Rubicon prevents autophagic degradation of GATA4 to promote Sertoli cell function.

Autophagy degrades unnecessary proteins or damaged organelles to maintain cellular function. Therefore, autophagy has a preventive role against various diseases including hepatic disorders, neurodegenerative diseases, and cancer. Although autophagy in germ cells or Sertoli cells is known to be required for spermatogenesis and male fertility, it remains poorly understood how autophagy participates in spermatogenesis. We found that systemic knockout mice of Rubicon, a negative regulator of autophagy, exhibited a substantial reduction in testicular weight, spermatogenesis, and male fertility, associated with upregulation of autophagy. Rubicon-null mice also had lower levels of mRNAs of Sertoli cell-related genes in testis. Importantly, Rubicon knockout in Sertoli cells, but not in germ cells, caused a defect in spermatogenesis and germline stem cell maintenance in mice, indicating a critical role of Rubicon in Sertoli cells. In mechanistic terms, genetic loss of Rubicon promoted autophagic degradation of GATA4, a transcription factor that is essential for Sertoli cell function. Furthermore, androgen antagonists caused a significant decrease in the levels of Rubicon and GATA4 in testis, accompanied by elevated autophagy. Collectively, we propose that Rubicon promotes Sertoli cell function by preventing autophagic degradation of GATA4, and that this mechanism could be regulated by androgens.

Autophagy is an intracellular membrane trafficking pathway that governs metabolic turnover via degradation of cytoplasmic constituents, thereby maintaining cellular homeostasis in various cell types [16,17]. Notably in this regard, our recent studies showed that Rubicon negatively regulates autophagy by interaction with PI3K complex that is essential for autophagy [18,19], and that loss of Rubicon ameliorates a variety of age-related diseases by upregulating autophagy [20,21]. As shown previously, autophagy regulates acrosome biogenesis and spermatid differentiation [22,23]. Autophagy in Sertoli cells is essential for ectoplasmic specialization assembly [24], and LC3-associated phagocytosis [25]. Although it is clear that autophagy is required for Sertoli cell homeostasis, it remains to be determined how Sertoli cell autophagy participates in spermatogenesis.
In this study, we found that Rubicon-null mice exhibited defective spermatogenesis and male subfertility, accompanied by upregulation of autophagy in testes. Importantly, genetic loss of Rubicon in Sertoli cells, but not in germ cells, caused defective spermatogenesis and promoted autophagic degradation of GATA4, which is crucial for Sertoli cell function. Furthermore, an antagonist of androgens, which are male steroid hormones, significantly decreased the levels of Rubicon and GATA4 in testes. On the basis of these findings, we propose that Rubicon inhibits autophagic degradation of GATA4 to promote Sertoli cell function, which could be maintained by androgen.

Rubicon is required for spermatogenesis and male fertility
In previous work, we showed that upregulation of autophagy by loss of Rubicon ameliorates age-related diseases, such as fatty liver, neurodegeneration, and renal fibrosis [20,21]. On the other hand, loss of Rubicon causes adipose tissue dysfunction due to excess autophagy [26]. Hence, to clarify the roles of Rubicon in other tissues, we examined systemic Rubicon knockout mice generated in a previous study [20]. Surprisingly, the knockout mice exhibited a significant reduction in testis weight (Fig 1A), accompanied by loss of Rubicon in testis (S1A Fig), suggesting an unexpected role of Rubicon in this organ. Histological analysis revealed that systemic Rubicon knockout mice had defective spermatogenesis (Fig 1B and 1C). Some of the knockout mice exhibited a more severe defect in testis (S1B Fig). In addition, the knockout mice had higher levels of TUNEL-positive testicular cells (Fig 1D and 1E). These data indicate that Rubicon maintains germ cell homeostasis. Consistent with this, Rubicon knockout caused a reduction in sperm motility (Fig 1F-1H and S1 and S2 Movies), but altered neither sperm number nor sperm morphology (S1C and S1D Fig). Systemic Rubicon knockout mice also had Error bars indicate means ± SEM. Data were analyzed by two-tailed Student's t-test (A, C, E-J). � P < 0.05, �� P < 0.01, ��� P < 0.001, ���� P < 0.0001. N.S., not significant.
https://doi.org/10.1371/journal.pgen.1009688.g001 fewer pups per litter (Fig 1I), but no decrease in the number of litters per male mouse (Fig 1J), suggesting that Rubicon knockout decreases male fertility without affecting sexual behavior. Importantly, systemic Rubicon knockout mice exhibited a significant reduction in the levels of the autophagic substrates p62 and NBR1 in testes (S1E-S1G Fig), suggesting that autophagy was upregulated by Rubicon knockout. Our results indicate that Rubicon plays a crucial role in germ cell homeostasis and male fertility that could be mediated by autophagy.

Rubicon in germ cells is dispensable for mouse spermatogenesis
To determine how Rubicon works in mouse testis, we performed in situ hybridization using an antisense probe, which revealed ubiquitous expression of Rubicon in mouse testis (Fig 2A). This signal was abolished in knockout mice, indicating that it represented a bona fide Rubicon signal. To further elucidate the expression pattern of Rubicon, we reanalyzed published singlecell transcriptome data from mouse testes [27], and defined the clusters as each cell type (S2A and S2B Fig). We found that Rubicon is expressed mainly in spermatocytes, but is expressed at some level in all testicular cell types (Fig 2B and 2C). Hence, we sought to determine whether Rubicon maintains germ cell homeostasis in a cell-autonomous manner. For this purpose, we employed a transplantation assay in which the germ cells derived from donor knockout mice were transplanted into the seminiferous tubules of germ cell-deficient W/Wv mice [28,29]. Surprisingly, Rubicon-deleted germ cells settled in the empty seminiferous tubules as efficiently as wild-type cells (Fig 2D and 2E). We also generated germ cell-specific Rubicon knockout mice (Rubicon Vasa-/mice) using Rubicon-floxed mice [21] and Vasa-Cre mice [30], and confirmed the decrease in the level of Rubicon in mouse testis (S2C Fig). Rubicon Vasa-/mice exhibited no defect in spermatogenesis relative to control mice ( Fig 2F). These data indicate that Rubicon in germ cells is dispensable for the maintenance of germ cell homeostasis, i.e., Rubicon participates in the spermatogenesis in a non-cell-autonomous manner.

Rubicon in Sertoli cells is crucial for SSC homeostasis
To determine which somatic cells are crucial for defective spermatogenesis in systemic Rubicon knockout mice, we examined gene expression profiles in the testis. The mRNA levels of Sertoli cell-related genes were significantly reduced in the knockout mice (Fig 3A), whereas those of Leydig cell-related genes (S3A Fig Importantly, the knockout mice did not exhibit a reduction in plasma levels of FSH, which regulates Sertoli cell function or proliferation (S4B Fig), suggesting that the Sertoli cell abnormality in systemic Rubicon knockout mice is independent of endocrine effects. Therefore, we hypothesized that Rubicon in Sertoli cells is crucial for germ cell homeostasis. To test this idea, we crossed Rubicon-floxed mice with Amh-Cre mice [32] to generate Sertoli cell-specific Rubicon knockout mice (Rubicon Amh-/mice). Strikingly, like systemic Rubicon knockout mice, Rubicon Amh-/mice had reduced testicular weight (Fig 3B), defective spermatogenesis ( Fig 3C and 3D), reduced male fertility (Fig 3E and 3F), and reduced levels of Rubicon in testes (S4C Fig). This finding suggests that Rubicon in Sertoli cells is required for normal spermatogenesis. Given that Sertoli cells maintain the niche for undifferentiated spermatogonia, including SSCs [1,2], we hypothesized that Rubicon in Sertoli cells plays a key role in the maintenance of undifferentiated spermatogonia. To test this, we performed immunohistochemistry for PLZF and GFRα1, which are markers of undifferentiated spermatogonia and SSCs, respectively [33][34][35]. PLZF-positive cells were significantly less abundant in Rubicon Amh-/mice ( Fig 3G and 3H), and the numbers of GFRα1-positive cells were also reduced in the knockout mice (Fig 3I and 3J).

Rubicon prevents autophagic degradation of GATA4 in Sertoli cells
Next, we sought to determine the mechanism by which Rubicon participates in Sertoli cell function. We focused on the transcription factor GATA4, which is essential for Sertoli cell function including SSC maintenance [13,14]. Because GATA4 is degraded by autophagy [36], we hypothesized that loss of Rubicon promotes autophagic degradation of GATA4, leading to a decline in Sertoli cell function. Notably, systemic Rubicon knockout mice had significantly reduced levels of GATA4 in testes (Fig 4A and 4B). Histological analysis revealed that SOX9-positive Sertoli cells were less abundant in the testes of Rubicon Amh-/mice, and that the positive rate of GATA4 in Sertoli cells was also decreased in the knockout mice ( Fig 4C-4E), suggesting that Rubicon deletion leads to a reduction in the levels of GATA4 and in the Sertoli cell number. The reduced number of Sertoli cells could lead to the reduction in SSC number in the knockout mice. This is consistent with a reduction in mRNA levels of Sertoli cell-related genes ( Fig 3A). To test whether GATA4 is degraded by autophagy in Rubicon knockout mice, we injected the mice with a lysosomal inhibitor chloroquine. We found that chloroquine treatment clearly rescued the reduction of GATA4 levels in the testis of systemic Rubicon knockout mice (Fig 4F and 4G). To further explore our hypothesis that Sertoli cell GATA4 is degraded by autophagy, we used 15P-1 cells, which are derived from mouse Sertoli cells [37]. Rubicon knockdown decreased the level of GATA4 protein in 15P-1 cells, whereas the lysosomal inhibitor Bafilomycin A1 had the opposite effect (Fig 5A and 5B). Bafilomycin A1 increased the Immunoblotting of the indicated proteins in testes from WT or Rubicon KO mice intraperitoneally injected with or without 100 mg/kg chloroquine (CQ) for 8 h. n = 3-6. (G) Quantification of the relative GATA4 levels in (F). Error bars indicate means ± SEM. Data were analyzed by two-tailed Student's t-test (B, D, E, G). � P < 0.05, �� P < 0.01, ��� P < 0.001, ���� P < 0.0001. N.S., not significant.
https://doi.org/10.1371/journal.pgen.1009688.g004 levels of GATA4, but no other proteins, in a time-dependent manner (Fig 5C-5E), suggesting that GATA4 is specifically degraded by lysosomal pathways such as autophagy and endocytosis. Because Rubicon negatively regulates both autophagic and endocytic pathways [18,19], we evaluated its role in 15P-1 cells. Rubicon depletion in 15P-1 cells caused a substantial increase in the autophagic degradation of LC3-II and p62 (S5A- S5C Fig), but no significant change in the endocytic degradation of EGFR (S5D and S5E Fig). Collectively, these results indicate that Rubicon prevents autophagic degradation of GATA4 in Sertoli cells. Furthermore, nuclearcytoplasmic fractionation assays revealed that Rubicon knockdown decreased the level of GATA4 protein not only in the cytoplasmic fraction, but in the nuclear fraction as well (Fig 5F  and 5G). This observation suggests that genetic loss of Rubicon can suppress the transcriptional activity of GATA4. Because bafilomycin A1 increased the level of GATA4 protein in the cytoplasmic fraction earlier than in the nuclear fraction (Fig 5H and 5I), it is conceivable that cytoplasmic GATA4 is degraded by autophagy in Sertoli cells.

Androgens maintain the levels of Rubicon and GATA4 in Sertoli cells
The results described above indicate that genetic suppression of Rubicon in Sertoli cells promotes autophagic degradation of GATA4. However, it remains unknown whether Rubicon in Sertoli cells could be downregulated under physiological conditions. In this regard, we focused on previous reports showing that blockade of male hormones known as androgens increases autophagic activity in prostate cancer cells with androgen receptors [38,39]. Moreover, Androgen receptor knockout in Sertoli cells causes a severe defect in mouse spermatogenesis [32,40,41], suggesting that androgens play pivotal roles in Sertoli cells. Therefore, we hypothesized that androgens regulate the levels of Rubicon in Sertoli cells to promote spermatogenesis. To test this, we treated 15P-1 Sertoli cells with an androgen antagonist. 15P-1 cells expressed Androgen receptor (Fig 5C). We found that anti-androgen treatment caused a time-dependent reduction in the levels of Rubicon and GATA4 in 15P-1 cells (Fig 6A-6C). Concomitant with this, autophagic flux assays using the autophagic substrates LC3-II and p62, revealed that antiandrogen treatment caused a time-dependent increase in autophagic activity (Fig 6E-6G). Because anti-androgen treatment also reduced levels of another Sertoli-cell-related protein, WT1 (Fig 6A and 6D), we examined whether GATA4 is specifically degraded by autophagy in Sertoli cells. Unlike WT1, GATA4 was significantly reduced during starvation (Fig 6H-6J). In addition, a lysosomal inhibitor bafilomycin A1 specifically increased levels of GATA4 ( Fig  5C-5E); therefore, it is conceivable that anti-androgen treatment promotes specific degradation of GATA4. Also in mice, anti-androgen therapy decreased the testicular levels of Rubicon, GATA4, and WT1 (Fig 7A-7D). Strikingly, prostate cancer patients receiving anti-androgen therapy exhibited a significant decrease in the levels of testicular Rubicon and GATA4 relative to a tumor-free patient or prostate cancer patients who did not receive such treatment ( Fig  7E-7G). Collectively, our findings suggest that androgen maintains the testicular levels of Rubicon and GATA4 both in mouse and human.

Discussion
Androgens are male steroid hormone that stimulate cytosolic androgen receptors, which subsequently translocate into the nucleus to promote cell type-specific expression in testis, thereby maintaining spermatogenesis [42,43]. Among various testicular cell types, androgen receptors in Sertoli cells are the most important for spermatogenesis and male fertility [32,40,41]. Fatherhood decreases the levels of androgen in humans [44], suggesting that downregulation of androgens in fatherhood could decrease male fertility, leading the male to focus on parenting existing offspring. These studies indicate that androgen plays a pivotal role in

PLOS GENETICS
Sertoli cells to regulate male reproduction. Previously, however, it remained largely unknown how androgen regulates Sertoli cell function in the context of spermatogenesis.
In this study, we found that androgen antagonists suppress testicular Rubicon, and that genetic loss of Rubicon in Sertoli cells but not in germ cells promotes autophagic degradation of GATA4, thereby decreasing spermatogenesis and stem cell maintenance. Our results suggest that androgens maintain the levels of Rubicon to control autophagic degradation of GATA4 to promote Sertoli cell function. Androgen receptor knockout in Sertoli cells caused more severe phenotypes than Rubicon knockout, implying the existence of other androgendependent pathways to be elucidated by future studies.
The lysosomal inhibitor bafilomycin A1 increases the levels of GATA4 at the basal state (Fig 5C and 5D), suggesting that GATA4 is constitutively degraded by autophagy. It remains to be determined why this is the case. GATA4 is an evolutionarily conserved transcription factor that is crucial for development of the heart, liver, and pancreas [45][46][47]. GATA4 is also essential for genital ridge formation [48], sex determination and differentiation [49][50][51], and Sertoli cell function [13,14]. Therefore, it is conceivable that Rubicon regulates autophagy in order to control the amount of GATA4 in various organs or tissues during development. If so, context-dependent regulators of Rubicon are of particular interest.
Our results suggest that genetic loss of Rubicon decreases male fertility. By contrast, we previously showed that Rubicon accumulates with age, and that loss of Rubicon extends lifespan by upregulating autophagy [20]. Fertility is negatively correlated with longevity in animals [52,53]; the regulation of autophagic degradation of GATA4 by Rubicon could be one of the underlying mechanisms that reciprocally regulates fertility and longevity. Indeed, the worm GATA homologs ELT-5 and ELT-6 accumulate with age, and knockdown of these genes extends lifespan [54], supporting the idea that autophagic degradation of GATA proteins could contribute to the longevity in Rubicon-ablated animals.
In summary, we propose that androgens maintain Rubicon levels in Sertoli cells to prevent autophagic degradation of GATA4 for spermatogenesis. In turn, excessive autophagy due to loss of Rubicon could contribute to the pathogenesis of idiopathic male infertility. Therefore, we anticipate that partial suppression of autophagy represents a promising therapeutic target for such diseases.

Ethics statement
The experimental procedures using mice were approved by the Institutional Committee of Osaka University (Approval number 02-009-000). The human studies were approved by the Institutional Review Board of Osaka University Hospital (IRB number 20225). We complied with all of the relevant ethical regulations, and informed consent was obtained from all subjects. Written consent was obtained from the participants.

Reagents and antibodies
The following antibodies were used for western blotting at the indicated dilutions: rabbit

Animals
C57BL/6J mice were obtained from CLEA Japan. Rubicon -/mice [20] and Rubicon-floxed mice [21] were previously generated in our laboratory. Vasa-Cre mice [30] and Amh-Cre mice [32] were obtained from Dr. Diego Castrillon (University of Texas Southwestern Medical Center) and Dr. Robert E. Braun (University of Washington School of Medicine), respectively. Vasa-Cre mice and Amh-Cre mice were crossed with Rubicon-floxed mice [21] to produce Rubicon flox/-; Vasa-Cre mice (Rubicon Vasa-/mice) and Rubicon flox/flox ; Amh-Cre mice (Rubicon Amh-/mice), respectively. Hemizygous Cre mice were used to avoid phenotypes resulting from homozygosity. Rubicon flox/mice and Rubicon flox/+ ; Amh-Cre mice were used as controls for Rubicon Vasa-/mice and Rubicon Amh-/mice, respectively. All mice used in this study, except for W and B6D2F1 mice, were maintained on the C57BL/6J background. The following primer sets were used for PCR genotyping: 5 0 -ACAACGACAATCACACAGAC-3 0 and 5 0 -TGACGA GGGGTAATGGATAG-3 0 for Rubicon WT and floxed allele; 5 0 -ACAACGACAATCACAC AGAC-3 0 and 5 0 -AATCCTTCGCCCCTTTTACC-3 0 for Rubicon deletion allele; 5 0 -GCATTA CCGGTCGATGCAACGAGTGATGAG-3 0 and 5 0 -GAGTGAACGAACCTGGTCGAAAT CAGTGCG-3 0 for Cre. Mice were maintained on a normal chow in 12-h light/12-h dark cycles. Ambient temperature and humidity were 23 ± 1.5˚C and 45 ± 15%, respectively. Food and water were provided ad libitum. Samples were obtained from male mice at 5-7 months of age for qRT-PCR, immunoblotting, or immunostaining. Testosterone EIA Kit (Cayman Chemical) was used to determine plasma testosterone levels. FSH ELISA Kit (Enzo) was used to determine plasma FSH levels. To examine male fertility, a 12-month-old WT or Rubicon KO male mouse was mated with three B6D2F1 females (CLEA Japan, 2-month-old) for 2 months, and the number of pups was counted at the day of birth. 100 μM Chloroquine in PBS was intraperitoneally injected into mice. 100 μM Bicalutamide in corn oil was intraperitoneally injected into mice once a day for six days. Control mice were injected with solvent only.

Transplantation assay
Spermatogonial transplantation was carried out by microinjection into the seminiferous tubules of infertile W mice via an efferent duct (Japan SLC) as previously described [55]. Briefly, the tunica albuginea was removed from the testis. The seminiferous tubules were incubated in HBSS containing 1 mg/ml Type IV collagenase (Sigma) at 37˚C, were then washed in HBSS, followed by incubation at 37˚C for 5 min in HBSS containing 1 mM EDTA and 0.25% trypsin. The activity of trypsin was terminated by adding fetal bovine serum. Following digestion, the cell suspension was filtered through a nylon mesh. The filtrate was centrifuged and the pellet was used as the donor cells. Approximately 4 μl of single-cell suspension from WT and Rubicon KO mice were transplanted into the recipient's left testis and right testis, respectively. 4 × 10 6 cells/testis were injected. Each injection filled 75-85% of all seminiferous tubules.

Human testis specimens
All testis specimens were obtained from living patients by surgery during the past 6 yr. A tumor-free specimen was obtained from a patient with testis trauma. Patients with prostate cancer were treated by castration for androgen deprivation therapy. The specimens were fixed in formalin, paraffinized, and processed for immunostaining as described below.

Sperm motility assay
Sperm motility assays were obtained using samples from 18-month-old male mice. Spermatozoa were collected from the cauda epididymis and suspended in Toyoda, Yokoyama, and Hoshi (TYH) medium [57]. Sperm motility was assessed at 10 min and 120 min after sperm suspension. One epididymis was used for each experiment. Sperm motility was measured and analyzed using a CEROS II sperm analysis system (software version 1.4; Hamilton Thorne Biosciences). Sperm morphology was observed on an Olympus BX-53 microscope (Olympus).

Histology
Tissues were fixed in 4% paraformaldehyde overnight, and then held in 70% ethanol until processing. Tissues were paraffinized and sectioned at 5 μm by microtome (Leica). The slides were stained with H&E. Immunohistochemical staining was performed on paraffin-embedded sections. After deparaffinization and rehydration, antigen retrieval was performed by microwaving in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) or Tris-EDTA buffer (10 mM Tris, 1 mM EDTA, pH 9.0) for 15 min, or by incubation in proteinase K solution (10 μg/ml in PBS) for 15 min at 37˚C. The sections were incubated in 3% hydrogen peroxide for 5 min at room temperature, and then blocked in 2.5% Normal Horse Serum (Vector Laboratories, S-2012) for 30 min at room temperature. The blocked sections were incubated with the primary antibody in 2.5% Normal Horse Serum for 60 min at room temperature, followed by incubation for 60 min at room temperature with the secondary antibody.
The sections were counterstained with hematoxylin. DAB staining was performed using the DAB Peroxidase Substrate Kit, ImmPACT (Vector Laboratories, SK-4015). TUNEL staining was performed using the In situ Apoptosis Detection Kit (Takara Bio). Images were acquired on a BZ-X700 microscope (Keyence). According to a previous report [58], defective tubules were defined as loss of germ cells along a significant portion of the seminiferous epithelium, germ cell sloughing, presence of large vacuoles, and tubular dilation.

In situ hybridization
Antisense and sense probes were generated from mouse Rubicon cDNA in pGBD using the DIG RNA Labelling kit (Roche, 11175025910) and the following primers: 5 0 -TAATACGACT CACTATAGGGCGTCCGGAGGGCGCGGGAATG-3 0 and 5 0 -ATTTAGGTGACACTATA GAAGGCTGTGACGTGGGCGTCACTCAG-3 0 . In situ hybridization was performed using the ISHR Starting Kit (Nippon Gene). Briefly, paraffin-embedded sections of mouse testis were deparaffinized and rehydrated and incubated in proteinase K solution (5 μg/ml in PBS) for 10 min at room temperature. The sections were acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine hydrochloride (pH 8.0) for 15 min at room temperature. After prehybridization with 50% formamide in 2× SSC buffer for 30 min at 45˚C, the sections were hybridized with the DIG-labeled probes in hybridization buffer (50% formamide, 2× SSC, 1 μg/μl tRNA, 1 μg/μl salmon sperm DNA, Denhardt's solution, and 10% dextran sulfate) overnight at 45˚C. After two washes with 50% formamide in 2× SSC buffer for 30 min at 45˚C, the sections were incubated in RNase A solution (20 μg/ml RNase A in NTE buffer) for 30 min at 37˚C. The sections were incubated in blocking buffer [1% Blocking Reagent (Roche, 11096176001) in 100 mM Tris-HCl, pH 7.5, 150 mM NaCl] for 30 min at room temperature, and then incubated with anti-DIG-AP antibody (Roche, 11093274910, 1:1000) in blocking buffer for 60 min at room temperature. The sections were visualized with NBT/BCIP solution [2% NBT/BCIP Stock Solution (Roche, 11681451001) in 100 mM Tris-HCl, pH 9.5, 100 mM NaCl] overnight at room temperature, and then counterstained with Nuclear Fast Red (Vector Laboratories, H-3403).

RNA isolation and quantitative PCR analyses
Mouse tissues were harvested in QIAzol (Qiagen) using a Precellys Evolution tissue homogenizer (Bertin). Total RNA was extracted using RNeasy Plus Mini kit (Qiagen). cDNA was generated using iScript (Bio-Rad). qRT-PCR was performed using Power SYBR Green (Applied Biosystems) on a QuantStudio 7 Flex Real-time PCR System (Applied Biosystems). Four technical replicates were performed for each reaction. Actb was used as an internal control. Sequences of qRT-PCR primers are shown in S1 Table.

Single-cell transcriptome analysis
A single-cell transcriptome data of murine testis was obtained from the previous report [27]. Rubicon expression and UMAP visualization of the cell clusters was re-analyzed by Scanpy [59]. The cell types of each cluster were identified manually by the expression of characteristic marker genes used in the previous study [27].

Immunoblotting
Mouse tissues were harvested in RIPA buffer [50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% w/v Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, protease inhibitor cocktail (Roche)] using a tissue homogenizer Precellys Evolution (Bertin). Cultured cells were lysed in the same RIPA buffer. After centrifugation, the supernatants were subjected to protein quantification by the Protein Assay BCA Kit (Nacalai Tesque). Protein lysates were mixed with 5× SDS sample buffer and boiled for 5 min, separated by 7% or 13% SDS-PAGE, and transferred to PVDF membranes. Membranes were stained with Ponceau-S, blocked with TBS-T containing 1% skim milk, and incubated with primary antibodies in TBS-T containing 1% skim milk. Immunoreactive bands were detected using HRP-conjugated secondary antibodies, visualized with Luminata Forte (Merck Millipore) or ImmunoStar LD (Wako), and imaged using ChemiDoc Touch (Bio-Rad). α-tubulin, β-actin, GAPDH, or Lamin B was used as a loading control. For quantification, the band intensity of each protein was normalized against the loading control. Band intensity was quantified using the ImageJ software (NIH).

Autophagic flux assay
Cells were incubated in normal medium with or without 125 nM bafilomycin A1 (BafA1) for 4 h at 37˚C in an atmosphere containing 5% CO 2 , and then lysed and immunoblotted for LC3 or p62. Autophagic flux was calculated by subtracting the densitometric values of LC3-II or p62 in BafA1-untreated samples from those in BafA1-treated samples.

Nuclear/cytoplasmic fractionation assay
Cells were lysed with 0.1% NP-40 and protease inhibitor cocktail (Roche) in PBS. An aliquot of each lysate was mixed with 5× SDS sample buffer and used as a whole-cell lysate. Another aliquot was centrifuged; the resultant supernatant was mixed with 5× SDS sample buffer and used as the cytoplasmic fraction. The pellet was washed once and lysed in SDS sample buffer, and used as a nuclear fraction.

EGFR degradation assay
Cells were incubated in DMEM without serum for 4 h. The cells were treated with 100 ng/ml EGF (Invitrogen, 53003-018) and lysed at 0, 15, 60, 120, and 180 min. Cell lysates were subjected to immunoblotting for EGFR.

Statistical analyses
All results are presented as means ± SEM. Statistical analyses were performed by two-tailed Student's t-test, one-way ANOVA followed by Tukey's test, or two-way ANOVA using Excel for Mac (Microsoft) and GraphPad Prism7 (GraphPad Software). Numerical data is available in S2 Table. Supporting information