NF-κB activation as a pro-survival signal from pharmacological inhibition of pyruvate dehydrogenase kinase 1 in non-small-cell lung carcinoma cell models.

Introduction
Targeting aberrant glucose metabolism in cancer cells represents a promising therapeutic approach [[1], [2], [3], [4]]. Pyruvate dehydrogenase complex (PDC) plays a crucial role in oxidative phosphorylation (OXPHOS) by decarboxylating pyruvate to form acetyl-CoA, which then enters the tricarboxylic acid (TCA) cycle and fuels metabolism through OXPHOS. However, the phosphorylation of PDC by pyruvate dehydrogenase kinase (PDK) leads to the inactivation of PDC, thereby redirecting glucose metabolism towards aerobic glycolysis. Inhibiting PDK activity could restore the function of PDC and the TCA cycle, facilitating the OXPHOS pathway, which could compromise the survival of cancer cells by restoring to normal metabolic state [[5], [6], [7]].
Recent studies have shown that inhibiting PDK induced apoptosis in cancer cells. Firstly, altering PDK expression by knocking out or knocking down PDK4 was found to induce apoptosis in cancer cells [8]. Inhibiting PDK4 expression using microRNA (miR-5683) was also effective in promoting apoptosis [9]. Secondly, compounds that inhibited PDK activation have shown promise in cancer therapy. Dichloroacetate (DCA) is a well-known PDK inhibitor has already been in clinical evaluations [[10], [11], [12]]. Moreover, PDK inhibitors, such as 2,2-dichloroacetophenone (DAP), have demonstrated apoptosis-inducing effects in acute myeloid leukemia [13]. Based on the knowledge of existing compounds, researchers have synthesized more potent and selective PDK inhibitors, such as 1,2,4-amino-triazine derivatives and arsenic-containing compound Aa-Z2 [14,15]. Finally, combining PDK inhibitors with other drugs was found to enhance apoptosis. For instance, the combination of DAP with Erlotinib and Shikonin derivatives E5 with Gefitinib had been shown to enhance apoptosis in NSCLC [16,17].
Several novel DAP-analogs had been reported as potential PDK1 inhibitors [18]. These compounds exhibited improved potency and did not affect the PI3K pathway. Among these analogs, 64 showed significant potential in inducing apoptosis in NSCLC at a much lower concentration as compared with DAP. Moreover, 64 was found to generate significant level of intracellular reactive oxygen species (ROS), which may further facilitate apoptosis. However, the molecular mechanism by which 64 influences the apoptosis pathway remains unclear [18,19].
This study aims to investigate the mechanism of action of a PDK1 inhibitor, 64, in NSCLC cells models, namely NCI-H1975 and NCI-H1650 cells. In particular, the involvements of ROS, AMPK, and MAPK pathways in 64-induced apoptosis will be thoroughly explored. Interestingly, RNA sequencing analyses following 64 treatment revealed NF-κB activation, suggesting a survival/compensatory mechanism of the cancer cells upon PDK1 inhibition. This was corroborated with the combined use of 64 and a NF-κB inhibitor in NSCLC cells models, exhibiting synergistic anticancer effects. Our research identified the survival/compensatory mechanism via NF-κB activation from PDK1 inhibition in NSCLC cell models, which might be overcome by simultaneous inhibition of PDK1 and NF-κB.

Materials and method

Reagents
The primary antibodies: Cleaved PARP (Asp214) (D64E10) XP® Rabbit mAb (5625S), PARP Antibody (9542S), Phospho-AMPKα (Thr172) (40H9) Rabbit mAb (2535S), AMPKα Antibody (2532S), Phospho-p38 MAPK (Thr180/Tyr182) (D3F9) XP® Rabbit mAb Cell (4511), p38 MAPK (D13E1) XP® Rabbit mAb (8690), Cleaved Caspase-9 (Asp330) (D2D4) Rabbit mAb (7237P), Caspase-9 (C9) Mouse mAb (9508P)
Bcl-2 (50E3) Rabbit mAb (2870S), Bax (D2E11) Rabbit mAb (5023S), Cleaved Caspase-3 (Asp175) (9661S), Caspase-3 Antibody (9662S), β-Actin Antibody (4967) were purchase from Cell Signaling Technology (Danvers, MA, USA). Anti-Cytochrome C antibody (ab13575) was purchased from abcam (Cambridge, UK). Anti-NFκB p65 Antibody (sc-8008), Anti-JNK Antibody (D-2) (sc-7345), Anti-p-JNK Antibody (G-7) (sc-6254), Anti-p-c-Jun Antibody (KM-1) (sc-822), Anti-c-Jun Antibody (G-4) (sc-74543), Anti-α Tubulin Antibody (B-7) (sc-5286), Anti-GAPDH Antibody (G-9) (sc-365062), COX4 Antibody (F-8) (sc-376731), Anti-Lamin A/C Antibody (sc-7292) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The secondary antibodies: Anti-mouse IgG, HRP-linked Antibody (7076S), Anti-rabbit IgG, HRP-linked Antibody (7074S), were purchase from Cell Signaling Technology (Danvers, MA, USA). ROS scavenger N-Acetyl-L-cysteine (NAC), A9165 was purchased from Sigma-Aldrich (Darmstadt, Germany). P38 MAPK inhibitor SB203580 (S1076) and NF-κB inhibitor JSH-23 (S7351) were purchased from SelleckChem (Houston, TX, USA). The PDK1 inhibitor 64 was synthesized as previously described [18,20].

Cell Culture
The human NSCLC cell lines, NCI-H1975 and NCI-H1650, were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). They were cultured in RPMI 1640 media supplemented with 10% fetal bovine serum and maintained in a Thermo 160i CO2 incubator (Thermo Fisher Scientific, Waltham, MA, USA) at 37°C with 5% CO2 and a humidified atmosphere.

Western blotting analysis of protein expression
NCI-H1975 and NCI-H1650 cells (3 × 105) were seeded into 6-well plates and incubated overnight for attachment. Cells were treated for 24 h with either a medium containing 64 or a medium containing 0.1% DMSO. The cells were harvested and lysed in 200 μL of RIPA buffer containing 1% Protease Inhibitor Cocktail (Sigma-Aldrich, Darmstadt, Germany) for 30 min. After centrifugation at 12,000 rpm (4°C) for 20 min, the supernatant was transferred to a fresh tube. Protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) and adjusted to the same level. The proteins were denatured by boiling at 100°C for 5 min with SDS loading buffer. Subsequently, the proteins were separated by SDS-PAGE gel and transferred onto a PVDF membrane. The membrane was blocked with 5% bovine serum albumin (BSA) (Sangon, Shanghai, China) for 2 h, followed by overnight incubation with suitable primary antibodies. Afterward, the membrane was incubated with secondary antibodies (HRP-conjugated anti-rabbit or anti-mouse IgG (Cell Signaling Technology, Danvers, MA, USA) for 2 h. Images were captured using the ChemiDoc MP Imaging System (Bio-Rad TM, Hercules, CA, USA) after incubation with the UltraSignal ECL Substrate (4A Biotech, Beijing, China).

Preparation of mitochondria and cytoplasmic extracts
NCI-H1975 or NCI-H1650 cells (3.5 × 106) were seeded into a 10-cm dish. Then, the treated or untreated cells were harvested. The pellets were transferred into a 2 μL microcentrifuge tube. Subsequently, the mitochondria and cytosol fractions were extracted using the Mitochondria Isolation Kit for Cultured Cells (Thermo Fisher Scientific, Waltham, MA, USA). Briefly, 800 µL of Mitochondria Isolation Reagent A was added to the tube and incubated on ice for 2 min. Then, 10 µL of Mitochondria Isolation Reagent B was added to the tube and incubated for another 5 min. Next, 800 µL of Mitochondria Isolation Reagent C was added to the system. After gentle mixing, the tubes were centrifuged at 700 g for 10 min at 4°C. The supernatants were transferred to microcentrifuge tubes and centrifuged at 12,000 g for 15 min at 4°C. Following this step, the cytosol fractions were in the supernatant. For the subsequent analysis, the mitochondria pellets were lysed in 100 µL of 2% CHAPS in TBS. After a 2-min high-speed centrifugation, the soluble mitochondrial protein was in the supernatant.

ROS determination
The intracellular levels of reactive oxygen species (ROS) were determined using DCFH-DA staining. Briefly, 3 × 105 NCI-H1975 or H1650 cells were seeded in 6-well plates and incubated overnight for cell attachment. The cells were then treated with various compounds or 0.1% DMSO for 4 h. After the incubation, the cells were washed three times with serum-free culture medium and incubated with 1 mL of 10 μM DCFH-DA (Beyotime, Shanghai, China) diluted in serum-free medium at 37°C for 20 min. Finally, the cells were washed three times with serum-free medium and imaged using the Invitrogen EVOS FL Auto Cell Imaging System (ThermoFisher Scientific, Waltham, MA, USA) for cell imaging.

Cell Viability Assay
Cell viability was determined using the 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. Briefly, 3 × 103 NCI-H1975 and NCI-H1650 cells in 100 μL were seeded into 96-well plates and incubated overnight to allow cell attachment. The cells were then treated with either media containing various compounds or a medium containing 0.1% DMSO for 24 h. After the incubation, the cells were incubated with 0.5 mg/mL MTT (Sigma-Aldrich, Darmstadt, Germany) in 100 μL of whole medium for 4 h. The solvent was replaced with 100 μL of DMSO to dissolve the formazan crystals. The optical density (O.D.) values were measured using a SpectraMax M5 Microplate Reader (Molecular Devices, San Jose, CA, USA) at a wavelength of 570 nm.

Assessment of drug synergy
The assessment of drug synergy was conducted by evaluating the drug combination synergy, i.e. the combination index (CI), based on the Chou-Talalay analysis [21].D1 and D2 represent the combined doses needed for a specific reaction, whereas (Dx)1 and (Dx)2 represent the individual drug doses necessary for a comparable response. Synergism, additive effect, or antagonism can be indicated by CI values < 1, = 1, or > 1, respectively. The relationship between Dx and the fraction of cancer cells affected (Fa) can be expressed through the following equation:Dm, represents the concentration (e.g., IC50) at which 50% inhibition of cancer cells occurs. The slope coefficient (m) determines the shape of the dose-effect curve in the dose-effect relationship. Value of m equal 1 indicates a hyperbolic dose-effect curve, while values greater than 1 and less than 1 represent sigmoidal and flat sigmoidal curves, respectively. The software package CalcuSyn was employed to calculate CI values (Biosoft, UK).

Fluorescence imaging for MMP detection
The alteration of mitochondrial membrane potential (MMP) was measured using the EVOS FL Auto Cell Imaging System with JC-1 (5,5,6,6-tetrachloro-1,1,3,3-tetraethylbenzimidazolycarcocyanine iodide, Sigma-Aldrich, Darmstadt, Germany) staining. Briefly, 3 × 105 NCI-H1975 or NCI-H1650 cells were seeded in 6-well plates and incubated overnight for cell attachment. The cells were then treated with various compounds or 0.1% DMSO for a defined time. Next, the cells were incubated with 2 µM of JC-1 at 37°C for 15 min. Then, the cells were washed three times with whole medium and imaged using the Invitrogen EVOS FL Auto Cell Imaging System (ThermoFisher Scientific, Waltham, MA, USA). The fluorescence intensity was analyzed using Image J software (NIH, Bethesda, USA).

Flow cytometry assay for cell apoptosis detection
The proportion of cell groups undergoing apoptosis was detected using Annexin V-FITC/Propidium Iodide (PI) dual staining and analyzed by flow cytometry. NCI-H1975 or NCI-H1650 cell lines were seeded in 6-well plates and incubated overnight for cell attachment (3 × 105). The cells were then treated with various compounds or 0.1% DMSO for 24 h. After harvesting the cells with trypsin (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and washing them twice with PBS, they were suspended in 500 μL of Binding Buffer containing 5 μL of Annexin V-FITC and 10 μL of PI (FITC Annexin-V Apoptotic Detection Kit, KeyGene, Nanjing, China) and incubated for 15 min. The fluorescence intensity was measured using a CytoFLEX Flow Cytometer (Beckman Coulter, Indianapolis, IN, USA).

Bioinformatics analyses
NCI-H1975 and NCI-H1650 cells (3 × 105) were seeded into 6-well plates and incubated overnight for attachment. The cells were treated with either a whole medium that contained 5μM of 64 or a whole medium containing 0.1% DMSO for 24 h. Three biological replicates were performed. All cells were harvested using TRIzol™ Reagent (Invitrogen™, Waltham, MA, USA). The RNA sequencing was performed by Beijing U-gene (Beijing, China). The cDNA libraries were sequenced on an Illumina Hi-Seq platform (Illumina, San Diego, CA, USA). The quality of the raw data and expression quantification were analyzed using TopHat 2.1.1 and Cufflinks 2.2.1, respectively. In paired samples, gene sets that were upregulated or downregulated by at least two-fold were filtered. Gene Set Enrichment Analysis (GSEA) was conducted using the clusterProfiler R package. Heatmaps of gene set changes were generated using R. Related pathways were predicted using the KEGG (Kyoto Encyclopedia of Genes and Genomes) Pathway Mapper.

Real time PCR analysis of cDNA expression
NCI-H1975 and NCI-H1650 cells (1 × 105) were seeded into 12-well plates and incubated overnight for attachment. The cells were treated for 24 h with either a whole medium containing various doses of 64 or a whole medium containing 0.1% DMSO. Total RNA was extracted using TRIzol™ Reagent (Invitrogen™, Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. Next, 1 μg of RNA was reverse transcribed into cDNA using the iScript™ cDNA Synthesis Kit (Bio-Rad ™, Hercules, CA, USA). The reaction system contained 1 μL of ten-fold diluted cDNA templates, 10 μL of iTaq Universal SYBR Green Supermix (Bio-Rad ™, Hercules, CA, USA), 0.4 μL of forward and reverse primers (10 μM each), and 8.2 μL of water. Table 1 shows the primers used in this study. The amplification of samples was carried out using the Bio-Rad CFX96 Real-time PCR System (Bio-Rad ™, Hercules, CA, USA) using the two-step method. Each sample was analyzed in triplicate. The 18S gene was selected as the internal reference. The relative gene expression levels were quantified using the comparative Ct method.

Preparation of nuclear and cytoplasmic extracts
Treated or non-treated cells were harvested using trypsin (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and washed with PBS twice. Then, the cells were lysed with a hypotonic buffer (10 mM HEPES-KOH [pH 7.9], 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, 0.2 mM Pefabloc® SC, 0.5% NP-40) for 5 min on ice. Cell lysates were centrifuged at 16,000 g (4°C) for 10 s. The cytoplasmic extracts (supernatant) were harvested after this step. Subsequently, the pellets were washed once with hypotonic buffer and dissolved in a high-salt buffer (50 mM Tris [pH 7.4], 450 mM NaCl, 1% NP-40, 1 mM PMSF, 0.2 mM Na3VO4, 5 mM β-glycerophosphate, 20% glycerol, 2 mM DTT), followed by incubation on ice for 10 min. Next, the cell lysates were centrifuged at 16,000 g for 15 min at 4°C. The nuclear extracts were presented in the supernatants and collected after this step.

Statistical analysis
All data were presented as mean ± SD. Differences among samples were assessed using SPSS 19 (ANOVA analysis), and a p-value of less than 0.05 was considered significant. * represents P< 0.05, ** represents P<0.01, and ***represents P<0.001.

Results

64 affected several key proteins in the apoptotic pathway
To further explore the effects of 64 on cellular apoptosis, we incubated the NCI-H1975 and NCI-H1650 cells with different concentrations of 64 for 24 h and evaluated the expression of key proteins in the caspase-dependent pathway. Our findings unveiled a compelling dose-dependent relationship between 64 and the ratio of cleaved caspase-3 to caspase-3, as well as cleaved PARP to PARP (Figs. 1a, b), indicative of 64-initiated apoptosis. Next, we conducted mitochondria and cytoplasm extraction experiments. Remarkably, our results demonstrated a significant increase in the translocation of cytochrome c (CytC) from the mitochondria to the cytoplasm upon treatment with 64 (Figs. 1c, d). Confocal images further bolstered our findings by visually depicting the localization of CytC in both the mitochondria and cytoplasm (Supplementary Material, Fig. S1). This suggests that 64 triggers the release of CytC, which is known to be involved in the activation of caspase-3 and subsequent apoptosis.

64 inhibited growth of NSCLC cell and induced apoptosis through the generation of ROS
Previous research demonstrated that 64 inhibited cell viability, depolarized mitochondrial membrane potential (MMP), and induced cell apoptosis, as well as increased the ROS level in NSCLC cell model [18]. However, detailed study of the roles of ROS in 64-induced apoptotic cell death was lacking. N-acetyl-l-cysteine (NAC) is a ROS scavenger that can inhibit the generation of intracellular ROS. Here, we used 64 (5μM) and NAC (5mM), either alone, or in combination, to treat NCI-H1975 and NCI-H1650 cells. As illustrated in Fig. 2, treatment with 64 stimulated intracellular ROS generation, inhibited cell viability, decreased MMP, and induced cell apoptosis, while the 64-NAC combo significantly attenuated these effects. Our results suggested that 64 induced NSCLC cell growth inhibition and cell apoptosis through ROS.

64 affected the activation of AMPK and MAPK pathway through ROS
The AMPK and MAPK pathways were reported to be activated through the phosphorylation of the corresponding proteins [22,23]. To measure how 64 activated these pathways, NCI-H1975 and NCI-H1650 cells were treated with varying concentrations of 64, either alone or in combination with NAC (5mM) for 24 h. The cells were then lysed and the protein expression levels were detected using a western blotting assay. The findings revealed a clear and consistent dose-dependent increase in the phosphorylation levels of AMPK protein induced by 64 (Fig. 3a). Likewise, 64 demonstrated a dose-dependent elevation in the phosphorylation of P38-MAPK, JNK, and c-Jun (Fig. 3c). These results were further supported by the accompanying histogram analysis, which clearly demonstrated a dose-dependent increase in the ratios of phosphorylated forms (P-AMPK, P-P38MAPK, P-JNK, and P-c-Jun) to their respective total proteins (AMPK, P38-MAPK, JNK, and c-Jun). Notably, when 64 was combined with NAC, a significant attenuation of these activation trends was observed (Figs. 3b, d). These suggested that 64 activated the AMPK and MAPK pathway through ROS.

64 affected NSCLC apoptotic pathway through ROS
To evaluate whether 64 affecting key proteins in the apoptotic pathway through ROS, the expression levels of apoptotic-related proteins was assessed in NSCLC cells following the combined treatment of 64 and NAC (5mM). The combo was found to attenuate the release of CytC from mitochondria into the cytoplasm triggered by 64 (Figs. 4a, b). As shown in Fig. 4c, 64 increases the expression of the Bcl-2-associated X protein (Bax) and decreases the expression of B-cell lymphoma 2 (BCL-2). Moreover, 64 alone can increase the expression of cleaved-caspase9, cleaved-caspase3, and cleaved-PARP (Fig.4 c). Additionally, the histogram in Fig. 4d, demonstrates the increasing ratios of cleaved forms (c-Caspase9, c-Caspase3, and c-PARP) to their respective total proteins (Caspase9, Caspase3, and PARP), further supporting this observation, while the 64-NAC combo would attenuate these trends (Figs. 4 c, d). Taken together, our data suggested that 64 induced cell mitochondria-dependent apoptosis through ROS in NSCLC cells.

64 activated NF-κB pathway through P38-MAPK
After demonstrating that 64 induces cell apoptosis through ROS, we performed RNA sequencing studies on NCI-H1975 and NCI-H1650 cells, either treated with or without 64 (5µM) for 24 h. Bioinformatics analyses revealed that the NF-κB pathway was the most significantly affected, followed by the MAPK pathway (Fig. 5a). This further substantiated by the heatmap depicting the 2-fold regulated genes in the NF-κB and MAPK pathways (Fig. 5b). To corroborate the RNA-seq results, we conducted real-time PCR to confirm the expression of key genes involved in the NF-κB pathway: (a) Mitogen-activated protein kinase 14 (MAP3K14) (NF-kappa-B-inducing kinase (NIK)) (NIK), (b) Baculoviral IAP Repeat Containing 3 (BRIC3) ((Cellular Inhibitors of Apoptosis 2 (c-IAP2)), (c) NF-κB2, and (d) v-rel reticuloendotheliosis viral oncogene homolog B (RELB) (Fig. 5c). Our results clearly demonstrated that 64 induced a dose-dependent increase in the expression of these genes, suggesting their potential to influence the transcription activity of NF-κB. Furthermore, we investigated the translocation of the NF-κB member p65 from the cytoplasm to the nucleus. Notably, 64 exhibited a dose-dependent decrease in cytoplasmic p65 protein levels coupled with an increase in nuclear p65 levels (Fig. 5d). This provides compelling evidence supporting the translocation of p65 induced by 64. Interestingly, we found that P38 MAPK, acting as an upstream effector of NF-κB, played a crucial role in the activation of the NF-κB pathway by 64. When 64 was combined with the P38 MAPK inhibitor SB203580 (20μM), the translocation of p65 induced by 64 was significantly attenuated (Fig. 5e). These results strongly suggested that 64 activated the NF-κB pathway through the involvement of P38 MAPK. In summary, our findings suggested that 64 triggered cell apoptosis through the generation of ROS and activated the NF-κB pathway via the P38 MAPK pathway.

The activation of NF-κB rescued of 64 induced inhibition of NSCLC cell growth and apoptosis
To assess how the activation of NF-κB affects NSCLC cell apoptosis, a NF-κB inhibitor, JSH-23, was utilized, either alone or in combination with 64 to treat the NSCLC cells. As shown in Fig. 6a, JSH-23 inhibits the activation of NF-κB. Moreover, the combination of JSH-23 (10µM) and 64 potentiated the inhibition of cell viability caused by 64 treatment alone (Fig. 6b). Furthermore, JSH-23 displayed anti-cancer synergy in combination with 64. The combination of 64 and JSH-23 significantly potentiated the inhibition of cell viability compared with treated alone (Fig. 6c). The calculated combination index (CI) values with all fraction affected points were less than 1 [24] (Fig. 6d). Fig. 6e shows that the 64-JSH-23 combo potentiates the decrease in Δψm (lower mitochondrial membrane potential) as compared with that induced by 64 alone. Finally, we observed that the 64-JSH-23 combo significantly potentiated 64-induced cell apoptosis (see Fig. 6f).

The activation of NF-κB pathway affected the alteration of key proteins caused by 64 in the apoptotic pathway of NSCLC cells
Attention was then direct to assess the effects of NF-κB on the apoptotic pathway upon the combined use of 64 (5μM) and JSH-23 (10μM) on NSCLC cells. The results demonstrated that the co-incubation of these two cell lines with 64 and JSH-23 led to a more pronounced decrease in CytC protein expression within the mitochondria, while simultaneously increasing its expression in the cytoplasm compared to incubations with 64 alone (Fig. 7a). Moreover, the histogram analysis revealed an elevated ratio of cytoplasmic CytC to mitochondrial CytC in the combination treatment of 64 and JSH-23, as compared with that of using 64 alone (Fig. 7b). These findings further supported the notion that inhibiting the activation of NF-κB enhanced the release of CytC from the mitochondria into the cytoplasm. Regarding caspase-dependent cell apoptosis, the combination of 64 and JSH-23 intensified the upregulation of Bax protein and downregulation of BCL-2 protein as compared with that induced by 64 alone (Fig. 7c, d). Additionally, the combination treatment enhanced the 64-induced increase in cleaved-caspase-3 and cleaved-PARP (Fig. 7c). The histogram analysis of the ratio between the cleaved form (c-Caspase3 and c-PARP) and the total form (Caspase3 and PARP) (Fig. 7d) further supported this observation. These findings suggested that the activation of NF-κB inhibited the release of CytC and caspase-dependent cell apoptosis induced by 64, suggesting that NF-κB activation can be rationalized as a rescue mechanism for 64-induced cell apoptosis in the NSCLC cell models. It has been reported that sustained JNK activation plays a crucial role in inducing cell apoptosis when the NF-κB pathway is deficient [25]. We next examined the activation of the JNK pathway under NF-κB inhibition. The results indicated that the combination of 64 and JSH-23 increased the phosphorylation levels of JNK and c-Jun as compared with that of using 64 alone (Fig. 7e). Our analysis demonstrated an increased ratio of phosphorylated forms (p-JNK and p-c-Jun) to total proteins (JNK and c-Jun) (Fig. 7f). These findings suggested that inhibiting NF-κB activation enhanced 64-induced cell apoptosis through JNK activation.

Discussion
It has been reported that 64 specifically targets the PDK1 enzyme located in the mitochondria and induced apoptosis in NSCLC cell models [18,20]. However the mechanism of action of 64 had not yet been fully elucidated. In the present study, first examined the impact of 64 on the mitochondrial-dependent apoptotic pathway (intrinsic pathway). It has been suggested that the release of CytC was the initiation step, and the activation of caspase 3 was the key execution step of intrinsic apoptosis [26,27]. As illustrated in Fig.1, 64 induced the release of CytC and activated caspase 3, confirming the mitochondrial-dependent apoptotic pathway.
We then attempted to investigate the upstream events of the apoptotic pathway induced by 64. As ROS primarily accumulate in mitochondria [28], 64 was shown to inhibit cell viability by generating intracellular ROS, as depicted in Fig. 2b. Consequently, our study aimed to examine the role of ROS in 64-induced apoptosis in NSCLC cells. Most cancers have a higher MMP (Δψm) than normal cells, which would be helpful for apoptosis resistance. Thus, mitochondrial depolarization could lead to cell apoptosis [10,29]. Our results have shown that 64 reduced Δψm and induced cell apoptosis through the accumulation of intracellular ROS in NSCLC cell models (Fig. 2c and 2d). Moreover upon 64 treatment we have observed CytC release, cleavage of caspase 9, caspase 3, and PARP, [[30], [31], [32]], which were also driven by ROS.
PDK1 inhibitors, such as 64 and other recently developed analogues, have been shown to affect NSCLC cell metabolism [33]. It has been reported that the activation of AMPK is crucial in cancer cell metabolism [34]. Moreover, previous research has shown that the activation of the MAPK pathway can induce cell apoptosis in NSCLC cells [23]. Therefore, inhibition of PDK1 by 64, along with its potential effects on AMPK, may contribute to cell apoptosis in NSCLC cells through the modulation of the MAPK pathway. There are three well-known MAPK pathways: extracellular-signal-regulated kinase (ERK), JNK, and p38 MAPK [35]. 64 has been reported not to significantly affect the activation of ERK [18]. Thus, our attention was directed to the JNK and P38 MAPK pathways. Our findings suggested that 64 exhibited significant activation of the AMPK, JNK and P38-MAPK pathways, as confirmed by the data presented in Fig. 3a and 3c. Moreover, it effectively induced the activation of the apoptotic pathway, as illustrated in Fig. 1, Fig. 4. Notably, previous studies have highlighted the role of ROS accumulation in triggering cell apoptosis through the activation of the AMPK and MAPK pathways [22,23]. In line with these insights, our combination experiment further strengthened the notion that 64 robustly activated the AMPK, JNK and P38 MAPK by promoting the generation of ROS (Fig. 3b, d). It is worth emphasizing that 64 actively stimulated ROS production, thereby instigating the activation of these critical signaling pathways. We have demonstrated that 64 effectively activated the AMPK and MAPK pathways by generating ROS, thereby initiating the apoptotic pathway. However, the interplays among JNK and P38 MAPK pathways in 64 induced cell apoptosis deserve further investigation.
The P38 MAPK pathway acts as an upstream regulator of NF-κB, and inhibiting P38 MAPK activation can reverse NF-κB activation in NSCLC [36]. In our study, 64 was found to induce the production of ROS in NSCLC cells, which triggered the activation of the P38 MAPK pathway (Fig. 3d), leading to the activation of the NF-κB pathway (Fig. 5d). The NF-κB pathway has been recognized as a pro-survival mechanism in cancer cells, most often upon external assaults [37]. It promoted cell proliferation, inhibited apoptosis, facilitated cell migration and invasion, and contributed to angiogenesis and metastasis [38,39]. We hypothesized that the activation of the NF-κB pathway served as a rescue mechanism to counteract apoptosis induced by 64 in NSCLC cells. As far as we aware this is the first report of possible survival mechanism of NSCLC therapy through PDK1 inhibition. To validate our hypothesis, we conducted a series of experiments using 64 in combination with JSH-23 (Fig. 6, Fig. 7). The results obtained from these experiments demonstrated that the combination of 64 and JSH-23 significantly potentiated the apoptotic effects of 64 on the cells. Moreover, our study revealed that the activation of JNK played a crucial role in promoting cell apoptosis. Notably, sustained JNK activation becomes crucial for inducing cell apoptosis when the NF-κB is deficient [25]. Interestingly, we found that inhibiting the activation of the NF-κB pathway enhanced 64-induced cell apoptosis through the potentiated activation of JNK. This finding suggests that the activation of the NF-κB pathway may inhibit the activation of JNK, which in turn leads to the rescue of cell apoptosis in NSCLC cells upon PDK1 inhibition.
It has been suggested that JNK and P38-MAPK pathways played intricate roles in tumorigenesis [40]. Although we have shown combined use of 64 and P38 MAPK inhibitor, SB203580, reduced the translocation of p65 induced by 64 (Fig. 5e), which could modulate 64-induced NF-κB activation. However, it remains to be seen whether such combination could lead to anticancer synergy in NSCLC cell models. Another clinical P38 MAPK inhibitor, losmapimod (GW856553X), could be a better candidate for the drug combination, which deserves further investigations.
To sum up, PDK1 inhibition via
64 in NSCLC cells led to significant generation/accumulation of intracellular ROS, which then activated the AMPK and JNK pathways, inducing cancer cell apoptosis. Moreover, the synergistic use of the PDK1 inhibitor 64 and the NF-κB inhibitor JSH-23 can enhance cell apoptosis in NSCLC cell models via the potentiated JNK pathway. Overall, the combination of 64 and the NF-κB inhibitor JSH-23 yields superior anti-cancer effects in NSCLC cell models. This study elucidated the mechanism underlying the combination of 64 and the NF-κB inhibitor JSH-23, which could contribute to the development of a novel therapeutic strategy for NSCLC.

Conclusion
We found that PDK1 inhibition via
64 stimulated the generation of intracellular ROS in NSCLC cells. This increase in ROS levels led to the inhibition of cell growth, reduction of MMP, and induction of cell apoptosis. Moreover, 64 activated the AMPK and MAPK pathways through intracellular ROS. The activation of the mitochondrial-dependent apoptotic pathway was confirmed as a result of ROS accumulation. Our studies revealed, for the first time, that 64 activated NF-κB through the P38 MAPK pathway, which was rationalized as a survival mechanism to counteract the cancer cell apoptosis induced by 64. This finding contributes to a better understanding of the intricate relationships between metabolic modulation, cancer cell apoptosis and the associated pro-survival signals which could be helpful for the development of therapeutic strategy for NSCLC through PDK1 inhibition.

Data availability
All data are available from the corresponding author upon reasonable request.

CRediT authorship contribution statement
Quan Liu: Investigation, Conceptualization, Visualization, Methodology, Formal analysis, Data curation, Writing – original draft. Maoxin Ran: Investigation, Methodology. Wenying Shan: Methodology, Investigation. Shao-Lin Zhang: Formal analysis. Kin Yip Tam: Supervision, Investigation, Writing – review & editing, Conceptualization, Project administration, Funding acquisition.

Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
Kin Yip Tam reports financial support was provided by Science and Technology Development Fund, Macau SAR. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.