SCF^FBXW5-mediated degradation of AQP3 suppresses autophagic cell death through the PDPK1-AKT-MTOR axis in hepatocellular carcinoma cells

ABSTRACT

AQP3 (aquaporin 3 (Gill blood group)), a member of the AQP family, is an aquaglyceroporin which transports water, glycerol and small solutes across the plasma membrane. Beyond its role in fluid transport, AQP3 plays a significant role in regulating various aspects of tumor cell behavior, including cell proliferation, migration, and invasion. Nevertheless, the underlying regulatory mechanism of AQP3 in tumors remains unclear. Here, for the first time, we report that AQP3 is a direct target for ubiquitination by the SCF^FBXW5 complex. In addition, we revealed that downregulation of FBXW5 significantly induced AQP3 expression to prompt macroautophagic/autophagic cell death in hepatocellular carcinoma (HCC) cells. Mechanistically, AQP3 accumulation induced by FBXW5 knockdown led to the degradation of PDPK1/PDK1 in a lysosomal-dependent manner, thus inactivating the AKT-MTOR pathway and inducing autophagic death in HCC. Taken together, our findings revealed a previously undiscovered regulatory mechanism through which FBXW5 degraded AQP3 to suppress autophagic cell death via the PDPK1-AKT-MTOR axis in HCC cells.

Abbreviation: BafA1: bafilomycin A₁; CQ: chloroquine; CRL: CUL-Ring E3 ubiquitin ligases; FBXW5: F-box and WD repeat domain containing 5; HCC: hepatocellular carcinoma; HSPA8/HSC70: heat shock protein family A (Hsp70) member 8; 3-MA: 3-methyladenine; PDPK1/PDK1: 3-phosphoinositide dependent protein kinase 1; RBX1/ROC1: ring-box 1; SKP1: S-phase kinase associated protein 1; SCF: SKP1-CUL1-F-box protein.

Introduction

Aquaporins (AQPs) are a family of transmembrane water channel proteins extensively present in diverse tissues across biological systems, which not only maintain water homeostasis by controlling cellular water transport [1,2], but also play a pivotal role in facilitating the movement of assorted molecules, such as glycerol and urea [2,3]. AQPs themselves are also subjected to regulation at different stages, such as transcription and degradation. Numerous previous investigations have illuminated the degradation of AQPs through the lysosomal system. For example, SNX27 (sorting nexin 27) regulates the lysosomal degradation of AQP2 protein in the kidney collecting duct [4]. In addition, VTA1 (vesicle trafficking 1) interacts with AQP2 and facilitates its lysosomal degradation in kidney [5]. Moreover, stress-induced kinase CSNK2 (casein kinase 2) phosphorylates Ser276 of AQP4 and subsequently enhances the lysosomal degradation of AQP4 [6]. Nevertheless, the degradation of AQPs proteins via the proteasomal system remains largely elusive.

AQP3 (aquaporin 3), belonging to the AQP family, serves as an aquaglyceroporin responsible for the conveyance of water, glycerol, and minor solutes across the cellular plasma membrane [1,7–9]. Besides its involvement in controlling fundamental biological functions in regular cells, AQP3 also plays an important role in tumorigenesis and tumor progression [10–14]. AQP3 functions as either a tumor suppressor or an oncoprotein in a cancer-type dependent manner. Prior investigations have demonstrated AQP3’s oncoprotein status across various tumor types. For instance, the upregulation of AQP3 intensified the proliferation of gastric cancer cells [15]. In vivo, AQP3 knockdown prevented lung metastasis from mammary orthotopic xenografts [16]. Conversely, AQP3 has also been implicated as a tumor suppressor in specific malignancies. For example, the upregulation of CDH1/E-cadherin and knockdown of RALA result in elevated AQP3 levels on the plasma membrane of prostate cancer cells, which significantly inhibits cell proliferation and enhances apoptosis [17]. The administration of 5-fluorouracil induced AQP3 expression, subsequently inducing cytotoxicity and cell cycle arrest in breast cancer [18]. In addition, analysis of the TCGA RNA-seq database revealed that AQP3 was expressed at a low level in hepatocellular carcinoma (HCC) tissues when compared with the adjacent normal tissues (Fig. S1). The pathological role of AQP3 in tumorigenesis and cancer progression is rather complicated, seemingly related to tissue type and cellular context. Consequently, elucidating the molecular mechanism of AQP3 regulation will help clarify the pathogenesis of HCC and may provide a basis for anti-HCC strategies.

The SKP1-CUL1-F-box protein (SCF) E3 ubiquitin ligase complex is the first identified CUL-Ring E3 ubiquitin ligases (CRL) [19–21], which is composed of four functional components, SKP1 (S-phase kinase associated protein 1), RBX1/ROC1 (ring-box 1), scaffold protein CUL1 and the substrate receptor F-box proteins. According to their structural characteristics, F-box family proteins can be divided into three subfamilies, FBXW, FBXL and FBXO [22]. FBXW proteins target different substrates for ubiquitination and play a large number of important biological functions, including cell proliferation, cell cycle regulation, cell invasion and immunity [23–29]. FBXW5 (F-box and WD repeat domain containing 5) is a member of the FBXW subclass of F-box proteins, which is responsible for the degradation of protein substrates related to a series of pathophysiological processes, including inflammation [30], mitotic progression [31,32], and tumorigenesis [33–35]. For example, CUL4A^FBXW5-mediated degradation of RHO GTPase activating protein DLC1 promoted the proliferation of non-small lung cancer cells. Besides, FBXW5 promoted proper mitotic progression by regulating the degradation of actin remodeler EPS8 in human osteosarcoma cells [32]. However, the influence of FBXW5 on cancers via the modulation of aquaporin proteins has yet to be revealed.

Autophagy, a cellular catabolic degradation process triggered by starvation or stress, primarily assumes four roles [36–47], protective, cytotoxic, cytostatic, and non-protective. Evidence has shown that autophagy may serve as a cell death mechanism and this autophagy-dependent cell death has been defined as “autophagic cell death” [46–48]. The AKT-MTOR pathway is a critical signaling axis in cell proliferation and cell survival, which has been proven to be a negative signal regulator of autophagy [49–51]. However, at present, the precise upstream mechanism driving autophagic cell death induced by the AKT-MTOR pathway has not been thoroughly elucidated.

Herein, we demonstrated that the SCF^FBXW5 E3 ligase complex specifically targeted AQP3 for ubiquitination and subsequent proteasome-dependent degradation in HCC cells. We further elucidated that FBXW5 knockdown-induced AQP3 accumulation restrained PDPK1 expression, thereby inducing autophagy via inactivating AKT-MTOR pathway. Our study revealed a new mechanism by which FBXW5 ubiquitinated and degraded AQP3 to modulate PDPK1-AKT-MTOR pathway to suppress autophagic cell death in HCC cells.

Results

FBXW5 interacted with AQP3 and inhibited AQP3 expression

By utilizing immunoprecipitation-based mass spectrometry screenings (Figure 1A), we have identified a number of putative interactors of FBXW5 in HepG2 cells. Notably, well-established FBXW5-interacting proteins, including Cullin1 and SKP1, were prominently identified in Figure 1B, lending strong credibility to the mass spectrometry analysis. Among these FBXW5-binding proteins, AQP3, one of the water channel proteins, aroused our interest (Figure 1B-C). To ascertain the interaction between FBXW5 and AQP3, confocal microscopy was employed to assess co-localization of AQP3 and FBXW5. Our observations demonstrated the interaction between AQP3 and FBXW5 in HepG2 and Huh7 cells (Figure 1D). In addition, we co-transfected HA-FBXW5 and FLAG-AQP3 to perform immunoprecipitation (IP) and found that AQP3 interacted with FBXW5 in HEK293T cell (Figure 1E). To explore the FBXW5’s regulatory role on AQP3 expression, we attenuated FBXW5 expression using siRNA and observed a notable upregulation of AQP3 upon FBXW5 knockdown (Figure 1F). We further ectopically expressed FLAG-FBXW5 in HepG2 cells to detect AQP3 expression and found that overexpression of FBXW5 promoted AQP3 degradation in a dose-dependent manner (Figure 1G). Given FBXW5’s pivotal role as a constituent of CRL E3 Ligases, we postulated that the SCF^FBXW5 E3 ligase complex similarly governed AQP3 expression. Subsequently, we silenced the constituents of the SCF complex, namely CUL1, RBX1/ROC1, or SKP1, to deactivate SCF activity, and subsequently assessed AQP3 expression. As shown in Fig. H-J, downregulation of CUL1, RBX1/ROC1 or SKP1 significantly induced AQP3 protein abundance. These findings collectively demonstrated that FBXW5 interacted with AQP3 and inhibited AQP3 expression.

Figure 1.

FBXW5 interacted with AQP3 and inhibited AQP3 expression. (A) Label-free quantitative proteomic assay of anti-FBXW5 immunoprecipitation (IP) in HepG2 cells. (B) Already known FBXW5 interacting proteins (CUL1, SKP1) were detected via mass spectrometry analysis. (C) Representative tandem MS peptide spectrum of AQP3. (D) Colocalization of FBXW5 and AQP3 in HepG2 and Huh7 cells. Confocal microscopy analysis of FBXW5 (red) and AQP3 (green) in HepG2 and Huh7 cells. Nuclei were stained with DAPI. White arrows indicated the place where FBXW5 and AQP3 were co-localized. Scale bar: 25 μm. (E) Exogenous FBXW5 interacted with AQP3. HA-tagged FBXW5 and flag-tagged AQP3 were constructed and transfected into HEK293T cells as indicated for 24 h. Total cell lysates were subjected to immunoprecipitation with anti-HA beads and immunoblotting with flag Ab or subjected to immunoprecipitation with anti-flag M2 beads and immunoblotting with HA Ab. (F) Downregulation of FBXW5 induced AQP3 accumulation in both HepG2 and Huh7 cells. Cells were transfected with siFBXW5 for 72 h and cells were collected and subjected to IB analysis for the expression of AQP3. (G) HepG2 cell was transfected with flag-FBXW5 (0, 0.25, 0.5, 1, 2 and 4 μg) for 36 h and subjected to IB analysis for the expression of AQP3. (H–J) Downregulation of endogenous CUL1, RBX1 or SKP1 induced the accumulation of AQP3 in HepG2 cell. HepG2 cells were transfected with siRNA oligoes targeting CUL1, RBX1 and SKP1. Cells were lysed after transfection for 72 h, followed by IB with the indicated antibodies.

AQP3 is a ubiquitin substrate of the SCF^FBXW5 E3 ligase complex

Given that FBXW5 functions as a crucial constituent of CRL E3 Ligases, we assumed that AQP3 as an ubiquitin substrate of the SCF^FBXW5 E3 ligase complex. To validate this hypothesis, the protein stability of AQP3 in FBXW5-silenced cells was measured in the presence of cycloheximide (CHX). Our results showed that FBXW5 knockdown dramatically extended the protein half-life of AQP3 in HepG2 and Huh7 cells (Figure 2A, S2A). Furthermore, we found that FBXW5 knockdown minimally affected the mRNA levels of AQP3 (Fig. S2B), excluding the possibility that altered the AQP3 expression level elicited by FBXW5 knockdown was attributable to transcriptional changes. Next, we knocked down the subunits of SCF complex CUL1, RBX1/ROC1 or SKP1 to inactivate SCF, and then detected the protein stability of AQP3 in the presence of CHX. Our results showed that CUL1, RBX1/ROC1 or SKP1 knockdown dramatically extended the protein half-life of AQP3 in HepG2 cells (Figure 2B-C, Fig. S2C). Since FBXW5 knockdown notably postponed AQP3 degradation, we proceeded to investigate whether FBXW5 knockdown hindered the ubiquitination process of AQP3. Our results showed that AQP3 ubiquitination was markedly impaired by FBXW5 knockdown in HepG2 and Huh7 cells (Figure 2D). Correspondingly, AQP3 ubiquitination was also dramatically inhibited in HepG2 cells upon downregulating SCF components CUL1, RBX1/ROC1, or SKP1, as shown in Figure 2E. The cumulative results strongly support the notion that AQP3 operates as a downstream substrate for ubiquitination by SCF^FBXW5.

Figure 2.

AQP3 was a ubiquitin substrate of the SCF^FBXW5 E3 ligase complex. (A) HepG2 cells were transfected with siControl or siFBXW5 for 72 h and then treated with 100 μg/mL of cycloheximide (CHX) for indicated period of time. The cells were collected for IB analysis. The protein level was quantified by densitometric analysis using ImageJ software. (B) HepG2 cells were transfected with siCUL1 for 72 h and then treated with 100 μg/mL of cycloheximide (CHX) for indicated period of time. The cells were collected for IB analysis. The protein level was quantified by densitometric analysis using ImageJ software. (C) HepG2 cells were transfected with siRBX1 for 72 h and then treated with 100 μg/mL of cycloheximide (CHX) for indicated period of time. The cells were collected for IB analysis. The proteinlevel was quantified by densitometric analysis using ImageJ software. (D) Downregulation of FBXW5 decreased the ubiquitination of AQP3. HepG2 and Huh7 cells were transfected with indicated siRNA, lysed under denatured lysis buffer containing 1% SDS, followed by immunoprecipitation using anti-AQP3. IB analyzes were performed using the indicated antibodies. (E) Downregulation of CUL1, RBX1 or SKP1 decreased the ubiquitination of AQP3. HepG2 cells were transfected with indicated siRNA, lysed under denatured lysis buffer containing 1% SDS, followed by immunoprecipitation using anti-AQP3. IB analyzes were performed using indicated antibodies. (F) Schematic depiction of the AQP3 architecture and its lysine variant mutations. (G) HepG2 cells, transfected with designated plasmids, were harvested for western blot analysis 24 h after transfection. (H) HepG2 cells, with overexpressed AQP3 or its K282R mutant, were treated with 50 µg/mL cycloheximide (CHX) for specified durations. The intensity of protein bands was quantified by densitometric analysis using ImageJ software. (I) An in vivo ubiquitination assay was performed on AQP3 lysine-to-arginine (K–R) mutants in HEK293 cells, in the absence and presence of FBXW5.

Having established that FBXW5 promotes the ubiquitination of AQP3, we proceeded to identify the specific lysine residues on AQP3 where ubiquitin is attached. To accomplish this, we mutated the five lysine sites on AQP3 to arginine, respectively, and examined whether FBXW5 induced ubiquitination on these mutants (Figure 2F). We observed that only the AQP3^K282R mutant was resistant to degradation by FBXW5 (Figure 2G). The co-expression of the K5R, K100R, K287R, or K289R mutants did not prevent the degradation of AQP3 (Figure 2G), suggesting that these lysine residues were not required for FBXW5-mediated ubiquitination of AQP3. Consistent with this observation, the half-life of the AQP3^K282R mutant was longer than that of wild-type AQP3 (Figure 2H). Furthermore, only the AQP3^K282R mutant failed to undergo ubiquitination by FBXW5, indicating that K282 is the primary site where FBXW5 catalyzes ubiquitin attachment (Figure 2I). Taken together, our results demonstrate that FBXW5 promotes the degradative ubiquitination of AQP3, with K282 being the main site for ubiquitin attachment.

Downregulation of FBXW5 induced autophagic death of HCC cells dependent on AQP3 accumulation

To elucidate the cellular phenotype and function regulated by FBXW5, our investigation revealed that silencing FBXW5 induced a punctate distribution of membrane-associated lipidated LC3 (LC3-II) in HepG2-EGFP-LC3 and Huh7-EGFP-LC3 cells (Figure 3A). By using transmission electron microscopy, we further observed the obvious double-membrane autophagosome in FBXW5-silenced HepG2 and Huh7 cells, as well as vacuoles with engulfed bulk cytoplasm and cytoplasmic organelles, which was another golden hallmark of autophagy (Figure 3B). Moreover, FBXW5 knockdown significantly induced the conversion of LC3-I to LC3-II, a classical marker of autophagy, (Figure 3C). In addition, we performed the autophagic flux analysis by treating cells with classical autophagy inhibitors, including chloroquine (CQ), bafilomycin A₁ (BafA1) and 3-methyladenine (3-MA), respectively. Notably, CQ and BafA1 enhanced, while 3-MA inhibited the accumulation of LC3-II, indicating that autophagic flux was induced by FBXW5 knockdown (Figure 3D). Collectively, these results provide compelling evidence that FBXW5 knockdown triggers autophagy in HCC cells.

Figure 3.

Downregulation of FBXW5 induced autophagic death of HCC cells dependent on AQP3 accumulation. (A) FBXW5 silencing at 72 h induced punctuative distribution of membrane-associated lipidated LC3-II in HepG2-EGFP-LC3 and Huh7-EGFP-LC3 cells, observed with fluorescence microscope. Arrows denoted punctate vesicle structure, indicating autophagy induction. Scale bar: 10 μm. (B) Obvious double-membraned autophagosome and vacuoles with engulfed bulk cytoplasm and cytoplasmic organelles in siFBXW5 HepG2 and Huh7 cells at 72 h. Scale bar: 1 μm. (C) downregulation of FBXW5 induced the conversion of LC3-I to LC3-II in both HepG2 and Huh7 cells. Cells were treated with siFBXW5 for 72 h and cells were collected and subjected to IB analysis for the expression of LC3. (D) Autophagic flux analysis. The HepG2 and Huh7 cells, post-transfection with siFBXW5 for 72 hours, were subsequently incubated with CQ (50 μM), BafA1 (50 nM) or 3 MA (5 mM) for 6 h. The treated cells were then subjected to IB analysis with actin as a loading control. (E) The HepG2 and Huh7 cells were transfected with siFBXW5 for 120 h and subjected to cell proliferation analysis using ATPlite assay. (F–G) FBXW5 silencing reduced colony formation in HepG2 and Huh7 cells. HepG2 and Huh7 cells were transfected with siFBXW5 for 12 days, followed by crystal violet staining and colony counting. Representative images were shown. (H) Trypan blue exclusion assays were performed at 24 h intervals over a period of 96 h subsequent to FBXW5 knockdown. (I) HepG2 and Huh7 cells were transfected with siFBXW5 and siATG5 for 96 h and cells were collected and subjected to IB analysis for the expression of LC3. (J) the inhibition of autophagic response by siATG5 reversed FBXW5 knockdown-induced apoptosis. (K–M) The inhibition of autophagic response by siATG5 obviously reversed FBXW5 knockdown-induced inhibition of proliferation (K) and colony formation (L–M) in HepG2 and Huh7 cells. (N) AQP3 is required for autophagy induced by downregulation of FBXW5 in HepG2 and Huh7 cells. HepG2 and Huh7 cells were transfected with siAQP3 and siFBXW5 for 72 h. AQP3 and FBXW5 knockdown efficiency and conversion of LC3-I to LC3-II were assessed by IB analysis. (O) Downregulation of AQP3 attenuated cell apoptosis-induced by downregulation of FBXW5 in HepG2 and Huh7 cells. (P) Downregulation of AQP3 attenuated proliferation inhibition-induced by downregulation of FBXW5 in HepG2 and Huh7 cells. HepG2 and Huh7 cells were transfected with siAQP3 and siFBXW5 for 120 h. The cell proliferation was detected by ATPlite assay. (Q) Downregulation of AQP3 attenuated colony formation inhibition-induced by downregulation of FBXW5 in HepG2 and Huh7 cells.

Given that autophagy played four roles [36–48], protective, cytotoxic, cytostatic, and non-protective, our subsequent investigation aimed to ascertain the biological implications of the autophagic response prompted by FBXW5 knockdown. To explore this issue, we initially assessed the viability of HCC cells and observed that FBXW5 silencing substantially impaired the proliferation of HepG2 and Huh7 cells (Figure 3E). Furthermore, FBXW5 depletion markedly decreased clonogenic survival in both HepG2 and Huh7 cells(Figure 3F-G). To elucidate the impact of FBXW5 inhibition on hepatic carcinoma cell growth through cell death induction, trypan blue exclusion assays were performed at 24-h intervals over a period of 96 h post-FBXW5 knockdown. The results revealed a significant, time-dependent elevation in cell mortality due to FBXW5 inhibition (Figure 3H, Fig. S3A), suggesting that FBXW5 downregulation inhibits liver cancer cell proliferation by promoting cell death. Further investigation into the specific death modality induced by FBXW5 downregulation revealed that only 6% of the cells underwent apoptosis, a modest increase compared to the control group (Fig. S3B-C). Interestingly, the apoptotic pathway inhibitor Z-VAD partially mitigated the reduction in hepatocellular carcinoma cell proliferation induced by FBXW5 silencing by obstructing apoptosis (Fig. S3D). This indicates that FBXW5 downregulation primarily prompts liver cancer cell death through non-apoptotic mechanisms. Investigating whether autophagy was implicated in the cell death resulting from FBXW5 knockdown, we employed siRNA to silence ATG5, thus inhibiting the autophagic process in HepG2 and Huh7 cells. The suppression of ATG5 significantly decreased the FBXW5 knockdown-induced conversion of LC3-I to LC3-II (Figure 3I), and consequently, ATG5 silencing markedly reduced the apoptosis initiated by FBXW5 depletion (Figure 3J, Fig. S3E). Moreover, inhibition of the autophagic response with siATG5 significantly reversed the suppression of proliferation and colony formation caused by FBXW5 knockdown in both HepG2 and Huh7 cells (Figure 3K-M). These findings demonstrate that autophagy, induced by FBXW5 knockdown, acts as a pro-death signal in HCC cells.

To determine the potential role of AQP3 in FBXW5 knockdown-induced autophagic cell death, the expression of AQP3 was downregulated via siRNA. First, we found that downregulation of AQP3 significantly decreased the conversion of LC3-I to LC3-II elicited by FBXW5 knockdown in both HepG2 and Huh7 cells, suggesting that FBXW5 knockdown induced autophagy dependent on AQP3 accumulation (Figure 3N). In addition, downregulation of AQP3 significantly rescued siFBXW5 induced-cell apoptosis (Figure 3O). Downregulation of AQP3 significantly rescued siFBXW5 induced-proliferation inhibition, as demonstrated by the significant decrease in proliferation and colony formation of HepG2 and Huh7 cells (Figure 3P-Q, Fig. S3F). Collectively, these findings demonstrated the crucial role of AQP3 accumulation induced in FBXW5 knockdown-induced autophagic cell death in HCC cells.

FBXW5 knockdown inactivated the AKT-MTOR pathway dependent on AQP3 accumulation

To investigate the molecular mechanism underlying FBXW5 knockdown-induced autophagy, we assessed the activation status of AKT-MTOR pathway. We found that, in HepG2 and Huh7 cells, FBXW5 knockdown inactivated mTOR, as demonstrated by the significant decrease in phosphorylated RPS6KB/p70S6K and EIF4EBP1 (Figure 4A). We next explored the upstream mechanism for MTOR inhibition upon FBXW5 knockdown and found that FBXW5 knockdown dramatically induced inhibition of AKT, as evidenced by the obvious decrease in phosphorylated AKT (p-AKT) (Figure 4A). These results indicated that FBXW5 knockdown suppressed the AKT-MTOR pathway and subsequently induced autophagy in HepG2 and Huh7 cells. To further validate the inhibitory effect of FBXW5 knockdown on the AKT-MTOR pathway, FGF or IGF, activators of the AKT-MTOR pathway, were added together with siFBXW5. The results showed that FBXW5 knockdown observably attenuated the induction of p-AKT and MTOR downstream effectors p-EIF4EBP1, p-RPS6KB/p-p70S6K upon FGF or IGF treatment, indicating that FBXW5 knockdown markedly suppressed the AKT-MTOR pathway (Figure 4B-C). These results collectively demonstrated that FBXW5 knockdown inactivated the AKT-MTOR pathway in HCC cells.

Figure 4.

FBXW5 knockdown inactivated the AKT-MTOR pathway dependent on AQP3 accumulation. (A) HepG2 and Huh7 cells were transfected with siFBXW5 for 72 h and then collected and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), AKT, LC3, FBXW5, p-TSC2(Ser1387), TSC2, p-GSK3(Ser9), GSK3, p-RPS6KB/p-p70S6K(Thr389), RPS6KB/p70S6K, p-EIF4EBP1(Thr37/46), and EIF4EBP1. (B) Huh7 cells were transfected with siFBXW5 for 72 h and then treated with FGF for 0, 15, 30 or 60 min. The cells were collected and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), AKT, p-RPS6KB(Thr389), RPS6KB, p-EIF4EBP1(Thr37/46), EIF4EBP1, and FBXW5. (C) Huh7 cells were transfected with siFBXW5 for 72 h and then added with IGF for 0 min, 15 min, 30 min or 60 min. The cells were collected and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), AKT, p-RPS6KB(Thr389), RPS6KB, p-EIF4EBP1(Thr37/46), EIF4EBP1, and FBXW5. (D) AQP3 silencing blocked FBXW5 knockdown-inhibited activation of AKT-mTOR pathway in HepG2 and Huh7 cells. HepG2 and Huh7 cells were transfected with siAQP3 and siFBXW5 for 96 h and then collected and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), p-RPS6KB (Thr389), RPS6KB, p-EIF4EBP1(Thr37/46), LC3, AQP3, and FBXW5. (E) Huh7 cell was transfected with flag-AQP3 (0, 0.75, 1.5 and 3 μg) for 48 h and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), p-RPS6KB(Thr389), p-EIF4EBP1(Thr37/46) and LC3.

We next determined whether FBXW5 knockdown-inactivated the AKT-MTOR pathway was dependent on AQP3 accumulation. The results showed that downregulation of AQP3 significantly rescued the inhibition of p-AKT and MTOR downstream effectors p-EIF4EBP1, p-RPS6KB/p-p70S6K, indicating that AQP3 accumulation was essential for the inactivation of the AKT-MTOR pathway by FBXW5 knockdown (Figure 4D). To explore whether ectopic overexpression of AQP3 resemble the phenotype of FBXW5 knockdown, AQP3 was ectopically expressed in HCC cells. Our findings indicated that AQP3 overexpression markedly suppressed cell proliferation (Fig. S4A) and promoted apoptosis (Fig. S4B), effects analogous to those resulting from FBXW5 knockdown in Huh7 cells. Moreover, overexpression of AQP3 also dramatically promoted autophagy in Huh7 cell (Figure 4E). Additionally, overexpression of AQP3 significantly inhibited PDPK1 expression and AKT-MTOR pathway in HCC cells (Figure 4E). These results collectively demonstrated that overexpression of AQP3 induced cell autophagy via PDPK1-AKT-MTOR pathway similar to FBXW5 knockdown in Huh7 cell.

Downregulation of FBXW5 induced lysosomal degradation of PDPK1 to inhibit AKT-MTOR pathway

To investigate potential downstream effectors of AQP3 that may contribute to the inactivation of AKT-MTOR pathway induced by FBXW5 knockdown, we conducted a label-free quantitative proteomic assay. This assay aimed to identify AQP3-interacting proteins in HepG2 cells, which could potentially play a role in the regulation of AKT-MTOR pathway. Mass spectrometry results revealed that PDPK1 ranked as the top potential interacting protein among all candidates associated with AQP3. Extensive research has documented that autophagy is activated through PDPK1-AKT-MTOR signaling pathway, implying the involvement of PDPK1 in the regulation of AKT-MTOR pathway inhibition triggered by FBXW5 knockdown. To address this, we first performed confocal microscopy to detect whether PDPK1 was co-localized with AQP3. We found that PDPK1 interacted with AQP3 in both HepG2 and Huh7 cells (Figure 5A). We further performed a GST-affinity-isolation assay to confirm the interaction between PDPK1 and AQP3 (Figure 5B). Additionally, we assessed whether FBXW5 knockdown influenced the expression of PDPK1. Our findings demonstrated that FBXW5 knockdown led to a notable reduction in PDPK1 expression levels in both HepG2 and Huh7 cells (Figure 5C). Building upon these findings, we formulated the hypothesis that the downregulation of PDPK1 resulting from FBXW5 knockdown was linked to the accumulation of AQP3. To verify this, we downregulated the expression of AQP3 via siRNA and found that AQP3 knockdown markedly attenuated the inhibition of PDPK1 expression, indicating that AQP3 was required for FBXW5 knockdown-induced PDPK1 reduction (Figure 5D). Furthermore, to ascertain the pivotal role of PDPK1 in the inhibition of the AKT-MTOR pathway and autophagic cell death induced by FBXW5 knockdown, we introduced ectopic PDPK1 expression in conjunction with FBXW5 knockdown in Huh7 cells. Remarkably, we observed that the expression of PDPK1 notably mitigated the suppression of proliferation, as evidenced by the dramatic decrease of proliferation and colony formation (Figure 5E-G).

Figure 5.

Downregulation of FBXW5 induced lysosomal degradation of PDPK1 to inhibit AKT-MTOR pathway. (A) Colocalization of AQP3 and PDPK1 in HepG2 and Huh7 cells. Confocal microscopy analysis of PDPK1 (red) and AQP3 (green) in HepG2 and Huh7 cells. Nuclei were stained with DAPI. White arrows indicate the place where AQP3 and PDPK1 were colocalized. Scale bar: 25 μm. (B) GST affinity isolation. GST-tagged PDPK1 were expressed in E. coli and purified. Bound proteins were eluted from the beads and examined by western blotting with anti-AQP3 antibody. (C) Downregulation of FBXW5 inhibited the expression of PDPK1. HepG2 and Huh7 cells were transfected with siFBXW5 for 72 h and cells were collected and subjected to IB analysis for the expression of PDPK1. (D) Downregulation of AQP3 rescued the reduction of PDPK1 expression-induced by downregulation of FBXW5. HepG2 and Huh7 cells were transfected with siAQP3 and siFBXW5 for 96 h and then collected and subjected to IB analysis for the expression of PDPK1. (E-G) overexpression of PDPK1 attenuated proliferation inhibition induced by downregulation of FBXW5. (E) Huh7 cells were transfected with siFBXW5 for 48 h and then transfected with flag-PDPK1 plasmid for 48 h. The cell proliferation was detected by ATPlite assay. (F–G) Statistical analysis and representative images of colony formation assays of Huh7 cells. (H) Overexpression of PDPK1 rescued the inhibition of AKT-MTOR pathway induced by downregulation of FBXW5. Huh7 cells were transfected with siFBXW5 for 48 h and then were transfected with flag-PDPK1 plasmid for 36 h. Huh7 cells were collected and subjected to IB analysis for the expression of p-AKT(Ser473), p-AKT(Thr308), p-RPS6KB (Thr389), p-EIF4EBP1, flag-PDPK1, LC3, and FBXW5. (I) PDPK1 expression increased in the presence of CQ or BafA1, two lysosome inhibitors, and did not increase in the presence of MG132, a classical proteasome inhibitor. HepG2 cells were treated with CQ, BafA1 or MG132 for indicated time and collected and subjected to IB analysis for the expression of PDPK1. (J) CQ induced PDPK1 accumulation in HepG2 cell. HepG2 cells were treated with 10 μM CQ for 0 h, 12 h, 24 h or 48 h and cell lysates were assessed by IB with specific antibody against PDPK1. (K) CQ rescued the reduction of PDPK1 expression-induced by downregulation of FBXW5. HepG2 cells were transfected with siFBXW5 for 72 h and then treated with 20 μM CQ for 24 h. The HepG2 and Huh7 cells were collected and subjected to IB analysis for the expression of PDPK1. (L) CQ reversed the degradation of PDPK1 induced by FBXW5 knockdown. HepG2 cells were transfected with siControl and DMSO, siFBXW5 and DMSO, siFBXW5 and CQ for 48 h and then treated with 50 μg/mL of CHX for indicated period of time. The cells were collected for IB analysis with specific antibody against PDPK1. (M) Downregulation of FBXW5 induced the expression of HSPA8. HepG2 cells were transfected with siFBXW5 for 72 h and cells were collected and subjected to IB analysis for the expression of HSPA8. (N) HepG2 cell was transfected with flag-FBXW5 (0, 0.75, 1.5 and 3 μg) for 48 h and subjected to IB analysis for the expression of HSPA8. (O) Immunoprecipitation and western blot analysis were used to verify the interaction between HSPA8 with AQP3, PDPK1 and LAMP2A. (P) Downregulation of HSPA8 rescued the reduction of PDPK1 expression-induced by downregulation of FBXW5. HepG2 cells were transfected with siHSPA8 and siFBXW5 for 96 h and then collected and subjected to IB analysis for the expression of PDPK1.

In mechanism, ectopic overexpression of PDPK1 significantly upregulated the phosphorylation status of AKT, RPS6KB/p70S6K and EIF4EBP1, indicating that the inhibition of the AKT-MTOR pathway was reversed upon PDPK1 upregulation (Figure 5H). In addition, PDPK1 expression significantly reversed the induction of autophagy, as demonstrated by the significant decrease in the conversion of LC3-I to LC3-II (Figure 5H). These findings demonstrated that downregulation of FBXW5 induced autophagic cell death through AKT-MTOR pathway by reducing PDPK1 in HCC cells.

To characterize the underlying mechanism of PDPK1 reduction by FBXW5 knockdown, we first determined the transcriptional expression of PDPK1 elicited by FBXW5 knockdown. As shown in Fig. S5A-B, FBXW5 knockdown did not significantly decrease the mRNA levels of PDPK1, suggesting that FBXW5 knockdown may regulate PDPK1 degradation at the protein level. Subsequently, we sought to determine whether the degradation of PDPK1 was mediated through the lysosome pathway or the proteasome pathway. As shown in Figure 5I, PDPK1 expression obviously increased in the presence of CQ and BafiA1, two lysosome inhibitors. However, PDPK1 expression did not increase in the presence of MG132, a classical proteasome inhibitor (Figure 5I). These results indicated that PDPK1 degradation was likely regulated via the lysosome pathway. Moreover, we found that the expression of PDPK1 time-dependently increased after CQ treatment in HepG2 cells (Figure 5J). To confirm whether the decrease of PDPK1 expression induced by FBXW5 knockdown attributed to the lysosomal degradation of PDPK1, we added the lysosomal inhibitor CQ for the rescue experiment. We found that CQ significantly reversed PDPK1 reduction induced by FBXW5 knockdown (Figure 5K, Fig. S5C). Furthermore, the protein stability of PDPK1 in FBXW5-silenced cells with lysosomal inhibitors was measured in the presence of CHX. We found that CQ significantly reversed PDPK1 degradation induced by FBXW5 knockdown (Figure 5L). These observations collectively indicate that FBXW5 knockdown predominantly triggers PDPK1 degradation through the lysosomal pathway.

We explored the mechanisms by which AQP3 negatively regulates PDPK1. Given that HSPA8/HSC70 (heat shock protein family A (Hsp70) member 8) is a key protein in chaperone-mediated autophagy, an essential lysosomal degradation pathway [52–54], we assumed that AQP3 facilitates PDPK1 degradation via HSPA8/HSC70. Consequently, we examined HSPA8/HSC70 expression following FBXW5 downregulation and observed significant accumulation of HSPA8/HSC70 (Figure 5M). In addition, overexpression of AQP3 significantly induced HSPA8/HSC70 expression (Figure 5N). Co-immunoprecipitation assays demonstrated interactions between HSPA8/HSC70 and AQP3, PDPK1, as well as LAMP2A on the lysosomal membrane, which aided in the translocation of target proteins into lysosomes for degradation (Figure 5O). We then investigated whether the inhibition of PDPK1 expression due to FBXW5 knockdown was depended on increased HSPA8/HSC70 levels. The data revealed that reducing HSPA8/HSC70 levels substantially mitigated the suppression of PDPK1 expression caused by FBXW5 knockdown (Figure 5P). In sum, these findings suggest that the promotion of PDPK1 degradation by AQP3 after FBXW5 knockdown is HSPA8/HSC70-dependent.

FBXW5 knockdown suppresses the growth of human HCC tumors in murine model

After determining the suppressive impact of FBXW5 knockdown on cancer cell proliferation, we proceeded to validate the therapeutic potential of targeting FBXW5 in inhibiting tumor growth in vivo. We first constructed FBXW5 and AQP3 respectively or simultaneously downregulated stable HepG2 cell lines. As shown in Figure 6A-C, compared with the negative control group, the tumor formation of FBXW5 downregulated group was dramatically inhibited, as revealed by the tumor growth curve (Figure 6B) and tumor weight analysis (Figure 6C). Consistently, tumor growth suppression induced by FBXW5 downregulation was rescued by simultaneous downregulation of AQP3, suggesting the inhibitory effect of targeting FBXW5 on tumor formation was mediated by AQP3 accumulation (Figure 6A-C). We further addressed whether FBXW5 downregulation induced autophagy through modulating AQP3-PDPK1 in vivo. As shown in Figure 6D, FBXW5 downregulation significantly induced AQP3 accumulation, PDPK1 inhibition and autophagy occurrence, which was rescued by simultaneous downregulation of AQP3. These observations indicated that FBXW5 downregulation promoted AQP3 accumulation to induce autophagy and inhibited the growth of liver cancer both in vitro and in vivo.

Figure 6.

FBXW5 regulates the degradation of AQP3 in vivo and an inverse correlation between FBXW5 and AQP3 in HCC. (A) Mice were sacrificed and tumor tissues were harvested and photographed at the end of study. (B) Tumor size of both models was determined by caliper measurement, and the data were converted to tumor growth curves. (C) Mice were sacrificed and tumor tissues were harvested and weighed at the end of study. (D) Proteins extracted from tumor tissues were analyzed by IB using anti-FBXW5, AQP3, PDPK1, and LC3 antibodies. GAPDH was used as a loading control. (E) HCC tissue arrays, comprising both normal and tumor specimens, were subjected to immunohistochemical staining to detect AQP3 expression. The stained tissues were categorized into four levels (+ to ++++) based on the intensity of AQP3 staining. Specific antibodies against AQP3 were utilized for the immunohistochemical staining of human HCC tissue arrays. (F) The distribution of staining intensities was quantified as percentages within the normal and tumor tissues. (G–H) Twelve paired HCC and adjacent non-tumorous tissues were lysed and analyzed by western blotting. “T” denotes tumor tissue, while “A” indicates adjacent non-tumorous tissue. (I) Schema of the mechanism for FBXW5 knockdown induced autophagic cell death in HCC cells.

FBXW5 and AQP3 are negatively correlated in liver cancer tissues

To examine the clinical significance of AQP3 in HCC, we assessed AQP3 expression in 88 paired samples of HCC and corresponding normal liver tissues using immunohistochemistry/IHC. The samples were stratified into four categories based on the intensity of immunohistochemistry staining, ranging from the lowest (+) to the highest (++++), as shown in Figure 6E. We observed that AQP3 expression in HCC tissues was markedly elevated compared to normal tissues in the + and ++ groups (Figure 6F). Conversely, in the +++ and ++++ groups, AQP3 expression was reduced in HCC tissues relative to normal liver tissues (Figure 6F). To elucidate the potential involvement of FBXW5 in the regulation of AQP3 degradation during tumorigenesis, we compared FBXW5 and AQP3 levels in tumor and adjacent tissues. An inverse correlation between FBXW5 and AQP3 was initially confirmed by Western blot analysis in 6 matched pairs of HCC and adjacent liver tissues (Figure 6G). Furthermore, correlation analysis revealed a pronounced inverse association between FBXW5 overexpression and reduced AQP3 levels in HCC patients (n = 12 pairs, Figure 6H). In conclusion, our findings highlight the negative regulatory role of FBXW5 on AQP3 expression in clinical outcomes of patients with HCC.

Discussion

Aquaporin 3 (AQP3), a member of the AQPs family, is an aquaglyceroporin which transports water, glycerol and small solutes across the plasma membrane. Apart from its involvement in regulating fundamental biological functions in normal cells, AQP3 assumes a critical role in tumorigenesis, encompassing aspects like tumor cell proliferation, migration, invasion, and tumorigenesis [10–14]. Nevertheless, the upstream signaling pathways regulating AQP3 protein stability remains largely elusive. In the present study, we reported for the first time to our knowledge that FBXW5 knockdown induced autophagic cell death in HepG2 and Huh7 cells. Mechanistically, FBXW5 knockdown induced the accumulation of AQP3 to inactivate PDPK1-AKT-MTOR pathway to initiate autophagy (Figure 6I). These findings revealed a pivotal role of the AQP3-PDPK1-AKT-MTOR axis in FBXW5 regulating the autophagic death of HCC cells.

AQP3 plays a crucial role in oncogenesis, influencing processes such as tumor cell proliferation, migration, invasion, and tumorigenesis. Its function as either a tumor suppressor or an oncoprotein is contingent upon the type of cancer. Previous studies have identified AQP3 as an oncoprotein across various tumors. For instance, AQP3 enhances cervical cancer cell invasion and metastasis by modulating NOX4-derived H₂O₂, which activates the SYK-PI3K-AKT signaling pathway [55]. Conversely, AQP3 acts as a tumor suppressor in specific cancers, especially in prostate and breast malignancies [17,18]. Our research revealed that curtailing FBXW5 expression induces AQP3 accumulation, thereby hindering hepatocellular carcinoma cell proliferation. In addition, AQP3 expression was substantially reduced in HCC tissues compared to normal hepatic tissues. We also uncovered a notable inverse correlation between FBXW5 overexpression and decreased AQP3 levels in HCC patients. Taken together, these findings imply that AQP3 functions as a potential tumor suppressor in HCC.

AQP3 serves as a membrane transporter responsible for the movement of water and glycerol across plasma membranes, assuming a crucial role in driving cancer progression, metastasis, and the epithelial-to-mesenchymal transition/EMT [56]. Recently, AQP3 has been reported as an initiator for autophagy [57,58]. Some observations support the notion that AQP3 enhances autophagy to promote the growth of gastric cancer cells, as well as that AQP3 overexpression promotes autophagy to induce resistance to chemotherapeutic drug cisplatin in gastric carcinoma cells [57,58]. These findings suggested that AQP3 regulated cell protective autophagy. However, in our study, we found that AQP3 regulated autophagic death of HCC induced by FBXW5 knockdown. We further elucidated the mechanism of AQP3 regulated autophagic death of HCC. We found that AQP3 accumulation by FBXW5 knockdown induced PDPK1 reduction, thereby inhibiting AKT-MTOR pathway to promote autophagic death of HCC. Therefore, we elucidated a brand-new mechanism of AQP3 promoting autophagic death in HCC cells.

Prior investigations have suggested that FBXW5 was associated with autophagy in cancer cells [59,60]. One previous study reported that FBXW5 regulated the degradation of TSC2 [35], a typical inhibitor of autophagy-related AKT-mTOR pathway [61,62], suggesting that FBXW5 knockdown may induce autophagy via AKT-MTOR pathway. Our results showed that downregulation of FBXW5 promoted autophagy by inactivating the AKT-MTOR pathway, but did not affect TSC2 expression in HCC cells (Figure 4A), which inspired us to further explore the other potential mechanism of FBXW5 knockdown affecting AKT-MTOR pathway-mediated autophagy. In our study, for the first time, we demonstrated that AQP3 was a potential substrate of the SCF^FBXW5 E3 ligase. More importantly, we found that FBXW5 knockdown induced AQP3 accumulation and eventually led to autophagic death of HCC cells by inactivating PDPK1-AKT-MTOR pathway, uncovering a novel mechanism of FBXW5 knockdown inducing autophagy through AKT-MTOR pathway. Intriguingly, we revealed that downregulation of AQP3 significantly restored the autophagy induced by FBXW5 knockdown, but not completely, suggesting other mechanisms may also be involved in the autophagic death of HCC cells upon FBXW5 downregulation.

PDPK1 (3-phosphoinositide-dependent protein kinase 1) appears to play a central regulatory role in autophagy signaling of several cancer cells [63–67]. For instance, it was reported that overexpression of mitochondrial membrane protein ANKRD22 inhibited kinase activity of PDPK1, promoted glycolysis and induced autophagy in colorectal cancer cells [67]. Furthermore, the serine protease HTRA2 was demonstrated to inhibit PDPK1 expression and promote autophagy via AKT-MTOR pathway in HCC cells [68]. However, the underlying upstream mechanism for the induction of PDPK1-elicited autophagic cell death remains largely unknown. In our study, we revealed that FBXW5 knockdown dramatically inhibited PDPK1 expression to promote autophagy dependent on AQP3 accumulation in HCC cells. We further elucidated that AQP3 accumulation induced by FBXW5 knockdown caused the degradation of PDPK1 in lysosomal pathway, which revealed a new mechanism for regulating PDPK1 induced autophagy.

In summary, our study emphasized the crucial involvement of the SCF^FBXW5 E3 ligase in mediating AQP3 ubiquitination and degradation, and unraveled a previously undisclosed mechanism through which FBXW5 curbed autophagic cell death via the AQP3-PDPK1-AKT-MTOR axis in HCC cells.

Materials and methods

Cell lines, culture and reagents

Human hepatocellular carcinoma cell lines HepG2 and Huh7 were cultured in Dulbecco’s Modifed Eagle’s Medium (Hyclone, SH30002.03), containing 10% FBS (Biochrom AG, S4115) and 1% penicillin-streptomycin solution, at 37°C with 5% CO₂. Chloroquine (CQ) (Selleck, S6999), bafilomycin A₁ (BafA1) (Sigma-Aldrich, 5.08409), 3-methyladenine (3-MA) (Sigma-Aldrich, M9281), cycloheximide (CHX) (Sigma-Aldrich 239,765).

Establishment of stable cell lines

HepG2 and Huh7 cells expressing EGFP-LC3 fusion protein were established as described. Briefly, cells were seeded in six-well plates and transfected with 3 μg pc3.1-EGFP-LC3 plasmid using Lipofectamine 2000 (Invitrogen 11,668,030). Cells with EGFP fluorescence were selected by MoFlo XDP Cell Sorter (Beckman Coulter, USA) and cultured in complete cell culture medium containing G418 (Merck 345,810) at 200 μg/ml. The autophagy was measured by appearance of punctate vesicle structure and photographed under fluorescence microscope (Leica, Wetzlar, Germany).

Cell viability

Cells were inoculated into 96-well plates (2 × 10³ cells per well) and transfected with the indicated siRNAs. According to the manufacturer’s protocol, cell proliferation was measured by ATPlite luminescence analysis kit (PerkinElmer 6,016,731). For clonogenic assay, the cells were seeded into six-well plates (500 cells per well) and cultured for 12 days. The colonies on the plate were fixed with 4% paraformaldehyde, stained with crystal violet and counted.

Detection of apoptosis

Cells were inoculated into 6-well plates (3 × 10⁵ cells per well) and transfected with the indicated siRNAs. Apoptosis was determined with the ANXA5-FITC/PI Apoptosis Kit (BD Biosciences 556,547) according to the manufacturer’s instructions.

Immunoblotting

Cell lysates were prepared for immunoblotting (IB) analysis with antibodies as follows. LC3 (Cell Signaling Technology, 2775), ATG5 (Cell Signaling Technology, 2630), p-AKT(308) (Cell Signaling Technology 13,038), p-AKT(473) (Cell Signaling Technology, 4060), AKT (Cell Signaling Technology, 2630), p-TSC2 (Cell Signaling Technology, 5584), TSC2 (Cell Signaling Technology, 4308), p-GSK3 (Cell Signaling Technology, 5558), GSK3 (Cell Signaling Technology 12,456), p-RPS6KB/p-p70S6K (Cell Signaling Technology, 9234), RPS6KB/p70S6K (Cell Signaling Technology, 2708), p-EIF4EBP1 (Cell Signaling Technology, 2855), EIF4EBP1 (Cell Signaling Technology, 9452), ubiquitin (Cell Signaling Technology, 2775), CUL1 (Cell Signaling Technology, 4995), RBX1/ROC1 (Cell Signaling Technology 11,922), SKP1 (Cell Signaling Technology, 2156), AQP3 (Abcam 125,219), LAMP2A (Abcam, ab25631), PDPK1/PDK1 (Abcam, ab52893), FBXW5 (Abclone, WG-00512D), HSPA8/HSC70 (abclonal A0415), FLAG (Sigma-Aldrich, F7425), GAPDH (Cwbiotech, CW0100M), ACTB/β-actin (Cwbiotech, CW0096M). ImageJ software was used for densitometric analysis.

Gene silencing using siRNA

HepG2 and Huh7 cells were transfected with siRNA oligonucleotides, synthesized by Ribobio (Shanghai, China), using Lipofectamine 2000. The sequences of siRNA are as follows:

siFBXW5–1: GGATGTGGAGTCAGAGAAC;

siFBXW5–2: CCACAGGCGCCAAGAGCAA;

siATG5: ATGTCCTCTGCGCTGGAAT;

siAQP3: GGAUCAAGCUGCCCAUCUA;

siCUL1-1: CUAGAUACAAGAUUAUACAUGCGG;

siCUL1-2: GGUUAUAUCAGUUGUCUAA;

siRBX1-1: GACTTTCCCTGCTGTTACCTAATT;

siRBX1-2: CTGTGCCATCTGCAGGAACCACATT;

siSKP1-1: CGCAAGACCUUCAAUAUCATT.

siSKP1-2: GCAAGACCTTCAATATCAA.

SiHSPA8: UAAUUCUAAGUACAUUGAGACCAGC

Immunoprecipitation and western blot

Cells were harvested and lysed in Nonidet p-40 lysate (Beyotime Biotechnology, P0013F) with protease inhibitors (PMSF; ThermoFisher Scientific 36,978), cocktail (Merck, p8340) and protein phosphatase inhibitors (ThermoFisher Scientific 78,420) for 20 min on ice, and centrifuged at 13,000 g for 15 min. The lysate was incubated with the indicated antibodies or anti-Flag M2 agarose (Merck, A2220) at 4°C with rocking overnight. Protein-A/G beads (Santa Cruz Biotechnology, sc-2003) were then added or not, and the incubation was continued for an additional 2 h. The proteins bound to the beads were washed three times with NETN buffer. The beads were boiled for 10 min at 95°C, followed by SDS-PAGE, and analyzed by immunoblotting with the indicated antibodies. For western blot, whole cell lysate was obtained by using RIPA buffer (50 mM Tris-HCl, pH 8.0, 0.1% SDS, 150 mM NaCl, 1% Nonidet p-40, 0.5% sodium deoxycholate [Merck, D6750] and protease inhibitor cocktail.

Plasmids construction and transfection

To generate Flag-FBXW5, HA-FBXW5, Flag-AQP3 and Flag-PDPK1 constructs, Human FBXW5, AQP3 or PDPK1 were amplified by PCR and cloned into the modified pc3.1-FLAG vector. Human FBXW5 was amplified by PCR and cloned into the modified pc3.1-HA vector. All constructs were verified by sequence analysis. Plasmids transfection were carried out using Lipofectamine 2000 following the manufacturer’s instructions.

CHX chase assay

For AQP3 or PDPK1 half-life assay, HepG2 and Huh7 cells were harvested at the indicated time after cycloheximide (100 μg/mL; Sigma-Aldrich 239,765) treatment and then subjected to immunoblotting analysis. HepG2 and Huh7 cells were transfected with the indicated siRNAs. Seventy-two h after transfection, cells were treated with cycloheximide (100 μg/mL), harvested at indicated time points and then subjected to IB analysis.

In vivo ubiquitination assay

For AQP3 ubiquitination analysis, HepG2 and Huh7 cells were transfected with sicontrol, siFBXW5, siCUL1 or siRBX1/ROC1 as indicated 72 h. After transfection, cells were lysed in 1% SDS lysis buffer (1% SDS, 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM DTT) and boiled for 10 min. For immunoprecipitation, the lysates were diluted 10-fold in 0.25% NETN buffer (0.25% Nonidet p-40, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA). Immunoprecipitation was performed using AQP3 antibody. Analyses of ubiquitination were performed by anti-ubiquitin blotting.

RNA isolation and Q-PCR

Total RNA was extracted by Ultrapure RNA kit (Cwbiotech, CW0581M). RNA (1.0 μg) was purified and reverse transcribed using PrimeScript® RT Master (Takara, RR047A) according to the manufacturer’s instructions. The cDNA was quantifed by real-time quantitative PCR using SYBR® Green Real-Time PCR Master Mixes (Applied Biosystems 4,309,155) and a 7500 Real-time PCR system (Applied Biosystems) according to the manufacturer’s instructions. The primer sequences were as follows: 5’-CGGAACAAGCCCTATGACG-3’ (forward primer) and 5’- GTTCTCTGACTCCACATCCTG- 3’ (reverse primer) for the human FBXW5 gene; 5’-CTCTGGACACTTGGATATGAT-3’ (forward primer) and 5’- AGCACACACACGATAAGG- 3’ (reverse primer) for the human AQP3 gene; 5’-TCAGGACGACGAGAAGCTGTAT-3’ (forward primer) and 5’- AACGTACTGCGCTGTTCCCACG- 3’ (reverse primer) for the human PDPK1 gene; 5’-CACTCTTCCAGCCTTCCTTC-3’ (forward primer) and 5’- GTACAGGTCTTTGCGGATGT- 3’ (reverse primer) for the human ACTB gene;

GST affinity isolation

Monoclonal cultures for GST-PDPK1 fusion protein production were selected and inoculated into 20 ml of LB medium supplemented with the appropriate antibiotics. These cultures were incubated at 37°C overnight. IPTG (ThermoFisher Scientific 15,529,019) was then added to a final concentration of 1 mM to induce protein expression for 4 h at the same temperature. Post-incubation, the cultures were transferred into centrifuge tubes and spun at 15,000 g for 1 min at 4°C. The supernatant was removed, and the pellet was resuspended thoroughly in 100 μl of lysis buffer. After a 10 min centrifugation at 15,000 g and 4°C, 20 μl of 50% BeyoGold™ GST-tag Purification Resin (Beyotime, P2262) was added to the homogenate. The mixture was agitated gently on a shaker at 4°C for 30 min to ensure complete binding of the GST-tagged target protein to the resin. The mixture was then centrifuged at 1000 g for 10 s at 4°C, allowing the resin to settle. Resins were washed widely with lysis buffer, and bound proteins were separated by 12% SDS-PAGE.

Generation of stable cell lines

FBXW5 knockdown was performed by FBXW5 sgRNA oligos: forward 5’-CACCGCCAGTTCGCGTCCTGCTCCA-3’ and reverse 5’-AAACTGGAGCAGGACGCGAACTGGC-3’ into LentiCRISPR v2 plasmid (Addgene 52,961; deposited by our laboratory). AQP3 knockdown was performed by AQP3 sgRNA oligos: forward 5’-CACCGTGATCCAGGGCTCACGAGCC-3’ and reverse 5’-AAACGGCTCGTGAGCCCTGGATCAC-3’ into LentiCRISPR v2 plasmid (Addgene 52,961; deposited by our laboratory). LentiCRISPR plasmid with sgRNA (5.0 μg, packaging plasmids psPAX2 (4.0 μg; Addgene 12,260; deposited by our laboratory) and pMD2.G (3.0 μg; Addgene 12,259; deposited by our laboratory) were transfected into HEK293T cells with Lipofectamine 2000, and virus supernatant was harvested 36 h post transfection and mixed with polybrene to increase infection efficiency. The infected HepG2 cells were treated with 2 μg/mL puromycin (Merck 540,411) for 1 week.

Subcutaneous xenograft in mice

Male BALB/c athymic nude mice (ages 5–7 weeks) were randomized into four groups and treated in accordance with established guidelines, and the protocol was approved by an internal animal protocol review committee (LHERAW-21070). Tumor xenografts were measured with a caliper every 3 days, and tumor volume was determined using the formula: (length × width²)/2. The investigators were blinded to the group allocation during the experiment and when assessing the outcome. Subcutaneous tumors were collected and subjected to immunoblotting analysis after mice were sacrificed.

Statistical analysis

The statistical significance of differences between groups was assessed using GraphPad Prism5 software. The student t test was used for the comparison of parameters between groups. Data are presented as mean ± standard deviation. The level of significance was set at p < 0.05. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. = not significant.

Funding Statement

This work was funded by the Shanghai Frontiers Science Center of Disease and Syndrome Biology of Inflammatory Cancer Transformation [2021KJ03-12], the National Natural Science Foundation of China [82002973, 82372984, 82272987, 82172933], the Chinese Minister of Science and Technology grant [2016YFA0501800], Science and Technology Commission of Shanghai Municipality [21ZR1482200], Shanghai Rising-Star Program [23QA1403000].

Disclosure statement

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

Data availability statement

The data analyzed during this study are included in this published article and the supplemental data files. Additional supporting data are available from the corresponding authors upon reasonable request.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2024.2353497.