Downregulation of low-density lipoprotein receptor mRNA in lymphatic endothelial cells impairs lymphatic function through changes in intracellular lipids

Abstract

Rationale: Impairment in lymphatic transport is associated with the onset and progression of atherosclerosis in animal models. The downregulation of low-density-lipoprotein receptor (LDLR) expression, rather than increased circulating cholesterol level per se, is involved in early atherosclerosis-related lymphatic dysfunction. Enhancing lymphatic function in Ldlr^-/- mice with a mutant form of VEGF-C (VEGF-C 152s), a selective VEGFR-3 agonist, successfully delayed atherosclerotic plaque onset when mice were subsequently fed a high-fat diet. However, the specific mechanisms by which LDLR protects against lymphatic function impairment is unknown. Methods and results: We have thus injected wild-type and Pcsk9^-/- mice with an adeno-associated virus type 1 expressing a shRNA for silencing Ldlr in vivo. We herein report that lymphatic contractility is reduced upon Ldlr dowregulation in wild-type mice only. Our in vitro experiments reveal that a decrease in LDLR expression at the mRNA level reduces the chromosome duplication phase and the protein expression of VEGFR-3, a membrane-bound key lymphatic marker. Furthermore, it also significantly reduced the levels of 18 lipid subclasses, including key constituents of lipid rafts as well as the transcription of several genes involved in cholesterol biosynthesis and cellular and metabolic processes. Exogenous PCSK9 only reduces lymphatic endothelial-LDLR at the protein level and does not affect lymphatic endothelial cell integrity. This puts forward that PCSK9 may act upon lymphatic muscle cells to mediate its effect on lymphatic contraction capacity in vivo. Conclusion: Our results suggest that treatments that specifically palliate the down regulation of LDLR mRNA in lymphatic endothelial cells preserve the integrity of the lymphatic endothelium and sustain lymphatic function, a prerequisite player in atherosclerosis.

Keywords: Atherosclerosis; LDLR; Lymphatic System; Lymphatic dysfunction; PCSK9.

Figure 1

Lymphatic function assessment in mice following a specific knockdown of LDLR in murine endothelial cells. Two weeks after the intraperitoneal injection of an adeno-associated virus type 1 (AAV1) containing a shRNA targeting LDLR, skin draining lymph nodes were harvested, digested and analyzed by flow cytometry to assess lymphatic endothelial cell (LEC)-specific knockdown of LDLR expression in mice. Membrane-bound LDLR expression was measured on (A) CD45^-CD31⁺Podoplanin⁺ and (B) CD45⁺ cells. Fluorescence minus one (FMO) control was used for LDLR expression, as depicted by the dotted line in the first histogram. White histogram, shSCR and dark histogram, shLldlr. LDLR expression was also determine by immunoblotting in (C) aorta isolated from female mice and (D) in liver from wild-type and Pcsk9^-/- male (full dot) and female (empty dot) mice. ShSCR in black and ShLdlr in grey. Total plasmatic cholesterol was measured following FPLC in each liproprotein subfraction in (E) wild-type and (F) Pcsk9^-/- female (dotted lines) and male (solid lines) mice treated with shSCR (black lines) and shLdlr (grey lines). Lymphatic contraction capacity was assessed by fluorescent in vivo imaging in (G) wild-type and (H) Pcsk9^-/- mice. Black histogram, ShSCR and grey histogram, ShLdlr. n=3-9. Statistics **p < 0,01.

Figure 2

Assessment of PCSK9 and LDLR expression in human LEC. (A) PCSK9 expression was measured by immunoblotting in either human LEC, HepG2 or HEK 293T cell lysates. (B) ELISA was used to measure PCSK9 levels in the cell culture supernatant of Huh7 (grey), HEK 293T (white) and human LEC (black). (C) Expression of LDLR was detected in protein lysates by immunoblotting of human LEC. Huh7 cells were used as a positive control and HEK 293T cells treated with siLDLR were used as a negative control. LDLR protein expression on human LEC was measured by (D) flow cytometry after extracellular staining of LEC (black line) and Huh7 (grey line) and by (E) immunofluorescence (Blue, DAPI; red, cholera toxin; green, anti-LDLR; yellow, colocalized voxels; scale bar, 20 µM). (F) Scatterplot of red and green pixel intensities of cholera toxin (red) and anti-LDLR (green) in human LEC. n = 4-9. Statistics: ****p < 0.0001.

Figure 3

LDLR regulation in human LEC. (A) Cell surface expression of LDLR was measured by flow cytometry after incubating human LEC with (ctl) or without FBS-containing media for 24 h. (B) Human LEC were incubated with Dil-LDL for 4 h and LDL uptake was measured by flow cytometry (dotted line, no LDL; grey line, Dil-LDL). (C) The percentage of LDL internalization compared to control as well as (D) cell surface expression of LDLR was then quantified by flow cytometry. Human LEC were incubated with exogenous human recombinant PCSK9 (6.5 µg/mL) for 16 h and LDLR protein expression was detected and quantified by (E) immunoblotting,(F) flow cytometry and (G) LDLR messenger RNA by qPCR (H) Immunofluorescence confirmed LDLR expression in human LEC. Representative scatterplot of red and green pixel intensities of cholera toxin subunit B (red) and anti-LDLR (green) in lymphatic endothelial cells and a reconstruction of colocalized voxels (scale bar, 20 µM). (I) Protein expression of VEGFR-3 was measured in human LEC by flow cytometry following a 16 h incubation with 6.5 µg/mL human recombinant PCSK9. n=3-9. Statistics: *p < 0.05, ***p < 0.001, ****p < 0.0001.

Figure 4

Human LEC-LDLR modulation at the mRNA level. Human LEC were transfected for 48 h with 25 nM of non-targeting siRNA (called siCtl) or siLDLR. The siRNA off-target effect of was verified by two siRNA sequences, herein called siLDLR2 and siLDLR3. Messenger RNA expression was assessed by qPCR in human LEC treated with either (A) siLDLR2 or (B) siLDLR3 or their respective control (siCtl). LDLR protein expression was measured by (C) immunoblotting and (D) flow cytometry. (E) LDLR, PROX-1, PDPN (podoplanin) and VEGFR-3 mRNA expression in human LEC was assessed by qPCR. (F) Representative scatterplot of red and green pixel intensities of cholera toxin (red) and anti-LDLR (green) in lymphatic endothelial cells and a reconstruction of colocalized voxels was performed for immunofluorescence experiments (scale bar, 20 µM). Protein expression of (G) VEGFR-3, (H) LYVE-1 and (I) podoplanin were measured by flow cytometry. (J) Percentage of cells in the S phase of the cell cycle. n = 3-10. Statistics: *p < 0.05, ****p < 0.0001.

Figure 5

Untargeted lipidomics in LDLR siRNA-treated human LEC. Human LEC were treated in vitro with 25 nM of siLDLR or siCtl for 48 h and MS-based lipidomics analysis of cellular extracts from siLDLR and control-treated cells was performed. (A) Total cholesterol was quantified and (B) 2139 MS features were obtained using as criteria of selection a p-value of 0.05 (y-axis where p-values are expressed as -LOG10) and a FC of 0.8 and 1.25 (x-axis where FC are expressed as LOG2), and (C) 18 unique lipid (sub)classes discriminating siLDLR- and siCtl- treated cells were further identified using MSMS analysis. Each dot represents a log2-tranformed siLDLR/siCtl signal intensity ratio. (D) Extracellular vesicles produced by human LEC in vitro were quantified in supernatant by flow cytometry and (E) their total cholesterol content was measured following a 30 min incubation with a detergent (Triton X-100) by spectrophotometry. EV, extracellular vesicles; CE, cholesteryl esters; Cer, ceramides; GPs, glycerophospholipids; PC, phosphatidylcholine; PE(O-), phosphatidylethanolamine (plasmanylethanolamine); SLs, sphingolipids; TGs, triglycerides. “_” in TGs refers to acyl chains for which the position remains to be ascertained. n = 3-9. Statistics: ***p < 0.001, threshold P-value < 0.01.

Figure 6

siRNA targeting LDLR expression modulates genes related to cholesterol biosynthesis and cellular and metabolic processes. Human LEC were transfected for 48 h with 25 nM of non-targeting siRNA (siCtl) or siLDLR and transcriptomic analysis was performed. (A) The efficiency of the treatment to decrease LDLR is represented. (B) Volcano plot of transcriptomic analysis performed in human LEC. Y axis = p value (-LOG10) of the gene expression, X axis = fold change siLDLR/siCtl (LOG2) of gene expression. Each point represents an individual gene. Significant upregulated and downregulated genes in siLDLR versus siCtl are depicted in red and blue, respectively. Using the R cluster Profiler package, function enrichment within significant genes (padj lower than 0.05 and fold change higher than 1.5), the following observed biological processes and enriched categories (using FDR adjustement to filter for 0.05 qValue cutoff) between siLDLR and siCtl was obtained. Biological processes implicated for (C) all the upregulated genes and (D) downregulated genes. (E) Protein-protein interaction network performed with the Search Tool for the Retrieval of Interacting Genes (STRING) database for all upregulated genes showed that cholesterol biosynthesis is the main biological process that is upregulated. n = 3. Statistics: *p < 0.05.