Abstract
Background/Aim: Oxidative phosphorylation (OXPHOS) is implicated in cancer progression and metastasis. However, its role in lung adenocarcinoma (LUAD) is unknown. We assessed OXPHOS in LUAD cases and cell lines and investigated the effect of OXPHOS inhibition on LUAD cells. Materials and Methods: The cases with high expression of OXPHOS-related genes and peroxisome proliferator-activated receptor gamma (PPAR-γ) were extracted using RNA-seq data from The Cancer Genome Atlas (TCGA) LUAD dataset and the clinicopathological features and survival were assessed. Resected LUAD specimens were stained for PPAR-γ. Real-time qPCR and western blot were used to examine the expression of OXPHOS- and glycolysis-related genes and proteins in four LUAD cell lines. Cell proliferation was evaluated in LUAD cells treated with OXPHOS inhibitors. Results: The TCGA database analysis revealed that cases with high OXPHOS or PPAR-γ expression had a worse prognosis (p=0.07 and p=0.01, respectively). High OXPHOS cases were associated with lymph node metastasis (p<0.01). PPAR-γ was expressed only in the peripheral area of the papillary component of LUAD. We identified A549, HTB181 and H322 as OXPHOS-high type cells and H596 as OXPHOS-low type cells. Oligomycin treatment inhibited cell proliferation in the OXHOS-high cells (0.72-, 0.69-, and 0.77- fold change in oligomycin vs. DMSO, for A549, HTB181, and H322 cells, respectively, p<0.01) but not in the OXPHOS-low cells. Conclusion: High expression of OXPHOS-related genes and PPAR-γ is a poor prognostic factor in LUAD. The levels of OXPHOS vary among cases and within different areas of the tumor. Targeting OXPHOS metabolism may represent a novel therapeutic approach for treating LUAD.
Lung adenocarcinoma (LUAD) is the most common histological type of lung cancer and a leading cause of cancer-related deaths worldwide (1, 2). The main treatment strategies for LUAD include surgery, radiotherapy, and chemotherapy (3-5). In recent years, chemotherapy for LUAD targeting tyrosine kinases and immune checkpoints has attracted increasing attention (6-9). However, the prognosis of patients with LUAD, particularly of those exhibiting metastases, remains unsatisfactory (1). Identification of novel therapeutic targets is crucial for further improving the prognosis of patients with advanced LUAD.
Previous studies have reported that glycolysis is up-regulated in cancer cells compared to normal cells, leading to the assumption that oxidative phosphorylation (OXPHOS) is universally down-regulated in various cancers (10, 11). This metabolic change is known as the Warburg effect (10). However, recent evidence suggests that mitochondrial metabolism is intact in some cancers, including leukemia, lymphoma, melanoma of the high-OXPHOS subtype, and in endometrial cancer (12, 13). In LUAD patients, it is unclear whether OXPHOS is impaired. A factor that impedes the analysis of metabolism in LUAD is intratumoral heterogeneity (14). LUAD is a tumor with varying degrees of malignancy, depending on the subtype (15). The lepidic component, a non-invasive component, is often located in the peripheral area of LUAD, which is an aerobic environment (16). In contrast, the invasive component is usually found in the center of the LUAD, often in a hypoxic environment (17). Recent studies suggest that tumors may be metabolically heterogeneous, and that cancer stem cells with high metastatic and tumorigenic potential depend more on OXPHOS metabolism (18, 19). Whether OXPHOS is used in the invasive component and whether LUADs use OXPHOS in the aerobic environment to which they are exposed when metastasizing are crucial questions for the control of LUAD progression. For pursuing a treatment strategy concept targeting OXPHOS metabolism, characterizing the metabolic activity of OXPHOS in LUAD is essential. Furthermore, several recent trials have highlighted mitochondrial metabolism as a target for antitumor therapy (see Additional File 1) (20, 21). However, the antitumor effects of targeting OXPHOS metabolism on LUAD remain unclear.
As the presence and localization of OXPHOS metabolism in LUAD and the potential antitumor effects of OXPHOS inhibitors remain unknown. Our study aimed to elucidate the genetic patterns and therapeutic potential of OXPHOS-related molecules in LUAD. To achieve this goal, we evaluated the clinical background characteristics and prognosis of LUAD patients based on OXPHOS-related gene status. Additionally, we assessed the antitumor efficacy of OXPHOS inhibitors targeting mitochondrial respiration in LUAD.
Materials and Methods
Clinical trial data. Data from clinical trials evaluating patients with lung cancer treated with OXPHOS inhibitors were collected from clinicaltrials.gov on 14 June 2021, using the following keywords: OXPHOS inhibitors (metformin, phenformin, IACS-010759, ME344, oligomycin) and lung cancer.
The Cancer Genome Atlas data collection and OXPHOS-related genes. Gene expression profiles and clinicopathological information, such as age, sex, tumor stage, lymph node metastasis, and survival data for LUAD were collected from the The Cancer Genome Atlas (TCGA) database (https://tcga-data.nci.nih.gov/tcga/). Patients in the TCGA-LUAD database were analyzed using RNA-seq data from TCGA raw count values normalized using edgeR (Ver 4.2.1). OXPHOS or glycolysis-associated genes were obtained from the Kyoto Encyclopedia of Genes and Genomes (https://www.genome.jp/pathway/hsa00190; glycolysis: https://www.genome.jp/pathway/hsa00010). Heatmaps were created with the scaled data and clustered by correlation coefficients (Pearson correlation) in both the gene and the sample directions. We defined the high and low OXPHOS groups by the sum of the Z scores of the OXPHOS-related genes (n=30, respectively). Clinicopathological data and overall survival (OS) were compared between the high- and low-OXPHOS gene expression groups.
We further validated the prognostic impact of peroxisome proliferator-activated receptor γ (PPARγ), a representative OXPHOS-related gene, on LUAD and lung squamous cell carcinoma (LUSC) in TCGA. The respective cutoffs were LUAD; 6.29 and LUSC; 5.04 according to the fragments per kilobase of exon per million reads mapped (FPKM) values.
Immunohistochemistry staining of PPARγ in resected LUAD. Fifty-five LUAD samples resected at Hiroshima University between 2013 and 2014 were enrolled in this study. For immunohistochemistry (IHC) staining, formalin-fixed paraffin-embedded sections (4 μm) were deparaffinized with xylene, rehydrated, and subjected to antigen retrieval in a microwave oven for 20 min. After inhibiting endogenous peroxidase activity, individual slides were incubated at 4°C with PPARγ antibody (ab59256; Abcam, Tokyo, Japan) for 8h. After the incubation, slices were washed and stained with an HRP-conjugated secondary antibody (1:2000; ab97051) for 1h at 20°C. IHC staining and evaluation were performed by an experienced pathologist (K. Kushitani).
Cell culture. The LUAD cell lines (A549, ECACC, Salisbury, UK, cat#86012804; HTB181, ATCC, Manassas, VA, USA, cat#HTB-181; H322, ECACC, cat#95111734, and H596, ATCC, cat#HTB-178) were incubated at 37°C in 5% CO2. A549 cells were cultured in DMEM (Thermo Fisher Japan, Tokyo, Japan) supplemented with 10% exosome-depleted fetal bovine serum (FBS; Gibco, Life Technologies, Tokyo, Japan) and 50 IU/ml penicillin (Gibco, Life Technologies). HTB181, H322, H596 cells were cultured in RPMI-1640 (Thermo Fisher) supplemented with 10% exosome-depleted FBS (Gibco) and 50 IU/ml penicillin (Gibco, Life Technologies) at 37°C in 5% CO2.
Quantitative real-time polymerase chain reaction. Total RNA was isolated using RNeasy Plus (QIAGEN, Tokyo, Japan) from A549, HTB181, H322, and H596 cell lines according to the manufacturer’s instructions. Quantitative real-time polymerase chain reaction (RT-qPCR) analysis was performed using the TaqMan Gene Expression Assay (Thermo Fisher). The expression assays were performed according to the manufacturer’s protocol. Primers were used to detect the OXPHOS-related genes PPARγ (cat#Hs011155513) and UCP2 (cat#Hs01075227) and the glycolysis-related genes, LDH (cat#Hs01378790) and HIF1α (cat#Hs00153153). All primers were purchased from Thermo Fisher Scientific.
Western blotting. Twenty-five μg of total protein was loaded onto SDS-PAGE on 10% gels (#4561033, Bio-Rad Laboratories, Tokyo, Japan) and transferred onto PVDF membranes. β-Actin was used as a loading control to ensure equal protein loading across samples. The membranes were blocked with 5% milk and incubated at 4°C for 16 h in Tris-buffered saline containing primary antibodies against the mitochondrial complex V (ATP5a; ab110411 Abcam), PPARγ (2442S; Cell Signaling Technology, Tokyo, Japan), and UCP-2 (89326; Cell Signaling Technology) followed by incubation with secondary antibodies (mouse anti-IgG pAb-HRP; Code 330 and rabbit anti-IgG pAb-HRP; Code 458, both from MBL Life science, Tokyo, Japan) at 20°C.
OXPHOS inhibitors and CyQUANT cell proliferation assay. LUAD cell lines (A549, HTB181, H322, H596) were treated with OXPHOS inhibitors, namely metformin, a mitochondrial complex I inhibitor (M605000 Sigma-Aldrich, Tokyo, Japan), or oligomycin, a mitochondrial complex V inhibitor (75351-5MG Sigma-Aldrich) or DMSO 1 μM as control. CyQuant cell proliferation assay experiments were repeated 3 times.
We treated LUAD cells (A549, HTB181, H322, H596) with oligomycin and metformin or DMSO and evaluated the growth curve on days 2, 4, and 6, and confirmed that the effect of OXPHOS inhibitors could be evaluated at day 4. The growth curve of all LUAD cell lines was evaluated following treatment with OXPHOS inhibitors for four days.
Statistical analysis. Data of TCGA are presented as numbers or mean values. We evaluated the differences between OXPHOS high group and OXPHOS low group using the Fisher’s exact test for categorical variables or the Mann–Whitney U-test for continuous variables. Survival was analyzed using the Kaplan–Meier method. Statistically significance was considered when p<0.05. Heatmaps of TCGA data were created using R (Version 4.2.1). Kaplan–Meier analysis, Fisher’s exact test, and Mann–Whitney U-test were performed using JMP Pro (Version 14.0; SAS Institute, Inc., Cary, NC, USA).
Ethical statement. The institutional review board of the participating institutions approved this study (Hiroshima University Hospital: E2022-0244).
Results
Microarray data analysis. We assessed the expression profile of genes related to OXPHOS and glycolysis and survival in patients with LUADs based on TCGA database. A heatmap was constructed of tumors expressing OXPHOS metabolism-related genes, including LUADs from TCGA (Figure 1A). Glycolysis-related genes were highly expressed in LUAD samples (Figure 1B). To identify tumors highly expressing OXPHOS-related genes, we sorted the z-score of OXPHOS-related gene expression from right to left, as indicated in Figure 1C. Glycolysis-related genes were also highly expressed in some tumors with high expression of OXPHOS genes. The clinicopathological factors were compared in 30 patients with the highest expression and in 30 patients with the lowest expression of OXPHOS genes (Table I). The high expression OXPHOS group had significantly more lymph node metastases than the low group (high 53.3% vs. low 13.3%, p<0.01), although the T factor, mainly defined by tumor size, and other clinical factors were not significantly different. The high-expression OXPHOS group also had a marginally worse prognosis (Figure 1D) (the 5-year OS, high 31.3% vs. low 70.7%, p=0.07).
Expression of genes related to oxidative phosphorylation and glycolysis in lung adenocarcinoma (LUAD). (A) Oxidative phosphorylation mRNA expression in LUAD patients. (B) Glycolysis mRNA expression in patients with LUAD. (C) Arranged to oxidative phosphorylation predominance in LUAD patients. (D) Overall survival (OS) in LUAD patients stratified by expression of oxidative phosphorylation mRNA.
Comparison of patient characteristics and pathological findings according to high or low OXPHOS-related gene expression.
The expression of PPARγ, an OXPHOS-related gene, was assessed in patients with LUAD and LUSC from TCGA database (Figure 2). Among patients with LUAD (n=500), high PPARγ expression group (FPKM value; >6.2 n=109) had a significantly worse prognosis than low PPARγ expression group (n=391) (5-year OS, 34% vs. 42%, p=0.01). In patients with stage I-IIIb (n=467) LUAD, high PPARγ expression group had significant worse OS (p=0.01), compared to the low PPARγ expression group, while in patients with stage I LUAD (n=268), high PPARγ expression was also associated with poor OS, though not statistically significant (p=0.15). In contrast, LUSC patients (n=494) with high PPARγ expression (n=102) had a comparable or slightly better prognosis than those with low PPARγ expression (n=392) (Figure 2B).
Peroxisome proliferator-activated receptor γ (PPARγ) expression in lung adenocarcinoma (LUAD). (A) PPARγ mRNA expression and overall survival in LUAD. (B) PPARγ mRNA expression and survival in lung squamous cell carcinoma (LUSC). (C) Lepidic components in LUAD presented PPARγ. (D) Papillary components in LUAD presented PPARγ. (E) PPARγ presented only in peripheral area of papillary component in LUAD.
Localization of PPARγ in LUAD samples. The localization of PPARγ was examined in surgically resected human LUAD tissues. PPARγ was highly expressed in the lepidic component, a noninvasive component of LUAD, in all cases (Figure 2C). In contrast, in the invasive component, the degree of staining for PPARγ varied depending on its localization in the individual specimen. The invasive subtype that most frequently expressed PPARγ was papillary adenocarcinoma (Figure 2D). PPARγ was differentially expressed within the invasive component, showing marked expression only in the peripheral area of the papillary component (Figure 2E).
RT-qPCR and western blotting in LUAD cell lines. We assessed the expression of OXPHOS- and glycolysis- related genes in LUAD cell lines (Figure 3A). RT-qPCR revealed that A549, HTB181, and H322 cells had significantly higher PPARγ expression than H596 cells (p<0.01, Figure 3A). UCP2, an OXPHOS-related gene, was highly expressed in A549 cells. Levels of HIF-1a and LDH, two glycolysis-related genes, were elevated in A549 and HTB181 cells. Western blotting (Figure 3B) demonstrated that PPARγ protein was expressed in A549 and HTB181 cells. ATP5a (ATP synthase) of mitochondrial complex V was more strongly expressed in A549, HTB181, and H322 cells than in H596 cells. Based on these results, we defined the A549, HTB181, and H322 cell lines as OXPHOS-high-type LUADs.
Inhibition of oxidative phosphorylation in lung adenocarcinoma (LUAD) cell lines. (A) Oxidative phosphorylation-related and glycolysis-related mRNA expression in LUAD cell lines. The data are expressed as fold-change relative to H596 cells. (B) Protein expression of oxidative phosphorylation-related genes in LUAD cell lines. The samples derived from the same experiment and those blots were processed in parallel. (C) Cell proliferation assay of A549 using oligomycin treatment. (D) CyQUANT cell proliferation assay using 4 days oligomycin treatment in LUAD cell lines. *p-Value <0.05 compared to the control (DMSO). (E) CyQUANT cell proliferation assay using metformin treatment in LUAD cell lines.
OXPHOS inhibitors for LUAD cell lines. Figure 3C shows the growth curves of A549 cells treated with oligomycin, an OXPHOS inhibitor targeting mitochondrial respiration complex V. Inhibition of cell proliferation was detected on day 4 after treatment. Specifically, after 4 days of treatment, the oligomycin-treated (1.0 μM) cells showed a 23.3-fold increase of cell proliferation, compared to a 32.7-fold change in the control (DMSO-treated) cells (p<0.01). Oligomycin significantly inhibited cell proliferation in the four high-OXPHOS LUAD cell lines at day4 (ratio of 1.0 μM oligomycin to control: 0.72 in A549 cells; 0.69 in HTB181 cells; 0.77 in H322 cells) (Figure 3D). However, oligomycin had no effects on H596 cells. In contrast, metformin, an OXPHOS inhibitor that targets complex I, did not inhibit proliferation in any of the LUAD cells at day4 (Figure 3E).
Discussion
In recent years, OXPHOS has been reportedly activated in several highly metastatic tumors, in part because tumors in metastatic foci are exposed to an aerobic environment (12, 13). In this study, we identified a subtype of LUAD patient samples characterized by high expression of OXPHOS-related genes, which is associated with worse prognoses, using the TCGA dataset. This OXPHOS-high expression type LUAD did not always interfere with glycolysis, which indicates that some high metabolic LUADs use both OXPHOS and glycolysis, which is consistent with previous reports describing melanoma (22). In melanoma, the aerobic environment during metastasis activates OXPHOS metabolism (22). Our study provides evidence to support this hypothesis in LUAD, as we found that OXPHOS-high expression type LUAD is associated with lymph node metastasis.
PPARγ, a representative OXPHOS-related molecule, has many biological functions that regulate mitochondrial turnover and energy metabolism (23). In TCGA database, patients with high PPARγ expression LUAD had a significantly poorer prognosis, similar to the prognostic outcome associated with the expression of OXPHOS-related genes. This association was not observed in patients with LUSC. Previous studies have reported that LUAD is more likely to develop distant metastasis than LUSC (24). These results indicate that the expression of OXPHOS-related genes, such as PPARγ, might be associated with poor prognosis in patients with LUAD.
Protein expression of PPARγ was also investigated in LUAD specimens to reveal its tissue localization and to assess the status of OXPHOS. The expression of PPARγ was localized only in the peripheral area of the papillary component of LUADs, suggesting that OXPHOS metabolism may be activated in invasive aerobic regions. Thus, OXPHOS may play an essential role in tumor survival in aerobic environments during tumor growth and metastasis. Therefore, we focused on inhibiting mitochondrial respiration, the main OXPHOS metabolic pathway, to control the cell progression of LUAD. Using LUAD cell lines, we found that A549, HTB181, and H322 cells express high levels of OXPHOS-related molecules. In these cell lines, we used OXPHOS inhibitors and evaluated the cell proliferation. The OXPHOS pathway generates ATP by transferring electrons to a series of transmembrane protein complexes in the inner mitochondrial membrane, a process known as the electron transport chain (ETC) (20). As electrons pass through the multiprotein ETC complexes I-IV, protons are pumped from the mitochondrial matrix to the intermembrane space by complexes I, III, and IV. OXPHOS activation induces protons from the intermembrane areas to the mitochondrial matrix via complex V, an ATP synthase, to stimulate ATP synthesis (25). Metformin, a complex I inhibitor, has been frequently evaluated in clinical trials in recent years. However, exposure to metformin did not inhibit LUAD cell proliferation in this study, suggesting that it may be ineffective as a single drug. In contrast, oligomycin, a complex V inhibitor, showed growth suppression effects in OXPHOS-high expression cell lines even at low concentrations (0.01 μM). Furthermore, no inhibitory effects were observed in OXPHOS-low expression type cells such as H596 cells. Thus, oligomycin treatment may be effective against OXPHOS high expression-type LUAD. Suganuma et al. examined the energy metabolism of leukemia cell lines using 2-deoxy-D-glucose (2-DG) as a glycolysis inhibitor and oligomycin (26). They defined THP-1 cells as an ‘OXPHOS’ leukemia cell line and found that THP-1 cells were resistant to 2-DG and sensitive to oligomycin. These previously reported findings and our results support that metabolic pathways differ according to the type of cancer and that the effect of oligomycin as an OXPHOS inhibitor depends on its dominant metabolic pathway for cell growth.
Study limitations. In terms of histopathological diagnosis, we focused on only PPARγ expression. PPARγ, a representative molecule of OXPHOS, was expressed in the peripheral area, suggesting an aerobic environment in LUAD. This finding supports the hypothesis that some LUADs utilize OXPHOS metabolism. Our study presents significant results on the metabolic type and heterogeneity of LUAD, suggesting that LUAD may possess different metabolic types even within the same tumor. Second, the toxicity of OXPHOS inhibitors in normal cells was not evaluated because of the inability to culture normal lung alveolar cells in vitro. The toxicity of OXPHOS inhibitors should be evaluated in future studies in murine models. However, this study indicated that oligomycin could be a potential treatment agent with low-toxicity, as oligomycin suppressed the proliferation of OXPHOS-high expression-type LUAD at very low concentrations, which do not appear to affect normal lung epithelial cells.
Conclusion
Our study demonstrated that OXPHOS is active in LUADs and correlates with nodal metastasis and poor prognosis. We observed elevated PPARγ expression in the peripheral regions of LUAD specimens, suggesting the reversibility of the Warburg effect and highlighting the metabolic heterogeneity of OXPHOS and glycolysis within the tumor. Inhibition of the mitochondrial respiration suppressed cell growth in LUAD cell lines expressing high levels of OXPHOS-related genes and proteins. These results suggest that OXPHOS genes, including PPARγ, may serve as a potential marker for high OXPHOS activity in LUADs. Moreover, OXPHOS inhibitors could be a promising therapeutic strategy. Further studies are needed to investigate the potential of PPARγ expression as a reliable marker of high OXPHOS activity in LUADs. Moreover, additional studies using in vivo models are essential to validate our current findings and explore their potential for developing targeted therapies against OXPHOS-associated cancer metabolism in LUAD. These investigations may contribute to the development of novel therapeutic approaches that specifically modulate OXPHOS metabolism in LUAD.
Acknowledgements
We gratefully acknowledge the work of past and present members of our laboratory and members of Cell Innovator in Kyushu-University.
Footnotes
Authors’ Contributions
MF, TM, and MO conceived the study. MF and TM analyzed the data pertaining to LUAD cells and TCGA. MF, TM, NT, YM, and MO interpreted the study data. KK and YT performed the IHC staining of LUAD and contributed significantly to the writing of the manuscript. MF, TM, and YM contributed to the final manuscript. All Authors read and approved the final manuscript.
Availability of Data and Materials
The TCGA data that support the findings of this study are openly available at [https://tcga-data.nci.nih.gov/tcga/]. The non-public data analyzed during the current study are available from the corresponding author on reasonable request.
Additional Files
Additional file 1 can be accessed at the following URL: https://zenodo.org/records/13997914
Conflicts of Interest
Not applicable.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
- Received May 14, 2024.
- Revision received October 30, 2024.
- Accepted November 4, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.