Relationship between mitochondrial DNA copy number and Lung cancer: a bidirectional Mendelian randomization study

doi:10.21203/rs.3.rs-4848222/v1

Download PDF

Article

Relationship between mitochondrial DNA copy number and Lung cancer: a bidirectional Mendelian randomization study

https://doi.org/10.21203/rs.3.rs-4848222/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Mitochondrial DNA copy number (mtDNA-CN) play a pivotal role in Lung cancer (LC). However, it is not clear that the relationship between exposure and outcome. Therefore, our database were obtained from a genome-wide association study (GWAS) summary statistics. By the univariate and multivariate bidirectional Mendelian randomization (MR), the results indicate that forward MR analysis showing that mtDNA-CN increased the risk of LC (IVW OR = 1.32, 95% CI: 1.08–1.61, P = 0.006; BW OR = 1.37, 95% CI: 1.12–1.68, P = 0.002). The results are consistent with subgroup. However, there was no statistical significance in reverse MR analysis. The mtDNA-CN was still at risk for LC after multivariate MR adjustment for confounders (IVW OR = 1.40, 95%CI: 1.13–1.75, P = 0.003; FDR adjust p = 0.005). Subgroup analysis suggest the same conclusion in multivariate MR Analysis. Sensitivity analyses found no evidence of horizontal pleiotropy and heterogeneity. In this study, it was found that mtDNA-CN can predict the risk of LC, especially in smokers. However, we did not find a reverse causal relationship. This imply that mtDNA-CN can be used as a biomarker in the clinical practice of LC.

Biological sciences/Cancer/Cancer metabolism

Biological sciences/Cancer/Cancer microenvironment

Biological sciences/Cancer/Lung cancer

Mitochondrial DNA copy number

Lung cancer

bidirectional univariate Mendelian randomization

multivariate Mendelian randomization

Bayesian weighting

Lung cancer (LC) remains the cancer with the highest incidence and mortality rates in the world, accounting for 12.4% and 18.7% of the cases, respectively, according to Global Cancer Statistics 2022¹. It was reported that 3.5 million people, approximately, die from LC, which is 2.5 times larger than colorectal cancer in the United States in 2023². Smoking has an incubation period of about 17 years for the development and the main cause of LC³. Therefore, smoking cessation is still the most effective way to reduce cancer incidence and mortality. Identifying novel biomarkers and understanding the molecular mechanisms underlying LC development are essential for early diagnosis and treatment^4,5. With the continuous research on LC, we urgently need to find new markers to improve the diagnosis rate of LC.

The unique metabolic pattern is a common feature of cancer, and mitochondria have many connections with cancer in biosynthesis, signal transduction, cell differentiation, apoptosis, cell cycle maintenance and cell growth control⁶. Mitochondrial DNA is extremely susceptible to mutation, resulting in changes in the mitochondrial genome due to reactive oxygen species(ROS)⁷. The mtDNA-CN can be a better indicator of mitochondrial genome quantity change⁸. The increase of mtDNA-CN not only represents an enhancement in mitochondrial activity, but also elevate the risk of cancer⁹. A prospective cohort study (227 cases) showed that mtDNA-CN in the blood increased the risk of LC¹⁰. Meng et al. came to the same conclusion from human white blood cells¹¹. However, other studies have shown that there is no positive relationship between mtDNA-CN and LC^12,13. Therefore, the relationship between mtDNA-CN and LC remains unclear. We need to further explore the relationship between them.

With the development of genomics and bioinformatics technology, MR provides new tools and methods for epidemiological research. This method can eliminate the influence of confounders, reverse causality and selection bias, thus improving the accuracy of causal inference^14,15. This study aims to explore the relationship between mtDNA-CN and LC by MR analysis, in order to gain new insights into the diagnosis and treatment of LC.

The forward MR results indicate that mtDNA-CN can increase the risk of LC, especially in smokers. As shown in Fig. 1, 49 SNPs were selected as the main research subjects, among which 41 SNPs were selected from the smoking population for subgroup analysis, and their F-statistics were all greater than 10. The results were almost unaffected by instrumental variable bias. The P values of Cochran's Q Test of inverse variance weighted (IVW) and MR-Egger were both greater than 0.05, indicating no heterogeneity. There is same conclusion in IVW analysis (OR = 1.32,95% CI: 1.08–1.61, P = 0.006) and Bayesian weighting (BW) analysis (OR = 1.37, 95% CI: 1.12–1.68, P = 0.002), and the MR-Egger and Weighted Median (WM) obtained the direction and amplitude similar to the IVW method. Thus, the greater mtDNA-CN predicted by genes, the higher the risk of LC. Subgroup analysis showed the same results after removing former non-smokers. IVW analysis (OR = 1.40, 95% CI: 1.10–1.79, P = 0.007) and BW analysis (OR = 1.40, 95% CI: 1.10–1.79, P = 0.007) also had the same result, and the MR-Egger and WM obtained the same direction and amplitude as the IVW. Note that, the value of OR increased from 1.32 to 1.40, indicating an increased risk of LC in smokers with increased mtDNA-CN after removing non-smokers.

Reverse MR Analysis (Fig. 2) showed no evidence that lung cancer increased the risk of mitochondrial DNA copy number. 10 SNPs were used as the main research subjects, and 6 SNPs were taken as subgroup analysis, with statistical values F greater than 10. The main analysis (IVW OR = 1.00, P = 0.607) and subgroup analysis (IVW OR = 1.02, P = 0.074) both yielded similar results. Both the P-values of the MR-Egger intercept and MR-PRESSO global tests were greater than 0.05, indicating that there was no horizontal pleiotropy in this study. Scatter plots (S1) show a positive correlation effect. Funnel plot (S2), indicating no horizontal pleiotropy. The leave-one-out (S3) found no abnormal SNPS that affected the outcome.

MVMR (Fig. 3) showed that mtDNA-CN increased the risk of LC (IVW OR = 1.40, 95%CI:1.13–1.75, P = 0.003, FDR P = 0.005) after adjusting for confounders such as smoking initiation, alcohol consumption and BMI. This result was also supported in subgroup analysis (IVW OR = 1.45,95%CI:1.11–1.89, P = 0.006, FDR P = 0.008). In addition, Egger, Median, Robust and Lasso also obtained similar trends and results. It is worth noting that the risk ratio of mtDNA-CN for LC increased from 1.40 to 1.45, which is similar to the trend of univariate Mendelian randomization after excluding non-smokers in the subgroup analysis.

This study explored the causal analysis between mtDNA-CN and the risk of LC through MR. The results show that mtDNA-CN can increase the risk of LC, and the mtDNA-CN can better reflect the risk of LC especially in smoking population. Our results are consistent with Hosgood, HR¹⁰ and Meng, S¹¹, and there is no reverse causal relationship.

The mtDNA-CN represents the energy states of different cells. Each mitochondrial genome maintains its independence, and the mtDNA-CN is strictly controlled, keeping it relatively stable to maintain homeostasis¹⁶. The mtDNA-CN vary in different tissue cells, with energy-intensive tissues such as heart and skeletal muscle containing 4,000 to 6,000 copies per cell, while liver, kidney and lung tissues contain an average of 500 to 2,000 copies¹⁶. With the increase of age, mitochondrial DNA copy number shows a decreasing trend¹⁷. Increased mtDNA-CN is generally associated with enhanced mitochondrial O₂ consumption¹⁸. The mtDNA-CN was depleted when ATP synthase was knocked out. Therefore, mtDNA-CN indirectly reflects the level of mitochondrial DNA synthesis and energy metabolism¹⁹. The previously reported "Warburg effect", in which tumor cells use anaerobic glycolysis to provide energy, does not explain the success of oxidative respiratory chain inhibitors in restraining tumor cell growth, so mitochondria play a crucial role in tumor cells^20,21. Cancer cells have highly complex metabolic pathways. Although the "Warburg effect" has played a crucial role in tumor metabolism, the intermediate products produced by mitochondria in Tricarboxylic Acid cycle are often an indispensable part of tumor cell survival²⁰. Thus, mtDNA-CN can be used to monitor the metabolic level of tumor cells.

Compared to nuclear genes, mitochondrial DNA is more susceptible to oxidative stress damage, and mtDNA CN compensates for oxidative stress damage. Due to the lack of intron and histone protection in mitochondria, as well as damage repair mechanisms, mitochondrial DNA is susceptible to endogenous or exogenous stress responses²². In smoking patients, cells are often damaged by a large amount of reactive nitrogen and oxygen species (RNOS) products, causing damage to mitochondrial DNA²³. On the one hand, mitochondria resist various stimuli by increasing mtDNA CN changes while maintaining their own stability²². On the other hand, the production of a large amount of ROS within cells can cause Ferroptosis. Excessive oxidative metabolites react with divalent iron to produce Fenton reaction, causing damage to mitochondrial membrane ²⁴and increasing mtDNA CN.

Mitochondria are prone to DNA mutations during energy metabolism. The oxidative metabolites produced by mitochondria induce mutations in mitochondrial DNA, leading to impaired electron transport chain function and the incorporation of defective subunits into the oxidative respiratory chain, resulting in a “vicious cycle” of further amplify in ROS production⁷. Kaori Ishikawa et al²⁵. transferred the mutated NADH dehydrogenase subunit 6 (ND6) gene into tumor cells with low metastasis and obtained high metastasis ability. This mutation reduces the activity of Respiratory Complex I and produces excess ROS²⁵. Therefore, mitochondrial DNA mutations can alter the physiological function of cancer cells through the production of ROS. Mutations are commonly present in mitochondria, and when mitochondrial DNA mutations reach a certain threshold²⁶, they often cause severe damage to the oxidative respiratory chain. In order to meet the high oxidative metabolism requirements of cancer cells, mitochondria can only correct the mitochondrial DNA mutation effect by increasing their mtDNA CN²⁷.

The causal association with LC was investigated from the point of view of serum markers in MR Analysis. Multivariate and reverse MR Analyses greatly reduce confounding and reverse causal bias. These analyses produced reliable results and further validated the science of previous studies. However, the study groups were all European, which limits the general applicability of the findings to other ethnic groups.

In short, mitochondria play a very important role in the field of tumor therapy. The mtDNA CN can be used as an easy to measure and capture indicator of mitochondrial function, and more and more studies have used this indicator to establish the association with diseases^8,13. The results of this study suggest that mtDNA CN plays a vital role in the development of LC and can be used as a biomarker for the diagnosis of LC. The results indicate that mtDNA CN can reflect the risk of LC, especially in smokers.

4.1. Study design

This MR Study are divided into two parts. The first one used bi-UVMR to assess the causal effect of mtDNA-CN and LC. We must follow three basic assumptions that (1) should be strongly correlated with exposure, (2) should be independent of confounders, and (3) should be independent of other factors that affect the outcome in MR Analysis to select the appropriate SNPs.

The second one, MVMR analysis was used to adjust the influence of relevant confounders on the outcome, and the independent effect of mtDNA-CN on LC was further analyzed. STRBOE-MR (strengthening the reporting if observational studies in epidemiology using Mendelian randomization) to ensure the integrity of this observational study²⁸.

4.2. Data sources

Chong et al. provides GWAS data on mitochondrial DNA copy numbers using an innovative automated mitochondrial replication analysis method (AutoMitoC). The study involved 395,781 participants from the UK Biobank for genome-wide and exome association analysis from blood samples and represents the most comprehensive dataset ever published²⁹.

The LC GWAS data comes from the TRICL project, which provides a cohort study covering a European population. These data have been incorporated into the IEU Open GWAS project, which comprises two distinct datasets. The dataset labeled as ieu-a-987, as the main analysis, covering 29,863 cases and 55,586 control samples. The data set coded ieu-a-985 was analyzed as a subgroup, including 23,848 cases and 16,605 controls, excluding the non-smoking population.

We obtained data related to smoking and alcohol consumption from a large GWAS conducted by Liu, M, et al., in 2019. The study included samples from up to 1.2 million European populations and identified 566 genetic variants across 406 loci³⁰.

These data are also included in the IEU Open GWAS project, data sets ieu-b-4877 (smoking initiation) and ieu-b-73 (weekly alcoholic beverage consumption). In addition, we obtained a dataset on Body mass index (BMI) from the IEU Open GWAS project, which contains a sample of 461,460 European populations, dataset number ukb-b-19953. All databases can be IEU Open GWAS official website(https://gwas.mrcieu.ac.uk/) in Table 1.

Table 1

Description of genome wide association studies used for each phenotype.
Trait	Dataset	Population	N case	N control	Sample size	Author	Consortium	PMID
Lung cancer	ieu-a-987	European	29,863	55,586	85,449	NA	TRICL	NA
Mitochondrial DNA copy number	ebi-a-GCST90026371	European	NA	NA	395,718	Chong M	NA	35023831
Subgroup
Lung cancer(ever smokers)	ieu-a-985	European	23,848	16,605	40,453	NA	TRICL	NA
Confounders
Alcoholic drinks per week	ieu-b-73	European	NA	NA	335,394	Liu M	GSCAN	30643251
Smoking initiation	ieu-b-4877	European	311,629	NA	607,291	Liu M	GSCAN	30643251
Body mass index (BMI)	ukb-b-19953	European	NA	NA	461,460	Ben Elsworth	MRC-IEU	NA

4.3 Instrumental variable selection

We apply strict criteria (P < 5×10^− 8) to identify single nucleotide polymorphisms (SNPs) associated with exposure, based on specific assumptions in bi-UVMR studies. We established a Linkage Disequilibrium (LD) threshold (r² < 0.001, within a 10,000 kb range), effectively eliminating SNPs with strong LD to ensure the independence of the selected SNPs. In addition, we searched through the LDlink website (https://ldlink.nih.gov/?tab=ldtrait) to further exclude SNPs associated with confounders and outcomes, such as those directly related to smoking, alcohol consumption, and cancer. We calculated the F-statistic to evaluate the strength of the selected instrumental variables (IVs), where an F-value greater than 10 indicates a significant correlation with mtDNA-CN. The formula for calculating the F-statistic is F = β ²/se².

The remaining SNPs that were screened were used as IVs to explore the potential causal relationship between mtDNA-CN and LC. Further, we adopted the MVMR method ³¹ and included BMI³², alcohol consumption³³, smoking initiation, and mtDNA-CN as exposure factors. SNPs of exposure factors were also strictly adopted with a P-value less than 5 ×10^− 8 and r² < 0.001, thresholds within the range of 10,000 kb. And adjusted their independent SNPs effects to delve deeper into the complex causal relationship between these factors and LC.

4.4 Statistical analysis

In bi-UVMR(Fig. 4), we mainly used IVW, WM³⁴, MR Egger³⁵, and BW³⁶. IVW and BW are considered the main methods³⁷, and if these four methods need to be consistent in direction, they are considered to produce consistent results in MR analysis. In the sensitivity analysis, IVW and MR-Egger are both used Cochran's Q test to quantify the heterogeneity among different IVs. When p < 0.05, it indicates the presence of heterogeneity, random effects model is applied. The MR-Egger intercept test uses the intercept term to assess pleiotropy, ensuring that exposure and outcome are independently associated with genetic variants that are calculated. In addition, MR-PRESSO global test is used to detect pleiotropy, and significant pleiotropy is considered when P < 0.05³⁸. Leave-one-out analysis to assess the stability of the results. This approach involves excluding each SNP in turn to assess its potential impact on the overall results, ensuring that our results are not biased by the removal of a single variable. At the same time, the scatter plot and the funnel plot can observe whether there are outliers and outliers. The forest map can visually represent the size of effect values between studies.

In MVMR (Fig. 5), the relevant confounders were carefully adjusted. IVW as the main method of analysis¹⁴. In order to ensure the robustness of the research results, Egger, Lasso, Median and Robust are also adopted for comprehensive analysis³⁹. Moreover, all results were corrected for FDR using the Benjamini-Hochberg method to control for the possibility of false positive results caused by multiple comparisons and enhance the credibility of the research conclusions.

All tests were conducted using bilateral validation and were carried out using the "TwoSampleMR", "MendelianRandomization", and "robustbase" packages in R software 4.4.0.

4.5 Ethical Analysis

This study utilized publicly available data. Each individual study in GWAS is approved by the corresponding institutional review committee, and participants or their authorized representatives provide informed consent.

Acknowledgements

We thank TRICL and GSCAN agencies as well as Chong M and Ben Elsworth for their GWAS data. We also thank the IEU OPEN GWAS and LDlink websites for providing data retrieval services.

Author contributions

Data curation, Weining Qi; Methodology, Shihao Tang; Software, Zhuo Xu; Supervision, Liqiang Liu;Validation, Tao Wang; Writing – original draft, Yangyang Qi; Writing – review & editing, Tao Jiang.

Data availability

The datasets presented in this study can be found in IEU Open GWAS official website(https://gwas.mrcieu.ac.uk/).

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Bray, F., et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA. Cancer J. Clin. 74, 229–263 (2024).
Siegel, R. L., Miller, K. D., Wagle, N. S. & Jemal, A. Cancer statistics, 2023. CA. Cancer J. Clin. 73, 17–48 (2023).
Mirahmadizadeh, A., Hassanzadeh, J., Moradi, A. M., Gheibi, Z. & Heiran, A. Projection of the prevalence of tracheal, bronchus, and lung cancer incidence using cigarette smoking prevalence in Iran from 1990 to 2018: a comparison of latent period-based models with standard forecasting models. BMC Public Health. 24, 1896 (2024).
Li, Y., Yan, B. & He, S. Advances and challenges in the treatment of lung cancer. Biomed. Pharmacother. 169, 115891 (2023).
Nooreldeen, R. & Bach, H. Current and Future Development in Lung Cancer Diagnosis. Int. J. Mol. Sci. 22, 8661 (2021).
Zong, W. X., Rabinowitz, J. D. & White, E. Mitochondria and Cancer. Mol. Cell. 61, 667–676 (2016).
Alexeyev, M. F. Is there more to aging than mitochondrial DNA and reactive oxygen species? FEBS. J. 276, 5768–5787 (2009).
Samuels, D. C., et al. Evaluation of mitochondrial DNA copy number estimation techniques. Plos One. 15, e0228166 (2020).
Hu, L., Yao, X. & Shen, Y. Altered mitochondrial DNA copy number contributes to human cancer risk: evidence from an updated meta-analysis. Sci. Rep. 6, 35859 (2016).
Hosgood, H. D., et al. Mitochondrial DNA copy number and lung cancer risk in a prospective cohort study. Carcinogenesis. 31, 847–849 (2010).
Meng, S., et al. Pre-diagnostic leukocyte mitochondrial DNA copy number and risk of lung cancer. Oncotarget. 7, 27307–27312 (2016).
Kazdal, D., et al. Prevalence of somatic mitochondrial mutations and spatial distribution of mitochondria in non-small cell lung cancer. Br. J. Cancer. 117, 220–226 (2017).
Chen, J., et al. Clinical application of plasma mitochondrial DNA content in patients with lung cancer. Oncol. Lett. 6, 7074–7081 (2018).
Yavorska, O. O. & Burgess, S. MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data. Int. J. Epidemiol. 46, 1734–1739 (2017).
Zhang, H. Pros and cons of Mendelian randomization. Fertil. Steril. 119, 913–916 (2023).
D'Erchia, A. M., et al. Tissue-specific mtDNA abundance from exome data and its correlation with mitochondrial transcription, mass and respiratory activity. Mitochondrion. 20, 13–21 (2015).
Mengel-From, J., et al. Mitochondrial DNA copy number in peripheral blood cells declines with age and is associated with general health among elderly. Hum. Genet. 133, 1149–1159 (2014).
Bentlage, H. A. & Attardi, G. Relationship of genotype to phenotype in fibroblast-derived transmitochondrial cell lines carrying the 3243 mutation associated with the MELAS encephalomyopathy: shift towards mutant genotype and role of mtDNA copy number. Hum. Mol. Genet. 5, 197–205 (1996).
Fukuoh, A., et al. Screen for mitochondrial DNA copy number maintenance genes reveals essential role for ATP synthase. Mol. Syst. Biol. 10, 734 (2014).
Vasan, K., Werner, M. & Chandel, N. S. Mitochondrial Metabolism as a Target for Cancer Therapy. Cell Metab.. 32, 341–352 (2020).
Hensley, Christopher T., et al. Metabolic Heterogeneity in Human Lung Tumors. Cell. 164, 681–694 (2016).
Lee, H. C., Yin, P. H., Lu, C. Y., Chi, C. W. & Wei, Y. H. Increase of mitochondria and mitochondrial DNA in response to oxidative stress in human cells. Biochem. J. 348, 425–432 (2000).
Durham, A. L. & Adcock, I. M. The relationship between COPD and lung cancer. Lung Cancer. 90, 121–127 (2015).
Dixon, Scott J., et al. Ferroptosis: An Iron-Dependent Form of Nonapoptotic Cell Death. Cell. 149, 1060–1072 (2012).
Ishikawa, K., et al. ROS-Generating Mitochondrial DNA Mutations Can Regulate Tumor Cell Metastasis. Science. 320, 661–664 (2008).
Filograna, R., et al. Modulation of mtDNA copy number ameliorates the pathological consequences of a heteroplasmic mtDNA mutation in the mouse. Sci. Adv. 5, eaav9824 (2019).
Filograna, R., Mennuni, M., Alsina, D. & Larsson, N. G. Mitochondrial DNA copy number in human disease: the more the better? FEBS Lett. 595, 976–1002 (2020).
Skrivankova, V. W., et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. Jama. 326, 1614–1621 (2021).
Chong, M., et al. GWAS and ExWAS of blood mitochondrial DNA copy number identifies 71 loci and highlights a potential causal role in dementia. eLife. 11, e70382 (2022).
Liu, M., et al. Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nat. Genet. 51, 237–244 (2019).
Sanderson, E., Davey Smith, G., Windmeijer, F. & Bowden, J. An examination of multivariable Mendelian randomization in the single-sample and two-sample summary data settings. Int. J. Epidemiol. 48, 713–727 (2019).
Hidayat, K., Du, X., Chen, G., Shi, M. & Shi, B. Abdominal Obesity and Lung Cancer Risk: Systematic Review and Meta-Analysis of Prospective Studies. Nutrients. 8, 810 (2016).
García Lavandeira, J. A., et al. Alcohol consumption and lung cancer risk in never smokers: a pooled analysis of case-control studies. Eur. J. Public Health. 28, 521–527 (2018).
Bowden, J., Davey Smith, G., Haycock, P. C. & Burgess, S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet. Epidemiol. 40, 304–314 (2016).
Burgess, S. & Thompson, S. G. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur. J. Epidemiol. 32, 377–389 (2017).
Zhao, J., et al. Bayesian weighted Mendelian randomization for causal inference based on summary statistics. Bioinformatics. 36, 1501–1508 (2020).
Burgess, S., Butterworth, A. & Thompson, S. G. Mendelian Randomization Analysis With Multiple Genetic Variants Using Summarized Data. Genet. Epidemiol. 37, 658–665 (2013).
Burgess, S., Bowden, J., Fall, T., Ingelsson, E. & Thompson, S. G. Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants. Epidemiology. 28, 30–42 (2017).
Grant, A. J. & Burgess, S. Pleiotropy robust methods for multivariable Mendelian randomization. Stat. Med. 40, 5813–5830 (2021).

No competing interests reported.

Download PDF

Editor invited by journal
08 Aug, 2024
Submission checks completed at journal
06 Aug, 2024
First submitted to journal
02 Aug, 2024

You are reading this latest preprint version

Relationship between mitochondrial DNA copy number and Lung cancer: a bidirectional Mendelian randomization study

Status:

Version 1

Abstract

Figures

1 Introduction

2 Results

3 Discussion

4 Materials and Methods

4.1. Study design

4.2. Data sources

4.3 Instrumental variable selection

4.4 Statistical analysis

4.5 Ethical Analysis

Declarations

Acknowledgements

Author contributions

Data availability

Competing interests

References

Additional Declarations

Supplementary Files

Status:

Version 1