Aldob expression is negatively correlated with Akt activation in human HCC and the expression of Aldob/p-Akt predicts overall survival of HCC patients

Our recent study has revealed a novel tumor suppressive role for the glycolytic enzyme Aldob in HCC through directly binding to G6PD and inhibiting its activity, acting as a metabolic switch in glucose metabolism and regulating the metabolic reprogramming [18]. Accumulating body of evidence has demonstrated that metabolic reprogramming in cancer cells to meet increased bioenergetic and biosynthetic requirements during tumorigenesis depends on multiple intracellular signaling pathways [19]. Among them, activated Akt signaling has a profound impact on metabolic reprogramming through up-regulating glycolytic enzymes and promoting aerobic glycolysis [20]. Therefore, we set out to examine the role of Akt signaling in HCC in the context of Aldob down-regulation and metabolic reprogramming. We first investigated the clinical relevance between Aldob expression and Akt activation in human HCC. As shown in Fig 1A, reduced Aldob expression and enhanced Akt phosphorylation levels at both the threonine 308 (pT308-Akt) and serine 473 (pS473-Akt) residues were observed in tumor tissues as compared to matched adjacent normal liver tissues. Next, we performed tissue microarray (TMA) analysis on paired clinical samples from HCC patients (Figs 1B and S1A and S1 Table, n = 70). Aldob expression was negatively correlated with Akt activation in human HCC tissues (Fig 1C, R² = 0.146, p = 0.001). We also found that low Aldob expression was significantly correlated with α-fetoprotein (AFP), albumin level, and tumor encapsulation (S1 Table). Kaplan–Meier survival analysis showed that the overall survival time for patients with low Aldob expression was notably shorter than those with high Aldob expression (Fig 1D, p = 0.016). Conversely, patients with high pT308-Akt level exhibited shorter overall survival compared with patients with low pT308-Akt level (Fig 1E, p = 0.029). In addition, an elevated level of p-Akt was significantly correlated with prothrombin time, number of tumors, and tumor encapsulation (S1 Table). More importantly, a low Aldob expression coupled with high pT308-Akt levels was associated with the worst overall prognosis (Fig 1F, p = 0.001). In conclusion, our study uncovers a novel inverse correlation between Aldob and p-Akt expression in human HCC and low Aldob expression with high p-Akt predicts the worst prognosis for HCC patients, suggesting an important role of Akt signaling in HCC in the context of Aldob downregulation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. A negative correlation between Aldob and p-Akt expression in human HCC specimens.

(A) IB analysis of WCLs derived from 10 pairs of human HCC tumors and matched adjacent nonmalignant tissues. (B) Representative IHC staining images of Aldob and pT308-Akt from the same HCC patients on tissue microarray were shown (original magnification ×200). (C) The relative IHC score of Aldob and pT308-Akt from tissue microarray of 70 HCC patients was plotted and assessed by a linear regression analysis. Protein expression score is classified into negative (0–20), weak (21–100), moderate (101–180), and strong (181–255). The score calculation formula is (weak + moderate + strong)/(negative + weak + moderate + strong) × 100%. (D and E) Overall survival of HCC patients (n = 70) grouped by high IHC score (≥50) or low IHC score (<50) of Aldob (D) or pT308-Akt (E) expression levels were conducted through the Kaplan–Meier analysis and log-rank test. The number of patients at risk was listed below each curve. (F) Overall survival of HCC patients (n = 70) grouped by different combination of Aldob and pT308-Akt levels were conducted through the Kaplan–Meier analysis and log-rank test. The number of patients at risk was listed under each curve. The data underlying this figure can be found in S1 Data. Aldob, aldolase B; HCC, hepatocellular carcinoma; IB, immunoblot; IHC, immunohistochemistry; WCL, whole cell lysate.

https://doi.org/10.1371/journal.pbio.3000803.g001

Aldob inhibits Akt kinase activity and downstream signaling in HCC

To investigate the relationship of Aldob down-regulation and activation of Akt signaling, we observed that, in liver-specific ALDOB knockout (KO) (ALDOB^f/f; Alb-Cre) mouse primary hepatocytes (PHs), Aldob deletion significantly augmented Akt activation compared to wild-type (WT) (ALDOB^f/f) PHs, as evidenced by increased levels of pT308-Akt, pS473-Akt, and Akt downstream targets, including pS9-GSK3β, pT389-S6K, hexokinase 1 (HK1), and HK2, without affecting the expression of Akt, GSK3β, and S6K (Fig 2A). Similarly, we observed that loss of Aldob resulted in elevated expression levels of pT308-Akt, pS473-Akt, pS9-GSK3β, and mechanistic target of rapamycin complex 1 (mTORC1) direct downstream target pT389-S6K in liver tumor tissues of global ALDOB KO mice compared to WT mice after diethyl nitrosamine (DEN) treatment (Figs 2B and S1B). However, there was no significant change of the phosphorylation level of mTORC2 downstream target protein kinase Cα (PKCα) (S1B Fig). Notably, the expression of HK1 and HK2 were significantly increased in ALDOB KO mice, while HK4 expression was decreased (Figs 2B and S1B). Previous study has demonstrated that HCC cells are metabolically distinct from normal hepatocytes through expressing the high-affinity HK1 and HK2 and inhibiting the low-affinity glucokinase (HK4) to increase glucose uptake and metabolism [21]. These results suggest that loss of Aldob may facilitate glucose metabolism through activation of Akt signaling in HCC. In combination with our recent study that loss of Aldob promotes DEN-induced HCC tumorigenesis in ALDOB KO mice [18], we hypothesized that Aldob exhibits a potential tumor-suppressive role through inhibiting oncogenic Akt activity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Aldob suppresses Akt phosphorylation and kinase activity.

(A) IB analysis of WCL derived from conditional ALDOB KO and WT mouse primary hepatocytes. (B) IB analysis of WCL derived from liver tumor tissues of ALDOB KO and WT mice after injection with DEN at postnatal day 14 to induce hepatocellular carcinoma for 10 months. (C) IB analysis of WCL derived from Huh7 and LM3 cells stably expressing Aldob via lentiviral infection (with Vector as a negative control). (D) IB analysis of WCL derived from Huh7-Aldob cells transfected with indicated siRNAs (si-NC as a negative control). (E and F) IB analysis of WCL derived from Huh7-Aldob cells transfected with indicated siRNAs. Resulting cells were serum-starved for 24 hours and then stimulated with 0.1 μM insulin (E) or 100 ng/ml EGF (F) at the indicated time before harvesting. (G) In vitro Akt1 kinase assay was conducted with recombinant GSK3α as a substrate and immunoprecipitated Akt1 from Huh7-Vector or Huh7-Aldob cells as the source of kinase. IgG was used as a negative control. The data underlying this figure can be found in S1 Data. Aldob, aldolase B; DEN, diethyl nitrosamine; EGF, epidermal growth factor; GSK3α, glycogen synthase kinase 3α; Huh7-Aldob, Aldob-overexpressing Huh7; IB, immunoblot; IgG, immunoglobulin G; KO, knockout; siRNA, small interfering RNA; WCL, whole cell lysate; WT, wild-type.

https://doi.org/10.1371/journal.pbio.3000803.g002

To further address the role of Aldob in the regulation of Akt phosphorylation in HCC, we generated stable Aldob-overexpressing liver cancer cell lines Huh7 and LM3 and found that ectopic expression of Aldob resulted in a marked reduction of Akt, GSK3β, and S6K phosphorylation levels (Fig 2C). Conversely, silencing Aldob by small interfering RNA (siRNA) restored Akt phosphorylation and its downstream targets (Fig 2D). Next, we analyzed the kinetics of Akt activation and attenuation in response to insulin (Fig 2E) and epidermal growth factor (EGF) (Fig 2F) stimulation. Down-regulation of Aldob led to increased levels of pT308-Akt and pS473-Akt with coordinate changes in immediate Akt downstream substrate pS9-GSK3β expression in a time-dependent manner (Fig 2E and 2F). Additionally, we performed an in vitro Akt kinase assay and demonstrated that overexpressed Aldob dramatically inhibited Akt kinase activity, as assessed by the decreased phosphorylation level of Akt substrate GSK3α (Fig 2G). Taken together, all these results strongly suggest that Aldob suppresses Akt phosphorylation and its kinase activity in HCC.

Aldob inhibits Akt kinase activity and suppresses HCC through impeding cell cycle progression and attenuating glycolysis and TCA metabolism

Given that aberrant activation of Akt signaling promotes cancer progression through regulating cell proliferation, metabolism, and survival [3], we next investigated the functional consequence of Aldob-mediated suppression of Akt kinase activity. We observed that enforced expression of Aldob inhibited cell viability and colony formation capability. Consistently, inhibition of Akt activity by MK2206, a highly selective allosteric Akt inhibitor [22], significantly blocked cell proliferation and eliminated the growth inhibitory effects of Aldob expression (Fig 3A and 3B). Furthermore, activation of Akt signaling by ectopic expression of Myc-Akt1 significantly enhanced cell survival and abolished Aldob overexpression-induced cell growth arrest (S2A and S2B Fig). These results indicate that Akt-dependent signaling plays an essential role in Aldob-mediated suppression of HCC cell proliferation. Meanwhile, Aldob overexpression or inhibiting Akt activity by MK2206 significantly increased the proportion of cells in the G1-phase, resulting in a decreased percentage of cells in S-phase, indicating that Aldob induces cell cycle arrest at the G1-to-S transition (Fig 3C). Consistent with this observation, up-regulated Aldob obviously reduced pT308-Akt, pS473-Akt, pS9-GSK3β, and CyclinD1 expression, while elevating the levels of cell cycle inhibitors p21 and p27 (Fig 3D). More importantly, either Akt inhibition using MK2206 or Akt activation by ectopic expression of Myc-Akt1, but not the Myc-Akt1 phospho-deficient mutant (Myc-Akt1-AA), dramatically diminished Aldob-mediated abovementioned tumor-suppressive effects (Figs 3C and 3D and S2C–S2F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Aldob-mediated inhibition of Akt activity is required for Aldob-induced tumor-suppressive effects.

(A) Huh7-Vector and Huh7-Aldob cells were treated with DMSO or MK2206 (2 μM) at the indicated time and cell viability was measured by CCK-8 assays. (B) Huh7-Vector and Huh7-Aldob cells (5,000 cells/well in 6-well plates) were cultured with or without MK2206 (2 μM) for 10 days before being fixed and stained. Representative graphs of cell colonies were shown. (C and D) Cells were treated with DMSO or MK2206 (2 μM) for 48 hours, and then monitored for cell cycle distribution by FACS (C) or subjected to IB analysis with the indicated antibodies (D). (E) Glucose levels in the culture medium of Huh7-Vector and Huh7-Aldob cells after treatment with DMSO or MK2206 (5 μM) at different time points. (F) Schematic presentation of ¹³C distribution in glycolysis and the first turn of the TCA cycle with 50% U-¹³C-glucose (labeled at all 6 carbons, red circles) and 50% ¹²C-glucose (unlabeled, black circles). (G and H) Fraction of the labeled metabolites of M+3 from ¹³C-glucose in glycolysis (G) and fraction of the labeled metabolites of M+2 from ¹³C-glucose in TCA cycle (H) by DMSO or MK2206 (5 μM) treatment for 12 hours in Huh7-Vector and Huh7-Aldob cells. (I–K) Huh7-Vector and Huh7-Aldob cells were injected subcutaneously into the flanks of nude mice, followed by treatment with Control solvent or MK2206 (n = 6 mice). Representative tumor images (I), xenograft tumor weight (J), and tumor size (K) in each group were recorded. Data are presented as mean ± SEM. * p < 0.05; ** p < 0.01 (Student t test). The data underlying this figure can be found in S1 Data. Aldob, aldolase B; CCK-8, Cell Counting Kit-8; FACS, fluorescence-activated cell sorting; F1,6P, fructose-1,6-bisphosphate; G3P, glyceraldehyde-3-phosphate; G6P, glucose-6-phosphate; Huh7-Aldob, Aldob-overexpressing Huh7; IB, immunoblot; TCA cycle, tricarboxylic acid cycle.

https://doi.org/10.1371/journal.pbio.3000803.g003

Metabolic reprogramming has been recognized as the core hallmark for various cancers [23]. Enhanced aerobic glycolysis, known as the Warburg effect, is frequently observed along with abnormally high rates of glucose uptake in rapidly proliferating cancer cells [24]. As the central route of oxidative phosphorylation, the tricarboxylic acid (TCA) cycle is also up-regulated to meet increased cellular energy, biosynthesis, and redox needs [25,26]. The downstream effectors of Akt signaling play a central role in cancer cell metabolic reprogramming through regulating glucose uptake and metabolism. To further determine the potential roles of Aldob-mediated inhibition of Akt activity on glucose metabolism, we observed that glucose levels were time-dependently decreased in the cell culture medium (Fig 3E). Interestingly, Huh7-Vector cells depleted glucose from the culture medium more efficiently than Huh7-Aldob cells. Furthermore, inhibition of Akt activity by MK2206 significantly suppressed glucose consumption and mitigated Aldob overexpression-induced inhibitory effects on glucose consumption (Fig 3E). Next, we used stable isotope labeled [U-¹³C₆] glucose as a tracer to track intracellular metabolic flux (Fig 3F). The M+3 (¹³C labeled at all 3 positions) fractions of enriched labeled metabolites in glycolysis including pyruvate, lactate, and alanine were significantly decreased in Huh7-Aldob cells compared with control (Fig 3G). Meanwhile, Aldob overexpression also observably reduced the M+2 (¹³C labeled at 2 positions) fractions of labeled metabolites in TCA cycle, such as fumarate, malate, aspartate, glutamate, and citrate (Fig 3H). Besides, blocking Akt activity by MK2206 efficiently weakened Aldob-induced inhibition of metabolic flux in glycolysis and TCA cycle (Fig 3G and 3H). These results suggest that Aldob suppresses Akt activity, resulting in the reduction in glucose consumption and cellular metabolism in glycolysis and TCA.

On the other hand, endoribonuclease-prepared siRNA (esiRNA)-mediated knockdown of endogenous Aldob expression in Huh7 cells significantly promoted tumor cell proliferation, glucose consumption, lactate production, and activation of Akt signaling (S3A–S3D Fig). Conversely, Akt inhibition using MK2206 reversed Aldob knockdown-induced oncogenic properties (S3A–S3D Fig). Furthermore, knockdown of Aldob via siRNA in Huh7-Aldob cells markedly enhanced Akt activity, promoted cell proliferation and cell cycle progression, increased glucose consumption and the levels of metabolites in glycolysis and TCA cycle, and rescued overexpressed Aldob-mediated inhibition of cell growth and metabolism (S4A–S4J Fig). Moreover, MK2206 exposure also reversed the growth-promoting effects of Aldob knockdown in Huh7-Aldob cells (S4A–S4J Fig). We found similar patterns of cell proliferation and cell cycle progression effects in LM3 cells (S5A–S5D Fig). These results suggest that Aldob impedes cell cycle progression and represses glucose metabolism through inhibiting Akt signaling, which contributes to Aldob-induced cell growth suppression.

We then established a subcutaneous xenograft mouse model to investigate the effects of Aldob expression on tumorigenesis in vivo. As illustrated in Fig 3I–3K, tumors derived from Huh7-Aldob cells exhibited slower growth and smaller sizes than those from Huh7-Vector cells. Inhibition of Akt activity by MK2206 alone significantly inhibited tumor growth. Additionally, a combination of Aldob overexpression and inhibition of Akt activity showed an additive effect on tumor growth inhibition. Interestingly, Huh7-Vector cells appeared to be more responsive to MK2206 than Huh7-Aldob cells in which Akt signaling was already suppressed at baseline. Moreover, Ki67 expression was consistent with the differences noted in tumor weights and size (S6A Fig). Likewise, western blot analysis revealed that Aldob overexpression and MK2206 treatment alone significantly inhibited Akt activation and its downstream signaling compared to control group, respectively, and the combination of the two had the lowest levels of Akt signaling-related proteins (S6B Fig). No significant changes in body weight were observed throughout the study (S6C Fig). Together, these data demonstrate that Aldob exhibits tumor suppressive effects through inhibition of Akt signaling in vitro and in vivo; inhibition of Akt activity by the allosteric inhibitor suppresses tumorigenesis caused by the loss of Aldob.

Aldob directly interacts with Akt to suppress Akt activity and HCC cell growth, whereas disruption of Aldob/Akt interaction in Aldob mutant cells restores Akt activity and cell proliferation

Next, we tested whether Aldob modulated Akt activity through a protein–protein interaction by performing immunoprecipitation (IP) experiments. As shown in Fig 4A, endogenous Aldob pulled down Akt1 in WT mice liver tissues. Reciprocally, Aldob was detected in the endogenous Akt1 immunocomplex. However, ALDOB KO significantly weakened immunoprecipitated interacting proteins, suggesting that Aldob and Akt1 bind with each other. Consistently, the interactions between overexpressed Aldob and endogenous or exogenous Akt1 were observed by co-IP assays in HCC cell lines (Figs 4B and 4C and S5E). Knockdown of Aldob or Akt1 expression by siRNA greatly decreased Aldob interaction with Akt1 (Fig 4B). Moreover, we detected a direct interaction between the recombinant glutathione S-transferase (GST)-Akt1 and His-Aldob performed by GST pull-down analysis (Fig 4D). In addition, Aldob and Akt2 combined with each other (S7A Fig). Collectively, these results demonstrate that Aldob physically interacts with Akt.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. The direct interaction of Aldob with Akt is essential for Aldob to suppress Akt activity and HCC cell growth.

(A) IB analysis of endogenous anti-Aldob and anti-Akt1 co-IP and WCL derived from liver tissues of ALDOB KO and WT mice. IgG was used as a negative control. (B) IB analysis of anti-Akt1 and anti-Flag IP and WCL derived from Huh7 cells transfected with the indicated constructs and siRNAs. IgG was used as a negative control. (C) IB analysis of co-IP and WCL derived from Huh7 cells transfected with the indicated constructs. (D) GST pull-down analysis to demonstrate the direct interaction between recombinant proteins GST-Akt1 and His-Aldob. Recombinant GST protein was used as a negative control. (E) IP analysis was performed with WCL derived from Myc-Akt1-transfected Huh7 cells and recombinant protein WT-His-Aldob or their various His-Aldob mutants harboring a lysine or arginine to alanine mutation in the identified active site residues of Aldob. The Aldob enzyme activity in Aldob mutants relative to WT-His-Aldob were calculated (bottom). (F) IB analysis of WCL derived from Huh7 cells transfected with WT or R304A mutant form of Flag-Aldob constructs. (G–I) Huh7 cells transfected with the indicated constructs were used to determine their effects on cell proliferation (G), colony formation (H), and cell cycle distribution (I). Data are presented as mean ± SEM. * p < 0.05; ** p < 0.01 (Student t test). The data underlying this figure can be found in S1 Data. Aldob, aldolase B; co-IP, co-immunoprecipitation; GST, glutathione S-transferase; HCC, hepatocellular carcinoma; IB, immunoblot; IgG, immunoglobulin G; IP, immunoprecipitation; KO, knockout; siRNA, small interfering RNA; WCL, whole cell lysate; WT, wild-type.

https://doi.org/10.1371/journal.pbio.3000803.g004

Hepatic Aldob is indispensable for glucose and fructose metabolism, and its deficiency as result of ALDOB gene mutation leads to HFI in humans [27]. The residues at Arg43, Arg46, Lys108, Lys147, Arg149, Lys 230, and Arg304 line within the active site pocket of Aldob, and play an important role in substrate binding and catalysis [28,29]. Mutations involved in aforementioned residues have been identified in the ALDOB gene of HFI patients [30,31]. To determine whether the enzymatic activity of Aldob is required for this novel interaction, we individually mutated the 7 positively charged amino acid residues to alanine and found that among the 4 point mutants (R43A, K108A, K230A, and R304A) with considerably reduced Aldob-Akt interaction, they exhibited varied Aldob enzymatic activities: R43A and K230A (completely dead), K108A (32.1%), and R304A (49.1%) of WT Aldob (Fig 4E). Among them, R304A mutant has the most significantly attenuated interactions with Akt. Moreover, the other 3 mutants (R46A, K147A, and R149A) with moderately attenuated Aldob-Akt interaction also had variable enzyme activities: R46A maintained 60.2% activity and the other two did not have significant enzyme activity (Fig 4E). Together, these data clearly indicate that the interaction of Aldob with Akt is independent of Aldob enzymatic activity.

To further identify the Akt interacting region on Aldob, a full-length and 2 truncated mutants of Aldob including a.a. 1–155 (amino acid 1–155 fragment containing clone) and a.a. 1–240 were constructed to examine their Akt-binding capability. Intriguingly, both Aldob truncated mutants failed to interact with Akt compared to the full-length Aldob, suggesting that the carboxyl-terminal region of Aldob (a.a. 241–364) is most likely the region for Akt interaction (S7B Fig), consistent with the observation that R304A has the weakest Aldob-Akt interaction. We next tested whether Aldob-Akt interaction is essential for Aldob regulation of Akt activity and HCC cell growth. Overexpression of WT Aldob significantly down-regulated Akt signaling, whereas disruption of Akt/Aldob interaction in R304A mutant abrogated Aldob-induced inhibition of Akt signaling (Fig 4F). Consistently, R304A mutant efficiently rescued cell viability, colony formation, and cell cycle progression in Aldob-overexpressing cells (Fig 4G–4I). Together, these data indicate that Aldob directly interacts with Akt to suppress Akt activity and that disruption of this interaction releases Akt kinase activity to promote cancer cell growth.

Aldob preferentially associates with Akt in a phosphorylation-dependent manner

Because the state of Akt phosphorylation is critical for the integration of various extracellular cues to downstream signaling events for multiple biological processes, we next studied whether the phosphorylation state of Akt played any role in the regulation of the protein–protein interaction between Akt and Aldob. To this end, we examined the effect of Akt1 phospho-deficient mutants (Akt1-T308A and Akt1-S473A) on Aldob-binding capability. Compared with Akt1-WT, both Akt1 phospho-deficient mutants displayed reduced binding with Aldob, suggesting that Akt1 kinase activity is critical for Akt1 and Aldob interaction (Fig 5A). Next, we treated Aldob- and Akt1-transfected Huh7 cells with insulin or EGF to further examine the correlation between Akt1 phosphorylation status and Aldob interaction with Akt1. Interestingly, the interaction of Aldob and Akt1 tracked with the kinetics of Akt1 activation and attenuation in response to either insulin or EGF, with the binding between Aldob and Akt1 almost peaking when the levels of phosphorylated Akt (p-Akt) were the highest (Fig 5B and 5C). Furthermore, blocking Akt phosphorylation through MK2206 treatment prominently decreased Aldob interaction with Akt1 (Fig 5D). Together, these results indicate that Aldob preferentially binds to the phosphorylated form of Akt.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Aldob preferentially interacts with the phosphorylated species of Akt1.

(A) IB analysis of anti-Flag IP and WCL derived from Huh7 cells transfected with the indicated constructs. T308A and S473A indicated that threonine at position 308 and serine at position 473 of Akt1 were replaced by alanine, respectively. (B and C) IB analysis of anti-Flag IP and WCL derived from Huh7 cells transfected with Flag-Aldob and Myc-Akt1. Resulting cells were serum-starved for 24 hours and then stimulated with 0.1 μM insulin (B) or 100 ng/ml EGF (C). At the indicated time points, WCL were harvested for Flag-IP and further for IB analysis (left). The relative intensity of Myc-Akt1 immunoprecipitated by Flag-Aldob was quantified (right). (D) IB analysis of anti-Flag IP and WCL derived from Flag-Aldob and Myc-Akt1 transfected Huh7 cells that were treated with DMSO or MK2206 (1 μM) for 24 hours before harvesting. The data underlying this figure can be found in S1 Data. Aldob, aldolase B; EGF, epidermal growth factor; IB, immunoblot; IP, immunoprecipitation; WCL, whole cell lysate.

https://doi.org/10.1371/journal.pbio.3000803.g005

Aldob promotes Akt interaction with PP2A and facilitates PP2A-mediated Akt dephosphorylation

Given that Akt interacted with Aldob in an Akt phosphorylation-dependent manner, we hypothesized that phosphatases of Akt may be involved in Aldob-induced repression of Akt activity. Interestingly, as shown in Fig 2, Akt phosphorylation at both T308 and S473 was significantly reduced in response to Aldob expression. Previous study suggested that PP2A directly dephosphorylates Akt at both T308 and S473 [32,33]. Moreover, phosphorylation of Akt at T308 has been identified to play a more crucial role in activating Akt [34,35]. Thus, we examined the possible involvement of the serine/threonine phosphatase PP2A in Aldob-mediated regulation of Akt activity. Strikingly, we observed an abundance of the catalytic subunit of PP2A (PP2A-C), but little PHLPP, in the exogenous Aldob immunocomplex (Fig 6A). Notably, neither interaction of Aldob with the Akt kinase PDK1, the structural subunit (PP2A-A), nor the regulatory subunit (PP2A-B55α) of PP2A was observed (Figs 6B and S7C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Aldob regulates Akt activity through protein phosphatase PP2A.

(A and B) IB analysis of anti-Flag IP and WCL derived from Huh7 cells transfected with the indicated constructs. (C) GST pull-down assays were performed with WCL derived from Flag-PP2A-C-transfected Huh7 cells and recombinant proteins GST-Akt1 and WT-Aldob or Aldob-R304A mutant. GST was used as the negative control. (D) In vitro PP2A dephosphorylation assays were performed with phosphorylated Myc-Akt1 and active PP2A in the presence or absence of recombinant WT-Aldob or Aldob-R304A mutant (top). The relative signal intensities of pT308-Akt and pS473-Akt were quantified (bottom). (E) IB analysis of WCL derived from Huh7 cells transfected with the indicated constructs. Resulting cells were treated for 4 hours with OA at 100 nM. (F) IB analysis of endogenous anti-Aldob and anti-Akt1 IP and WCL derived from 2 pairs of human HCC samples. IgG was used as the negative control. The data underlying this figure can be found in S1 Data. Aldob, aldolase B; IB, immunoblot; IP, immunoprecipitation; GST, glutathione S-transferase; IgG, immunoglobulin G; OA, okadaic acid; PP2A, protein phosphatase 2A; WCL, whole cell lysate; WT, wild-type.

https://doi.org/10.1371/journal.pbio.3000803.g006

Since PP2A has been shown to directly bind to Akt and dephosphorylate Akt [36], we speculated that Aldob might nucleate the Akt interaction with PP2A to trigger Akt dephosphorylation resulting in inhibition of downstream signaling. To support this hypothesis, we performed GST pull down assay and found that WT Aldob, but not the Akt1-binding deficient Aldob-R304A mutant, significantly enhanced Akt1 interaction with PP2A-C (Fig 6C). Moreover, IP analysis showed that Aldob overexpression markedly facilitated PP2A-C interaction with Akt1 and Aldob without a significant impact on the interactions between PP2A-C and PP2A-A or B55α subunits (S7D Fig). Similarly, endogenous Akt1 pulled down both Aldob and PP2A-C in the liver tissue of WT mice, whereas loss of Aldob drastically diminished Akt1-immunoprecipitated Aldob and PP2A-C in ALDOB KO mice (Fig 4A), further suggesting that Aldob acts as a scaffold protein for the Akt/PP2A complex. Next, we performed in vitro PP2A phosphatase activity assay using phosphorylated Akt as a substrate and observed a time-dependent decrease of p-Akt at both T308 and S473 residues upon incubation with active PP2A (Fig 6D). More importantly, WT Aldob remarkably augmented the PP2A-induced reduction of p-Akt expression, while no significant effect was found on the Akt1-binding deficient Aldob-R304A mutant (Fig 6D). Furthermore, we examined the effect of okadaic acid (OA) on these processes. OA is an inhibitor of serine/threonine protein phosphatase PP1 and PP2A, exhibiting greater selectivity toward the latter up to concentration of 100 nM [37]. As shown in Fig 6E, OA treatment effectively restored Akt phosphorylation levels in Aldob-overexpressing cells, which almost comparable to untreated control cells, suggesting that Aldob-induced inhibition of Akt activity is dependent on PP2A. To further verify these results in human HCC samples, we randomly selected 2 pairs of HCC tissue samples for IP assays. The endogenous interaction between Aldob and Akt1 was identified in peripheral normal liver tissues of HCC patients (Fig 6F). However, down-regulated Aldob in tumor tissues significantly decreased Akt1-immunoprecipitated Aldob and PP2A-C when compared to corresponding adjacent normal tissues (Fig 6F), further suggesting that Aldob/Akt/PP2A signaling is involved in human HCC progression. Taken together, Aldob interacts with p-Akt to promote the recruitment of the phosphatase PP2A and accelerate PP2A-mediated dephosphorylation, which is primarily responsible for Aldob-induced down-regulation of Akt phosphorylation.

PP2A activation using SMAP down-regulates Akt activity and suppresses HCC progression

PP2A is generally considered as a tumor suppressor that dephosphorylates multiple critical oncogenic proteins, such as Akt, extracellular signal-regulated kinase (ERK), and MYC [38]. Functional inactivation of PP2A has been linked to tumor development in many cancers [39]. Thus, pharmacologic restoration of PP2A phosphatase activity has emerged as an attractive strategy for cancer therapy [40]. Recent studies have reported that a series of specific SMAPs reengineered from tricyclic neuroleptics effectively activate PP2A, resulting in the dephosphorylation of key targets Akt and ERK, and blocking tumor growth of lung, prostate cancer, and pancreatic neuroendocrine tumors both in vitro and in vivo [41–44]. The validation of their target specificity, mechanism of action, and pharmaceutic properties have been extensively documented [41–44]. Here, we treated Huh7 and LM3 cells with increasing concentrations of SMAP. SMAP-induced PP2A activation was demonstrated by parallel reduction of Akt, ERK, and GSK3β phosphorylation levels, which was accompanied by decreased CyclinD1 expression (S8A Fig). Moreover, SMAP treatment reduced cell viability in both cell lines in a dose-dependent manner with IC₅₀ of approximately 9 μM (Huh7) and 12 μM (LM3), similar to those reported in other tumor cell types (S8B Fig). SMAP treatment induced visible HCC cell death. These results indicate that PP2A activation with SMAP results in growth inhibition and cell death in HCC cell lines.

To evaluate the utility of PP2A activation toward Aldob-mediated suppression of Akt activity, Huh7-Vector and Huh7-Aldob cells were treated with 10 μM SMAP for further functional studies. We found that PP2A activation with SMAP effectively reversed the promoting effects of esi-RNA-mediated Aldob knockdown on cell proliferation, glucose consumption, lactate production, and activation of Akt signaling (Fig 7A–7D). Furthermore, Aldob overexpression or SMAP treatment alone inhibited cell proliferation, and the combination of the two showed an additive effect on repressing cell proliferation, as measured by cell viability and clonogenic assays (Fig 7E and 7F). Additionally, cell cycle analysis revealed that SMAP treatment resulted in a significant increase in the G1 phase with a coordinate loss of the S population, and exacerbated Aldob-induced G1-to-S phase cell cycle arrest (Fig 7G). Western blot analysis showed Aldob overexpression had no effect on ERK phosphorylation in the presence of either DMSO or SMAP, whereas Aldob-induced down-regulation of Akt/GSK3β/CyclinD1 signaling was potentiated by SMAP-mediated PP2A activation (Fig 7H). Notably, relative to Huh7-Aldob cells that contain diminished Akt activity, Huh7-Vector cells were more sensitive to SMAP treatment. Together, these results indicate that PP2A plays an essential role in Aldob-induced growth inhibition and that small molecule mediated PP2A activation is beneficial to Aldob-mediated suppression of Akt signaling and HCC cell growth.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Activation of PP2A with SMAP potentiates Aldob-mediated tumor-suppressive role.

(A–D) Huh7 cells were transfected with esiRNA for knockdown of endogenous Aldob, and cell proliferation (A), glucose consumption (B), lactate production (C), and the protein expression levels in Akt signaling pathway (D) were determined in the presence of control DMSO or SMAP (10 μM). (E) Cell proliferation of Huh7-Vector and Huh7-Aldob cells in the presence of DMSO or SMAP (10 μM). (F) Representative graphs from colony formation assay of Huh7-Vector and Huh7-Aldob cells after exposure to DMSO or SMAP (10 μM) for 14 days. (G) Huh7-Vector and Huh7-Aldob cells were treated with 10 μM SMAP for 24 hours, and cell cycle distribution was monitored by FACS. (H) IB analysis of WCL derived from Huh7-Vector and Huh7-Aldob cells treated for 20 hours with control DMSO or SMAP (10 μM). (I–K) Representative tumor images (I), xenograft tumor volume (J), and tumor weight (K) of Huh7-Vector and Huh7-Aldob xenograft tumors from control solvent and SMAP-treated nude mice (n = 6 mice). Data are presented as mean ± SEM. * p < 0.05; ** p < 0.01 (Student t test). The data underlying this figure can be found in S1 Data. Aldob, aldolase B; esiRNA, endoribonuclease-prepared siRNA; FACS, fluorescence-activated cell sorting; IB, immunoblot; PP2A, protein phosphatase 2A; SMAP, small-molecule activator of PP2A; WCL, whole cell lysate.

https://doi.org/10.1371/journal.pbio.3000803.g007

We further investigated the efficacy of SMAP on HCC tumor progression in vivo. SMAP treatment alone resulted in dramatic inhibition of tumor growth, similar to Aldob overexpression-induced growth-inhibitory effects, whereas overexpression of Aldob and SMAP treatment achieved additive tumor suppressive effects (Fig 7I–7K). It is noteworthy that there was no significant weight loss or behavioral abnormalities during the 3-week treatment, suggesting that SMAP is well tolerated in this mouse model (S8C Fig). In addition, western blot analysis in SMAP-treated tumors detected the simultaneous decrease in Akt/GSK3β/CyclinD1 signaling and ERK phosphorylation levels, suggesting that SMAP induces PP2A-dependent inhibition of these signaling pathways resulting in suppression of tumor growth (S8D Fig). Immunohistochemistry (IHC) analysis validated lower protein levels of Ki67 in SMAP-treated tumors compared with control vehicle-treated tumors (S8E Fig). More importantly, up-regulated Aldob-induced attenuation of Akt signaling and tumor growth was greatly enhanced in response to SMAP treatment (Figs 7I–7K and S8D and S8E). Collectively, these results support the potential therapeutic application of PP2A phosphatase reactivation for the treatment of HCC with Aldob deficiency.

Ethics statement

This study was approved by the Clinical Research Ethics Committee of the Eastern Hepatobiliary Surgery Hospital, Shanghai, China. Written informed consent was obtained from all patients. All specimens were pathologically confirmed at the Eastern Hepatobiliary Surgery Hospital. Animal studies were approved by the Institutional Animal Care and Use Committee of the Shanghai Institutes for Biological Sciences of the Chinese Academy of Sciences (approval number 2015-AN-2).

Mouse models

Mice were housed in pathogen-free animal care facilities in Shanghai Institutes for Biological Sciences (SIBS), Chinese Academy Science (CAS) and allowed access to water and chow diet ad libitum. All animal experimental procedures were conformed to SIBS Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of SIBS, CAS. For subcutaneous xenograft growth model, Huh7-Aldob or Vector cells (2 × 10⁶) mixed in serum-free medium containing 50% Matrigel (BD Biosciences, San Jose, California, United States of America) were subcutaneously injected into the left and right flanks of 5-week-old male BALB/C nude mice. About 10 days after cell injection, when the tumors were visible and reached 100 mm³, the mice were randomly assigned to control solvent and agent treatment groups (n = 6 per group). MK2206 (40 mg/kg, once every other day) and SMAP (5 mg/kg, twice a day) were administered via oral gavage. Tumor growth and body weight were monitored every other day for 3 weeks. Tumor volume was estimated using the following formula: volume = 0.52 × length × width². Animals were observed for signs of toxicity (weight loss, abdominal stiffness, and ruffled fur). Tumors were harvested, photographed, and weighted 2 hours after the final dose of treatment, following snap-frozen in liquid nitrogen for immunoblotting or fixation in 4% paraformaldehyde (PFA) for IHC. IHC staining was performed to examine the protein expression of proliferation marker gene Ki67 in the tumor tissues.

The whole-body ALDOB KO mouse model and the liver-specific ALDOB KO (ALDOB^f/f;Alb-Cre) mouse model were generated as previously described [18]. ALDOB KO and WT mice were injected with a single dose of DEN (40 mg per kg body weight) at postnatal day 14. These mice were sacrificed at the age of 40 weeks, and mouse liver, tumor tissues, and plasma were collected.

Cell culture and HCC tissue samples

The human HCC cell lines Huh7 and HCC-LM3 were purchased from Cell Bank of CAS. Primary hepatocytes were isolated from the livers of male liver-specific ALDOB KO and WT mice using collagenase perfusion as previously described [66]. HCC cell lines and primary hepatocytes were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Hyclone, Logan, Utah, USA) plus with 10% fetal bovine serum (FBS, Hyclone) and 1% penicillin-streptomycin (Gibco, Carlsbad, California, USA) in a 37°C incubator with 5% CO₂ in air.

Human HCC tissues were collected immediately after hepatectomy from the Eastern Hepatobiliary Surgery Hospital, Shanghai. Collection of patient specimens with informed consent to an established protocol and all experimental procedures were approved by the Research Ethics Committee of Eastern Hepatobiliary Surgery Hospital. The fresh specimens were snap-frozen in liquid nitrogen and stored at −80°C until analysis.

Reagents

MK2206 (Selleck Chemicals, Houston, Texas, USA) was dissolved in DMSO (Sigma-Aldrich, St. Louis, Missouri, USA) and diluted in DMEM with a final concentration of 0.1% DMSO for in vitro studies or in the vehicle solution containing Cremophor (Sigma-Aldrich), 95% ethanol and water in a ratio of 1:1:6 for in vivo studies. EGF (Sigma), insulin (Invitrogen), and OA (Cell Signaling Technology, Boston, Massachusetts, USA) were used at the indicated doses. The PP2A activator SMAP compound (provided by Goutham Narla, The University of Michigan, Ann Arbor, Michigan) was dissolved in DMSO and stored at room temperature and used for up to 1 month for in vitro studies. For in vivo studies, SMAP was prepared in an N, N-Dimethylacetamide (DMA)/Kolliphor HS-15 (Solutol)/diH₂O solution.

Plasmids and viruses

Aldob, Akt1, and Akt2 expression plasmids were constructed by cloning the open reading frame of each cDNA into the multiple cloning site of pCDNA3.0 vector, and their different Flag or Myc tag was constructed through design in 1 primer. The different gene fragments were amplified using their primers. The pcDNA3.0 vector and each PCR fragments were both digested with their restriction enzymes EcoRI and XhoI or BamHI and XhoI or EcoRI and XhoI, and then ligated with T4 ligase. Transformation was performed using DH5α competent cells. Single clones were selected and amplified to extract plasmids. All clones were sequenced to confirm the results. Flag-PP2A-C plasmids were obtained from Dr. Xie Dong (SIBS, CAS). Plasmid transfection and virus infection: Expression constructs were transfected into cells using Lipofectamine 2000 Transfection Reagent (Invitrogen) following the manufacturer’s protocol. Lentivirus expression clone pLVX-IRES-GFP-puromycin-Aldob with/without Flag-tag was constructed with the Aldob fragment digested from pcDNA3.0-Aldob mentioned above via restriction enzyme EcoRI and XhoI. This fragment was inserted into pLVX-IRES-GFP-Puro vector and sequenced to verify the results. Lentivirus was produced by co-transfection of 293T cells using Lipofectamine 2000 Transfection Reagent (Invitrogen) with 3 μg pCMV-dR8.91, 2 μg pCMV-VSV-G, and the target gene expression vectors (such as 5 μg pLVX-Aldob-Flag-puro). Fresh medium was transferred into the dish after transfection of 6 hours. After 48 hours, the virus was harvested from the medium filtered by 0.45 mm Steriflipfilter (Millipore, Boston, Massachusetts, USA). Virus infection was carried out by incubating cells with cell medium containing indicated virus and 5 μg/ml polybrene (Sigma) for 8 to 12 hours. Cells were allowed to recover in complete medium for 24 hours and then selected with antibiotic (such as puromycin) for 10 to 15 days until stable cell lines were screened and constructed.

RNA-mediated interference (RNAi)

siRNA-Aldob-2-F: 5′-ACCCUCUACCAGAAGGACAdTdT-3′, siRNA-Aldob-2-R: 5′-UGUCCUUCUGGUAGAGGGUdTdT-3′, siRNA-Aldob-3-F: 5′-GUAUGUUCACACGGGUUCUdTdT-3′, siRNA-Aldob-3-R: 5′-AGAACCCGUGUGAACAUACdTdT-3′. siRNA-Akt1-F: 5′-GCUACUUCCUCCUCAAGAAdTdT-3′, siRNA-Akt1-R: 5′-UUCUUGAGGAGGAAGUAGCdTdT-3′. The esiRNA targeting human ALDOB (EHU001201) was obtained from Sigma-Aldrich. The siRNA was transfected into cells using Lipofectamine 2000 Transfection Reagent (Invitrogen) for at least 48 hours and then subjected to different assays.

Western blotting and antibodies

Cells or tissues were lysed in ice-cold lysis buffer (25 mM Tris (pH 8.0), 150 mM NaCl, 1 mM CaCl₂, 1% Triton X-100) with EDTA-free protease inhibitors (Bimake, Houston, Texas, USA) and phosphatase inhibitors (Bimake) and determined the protein concentration though BCA protein assay kit (Beyotime Biotechnology, Shanghai, China). Equal amounts of total protein were subjected to 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and then transferred to polyvinylidene fluoride (PVDF, Millipore) membranes and incubated with the indicated antibodies. Antibodies to Aldob (18065-1-AP), GAPDH (60004-1-Ig), Tubulin (11224-1-AP), p70S6K (14485-1-AP), PHLPP (22789-1-AP), p21 (10355-1-AP), p27 (25614-1-AP), PDK1 (17086-1-AP), PP2A-C subunit (13482-1-AP), GST-tag (66001-2-AP), Hexokinase1 (19662-1-AP), Hexokinase2 (22029-1-AP), and Hexokinase4 (19666-1-AP) were purchased from Proteintech (Rosemont, USA). Antibodies to pT308-Akt (13038), pS473-Akt (4060), Akt1 (2938), pS9-GSK3β (5558), GSK3β (12456), pT389-S6K (9205), Myc-tag (2276), IgG (2729), pT202/Y204-ERK (4370), ERK (4695), pT638/641-PKCα (9375), PP2A-A subunit (2041), PP2A-B subunit (2290), and p70S6K (2708) were purchased from Cell Signaling Technology. Antibodies to Flag-tag (F3165) were purchased from Sigma. Antibodies against PKCα (sc-17769) were obtained from Santa Cruz. Antibodies to pT308-Akt (ab38449, for IHC use), CyclinD1 (ab16663), and Ki67 (ab15580, for IHC use) were purchased from Abcam (Cambridge, United Kingdom). After extensive rinse with TBST, immunoreactive bands were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies and visualized using a chemiluminescence (ECL) kit (Proteintech) through Tanon 5200 Chemiluminescent Imaging System. All western blots were analyzed with ImageJ for quantitative measurements.

Cell viability assay

The viability assay was performed by seeding cells in a 96-well plate at a density of 2,000 cells/well for 6 repeats. After 7 hours, cell media were replaced with fresh media with various concentrations of test agents or control DMSO, and cells were proceeded to grow. CCK8 (OBIO Cell Counting Kit) then was added to separate plates at the indicated time according to the manufacturer's instruction.

Colony formation assay

Colony formation assays were adapted from a protocol reported previously [34]. Cells of different groups were seeded (5,000 cells/well) into 6-well plates and cultured for 7 to 10 days. Medium were changed every other day. Colonies were washed with PBS, fixed with 10% acetic acid/10% methanol for 20 minutes, and then stained with 0.4% crystal violet in 20% ethanol for 20 minutes. After staining, the plates were washed and air-dried for colony visualization.

Cell cycle analysis

Cells of different groups were plated (2 × 10⁵ cells/well) into 6-well plates and subjected to indicated treatment. Resulting cells were detached with trypsin, washed with ice cold PBS, and fixed in 70% ethanol overnight. About 1 × 10⁶ cells were detected using the cell cycle analysis kit (Beyotime Biotechnology, C1052). Then, cells were incubated with propidium iodide (PI) and RNAase for 30 minutes at 37°C, followed by analysis via the flow cytometry (BD Bioscience). The results were presented as the proportion of cell population in each phase of the cell cycle, which was determined using FlowJo 7.6 software (Treestar, USA).

Measurement of glucose consumption and lactate production

Cells were seeded and cultured with the medium containing 2 g/l glucose. The culture medium was collected at indicated time points. The D-Glucose Kit (#10716251035, R-Biopharm) and the L-Lactic Acid Kit (#10139084035, R-Biopharm) were used to detect the concentration of glucose and lactate in the culture medium and plasma according to the manufacturer’s instructions.

Aldolase activity assay

Aldolase catalyzes the reversible reaction of fructose-1,6-bisphosphate (F-1,6-BP) to glyceraldehyde-3-phosphate and dihydroxyacetone phosphate. The measurement of Aldolase activity was performed using F-1,6-BP as the substrate by Aldolase Activity Colorimetric Assay Kit (#K665-100; BioVision, San Francisco, USA) according to the manufacturer’s instructions.

Metabolic flux experiments using [U-¹³C₆]-glucose and Gas chromatography/mass spectrometry

The metabolic flux analysis using [U-¹³C₆]-glucose followed our previously published protocol [18,63,67]. In short, cells of different groups were seeded at a density of approximately 2 × 10⁶ cells per 10-cm dish. After cells grew to approximately 60% confluence, the unlabeled medium was replaced with [U-¹³C₆]-glucose medium, which composed of low glucose DMEM (Gibco, 11054–020: 1 g/L unlabeled glucose, no glutamine) with supplement of 1 g/L [U-¹³C₆]-glucose (Cambridge Isotope Laboratory), 10% (v/v) fetal bovine serum, 1 mM pyruvate, 2 mM unlabeled L-glutamine, and 1% (v/v) penicillin-streptomycin with or without MK2206 treatment (5 μM) for 12 hours.

Immunoprecipitation (IP) and GST pull-down assay

A total of 1,000 μg cell lysates were incubated with the indicated antibody (1 to 2 μg) overnight at 4°C with gentle rotation, followed by 6-hour incubation with Protein A/G Sepharose beads (Santa Cruz). Then, beads were washed 4 times with IP lysis buffer (150 mM Tris-HCl, 400 mM NaCl, 0.8%Triton X-100 (pH 7.4)) and boiled in sample buffer before being resolved by SDS-PAGE and immunoblotted with indicated antibodies. For the Flag or Myc-tag labeled IP experiments, cell lysates were incubated with antibody-Flag or antibody-Myc affinity gel (Bimake) overnight at 4°C. Affinity gel were collected after 300g centrifuging for 4 minutes at 4°C, and then washed 4 times. The subsequent operations were similar to western blot analysis. For GST pull-down assay, recombinant GST-Akt1 or GST proteins were incubated with His-Aldob proteins or cell lysates derived from cells transfected with indicated constructs in binding buffer (25 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, 0.5% NP-40, 10 μg/μl BSA, 1 mM DTT (pH 7.3)). A total of 100 μl prepared Glutathione Sepharose 4B (GE Healthcare, Little Chalfont, Buckinghamshire, UK) were added to the mixture and incubated overnight. After washed in ice-cold binding buffer for 3 times, bound proteins were released by boiling in sample buffer and resolved by SDS-PAGE electrophoresis.

Expression and purification of recombinant proteins

The full-length Aldob and Akt1 genes were inserted into the pET-30a (Novagen, Madison, Wisconsin, USA) and pGEX4T1 (GE Healthcare) expression plasmids, respectively. The E. coli BL21 (DE3) Codon-Plus strain (Novagen) transformed with the reconstructed expression plasmid was cultured at 37°C in LB medium containing 0.05 mg/ml ampicillin until OD600 of medium reached 0.8, and addition of 0.1 mM IPTG induced the target protein expression at 16°C for 24 hours. The cells were further harvested by centrifugation and lysed by sonication in the lysis buffer (20 mM Tris-HCl, 300 mM NaCl, and 1 mM PMSF). The recombinant proteins were purified using GSH Magnetic Beads for GST tag protein purification (Bimake, B23702) and Nickel Magnetic Beads for His tag protein purification (Bimake, B23602) according to the manufacturer’s protocol. The purified proteins were determined by SDS-PAGE with high purity (>95%).

In vitro Akt kinase assay

The in vitro kinase assay (Abcam, ab65786) was performed according to the manufacturer’s instructions. Cell lysates were immunoprecipitated with either anti-Akt1 or anti-IgG antibody along with protein A/G beads overnight, and the immunoprecipitates were incubated with substrate GSK3α Protein/ATP mixture in kinase reaction buffer for 30 minutes at 30°C. The reactions were terminated by 6×SDS loading buffer and subjected to western blot analysis with antibody against pS21-GSK3α.

In vitro PP2A dephosphorylation assay

The in vitro PP2A dephosphorylation assay was performed using phosphorylated Myc-Akt1 as the substrate. Myc-Akt1 transfected Huh7 cells were serum-starved for 24 hours and then stimulated with 0.1 μM insulin for 1 hour before harvesting. Equal amounts of cell lysates were incubated with antibody-Myc affinity gel overnight at 4° C. The immunoprecipitates were washed 3 times in IP lysis buffer, 3 times in PP2A phosphatase assay buffer (20 mM HEPES (pH 7.2), 100 mM NaCl, and 3 mM DTT) and finally resuspended in PP2A phosphatase assay buffer. Then, immunoprecipitated phospho-Akt were reacted with recombinant active PP2A (BPS Bioscience, 30056) at 30°C for indicated time periods in the presence or absence of recombinant WT-Aldob or Aldob-R304A mutant. Dephosphorylation reactions were terminated by 6×SDS loading buffer and subjected to western blot analysis.

Tissue microarray and immunohistochemistry

IHC was constructed on the TMA with a 2-step immunoperoxidase technique [68]. In short, tissues were fixed overnight in 4% PFA, then dehydrated in ethanol and embedded in paraffin. The paraffin-embedded tissue pieces were cut into 5-μm thick sections, which were stained with hematoxylin and eosin (H&E) and processed for IHC. For IHC, specimen sections were dewaxed, rehydrated in ethanol, and subjected to antigen extraction. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide in methanol for 30 minutes. Sections were blocked by 2% BSA in PBS for 1 hour at 37°C, and then incubated with the indicated primary antibody overnight at 4°C. After washing 3 times with PBS the next day, the samples were further incubated for 1 hour at 37°C using the corresponding secondary antibodies. The slides were displayed by DAB staining. Reading of IHC slides and TMA was performed using Vectra 2 (Perkin Elmer, Waltham, Massachusetts, USA) to quantify protein expression in all samples. For TMA in Fig 1B, A Leica APERIO AT TURBO instrument was used for scanning and analysis. Protein expression intensity was stratified into negative (0 to 20 score), weak (21 to 100 score), moderate (101 to 180 score), and strong (181 to 255 score). The calculation formula was (weak + moderate + strong)/(negative + weak + moderate + strong) × 100%. Quantification of IHC staining was performed blindly. TMA antibody characterization of Aldob and pT308-Akt was shown in S1A Fig. For tissue IHC staining of Ki67 in S6A and S8E Fig, Ki67 IHC score was estimated using the following formula: IHC score = intensity × area. The intensity was graded as 0, negative; 1, weak; 2, moderate; and 3, strong. The area of positive staining was assessed by 0, 1% to 25%, 26% to 50%, 51% to 75%, and 76% to 100%, corresponding to 0, 1, 2, 3, and 4 grades, respectively.

Statistical analysis

All data are presented as the means ± standard errors of the mean from at least 3 independent experiments. The survival curve was calculated according to the Kaplan–Meier method, and differences were assessed by the log-rank test. Differences between measurable variants of 2 groups were calculated by the unpaired Student t test or 2-way ANOVA using GraphPad Prism 5.0, and the value of p < 0.05 was considered statistically significant.