Skip to main content
NCBI home page
iScience logoLink to iScience
. 2020 Jun 7;23(6):101247. doi: 10.1016/j.isci.2020.101247

Mitoribosome Defect in Hepatocellular Carcinoma Promotes an Aggressive Phenotype with Suppressed Immune Reaction

So Mee Kwon 1,2,3,5,∗, Young-Kyoung Lee 2,3, Seongki Min 2,3, Hyun Goo Woo 1,3, Hee Jung Wang 4, Gyesoon Yoon 2,3,∗∗
PMCID: PMC7306587  PMID: 32629612

Summary

Mitochondrial ribosomes (mitoribosomes), the specialized translational machinery for mitochondrial genes, exclusively encode the subunits of the oxidative phosphorylation (OXPHOS) system. Although OXPHOS dysfunctions are associated with hepatic disorders including hepatocellular carcinoma (HCC), their underlying mechanisms remain poorly elucidated. In this study, we aimed to investigate the effects of mitoribosome defects on OXPHOS and HCC progression. By generating a gene signature from HCC transcriptome data, we developed a scoring system, i.e., mitoribosome defect score (MDS), which represents the degree of mitoribosomal defects in cancers. The MDS showed close associations with the clinical outcomes of patients with HCC and with gene functions such as oxidative phosphorylation, cell-cycle activation, and epithelial-mesenchymal transition. By analyzing immune profiles, we observed that mitoribosomal defects are also associated with immunosuppression and evasion. Taken together, our results provide new insights into the roles of mitoribosome defects in HCC progression.

Subject Areas: Immunology, Cancer Systems Biology, Cancer

Graphical Abstract

graphic file with name fx1.jpg

Open in a new tab

Highlights

  • •

    A set of down-regulated MRPs in HCC cause mitoribosomal defects

  • •

    Mitoribosomal defects are linked to aggressive molecular features and poor prognosis

  • •

    Mitoribosomal defects in HCC are associated with immunosuppression and evasion

  • •

    TGF-β signaling pathway is a crucial mechanism to mediate mitoribosomal defects in HCC

Immunology; Cancer Systems Biology; Cancer

Introduction

Mitochondria are key organelles with pivotal roles in the bioenergetic and biosynthetic processes of cells. The main function of mitochondria, which are often regarded as the powerhouse of cells, is ATP production through oxidative phosphorylation (OXPHOS). Mitochondrial OXPHOS plays a central role in energy production, which is vital to maintain cellular function and physiology (Zong et al., 2016). There is accumulating evidence for impaired OXPHOS activity in many solid tumor cells (Chandra and Singh, 2011). Furthermore, OXPHOS defects have been described as one of the most common and prominent phenotypes of most cancers (Chandra and Singh, 2011, Warburg, 1956) and are often implicated in several cancer hallmarks (Hanahan and Weinberg, 2011), such as genomic instability (Martinez-Outschoorn et al., 2010), tumor-promoting inflammation (Sukumar et al., 2017, Colegio et al., 2014), and metastasis (LeBleu et al., 2014, Nunes et al., 2015). Mitochondria possess their own mitochondrial DNA (mtDNA) genome and the corresponding replication, transcription, and translation machinery required for its maintenance and expression. Human mtDNA encodes 13 hydrophobic core subunits of the OXPHOS system, as well as 2 rRNAs and 22 tRNAs. These 13 proteins are synthesized in the specialized translation machinery called mitochondrial ribosome, also known as the mitoribosome (De Silva et al., 2015). Mitoribosome is a macrostructure of dual genetic origin, formed by two mitoribosomal RNA components encoded by mtDNA and 82 mitoribosomal proteins (MRPs) encoded by nuclear DNA. Therefore, the balanced expression of rRNAs and MRPs and their well-coordinated assembly are essential for regulation of OXPHOS activity and the resultant energy production.

Recent advances in genomics and proteomics have provided information on the contribution of mitoribosome composition and function in the initiation and promotion of tumors. Many studies have shown that altered expression of several MRPs is closely associated with cancer development and progression as well as the metastasis of many cancer types (Loo et al., 2012, Pu et al., 2017, Sorensen et al., 2017, Wei et al., 2019). For example, MRPL11 was decreased in the primary tumor tissues of head and neck squamous cell carcinoma (HNSCC) and was considered a potential biomarker for HNSCC (Koc et al., 2015). In addition, the alternative role of several MRPs, such as DAP3 (Levy-Strumpf and Kimchi, 1998), MRPS30/PDCD9 (Carim et al., 1999), and MRPL41 (Chintharlapalli et al., 2005), as apoptosis-inducing factors has been reported. However, the mechanisms underlying the contribution of MRP alterations to cancer progression remain poorly understood. Moreover, these studies have a limitation on substantiating how MRP alterations affect mitochondrial functions systematically.

As a central organ in a variety of critical biological functions such as homeostasis of carbohydrate, lipid, amino acid, and protein synthesis, the liver is heavily enriched with mitochondria in terms of number and density (Degli Esposti et al., 2012), compared with other organs. In addition, most liver diseases including alcoholic liver disease, non-alcoholic fatty liver disease, non-alcoholic steatohepatitis, viral hepatitis, and HCC are characterized by mitochondrial dysfunction and are associated with accumulation of damaged mitochondria (Auger et al., 2015). Therefore, MRP alteration might play critical roles in the pathophysiology of liver diseases such as HCC. We have previously reported that reduced expression of MRPL13 is a cause of OXPHOS defects in HCC and promotes the invasiveness of OXPHOS-defective HCC cells, suggesting the potential involvement of mitoribosomal defects in liver cancer progression (Lee et al., 2017). However, it is unclear whether MRPL13 depletion is the only cause for cancerous OXPHOS defects and the associated aggressive phenotype of HCC.

In this study, we demonstrated that patients with HCC with mitoribosome defects had poor prognostic outcomes. In addition, we observed that HCCs with mitoribosome defects exhibited suppressive immune responses, which may allow tumor cells to evade immune surveillance and develop.

Result

Mitoribosomal Defect Gene Signature Is a Good Indicator of HCC Prognosis

To elucidate whether mitoribosomal defects are the only primary cause for the OXPHOS defects seen in HCC, we compared mitochondrial biogenesis by investigating mitochondrial translation, transcription, and replication between paired tumors (T) and surrounding tissues (ST) obtained from 15 patients with HCC (Table S1). Among 15 tumor tissues, 8 tissues (53%) showed lower protein levels of COX2, which is an important mtDNA-encoded core subunit of the OXPHOS complex IV, compared with their paired surrounding tissues (Figures S1A–S1E). Unexpectedly, among those 8 tissues, only one case showed decreased protein levels of mitochondrial transcription factor A (TFAM), a key mitochondrial transcription factor. However, the other cases harbored even higher TFAM levels (Figures S1A and S1C). Moreover, when we further examined the mtDNA copy number, only one case of eight tumor tissues showed a slight decrease in mtDNA levels (Figures S1A and S1D). These results imply that mitochondrial replication and its transcription activity are not the main contributors to the low expression of an mtDNA-encoded protein (Figure S1E).

As the mitoribosome, comprising 82 MRPs (Pietromonaco et al., 1991), is a huge and complex structure, it is plausible that co-expression of MRPs may play a critical role in the mitoribosome assembly and translation into functional proteins. With this concern, we examined the expression of MRPs in HCC by analyzing The Cancer Genome Atlas (TCGA) transcriptome data (n = 371, Figure 1A). Interestingly, we observed that the distribution of overall expression levels of MRPs was not different between non-tumor (NT) and primary tumor (PT) tissues (Figure 1B). However, the NT group showed a well-coordinated expression of MRPs in each sample, whereas the PT group showed more diverse expression levels of MRPs in each sample (Figures 1C and 1D). Moreover, the NT group, compared with the PT group, showed a closer correlation among the expression levels of MRPs (Figures S2A and S2B). These results imply that the NT group had intact mitoribosomal integrity, showing a well-coordinated expression of MRPs, whereas the PT group lost its mitoribosomal integrity and showed dysregulated expression of MRPs. Thus, we suggest that the dysregulated MRP expression in HCC may play important roles in HCC development or progression.

Figure 1.

Figure 1

Open in a new tab

Mitoribosomal Defect Gene Signature Is a Good Indicator of HCC Prognosis

(A) Schematic view of the analysis of MRPs to define three distinctive signatures (Up-MRPs, Dn-MRPs, and Other-MRPs).

(B) Comparison of MRPs' expression between NT and PT.

(C) Heatmap shows the expression of 82 MRPs in 371 PT and 50 NT samples from TCGA-LIHC cohort.

(D) Comparison of maximum absolute deviation (MAD) among MRPs' expression between NT and PT.

(E) (Top) Enrichment of three distinctive signatures (Up-MRPs, Dn-MRPs, and Other-MRPs) among PT. (Bottom) Heatmap shows the expression of variable MRPs (n = 38) among PT. (Right) Bar plot indicates the fold difference in the expression of variable MRPs (n = 38) between PT and NT.

(F and G) Mitoribosome defect scores (MDSs) were calculated based on the three distinctive signatures (Up-MRPs, Dn-MRPs, and Other-MRPs) and total MRPs; MDSup, MDSdn, MDSother, and MDSall, respectively. Heatmap indicates the association of MDSup and MDSdn with Cancer Hallmark gene sets. According to the direction of the association between MDSup and MDSdn, gene sets were classified into the opposite, Only-MDSup, Only-MDSdn, or common, which were associated with MDSup and MDSdn in the opposite manner, MDSup specifically, MDSdn specifically, or in the same direction, respectively. A positive or negative association is shown in red or blue, respectively. Gene set with a non-significant association (p > 0.001) is shown in the blank (F). Forest plot indicates hazard ratios for MDSs based on the Cox-regressional univariate survival analysis (G).

(H) Comparison of MDS between male (n = 250) and female (n = 121) samples from TCGA-LIHC cohort. Boxplots of MDSs of males and females are shown as first quartile, median, and third quartile (bottom box, middle line, and top box, respectively). Whiskers represent the minimum and maximum values.

See also Figures S1 and S2; Table S2 and Table S3

To evaluate whether the expression of MRPs contributes to the development and progression of HCC, we generated three unique MRP signatures, 12 up-MRPs, 6 down-MRPs, and 20 other-MRPs, by comparing the MRP expression between PT and NT (permuted Student's t test p < 0.05 and fold difference >0.3, Transparent Methods; Figures 1A and 1E; Table S2). Thereafter, by performing gene set enrichment analysis (GSEA), the expression enrichment of these signatures was measured in each of the tumor samples (Figure 1E). Interestingly, most HCC samples demonstrated higher expression of up-MRPs and lower expression of the down-MRPs. This may indicate that the expression of each of the up-MRPs and down-MRPs is well coordinated in tumor cells.

To estimate the degree of mitoribosome defects in individual samples, we then designated the enrichment scores (NES, normalized enrichment score) of the three MRP signatures (up-MRPs, down-MRPs, or other-MRPs) and transformed the NES as mitoribosome defects scores (MDS), MDSup, MDSdn, and MDSother, to reflect the direction of the up and down signatures (see Transparent Methods). When evaluating whether these MDSs were associated with cancer-related features using 50 cancer hallmark gene sets (Liberzon et al., 2015) (see Transparent Methods), we observed that most of the cancer gene sets showed closely correlated expressions with MDSdn and negatively correlated expression with MDSup (Figure 1F; Table S3). In addition, MDSdn showed significant correlations with the expression of gene sets for the unfolded protein response (r = 0.417, p = 5.083 × 10−17) and glycolysis (r = 0.308, p = 1.303 ×10−9) and negative correlations with lipid metabolism-related gene sets (i.e., fatty acid metabolism and adipogenesis) (Figure 1F; Table S3). Remarkably, MDSdn showed the strongest negative correlation with OXPHOS (r = −0.565, p = 1.020 ×10−32). This may imply that mitoribosomal defects induced by the suppression of down-MRPs are closely associated with OXPHOS dysfunction. Moreover, survival analysis based on the Cox regressional hazard model showed that only MDSdn was significantly associated with poor prognosis (p = 8.479 × 10−3) (Figure 1G). These results suggest that mitoribosomal defects invoked by the decreased expression of down-MRPs may play a critical role in HCC prognosis. In addition, these findings indicate that MDSdn is a good indicator of mitochondrial dysfunction and HCC prognosis, emphasizing that our MDS scoring system is an important new tool that successfully reflects the mitochondrial dysfunction-related cancer features in HCC. Furthermore, there was no significant difference found in the comparison of MDS between males (n = 250) and females (n = 121), implying that mitoribosomal defects may not be affected by sex differences (Figure 1H).

Identification of Molecular Features Linked to Mitoribosomal Defects

To advance the clarification of the underlying molecular features related to mitoribosomal defects in HCC, we stratified TCGA HCC samples into high- and low-mitoribosome defect (MD) subtypes, i.e., H-MD (>upper quartile MDSdn) and L-MD (<lower quartile MDSdn), respectively. GSEA analysis (MsigDB. V. 6.1) revealed that H-MD was depleted in the mitochondrial function-related gene sets, such as OXPHOS, fatty acid metabolism, and reactive oxygen species (ROS) pathway, as well as the mitochondrial structure-related gene sets, such as mitochondrial matrix, envelope, membrane, and protein complex (Table S4; Figures 2A, 2B, and S3A–S3C). In contrast, the H-MD group revealed enriched expression of cell cycle-related gene sets (i.e., mitotic spindle, PI3K/AKT/mTOR axis signaling). Other cancer-related gene sets were also enriched in the H-MD, including transforming growth factor (TGF)-β signaling (Caja et al., 2018), WNT/β-catenin signaling (Khalaf et al., 2018), nuclear factor-κΒ signaling (Luedde and Schwabe, 2011), epithelial to mesenchymal transition (EMT), inflammatory response, and interleukin (IL)-6/JAK/STAT3 axis signaling (Table S4; Figures 2C and S4A–S4C). These results indicate that mitochondrial dysfunction, as a result of mitoribosomal defects, may be critically involved in cell transformation and progression of aggressive HCC. Furthermore, in the Kaplan-Meier survival analysis, the H-MD group showed worse prognostic outcomes of overall survival compared with L-MD (Figure 2D). By comparing the gene expression between H-MD and L-MD, we identified 83 upregulated and 161 downregulated differentially expressed genes (DEGs) in H-MD (permutated Student's t test false discovery rate [FDR] <0.005 and fold change >1, Figures 2E–2G; Table S5). Among the upregulated DEGs, immune response-related genes, such as secreted phosphoprotein 1 (SPP1), integrin subunit beta 1 (ITGB1), and immunoglobulin heavy constant gamma 1 (IGHG1), as well as cell adhesion-related genes such as CD24, stratifin (SFN), and epithelial cell adhesion molecule (EPCAM), were the most highly expressed in H-MD (Figure 2E, pink-colored box). On the contrary, genes related to metabolism, such as carbamoyl-phosphate synthase 1 (CPS1), serum amyloid A1 (SAA1), cytochrome P450 family 8 subfamily B member 1 (CYP8B1), and tyrosine aminotransferase (TAT), showed the lowest expression in the H-MD group (Figure 2E, blue-colored box). Interestingly, according to the Human Protein Atlas (Uhlen et al., 2015) (https://www.proteinatlas.org/), more than half of the downregulated DEGs (82/161) are liver tissue-specific genes, including CPS1, SAA1, TAT, and HPD, indicating that mitoribosomal integrity is also essential in maintaining the liver's intrinsic functions (Table S5). Similarly, Gene Ontology (GO) analysis of DEGs demonstrated that genes involved in fundamental mitochondrial functions such as catabolic, biosynthetic process and oxidation-reduction processes were significantly downregulated, whereas those linked to immune cell infiltration, inflammatory response, and extracellular matrix (ECM) organization were highly upregulated in H-MD (Figure 2H). These results indicate the potential impact of mitoribosome defects on the tumor microenvironment concerning the immune response.

Figure 2.

Figure 2

Open in a new tab

Identification of Molecular Features Linked to Mitochondrial Defects

Based on the MDSdn, TCGA-LIHC samples were stratified into the subgroup with higher mitoribosome defect (H-MD) and one with lower mitoribosome defect (L-MD), and molecular features associated with H-MD or L-MD were compared.

(A–C) GSEA results based on the OXPHOS (A), mitochondria respiratory chain complex assembly (B), and TGF-β signaling (C) were shown. Normalized enrichment scores (NES) and FDR for each gene set are noted.

(D) Overall survival time of H-MD and L-MD was compared based on the Kaplan-Meier survival analysis.

(E) Volcano plot indicates fold change (FC) and FDR based on the permutation t test between H-MD and L-MD groups. Differentially expressed genes (DEGs) were marked with red- or blue-colored points (FC > 1 or < −1 & FDR <0.005, respectively).

(F) The expression of upregulated (n = 83) or downregulated (n = 161) DEGs in either H-MD or L-MD is shown. Samples are represented in columns, grouped by H-MD or L-MD.

(G) Enrichment plots based on the 83 upregulated genes and 161 downregulated genes are shown in the left and right panels, respectively. NES and FDR for each gene set are noted.

(H) Gene ontology (GO) analysis was performed based on the up and down DEGs. The -log10(p value) is shown in red and blue bars for up- and downregulated genes, respectively.

See also Figure S3; Table S4 and S5.

Mitoribosomal Defects Are Closely Associated with Immune Cell Response in HCC

Based on GO analysis, we interrogated the relevance of mitoribosomal defects to immune response signaling in detail. First, we employed the recently reported immune signatures (n = 112) derived from an immune-specific subclass of HCC (Sia et al., 2017). The H-MD group showed higher enrichment for the overall HCC immune signature (Sia et al., 2017) than did the L-MD group (Figure 3A). Furthermore, among the immune system genes with significantly higher expression in H-MD (n = 21, permutated Student's t test FDR <0.01 and fold change >0.5; Figure S4A), COL1A1, COL1A2, and POSTN were prominently expressed in the H-MD group (Figure S4A). Considering their function in ECM, we investigated whether or not mitoribosomal defects were associated with cancer-associated ECM (C-ECM). It was observed that the H-MD group was enriched with the C-ECM signature (Chakravarthy et al., 2018) (Figure 3B). Furthermore, the H-MD group was enriched with the activated stroma-related genes (Moffitt et al., 2015), but not with normal stroma-related genes (Moffitt et al., 2015) (Figures 3C and 3D). These results are indicative of the possibility that mitoribosome defects may selectively affect the neighboring cells to perturb the HCC microenvironment. To delineate the repertoire of neighboring cells in the microenvironment, we used the inferred proportion for 64 immune and stromal cell types based on xCell analysis (Aran et al., 2017) and compared the proportion of infiltrated immune cells between H-MD and L-MD (see Transparent Methods, and Figures 3E and 3F; Table S5). Interestingly, suppressive immune cells favorable for cancer cells including mast cells, Th2 cells, and regulatory T cells showed more infiltration (Figure 3E), whereas stimulatory immune cells related to anti-cancer immunity such as CD8+ naive T cells, NK T cells (NKT), and Th1 cells were less infiltrated in the H-MD than in the L-MD subtype (Figure 3F). These results suggest that mitoribosome defects may contribute to the creation of a more favorable microenvironment by suppressing immune responses and promoting evasion from immune surveillance. Next, we examined the link of mitoribosomal defects with immune-modulatory cytokines using 25 immune inhibitors, 28 immune stimulators, 6 major histocompatibility complex (MHC) class I, 10 MHC class II, and 3 MHC non-classified molecules (Transparent Methods, Table S7). This analysis revealed that most cytokines, except for IL6R and ICOSLG (green-colored box), were more highly expressed in the H-MD than in the L-MD subtype and were positively associated with mitoribosomal defects (Figure 3G), implying a more active immune contexture of the H-MD group.

Figure 3.

Figure 3

Open in a new tab

Mitoribosomal Defects Are Closely Associated with Immune Cell Response in HCC

(A–D) preRanked GSEA was performed based on the HCC immune-related signature (A), cancer-associated ECM signature (B), activated stroma-associated signature (C), and normal stroma-associated signature (D). NES for each gene set is compared between H-MD and L-MD.

(E and F) The proportion of immune cells with pro-cancer (Mast cells, Th2 cells, and Tregs) (E) and with anti-cancer properties (CD8+ T cells, NK-T cells, and Th1 cells) (F) are compared between H-MD and L-MD based on the xCell analysis output.

(G) The average expression values of immune-modulatory cytokines in either H-MD or L-MD group are shown in colored scale. According to their immune response-related properties, cytokines were classified as inhibitory; stimulatory; MHC classes I, II, and non-classified (NC) and marked with green-, purple-, sky blue-, navy blue-, and gray-colored bars, respectively. A significant association with MDSdn is shown as a colored scale, and non-significant association is shown as blank.

(H) Permutation t test was performed between H-MD and L-MD among the immune-modulatory cytokines. The x axis indicates the expression fold change between H-MD and L-MD, and the y axis indicates the -log10FDR for each cytokine. Cytokines significantly associated with MDSdn are depicted in color coded in (G) according to their immune-related properties.

(I) TGFB1 expression is compared between H-MD and L-MD. Boxplots are shown as first quartile, median, and third quartile (bottom box, middle line, and top box, respectively) with Welch two-sample t test p values. Whiskers represent the minimum and maximum values.

See also Figure S4 and S5; Tables S6 and S7.

Among the immune molecules, TGFB1 was the most highly expressed gene in H-MD with a significant association with mitoribosomal defects (Figures 3G–3I). This result is further supported by our previous report that TGFB1 is one of the common mitochondrial defect signature genes identified from various mitochondrial defect models (Lee et al., 2015). Furthermore, as shown in the GSEA result in Figure 2C, TGF-β signaling was significantly enriched in H-MD. These findings raise the possibility that TGFB1 may play a key role in mediating the effect of mitoribosomal defects within the tumor microenvironment. TGF-β signaling is known to exhibit either tumor-suppressive or oncogenic properties, depending on the tumor stage. Moreover, HCC tissues enriched with a late-responsive TGF-β signature showed a significantly poorer prognosis with shorter mean survival time, displaying an invasive phenotype and increased tumor recurrence compared with the tissues enriched with an early responsive TGF-β signature (Coulouarn et al., 2008). Consequently, we examined the relevance of mitoribosomal defects with the early- and late-responsive TGF-β signatures. In this analysis, we found that the late TGF-β signature was more enriched in H-MD, whereas few differences in the enrichment of the early TGF-β signature were observed between subtypes (Figures S4B and S4C), indicating a close association of mitoribosomal defects with the tumor-promoting TGF-β signature. To elucidate the effect of mitoribosomal defects on the tumor microenvironment, we assessed the TGF-β responses in endothelial cells, fibroblasts, T cells, and macrophages based on the cell type-specific TGF-β response signatures (TBRS) derived from cells cultured with TGF-β (Calon et al., 2012) (i.e., E-TBRS, F-TBRS, T-TBRS, or M-TBRS). In this analysis, all the TBRS were highly enriched in the H-MD group and were strongly associated with mitoribosomal defects (Figures S4D and S4H). Furthermore, we found a robust direct correlation among them, implying their concurrent expression among samples with mitoribosomal defects to a large extent (Figure S4H). These results suggest that mitoribosomal defect-mediated TGF-β expression may be an essential regulator of various cell types in the tumor microenvironment, thereby promoting HCC progression.

Mitoribosomal Defects Regulate Immune Modulator Expression in HCC Cells

Next, to validate the involvement of mitoribosomal defects in HCC immune modulator expression, we analyzed the transcriptome data of HCC cell lines from the Cancer Cell Line Encyclopedia (CCLE) (Transparent Methods). By applying the same approach in HCC tissue samples, we inferred MDSdn in cancer cell lines and categorized cancer cell lines into the H-MD type or L-MD type, respectively (Figure 4A and Table S8). First, to validate the strong association of mitoribosomal defects with immune modulators (Figure 3G), we compared their expression between H-MD and L-MD. This analysis confirmed the upregulation of TGFB1 and colony-stimulating factor-1 (CSF1) and the downregulation of ICOSLG in H-MD-type cells (Figures 4B, 4C, and S5A). These findings suggest that they are key immune modulators mediating the immune response induced by mitoribosomal defects in HCC cells. As CSF1 is involved in the immunosuppressive nature of the HCC microenvironment via the SPP1/CSF1/CSF1R axis (Zhu et al., 2019), we assessed the association of mitoribosomal defects with the CSF1 signaling axis. In addition to the upregulated expression of secreted phosphoprotein 1 (SPP1), the CSF1 axis was positively associated with mitoribosomal defects showing higher enrichment in H-MD-type cells (Figures 4D, 4E, and S5B). As a proof of concept, we selected two H-MD-type (JHH4 and SNU475) and two L-MD-type cell lines (HepG2 and JHH5) and confirmed the low to rare expression of a few representative down-MRPs in the H-MD-type compared with L-MD-type cells (Figure 4F). Following the unified nomenclature for MRP (De Silva et al., 2015), we used mS31 and mL46 for proteins of MRPS31 and MRPL46. We further examined whether down-MRPs were truly linked to the integrity of mitoribosome. When cellular components were subjected to sucrose gradient sedimentation, JHH4 showed a lower level of mitochondrial ribosomal subunit assembly (28S small, 39S large, and 55S total subunits) compared with HepG2 (Figure 4G). The assembly profile of JHH4 was not clearly distorted, indicating overall assembly was decreased due to the low MRP expression. Moreover, the protein levels of mitochondria-encoded genes, MT-CO2 and MT-ND6, were lower in the H-MD than in the L-MD type, whereas the protein level of the nuclear-encoded gene (succinate dehydrogenase complex flavoprotein subunit A; SDHA) was unchanged (Figure 4F). When HCC cells were exposed to MG132, which has been verified as an inhibitor of mitochondrial protein (Lavie et al., 2018), MT-COX2 level of H-MD-type cells was not restored to the basal level of L-MD-type cells, whereas that of the L-MD-type cell was further augmented (Figure 4H). These results indicate that the two H-MD-type cell lines support a decreased mitochondria-specific translational capacity, which is associated with mitoribosomal defects, rather than weak protein stability. We further validated that the two H-MD-type HCC cells, JHH4 and SNU475, harbored increased TGFB1 and CSF1 mRNA levels, compared with the L-MD-type cells (Figures 4I and 4J). When L-MD-type cell lines, HepG2 and JHH5, were exposed to doxycycline, which is a mitoribosome-specific translation inhibitor, the mtDNA-encoded protein expression of COX2 and ND6 was effectively decreased. However, TGFB1 and CSF1 mRNAs were significantly increased (Figures 4K–4P). These results clearly indicate that mitoribosomal defects in HCC cells themselves induce the expression and release of TGFB1 and CSF1, suggesting their crucial involvement in fostering an immune-suppressive environment.

Figure 4.

Figure 4

Open in a new tab

Mitoribosomal Defects Regulate HCC Immune Modulator Expression in HCC Cells

For validation in the HCC cell lines, transcriptome data for HCC cell line was obtained from CCLE dataset. MDSdn for each HCC cell line was calculated (Methods).

(A) According to the increasing order of MDSdn, MDSdn of each HCC cell line is plotted. HCC cell lines representing H-MD or L-MD subgroup are marked in red or blue color, respectively.

(B–E) Messenger RNA expression of TGFB1 (B), CSF1 (C), and SPP1 (D) are compared between H-MD-type and L-MD-type HCC cell lines of (A). The preRanked GSEA was performed based on the SPP1/CSF1/CSF1R/PDL1 signal axis, which is related to a sensitivity of anti-PD-1 immunotherapy in patients with HCC (E). The NES was compared between D- and ND-type cell lines (E). Boxplots are shown as first quartile, median, and third quartile (bottom box, middle line, and top box, respectively) with Welch two-sample t test p values. Whiskers represent the minimum and maximum values.

(F) Protein levels of Dn-MRPs (MRPS31 and MRPL46) and mitochondrial-encoded genes (MT-CO2 and MT-ND6) were validated in representative H-MD-type (JHH4 and SNU475) and L-MD-type cell line (HepG2 and JHH5) by western blot. According to new nomenclature for mitoribosomal protein, we used mS31 and mL46 for proteins of MRPS31and MRPL46.

(G) Sucrose gradient sedimentation analysis of MRPs from whole-cell lysates of the indicated cell lines. According to the new nomenclature for mitoribosomal protein, we used mS31, uS22, uS15, uL15, uL13, uL11, and mL46 for proteins of MRPS31, MRPS22, MRPS15, MRPL13, MRPL11, and MRPL46.

(H) HCC cells were exposed to various doses of MG132 for 12 h and subjected to western blot analysis. DMSO was used as a vehicle (V).

(I and J) Messenger RNA expression of TGFB1 (I) and CSF1 (J) in two H-MD-type (JHH4 and SNU475) and two L-MD-type cell lines (HepG2 and JHH5) were monitored by qRT-PCR.

(K–P) L-MD-type cell lines (HepG2 and JHH5) were exposed to various doses of mitochondria-specific translation inhibitor, doxycycline (DOX) for 72 h. JHH5 cells showed higher sensitivity to DOX than HepG2. Western blot analysis for the expression level of mtDNA-encoded proteins, COX2 and ND6, and nuclear DNA-encoded SDHA protein was also examined to prove the inhibitor's specificity (K and N). Messenger RNA expression of TGFB1 (L and O) and CSF1 (M and P) were examined by qRT-PCR. Bar plots are represented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005 (Student's t test, DOX treated group versus non-treated group).

See also Table S8.

TBRS Induced by MRP Defects Mediates an Aggressive Phenotype in HCC

Besides suppressing the immune response, TBRS modulates cancer cell activity mainly through EMT in HCC (Chen et al., 2019, Reichl et al., 2012). Therefore, we evaluated the association of TBRS with mitoribosomal defects in HCC cells. When the TBRS of epithelial cells (Padua et al., 2008) was applied to the transcriptome data of HCC cells, mitoribosomal defects showed a strong association with TBRS, as corroborated by the increased enrichment of TBRS in H-MD-type than L-MD-type cells (Figures 5A and 5B). We also found that among the common HCC molecular subclass described by Hoshida et al. (2009) (Hoshida et al., 2009), mitoribosomal defects had a strong positive association with the S1 subtype, yet negative associations with both the S2 and the S3 signatures (Figure 5C). These findings suggest that mitoribosomal defects are selectively associated with the molecular features of the S1 subtype. As vascular invasion is a well-defined clinical feature for the S1 subtype, we assessed the association of mitoribosomal defects with an invasive signature from a multi-cancer study (Anastassiou et al., 2011). Consistently, mitoribosomal defects were positively associated with cancer cell invasiveness (Figure 5D). Furthermore, invasive features were highly enriched in H-MD-type cells (Figure 5E). To examine whether the invasion ability of HCC cells was influenced by mitoribosomal defect-mediated TGF-β signaling, we implemented the invasion assay using H-MD-type cells (JHH4 and SNU475) and L-MD-type (HepG2 and JHH5) HCC cells. H-MD-type cells showed higher invasion activity than L-MD-type cells despite their delayed cell growth rate (Figures 5F and 5G). When we incubated H-MD-type cells with culture medium containing TGF-β-neutralizing antibodies to block TGF-β signaling, TBRS was effectively abrogated as evidenced by decreased Smad2 phosphorylation (Figures 5H and 5I). This indicated that H-MD-type cells truly released TGF-β and that their TBRS was active. Moreover, we found that the invasion capacity of H-MD-type cells was significantly reduced by eliminating TBRS (Figures 5J and 5K). These results indicate that mitoribosomal defect-mediated TGF-β release regulated the invasion activity of H-MD-type cells in an autocrine manner.

Figure 5.

Figure 5

Open in a new tab

TBRS Induced by MRP Defects Mediates an Aggressive Phenotype in HCC

(A and B) The association of mitoribosomal defects in HCC with TBRS was recapitulated in HCC cell lines. For TBRS in HCC cell line, signatures of epithelial cells (Epi_TBRS) were used. The MDSdn shows a significant correlation with Epi_TBRS (A). NES based on Epi_TBRS was compared between H-MD- and L-MD-type HCC cell lines (B).

(C) MDSdn shows differential associations with HCC sub-classification signatures (Hoshida_S1, S2, and S3).

(D and E) The association of mitoribosomal defect with cancer cell invasiveness was examined. preRanked GSEA was performed based on the invasion signature defined by Anastassiou et al. MDSdn shows a significant association with cancer cell invasiveness (D). NES based on cancer cell invasiveness signature (Anastassiou et al.) was compared between H-MD and L-MD-type HCC cell lines (E).

(F) Cell invasion activity was assayed using Matrigel-coated Transwell as described in the Transparent Methods section. Representative images are shown in the right panel.

(G) Cell growth rates were monitored by counting the trypan blue-negative viable cells using the Countess automated cell counter.

(H–K) TGF-β signaling in H-MD-type HCC cell lines (SNU475 and JHH4) was abolished by exposure to various concentrations of neutralizing antibodies for TGF-β (GTX14052, GeneTex Inc., Irvine, CA) for 24 h. Normal mouse IgG (sc-2025, Santa Cruz Biotechnology, Inc., Dallas, TX) was used as control. Phosphorylated SMAD2 (P-smad2) and total SMAD2 (T-smad2) levels were examined by western blot to validate the extent of TGF-β signaling inhibition (H and I). Then, the cell invasion assay was performed (J and K). Representative images for invaded cells are shown in the right panels. Boxplots are shown as first quartile, median, and third quartile (bottom box, middle line, and top box, respectively) with Welch two-sample t test p values. Whiskers represent the minimum and maximum values. Bar plots are represented as mean ± SEM. ∗p<0.05, ∗∗p<0.01, ∗∗∗p<0.005 (Student’s t-test, DOX treated group vs. non-treated group).

See also Figure S6; Table S9.

Last, we postulated that the mitoribosomal defect-mediated invasion capacity of H-MD-type cells was also closely linked with HCC recurrence. To test this hypothesis, we performed GSEA using HCC recurrence signatures (Woo et al., 2008, Hoshida et al., 2008, Iizuka et al., 2003, Yoshioka et al., 2009, Kurokawa et al., 2004) and examined their association with mitoribosomal defects (Table S8). Based on the upregulated or downregulated genes in the recurred HCC samples from Woo et al. (2008) and Hoshida et al. (2008), we investigated the association of mitoribosomal defects with HCC recurrence. In this analysis, the H-MD subtype showed increased enrichment of upregulated genes and decreased enrichment of downregulated genes, indicating that mitoribosomal defects in HCC cells were strongly associated with HCC recurrence (Table S9 and Figures S6A–S6D). Although early recurrence was one of the clinical features of the S1 subtype (Hoshida et al., 2009), there was no significant association of early recurrence signatures with mitoribosomal defects (Table S9 and Figures S6E and S6F). Considering that the distinct genetic features were involved in the early and late recurrence of HCC, these results implied that mitoribosomal defects were selectively associated with molecular features of late recurring HCC.

Taken together, these results indicate that the mitoribosomal defects caused by aberrant expression of down-MRPs are one of the causes of OXPHOS impairment and the related overall mitochondrial dysfunction, further introducing a novel mitochondria-related HCC hallmark. Also, our findings demonstrate that mitoribosomal defects actively participate in establishing a favorable microenvironment by recruiting suppressive immune cells and activating the ECM to increase the invasion capacity of tumor cells. In detail, the TGF-β signaling pathway is a crucial mechanism to mediate mitoribosomal defects with the diverse aggressive molecular features of HCC. This implies that a set of down-MRPs may be used as good prognostic markers for HCC and that therapeutics co-targeting TGF-β and down-MRPs can be a more selective strategy for the treatment of HCC with mitoribosomal defects.

Discussion

As a specialized mitochondrial translational machinery, the mitoribosome integrity is essential for OXPHOS function including ATP production in an aerobic manner. Thus, aberrant expression of the MRPs comprising mitoribosomes may result in critical defects in energy production and in the overall regulation of cellular respiration, further resulting in deregulated cellular energetics, which has recently been added as a cancer hallmark (Hanahan and Weinberg, 2011). There are several underlying mechanisms such as mtDNA mutations, mitochondrial enzyme defects, or altered oncogenes/tumor suppressors explaining the close association of mitochondrial dysfunction with deregulated cellular energetics (Wallace, 2012, Cairns et al., 2011). However, to date, the contribution of mitoribosomal defects on deregulated cellular energetics has not been studied well. Moreover, metabolic reprogramming in hepatocytes has been proposed as a key component in the progression to malignant HCC cells (Cancer Genome Atlas Research Network. Electronic address and Cancer Genome Atlas Research Network, 2017). Thus, we hypothesized that mitoribosomal defects induced by MRP aberration may contribute to the initiation and progression of HCC. To the best of our knowledge, this study is the first to systematically profile the MRPs in HCC and elucidate how mitoribosomal defects affect HCC progression. In this study, we identified a signature reflecting mitoribosomal defects, consisting of six downregulated MRPs, MRPL39, MRPL46, MRPL34, MRPL54, MPLS6, and MPLS31, and devised a scoring metric of the mitoribosome defect using these genes and designated it as MDS. Interestingly, we found that MDS had close associations with not only essential mitochondrial functions such as OXPHOS but also aggressive phenotypes such as the cell cycle and EMT. Moreover, MDS also had prognostic significance in indicating shorter overall survival time. These results demonstrated that the MDS could reflect the oncogenic potential of mitoribosomal defect. In addition, the differential associations of MDS with infiltrated immune cells, associations such as higher MDS in the tumor showed increased enrichment with immune cells having suppressive characteristics and less enrichment with immune cells having stimulatory characteristics, indicated that mitoribosomal defects attract a differential repertoire of immune cells and provide a favorable microenvironment for cancer cells by recruiting immune cells with suppressive activity and promoting escape from immune surveillance.

Among suppressive cytokines, TGFB1 showed a significant association with mitoribosomal defects as in vitro analysis using the H-MD-type and L-MD-type HCC cell lines showed increased TGFB1 expression in the H-MD type and mitoribosome defect-induced TBRS activation affected cancer cell invasion. Based on our findings, we postulate that TGFB1 plays a crucial role in constructing a favorable microenvironment that is induced by the mitoribosomal defects. As a pleiotropic cytokine, TGF-β regulates cellular proliferation, differentiation, and ECM production. Aberrant expression of TGF-β or dysregulated TBRS has been involved in the pathogenesis of a variety of diseases, including cancer and fibrosis. Moreover, TGF-β was elevated in both the plasma and tumor tissues of patients with HCC (Giannelli et al., 2002), and TBRS activation was considered to engage in hepatocarcinogenesis and to contribute to EMT in HCC models (Giannelli et al., 2005). Interestingly, Yi et al. recently reported a direct association between TGF-β with mitochondrial dysfunction in HCC, demonstrating that TGFB1 expression was augmented in mitochondria-depleted ρ0 HCC cell lines and that TBRS was involved in the EMT induced by mitochondrial dysfunction (Yi et al., 2015). In addition, a recent study using human peripheral and tumor-associated lymphocytes, tumor-derived TGF-β, demonstrated the suppression of a key antitumor function of CD4+ T cells and interferon-γ production by inhibiting the mitochondrial respiration (Dimeloe et al., 2019). Considering these findings, we suggest TGF-β as a key effector mediating the effect of mitoribosomal defect on HCC progression and immune evasion.

However, we could not clearly present how mitoribosome defects in HCC upregulate the expression of TGFB1 and activate TBRS, or clarify how activated TBRS fosters a cancer-favorable microenvironment and suppresses anti-tumor immunity. One plausible explanation for this can be the rise of mitochondrial ROS as a result of OXPHOS dysfunction in mitoribosome-defective HCC. Furthermore, Jain et al. reported that mitochondrial ROS regulated TBRS and augmented TGF-β-mediated transcription in primary normal human lung fibroblasts (Jain et al., 2013). Depending on their concentration and cellular compartmentalization, ROS behave ambivalently in cancer progression. ROS facilitate carcinogenesis and cancer progression under mild-to-moderate elevated levels, whereas excessive ROS damages cancer cells dramatically, causing cell death (Zhang et al., 2016, Tafani et al., 2016). Furthermore, as signaling messengers in the immune system, ROS are associated with tumor-induced immunosuppression in the tumor microenvironment by extensively participating in T cell activation, apoptosis, and hyporesponsiveness. Owing to this, we consider that ROS evoked from the mitoribosomal defects may play a crucial function in the suppressive immunity of the cancer microenvironment via TBRS. Meanwhile, TGF-β, which can promote ROS generation in the mitochondria of damaged hepatocytes, has been reported by directly downregulating antioxidative systems (Albright et al., 2003). In this way, both TGF-β and ROS are building a strong interaction that provides an advantage to cancer cells to increment their malignancy. The mutual collaboration of TGF-β and ROS in mitoribosomal defect may be difficult to probe, due to their inherently complex and contradictory role in cancer progression (Krstic et al., 2015).

Although it is not clear how mitoribosome defects in HCC foster a cancer-favorable microenvironment and suppress anti-tumor immunity, we found that SPP1, CSF1, as well as TGFB1 was highly enriched in the mitoribosome-defective tumor. Furthermore, activation of the SPP1/CSF1/CSF1R axis has been reported to provide the immunosuppressive nature of the HCC microenvironment and to be correlated with the increased expression of programmed cell death protein 1 (PD-1) (Zhu et al., 2019). Furthermore, the expression level of PD-1 is correlated with cellular ROS production and oxidative metabolism (Tkachev et al., 2015). PD-1 is considered an immune checkpoint that controls T cell function, and the high expression of PD-1 is currently under investigation as a potential predictor of responses to anti-PD1 therapy (Herbst et al., 2016, Garon et al., 2015, Topalian et al., 2012). As PD-1 maintains immune homeostasis by negatively regulating T cell function and survival, PD-1 blockade has been recently approved to treat patients with advanced-stage cancers by enhancing antitumor T cell immunity (Tan and Quintal, 2015). Furthermore, TGF-β inhibitors showed promise in several preclinical in vitro and in vivo studies (Faivre et al., 2019). Therefore, it would be interesting to stratify patients with HCC based on the magnitude of mitoribosomal defects and to explore the potential of combining ROS scavengers or TGF-β inhibitors with PD-1 signaling blockade for rapid clinical translation. Collectively, our findings demonstrate that mitoribosomal defect assessment may facilitate a survey of liver tumor immunity and stratification of patients for immune-based therapy.

Limitations of the Study

In this study, we could not present the precise underlying mechanism of how mitoribosome defects in HCC upregulate the expression of TGFB1 and activate TBRS. Also, we could not clarify how activated TBRS fosters a cancer-favorable microenvironment and suppresses anti-tumor immunity.

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by Lead Contact, So Mee Kwon (smkwon@ajou.ac.kr)

Material Availability

This study did not generate new unique reagents. RNA sequencing data for TCGA liver cancer cohorts and HCC cell lines were obtained from Genomic Data Commons Data portal (https://docs.gdc.cancer.gov) and CCLE (Cancer Cell Line Encyclopedia, https://portals.broadinstitute.org/ccle).

Data and Code Availability

No previously unreported algorithms were used to generate the results.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.

Acknowledgments

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (MSIT) (G.Y., NRF-2015R1A2A1A10055038; H.G.W., NRF-2017R1E1A1A01074733 and NRF-2017M3A9B6061509) and also supported by a fellowship from the Basic Science Research Program through the NRF grant funded by the Ministry of Education (S.M.K., NRF-2017R1A6A3A11030191).

Author Contributions

S.M.K and G.Y. conceived the idea. S.M.K. carried out most of the data analysis. Y.-K.L. and S.M. performed all experiments. H.G.W. and H.J.W advised data analysis approaches. S.M.K., H.G.W., and G.Y. interpreted data and prepared the manuscript. All authors read, advised, and approved the manuscript.

Declaration of Interests

The authors declare no competing interests.

Published: June 26, 2020

Footnotes

Supplemental Information can be found online at https://doi.org/10.1016/j.isci.2020.101247.

Contributor Information

So Mee Kwon, Email: smkwon@ajou.ac.kr.

Gyesoon Yoon, Email: ypeace@ajou.ac.kr.

Supplemental Information

Document S1. Transparent Methods, Figures S1–S6, and Tables S1–S3 and S6–S9
mmc1.pdf (3MB, pdf)
Table S4. Top 20 Positively Enriched Gene Sets for either H-MD or L-MD Group in GSEA Analysis, Related to Figures 2A–2C
mmc2.xlsx (33.1KB, xlsx)
Table S5. Differentially Expressed Genes (DEGs) Between H-MD and L-MD Subtypes, Related to Figure 2E
mmc3.xlsx (50.1KB, xlsx)

References

  1. Albright C.D., Salganik R.I., Craciunescu C.N., Mar M.H., Zeisel S.H. Mitochondrial and microsomal derived reactive oxygen species mediate apoptosis induced by transforming growth factor-beta1 in immortalized rat hepatocytes. J. Cell Biochem. 2003;89:254–261. doi: 10.1002/jcb.10498. [DOI] [PubMed] [Google Scholar]
  2. Anastassiou D., Rumjantseva V., Cheng W., Huang J., Canoll P.D., Yamashiro D.J., Kandel J.J. Human cancer cells express Slug-based epithelial-mesenchymal transition gene expression signature obtained in vivo. BMC Cancer. 2011;11:529. doi: 10.1186/1471-2407-11-529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Aran D., Hu Z., Butte A.J. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. 2017;18:220. doi: 10.1186/s13059-017-1349-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Auger C., Alhasawi A., Contavadoo M., Appanna V.D. Dysfunctional mitochondrial bioenergetics and the pathogenesis of hepatic disorders. Front. Cell Dev Biol. 2015;3:40. doi: 10.3389/fcell.2015.00040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cairns R.A., Harris I.S., Mak T.W. Regulation of cancer cell metabolism. Nat. Rev. Cancer. 2011;11:85–95. doi: 10.1038/nrc2981. [DOI] [PubMed] [Google Scholar]
  6. Caja L., Dituri F., Mancarella S., Caballero-Diaz D., Moustakas A., Giannelli G., Fabregat I. TGF-beta and the tissue microenvironment: relevance in fibrosis and cancer. Int. J. Mol. Sci. 2018;19:1294–1317. doi: 10.3390/ijms19051294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Calon A., Espinet E., Palomo-Ponce S., Tauriello D.V., Iglesias M., Cespedes M.V., Sevillano M., Nadal C., Jung P., Zhang X.H. Dependency of colorectal cancer on a TGF-beta-driven program in stromal cells for metastasis initiation. Cancer Cell. 2012;22:571–584. doi: 10.1016/j.ccr.2012.08.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cancer Genome Atlas Research Network. Electronic address: wheeler@bcm.edu. Cancer Genome Atlas Research Network. Comprehensive and Integrative Genomic Characterization of Hepatocellular Carcinoma. Cell. 2017;169:1327–1341.e23. doi: 10.1016/j.cell.2017.05.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carim L., Sumoy L., Nadal M., Estivill X., Escarceller M. Cloning, expression, and mapping of PDCD9, the human homolog of Gallus gallus pro-apoptotic protein p52. Cytogenet. Cell Genet. 1999;87:85–88. doi: 10.1159/000015397. [DOI] [PubMed] [Google Scholar]
  10. Chakravarthy A., Khan L., Bensler N.P., Bose P., De Carvalho D.D. TGF-beta-associated extracellular matrix genes link cancer-associated fibroblasts to immune evasion and immunotherapy failure. Nat. Commun. 2018;9:4692. doi: 10.1038/s41467-018-06654-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chandra D., Singh K.K. Genetic insights into OXPHOS defect and its role in cancer. Biochim. Biophys. Acta. 2011;1807:620–625. doi: 10.1016/j.bbabio.2010.10.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chen J., Gingold J.A., Su X. Immunomodulatory TGF-beta signaling in hepatocellular carcinoma. Trends Mol. Med. 2019;25:1010–1023. doi: 10.1016/j.molmed.2019.06.007. [DOI] [PubMed] [Google Scholar]
  13. Chintharlapalli S.R., Jasti M., Malladi S., Parsa K.V., Ballestero R.P., Gonzalez-Garcia M. BMRP is a Bcl-2 binding protein that induces apoptosis. J. Cell Biochem. 2005;94:611–626. doi: 10.1002/jcb.20292. [DOI] [PubMed] [Google Scholar]
  14. Colegio O.R., Chu N.Q., Szabo A.L., Chu T., Rhebergen A.M., Jairam V., Cyrus N., Brokowski C.E., Eisenbarth S.C., Phillips G.M. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513:559–563. doi: 10.1038/nature13490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Coulouarn C., Factor V.M., Thorgeirsson S.S. Transforming growth factor-beta gene expression signature in mouse hepatocytes predicts clinical outcome in human cancer. Hepatology. 2008;47:2059–2067. doi: 10.1002/hep.22283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. De Silva D., Tu Y.T., Amunts A., Fontanesi F., Barrientos A. Mitochondrial ribosome assembly in health and disease. Cell Cycle. 2015;14:2226–2250. doi: 10.1080/15384101.2015.1053672. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Degli Esposti D., Hamelin J., Bosselut N., Saffroy R., Sebagh M., Pommier A., Martel C., Lemoine A. Mitochondrial roles and cytoprotection in chronic liver injury. Biochem. Res. Int. 2012;2012:387626. doi: 10.1155/2012/387626. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dimeloe S., Gubser P., Loeliger J., Frick C., Develioglu L., Fischer M., Marquardsen F., Bantug G.R., Thommen D., Lecoultre Y. Tumor-derived TGF-beta inhibits mitochondrial respiration to suppress IFN-gamma production by human CD4(+) T cells. Sci. Signal. 2019;12:eaav3334. doi: 10.1126/scisignal.aav3334. [DOI] [PubMed] [Google Scholar]
  19. Faivre S., Santoro A., Kelley R.K., Gane E., Costentin C.E., Gueorguieva I., Smith C., Cleverly A., Lahn M.M., Raymond E. Novel transforming growth factor beta receptor I kinase inhibitor galunisertib (LY2157299) in advanced hepatocellular carcinoma. Liver Int. 2019;39:1468–1477. doi: 10.1111/liv.14113. [DOI] [PubMed] [Google Scholar]
  20. Garon E.B., Rizvi N.A., Hui R., Leighl N., Balmanoukian A.S., Eder J.P., Patnaik A., Aggarwal C., Gubens M., Horn L. Pembrolizumab for the treatment of non-small-cell lung cancer. N. Engl. J. Med. 2015;372:2018–2028. doi: 10.1056/NEJMoa1501824. [DOI] [PubMed] [Google Scholar]
  21. Giannelli G., Bergamini C., Fransvea E., Sgarra C., Antonaci S. Laminin-5 with transforming growth factor-beta1 induces epithelial to mesenchymal transition in hepatocellular carcinoma. Gastroenterology. 2005;129:1375–1383. doi: 10.1053/j.gastro.2005.09.055. [DOI] [PubMed] [Google Scholar]
  22. Giannelli G., Fransvea E., Marinosci F., Bergamini C., Colucci S., Schiraldi O., Antonaci S. Transforming growth factor-beta1 triggers hepatocellular carcinoma invasiveness via alpha3beta1 integrin. Am. J. Pathol. 2002;161:183–193. doi: 10.1016/s0002-9440(10)64170-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hanahan D., Weinberg R.A. Hallmarks of cancer: the next generation. Cell. 2011;144:646–674. doi: 10.1016/j.cell.2011.02.013. [DOI] [PubMed] [Google Scholar]
  24. Herbst R.S., Baas P., Kim D.W., Felip E., Perez-Gracia J.L., Han J.Y., Molina J., Kim J.H., Arvis C.D., Ahn M.J. Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet. 2016;387:1540–1550. doi: 10.1016/S0140-6736(15)01281-7. [DOI] [PubMed] [Google Scholar]
  25. Hoshida Y., Nijman S.M., Kobayashi M., Chan J.A., Brunet J.P., Chiang D.Y., Villanueva A., Newell P., Ikeda K., Hashimoto M. Integrative transcriptome analysis reveals common molecular subclasses of human hepatocellular carcinoma. Cancer Res. 2009;69:7385–7392. doi: 10.1158/0008-5472.CAN-09-1089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hoshida Y., Villanueva A., Kobayashi M., Peix J., Chiang D.Y., Camargo A., Gupta S., Moore J., Wrobel M.J., Lerner J. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N. Engl. J. Med. 2008;359:1995–2004. doi: 10.1056/NEJMoa0804525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Iizuka N., Oka M., Yamada-Okabe H., Nishida M., Maeda Y., Mori N., Takao T., Tamesa T., Tangoku A., Tabuchi H. Oligonucleotide microarray for prediction of early intrahepatic recurrence of hepatocellular carcinoma after curative resection. Lancet. 2003;361:923–929. doi: 10.1016/S0140-6736(03)12775-4. [DOI] [PubMed] [Google Scholar]
  28. Jain M., Rivera S., Monclus E.A., Synenki L., Zirk A., Eisenbart J., Feghali-Bostwick C., Mutlu G.M., Budinger G.R., Chandel N.S. Mitochondrial reactive oxygen species regulate transforming growth factor-beta signaling. J. Biol. Chem. 2013;288:770–777. doi: 10.1074/jbc.M112.431973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Khalaf A.M., Fuentes D., Morshid A.I., Burke M.R., Kaseb A.O., Hassan M., Hazle J.D., Elsayes K.M. Role of Wnt/beta-catenin signaling in hepatocellular carcinoma, pathogenesis, and clinical significance. J. Hepatocell Carcinoma. 2018;5:61–73. doi: 10.2147/JHC.S156701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Koc E.C., Haciosmanoglu E., Claudio P.P., Wolf A., Califano L., Friscia M., Cortese A., Koc H. Impaired mitochondrial protein synthesis in head and neck squamous cell carcinoma. Mitochondrion. 2015;24:113–121. doi: 10.1016/j.mito.2015.07.123. [DOI] [PubMed] [Google Scholar]
  31. Krstic J., Trivanovic D., Mojsilovic S., Santibanez J.F. Transforming growth factor-beta and oxidative stress interplay: implications in tumorigenesis and cancer progression. Oxid Med. Cell Longev. 2015;2015:654594. doi: 10.1155/2015/654594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kurokawa Y., Matoba R., Takemasa I., Nagano H., Dono K., Nakamori S., Umeshita K., Sakon M., Ueno N., Oba S. Molecular-based prediction of early recurrence in hepatocellular carcinoma. J. Hepatol. 2004;41:284–291. doi: 10.1016/j.jhep.2004.04.031. [DOI] [PubMed] [Google Scholar]
  33. Lavie J., De Belvalet H., Sonon S., Ion A.M., Dumon E., Melser S., Lacombe D., Dupuy J.W., Lalou C., Benard G. Ubiquitin-dependent degradation of mitochondrial proteins regulates energy metabolism. Cell Rep. 2018;23:2852–2863. doi: 10.1016/j.celrep.2018.05.013. [DOI] [PubMed] [Google Scholar]
  34. LeBleu V.S., O'Connell J.T., Gonzalez Herrera K.N., Wikman H., Pantel K., Haigis M.C., de Carvalho F.M., Damascena A., Domingos Chinen L.T., Rocha R.M. PGC-1alpha mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis. Nat. Cell Biol. 2014;16:992–1003. doi: 10.1038/ncb3039. 1-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lee Y.K., Jee B.A., Kwon S.M., Yoon Y.S., Xu W.G., Wang H.J., Wang X.W., Thorgeirsson S.S., Lee J.S., Woo H.G., Yoon G. Identification of a mitochondrial defect gene signature reveals NUPR1 as a key regulator of liver cancer progression. Hepatology. 2015;62:1174–1189. doi: 10.1002/hep.27976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lee Y.K., Lim J.J., Jeoun U.W., Min S., Lee E.B., Kwon S.M., Lee C., Yoon G. Lactate-mediated mitoribosomal defects impair mitochondrial oxidative phosphorylation and promote hepatoma cell invasiveness. J. Biol. Chem. 2017;292:20208–20217. doi: 10.1074/jbc.M117.809012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Levy-Strumpf N., Kimchi A. Death associated proteins (DAPs): from gene identification to the analysis of their apoptotic and tumor suppressive functions. Oncogene. 1998;17:3331–3340. doi: 10.1038/sj.onc.1202588. [DOI] [PubMed] [Google Scholar]
  38. Liberzon A., Birger C., Thorvaldsdottir H., Ghandi M., Mesirov J.P., Tamayo P. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. 2015;1:417–425. doi: 10.1016/j.cels.2015.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Loo L.W., Cheng I., Tiirikainen M., Lum-Jones A., Seifried A., Dunklee L.M., Church J.M., Gryfe R., Weisenberger D.J., Haile R.W. cis-Expression QTL analysis of established colorectal cancer risk variants in colon tumors and adjacent normal tissue. PLoS One. 2012;7:e30477. doi: 10.1371/journal.pone.0030477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Luedde T., Schwabe R.F. NF-kappaB in the liver--linking injury, fibrosis and hepatocellular carcinoma. Nat. Rev. Gastroenterol. Hepatol. 2011;8:108–118. doi: 10.1038/nrgastro.2010.213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Martinez-Outschoorn U.E., Balliet R.M., Rivadeneira D.B., Chiavarina B., Pavlides S., Wang C., Whitaker-Menezes D., Daumer K.M., Lin Z., Witkiewicz A.K. Oxidative stress in cancer associated fibroblasts drives tumor-stroma co-evolution: a new paradigm for understanding tumor metabolism, the field effect and genomic instability in cancer cells. Cell Cycle. 2010;9:3256–3276. doi: 10.4161/cc.9.16.12553. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Moffitt R.A., Marayati R., Flate E.L., Volmar K.E., Loeza S.G., Hoadley K.A., Rashid N.U., Williams L.A., Eaton S.C., Chung A.H. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat. Genet. 2015;47:1168–1178. doi: 10.1038/ng.3398. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Nunes J.B., Peixoto J., Soares P., Maximo V., Carvalho S., Pinho S.S., Vieira A.F., Paredes J., Rego A.C., Ferreira I.L. OXPHOS dysfunction regulates integrin-beta1 modifications and enhances cell motility and migration. Hum. Mol. Genet. 2015;24:1977–1990. doi: 10.1093/hmg/ddu612. [DOI] [PubMed] [Google Scholar]
  44. Padua D., Zhang X.H., Wang Q., Nadal C., Gerald W.L., Gomis R.R., Massague J. TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell. 2008;133:66–77. doi: 10.1016/j.cell.2008.01.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pietromonaco S.F., Denslow N.D., O'Brien T.W. Proteins of mammalian mitochondrial ribosomes. Biochimie. 1991;73:827–835. doi: 10.1016/0300-9084(91)90062-6. [DOI] [PubMed] [Google Scholar]
  46. Pu M., Wang J., Huang Q., Zhao G., Xia C., Shang R., Zhang Z., Bian Z., Yang X., Tao K. High MRPS23 expression contributes to hepatocellular carcinoma proliferation and indicates poor survival outcomes. Tumour Biol. 2017;39:1–12. doi: 10.1177/1010428317709127. [DOI] [PubMed] [Google Scholar]
  47. Reichl P., Haider C., Grubinger M., Mikulits W. TGF-beta in epithelial to mesenchymal transition and metastasis of liver carcinoma. Curr. Pharm. Des. 2012;18:4135–4147. doi: 10.2174/138161212802430477. [DOI] [PubMed] [Google Scholar]
  48. Sia D., Jiao Y., Martinez-Quetglas I., Kuchuk O., Villacorta-Martin C., Castro de Moura M., Putra J., Camprecios G., Bassaganyas L., Akers N. Identification of an immune-specific class of hepatocellular carcinoma, based on molecular features. Gastroenterology. 2017;153:812–826. doi: 10.1053/j.gastro.2017.06.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Sorensen K.M., Meldgaard T., Melchjorsen C.J., Fridriksdottir A.J., Pedersen H., Petersen O.W., Kristensen P. Upregulation of Mrps18a in breast cancer identified by selecting phage antibody libraries on breast tissue sections. BMC Cancer. 2017;17:19. doi: 10.1186/s12885-016-2987-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sukumar M., Kishton R.J., Restifo N.P. Metabolic reprograming of anti-tumor immunity. Curr. Opin. Immunol. 2017;46:14–22. doi: 10.1016/j.coi.2017.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Tafani M., Sansone L., Limana F., Arcangeli T., De Santis E., Polese M., Fini M., Russo M.A. The interplay of reactive oxygen species, hypoxia, inflammation, and sirtuins in cancer initiation and progression. Oxid Med. Cell Longev. 2016;2016:3907147. doi: 10.1155/2016/3907147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tan M., Quintal L. Pembrolizumab: a novel antiprogrammed death 1 (PD-1) monoclonal antibody for treatment of metastatic melanoma. J. Clin. Pharm. Ther. 2015;40:504–507. doi: 10.1111/jcpt.12304. [DOI] [PubMed] [Google Scholar]
  53. Tkachev V., Goodell S., Opipari A.W., Hao L.Y., Franchi L., Glick G.D., Ferrara J.L., Byersdorfer C.A. Programmed death-1 controls T cell survival by regulating oxidative metabolism. J. Immunol. 2015;194:5789–5800. doi: 10.4049/jimmunol.1402180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Topalian S.L., Hodi F.S., Brahmer J.R., Gettinger S.N., Smith D.C., McDermott D.F., Powderly J.D., Carvajal R.D., Sosman J.A., Atkins M.B. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N. Engl. J. Med. 2012;366:2443–2454. doi: 10.1056/NEJMoa1200690. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Uhlen M., Fagerberg L., Hallstrom B.M., Lindskog C., Oksvold P., Mardinoglu A., Sivertsson A., Kampf C., Sjostedt E., Asplund A. Proteomics. Tissue-based map of the human proteome. Science. 2015;347:1260419. doi: 10.1126/science.1260419. [DOI] [PubMed] [Google Scholar]
  56. Wallace D.C. Mitochondria and cancer. Nat. Rev. Cancer. 2012;12:685–698. doi: 10.1038/nrc3365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Warburg O. On the origin of cancer cells. Science. 1956;123:309–314. doi: 10.1126/science.123.3191.309. [DOI] [PubMed] [Google Scholar]
  58. Wei Z., Jia J., Heng G., Xu H., Shan J., Wang G., Liu C., Xia J., Zhou H., Wu M. SIRT1/MRPS5 axis enhances the metabolic flexibility of liver cancer stem cells. Hepatology. 2019;70:1197–1213. doi: 10.1002/hep.30622. [DOI] [PubMed] [Google Scholar]
  59. Woo H.G., Park E.S., Cheon J.H., Kim J.H., Lee J.S., Park B.J., Kim W., Park S.C., Chung Y.J., Kim B.G. Gene expression-based recurrence prediction of hepatitis B virus-related human hepatocellular carcinoma. Clin. Cancer Res. 2008;14:2056–2064. doi: 10.1158/1078-0432.CCR-07-1473. [DOI] [PubMed] [Google Scholar]
  60. Yi E.Y., Park S.Y., Jung S.Y., Jang W.J., Kim Y.J. Mitochondrial dysfunction induces EMT through the TGF-beta/Smad/Snail signaling pathway in Hep3B hepatocellular carcinoma cells. Int. J. Oncol. 2015;47:1845–1853. doi: 10.3892/ijo.2015.3154. [DOI] [PubMed] [Google Scholar]
  61. Yoshioka S., Takemasa I., Nagano H., Kittaka N., Noda T., Wada H., Kobayashi S., Marubashi S., Takeda Y., Umeshita K. Molecular prediction of early recurrence after resection of hepatocellular carcinoma. Eur. J. Cancer. 2009;45:881–889. doi: 10.1016/j.ejca.2008.12.019. [DOI] [PubMed] [Google Scholar]
  62. Zhang L., Li J., Zong L., Chen X., Chen K., Jiang Z., Nan L., Li X., Li W., Shan T. Reactive oxygen species and targeted therapy for pancreatic cancer. Oxid Med. Cell Longev. 2016;2016:1616781. doi: 10.1155/2016/1616781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Zhu Y., Yang J., Xu D., Gao X.M., Zhang Z., Hsu J.L., Li C.W., Lim S.O., Sheng Y.Y., Zhang Y. Disruption of tumour-associated macrophage trafficking by the osteopontin-induced colony-stimulating factor-1 signalling sensitises hepatocellular carcinoma to anti-PD-L1 blockade. Gut. 2019;68:1653–1666. doi: 10.1136/gutjnl-2019-318419. [DOI] [PubMed] [Google Scholar]
  64. Zong W.X., Rabinowitz J.D., White E. Mitochondria and cancer. Mol. Cell. 2016;61:667–676. doi: 10.1016/j.molcel.2016.02.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Transparent Methods, Figures S1–S6, and Tables S1–S3 and S6–S9
mmc1.pdf (3MB, pdf)
Table S4. Top 20 Positively Enriched Gene Sets for either H-MD or L-MD Group in GSEA Analysis, Related to Figures 2A–2C
mmc2.xlsx (33.1KB, xlsx)
Table S5. Differentially Expressed Genes (DEGs) Between H-MD and L-MD Subtypes, Related to Figure 2E
mmc3.xlsx (50.1KB, xlsx)

Data Availability Statement

No previously unreported algorithms were used to generate the results.

Articles from iScience are provided here courtesy of Elsevier

RESOURCES