Common and Differential Traits of the Membrane Lipidome of Colon Cancer Cell Lines and their Secreted Vesicles: Impact on Studies Using Cell Lines

Abstract

Colorectal cancer (CRC) is the fourth leading cause of cancer death in the world. Despite the screening programs, its incidence in the population below the 50s is increasing. Therefore, new stratification protocols based on multiparametric approaches are highly needed. In this scenario, the lipidome is emerging as a powerful tool to classify tumors, including CRC, wherein it has proven to be highly sensitive to cell malignization. Hence, the possibility to describe the lipidome at the level of lipid species has renewed the interest to investigate the role of specific lipid species in pathologic mechanisms, being commercial cell lines, a model still heavily used for this purpose. Herein, we characterize the membrane lipidome of five commercial colon cell lines and their extracellular vesicles (EVs). The results demonstrate that both cell and EVs lipidome was able to segregate cells according to their malignancy. Furthermore, all CRC lines shared a specific and strikingly homogenous impact on ether lipid species. Finally, this study also cautions about the need of being aware of the singularities of each cell line at the level of lipid species. Altogether, this study firmly lays the groundwork of using the lipidome as a solid source of tumor biomarkers.

Keywords: cell lines; colorectal cancer; extracellular vesicles; lipid biomarkers; lipidomics; plasmalogens).

Figure 1

Analysis of the main membrane lipid classes of the commercial cell lines analyzed. (A) PCA using membrane lipid levels expressed as % of total membrane lipids. Explained Variance = 83.4%; (B) Loading plot after PCA of the main membrane lipid classes. For clarity, only the most influential species are indicated in each variable PCA analysis; (C) Membrane lipid composition. Values are expressed as % of total membrane lipids (mean ± SD), n = 3–6. Statistical significance was assessed using one-way ANOVA followed by Bonferroni post-test. For clarity, only statistical differences between primary and cancer cells are represented. The asterisk (*) indicates a significant difference between cancer cell lines and the primary cell line. * p < 0.05; ** p < 0.01; *** p < 0.001. Detailed results showing all comparisons are included in Table S1.

Figure 2

Cell lipidome segregates cell lines according to their malignancy. (A) PCA using the levels of all lipid species expressed as % of total lipid class. Explained variability 54.6%; (B) Loading plot after PCA of the main membrane lipid classes. For clarity, only the most influential species are included.

Figure 3

Membrane lipid fingerprint of primary, in situ, and metastatic cancer cell lines. Bar diagrams comparing changes in lipid composition of (A) PC, (B) PE, (C) PE plasmalogens, (D) PI, (E) PS, (F) SM, (G) Cer, and (H) HexCer at the molecular species level in primary, HT29, LS174t, SW480, and Colo 201 cell lines. Values are expressed as percentage of total fatty acid (mole %) and represent mean ± SD, n = 3–6. Statistical significance was assessed using one-way ANOVA followed by Bonferroni post-test. For clarity, only significance with respect to primary cells are expressed, * p < 0.05; ** p < 0.01; *** p < 0.001; and only species accounting for <5% of total membrane lipid class are included in the graph. Detailed results of all comparisons and all lipid species are included in Table S2.

Figure 4

Specific shift of PC and PE molecular species to sn-1 saturated /sn-2 AA or DHA - containing PE plasmalogens in cancer cells. The distribution of the total amount of a particular fatty acid combination within each membrane phospholipid class was evaluated. (A) 38:4 and 40:6-containing phospholipids; (B) 38:5 and 40:7-containing phospholipids; (C) 34:1 and 36:2-containing phospholipids; (D) 38:3 and 40:5-containing phospholipids; Values are expressed as a percentage of the total amount of the selected fatty acid combination (mole %) and represent mean± SD, n = 3–6. Statistical significance was assessed using one-way ANOVA followed by Bonferroni post-test. Only significance with respect to primary cells are expressed. * p < 0.05; ** p < 0.01; *** p < 0.001. Detailed results of all comparisons are included in Table S3. Minor species are included in Figure S2.

Figure 5

Protein and gene expression of ether lipid synthetic enzymes in primary and cancer colon cell lines. (A–D) Protein expression of ether lipid synthesis enzymes: FAR1 and FAR2 (fatty acyl-CoA reductases 1 and 2), AGPS (alkyl-glycerone-3-phosphate synthase), and GNPAT (glyceronephosphate O-acyltransferase) in primary (Pr) and cancer colon cell lines (Co, Colo-201; HT, HT29; LS, LS174T; SW, SW480). Values are expressed as a percentage of control and represent the mean ± SEM, n = 3–5; (E–H) Gene expression of ether lipid synthesis enzymes in primary (Pr) and cancer colon cell lines (Co, Colo-201; HT, HT29; LS, LS174T; SW, SW480). Values are expressed as a percentage of control and represent the mean ± SEM, n = 5. To assess statistical differences, one-way ANOVA and Bonferroni post-test were applied. For simplicity only significance with respect to primary cells are expressed. * p < 0.05; ** p < 0.01; *** p < 0.001. Detailed results of all comparisons are included in Table S5. Original values and densitometry values are included in Figure S4 and Table S4, respectively.

Figure 6

Membrane lipid composition of EVs isolated from colon commercial cell lines. (A) PCA of the EV lipid composition at the level of membrane lipids classes; (B) Loading plot after PCA of the major membrane lipid classes. For clarity, only the most influential species are indicated at each variables PCA analysis; (C) EV membrane lipid composition. Values are expressed as % of total membrane lipids (mean ± SD), n = 3–6. Statistical significance was assessed using one-way ANOVA followed by Bonferroni post-test. For clarity, only statistical differences between primary and cancer cells are represented. The asterisk (*) indicates a significant difference between cancer cell lines and the primary cell line. * p < 0.05; ** p < 0.01; *** p < 0.001. Detailed results showing all comparisons are included in Table S6. (D) Membrane lipid class segregation between cells and cell-derived EVs. Enrichment of lipid classes in cells or exosomes calculated as mol% of lipids in these samples.

Figure 7

Molecular species composition of the main membrane lipids of EVs isolated from commercial colon cell lines. Bar diagrams comparing levels of (A) PC, (B) PI, (C) PE, (D) PE plasmalogens, (E) PS, (F) SM, and (G) Cer at the molecular species levels in primary, HT29, LS174t, SW480, and Colo 201 cells. Values are expressed as a percentage of total fatty acid (mole %) and represent the mean ± SD, n = 3–6. Statistical significance was assessed using one-way ANOVA comparing primary to cancer cells. For clarity, only species accounting for <5% of the total lipid class were included in the graphs. * p < 0.05; ** p < 0.01; *** p < 0.001. Detailed results of all comparisons are included in Table S7.

Figure 8

Model describing the impact of the most consistent lipid changes observed in cancer cells—increase in PE plasmalogen- and AA-containing phospholipids (in particular PI)—on the Akt signaling pathway, a canonical cell differentiation and proliferation pathway. (A) In healthy cells, phosphatidylinositol-3–kinase (PI3K) phosphorylates PIP2 to PIP3, which recruits Akt directly via a PH-pleckstrin domain. Despite the lack of direct evidence indicating the preference of PI3K enzymatic for AA-containing PIP2, this specificity was shown for PI4K [68,69]; in addition, both PIP2 and PIP3 are enriched in AA [70]. Altogether, it could be speculated that PI3K, may prefer AA-containing substrates. Thus, other papers show that plasmalogens are needed to maintain Akt linked to the membrane [67,71], which is crucial for its activation via phosphorylation by PDK1 and PDK2 (among others). Once phosphorylated, Akt shuttles back to the cytosol where it phosphorylates a myriad of targets, activating downstream pathways that culminate in cell proliferation. (B)—In cancer cells, including colorectal cancer cells, PI3K and Akt are overexpressed at the protein level [72]. Therefore, the presence of high levels of AA-containing phospholipid, and plasmalogen [8,15,45,46,47] in cancer cells would provide the substrate and necessary environment to sustain enhanced and uncontrolled cell division.