The retinoids used in this study are grouped into three categories: (1) carotenoids including β-carotene, lycopene, and lutein; (2) classic retinoids including all-trans retinol palmitate, retinol acetate, 9-cis retinaldehyde, 13-cis retinol, 13-cis retinaldehyde, 13-cis retinoic acid, and fenretinide; (3) receptor-specific retinoids including all-trans retinoic acid (ligand for RAR), 9-cis retinoic acid (ligand for both RAR and RXR), and TTNPB (4-(E-2-[5,6,7,8-tet-rahydro-5,5,8,8-tetramethyl-2-naphthalenyl]-1-propenyl) benzoic acid) (ligand for RAR). β-Carotene, lycopene, all-trans retinol palmitate, 9-cis retinaldehyde, 13-cis retinol, fenretinide, all-trans retinoic acid, 9-cis retinoic acid, and TTNPB were purchased from Sigma-Aldrich (St. Louis, MO). Lutein was purchased from US Biological (Swampscott, MA). Retinol acetate and 13-cis retinaldehyde were purchased from Toronto Research Chemicals (North York, Canada). 13-cis retinoic acid was purchased from BIOMOL (Plymouth Meeting, PA). Retinoids were dissolved in dimethyl sulfoxide (DMSO) at 10 mM as the stock solution and stored at -80°C. Retinoids were diluted with serum-free medium to a 10 μM final concentration immediately before use. The final concentration of DMSO in the culture medium was 0.1% in all treatments. Because retinoids are light sensitive, all retinoid treatments were conducted under dim light.

Huh-7 cells were cultured in Dulbecco's Modification of Eagle's Medium and HepG2 and Hep3B cells were cultured in Minimum Essential Medium (Mediatech, Herndon, VA). The media were supplemented with 10% fetal calf serum (FBS) (Atlanta Biologicals, Lawrenceville, GA). Cells were cultured at 37°C in 5% CO₂ atmosphere with a relative humidity of 95%. Cells were plated with approximately 1 × 10⁶ cells per T-25 flask and cultured overnight. The next morning cells were rinsed with PBS to remove FBS and incubated with individual retinoids (10 μM) in serum-free media for three days or as otherwise indicated. Fresh medium containing individual retinoids was provided every 24 hours. Cell viability was determined by trypan blue exclusion counting with a hemocytometer. Every sample was counted in triplicates.

Detached cells from individual retinoid treatments were collected every 24 hours and combined at the end of the treatment. Cells were lysed with lysis buffer (50 mM Tris·Cl pH 7.4, 150 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% (V/V) NP-40 with protease and phosphatase inhibitors (Pierce, Rockford, IL)). Equal amounts of lysates (50 μg total protein) were run on SDS-PAGE and electroblotted onto PVDF membrane (Bio-Rad, Hercules, CA). The membranes were first incubated with PBS supplemented with 0.1% Tween 20 and 5% nonfat dry milk (PBST-milk) for 1 hour at room temperature to block nonspecific binding sites. Immunostaining was performed by incubating the membranes with primary antibodies for caspase-3 (Cell Signaling, Beverly, MA) or β-actin (Santa Cruz, Santa Cruz, CA) in PBST-milk overnight at 4°C. After three washes in PBST, membranes were incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies for 1 hour in PBST-milk followed by three washes. Signal was detected using the ECL system SuperSignal West Pico Chemiluminescent Substrates (Pierce, Rockford, IL).

Cells (5 × 10⁴) were plated into chamber slides (BD, Franklin Lakes, NJ) in medium supplemented with 10% FBS and cultured overnight to attach. The next morning, cells were washed with PBS and incubated with fenretinide (10 μM) in serum-free medium for 24 hours followed by TUNEL staining using an in situ cell death detection kit (Roche Applied Science, Indianapolis, IN) according to the manufacturer's instruction.

Cells (1 × 10⁶) were plated into T-25 flasks in medium supplemented with 10% FBS and cultured overnight to attach. The next morning, cells were washed with PBS and incubated with fenretinide (10 μM) for 24, 48, or 72 hours. Medium containing fresh fenretinide were provided every 24 hours. Attached cells were collected at each time point and processed for Annexin V-FITC and propidium iodide double staining using Annexin V-FITC Apoptosis Detection kit (BD Biosciences, San Jose, CA) according to the manufacturer's instruction. Samples were then analyzed for Annexin V-FITC positive cells on a Fluorescence Activated Cell Sorter (FACS) (BD Biosciences, San Jose, CA).

Total RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instruction. RNA was quantified and assessed for purity on a UV spectrophotometer.

Total RNA (1 μg) was reverse-transcribed with oligo (dT) primer and M-MLVRT reverse transcriptase (Invitrogen, Carlsbad, CA) for cDNA synthesis. cDNA corresponding to 32 ng total RNA was used as the template in a 20 μl real-time PCR reaction with the ABI TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA) and the appropriate primer pair and Taqman probe. All primer pairs and Taqman probes were designed with Primer Express software v2.0 (Table 1). Real-time PCR was conducted using the ABI Prism 7300 real-time PCR system (Applied Biosystems, Foster City, CA). The quantification analysis for target gene expression was performed using the relative quantification comparative CT method 9.

ChIP was performed using the ChIP assay kit from Upstate (Charlottesville, VA) and antibodies specific for RARβ (1:600, Santa Cruz, Santa Cruz, CA). Control ChIP was performed using a normal rabbit IgG (Santa Cruz, Santa Cruz, CA). The immunoprecipitated DNA fragments were amplified by PCR with primer pairs encompassing the proximal RARE of human RARβ or CYP26A1 gene (RARβ: sense 5'-TGGGTCATTTGAAGGTTAG-3', antisense 5'-GTTCTCGGCATCCCAGTC-3'; CYP26A1: sense 5'-CCGCAATTAAAGATGAACT-3', antisense 5'-TACAGGTCCCAGAGCTTGAT-3').

Scramble siRNA and pre-designed siRNA for human RARβ gene were purchased from Ambion (Austin, TX). Huh-7 cells were transfected with siRNA (10 nM per 1 × 10⁵ cells) using Lipofectamine™ RNAiMAX Transfection Reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instruction. Cells were harvested 48 hours post-transfection for evaluation of RARβ knockdown efficiency.

Data are presented as mean ± S.E.M. Statistical analysis was performed using Student's t-test or one-way ANOVA. Significance was defined by p < 0.05.

Several studies have shown that retinoids have anti-proliferation or apoptotic effects in certain cancer cells 4. To assess the effect of retinoids on HCC cells, we examined cell viability and caspase-3 cleavage induced by individual retinoids in three human HCC cell lines (Figure 1). As an initial screening, 10 μM was used for all the tested retinoids. This dose might be high for certain retinoids, however, besides apoptosis, no obvious cytotoxicity was noted during the 3-day treatment. In Huh-7 cells, nine out of thirteen retinoids decreased cell viability, with fenretinide being the most effective one (79% decrease in cell number compared with DMSO treatment). Fenretinide also induced the strongest caspase-3 cleavage in detached Huh-7 cells and 9-cis retinoic acid caused a modest induction (Figure 1A). In HepG2 cells, although all retinoids examined significantly decreased cell viability, only 9-cis retinoic acid induced weak caspase-3 cleavage (Figure 1B). In Hep3B cells, six out of thirteen retinoids decreased cell viability, whereas another three retinoids increased cell number (Figure 1C). 9-cis retinoic acid, fenretinide, TTNPB, and lutein induced strong caspase-3 cleavage in Hep3B cells (Figure 1C). These findings indicate that fenretinide induces apoptosis in both Huh-7 and Hep3B cells, but not in HepG2 cells.

To further confirm the differential responses of Huh-7 and HepG2 cells to fenretinide, both cell lines were treated with fenretinide and assessed for caspase-3 cleavage, DNA double-strand breaks, and phosphatidylserine translocation in a time course study. In Huh-7 cells, caspase-3 cleavage was detected at as early as 24 hours after treatment, and the induction was sustained at 48 and 72 hours. In HepG2 cells, however, even after 72 hours treatment, no obvious caspase-3 cleavage was detected (Figure 2A). DNA double-strand breaks, another hallmark of apoptosis, assessed by the TUNEL assay, were detected in Huh-7 cells after treatment (Figure 2B). In contrast, no significant increase in DNA double-strand breaks was detected in HepG2 cells after treatment (Figure 2B). Some background TUNEL staining was detected in HepG2 cells possibly due to the endogenous peroxidase activity. In addition, during apoptosis, membrane lipid phosphatidylserine translocates from the inner leaflet of the plasma membrane to the outer leaflet, resulting in loss of cell membrane asymmetry. Fenretinide induced phosphatidylserine translocation in Huh-7 cells in a time-dependent manner, reaching 17-fold after 72 hours (Figure 3A). However, fenretinide failed to induce such changes in HepG2 cells even after a 3-day treatment (Figure 3B). Taken together, these findings convincingly demonstrate that Huh-7 cells are susceptible to fenretinide-induced apoptosis, but HepG2 cells are resistant.

To determine whether nuclear receptors mediate the apoptotic effect of fenretinide in HCC cells, the basal mRNA levels of twelve nuclear receptors were assessed by real-time PCR (Figure 4A). Among the three cell lines, Huh-7 cells have the highest basal mRNA levels of RARβ, RXRα, and an orphan nuclear receptor Nurr1 (also known as NR4A2). Hep3B cells have the second highest mRNA levels of RARβ and Nurr1. In contrast, in HepG2 cells, the basal mRNA level of RARβ is undetectable. On the other hand, in HepG2 cells, the basal mRNA levels of SXR (steroid and xenobiotic receptor) and CAR (constitutive androstane receptor), two xenobiotic sensors that mediate many xenobiotic responses, are the highest among the three cell clines.

The regulation of these twelve nuclear receptor genes by fenretinide were then evaluated in Huh-7 and HepG2 cells by real-time PCR. The induction fold was calculated by comparing the mRNA level of each nuclear receptor gene between DMSO and fenretinide treatment at each time point. Only those genes that showed marked changes in expression were presented. In Huh-7 cells, fenretinide caused a continuous induction of RARβ mRNA level (Figure 4B). After 48 hours of treatment, mRNA levels of RARα and γ were also highly induced in Huh-7 cells (Figure 4B). In contrast, a 9-fold induction of RARβ was detected 6 hours after fenretinide treatment in HepG2 cells, and then the induction dropped down to 3–5 fold later on. NOR1 (NR4A3) mRNA was induced 6-fold after 48 hours treatment in HepG2 cells (Figure 4C). Furthermore, by comparing the RARβ mRNA level between Huh-7 and HepG2 cells after fenretinide treatment, our data revealed a dramatic difference in RARβ mRNA level between these two cell lines (Figure 4D). Taken together, these data clearly depict a positive correlation between RARβ mRNA level and susceptibility to fenretinide-induced apoptosis, which suggests that RARβ may play an important role in mediating fenretinide-induced apoptosis in HCC cells.

It is known that RARβ induces it own expression upon stimulation by RAR ligands 10. Since fenretinide is a synthetic retinoid whether it can directly activate RARβ remains to be tested. So we examined whether fenretinide activates RARβ. We first tested whether fenretinide can activate the RXRα/RARβ-mediated pathway by transactivation assay (Figure 5). Fenretinide caused a marked induction of luciferase activity in Huh-7 cells (43-fold) and in CV-1 cells (13-fold) (Figure 5A and 5B). In contrast, fenretinide did not significantly increase luciferase activity in HepG2 cells (Figure 5C). These data indicate that fenretinide can transactivate RXRα/RARβ-mediated pathway in Huh-7 but not in HepG2 cells. As shown in Figure 4C, a modest induction of RARβ mRNA was seen in HepG2 cells 6 hours after fenretinide treatment, but no sustained induction at 48 hours. Consistently, no significant increase in luciferase activity, which was measured 48 hours after fenretinide treatment, was detected in HepG2 cells.

The most direct evidence of RARβ activation is the increased binding of RARβ to the response elements in retinoid target genes. It is known that RARβ binds to the RARE in its own promoter upon RA treatment 1011. One of the classic RAREs, a 5-bp spaced direct repeat (DR5), was found in the promoter of RARβ 10. Another well established retinoid target gene is cytochrome P450 26A1 (CYP26A1), an enzyme that catalyzes the breakdown of retinoic acid to more polar metabolites 10. Two RAREs have been found in the promoter of CYP26A1, one in the proximal region and the other in the distal region 12. Using chromatin immunoprecipitation (ChIP) assay, the direct binding of RARβ to the RAREs in RARβ and CYP26A1 upon fenretinide treatment was revealed. Fenretinide increased the binding (14-fold) of RARβ to its own promoter compared with DMSO treatment (Figure 6A). An even higher increase (27-fold) of RARβ binding to the CYP26A1 promoter was also noted (Figure 6B). Together, these results demonstrate that fenretinide directly activates RARβ.

To determine the role of RARβ in mediating fenretinide-induced apoptosis, the endogenous RARβ mRNA expression in Huh-7 cells was knocked down using siRNA. The knockdown efficiency of RARβ by three sequence-independent siRNA oligonucleotides was evaluated by real-time PCR. Three siRNAs silenced RARβ gene expression to different extents, the most efficient knockdown being 86% (siRNA #4124) compared with scramble siRNA, followed by 81% (siRNA #3935) and 68% (siRNA #4030) (Figure 7A). The apoptotic effect of fenretinide was then evaluated in these RARβ-deficient cells. Our results showed that DNA double-strand breaks induced by fenretinide were markedly reduced in RARβ-deficient Huh-7 cells (Figure 7B). Consistent with RARβ knockdown level, the greatest reduction of apoptosis (88.6%) was seen in the cells with the lowest endogenous RARβ mRNA level (cells transfected by siRNA #4124 with 86% knockdown of RARβ mRNA), followed by 83.1% in cells transfected by siRNA # 3935 with 81% knockdown of RARβ mRNA, and 70.7% in cells transfected by siRNA # 4030 with 68% knockdown of RARβ mRNA. These data clearly demonstrate that fenretinide-induced apoptosis of Huh-7 cells is RARβ dependent.