stained with 0.5% crystal violet (Sigma Aldrich, 61135) for 30 min at room temperature. After three washes with PBS, dishes were air-dried and images acquired using a digital camera. Finally, crystal violet was eluted with 10% acetic acid (Sigma-Aldrich, A6283) and quantification was performed by measuring the absorbance at 595 nm using a microplate reader.

Wound healing assay
Cells were transfected as required and let grow until confluent. Then, cells were serum-starved overnight to prevent proliferation, medium replaced with one containing 1% FBS (Euroclone, ECS0180L) and a wound created using a pipette tip. Images were taken using an inverted microscope connected with a digital camera at the beginning (T 0 ) and end of the experimental time (T 12-24h ). Quantification of gap closure was performed using the ImageJ software. In selected experiments inhibitors were added 1 h before T 0 .

Transwell migration assay
Cells were transfected with either a siScr or siLAMP2A and let grow for 48 h, after which time they were trypsinized, counted and resuspended in serum-free medium. SiScr and siLAMP2A cells (1 x 10 5 ) were transferred to transwell inserts with 8-μm pore size (VWR, 734-1574). The outer chambers were filled with complete medium. After 16 h, cells were fixed with 10% formalin solution (Sigma-Aldrich, HT5011) and stained with 0.5% crystal violet (Sigma Aldrich, 61135) for 30 min at room temperature. Non-migrated cells were removed with a cotton swab and images were taken using an inverted microscope connected with a digital camera. Migrated cells were counted using the ImageJ software.

Statistical analysis
Values are expressed as means ± SEM of at least 3 independent experiments; unless otherwise indicated, p values were calculated using the two-tailed Student's t-test. A p-value < 0.05 is considered statistically significant and a p-value < 0.01 is considered highly statistically significant.

Disclosure statement
No potential conflict of interest was reported by the author(s).