Rare tumors of solid organs remain some of the most difficult pediatric cancers to cure. These difficult tumors include rare pediatric renal malignancies, such as malignant rhabdoid kidney tumors (MRKT) and non-osseous renal Ewing sarcoma, and hepatoblastoma, a pediatric liver tumor that arises from immature liver cells. There are data in adult renal and hepatic malignancies demonstrating the efficacy of retinoid therapy. The investigation of retinoic acid therapy in cancer is not a new strategy, but the widespread adoption of this therapy has been hindered by toxicities. Our laboratory has been investigating a novel synthetic rexinoid, UAB30, which exhibits a more favorable side-effect profile. In this study, we hypothesized that UAB30 would diminish the growth of tumor cells from both rare renal and liver tumors in vitro and in vivo. We successfully demonstrated decreased cellular proliferation, invasion and migration, cell-cycle arrest, and increased apoptosis after treatment with UAB30. Additionally, in in vivo murine models of human hepatoblastoma or rare human renal tumors, there were significantly decreased tumor xenograft growth and increased animal survival after UAB30 treatment. UAB30 should be further investigated as a developing therapeutic in these rare and difficult-to-treat pediatric solid organ tumors. Mol Cancer Ther; 15(5); 911–21. ©2016 AACR.
Rare tumors of solid organs remain some of the most difficult pediatric cancers to cure. These difficult tumors include pediatric renal malignancies such as malignant rhabdoid kidney tumors (MRKT) and non-osseous renal Ewing sarcoma. MRKTs make up 2% of pediatric kidney tumors (1) but, unfortunately, the overall 10-year survival rate is less than 30% (2). Renal Ewing sarcoma represents another aggressive renal tumor with 25% to 50% of children presenting with advanced disease (3, 4). Hepatoblastoma, a pediatric liver tumor that arises from immature liver cells, like these rare renal malignancies, also provides a significant therapeutic challenge to the pediatric cancer caregiver, having a dismal survival rate of only 50% (5). Novel and innovative therapeutic interventions are needed for the treatment of these aggressive solid organ malignancies, especially high-risk, nonresponsive, and relapsing tumors.
Retinoid therapy is currently the standard of care for maintenance treatment in other pediatric malignancies, including acute promyelocytic leukemia and the solid tumor, neuroblastoma (6). A novel retinoid, 9-cis-UAB30 (UAB30), a synthetic analogue of 9-cis-retinoic acid, binds selectively to the retinoid X receptor (RXR), leading to activation of genes involved in induction of differentiation and apoptosis (7, 8). UAB30 has minimal toxicity compared with other retinoids (7, 9–11) and has recently been found to be effective in decreasing xenograft tumor growth in neuroblastoma (12). Investigators have demonstrated the effectiveness of retinoids against adult liver (13) and kidney tumors (14). These data in neuroblastoma (12), and the prior results with retinoids in adult renal and hepatic malignancies (13, 14), led us to believe that UAB30 may have an effect upon tumorigenicity in rare pediatric renal and hepatic malignancies.
We hypothesized that the treatment of the rare pediatric renal tumors, MRKT and renal Ewing sarcoma, and hepatoblastoma with UAB30 would affect tumor growth in vitro and in vivo. We demonstrated cell-cycle arrest, decreased proliferation and motility, and apoptosis in vitro and decreased xenograft growth in vivo with UAB30 treatment. These studies suggested a potential role for UAB30 in the treatment of aggressive pediatric solid organ tumors.
Materials and Methods
Cells and cell culture
Cell lines were maintained at 37°C and 5% CO2. The human hepatoblastoma cell line HuH6 was from Thomas Pietschmann (Hannover, Germany; ref. 15) and maintained in Dulbecco's Modified Eagle's Medium with 10% fetal bovine serum (FBS; Hyclone), 1 μg/mL penicillin/streptomycin (Gibco), and 2 mmol/L l-glutamine (Thermo Fisher Scientific). The MRKT cell line G401 (CRL-1441) was from ATCC (16) and maintained in McCoy's medium containing 10% non–heat-inactivated FBS (Hyclone), 1 μg/mL penicillin/streptomycin (Gibco), and 2 mmol/L l-glutamine (Fisher). The renal Ewing sarcoma cell line SK-NEP-1 (ATCC, HTB-48; ref. 17) was maintained in McCoy's medium (30-2007, ATCC) containing 15% FBS (Hyclone), 1 μg/mL penicillin/streptomycin (Gibco), and 2 mmol/L l-glutamine (Fisher). Due to nonadherent nature, SK-NEP-1 cells were passed by centrifugation at 300 rpm for 5 minutes, aspiration of the supernatant, and resuspension of the pelleted live cells in medium. All cell lines were obtained within last 4 years and were mycoplasma free, but were not characterized.
Antibodies and reagents
Rabbit polyclonal anti-PARP (9542S), anti-caspase-3 (9662), anti-AKT (9272), anti-phospho AKT (S473, 9271), anti-ERK1/2 (9102), and mouse monoclonal anti-β-actin antibodies were from Cell Signaling Technology. Rabbit anti-phospho-ERK1/2 (05-797R), mouse anti-FAK (4.47), and anti-phospho-Src (Y416, 05-677) were from Millipore (EMD Millipore). Mouse monoclonal anti-RXR was from Abcam (clone MOK13-17), rabbit anti-Src from Santa Cruz Biotechnology, and anti-phospho FAK (Y397) from Invitrogen (Invitrogen Life Technologies). 9-cis-UAB30 (UAB30) was synthesized as described previously (18).
Antibodies for immunofluorescence: primary anti-RXR antibody (Abcam, 1:1,000) and secondary goat anti-mouse Alexa Fluor 594 antibody (A-11045, Thermo Fisher, 1:33 dilution).
Western blot analyses were performed as described (19). Whole-cell lysates were isolated using radioimmunoprecipitation assay (RIPA) buffer supplemented with protease inhibitors (Sigma Aldrich), phosphatase inhibitors (Sigma), and phenylmethanesulfonylfluoride. Lysates were centrifuged at 14,000 rpm for 30 minutes at 4°C, protein concentrations were determined with BCA Protein Assay Reagent (Pierce) and separated by electrophoresis on sodium dodecyl sulfate polyacrylamide (SDS-PAGE) gels. Antibodies were used according to the manufacturer's recommendations. Molecular weight markers (Precision Plus Protein Kaleidoscope Standards, Bio-Rad) confirmed the expected size of the target proteins. Immunoblots were developed with Luminata Classico or Crescendo ECL (EMD Millipore). Blots were stripped with stripping solution (Bio-Rad) at 37°C for 15 minutes and then re-probed with selected antibodies. Equal protein loading was confirmed with β-actin.
Cell viability/proliferation assays
Cell viability was measured with alamarBlue assays. Cells (1.5 × 103 per well) were plated on 96-well culture plates, allowed to attach, and treated with UAB30 at increasing concentrations for 48 hours. Following treatment, 10 μL of alamarBlue dye (Invitrogen) was added, and absorbance (595 nm) was measured using a microplate reader (BioTek Gen5, BioTek Instruments). Viability was reported as fold change. Cell proliferation was measured with CellTiter96 AQueous One Solution (Promega) and trypan blue staining. Cells (5 × 103 cells per well) were plated, allowed to attach, and treated with UAB30 for 48 hours. Following treatment, 10 μL CellTiter96 AQueous One Solution (Promega) was added, and the absorbance at 490 nm was measured using a kinetic microplate reader (BioTek). For trypan blue, cells (1.5 × 103 cells per well) were plated, treated, stained with trypan blue, counted with a hemacytometer, and reported as fold change in cell count and fold change in the ratio of dead to viable cells.
Multiple methods were used to detect apoptosis. Cells were treated with increasing concentrations of UAB30, and whole-cell lysates were immunoblotted for cleavage of poly ADP-ribose polymerase (PARP). Detection of PARP cleavage indicated apoptosis. In addition, apoptosis was shown with a caspase-3 activation kit (KHZ0022, Invitrogen). Briefly, cells (1.5 × 106) were plated and treated with increasing concentrations of UAB30 for 48 hours, lysed, and protein determination assays were completed to ensure equal protein loading. Reagents were added to the proteins; the colorimetric changes were detected using a microplate reader (405 nm, BioTek), and results were reported as fold change in caspase-3. Caspase-3 was also detected with immunoblotting. Decreased total protein or increased cleavage products indicated apoptosis.
Cells (1.0 × 106) were allowed to attach and then treated with UAB30 (10 μmol/L, 48 hours). Cells were trypsinized, washed with phosphate-buffered saline (PBS), and fixed in 100% ethanol. Ethanol was removed and cells were stained with propidium iodide (PI) solution containing 0.3 μmol/L PI (Invitrogen) in 0.1% Triton X and RNAse A (Qiagen) for 30 minutes at room temperature (RT), and analyzed with fluorescence-activated cell sorting (FACS) using a FACSCalibur Flow Cytometer (Becton Dickinson Biosciences). Data were analyzed with ModFit LT software (Verity Software House Inc.). Negative controls were included in each FACS run.
Cellular invasion assay
Twelve-well culture plates (TransWell, Corning Inc.) with 8-μm micropore inserts were coated with Matrigel (BD Biosciences; 2 mg/mL, 50 μL for 4 hours at 37°C) on the top side. Cells (G401, HuH6) were treated with UAB30, plated into the upper well (4 × 104 cells), and allowed to invade for 48 hours. The inserts were fixed with 3% paraformaldehyde, stained with crystal violet, cells counted with a light microscope, and invasion was reported as fold change.
Attachment-independent growth assay
Attachment-independent growth was determined by the soft-agar assay. A base layer of culture media in 1% noble agar was established, and SK-NEP-1, G401, and HuH6 cells (1 × 104) were plated in the top layer of culture media and agar. Dishes were treated with UAB30 (0, 10 μmol/L) every 4 days. After for 4 to 6 weeks, colonies were imaged and quantified using Gel Dock Imager (Bio-Rad) and Quantity One Software (Bio-Rad) and colony counts were reported as fold change.
Similar to invasion, culture plates (TransWell) with 8-μm micropore inserts were used with collagen type I (10 mg/mL, 50 μL for 4 hours at 37°C) coating the bottom side of the insert. Cells (G401, HuH6; 4 × 104) were treated with UAB30, placed into the upper well, and allowed to migrate for 24 hours. The inserts were then fixed with 3% paraformaldehyde, stained with crystal violet, and migrated cells counted with a light microscope. Migration was reported as fold change.
Cellular migration was also measured utilizing a cell monolayer wounding (scratch) assay. G401 and HuH6 cells (3.5 × 105) were plated and allowed to attach. A 200-μL pipette tip created a uniform scratch in the cell layer, and photos (Photometrics CoolSNAP HQ2 CCD camera attached to a Nikon Eclipse Ti microscope) were obtained at time zero. Cells were treated with UAB30 (0–20 μmol/L) for 24 hours and photos were repeated. The area of the scratch was quantified by comparing the pixel count of the scratched area to the pixel count of the same plate at time zero and was reported as fold change in closure. Invasion, migration, and wounding assays were not performed utilizing SK-NEP-1 cells because they do not propagate in an adherent fashion, and floating cells do not lend well to these assays (20).
Immunofluorescence staining was utilized to detect the movement of RXR into the nucleus following UAB30 (10 μmol/L). Cells were plated on glass chamber slides, allowed to attach, and treated with UAB30. The SK-NEP-1 cell line did not grow well on these slides because they propagate as floating cells and were therefore not analyzed. After 48 hours, cells were fixed with 3% paraformaldehyde, permeabilized with 0.15% Triton X-100, and primary antibody was added at room temperature for 1 hour. The Alexa Fluor 594 secondary antibody was added (room temperature, 45 minutes). Prolong Gold with DAPI (P36931, Invitrogen) was used for mounting. Imaging was performed with a Zeiss LSM 710 Confocal Scanning Microscope with Zen 2008 software (Carl Zeiss MicroImaging, LLC) using a 63× objective with a zoom of 0.9. MetaMorph Microscopy Image Analysis Software (Ver. 7.6, Analytical Technologies, Molecular Devices) analyzed the images.
In vivo tumor growth
Six-week-old, female, athymic nude mice (Harlan Laboratories, Inc.) were maintained in the specific pathogen-free facility with standard 12 hours light/dark cycles with chow and water ad libitum. Experiments were approval by the Institutional Animal Care and Use Committee (140209355) and conducted within institutional, national, and NIH guidelines. Human MRKT cells G401 (2.5 × 106 cells in 25% Matrigel; BD Biosciences) or human hepatoblastoma cells HuH6 (2.5 × 106 cells in 25% Matrigel; BD Biosciences), were injected into the right flank. On the day of injection, mice were randomized to receive daily administration of vehicle-treated or UAB30-treated (100 mg/kg body weight/day) chow (n = 10 per group). Previous experiments proved this dosage to be well tolerated (21). Tumor volumes were measured with calipers and calculated with the standard formula (width2 × length)/2, where length is the largest measurement. Animals were followed until IACUC parameters for euthanasia were met (6 or 7 weeks, HuH6 and G401, respectively), at which point all animals were euthanized with CO2 and bilateral thoracotomy and the tumors harvested.
Human renal Ewing sarcoma cells SK-NEP-1 (1.5 × 106 cells/50 μL sterile PBS) were injected into the sub-capsular space of the left kidney in 6-week-old, female, athymic nude mice (Harlan Laboratories). When injected into the orthotopic position, these cells will form lung metastasis (20, 22). At injection, animals were randomized to daily administration of vehicle-treated or UAB30-treated (100 mg/kg body weight/day) chow (n = 10 per group). After 4 weeks of treatment, the animals were euthanized, and the kidney tumors and lungs were harvested.
Detection of metastases
Three levels from each lung of the animals with SK-NEP-1 xenografts were formalin fixed (10% buffered formalin), paraffin embedded, and stained with hematoxylin and eosin. These specimens were examined by a board-certified pediatric pathologist (E. Mroczek-Musulman) to determine the presence or absence of metastases.
Formalin-fixed, paraffin-embedded human specimens and xenograft tumor specimens were cut (6 μmol/L sections), baked at 70°C for 1 hour, deparaffinized, rehydrated, and steamed. Sections were quenched with 3% hydrogen peroxide and blocked with blocking buffer (BSA, powdered milk, Triton X-100, PBS) for 30 minutes at 4°C. The primary RXR (ab2815, Abcam) or anti-Ki67 antibody (ab15580, Abcam) was added (1:200) and incubated overnight at 4°C. Secondary antibody (1:400, rabbit anti-mouse SuperPicture Polymer HRP, Invitrogen, or donkey anti-rabbit, Jackson ImmunoResearch Laboratories, respectively) was added for 1 hour at 22°C. Staining was developed with the VECTASTAIN Elite ABC kit (PK-6100, Vector Laboratories), TSA (biotin tyramide reagent, 1:400, PerkinElmer, Inc.), and 3,3′-diaminobenzidine (DAB; ImmPACT DAB, Vector Laboratories). Slides were counterstained with hematoxylin. Negative controls [mouse IgG (1 μg/mL, Invitrogen) or rabbit IgG (1 μg/mL, EMD Millipore)] were included with each run.
A board-certified pathologist (E. Mroczek-Musulman) blinded to the treatment groups completed the Ki67 quantification. The area chosen for analysis was of greatest immunoreactivity. Five hundred cells were counted, and the ratio of immunopositive to total cells was reported as a percentage of positive staining (23, 24).
Experiments were repeated at least in triplicate, and data were reported as mean ± standard error of the mean. Student t test, Fisher exact test, χ2 test, or ANOVA was used as appropriate, with statistical significance determined at P < 0.05.
To determine whether UAB30, an RXR agonist, would have a receptor target, the expression of RXR receptors was evaluated. Immunoblotting detected RXR expression in all three cell lines (Fig. 1A). In addition, UAB30 treatment led to increased percentage of RXR staining in the nucleus in the G401 and HuH6 cell lines (P ≤ 0.03; Fig. 1B; Supplementary Fig. S1), indicating that UAB30 functioned as an RXR agonist. RXR staining in human renal tumors and hepatoblastoma specimens was examined. Human specimens were obtained from the tumor repository at our institution (IRB: X10093009; X110825022). Positive staining was seen in 8 of 8 hepatoblastoma specimens and in both of the MRKT and renal Ewing tumors examined. Representative photomicrographs at 40× are presented (Fig. 1C). Negative controls were included with each run (inserts on left corners).
UAB30 led to cell death
Next, the effects of UAB30 treatment upon cell survival were evaluated with alamarBlue assays. UAB30 treatment resulted in significant cell death in the renal tumors and hepatoblastoma cell lines (Fig. 2A and B, P ≤ 0.05, respectively). The LD50 for UAB30 was calculated: G401, 30 ± 4 μmol/L; SK-NEP-1, 22 ± 2 μmol/L; and HuH6, 78 ± 9 μmol/L. To determine whether UAB30-induced cell death involved apoptosis, immunoblotting was performed for PARP cleavage. As demonstrated by increased PARP cleavage (Fig. 2C), apoptosis was present. Reduction of total caspase-3 was detected by immunoblotting in all three cell lines at 100 μmol/L UAB30 (Supplementary Fig. S2) and activation of caspase-3 was significantly increased in all three cell lines (P ≤ 0.05) by treatment with UAB30 (50 μmol/L; Fig. 2D), further demonstrating apoptosis.
UAB30 resulted in cell-cycle arrest
Because retinoids are known to arrest cell cycle (25), we wished to determine whether UAB30 would have similar effects in these cell lines. Concentrations of UAB30 (10 μmol/L) were chosen that were below the calculated LD50 to demonstrate early morphologic changes rather than cell death. Representative FACS histograms are presented both pre- (control, left) and post-UAB30 (right) treatment (Fig. 3A). UAB30 led to arrest in cell-cycle G0–G1 progression in all cell lines with an increased percentage of cells in G1 (P ≤ 0.05) and decreased percentage of cells in the S phase (P ≤ 0.05; Fig. 3B).
After finding cell-cycle arrest, we wished to determine if UAB30 would also affect cellular proliferation. CellTiter96 and trypan blue exclusion was performed to measure proliferation. In all three cell lines, proliferation decreased following UAB30 treatment (10 μmol/L, 48 hours; P ≤ 0.05; Fig. 3C). In the trypan blue assays, there was a decrease in total cell number (Supplementary Fig. S3A) and the ratio between dead cells and viable cells significantly increased (P ≤ 0.05; Supplementary Fig. S3B). Together, these data indicated that UAB30 diminished cellular proliferation.
UAB30 decreased cell invasion and migration
Next, we wished to determine if UAB30 would alter other phenotypic features of the tumor cells. The hallmark of aggressive tumor cells is their ability to invade and migrate. To study the effects of UAB30 upon cellular invasion and migration, G401 and HuH6 cells were treated with low concentrations of UAB30 (up to 10 μmol/L) for 48 hours. There was a significant decrease in cellular invasion at 5 μmol/L concentration of UAB30 in both cell lines that was even more enhanced at 10 μmol/L (P ≤ 0.01; Fig. 4A).
Attachment-independent growth (soft-agar assays) is considered one of the best measures of cellular invasion, and was used to measure invasion in the G401, SK-NEP-1, and HuH6 cell lines. Cells were treated with UAB30 (10 μmol/L), placed into soft agar, and colonies were allowed to grow for 4 to 6 weeks. The number of colonies detected at the end of the studies decreased by 34% (P ≤ 0.01), 26% (P ≤ 0.05), and 86% (P ≤ 0.01) in the UAB30-treated G401, SK-NEP-1, and HuH6 cell lines, respectively, compared with controls (Fig. 4B).
The migration of tumor cells was studied in a fashion similar to invasion. Again, concentrations of UAB30 used to treat TransWell plates (0–10 μmol/L) were significantly less than the LD50. Treated cells were allowed to migrate for 24 hours. There was a significant decrease in the number of migrating cells beginning at 5 μmol/L UAB30 when compared with controls (P ≤ 0.01; Fig. 4C). Representative photographs of the stained inserts demonstrate decreased migration following UAB30 treatment (Fig. 4D). Migration was also studied with a monolayer scratch assay. The UAB30-treated cells showed a significant decrease in their migration across a monolayer scratch injury (Supplementary Fig. S4A and S4B). The SK-NEP-1 cell line was not studied with TransWell invasion or migration or scratch assay as these cells propagated in both an adherent and nonadherent fashion, which limited the evaluation of this cell line by these methods.
UAB30 decreased tumor growth in vivo in a nude mouse model of pediatric renal tumors
To more easily monitor tumor growth, a flank xenograft model was used to study the effects of UAB30 upon malignant rhabdoid kidney tumor growth. G401 cells (2.5 × 106) were injected into the right flank of female athymic nude mice (n = 20). On the day of injection, animals were randomized to either standard chow (control, vehicle) or chow with UAB30 added (n = 10/group). UAB30 was administered at a dose (100 mg/kg body weight/day) previously shown to be well tolerated by this species (21). Tumors were measured for 7 weeks. The tumors in the G401 control-treated animals grew more rapidly and were significantly larger at 7 weeks than the UAB30-treated animals (662 ± 189 mm3 vs. 36 ± 13 mm3; control vs. UAB30, P = 0.003; Fig. 5A). Xenograft tumors were stained for Ki67 to determine proliferation. Ki67 decreased in the UAB30-treated tumors, but differences were not statistically significant. This finding was likely secondary to the fact that the control tumors were large and necrotic, leaving fewer cells that could proliferate, compared with the UAB30 tumors that, although small, had more viable cells and, therefore, more cells capable of proliferation.
An orthotopic nude mouse model was used to evaluate UAB30 upon the growth of renal Ewing sarcoma cells and their ability to form lung metastasis (20, 22). SK-NEP-1 cells (1.5 × 106) were injected into the subcapsular space of the left kidney of female athymic nude mice (n = 28). On the day of injection, mice were randomized to standard chow (control, vehicle, n = 13) or chow with UAB30 (100 mg/kg body weight/day; n = 15). After four weeks, all animals were euthanized and kidney tumors were harvested (Fig. 5B). Previously, we showed that four weeks were required for control animals to reach pre-moribund tumor growth (20). In addition, the un-injected kidney (right) was also weighed to serve as a control. Animals treated with UAB30 had significantly smaller tumor weights (0.6 ± 0.1 g vs. 1.5 ± 0.2 g, UAB30 vs. control, P = 0.003; Fig. 5C) and volumes (204 ± 50 mm3 vs. 630 ± 134 mm3, UAB30 vs. control, P = 0.007; Fig. 5D) than controls. Animals that received UAB30 demonstrated less weight gain than the control animals; therefore, a tumor weight to animal weight ratio was calculated to determine if the lack of weight gain could explain the differences in tumor weight. The UAB30-treated animals had a significantly smaller ratio than the control animals (Supplementary Fig. S5), which demonstrated that the treatment, rather than the smaller animal size, accounted for the decreased kidney tumor growth. In addition, there was no difference between the weights of the control, contralateral right kidney, between the two groups, also providing evidence that the lack of weight gain was not responsible for the significant difference in tumor growth. Ki67 staining was completed on these xenografts tumors, and again tended to be higher in the untreated tumors, but did not reach statistical significance. Because SK-NEP-1 tumors are known to metastasize to the lungs (20, 22), lung tissues were stained with H&E to detect metastases. A representative example of a pulmonary metastasis (40×) in a control-treated animal is presented in Fig. 5E. The samples were evaluated by a board-certified pathologist (E. Mroczek-Musulman) blinded to the treatment groups. Any sized metastasis was counted as a positive. Lung metastases were significantly more common in control animals (54% vs. 13%, P ≤ 0.05, control vs. UAB30).
UAB30 decreased tumor growth in a nude mouse model of hepatoblastoma
An in vivo model of hepatoblastoma tumor growth following UAB30 treatment was used using female athymic nude mice. HuH6 hepatoblastoma cells (2.5 × 106 in Matrigel) were injected into the right flank of each mouse (n = 20). On the day of injection, mice were randomized to standard chow (control, vehicle) or chow with UAB30 (100 mg/kg body weight/day; n = 10/group). Tumors were measured for 6 weeks, at which time all the control tumors reached IACUC parameters for euthanasia, and all animals were euthanized. The tumors in the HuH6 control-treated animals were significantly larger at 6 weeks than those from the UAB30-treated animals (2,440 ± 366 mm3 vs. 657 ± 175 mm3, control vs. UAB30, P = 0.0002; Fig. 6A). These tumor xenografts were stained for Ki67 to measure cellular proliferation. Representative photomicrographs of Ki67 staining are presented (Fig. 6B). There was a decrease in cell proliferation in the UAB30-treated animals compared with controls, but it did not reach statistical significance (80% vs. 58%, control vs. UAB30, P = 0.1). Again, the control tumors were large and mostly necrotic, leaving fewer viable cells available for proliferation than the much smaller UAB30-treated tumors, likely leading to these findings.
The investigation of retinoids for the treatment of cancer has been ongoing for over 25 years, but other than in neuroblastoma (26), it has not become widely used for pediatric solid tumors. In recent literature, retinoids have been shown to have some efficacy for adult renal cell carcinoma (RCC) prolonging both median time to disease progression and overall survival (27, 28). In hepatocellular carcinoma (HCC), retinoids are currently being studied for chemoprevention (29–31). These studies prompted the current investigation of UAB30 in these difficult to treat pediatric solid tumors involving the kidney and liver.
In this study, we demonstrated that the nuclear receptor RXR was present in human hepatoblastoma specimens and in the HuH6 hepatoblastoma cell line. To our knowledge, this finding is novel. The RXR receptors have been documented in HCC cell lines (32–34), a primarily adult hepatic malignancy, so it was not completely unexpected that this nuclear receptor would be present in hepatoblastoma, and abnormalities in the RXR receptors have been highly associated with HCC progression (35, 36). There were not significant differences in RXR staining in the human hepatoblastoma specimens and therefore staining could not be correlated with patient survival or disease progression. However, demonstration of RXR expression was important to providing evidence that this therapy was clinically reasonable for hepatoblastoma.
Likewise, RXR receptor protein was found in MRKT and non-osseous renal Ewing sarcoma specimens and cell lines, also not previously documented. RXR receptors have been studied in RCC, a predominantly adult renal malignancy. Buentig and colleagues found that nuclear location of RXR inferred a survival advantage (37), implying that functional RXR receptors decreased RCC tumorigenicity. Further, Obara showed that the loss of RXR expression was associated with advanced RCC (38). Similar to hepatoblastoma specimens, the renal tumors specimens had fairly uniform staining, but the lack of a large number of specimens for review prevented a comparison between stain score and patient outcomes. Like hepatoblastoma, the presence of RXR in these specimens did lend credence to the potential for use of an RXR agonist in these tumors.
Multiple kinases have been hypothesized to be involved in retinoic acid effects upon cellular differentiation and apoptosis. Zhang and colleagues showed that all-trans-retinoic acid (ATRA) combined with cisplatin reduced AKT phosphorylation of HCC cells (39). ERK1/2 activation has also been shown to be induced by fenretinide in HepG2 (HCC) cells (39), but these same investigators showed that fenretinide decreased phosphorylation of ERK 1/2 in other HCC cell lines (40). Others showed that ATRA treatment of HCC cells resulted in increased JNK phosphorylation (41). These findings highlight the conflicting data that exist regarding retinoids and downstream pathways. Extensive investigations of retinoid-induced kinase activation have not been completed in renal tumor cell lines, but data from other tumor types have indicated that focal adhesion kinase (FAK) or Src may be involved. Our studies found no changes in phosphorylation in a number of these kinases previously shown to be relevant to retinoid treatment of hepatic or renal tumors (Supplementary Fig. S6). Our data did indicate that the therapeutic activity of UAB30 may be related to increasing the RXR activity in the nucleus (Fig. 1B). Other investigators have demonstrated that increases in retinoid receptor mRNA abundance or proteins correlated with retinoid-induced growth inhibition in RCC (42) and HCC (34, 43). Therefore, the mechanisms are likely more related to restoration of retinoid receptors responsible for controlling the expression of differentiation and other retinoic acid-responsive genes than to changes in kinase signaling.
The number of cell lines available for study of these rare, deadly, pediatric solid tumors is limited. One of the cell lines that we chose was the SK-NEP-1 cell line. This cell line was originally thought to be Wilms tumor, but has since been characterized as a renal Ewing sarcoma (17). SK-NEP-1 was chosen because renal Ewing sarcomas, similar to MRKTs, are clinically much more difficult to treat and carry a significantly worse prognosis than the standard Wilms tumor (44).
The dose of UAB30 chosen for the in vivo treatments was based upon our previous experience. The animals were given UAB30 a dose of 400 mg/kg diet, which translated to 100 mg/kg body weight/day (45). The animals tolerated this dose without significant changes in mucous membranes or skin. The animals with the orthotopic kidney tumor implant did not gain weight as quickly as their control counterparts; but their smaller size did not result in a difference in tumor:body weight ratios (Supplementary Fig. S5), indicating that the differences in body growth were not responsible for the decreased tumor size. In the adult trials that have been conducted, the patients given UAB30 did not experience changes in appetite or weight loss. Although the adults did not lose weight, unlike children, they were not required to gain weight and grow during their exposure to UAB30. The decreased growth in the animals will raise the question as to the effects of UAB30 upon growth in children and will require investigation as this treatment moves forward into the clinical arena.
Currently, retinoic acid is utilized in pediatric cancer treatment algorithms in the course of treatment other than established disease (26). Therefore, we chose to incorporate this same model in our animal studies, looking at the ability of UAB30 to prevent tumor establishment or metastasis. The findings that UAB30 is a better differentiating agent than cytotoxic agent, similar to other retinoids, also made this model a more attractive one for the current studies. Future studies combining UAB30 with cytotoxic agents will obviously require utilization of an established disease model.
In the current study, we demonstrated that a novel rexinoid, UAB30, was able to decrease cell survival in both kidney and hepatic tumor cell lines. In addition, UAB30 led to changes in cellular phenotypes that resulted in cell-cycle arrest and decreased migration and invasion in vitro. There was also significantly decreased tumor growth in flank xenograft and orthotopic murine tumor models. These results suggested that UAB30 may have a therapeutic role in these difficult to treat pediatric solid malignancies.
Disclosure of Potential Conflicts of Interest
D.D. Muccio and V.R. Atigadda hold a US Patent for the methods of making UAB30 and the uses thereof. Patent number 11/661,030, August 23, 2005. No potential conflicts of interest were disclosed by the other authors.
Conception and design: A.M. Waters, V.R. Atigadda, C.J. Grubbs, E.A. Beierle
Development of methodology: A.M. Waters, E.A. Beierle
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.M. Waters, J.E. Stewart, E. Mroczek-Musulman, E.A. Beierle
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.M. Waters, E. Mroczek-Musulman, E.A. Beierle
Writing, review, and/or revision of the manuscript: A.M. Waters, D.D. Muccio, C.J. Grubbs, E.A. Beierle
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.M. Waters, E.A. Beierle
Study supervision: E.A. Beierle
Other (refining of protocols used in experimental procedures related to the treatment and IHC staining): J.E. Stewart
The authors specially thank Shawn Williams and the UAB High Resolution Imaging Core for confocal microscopy assistance.
This work was funded by the UAB Comprehensive Cancer Center Mortimer A. and Josephine B. Cohen Research Acceleration and Innovation Fund (E.A. Beierle) and in part by institutional grants from the National Cancer Institute, including T32CA091078 (A.M. Waters), flow cytometry core grants P30 AR048311 (UAB) and P30 AI027767 (UAB), and high resolution imaging core grants P30 CA013148 (UAB) and P30 AR048311 (UAB).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.