Metabolic Reprogramming Into a Glycolysis Phenotype Induced by Extracellular Vesicles Derived From Prostate Cancer Cells

Abstract

Most cancer cells adopt a less efficient metabolic process of aerobic glycolysis with high level of glucose uptake followed by lactic acid production, known as the Warburg effect. This phenotypic transition enables cancer cells to achieve increased cellular survival and proliferation in a harsh low-oxygen tumor microenvironment. Also, the resulting acidic microenvironment causes inactivation of the immune system such as T-cell impairment that favors escape by immune surveillance. While lots of studies have revealed that tumor-derived EVs can deliver parental materials to adjacent cells and contribute to oncogenic reprogramming, their functionality in energy metabolism is not well addressed. In this study, we established prostate cancer cells PC-3AcT resistant to cellular death in an acidic culture medium driven by lactic acid. Quantitative proteomics between EVs derived from PC-3 and PC-3AcT cells identified 935 confident EV proteins. According to cellular adaptation to lactic acidosis, we revealed 159 regulated EV proteins related to energy metabolism, cellular shape, and extracellular matrix. These EVs contained a high abundance of glycolytic enzymes. In particular, PC-3AcT EVs were enriched with apolipoproteins including apolipoprotein B-100 (APOB). APOB on PC-3AcT EVs could facilitate their endocytic uptake depending on low density lipoprotein receptor of recipient PC-3 cells, encouraging increases of cellular proliferation and survival in acidic culture media via increased activity and expression of hexokinases and phosphofructokinase. The activation of recipient PC-3 cells can increase glucose consumption and ATP generation, representing an acquired metabolic reprogramming into the Warburg phenotype. Our study first revealed that EVs derived from prostate cancer cells could contribute to energy metabolic reprogramming and that the acquired metabolic phenotypic transition of recipient cells could favor cellular survival in tumor microenvironment.

Keywords: cancer; exosomes; extracellular vesicles; metabolism; proteomics.

Graphical abstract

Fig. 1

Isolation and characterization of PC-3 and PC-3AcTEVs.A, Western blotting showing increased expression of HK2 and PFKP with activated ERK signaling pathway (HK2 and PFKP are rate-limiting enzymes in glycolysis). B, expression levels of mitochondrial OXPHOS proteins not changed in PC-3AcT cells. C, diagram showing a schematic workflow of the isolation of EVs for proteomics. D and E, NTA showing size and released number of EVs in PC-3 and PC-3AcT EVs. F, Western blotting showing relative enrichment of canonical EV marker proteins such as Syntenin-1, CD9, and CD81 but depletion of non-EV protein cytochrome c. Western blotting analyses were conducted using independent biological replicate of isolated EVs, distinct from those used in proteomics. G, TEM images showing a spherical round shape of PC-3 and PC-3AcT EVs consistent with sizes measured by NTA.

Fig. 2

Quantitative proteomics of PC-3 and PC-3AcT EVs.A, comparison of identified EV proteins with Vesiclepedia (www.microvesicles.org). Venn diagram indicated that the majority of both EV proteins were previously known EV proteins. B, most of top 100 EV proteins frequently identified in Vesiclepedia are included in our EV proteome. C, heatmap showing enrichment of EV marker proteins related to endosome, tetraspanins, membrane-binding proteins, and EV-associated proteins, which are canonical EV proteins by MISEV guidelines (36). D, differentially regulated proteins in PC-3 and PC-3AcT EVs are indicated in a volcano plot. Representative proteins validated by Western blotting are indicated by a green arrow. E, Western blotting showing differential expression of proteins including CTNNB1, PKM, FN1, and TSG101 as quantified in proteomics. Additionally, CTNNB1 was further validated in independently isolated EVs to confirm its differential expression observed in quantitative proteomics (supplemental Fig. S4). Western blotting analyses were performed using independent biological replicate of isolated EVs, separate from those used in proteomics. F, GO analyses for biological process, cellular component, and molecular function in significant and nonsignificant regulated proteins between PC-3 and PC-3AcT EVs (p-value <0.1).

Fig. 3

Enrichment of glycolysis-related enzymes in PC-3 and PC-3AcT EVs.A, glycolysis pathway was indicated by the blue letter of identified EV proteins. B, a scatter plot showing quantitative values of proteins in PC-3 and PC-3AcT EVs. Identified glycolytic enzymes in EVs are indicated by a red color–filled circle. Note that proteins identified only in PC-3 and PC-3AcT EVs are not plotted. C, heatmap showing abundance enrichment of glycolysis-related enzymes in EVs.

Fig. 4

Metabolic reprogramming of PC-3 cells treated by PC-3AcT EVs.A, MTT assay showing significantly increased cellular proliferation of PC-3 cells treated by PC-3AcT EVs in acidic culture media induced by lactic acid. Control is PC-3 cells without PC-3AcT EV treatment. B, HK1, HK2, and PFKP are overexpressed in reprogrammed PC-3 cells by PC-3AcT EVs. C–E, HK activity, glucose consumption, and ATP generation are significantly increased in reprogrammed PC-3 cells by PC-3AcT EVs. p-values (∗∗∗∗ < 0.0001; ∗∗∗ < 0.001, ∗∗ < 0.01, ∗ < 0.05).

Fig. 5

Transcriptomics of PC-3 and PC-3 treated by PC-3AcT EVs (EV-treated PC-3).A, significantly altered mRNA were selected with a fold change greater than 2 and a p-value less than 0.05. In volcano plots, significantly downregulated mRNA was indicated with blue numbers, and upregulated mRNA with red numbers. B, a total of 792 significantly regulated proteins was analyzed by GO biological process and KEGG pathway analysis using the DAVID bioinformatics database (davidbioinformatics.nih.gov). C, differentially regulated EV proteins were mapped in KEGG HIF-1 signaling pathway annotation. D, glycolysis pathway–related mRNA is mapped based on their fold change.

Fig. 6

Increased uptake of PC-3AcT EVs and regulated proteins related to their uptake by PC-3 cells.A, far red–labeled EVs at 2.0 × 10⁹ particles/ml were used to treat PC-3 cells and their taken fluorescent EVs were measured by fluorescence microscope. PC-3AcT EVs showed better uptake than PC-3 EVs in PC-3 cells. Please note that fluorescent spots that exceed 1 μm represent clusters of labeled EV-derived components or aggregates in cellular compartments (e.g., endosomes or lysosomes). B, differentially regulated EV proteins were mapped in KEGG ECM–receptor interaction annotation. Collagen COLA1, FN1, and other ECMs were overexpressed in PC-3AcT EVs. C, protein–protein interaction network among 1.5-fold upregulated PC-3AcT EV proteins was mapped in the BioGRID protein interaction database and their interactions were visualized by Cytoscape (cytoscape.org). EV proteins were grouped based on their main functionality and related proteins as indicated by a dashed line box. p-values (∗∗∗∗ < 0.0001).

Fig. 7

Endocytic uptake of PC-3AcT EVs by vesicular APOB recognized by cellular LDLR.A, DiR-labeled EVs at 1.0 × 10⁹ particles/ml were used to treat PC-3 cells with or without dynasore, an endocytosis inhibitor, and fluorescent EVs taken were measured by fluorescence microscope. Increased PC-3AcT EV uptake in PC-3 cells was inhibited by dynasore treatment. B, anti-LDLR neutralizing antibody treatment to PC-3 cells inhibited the uptake of PC-3AcT EVs at a similar extent to that obtained in the presence of dynasore, implying that this endocytic uptake depended on vesicular APOB recognition by cellular LDLR. Please note that fluorescent spots that exceed 1 μm represent clusters of labeled EV-derived components or aggregates in cellular compartments (e.g., endosomes or lysosomes). C, MTT assay showing that increased cellular proliferation of reprogrammed PC-3 cells by PC-3AcT EVs was inhibited by the presence of dynasore, representing that APOB-carrying EVs were functional regulators for energy metabolic reprogramming in recipient PC-3 cells. D, staining for fluorescent organelle trackers with taken DiR-labeled PC-3AcT EVs revealed that PC-3 cells could concentrate ingested EVs in the lysosome and to a lesser extent in the ER. p-values (∗∗∗∗ < 0.0001; ∗∗∗ < 0.001; ∗∗ < 0.01).

Fig. 8

A scheme for energy metabolic reprogramming of prostate cancer PC-3 cells mediated by PC-3AcT EVs. PC-3AcT EVs could reprogram recipient PC-3 cells taken by vesicular APOB recognition by cellular LDLR, resulting in the activation of HK and PFKP to promote the activation of glycolysis favoring ATP generation caused by increased glucose consumption.