Global lipid remodelling by hypoxia aggravates migratory potential in pancreatic cancer while maintaining plasma membrane homeostasis

Abstract

Hypoxia plays an important role in pancreatic cancer progression. It drives various metabolic reprogramming in cells including that of lipids, which in turn, can modify the structure and function of cell membranes. Homeostatic adaptation of membranes is well-recognized, but how and if it is regulated in hypoxic pancreatic cancer and its relation to aggressive phenotype and metastasis remains elusive. Here we show hypoxia-induced extensive global lipid remodelling spanning changes in lipid classes, unsaturation levels, glyceryl backbone and acyl chain lengths. No major modulation of plasma membrane biophysical properties revealed a decoupling of lipidome modulation from membrane properties under hypoxia. This was supported by observing minor changes in the lipidome of plasma membranes under hypoxia. Further, hypoxia increased migration and invasion underpinned by reduced actin volume, cell cortical stiffness and facile tether dynamics. In conclusion, we demonstrate buffering of the lipidome alterations leading to a homeostatic membrane response. These findings will help to understand the hypoxic regulation of pancreatic membrane homeostasis and identify tangible theranostic avenues.

Keywords: Cell membranes; Hypoxia; Lipidomics; Membrane biophysics; Pancreatic cancer.

Fig. 1

(A) Volcano plot depicting the significant fold change of different lipid species in PANC-1 cells cultured in the absence and presence of hypoxia in three independent experiments. Lipids with a min. fold change of 2 and adjusted P-value <0.05 are displayed in a colour-coordinated manner. (B) Mol% total lipid abundance at the lipid class level shows differential distributions. Data represent the average mol% within total lipid abundance for each lipid class ± standard error of the mean of three independent experiments. PC, LPE and TG in panel B represents both Ester- and Ether-linked lipids. (C) Heatmap showing a distinct pattern in LC-MS-based protein label free quantification (LFQ) intensity of selected targets in PANC-1 control and hypoxia induced cells for three independent experiments. Proteins with a min. log₂ fold change of 1.5 and P-value <0.05 are displayed. (D) Lipid droplet staining with Bodipy showed a higher number of LDs in the cancer cells in the presence of hypoxia compared to normoxia and normal cell line HPNE. Data for LDs are presented as mean ± SEM of two independent experiments. Total 20 fields of view were taken per replicate and 60 cells were considered (Inset showing zoomed in lipid droplets; Scale bar: 25 μm). PANC-1 and HPNE are represented by blue and light green colour codes, respectively. Bar plots with patterns represent hypoxic conditions. (E) The mol% of each glycerol backbone linkage within lipid class. Alkyl ether (O) and plasmalogen (P) were combined as ether (E) lipids. (F) The mol% of lipids with varying lengths in their FA chain components (SCFA: small chain length fatty acid; <20, MCFA: medium chain length fatty acid; C20-C39, LFCA: long chain length fatty acid; C40-C60) within lipid category under control and hypoxia conditions. In all the cases, bar plots with patterns represent hypoxic conditions. Statistical analysis was determined using a student’s t-test (* P value <0.05, ** P value <0.01, and ***P value <0.001). To control false discovery two-stage linear step-up method of Benjamini, Krieger, and Yekutieli were performed, with Q = 1 %.

Fig. 2

(A) Heatmap depicting the Mol% total lipid abundance distribution at the lipid category level as per the double bond index (0,1,2,3,4,5 or >6 bonds), colour coded by row with green indicating high mol% abundance and red indicating low mol% abundance. (B) Mol% total lipid abundance distribution at the lipid class level of PANC-1 cells were summed up as per double bond index (0,1,2,3,4,5,6 & >6 bonds) and plotted using graphpad prism 8. In all the cases, bar plots with patterns represent hypoxic conditions. Statistical significance was determined by ANOVA followed by Bonferroni’s multiple comparison test, with multiplicity adjusted P value <0.01. * Adjusted P value <0.01, ** adjusted P value <0.001, and *** adjusted P value <0.0001.

Fig. 3

(A) GP imaging (N = 3, n = 70) of di-4-ANEPPDHQ-labeled cell membranes of PANC-1 cell line and the corresponding GP values. (B) Mol% total lipid abundance (N = 3) at the lipid class level of the plasma membrane (PM), after removing TAGs and CEs, of PANC-1 cells cultured in the absence and presence of hypoxia. (C) Elastic moduli distribution of PANC-1 cells in the presence and absence of hypoxia, calculated using in a total of 200 force curves in two independent experiments. (D) Representative isosurface and Phalloidin-TRITC images (N = 3, n = 50) were used to quantify actin abundance. The bar graph represents the volume of actin per unit area in control and hypoxia-induced cells. Scale bar, 10 μm; 63× oil objective. (E) Mol% PC lipid subclass abundance distribution at the glycerol backbone linkage level of PM extracted lipid. Alkyl ether (O) and plasmalogen (P) were together abbreviated as ether lipids. (F) FLIM imaging (N = 2, n = 30) of PANC-1 cells with FliptR dye showed a decrease in average fluorescence lifetime after 24 h of hypoxia induction. (G) Relative frequency distributions of membrane tether forces, (H) tether number, and (I) tether length in the presence and absence of hypoxia. The mean values are provided in the bar graph. In all the cases, bar plots with patterns represent hypoxic conditions, and plasma membrane lipidomics data of PANC-1 is given in Magenta colour. The significance test was performed using a student’s t-test (*P < 0.01, two-tailed Student’s t-test). To control false discovery two-stage linear step-up method of Benjamini, Krieger, and Yekutieli were performed, with Q = 1 %.

Fig. 4

(A) Representative images of transwell migration based on the visualization and quantification of the number of migratory cells. (B) Transwell matrigel invasion assay to show the invasive potential of hypoxia-treated PANC-1 cells. The significance test of two independent experiments of 4 randomly selected field per replicates were performed using a student’s t-test (*P < 0.01, two-tailed Student’s t-test). In all the cases, bar plots with patterns represent hypoxic conditions.