Phosphatidic Acid (PA) can Displace PPARα/LXRα Binding to The EGFR Promoter Causing its Transrepression in Luminal Cancer Cells

Abstract

The expression of the epidermal growth factor receptor (EGFR) is highly regulated in normal cells, whereas some cancer cells have high constitutive levels. Understanding naturally-occurring ways of downregulating EGFR in cancer cells was investigated. Phosphatidic acid (PA) or Nuclear Receptors (NR) PPARα/RXRα/LXRα, enhance EGFR expression, mediated by the promoter region -856(A) to -226(T). Unexpectedly, the combination of NRs and PA caused repression. PA induces a conformational change in the nuclear receptor PPARα (increase of alpha-helices at the expense of decreasing beta-sheets), as evidenced by circular dichroism. This represses the naturally-enhancing capability of PPARα on EGFR transcription. PPARα-overexpressing cells in the presence of PA > 300 nM or the enzyme that produces it, phospholipase D (PLD), downregulate EGFR expression. The reasons are two-fold. First, PA displaces PPARα binding to the EGFR promoter at those concentrations. Second, NR heterodimer-dependent promoter activity is weakened in the presence of PA in vivo. Since other genes considered (β-catenin, cyclin D3, PLD2 and ACOX-1) are also downregulated with a PA + PPARα combination, the transrepression appears to be a global phenomenon. Lastly, the reported effect is greater in MCF-7 than in MDA-MB-231 breast cancer cells, which could provide a novel basis for regulating excessive expression of EGFR in luminal cancer cells.

Figure 1. Effect of PPAR family of nuclear receptors on EGFR expression.

(A) Silencing of PPARα (endogenous) with 200 nM siRNA for four days. Effect on EGFR expression and PPARα (control of silencing) by qPCR. (B) Western blot of endogenous protein silencing. (C) Effect of transcription of 3 different nuclear receptors (1 μg DNA) PPARα, RXRα or LXRα alone or in combination, on EGFR gene expression by qPCR. (D) Detection of EGFR protein mass by Western blot of cell lysates prepared from cell transfected with plasmids as in (C). (E) Densitometric analysis of three Western blots with similar experimental conditions as the one shown in (D). (F–I) Results of gene expression analyses for four target genes: EGFR (F), PLD2 (G), cyclin D3 (H) and β-catenin (I) with RNAs derived from cells overexpressing (1 μg DNA each) PPARα + RXRα or PPARα + LXRα. The blots presented in B have been cropped to depict the region around 50 kDa; and in (D), regions around 50 kDa and 175 kDa as indicated; all gels were run under the same experimental conditions. Experiments in this figure were performed in triplicate for at least 3 independent sets in total (n = 9). Results are mean +/- SEM and are expressed in terms of gene expression. The * symbols denote statistically significant (P < 0.05) ANOVA increases between samples and controls. The # symbols denote statistically significant (P < 0.05) ANOVA decreases between samples and controls.

Figure 2. Effect of increasing concentrations of PA on EGFR gene or protein expression in the absence or presence of NRs heterodimers.

(A) COS-7 cells overexpressing (1 μg DNA each) PPARα + RXRα or PPARα + LXRα were incubated with the indicated concentrations of 1,2,-dioleyl phosphatidic acid (DOPA), after which cells were used for RNA isolation, reversed transcribed to cDNA and analyzed by Q-PCR with EGFR primers and probe. (B,C) Western blot analyses of EGFR protein expression in COS-7 (B) or MCF-7 cells (C) overexpressing nuclear receptors in the presence of increasing concentrations of DOPA. (D) Effect of PA (alone, i.e., no NRs transfected) on EGFR protein expression using Western blot analysis. (E) Cell viability (by trypan blue) of PPARα + RXRα or PPARα + LXRα transfected cells treated with increasing concentrations of DOPA. (F,G) Effect of increasing concentrations of PA on EGFR and ACOX-1 gene or protein expression in the absence or presence of NRs heterodimers. COS-7 cells overexpressing (1 μg DNA each) PPARα + RXRα or PPARα + LXRα were incubated on the plates with the indicated concentrations of DOPA for 4 hours, after which cells were used for either RNA isolation, reverse transcription to cDNA and analysis by Q-PCR with EGFR or ACOX-1 primers and probes (F) or they were used to obtain cell lysates for Western blot analyses with anti-ACOX-1 antibodies (G). The blots presented have been cropped to depict the regions around 50 kDa and 175 kDa; and regions around 34 kDa, 50 kDa, 100 kDa and 105 kDa as indicated; all gels were run under the same experimental conditions. Experiments in this figure were performed in triplicate, and statistics and symbols are as indicated in the legend to Fig. 1.

Figure 3. Effect of increasing PA concentration on PPARα-mediated protein expression of PLD2, cyclin D3 and β-catenin.

(A) Western blot analysis of PLD2, cyclin D3 or β-catenin was performed with either COS-7 or MCF-7 cells that were subjected to the same conditions as in Fig. 2B,C. (B) PLD2 protein expression detected by Western blot analysis from cells treated as indicated, with either PA alone or in combination with NRs. (C) Effect of PLD2 transfection on EGFR gene expression in the absence or presence of silencing or overexpressing PPARα. The blots presented in panels A-B have been cropped to depict the regions around 50 kDa and 175 kDa; and panel E, regions around 34 kDa, 50 kDa, 100 kDa and 105 kDa as indicated; all gels were run under the same experimental conditions. Experiments in this figure were performed in triplicate, and statistics and symbols are as indicated in the legend to Fig. 1.

Figure 4. Transactivation activity of the EGFR promoter upon biding of nuclear receptors.

(A) Schematic of EGFR promoter cloned into the pEZX-PG04 backbone. (B) EGFR promoter with putative PPAR response elements shaded in yellow and conserved bases of Response Element (RE) motifs shaded in red. (C) Luciferase assay of cells overexpressing EGFR promoter reporter plasmid (pEZX-PG04-EGF-R promoter) in the presence of co-transfected nuclear receptors. (D) Luciferase assay with TK-ACOX1-Luc, a positive control for trans-activation, to indicate validity of reagents and cells. (E) Effect of PLD2 (WT or lipase-inactive K758R mutant) and PA on the PPARα-mediated transactivation of EGFR promoter.

Figure 5. Binding and circular dichroism of PPARα in the presence of PA.

(A) Phospholipase assay was performed using recombinant PLD2 and cells overexpressing PLD2 in the presence of several phospholipids: potential PLD substrates, 1,2-dioctanoyl-sn-glycero-3-phosphocholine (DOPC); 1,2-dimirystoyl-sn-glycero-3-phosphocholine (DMPC); 1,2-diarachidonoyl-sn-glycero-3-phosphocholine (AraPC); 1-oleoyl-2 -hydroxy-sn-glycero-3-phosphatidic acid (lyso-PC) and oxidized 1-palimitoyl, 2-arachidonoyl—sn-glycero-3-phosphocholine (OxPAPC). DOPC is considered the best substrate (positive control). (B) Protein-lipid overlay assays to PVDF membranes were performed with recombinant PPARα. PIP₂ was used as a positive control for biding to PPARα and cholesterol was used as a negative control. (C–E) Protein-lipid binding by quenching of intrinsic aromatic amino acid fluorescence using the lipids: PC (C), DOPA (D) or AraPA (E) with recombinant PPARα in vitro. (F,G) Circular dichroism of PPARα upon binding to (F) AraPA (yellow circles) or lysoPA (blue circles), or (G) DOPA (red circles) or DMPA (green circles). (H) Positive control for Circular Dichroism; PPAR bound to C18:0-CoA, its strongest lignad. (I) Secondary structure analysis to ascertain the percentage of alpha helices, beta sheets, turns and undetermined structures in PPAR in the presence of PC or PA. C18:0-CoA was used as a positive control.

Figure 6. Effect of EGF on PPARα recruitment and binding to the EGR promoter in the absence or presence of PA.

(A) Effect of EGF on PPARα/RXRα/LXRα mediated transactivation of EGFR promoter. Pre-incubation with 3 nM EGF enhanced the NRs-mediated positive effect on EGFR promoter activity, especially when heterodimers are present. The presence of 300 nM PA has an overall negative effect irrespective of the absence or presence of EGF. (B) In vitro binding of dsDNA with a putative RE binding sequence to PPARα (derived from Fig. 4B). In vitro assay was performed as indicated in Materials & Method section. Control had 30 ng/well BSA instead of PPARα. PA was used at the concentration of 300 nM.

Figure 7. Effect of PA on MDA-MB-231 cells overexpressing NR heterodimers.

(A) Basal breast cancer cells MDA-MB-231 were transfected with either PPARα + RXRα or PPARα + LXRα plasmids for 2 days. Cells that were incubated with the indicated concentrations of DOPA were analyzed by Western blots for EGFR protein expression. (B) Graphical quantification of the EGFR densities shown in (A) relative to actin staining. (C) In parallel, cells treated similarly, were used for RNA extraction and Q-PCR to quantify EGFR gene expression. Experiments were performed in triplicate, and statistics and symbols are as indicated in the legend of Fig. 1. (D) Western blot analysis of cells after subcellular fractionation and ultracentrifugation, indicating that PLD2 is present mainly in the cytoplasm, but also in the nucleus. The histone-3 (H3) is a nuclear marker. Experiments in this figure were performed in triplicate, and statistics and symbols are as indicated in the legend to Fig. 1.

Figure 8. Model on how PLD/PA + NRs could repress EGFR expression.

In the absence of PLD/PA (a), PPARα-LXRα or PPARα-RXRα heterodimers exert a positive effect on EGFR promoter activity (b), as well as in cyclin D3, β-catenin and PLD2 genes. When PLD2-generated PA (c) binds to PPARα it causes conformational changes in PPARα (d) that increase the α-helices content at the expense of the β-sheets content. Either this conformational change or the presence of PA at concentrations >300 nM, diminish the ability of PPARα to bind to the EGFR promoter (e) and/or the recruitment of RXRα to form functional dimers (f) leading to a repression of EGFR expression (g).