Hypercholesterolemia is a strong risk factor for the development of atherosclerosis, a hallmark of cardiovascular disorders [1]. Familial hypercholesterolemia suggests that hepatic expression of LDLR, a receptor for LDL-cholesterol, is crucial for the regulation of blood cholesterol levels [2,3] . Extensive studies indicate that statins improve lipid profiles by inhibiting the activity of HMGCR, a rate-limiting enzyme of cholesterol synthesis, thereby increasing expression of LDLR [4]. Even so, statins only reduce cardiovascular risk by about 20 to 40% [5]. Diet alone successfully lowers blood LDL-cholesterol to a degree comparable to statin treatment [6]. Combination therapy using statins and a dietary regime may further reduce blood cholesterol levels [5]. Moreover, the neutral antagonist treatment of type 1 cannabinoid receptor may improve blood lipid profile [7,8].

Sterol regulatory element-binding protein 2 (SREBP-2) is a key transcription factor in cholesterol metabolism. SREBP-2 regulates the transcription of LDLR for cellular uptake of LDL cholesterol. At a high concentration of sterol, precursor SREBPs (pre-SREBPs) complex with SCAP and Insig proteins and are retained in the endoplasmic reticulum (ER) [9 ,10]. At a low concentration of sterol, Insig proteins are no longer bound to SCAP and the pre-SREBP complex, causing them to translocate to the Golgi apparatus. There, pre-SREBPs undergo proteolytic cleavage by serine proteases, resulting in liberation of the N-terminal region, which is a nuclear SREBP. Nuclear SREBP enters the nucleus and binds a sterol-regulatory element (SRE) in the promoter region of target genes such as LDLR [11-14]. SREBP-regulated gene expression has its own negative intrinsic regulatory mechanism.

Dietary soy isoflavones, including genistein, have received great attention as anti-atherogenic foods due to their lipid-improving effects [15-19], but in vivo and in vitro findings regarding their total and/or LDL cholesterol-lowering effects have been inconsistent [20-25]. Some studies report that dietary supplementation with genistein, a strong bioactive soy isoflavone, lowers total and LDL-cholesterol levels [23,24], while others describe no effect of soy isoflavones on the concentration of LDL-cholesterol even when vascular function is improved [25].

In the current study, we sought to understand the molecular actions of genistein on the expression of LDLR in hepatocytes and implications for observed hypocholesterolemic or anti-atherogenic effects. Although LDLR is expressed in nearly all tissues, liver LDLR plays a pivotal role in the clearance of LDL-cholesterol [26].

Cell culture

Human hepatoblastoma (HepG2) cells were purchased from the Korean Cell Line Bank (Seoul, Korea) and grown in 5% CO₂ at 37°C in high glucose DMEM medium (Welgene, Daegu, Korea) containing 10% (v/v) FBS (Cellgro, Manassas, VA, USA) and 5% AA (Invitrogen, Carlsbad, CA, USA). For experiments, cells were plated at a density of 1.6×10⁶ cells per 60mm dish. The next day, the medium was changed to a serum-starvation medium containing 0.5% (v/v) FBS. After overnight incubation, the cells were treated with genistein at a final concentration of 10 μM or 20 μM for the indicated periods of time. For JNK inhibition experiments, cells were pre-treated with the JNK inhibitor SP600126 (Calbiochem, Darmstadt, Germany) at a concentration of 10 μM for 30 min prior to the addition of genistein. The inhibitors 0.3 μM AEBSF (4-(2-aminoethyl)benzenesulfonyl fluoride) (Sigma, St. Louis, MO, USA), 1 μg/ml brefeldin A (Sigma, St. Louis, MO, USA), and 10 μM MG132 (Calbiochem,Darmstadt, Germany) were used in our study.

Plasmids and cloning

A previously described reporter plasmid pSRE-luc was a gift from Dr. Shimano [27]. A luciferase reporter gene plasmid containing the LDLR promoter region was constructed. A promoter fragment of the human LDLR gene from -805 to +50 relative to the transcription start site was amplified from genomic DNA extracted from HepG2 cells by PCR using the following primers: forward 5’- AACTCGAGTTGGTCTCCACCAGCTCTCT- 3’ and reverse 5’- TGAAGCTTTCACGACCTGCTGTGTCCTA -3’. The PCR product was subcloned into the XhoI and HindIII sites of the pGL3basic luciferase vector (Promega, Madison, WI, USA) to generate pLDLR-luc (-805 to +50). Using the parental clone (pLDLR-luc (-805 to +50)), a 5’ deletion construct containing the region from -171 to +50 (pLDLR (-171 to +50)) of the human LDLR gene was generated by PCR using the following primers: forward 5’ – AACTCGAGGGACTGGAGTGGGAATCAGA- 3’ and reverse 5’- TGAAGCTTTCACGACCTGCTGTGTCCTA -3’. SRE deletion constructs of pLDLRdelSRE-luc (-805 to +50) and pLDLRdelSRE (-171 to +50) were generated using the following primers: forward 5’- TGAAGACATTTGAAATGCAAACTCCTCCCCCTGCT-3’ and reverse 5’- GGGGAGGAGTTTGCATTTCAAATGTCTTCACCTCACTGC- 3’.

Transient transfection and luciferase assay

HepG2 cells were seeded in 6-well plates (Corning Costar Corp., Tewksbury, MA, USA) at 1.6×10⁵/ well 24 h prior to transfection. Transfections were performed with 0.5 μg of each DNA construct and pRLSV40 (Promega, Madison, WI, USA) using Lipofectamin2000 (Life Technologies, Inc., Grand Island, CA, USA) according to the manufacturer’s instructions. Twenty-four hours later, genistein was added to the medium. The next day, the cells were washed 3 times in PBS and lysed with 100 μl of passive lysis buffer. The clear cell lysate was used for the measurement of luciferase activity using the Dualluciferase assay system (Promega, Madison, WI, USA) and a Promega Glomax luminometer (Promega BioSystems Sunnyvale Inc., Sunnyvale, CA, USA).

RNA extraction and quantitative RT-PCR

Trizol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA from samples according to the manufacturer’s protocol. cDNA was synthesized from 1 μg of total RNA after being primed by random hexamers through reverse-transcription by ImprompII reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer’s directions. Quantitative PCR was performed on a CFX96 sequence detection system (Biorad, Hercules, CA, USA) using EvaGreen qPCR mix plus (Solis BioDyne, Estonia). Levels of expression were normalized relative to the amount of 18S rRNA. Relative mRNA levels were calculated by differences in Ct values and are expressed as the fold change.

Western blot analysis

HepG2 cells were harvested and lysed in RIPA buffer containing leupeptin, 1mM PMSF, Na₃VO₄, and 0.1% protease inhibitor cocktails (Sigma, St. Louis, MO, USA). After 30 min incubation on ice, the cells were centrifuged at 13,000 rpm for 20 min at 4°C. The supernatant was mixed with 5X sample buffer and then boiled for 10 min. Equal amounts of protein cell lysates were loaded onto an SDS-polyacrylamide gel for electrophoresis. PVDF membrane (Millipore, Billerica, MA, USA) was used to transfer proteins from the gel. Mouse anti-SREBP-2 (Santa Cruz Biotech., Santa Cruz, CA, USA), rabbit anti-LDLR (BioVision, Milpitas, CA, USA), rabbit anti-p-JNK (Merck Millipore, Billerica, MA, USA), rabbit anti-p-p38 (Merck Millipore, Billerica, MA, USA), mouse antip- ERK (Santa Cruz Biotech., Santa Cruz, CA, USA), and rabbit anti- GAPDH (Signalway antibody, Pearland, TX, USA) antibodies were used to detect the proteins.

Chromatin immunoprecipitation assay (CHIP)

Cells were crosslinked with 0.4% ice cold formaldehyde (Sigma- Aldrich, MO, USA) for 15 min at room temperature. The reaction was stopped by adding 0.2M of glycine and incubating for 5 min at room temperature. After being rinsed with cold PBS twice, the cells were lysed with 250 μl RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris (pH 8), 5 mM EDTA). The lysates were sonicated 4 times, 30s each time, with a sonicator (Sonifier 250, Branson, CT, USA). After centrifugation at 13,000 rpm for 15 min at 4°C, the clear supernatant was used for immunoprecipitation. The precleared protein lysates were gently rotated with anti-SREBP-2 antibody (Santa Cruz Biotech., Santa Cruz, CA, USA) overnight at 4°C and then rotated with a 50% slurry of blocked protein A/G sepharose beads (Santa Cruz, CA, USA) for 2h at 4°C. Samples were washed with RIPA buffer 2 times, IP buffer (100 mM Tris (pH 8), 500mM LiCl, 1% NP-40, 1% sodium deoxycholate) 4 times, RIPA buffer 2 times and TE buffer 2 times. Every wash process was carried out for 5 min at 4°C. IP material was eluted by adding 200 μl of elution buffer (70 mM Tris (pH 8), 1 mM EDTA, 1.5% SDS) in the remaining 100 μl of TE buffer (10 mM Tris, 1mM EDTA, pH 7.5) from the last wash step and incubating in a 65°C water bath for 10 min. Chromatin was reverse-crosslinked by adding 13 μl of 4M NaCl to 237 μl of the sample and incubating in a 65°C water bath for 5 h. DNA fragments were purified using a DNA extraction kit (iNtRON Biotechnology, Korea) and PCR was performed with the following primers: LDLR-SRE forward, 5’- TCCTCTTGCAGTGAGGTGAA-3’, reverse, 5’- TTTCTAGCAGGGGGAGGAGT-3’. The PCR generated a 66-bp fragment containing a sterol regulatory element (SRE) of the human LDLR promoter.

Statistical Analyses

SPSS was used to perform statistical analyses (Statistical Package for the Social Sciences; SPSS Inc., Chicago, IL, USA). One-way analysis of variance (ANOVA) followed by Duncan’s multiple comparisons test was used to determine statistically significant differences among the experimental groups. P values < 0.05 were considered statistically significant.

Genistein increases LDLR expression

LDLR protein expression was examined in HepG2 cells after incubation with varying concentrations of genistein. As the concentration increased from 1 μM to 20 μM, the protein levels of LDLR increased 1.8 fold to 2.9 fold (p<0.05) (Figure 1A). In time course experiments, a significant increase in LDLR mRNA levels was detected approximately 4 h after the addition of 10 μM genistein (Figure 1B). Significant elevation of LDLR mRNA levels was detected around 6 h of treatment with genistein (Figure 1C). Furthermore, when LDLR promoter activity was investigated using a luciferase construct containing the proximal region of the LDLR gene, genistein increased transcriptional activity of the construct (Figure 2A). These results suggest that genistein increases the expression of LDLR at the transcriptional level.

Figure 1: Genistein stimulation of LDLR expression at the transcriptional level. HepG2 cells were serum-starved and then stimulated with genistein. Protein levels of LDLR in response to 1, 5, 10 and 20 μM of genistein for 24 h (A), changes in protein (B) and mRNA levels (C) of LDLR after incubation with 10 μM of genistein within 24 h. Results are expressed as mean ± SD of at least three independent experiments. Different letters (a-c) represent significant differences (p<0.05) between all groups by ANOVA.

Figure 2: SRE-dependent effects of genistein on LDLR promoter activity.
Luciferase activities of phLDLR-luc(-805+50) and its corresponding SRE deletion construct (A) and pSRE-luc construct (B) upon the treatment of 10 or 20 μM genistein (G10 μM or G20 μM) and DMSO (G0 μM) for 24 h. CHIP assay for binding of SREBP-2 to SRE regions of the human LDLR promoter region (C). The results are expressed as mean ± SD of at least three independent experiments. Different letters (a-c) represent significant differences (p<0.05) for the comparison of all groups by ANOVA (A and B) or two groups by student’s t-test (C).

Genistein-induced expression of LDLR requires SREBP-2 transcriptional activity

We investigated whether SREBP-2 is involved in the genisteinmediated up-regulation of LDLR, based on the fact that SREBP-2 is a crucial transcription factor for the LDLR gene. In order to examine whether SREBP-2 transcriptional activity is required for genisteinmediated up-regulation of the LDLR gene, luciferase reporter constructs were generated. Genistein significantly increased luciferase activity, and this effect was counteracted by an SRE deletion construct (Figure 2A). The luciferase construct driven by SRE alone, which originated from the human SREBP-2 promoter region, responded to genistein by increasing luciferase activity (Figure 2B). In addition, the DNA binding activity of SREBP-2 upon genistein treatment was higher, as tested by CHIP (Figure 2C). These results clearly suggest that SREBP-2 transcriptional activity is required for the genistein-driven up-regulation of the LDLR gene.

Genistein promotes SREBP-2 processing

Protein levels of the mature and nuclear forms of SREBP-2 significantly increased within 6 h of genistein treatment (Figure 3A). Regulatory steps affected by genistein in the SREBP-2 maturation process were examined by the use of chemical inhibitors AEBSF (4-(2-aminoethyl)benzenesulfonyl fluoride) and brefeldin A. AEBSF significantly lowered the nuclear form of SREBP-2 in both the absence and presence of genistein (Figure 3B), suggesting that the step regulated by genistein occurs prior to proteolytic cleavage in the Golgi apparatus. Brefeldin A increased nuclear SREBP-2 levels (Figure 3C). Genistein did not further elevate nuclear SREBP-2 levels beyond the increase caused by brefeldin A (Figure 3C). In addition, Insig-1 protein levels were not down-regulated within 12 h after genistein treatment, which presumably triggered SREBP-2 processing (Figure 3D). These data suggest that genistein may promote the movement of pre-SREBP-2 from the ER to the Golgi apparatus without significantly affecting Insig-1 degradation.

Figure 3: Genistein-stimulated SREBP-2 processing.
HepG2 cells were stimulated with genistein for the indicated period of times and analyzed by Western blot. Formation of nuclear SREBP-2 over time after genistein treatment (A), the effects of AEBSF (B) or brefeldin A (C) on the formation of nuclear SREBP-2 by genistein. Protein levels of Insig-1 over the course of genistein treatment for 24 h (D).

Genistein activates the JNK signaling pathway for SREBP-2 processing

Next, we asked whether mitogen-activated protein kinase (MAPK) signaling pathways are involved in genistein-mediated LDLR expression. Genistein didn’t significantly alter levels of the phosphorylated form of p38 and pERK proteins (Figure 4A), but phosphorylation of JNK protein increased gradually over the course of the 24 h treatment (Figure 4A and B). To assess if JNK is involved in the regulation of the LDLR gene by genistein, we treated cells with a JNK specific inhibitor, SP600126. SP600126 greatly reduced the phosphorylation of JNK and LDLR protein levels that had been increased by genistein (Figure 4C and D). In addition, pretreatment with SP600126 greatly lowered the levels of nuclear SREBP-2 (Figure 4E). In order to minimize the effects of c-jun, a transcription factor responding to JNK activation, we generated a construct harboring 171 base pairs of the upstream region of the human LDLR gene that did not contain the putative c-jun binding site but did have an SRE (pLDLR-luc (-171 to +50)). We also generated an SRE deletion construct. Genistein increased luciferase activity with the construct (pLDLR-delSRE-luc (-171 to +50)), and the SRE deletion construct counteracted the effects of the genistein (Figure 4F). These results indicate that the activation of JNK signals is necessary for genistein-induced SREBP-2 maturation and, thereby, LDLR expression.

Figure 4: JNK activation by genistein involves SREBP-2 processing and LDLR expression.
Phosphorylated levels of JNK, p38 and ERK upon 10 μM genistein treatment for the indicated period of times (A) and quantification of p-JNK levels (B). LDLR and p-JNK protein levels with pre-treatment of SP600126 or vehicle DMSO (C) and quantification of LDLR protein levels (D). Protein levels of pre-SREBP-2 and nuclear SREBP-2 in cells treated with genistein for 6 or 24 h after 30 min incubation with SP600126 or DMSO (E). Luciferase activities of pLDLR-luc (-171 to +50) and pLDLRdelSRE-luc (-171 to +50) upon 24 h incubation of 10μM genistein (G10μM) or DMSO (G0μM). The results are expressed as mean ± SE of at least three independent experiments. Different letters (a-c) represent significant differences between all groups (p<0.05) by ANOVA.

Genistein didn’t up-regulate SREBP-1c mRNAs and cholesterol prevented genistein effects

Genistein didn’t significantly up-regulate SREBP-2 mRNAs within 12 hr after the addition but 24 hr treatment of genistein appear to up-regulate SREBP-2 mRNA levels (Figure 5A). Although genistein markedly increased LDLR mRNA levels, 24 h incubation with geneistein didn’t significantly up-regulate HMGCR mRNA levels (Figure 5B). SREBP-1c didn’t change its expression levels in response to genistein (Figure 5C). Together, it is likely that there are other regulatory factors affecting the expression levels of HMGCR and SREBP-1c. Then we asked if cholesterol affects the regulation of genistein on SREBP-2 processing and LDLR expression. Cholesterol lowered nuclear levels of SREBP-2 and subsequently LDLR protein levels (Figure 5D). Even when the concentration of genistein was increased to 20 μM cholesterol abolished the genistein effect on SREBP-2 processing (data not shown). These data suggest that elevation in intracellular cholesterol concentration prevent the ability of genistein to promote SREBP-2 processing.

Figure 5: Transcript levels of SREBPs and HMGCR and cholesterol effects on nuclear SREBP-2 levels in genistein-treated cells.
mRNA levels of SREBP-2 (A), HMGCR (B) and SREBP-1c (C) in cells treated with 10 μM genistein for 24 h. Protein levels of LDLR, pre-SREBP-2, nuclear SREBP-2 and GAPDH in cells incubated with 10 μM of genistein in the presence of either cholesterol (5 μg/ml) or solvent alone (0.1% ethanol) for 24 h (D).

In a previous animal study, the anti-atherogenic effects of genistein appeared to mediate LDLR based on a lack of the effects in LDLR knock-out mice [19], but the exact molecular mechanisms behind the actions of genistein were not investigated. Here, we demonstrate that genistein activates the JNK signaling pathway and increases SREBP-2 processing for SRE-dependent transcription of the LDLR gene in human hepatocytes. Because hepatic LDLR plays a crucial role in the regulation of blood LDL cholesterol, up-regulation of LDLR in hepatocytes by genistein is likely to contribute to its anti-atherogenic effects.

LDLR is regulated by various cellular stimuli such as growth factors, insulin [28] and estradiol [28,29]. SRE is required for insulin or estradiol- induced expression of the LDLR gene [28]. Maturation of SREBP by proteolytic cleavage has been implicated in regulation by growth factors such as platelet-derived growth factor [30]. Our data showed that genistein also up-regulates LDLR expression at transcriptional levels, which requires SRE-mediated transcriptional activity of SREBP-2.

The genistein-induced transcriptional activity of SREBP-2 appears to be due to an increase in the amount of nuclear SREBP-2. Nuclear localization of SREBP-2 can be regulated in at least two places: trafficking between the ER and Golgi apparatus and the proteolytic release of the nuclear form of SREBP-2 from the precursor form in the Golgi apparatus. We utilized AEBSF and brefeldin A to isolate the location of genistein-regulated SREBP-2 processing. AEBSF is a serine protease inhibitor that directly blocks the activity of Site-1 protease (S1P), a processing enzyme present in the Golgi apparatus [31]. Brefeldin A inhibits protein transport from the endoplasmic reticulum (ER) to the Golgi apparatus, resulting in the retention of Golgi proteins, including the processing enzymes, in the ER [32,33]. Co-treatment with genistein and AEBSF yielded a lack of nuclear SREBP-2, while the addition of Brefeldin A did not alter the effects of genistein, suggesting that genistein promotes pre-SREBP-2 trafficking from the ER to the Golgi apparatus. Even so, we cannot completely exclude the possibility of genistein increasing the proteolytic cleavage of pre-SREBP-2 in the Golgi apparatus. Genistein-induced SREBP-2 processing is not likely to be mediated by Insig-1 degradation, because no significant changes in Insig-1 protein levels occurred upon genistein treatment. Genistein could have induced a structural change in the SREBP-SCAP complex that weakened interactive forces with Insig proteins to initiate movement to the Golgi.

Intracellular signaling mechanisms involved in soy isoflavonemediated up-regulation of the LDLR gene have not been extensively elucidated. MAP kinase-related signals has been implicated in the transcriptional activation of SREBPs [34]. Here we provide evidence of the role of JNK activation on SREBP-2 processing and LDLR expression. JNK inhibition greatly reduced genistein-increased nuclear SREBP-2 as well as its target gene expression of LDLR, suggesting that the JNK pathway links the effects of genistein and the activation of SREBP-2 for LDLR expression. Involvement of JNK activation in LDLR expression has been previously reported in berberine-treated cells [35]. However the effect of berberine on JNK activation and LDLR expression is SREindependent and mediated by the transcriptional activity of c-jun, which is phosphorylated upon JNK activation [35]. Thus, our findings regarding JNK-dependent SREBP-2 activation of LDLR expression by genistein are novel. JNK activation is better understood in the regulation of SREBP-1c-mediated gene expression. JNK2 increases insulin-mediated activation of SREBP-1c for de novo fatty acid synthesis [34]. Another MAPK, ERK, mediates the signals initiated by insulin and platelet-derived growth factor [36]. ERKs directly phosphorylate nuclear SREBPs and lower the recruitment of SUMO-mediated HDAC, resulting in the enhancement of SREBP transcriptional activity [37].

SREBP-2 activation by genistein could result in expression of other target genes such as HMGCR and SREBP-2 itself. We observed that HMGCR mRNA levels were not significantly up-regulated by genistein, which may indicate that factors other than SREBP-2 as affected by genistein influence regulation of HMGCR and LDLR expression. SREBP-2 mRNA levels were elevated only at the time point of 24 h after genistein treatment, suggesting that a positive feedback loop of SREBP-2 enhanced SREBP-2 mRNA production. On the other hand SREBP-1c, which is under the control of sterol-regulated processing similar to that of SREBP-2, was not significantly up-regulated by genistein.

LDLR expression is controlled in a tight negative feedback mechanism by intracellular sterol levels [26]. The LDLR-lowering effects of genistein were sensitive to the presence of cholesterol. Cholesterol abolished genistein-induced SREBP-2 processing and LDLR expression. This outcome suggests that elevation of intracellular cholesterol concentration by genistein via LDLR could initiate a negative feedback mechanism to blunt the effects of the genistein. In addition, the sensitivity of genistein to the presence of cholesterol may indicate that genistein acts at the step at which sterol exerts its effect during SREBP-2 processing. Sterol prevents pre-SREBP-2 from trafficking from the ER to the Golgi apparatus by being associated with Insig-1 and SCAP proteins [38]. Because of plausible negative feedback mechanisms, the cholesterol-lowering effects of genistein could diminish over time, further limiting uptake of LDL-cholesterol elicited by genistein. This loop may explain the inconsistent ability of genistein to lower cholesterol.

In conclusion, our results demonstrate that genistein, a soy isoflavone, could act as an LDL-cholesterol-lowering agent via JNK activation and SREBP-2 processing in hepatocytes. The long-term hypocholesterolemic effects of genistein may require additional intervention to restrict the negative feedback mechanisms initiated by elevation of cholesterol levels.

We greatly appreciate SB Lee and SY Yim for their excellent technical support.