A Multi-Omics Analysis of a Mitophagy-Related Signature in Pan-Cancer

Abstract

Mitophagy, an essential process within cellular autophagy, has a critical role in regulating key cellular functions such as reproduction, metabolism, and apoptosis. Its involvement in tumor development is complex and influenced by the cellular environment. Here, we conduct a comprehensive analysis of a mitophagy-related gene signature, composed of PRKN, PINK1, MAP1LC3A, SRC, BNIP3L, BECN1, and OPTN, across various cancer types, revealing significant differential expression patterns associated with molecular subtypes, stages, and patient outcomes. Pathway analysis revealed a complex interplay between the expression of the signature and potential effects on the activity of various cancer-related pathways in pan-cancer. Immune infiltration analysis linked the mitophagy signature with certain immune cell types, particularly OPTN with immune infiltration in melanoma. Methylation patterns correlated with gene expression and immune infiltration. Mutation analysis also showed frequent alterations in PRKN (34%), OPTN (21%), PINK1 (28%), and SRC (15%), with implications for the tumor microenvironment. We also found various correlations between the expression of the mitophagy-related genes and sensitivity in different drugs, suggesting that targeting this signature could improve therapy efficacy. Overall, our findings underscore the importance of mitophagy in cancer biology and drug resistance, as well as its potential for informing treatment strategies.

Keywords: BECN1; BNIP3L; MAP1LC3; OPTN; PARK2; PINK1; PRKN; SRC; drug resistance; drug sensitivity; mitophagy; multi-omics; pan-cancer.

Figure 1

Mechanisms of mitophagy and associated mitochondrial pathways. (A) PINK1/Parkin-mediated mitophagy: Under mitochondrial stress, PINK1 accumulates on the outer mitochondrial membrane and recruits the E3 ubiquitin ligase Parkin. Parkin ubiquitinates mitochondrial outer membrane proteins such as Mfn1/2, VDAC1, and Miro1. These ubiquitinated proteins are subsequently recognized by autophagy receptors (NDP52, p62, and OPTN), which facilitate the recruitment of the autophagy machinery, leading to mitochondrial degradation. This pathway is regulated by phosphorylation events mediated by TPK1. (B) BNIP3/NIX-mediated mitophagy and cell death pathways: BNIP3 and its homolog NIX, both regulated through phosphorylation, can promote mitophagy by directly interacting with autophagy machinery. Alternatively, BNIP3 induces mitochondrial permeabilization through Bax/Bak activation, leading to the opening of the mitochondrial permeability transition pore (mPTP) and dissociation of the COX1-UCP3 complex. This can trigger apoptosis, necrosis, or pyroptosis. In pyroptosis, BNIP3 activation is linked to Caspase-3/GSDME activity. (C) FUNDC1-mediated mitophagy: FUNDC1, a mitochondrial outer membrane protein, undergoes phosphorylation-dependent regulation to mediate mitophagy by interacting with autophagic components, enabling the selective degradation of damaged mitochondria.

Figure 2

Differential expression of the mitophagy-related signature in pan-cancer. (a) The bubble plot illustrates the log₂ fold change (FC) in the expression of the mitophagy-related genes across a spectrum of cancer types, with significance denoted by the false discovery rate (FDR) values, being represented by the color and size of the bubbles. Blue indicates downregulation, while red indicates upregulation of each gene in the tumor versus the normal tissues. The size of the circles is associated with FDR significance. Notably, PRKN and PINK1 show substantial upregulation in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC), while SRC and OPTN are significantly elevated in breast cancer (BRCA) and colon adenocarcinoma (COAD), respectively. BECN1 was significantly downregulated in KIRC, in contrast to BNIP3L, which was upregulated in the tumor. (b) Boxplots for selected cancer types, showcasing significant differences in gene expression between normal and tumor tissues. (c) Subtype-specific expression differences, emphasizing that certain genes exhibit distinct expression patterns within cancer subtypes, such as SRC in kidney renal clear cell carcinoma (KIRC) and OPTN in BRCA. Red large circles represent deregulation in mRNA expression that is statistically significant. (d) The boxplots depict expression differences (log₂ RSEM) across molecular subtypes in different cancer types (PINK1 in LUSC, OPTN in BRCA, SRC in KIRC, MAP1LC3A in GBM, BNIP3L and BECN1 in BRCA, and PRKN in LUAD). (e) The heatmap shows a general trend of stable or decreased (in some cases like SRC in BLCA or PINK1 and OPTN in ACC) expression for the mitophagy-related signature with advancing tumor stages. (f) The Kaplan–Meier curves show the survival rates in different types of cancer, according to gene expression. Higher expression of PRKN in LUAD and of PINK1 in LUSC is associated with poorer prognosis, thereby underscoring the clinical relevance of these mitophagy-related genes in cancer progression and patient survival.

Figure 3

Pathway activity associated with differential expression of mitophagy-related genes. (a) Heatmap depicting the potential activation (A) or inhibitory (I) effects of the mRNA levels of the mitophagy-related gene signature on the activity of 10 cancer-related pathways in pan-cancer. The color scale indicates the percentage of pathway activation (red) or inhibition (blue). The percentages represent the frequency of gene association with pathway regulation in various types of cancer. (b) The boxplots compare the pathway activity scores (PAS) between high and low expression groups of PRKN and OPTN in BRCA. The FDR values indicate the significance of the differences in PAS found between the high and low expression groups.

Figure 4

Correlation between mitophagy-related gene expression and immune cell infiltration in pan-cancer. (a) The heatmap represents the Spearman’s correlation between the infiltration of 24 immune cell types evaluated through ImmuCellAI, and the expression of the mitophagy-related gene signature across different cancer types. The color scale indicates the strength and direction of Spearman’s correlation (blue for negative, red for positive). #, Spearman’s rho <−0.4 or >0.4; *, p < 0.01. (b) Correlation between the expression of the mitophagy-related signature and immune cell infiltrates in skin melanoma (SKCM). The size of the dots represents the significance (−log₁₀FDR values), while the color indicates the correlation coefficient (blue for negative, red for positive). (c) Scatter plots depicting the Spearman’s correlation between OPTN expression and specific immune cell infiltrates (Th1, neutrophils, monocytes, and central memory infiltrates) in SKCM, with trend lines and correlation coefficients. OPTN expression is positively correlated with the infiltration of Th1 cells and the central memory infiltrate score, and negatively correlated with the infiltration of neutrophils and monocytes in SKCM.

Figure 5

DNA methylation analysis of mitophagy-related genes in pan-cancer and its correlation with their mRNA expression. (a) The dot plot shows differential DNA methylation levels (tumor vs. normal) for mitophagy-related genes (SRC, OPTN, PINK1, PRKN, MAP1LC3A, BECN1, and BNIP3L) across various cancer types. The color scale represents the methylation difference (tumor–normal), and the size of the dots indicates the significance (FDR values). (b) The dot plot illustrates the correlation between DNA methylation levels and mRNA expression for the same set of genes across multiple cancer types. The color scale represents Spearman’s correlation coefficient, with the size of the dots indicating the significance (FDR values). (c) Integrative genomic analysis displaying the association between PINK1 and MAP1LC3A methylation, expression and CNV, and various clinical and demographic features in KIRC. The top panel summarizes patient data [age at diagnosis, hemoglobin levels, histological type, tumor recurrence, smoking history, gender, tumor stage, sample type, and overall survival (OS)]. The middle panel shows the distribution of PINK1 expression levels across different copy number alterations. Sections marked with ‘−2’ indicate homozygous deletions, while ‘+1’ or ‘+2’ would indicate low-level or high-level amplifications, respectively. Higher or lower expression levels of PINK1 and MAP1LC3A are indicated, showing how these levels align with methylation patterns and CNVs. The bottom panel presents a detailed view of the PINK1 (left) and MAP1LC3A (right) regions, showing the relationship between DNA methylation sites, CpG islands, gene structure, and expression levels. The color coding of the CpG sites (vertical lines) likely indicates their methylation status, with different shades representing varying levels of methylation.

Figure 6

Mutation landscape of PRKN, OPTN, PINK1, SRC, MAP1LC3A, BECN1, and BNIP3L across various cancer types. (a) The waterfall plot presents the tumor mutation burden (TMB) per cancer sample and the mutation distribution of the mitophagy-related gene signature across 400 cancer samples in the TCGA. Each row represents a gene, and each column represents a cancer sample. Different colors indicate different types of mutations, including missense mutations, nonsense mutations, frame-shift deletions, frame-shift insertions, in-frame deletions, splice-site mutations, and multi-hit mutations. The bar plot on the right shows the percentage of samples with alterations in each gene. The bottom annotation panel indicates the cancer type for each sample. (b) Summary of variant classifications and types. Bar plots show the number of different mutation types (missense, nonsense, etc.) and variant types (SNP, insertion, deletion) for the five genes. The SNV class distribution is displayed, highlighting the most frequent base substitutions (184 C>T and 138 C>A). The bottom left plot shows the distribution of variants per sample, with a median of 1 variant per sample. The bottom right plot shows the variant classification summary for each gene, with the percentage of samples altered. (c) The heatmap shows the mutation frequency of the mitophagy-related genes across different cancer types. The numbers in each cell represent the percentage of samples with mutations in the respective gene and cancer type. The intensity of the color corresponds to the mutation frequency, with a darker red indicating a higher frequency.

Figure 7

Copy number variations in PRKN, OPTN, PINK1, SRC, BNIP3L, BECN1, and MAP1LC3A across different cancer types. (a) The pie charts show the frequency and types of CNVs in PRKN, OPTN, PINK1, SRC, BNIP3L, BECN1, and MAP1LC3A across various cancer types. Each pie chart represents a cancer type, with different colors indicating the type of CNV. (b) The dot plots illustrate the distribution of heterozygous amplifications (left) and heterozygous deletions (right) in PRKN, OPTN, PINK1, SRC, BNIP3L, BECN1, and MAP1LC3A across different cancer types. The size of the dots corresponds to the percentage of samples with the respective CNV type. (c) The dot plots show the distribution of homozygous amplifications (left) and homozygous deletions (right) in PRKN, OPTN, PINK1, SRC, BNIP3L, BECN1, and MAP1LC3A across various cancer types. The size of the dots represents the percentage of samples with the respective CNV type.

Figure 8

Correlation of MAP1LC3A, OPTN, SRC, PINK1, BNIP3L, BECN1, and PRKN expression with drug sensitivity (IC50) in pan-cancer, using the GDSC (a) and CTRP (b) drug databases. The color, from red to blue, depicts the correlation between each gene’s mRNA expression and IC50. Also, the bubble size represents the false discovery rate (FDR), with larger circles indicating stronger statistical significance. The color gradient indicates the direction and magnitude of correlation. Blue color, negative correlation; red color, positive correlation. Significant correlations (FDR < 0.05) are emphasized with bold outlines, highlighting the most critical interactions. (c) Regulator prioritization of the mitophagy-related gene signature. Each column is a data cohort. Genes are ranked based on their average score with multiple cohorts. The colors correspond to different score values, ranging from −3 (blue) to +3 (red).