PLK1-Mediated PDHA1 Phosphorylation Drives Metabolic Reprogramming in Lung Cancer

Abstract

Although the involvement of polo-like kinase 1 (PLK1) in metabolic reprogramming from oxidative phosphorylation (OXPHOS) to glycolysis has been previously described, the underlying molecular mechanism remains unclear. Pyruvate dehydrogenase (PDH) catalyzes the conversion of pyruvate into acetyl-CoA, the starting material for the tricarboxylic acid (TCA) cycle. In a companion study by Zhang et al., we demonstrated that PLK1 phosphorylation of PDHA1 at threonine 57 (PDHA1-T57) drives its protein degradation via mitophagy activation. Using a stable-isotope resolved metabolomics (SIRM) approach, we now show that PLK1 phosphorylation of PDHA1-T57 results in metabolic reprogramming from OXPHOS to glycolysis. Notably, cells mimicking PDHA1-T57 phosphorylation rely more on the aspartate-malate shuttle than on glucose-derived pyruvate to sustain the TCA cycle. This metabolic shift was also observed in mouse embryonic fibroblasts (MEFs) and transgenic mice conditionally expressing the PDHA1-T57D variant, highlighting the role of PLK1 in metabolic reprogramming in vivo. It is well-established that pyruvate dehydrogenase kinase (PDK)-mediated phosphorylation of PDH leads to its inactivation and that dichloroacetic acid (DCA), a PDK inhibitor, has been investigated in preclinical and early clinical studies as a potential therapeutic agent for lung cancer. We demonstrated that DCA combined with Onvansertib, a PLK1 inhibitor, synergistically inhibits lung tumor growth by enhancing mitochondrial ROS, inhibiting glycolysis, and inducing apoptosis. This study aims to elucidate how PLK1-associated activity drives the metabolic reprogramming from OXPHOS to glycolysis during cellular transformation, thereby contributing to lung carcinogenesis. Our results provide support for a clinical trial to evaluate the efficacy of Onvansertib plus DCA in treating lung cancer.

Introduction

Lung cancer includes small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), which is further classified into lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). The genetics of lung cancer is highly complex. For example, SCLC is characterized by TP53 mutations in approximately 90% of cases. In LUSC, CDKN2A is frequently inactivated, while 20% of cases harbor PIK3CA mutations that activate the PI3K/AKT/mTOR pathway. In LUAD, approximately 30% of cases carry KRAS mutations, and 15% carry EGFR mutations (1). As such, targeted therapies against KRAS or EGFR in lung cancer have achieved limited success. This highlights the urgent need to identify new pathways whose inhibition can induce lung cancer cell death, regardless of the tumor’s genetic background.

One such universally altered hallmark in lung cancer is the Warburg effect, a metabolic shift from oxidative phosphorylation (OXPHOS) to aerobic glycolysis (2, 3). Targeting metabolic pathways, therefore, presents a promising therapeutic strategy across diverse lung cancer subtypes. Pyruvate dehydrogenase (PDH) serves as a critical regulator of this metabolic switch. As a gatekeeping mitochondrial enzyme, PDH determines whether glucose-derived pyruvate enters the tricarboxylic acid (TCA) cycle for mitochondrial respiration or is diverted to lactate production via glycolysis.

PDH catalyzes the conversion of pyruvate to acetyl-CoA, which fuels the TCA cycle and generates NADH and FADH₂—electron donors for the respiratory chain. Electron transport through this chain drives ATP synthesis via the mitochondrial membrane potential (ΔΨm), but also generates reactive oxygen species (ROS). PDH activity is negatively regulated by pyruvate dehydrogenase kinases (PDKs) through phosphorylation (4). Additionally, under hypoxic conditions, the PDH-E1β subunit is transcriptionally downregulated in a HIF1-dependent manner (5). Notably, dichloroacetic acid (DCA), a PDK inhibitor, has been explored in preclinical and early clinical studies as a potential therapeutic agent for lung cancer (6).

Polo-like kinase 1 (PLK1), a molecule involved in critical cell cycle events, also plays a well-documented role in lung cancer (7). The expression level of Plk1 is significantly higher in LUAD (about 9.7-fold) and LUSC (about 20.8-fold) compared to adjacent normal tissue (7, 8). Patients with higher PLK1 expression levels in lung cancer have significantly worse overall survival prognosis than those with lower expression levels (7, 8). PLK1 elevation significantly accelerates oncogenic Kras expression-induced LUAD in a genetically engineered mouse (GEM) model (9). Furthermore, various lung cancer cell lines with diverse genetic backgrounds all exhibit elevated levels of PLK1. Targeting PLK1 leads to cell death and inhibition of tumor growth in lung cancer (10). Of note, Onvansertib, a PLK1 inhibitor, is currently being tested in clinical trials for the treatment of both NSCLC and SCLC (NCT05450965).

Despite these advancements, whether PLK1 is involved in the metabolic reprogramming of lung cancer is not known. In the companion study by Zhang et al, we demonstrated that PLK1 is upregulated in Cr(VI)-transformed BEAS-2B cells, promoting mitochondrial dysfunction and mitophagy, which in turn enhances cell proliferation. Mechanistically, PLK1 directly phosphorylates PDHA1 at threonine 57 (T57), leading to destabilization of the PDH complex, impaired oxidative phosphorylation (OXPHOS), and mitochondrial dysfunction. This initiates a mitophagy-dependent feedback loop that further accelerates PDHA1 degradation and exacerbates mitochondrial damage(11).

In the current study, we expand on this mechanism by performing stable isotope-resolved metabolomics (SIRM) in lung cells expressing phospho-deficient (T57A) or phospho-mimetic (T57D) PDHA1 mutants. We also generated a novel knock-in mouse model expressing PDHA1-T57D to evaluate the in vivo metabolic consequences. Collectively, our findings establish a critical role of PLK1-mediated PDHA1 phosphorylation in promoting the Warburg effect and metabolic reprogramming in lung cancer.

Importantly, unlike PDK-mediated phosphorylation that inhibits PDH activity, PLK1-dependent phosphorylation leads to PDHA1 protein degradation. Based on these findings, we hypothesized that co-targeting PLK1 (using Onvansertib) and PDKs (using DCA) could restore mitochondrial function and suppress tumor growth, independent of tumor genotype.

Our working model proposes that during malignant transformation, elevated PLK1 promotes PDHA1 degradation, leading to a shift from OXPHOS to glycolysis and supporting tumorigenesis. This study aims to elucidate the metabolic impact of PLK1-PDHA1 signaling and assess the therapeutic efficacy of dual inhibition with Onvansertib and DCA. Our results provide preclinical rationale for a clinical trial testing this combinational approach in lung cancer patients.

Results

PLK1-mediated PDHA1-T57 phosphorylation affects glucose metabolism and the TCA cycle

Glycolysis-derived pyruvate is oxidized to acetyl-CoA by the PDH complex and utilized as one of the major carbon sources to maintain the TCA cycle. Since PLK1 phosphorylation of PDHA1 at T57 results in its degradation (as shown in the companion study by Zhang et al.(11)), we aimed to determine the importance of PLK1-mediated phosphorylation of PDHA1 in glucose metabolism and the TCA cycle. We used the human bronchial epithelial cell line BEAS-2B, a system that has been used extensively to study heavy metal-induced lung cancer (12). It has been documented that chronic exposure of Cr(VI) at concentrations of 0.1250αM (30ppb) to 0.5αM (130ppb) for 3 months results in transformation of BEAS-2B cells (13). The resulting cells were named as Cr(VI)-transformed BEAS-2B (CrT) cells. Accordingly, we carried out a SIRM analysis of CrT cells expressing PDHA1 variants (T57A and T57D).

The cells were cultured with uniformly labeled ¹³C-glucose ([U-¹³C]-glucose), followed by sample processing and metabolite extraction for NMR and IC-FTMS (Figure 1). ¹³C-enriched metabolites from glycolysis, the TCA cycle, and amino acid synthesis were identified, and their total amounts were determined. While PDHA1-T57 phosphorylation did not affect glucose uptake (Figure 2A), the level of ¹³C-labeled pyruvate was significantly increased in the culture media of cells expressing the PDHA1-T57D (Figure 2B). Detailed analysis of metabolites involved in glycolysis showed that cells with the PDHA1-T57D variant failed to utilize ¹³C-labeled glucose to fulfill the carbon influx for the TCA cycle, as indicated by the accumulation of ¹³C-labeled pyruvate in both cells and medium during the production of acetyl-CoA by PDH. In particular, cells expressing the PDHA1-T57D variant exhibited low levels of G6P, F6P, PEP, and lactate derived from ¹³C-labeled glucose but significantly higher levels of pyruvate compared to cells expressing the PDHA1-T57A variant (Figures 2C, 2D, 2E). Taking into account that the level of ¹³C-labeled pyruvate was increased both in the culture media (Figure 2B) and cell extracts (Figure 2C, last panel) of cells expressing the PDHA1-T57D variant, this confirms the blockage of pyruvate metabolism.

Figure 1.

Schematic of stable isotope-resolved metabolomics (SIRM) analysis of CrT cells expressing different PDHA1 variants (T57A or T57D). The cells were cultured with uniformly labeled ¹³C-glucose ([U-¹³C]-glucose), harvested for sample processing, and subjected to metabolite extraction. The analysis was conducted using nuclear maFnetic resonance (NMR) and ion chromatography-Fourier transform mass spectrometry (IC-FTMS). Black circle represents ¹²C, while red circle and blue circles indicate ¹³C contributions from the PDH and PC-derived TCA cycle, respectively. G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; PEP: phosphoenolpyruvic acid; PC, Pyruvate carboxylase.

Figure 2.

PLK1-mediated PDHA1-T57 phosphorylation impacts glucose metabolism. (A) Time courses of ¹³C enrichment in glucose in the culture media. Representative NMR analysis of ¹³C-enriched glucose in the media from CrT cells expressing the PDHA1-T57A or -T57D variant. Media samples were collected at 3, 6, 12, and 24 hours and processed for NMR analysis. (B) Representative NMR analysis of ¹³C-enriched pyruvate in the media. (C) Total amounts of isotopologues in glycolytic metabolites. The x-axis indicates the number of ¹³C atoms in each compound. Total amounts of isotopologues are presented as means ± standard deviations (n = 3). (D) Representative IC-MS analysis of the fractional enrichment of ¹³C-labeled glycolytic metabolites in CrT cells expressing the PDHA1-T57A or PDHA1-T57D variant. The x-axis indicates the number of ¹³C atoms in each compound. Fractional enrichment of isotopologues is presented as means ± standard deviations (n = 3). (E) Total amounts of isotopologues in lactate. *P < 0.05; **P < 0.01; ***P < 0.001 by unpaired t-test.

To assess whether phosphorylation of PDHA1 at T57 affects downstream metabolism of pyruvate, we measured the amounts of metabolites in the TCA cycle. The first turn of the TCA cycle produces metabolites with m+2, and this was not affected by the expression of the T57D variant. However, subsequent turns, which produce metabolites with m+3, m+4, and m+5, showed decreases upon expression of the T57D variant. Most of the significant ¹³C-enriched isotopologues of citrate, cis-aconitate, and isocitrate (Figures 3A, 3B) were significantly decreased in cells with PDHA1-T57D compared to the phospho-null mimics, indicating that phosphorylation at T57 of PDHA1 blocks carbon influx from ¹³C-labeled pyruvate. Notably, the contribution from non-glucose substrates, measured through m+0 levels, was considerably higher in cells expressing PDHA1-T57D variant compared to the phospho-null mimics. This demonstrates that PLK1-mediated phosphorylation of PDHA1 at T57 prevents cells from utilizing carbon derived from ¹³C-labeled glucose. In contrast, cells expressing the phospho-null mimics of PDHA1 efficiently consumed ¹³C-enriched pyruvate derived from [U-¹³C]-glucose, leading to ¹³C-labeled carbon entering the TCA cycle (Figures 3A, 3B).

Figure 3.

Effects of PLK1-mediated PDHA1-T57 phosphorylation on the TCA cycle. (A) Total amounts of isotopologues in citrate, cis-aconitate, and isocitrate in CrT cells expressing the PDHA1-T57A or PDHA1-T57D variant. (B) Fractional enrichment of ¹³C-labeled citrate, cis-aconitate, and isocitrate. (C) Total amounts of the isotopologues in α-ketoglutarate (α-KG), succinate, fumarate, and malate in CrT cells expressing the PDHA1-T57A or PDHA1-T57D variant. (D) Fractional enrichment of ¹³C-labeled α-KG, succinate, fumarate, and malate. (E, F) Total amounts and fractional enrichment of ¹³C-labeled glutamate and aspartate. The x-axis indicates the number of ¹³C atoms in each compound. Total amounts of isotopologues are presented as means ± standard deviations (n = 3), *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 by unpaired t-test.

Although most of the ¹³C-labeled carbon in TCA precursor metabolites, such as citrate, cis-aconitate, and isocitrate (Figures 3A, 3B), originated from [U-¹³C]-glucose in cells expressing PDHA-T57A variant, we still observed a shift. Nearly half of α-KG, succinate, fumarate, and malate (Figures 3C, 3D) came from non-¹³C labeled sources, indicating a dominant contribution from carbon sources other than glucose. Similar effects were observed in cells expressing the PDHA1-T57D variant, where more than 40% of TCA precursor metabolites (citrate, cis-aconitate, and isocitrate) and 60% of α-KG, succinate, fumarate, and malate relied on unlabeled carbon sources. In addition to glucose, glutamate and aspartate can contribute to carbon to the TCA cycle through transamination. To determine whether glutamate and aspartate account for the unlabeled TCA cycle metabolites, we measured the levels of aspartate and glutamate in cells expressing the PDHA1-T57A or -T57D variants. We found that both cell lines had higher levels of unlabeled aspartate and glutamate than their labeled isotopologues (Figures 3E, 3F). More surprisingly, cells expressing the PDHA1-T57D variant had higher levels of both aspartate and glutamate compared to cells with the phospho-null mimics (Figures 3E, 3F). This suggests that PLK1-mediated phosphorylation of PDHA1 may enhance the metabolic reliance on glutamate and aspartate to compensate for the reduced carbon input from pyruvate due to the inhibition of PDH function.

In addition to being reduced by NADH to lactate by LDH, pyruvate mainly relies on PDH and PC to enter the TCA cycle and provide carbon for subsequent reaction. We therefore assessed the relative activity of PDH and PC in cells expressing PDHA1-T57A or -T57D variants. The citrate m+2 / pyruvate m+3 ratio serves as a surrogate for PDH activity, while the citrate m+3 / pyruvate m+3 ratio serves as a surrogate for PC activity. Our findings support the hypothesis that phosphorylation of PDHA1 at T57 leads to suppressed PDH activity (Figure 4A). Surprisingly, PLK1-mediated phosphorylation of PDHA1 reduced the catalytic capacity of PC by half, even though carbon entry into the TCA via PC was limited in this case (Figure 4A).

Figure 4.

PDHA1-T57 phosphorylation shifts the carbon source from glucose to glutamate and aspartate. (A) Citrate m+2 / pyruvate m+3 ratio and citrate m+3 / pyruvate m+3 ratio are shown as surrogates for the activity of PDH and PC, respectively, in cells expressing the PDHA1-T57A or PDHA1-T57D variant. (B) Total fractional enrichments of ¹³C-labeled isotopologues. (C) Heat map showing the total amounts of ¹³C-enriched metabolites associated with glycolysis, the TCA cycle, purine synthesis, pyrimidine synthesis, and amino acid metabolism.

To better visualize the carbon influx affected by PLK1-mediated PDHA1 phosphorylation, we calculated several major ¹³C-enriched metabolic intermediates in glycolysis and the TCA cycle according to the carbon metabolic pathways in the cell. We then displayed the major carbon flows based on their fraction labeled by ¹³C in cells expressing the PDHA1-T57A or -T57D variant (Figure 4B). Combining the actual content of these metabolic intermediates, we found that more than 95% of G6P, F6P, and PEP (Figure 2D) are derived from [U-¹³C]-glucose. However, these metabolites are metabolized and consumed in cells expressing PDHA1-T57D variant at a faster rate than in cells expressing the PDHA1-T57A variant (Figure 2C). For pyruvate, although ¹³C-enriched pyruvate accounted for 80% of the total content in both cell lines expressing PDHA1-T57A or -T57D variants (Figure 2D), the actual content of both labeled and unlabeled pyruvate in cells expressing the PDHA1-T57D variant was three times higher than in cells expressing its phospho-null mimics (Figures 2B, 2C, 4B). Meanwhile, several TCA intermediates in cells expressing the PDHA1-T57A variant, such as citrate, cis-aconitate, and isocitrate, are derived from [U-¹³C]-glucose. However, the corresponding fraction of ¹³C-labeling in cells expressing the phospho-mimics is only 60% (Figures 3B, 4B). For metabolites downstream of α-KG, such as α-KG, succinate, fumarate, and malate, only about 40% of them are derived from [U-¹³C]-glucose in cells expressing PDHA1-T57D (Figures 3D, 4B). This suggests that in cells expressing PDHA1 phospho-mimics, the carbon carriers derived from U¹³C-glucose cannot efficiently enter the TCA cycle through the pyruvate metabolic pathway. In contrast, cells expressing phospho-null mimics can use the [U-¹³C]-glucose-derived pyruvate to provide sufficient carbon for the TCA cycle. This further indicates that PLK1 affects carbon influx by phosphorylating PDHA1 at T57, ultimately impacting oxidative phosphorylation.

To gain a more comprehensive understanding of the potential impact of PLK1 phosphorylation of PDHA1 on cellular metabolism, we integrated the actual content of ¹³C-enriched metabolites in cells expressing the PDHA1-T57A or PDHA1-T57D variants related to glycolysis, the TCA cycle, purine synthesis, pyrimidine synthesis, and amino acid synthesis (Figure 4C). As expected, PLK1-mediated phosphorylation of PDHA1 resulted in a reversal of pyruvate metabolism, while upregulation of metabolites related to the aspartate-malate shuttle suggested a shift in the carbon source from pyruvate to amino acids such as glutamate. Surprisingly, we also found that certain purine and pyrimidine metabolites were affected by PLK1-mediated phosphorylation of PDHA1, such as GMP, GTP, UMP, and CMP, further demonstrating the close relationship between PLK1 and cellular metabolism. Taken together, considering that PLK1-meidated PDHA1 phosphorylation triggers PDHA1 protein degradation (see companion manuscript by Zhang et al. for details(11)), cells expressing the PDHA1-T57D variant can no longer efficiently consume pyruvate and utilize ¹³C-enriched pyruvate as a carbon source to drive the TCA cycle. This results in an accumulation of pyruvate and a shift in dependence from pyruvate to backup carbon sources such as glutamate and aspartate. Hence, PLK1 is tightly involved in glucose metabolism and negatively regulates the TCA cycle through phosphorylation of PDHA1 at T57.

Introduction of the PDHA1-T57D mutation in endogenous protein results in metabolic reprogramming in vivo

To better understand the potential effects of PLK1-mediated PDHA1 T57 phosphorylation on energy metabolism, we generated a PDHA1-T57D conditional knock-in (KI) transgenic mouse line. Mouse embryonic fibroblasts (MEFs) were isolated from embryos of pregnant transgenic mice conditionally expressing the PDHA1-T57D variant and Cre-recombinase. To assess whether PDHA1 T57 phosphorylation leads to metabolic reprogramming in vivo, primary MEFs were transfected with Lenti-GFP or Lenti-GFP-Cre, followed by glycolytic rate assays. The results indicate that MEFs in the PDHA1-T57D induction group (Lenti-GFP-Cre) exhibit higher levels of both basal and compensatory glycolysis compared to the control group (Lenti-GFP) (Figure 5A). Consistent with findings from a companion study by Zhang et al., which showed that PDHA1-T57D induction reduces maximal respiratory capacity and ATP production(11). These data suggest that MEFs with Cre induction compensate for diminished mitochondrial ATP production by shifting their reliance from oxidative phosphorylation to glycolysis. Since PDHA1-T57 phosphorylation inhibits carbon influx from pyruvate into the TCA cycle, accumulated pyruvate is preferentially reduced, facilitating NAD+ regeneration. This metabolic shift leads to a 50% increase in proton efflux rate (PER) derived from glycolysis in MEFs with Cre induction.

Figure 5.

The PDHA1-T57D mutation alters energy metabolism in the mouse model. (A) Glycolytic rate assay in PDHA1-T57D conditional KI MEFs. MEFs from PDHA1-T57D conditional KI mice were transfected with Lenti-GFP-Control or Lenti-GFP-Cre for 48 hours, followed by glycolytic rate assay. Quantification of metabolic parameters is presented as means ± standard deviations (n = 6). **P < 0.01 by unpaired t-test. Indirect calorimetry was conducted in PDHA1-T57D KI male and female mice. The PDHA1-T57D KI was activated via tamoxifen injection in male mice (T-M, n = 3) and female (T-F, n = 3) mice. Mice without tamoxifen treatment served as control groups (N-M and N-F, n = 3). Oxygen consumption (B), carbon dioxide production (C), energy expenditure (D), and respiratory exchange ratio (E) were monitored for all groups over 84 hours at room temperature. Results were presented as means ± standard deviations, *P < 0.05 by unpaired t-test.

Next, we performed genetic crosses to incorporate Cre recombinase-estrogen receptor T2 (Cre-ER^T2) into the PDHA1-T57D conditional KI mouse model, allowing for the expression of the T57D mutant upon tamoxifen administration. We extended our investigation to examine whole-body metabolism in the mouse model with the mutated PDHA1, starting with tamoxifen-treated mice that expressed the PDHA1-T57D variant to mimic PLK1-mediated phosphorylation. In this study, we measured the effects of PDHA1-T57D on oxygen consumption, carbon dioxide production, respiratory exchange ratio, and energy expenditure during both the light and dark phases. Tamoxifen-induced PDHA1-T57D expression suppressed oxygen consumption and energy expenditure in both male and female mice during both phases (Figures 5B, 5D), indicating that tamoxifen-induced mice had a lower metabolic rate compared to mice without induction. However, contrary to our prediction, the expression of PDHA1-T57D did not affect carbon dioxide production. Regardless of the phase, tamoxifen-induced PDHA1-T57D expression had no significant effect on carbon dioxide production in either male or female mice (Figure 5C). Additionally, we observed a significant increase in the respiratory exchange ratio in tamoxifen-injected mice (Figure 5E), suggesting that mice with the PDHA1-T57D KI relied more on carbohydrate oxidation than lipid oxidation to produce energy. This greater reliance on carbohydrates did not compensate for the suppression of metabolic rate in the mice due to PDHA1-T57D expression. This is consistent with our in vitro findings, where PLK1-mediated PDHA1-T57 phosphorylation inhibits oxidative phosphorylation, forcing cells to produce more ATP via glycolysis. The incomplete metabolism of carbohydrates reduces the energy metabolic efficiency of cells, which is reflected in the lower oxygen consumption and energy expenditure in mice when PDHA1-T57D is introduced.

PLK1 inhibition enhances the efficacy of DCA

We hypothesized that inhibiting PLK1 could restore PDHA1 protein levels, promote mitochondrial ROS generation, suppress tumor cell proliferation, and enhance apoptosis. Since DCA activates PDHA1 enzymatic activity by inhibiting PDK, we further proposed that combining DCA with PLK1 inhibition would synergistically enhance anti-cancer effects by both stabilizing and activating PDHA1.

To test this, we treated CrT cells (shCtrl or shPLK1) with varying doses of DCA. As shown in Figure 6A, PLK1 knockdown significantly enhanced the growth-inhibitory effect of DCA—by more than fourfold compared to control cells. Additionally, PLK1 depletion increased acetyl-CoA levels in both CrT cells and corresponding xenograft tumors (Figures S1A, S1B). Using MitoSOX Red staining, we observed that PLK1-depleted CrT cells exhibited elevated mitochondrial ROS generation upon DCA treatment, while control cells did not (Figure 6B). Similarly, colony formation was markedly reduced in CrT shPLK1 cells upon DCA exposure, indicating increased chemosensitivity (Figure 6C).

Figure 6.

PLK1 inhibition enhances the chemosensitivity of DCA in vitro. (A) MTT analysis of CrT cells (shCtl or shPLK1) treated with increasing doses of DCA. Results are presented as means ± standard deviations (n = 3). (B) Quantification of mitochondrial ROS levels in CrT cells (shCtl or shPLK1) treated with 2 mM DCA. Results are presented as means ± standard deviations (n=3). **P < 0.01; ns P > 0.05 by unpaired t-test. (C) Representative images of colony formation analysis of CrT cells (shCtl or shPLK1) after 2 mM DCA treatment. (D-F) MTT analysis of CrT(D), H1299(E), PC9(F) cells treated with treated with 5 mM DCA and/or 50 nM Onvansertib. Results are presented as means ± standard deviations (n = 6). ****P < 0.0001, ***P < 0.001**P < 0.01 by two-way ANOVA. (G-I) Representative images of colony formation analysis of CrT(G), H1299(H), PC9(I) cells treated with treated with 2.5 mM DCA and/or 25 nM Onvansertib. Results are presented as means ± standard deviations (n = 6). (J) Immunoblotting analysis of PARP,c-PARP, Caspase3, cleaved Caspase3 protein levels in CrT, H1299 and PC9 cells treated with 5 mM DCA and/or 50 nM Onvansertib. (K) Flow cytometry apoptosis detection using Annexin-V FITC/7AAD staining in Crt cells treated with 5 mM DCA and/or 50 nM Onvansertib. Results are presented as means ± standard deviations (n = 3). ***P < 0.001 by one-way ANOVA. (L) Dose-response map showing the inhibition rate of cell viability in CrT cells treated with DCA + Onvansertib. (M) Synergy matrix plot showing ZIP score for CrT cells treated with DCA + Onvansertib. The average ZIP synergy score is 9.58, with a maximum ZIP synergy score of 22.7.

We then investigated whether pharmacological PLK1 inhibition could replicate these effects. Treatment with the PLK1 inhibitor Onvansertib (50 nM) combined with DCA (5 mM) significantly suppressed cell viability (Figures 6D–6F) and colony formation (Figures 6G–6I) in CrT, H1299, and PC9 cells—demonstrating that this combination exerts a broad anti-proliferative effect across lung cancer cell lines. To determine whether the combination induces apoptosis, we examined apoptosis markers by immunoblotting. Co-treatment with DCA and Onvansertib led to robust induction of cleaved PARP and cleaved Caspase-3, compared to either drug alone (Figure 6J). Consistently, Annexin V/7-AAD staining confirmed a significant increase in apoptotic cell populations in the combination group relative to controls and monotherapies (Figure 6K). We further assessed synergy between DCA and Onvansertib using a 6×6 dose matrix and calculated synergy using the ZIP model. As shown in Figures 6L, 6M, strong synergy was observed across a range of concentrations, with a maximum ZIP synergy score exceeding 22, indicating 22% greater inhibition than expected under an additive model.

Given that PLK1 knockdown elevated mitochondrial ROS in the presence of DCA, we hypothesized that mitochondrial ROS might contribute to the synergistic effect. To test this, we evaluated mitochondrial ROS using MitoSOX Red staining following DCA and/or Onvansertib treatment. The combination group showed significantly higher ROS levels than either monotherapy (Figure 7A). To assess the role of mitochondrial ROS in apoptosis, we performed a rescue experiment using MitoTEMPO, a mitochondria-targeted antioxidant. MitoTEMPO effectively suppressed ROS levels (Figure 7B) and attenuated DCA + Onvansertib–induced apoptosis, as shown by reduced c-PARP and c-Caspase-3 levels in both CrT and H1299 cells (Figure 7C). These findings support that mitochondrial ROS is a key mediator of apoptosis induced by the combination therapy.

Figure 7.

PLK1 inhibition enhances the chemosensitivity of DCA through enhancing mitochondrial ROS, inhibiting glycolysis. (A, B) Quantification of mitochondrial ROS levels in CrT cells treated with treated with 5 mM DCA and/or 50 nM Onvansertib(A). CrT cells treated with or without MitoTEMPO (5μM) Results are presented as means ± standard deviations (n=4). * P <0.05 by one-way ANOVA; **P < 0.01 by unpaired t-test. (C) Immunoblotting analysis of PARP,c-PARP, Caspase3, cleaved Caspase3 protein levels in CrT and H1299 cells treated with 5 mM DCA, 50 nM Onvansertib, 5μM MitoTEMPO. (D) Mitochondrial Stress Test of CrT cells. CrT cells treated with 5 mM DCA and/or 50 nM Onvansertib for 16h, followed by Mitochondrial Stress Test. Quantification of metabolic parameters is presented as means ± standard deviations (n=6). *P<0.05; **P<0.01 by , by one-way ANOVA. (E) Glycolytic Rate Assay of CrT cells. CrT cells treated with treated with 5 mM DCA and/or 50 nM Onvansertib, followed by Glycolytic Rate Assay. Quantification of metabolic parameters are presented as means ± standard deviations (n=6). ****P < 0.0001, ***P < 0.001**P < 0.01, *P<0.05, ns. not statistically significant, by one-way ANOVA.

Since metabolic reprogramming from OXPHOS to glycolysis is a hallmark of cancer progression and a central theme in our study, we performed additional assays to evaluate how the combination of DCA and Onvansertib affects this process.

Specifically, we conducted a Mitochondrial Stress Test (MST) and a Glycolytic Rate Assay (GRA) in CrT cells treated with DCA and/or Onvansertib. The combination treatment led to an increase in Spare Respiratory Capacity compared to either treatment alone, suggesting improved mitochondrial oxidative function (Fig. 7D). Consistently, the GRA revealed a decrease in Basal Glycolysis, Compensatory Glycolysis, and the percentage of Proton Efflux Rate (PER) from glycolysis in cells receiving the combination treatment (Fig. 7E), indicating a suppression of the glycolytic phenotype and a shift back toward mitochondrial respiration. Together, these results support the concept that co-targeting PDHA1 and PLK1 can inhibit the metabolic reprogramming, potentially leading to enhanced apoptosis and reduced proliferation of lung cancer.

To validate the in vitro findings in vivo, we evaluated the combination therapy in nude mice. Mice were treated with vehicle, DCA, Onvansertib, or both drugs for three weeks. While DCA or Onvansertib alone had minimal effects, the combination treatment dramatically suppressed tumor growth and volume (Figures 8A, 8B). The reduction in tumor numbers at the endpoint occurred primarily in the placebo and monotherapy groups, due to humane euthanasia following rapid tumor progression. Histological analysis revealed lower cellular density and increased apoptotic bodies in tumors from the combination group (Figure 8C). IHC staining confirmed that the combination significantly reduced proliferation (Ki67) and increased apoptosis (cleaved Caspase-3) (Figures 8D–E). Notably, body weight remained stable across all groups (Figure S2), indicating treatment was well tolerated.

Figure 8.

DCA plus Onvansertib synergistically inhibits tumor growth in vivo. Nude mice were inoculated with CrT cells, randomized into four groups (n = 5 per group), and treated with placebo, DCA, Onvansertib, or a combination for 4 weeks. (A) Tumor volumes over the 4-week treatment period. *P < 0.05 by two-way ANOVA. (B) Representative images of xenograft tumors at harvest. (C) Representative hematoxylin and eosin (H&E) staining of xenograft tumors derived from CrT cells. Red arrows indicate apoptotic cells. Images captured at 20x (scale bars, 200 μm) and 40x magnification (scale bars, 100 μm). (D) Immunofluorescence (IF) staining for cleaved-caspase 3, a marker of apoptosis, reveals increased apoptotic cells in tumors treated with DCA and Onvansertib. Quantification of cleaved-caspase 3 staining is shown on the right. ****P < 0.0001 by one-way ANOVA. Scale bars, 50 μm. (E) IF staining for Ki67 demonstrates reduced proliferation in tumors treated with DCA and Onvansertib. Quantification of Ki67 staining is shown on the right. ****P < 0.0001 by one-way ANOVA. Scale bars, 50 μm.

Together, these findings demonstrate that PLK1 inhibition sensitizes lung cancer cells to DCA treatment by enhancing mitochondrial ROS, inhibiting glycolysis, and inducing apoptosis, and that the combination therapy has potent anti-tumor activity in vitro and in vivo.

High PLK1 and low PDHA1 predict poor clinical outcomes in NSCLC

Aberrant upregulation of PLK1 in cancer cells influences oxidative phosphorylation by promoting PDHA1 degradation, leading to mitochondrial dysfunction and enhanced cancer progression. To investigate whether this PDHA1-dependent mechanism of metabolic reprogramming by PLK1 is relevant to clinical setting, we performed IHC staining with anti-PLK1 and anti-PDHA1 antibodies on a tissue microarray (TMA) containing 216 NSCLC patient samples from a clinical cohort. As shown in Figures 9A and 9B, higher PLK1 levels were associated with poorly differentiated tumors. Although, the difference in PDHA1 levels between the moderately and poorly differentiated tumor groups did not reach statistical significance (P = 0.086), the trend toward higher PDHA1 expression in moderately differentiated tumors may still suggest a biological association between PDHA1 levels and tumor grade. suggesting that the PLK1/PDHA1 axis plays a significant role in tumor differentiation and aggressiveness. This conclusion was further supported by Kaplan-Meier analysis (Figure 9C). Survival curves demonstrated that elevated PLK1 expression was significantly correlated with worse overall survival in NSCLC patients. Similarly, low PDHA1 expression was associated with poorer overall survival outcomes. In a companion study by Zhang et al., we also found a negative correlation between PLK1 and PDHA1 protein levels, supporting the notion that PLK1 mediates PDHA1 downregulation both in vitro and in patient samples(11). In summary, these findings highlight the potential of the PLK1/PDHA1 axis as a diagnostic target for predicting clinical outcomes in NSCLC patients.

Figure 9.

High PLK1 and low PDHA1 expression predict poor clinical outcomes in NSCLC. (A, B) Immunohistochemistry (IHC) staining of PLK1 and PDHA1 was performed on a tissue microarray (TMA) containing 216 NSCLC patient samples from a clinical cohort. (A) Protein expression levels in NSCLC samples with well, moderate, and poor differentiation are shown. A statistically significant (*P < 0.05 by one-way ANOVA) increase in PLK1 staining is observed as differentiation decreases. (B) Representative images demonstrate the reciprocal staining patterns of PDHA1 and PLK1 in moderately and poorly differentiated NSCLC samples. PDHA1 expression shows granular cytoplasmic staining, while PLK1 expression shows both nuclear and cytoplasmic staining. (C) Kaplan-Meier 10-year overall survival analysis of NSCLC patients based on PLK1 and PDHA1 expression levels. High PLK1 expression is associated with significantly reduced survival probability (***P < 0. 001), and low PDHA1 expression is similarly associated with poorer outcomes (**P < 0. 01).

Discussion

Accumulating evidence highlights the critical role of PLK1 in various cellular processes. Over the past decade, our lab has made significant contributions to understanding the role of PLK1 in cancer progression and therapeutics(14, 15). In the case of lung cancer, elevated PLK1 expression markedly accelerates oncogenic Kras-induced LUAD in a GEM model by activating rearranged during transfection (RET) signaling (9). Additionally, PLK1-mediated phosphorylation of aryl hydrocarbon receptor (AHR) promotes LUAD metastasis via the type 2 deiodinase (DIO2) – thyroid hormone (TH) signaling pathway (16). Single-cell analysis further identified PLK1 as a driver of immunosuppressive tumor microenvironment in LUAD (17). In prostate cancer, we demonstrated that PLK1 kinase activity activates androgen receptor (AR) signaling through the SREBP (sterol regulatory element-binding proteins) pathway. Notably, PLK1 inhibition enhances the efficacy of enzalutamide in castration-resistant prostate cancer (CRPC) (18). Moreover, we showed that PLK1 phosphorylation of CLIP-170 and p150^Gued, two microtubule plus-end-binding proteins, increases microtubule dynamics, ultimately contributing to docetaxel resistance in prostate cancer (19–21). We also found that PLK1 phosphorylation of Mre11 leads to premature termination of the DNA damage checkpoint and impaired DNA repair, which can predict the efficacy of olaparib (22). More recently, we demonstrated that PLK1 phosphorylation of BRD4 results in its protein degradation, thereby influencing the efficacy of inhibitors for bromodomain and extra-terminal domain (BET) (23).

The involvement of PLK1 in energy metabolism was previously reported by our team (24). In that study, we demonstrated that PLK1 phosphorylation of the PTEN tumor suppressor activates the PI3K/AKT/mTOR pathway, resulting in a metabolic switch from OXPHOS to glycolysis in prostate cancer (24). However, a limitation of the previous study was that the metabolic shift caused by PLK1-mediated PTEN inactivation represented an indirect effect. Consequently, whether PLK1-associated metabolic reprogramming occurs through its direct phosphorylation of key enzymes involved in energy metabolism remains unknown. In the companion manuscript by Zhang et al, we identified PDHA1 as a substrate of PLK1 and demonstrated that PLK1 phosphorylation of PDHA1 promotes its protein degradation via mitophagy activation(11). Given the critical role of PDHA1 in energy metabolism, we propose that this discovery unveils a direct mechanism underlying PLK1-mediated metabolic reprogramming.

Our findings, derived from transformed lung epithelial cells expressing the PDHA1-T57A variant and PDHA1-T57D variant, strongly support the hypothesis that PLK1-mediated phosphorylation of PDHA1 triggers a metabolic shift from OXPHOS to glycolysis in vitro. Notably, cells mimicking PDHA1-T57 phosphorylation rely more on the malate-aspartate shuttle (MAS) than on glucose-derived pyruvate to maintain the TCA cycle. The MAS plays a crucial role in maintaining redox balance and energy production. It may be a general compensatory mechanism of MAS, but the metabolic stress of cancer varies in contexts, especially under MDH1/2 mutation. Glutaminolysis could be a major contributor in our PDHA1-T57D cell line due to the large amount and higher fraction of non-13C labeled glutamate (Figure 3E and 3F), which was probably derived from glutamine and L-Glutamine-N2–15N will be the best tracer for future studies.

This metabolic shift was also observed in MEFs and transgenic mice conditionally expressing the PDHA1-T57D variant, emphasizing the role of PLK1 in metabolic reprogramming in vivo. While no significant changes in oxidative phosphorylation were observed following Cre induction, glycolysis was markedly promoted. In MEF cells, approximately 70% of energy is supplied by mitochondria, with pyruvate serving just one of several carbon sources for the TCA cycle. Thus, the inhibition of pyruvate metabolism via PLK1-mediated PDHA1 phosphorylation may not be sufficient to completely disrupt OXPHOS. Similar results were obtained in whole-body metabolism tests and calorimetry experiments in mice.

These findings underscore the compensatory role of the aspartate-malate shuttle under conditions of PLK1-mediated PDHA1-T57 phosphorylation. This mechanism leads to pyruvate accumulation and increased lactate production, ultimately tipping the energy balance towards glycolysis. Notably, these metabolic changes are accompanied by alterations in mitochondrial function and cell cycle-related activities, as detailed in the companion manuscript by Zhang et al(11). Previous studies have shown that mitochondrial function and morphology are dynamically regulated during the cell cycle. This regulatory interplay is closely associated with cancer development, as PLK1 plays a dual role in cell cycle regulation and oncogenesis. Elucidating the function of PLK1 in OXPHOS within cancer cells is expected to advance our understanding of the dynamic relationship between the cell cycle and cell metabolism.

Considering the inhibitory effect of PDHA1 on cancer progression, we propose that simultaneously targeting PLK1-dependent PDHA1 degradation and PDHA1 activity could serve as a viable approach for cancer therapy. Previous studies have shown that DCA enhances certain cancer treatments by upregulating PDH activity and promoting mitochondria-dependent apoptosis. However, despite efforts to incorporate DCA into cancer treatment regimens, its clinical application has been limited due to the excessively high doses required for effectiveness and the rick of peripheral neuropathy (25). Given that PLK1 mediates PDHA1 protein degradation and mitochondrial dysfunction, we hypothesize that PLK1 inhibition could enhance the therapeutic efficacy of DCA by upregulating PDHA1 protein levels. Notably, our results demonstrate that cancer cells with reduced PLK1 expression are four times more sensitive to DCA treatment, supporting the feasibility of a novel combination therapy involving a PDHA1 activator, such as DCA, and a PLK1 inhibitor.

Several very recent studies have reported ZIP synergy scores in the 10–20 range for effective drug combinations in lung cancer models: Ye et al. showed that cannabidiol (CBD) combined with dasatinib induced significant synergistic apoptosis in A549 cells, with a ZIP score of 10.19, and demonstrated in vivo efficacy(26). Li et al. reported the combination of syrosingopine and UK-5099 in NSCLC models, yielding average ZIP scores of 11.31 (H661) and 12.73 (PC-9), and maximum ZIP scores of 16.97 and 19.04, respectively(27). Noguchi et al. demonstrated an average ZIP synergy score of 14.25 for the combination of a mitochondria-shaping protein Opa1 inhibitor (MYLS22) with gefitinib in PC9M2 cells(28). Wang et al. found that betulinic acid combined with EGFR-TKIs (gefitinib/osimertinib) showed ZIP scores >10 and strong synergistic effects in A549 and H1299 cells, with synergy-associated antitumor responses exceeding 25% in some cases(29). Compared to these examples, the ZIP score of 22 observed in our study suggests a highly synergistic interaction between Onvansertib and DCA, reinforcing the therapeutic potential of this co-targeting strategy. Furthermore, our in vivo data from xenograft models corroborate this hypothesis, highlighting the potential of such a combinatorial approach. These findings provide a strong rationale for future clinical trials evaluating the efficacy of combining a PLK1 inhibitor and a PDHA1 activator as a promising strategy for cancer therapy.

Materials and Methods

Cell culture, chemicals and reagents

The human bronchial epithelial cell line BEAS-2B (CRL-3588) and A549 (CCL-185), H1299 (CRL-5803) cells were obtained from the American Type Culture Collection (Rockville, MD). PC9 cell lines were kindly provided by Dr. Qiou Wei at the University of Kentucky. Chromium-transformed BEAS-2B (CrT) cells were generated as previously described (30). CrT and A549 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, D5796) supplemented with 10% fetal bovine serum (FBS) (R&D, S11550H) and 1% penicillin/streptomycin (Sigma-Aldrich, A5955) at 37°C in a humidified atmosphere containing 5% CO₂. Mouse embryonic fibroblasts (MEFs) were isolated from 13.5-day embryos of female C57BL/6 mice with PDHA1 wide-type or PDHA1-T57D conditional Knock-in/Cre-ERT2 genotype. The embryos were minced and digested with trypsin (Sigma-Aldrich, T4049) for 20 minutes at 37°C. The tissue cells were centrifuged at 600×g for 5 minutes and resuspended in complete DMEM containing 10% FBS and 1% penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂. When necessary, MEFs were treated with 1 μM Tamoxifen (MCE, HY-13757A) for 48 hours to induce Cre recombination. Sodium dichloroacetate (DCA) (Sigma-Aldrich,347795) was obtained from Sigma-Aldrich. Mito-TEMPO (MCE, HY-112879) was purchased from MCE. Onvansertib was kindly provided by Cardiff Oncology.

Immunoblotting and Antibodies

Cells were harvested and lysed in ice-cold lysis buffer containing 20 mM Tris-HCl (pH 8.0), 0.5% NP-40, 5 mM EGTA, 1.5 mM EDTA, 0.5 mM sodium vanadate, and 150 mM NaCl, supplemented with protease and phosphatase inhibitor cocktails. Lysates were briefly sonicated and centrifuged to remove debris. Protein concentrations were determined using the Protein Assay Dye Reagent (Bio-Rad) according to the manufacturer’s instructions. Equal amounts of total protein were denatured in SDS loading buffer, separated by SDS-PAGE, and transferred onto PVDF membranes. Membranes were then blocked and incubated with specific primary antibodies, followed by appropriate HRP-conjugated secondary antibodies. Signal detection was performed using enhanced chemiluminescence (ECL) reagents.

Antibodies against PARP (Cell Signaling Technology, 9542), full-length (Cell Signaling Technology, 9662) and cleaved Caspase-3 (Cell Signaling Technology, 9661), and α-Tubulin (Cell Signaling Technology, 2144) were purchased from Cell Signaling Technology. Antibody against Actin (Santa Cruz, sc-8432) was obtained from Santa Cruz.

Cell Viability Assay

Cells were plated in 96-well plates and allowed to adhere overnight. The following day, cells were treated with the indicated compounds for 72 hours. Cell viability was assessed using the AquaBluer reagent (MultiTarget Pharmaceuticals LLC) according to the manufacturer’s instructions, which measures the metabolic activity of viable cells. Each condition was tested in six technical replicates, and experiments were independently repeated at least three times. Fluorescence values were background-corrected, and results are presented as the percentage of viable cells relative to untreated controls.

Colony Formation Assay

Cells were seeded in 6-well plates at a density of 1500/well and allowed to attach overnight. The next day, cells were treated with the indicated agents and maintained in culture for 10–14 days to allow colony formation. At the endpoint, colonies were fixed with 4% paraformaldehyde for 15 minutes and stained with 1% crystal violet for 30 minutes at room temperature. Excess stain was rinsed off with water, and plates were air-dried. Experiments were repeated independently at least three times.

Apoptosis Assay

Apoptosis was evaluated using the Annexin V-FITC/7-AAD staining kit according to the manufacturer’s instructions. Briefly, cells were harvested after indicated treatments, washed twice with cold PBS, and resuspended in binding buffer at a concentration of 1 × 10⁶ cells/mL. A total of 100 μL of the cell suspension was incubated with 5 μL of Annexin V-FITC and 5 μL of 7-AAD for 15 minutes at room temperature in the dark. After incubation, 400 μL of binding buffer was added to each tube, and samples were analyzed by flow cytometry within one hour. Data were analyzed using FlowJo software. Annexin V⁺/7-AAD⁻ cells were considered early apoptotic, while Annexin V⁺/7-AAD⁺ cells were considered late apoptotic or necrotic.

Mitochondrial ROS Detection

Cells were seeded in 96-well plates and were treated with the indicated agents. To detect mitochondrial superoxide, cells were incubated with 5 μM MitoSOX^™ Red reagent (Thermo Fisher Scientific) in HBSS for 30 minutes at 37°C in the dark. MitoSOX Red selectively reacts with mitochondrial superoxide (O₂•⁻), producing red fluorescence upon oxidation. After incubation, cells were washed 2 times with PBS. Fluorescence intensity was measured using a microplate reader with excitation/emission settings of 396/610 nm.

Mouse xenograft model

Two weeks after CrT cells (1 × 10⁶) were inoculated subcutaneously into NU/J mice, the mice were treated with various drugs. Onvansertib and DCA were administered orally twice a week at doses of 45 mg/kg and 100 mg/kg, respectively. Tumor volumes were calculated using the formula V=L×W²/2 (where V is volume [cubic millimeters], L is length [millimeters], and W is width [millimeters]). Treatment administration and outcome assessment were conducted by separate team members. The individual responsible for measuring tumor size and recording data was blinded to the treatment assignments of each group. Mice showing clear signs of illness or pain—including weight loss exceeding 15% of normal body weight, abnormal posture, lethargy, moribund behavior, or dehydration—were excluded from the analysis. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Kentucky (Protocol No. 2020–3681).

Indirect calorimetry

Indirect calorimetry was performed on 8 to 12-week-old male and female C57BL/6 mice with the PDHA1-T57D conditional Knock-in/Cre-ERT2 genotype. Pdha1 T57D KI was induced by tamoxifen injection in both male (T-M, n=3) and female (T-F, n=3) mice. The mice were maintained at the standard housing room temperature for rodents (74 ± 2°F or 23 ± 1°C) with a 12-hour light/12-hour dark cycle. Oxygen consumption, carbon dioxide production, respiratory exchange ratio, and energy expenditure were measured using the Sable Promethion system (Sable Systems International, North Las Vegas, NV). Indirect calorimetry measurements were recorded for each mouse (n=12) over a minimum of 84 hours, covering both the pre-acclimation and post-acclimation periods. Data generated were analyzed using CalR, as described by Mina et al. (31).

Stable isotope tracer metabolic analysis

Cells were grown in DMEM medium containing 25 mM [U-13C]-glucose on 10 cm plates in triplicate. Media samples (50 μL) were collected at 3, 6, 12, and 24 hours and immediately flash- frozen in liquid nitrogen. Cells were quenched using CH₃CN and extracted using a modified Folch method, as previously described (32). Chloroform was added to the CH3CN-water mixture, and the polar fraction, lipid fraction, and protein were enriched and extracted from CH3CN-chloroform mixture for stable isotope tracer metabolic analyses.

Nuclear magnetic resonance (NMR)

Polar extracts were reconstituted in 35 μL D₂O containing 17.5 nmol DSS as internal chemical shift reference and concentration standard. ¹H PRESAT and ¹H{¹³C}-HSQC spectra were recorded at 15°C on a 16.45 T Bruker Avance III spectrometer equipped with a 1.7 mm inverse triple resonance cryoprobe as previously described (33). Raw data were processed and analyzed using MRestNova. Metabolites and isotopomers were identified using in-house databases (34), and quantified as μmol/mg protein and as fractional enrichments F as: F = A(¹³C)/[A(¹³C)+A(¹²C)]. Where A(¹³C) is the peak area of a proton attached to ¹³C, and A(¹²C) is the peak area of a proton attached to ¹²C.

UHR IC-FT-MS analysis

Ion chromatography-ultra high-resolution-Fourier transform-MS (IC-UHR-FTMS) was performed as previously described (35). Briefly, polar extracts were reconstituted in 20 μL Nanopure water, and analyzed by a Dionex ICS-5000+ ion chromatograph interfaced to an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific, San Jose) operating at a resolution setting of 500,000 (FWHM at m/z 200) on MS1 acquisition to capture all ¹³C isotopologues. The chromatograph was outfitted with a Dionex IonPac AG11-HC-4 μm RFIC&HPIC (2 × 50 mm) guard column upstream of a Dionex IonPac AS11-HC-4 μm RFIC&HPIC (2 × 250 mm) column. Chromatography and mass spectrometric settings were the same as described previously (36) with an m/z range of 80 to 700. Metabolites and their isotopologues were identified and their peak areas were integrated and exported to Excel via the TraceFinder 3.3 (Thermo) software package. Peak areas were corrected for natural abundance as previously described (37), after which fractional enrichment and μmoles metabolites/g protein were calculated to quantify ¹³C incorporation into various metabolites.

Glycolytic Rate Assay (GRA)

Glycolytic Rate Assay was performed using a Seahorse XF96 Extracellular Flux Analyzer (Agilent Technologies) according to the manufacturer’s protocol. A total of 1 × 10⁴ Crt cells were seeded in Seahorse XF96 V3 PS cell culture microplates and treated with DCA and/or Onvansertib for 16h. Prior to the assay, cells were washed and incubated in Seahorse XF Base Medium supplemented with 2 mM glutamine and 10 mM glucose (pH adjusted to 7.4) in a non-CO₂ incubator for 1 hour. Basal Proton Efflux Rate (PER) was measured, followed by sequential injections of Rotenone/Antimycin A (to inhibit mitochondrial respiration) and 2-Deoxy-D-glucose (2-DG) to inhibit glycolysis. Basal glycolysis, compensatory glycolysis, and the percentage of PER derived from glycolysis were calculated using Wave software based on standard Agilent GRA analysis algorithms.

Mitochondrial Stress Test (MST)

Mitochondrial stress test was performed using a Seahorse XF96 Extracellular Flux Analyzer (Agilent Technologies). A total of 1 × 10⁴ Crt cells were seeded in Seahorse XF96 V3 PS cell culture microplates and treated with DCA and/or Onvansertib for 16h. On the day of the assay, cells were washed and switched to Seahorse XF Base Medium supplemented with 10 mM glucose, 1 mM pyruvate, and 2 mM glutamine (pH adjusted to 7.4), followed by a 1-hour incubation in a non-CO₂ incubator. Basal oxygen consumption rate (OCR) was measured, followed by sequential injections of oligomycin (ATP synthase inhibitor), FCCP (uncoupler), and rotenone/antimycin A (complex I and III inhibitors). Basal and maximal OCR values were calculated based on the area under the curve (AUC) before and after FCCP injection, respectively, according to standard Seahorse XF Cell Mito Stress Test analysis.

Tissue Microarray (TMA) and IHC Staining

TMA slides containing 216 patient samples of NSCLC from a clinical cohort were obtained from the University of Kentucky Markey Cancer Center. IHC staining for PLK1 (MILLIPORE, 05–844) and PDHA1 (Cell Signaling Technology, 3205) was performed by the Biospecimen Procurement and Translational Pathology Shared Resource Facility at the Markey Cancer Center. Evaluation and scoring were conducted by Dr. Derek Allison, a pathologist in the Department of Pathology and Laboratory Medicine at the University of Kentucky. Approval for the use of human tissues was obtained from the University of Kentucky IRB under Dr. Derek Allison’s name.

Online survival analysis

The Kaplan-Meier plotter online survival analysis tool [32] (http://kmplot.com/analysis/), which integrates clinical and gene expression data for lung, breast, gastric, and ovarian cancers, was used to assess the prognostic significance of various genes in NSCLC. Based on the median expression value of each gene, patient samples were categorized into high- and low-expression groups. In this study, the analysis was performed using the tool’s default parameters for histology, stage, grade, gender, and smoking history. AJCC stage T=2, N=0, and M=0 were applied. “Univariate” Cox regression and “exclude biased arrays” options were selected for array quality control. For each survival plot, the log-rank p-value and hazard ratio (HR) with 95% confidence intervals were calculated and displayed on the main plot.

Statistical analysis

Data from cell viability assays, colony formation, IF staining, IHC staining, and tumor mass measurements were analyzed using an unpaired Student’s t-test. Tumor growth among different groups was analyzed with a two-way ANOVA. All data are presented as mean ± SEM. Statistical analyses were performed using the GraphPad Prism 9 software package. Statistical significance is indicated as follows: P < 0.05; P < 0.01; *P < 0.001; **P < 0.0001; n.s., not significant.

Data availability statement:

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.