Identification of a hormone response element that mediates suppression of APOF by LXR and PPARα agonists

Abstract

Apolipoprotein F (ApoF) regulates cholesteryl ester transfer protein activity. We previously observed that hepatic APOF mRNA levels are decreased by high fat, cholesterol-enriched diets. Here we show in human liver C3A cells that APOF mRNA levels are reduced by agonists of LXR and PPARα nuclear receptors. This negative regulation requires co-incubation with the RXR agonist, retinoic acid. Bioinformatic analysis of the ~2 kb sequence upstream of the APOF promoter identified one potential LXR and 4 potential PPARα binding sites clustered between nucleotides -2007 and -1961. ChIP analysis confirmed agonist-dependent binding of LXRα, PPARα, and RXRα to this hormone response element complex (HREc). A luciferase reporter containing the 2 kb 5' APOF sequence was negatively regulated by LXR and PPARα ligands as seen in cells. This regulation was maintained in constructs lacking the ~1700 nucleotides between the HREc and the APOF proximal promoter. Mutations of the HREc that disrupted LXRα and PPARα binding led to the loss of reporter construct inhibition by agonists of these nuclear receptors. siRNA knockdown studies showed that APOF gene regulation by LXRα or PPARα agonists did not require an interaction between these two nuclear receptors. Thus, APOF is subject to negative regulation by agonist-activated LXR or PPARα nuclear receptors binding to a regulatory element ~1900 bases 5' to the APOF promoter. High fat, cholesterol-enriched diets likely reduce APOF gene expression via these receptors interacting at this regulatory site.

Keywords: Apolipoprotein F; Apolipoproteins; Cholesteryl ester transfer protein; Gene expression; Nuclear receptors/lipid ligands; Transcription.

Figure 1 –. Effect of nuclear receptor ligands on APOF expression.

Panels A and B: Confluent C3A cells were incubated for 2 hr with 10 μM GW3965, 300 μM clofibrate (CFB), 25 μM 22(R)-hydroxycholesterol (22(R)HC), 25 μM 22(S)-hydroxycholesterol (22(S)HC), 0.1 μM GW7647 or 10 μM chenodeoxycholic acid (CDCA) ± 10 μM 9-cis retinoic acid (RA) as indicated. DMSO vehicle was added to control cells receivin no agonist. APOF mRNA was quantified by qPCR. Mean ± SD (n=3). Results are typical of 2-3 similar experiments. ** - significantly different (p< 0.02) from cells treated ± RA only. Panel C: C3A cells were grown to confluence, washed with PBS, and then switched to Opti-MEM media containing 9-cis retinoic acid plus the indicated agonist. After 48 hr, conditioned media was collected and concentrated 10-fold by cold acetone precipitation. Concentrated media was fractionated on 4-20% SDS PAGE gels and ApoF protein detected by immunoblot. Panels D and E: Expression level of the indicated gene in cells incubated with 9-cis retinoic acid ± the indicated agonist as described in above. * - p< 0.05, ** - p<0.02 versus control.

Figure 2 –. APOF mRNA turnover in agonist-treated C3A cells.

C3A cells were treated with 1 μg/ml actinomycin D plus retinoic acid and the indicated agonist as described in the legend for Figure 1. At the indicated time, cells were harvested and their content of APOF mRNA determined by qPCR. Mean ± SD (n = 2-4). Results are typical of 2 experiments. 22-HC, 22(R)-hydroxycholesterol

Figure 3 –. APOF HRE complex sequence and reporter constructs.

Panel A: Nucleotide sequence of the predicted nuclear hormone response element complex (HREc). Predicted binding half-sites are highlighted with red boxes. DR – direct repeat. DR1 and DR2 – PPAR binding sites, DR4 – LXR binding site. Panel B: – Luciferase promoter reporter constructs. The APOF promoter (PR) resides at nucleotides −98 to −46 and the predicted HREc is at nucleotides −2007 to −1961 relative to the transcription start site. Schematics illustrate the portions of 5’ APOF sequence incorporated into each pGL3 luciferase construct. Constructs containing APOF sequence extend 70 nucleotides beyond the transcription start site but do not include the APOF translation start site.

Figure 4 –. Chromatin immunoprecipitation (ChIP) assay.

Panel A: ChIP analysis of the SREBF1 promoter element from C3A cells treated with 10 μM GW3956 + 10 μM RA (LXRα IP) and the HMGCR promoter element from cells treated with 300 μM clofibrate + RA (PPARα IP) versus control cells receiving DMSO vehicle (IgG IP). Mean ± SE, n=3. Panels B and C: ApoF ChIP analysis in cells treated with RA + GW3956 or clofibrate plus 10 μM RA. Panel B shows an agarose gel of ChIP PCR products. NC is a negative control DNA sequence. In panel C, qPCR quantification of APOF HREc and NC sequence in the indicated immunoprecipitates. As indicated, some C3A cells were transiently transfected with plasmid expressing LXRα or PPARα prior to agonist treatment (n=4). Panel D: Fold enrichment of LXR and PPARα in the APOF HREc from cells incubated with agonist as indicated. (n=3). Panel E: Fold enrichment of RXRα in the APOF HREc after agonist treatment. (n= 3) * - p< 0.05, ** - p<0.02 versus control. NC – negative control sequence, RA - retinoic acid, CFB - clofibrate.

Figure 5 –. Regulation of APOF reporter constructs by nuclear receptor ligands.

Cells were transiently transfected with the p2169-Luc reporter (Panels A and B), or the p394-Luc reporter, (Panel C). Construct details are shown in Figure 3. After 24 hr, cells were washed and then treated with 10 μM retinoic acid (RA), 10 μM GW3965 or 300 μM clofibrate (CFB) as shown. After 48 hr, cells were lysed and luciferase activity was determined. Data are shown for C3A cells (Panel A) and HEK293 cells (Panels B and C). Results are typical of at least 2 experiments. Mean ± SD (n=4, (Panels A and B) or n=2 (Panel C)). * - p< 0.05, ** - p< 0.02 versus control cells.

Figure 6 –. Electromobility shift assay.

Panel A: Analysis of LXRα and PPARα binding to the APOF HREc. HEK293 cells were transiently transfected with expression plasmids coding for LXRα, PPARα or RXα. Double stranded, ³²P-labeled DNA oligonucleotide probes were incubated with nuclear extracts (3 μg protein) from these cells ± unlabeled probe (200-fold), ± LXRα or ± PPARα antibody (5 μg), as indicated, at room temperature. Controls received nuclear extract from non-transfected HEK293 cells. After 30 min, samples were separated on 5% acrylamide gels and dried gels exposed to x-ray film overnight at −70°C. Results are typical of 3 experiments. Panel B: LXRα and PPARα binding to the native and mutated APOF HREc. See Panel A for experimental details. Results are typical of 2 experiments. Several lanes from the original gel were omitted for a clearer presentation. Sequences of the native, mutated and idealized HREc probes are shown in Table 2.

Figure 7 –. Luciferase promoter assay of native and mutated APOF reporter constructs.

HEK293 cellswere transiently transfected and treated with agonists as described in Figure 5. Panel A: Luciferase activity in HEK293 cells transfected with native or mutated p394-Luc reporter construct. For the mutant, the HREc sequence (see Figure 3) was mutated to TGGGGAATGGACAGTTTGACAGTATGGGACAATGGTGAGATGGACA. Data for both native and mutant constructs are expressed relative to that measured in the native construct without agonist. Note that the y-axis scales are different. Results are typical of 2 experiments. ** - p< 0.02 versus native construct without agonist, ^# - p< 0.01 versus mutant p394-Luc construct without agonist. Panel B: Luciferase activity in control cells and in cells transfected with the indicated pGL3 construct was measured. See Figure 3 for construct details. ** - p< 0.02 versus p0-Luc construct, # - p< 0.01 versus the p394-Luc construct. Results are typical of 2 experiments.

Figure 8 –. Effect of LXRα and PPARα knockdown on gene expression.

C3A cells were treated with control, LXRα or PPARα siRNA as described in the Methods. Panels A and B: LXRα or PPARα mRNA levels following siRNA treatment. Panels C and D: Target gene mRNA levels in siRNA-treated cells in response to exposure to GW3965 or clofibrate, as indicated. Panel E: Agonist-induced alterations of APOF mRNA levels in siRNA treated cells. Panels C-E: Agonist-treated cells received 10μM RA (control) or RA plus 10 μM GW3965 or 300 μM clofibrate. Mean ± SEM, n=3. * - p< 0.05 and ** - p< 0.02 versus control siRNA treated cells. ns-not significantly different from control. cntrl – control.