Ceramide (Cer) is an active cellular sphingolipid that can induce apoptosis or proliferation-arrest of cancer cells. Nanoparticle-based delivery offers an effective approach for overcoming bioavailability and biopharmaceutics issues attributable to the pronounced hydrophobicity of Cer. Missense mutations of the protein p53, which have been detected in approximately 42% of cancer cases, not only lose the tumor suppression activity of wild-type p53, but also gain oncogenic functions promoting tumor progression and drug resistance. Our previous works showed that cellular Cer can eradicate cancer cells that carry a p53 deletion-mutation by modulating alternative pre-mRNA splicing, restoring wild-type p53 protein expression. Here, we report that new ceramide–rubusoside (Cer–RUB) nanomicelles considerably enhance Cer in vivo bioavailability and restore p53-dependent tumor suppression in cancer cells carrying a p53 missense mutation. Natural RUB encapsulated short-chain C6-Cer so as to form Cer–RUB nanomicelles (∼32 nm in diameter) that substantially enhanced Cer solubility and its levels in tissues and tumors of mice dosed intraperitoneally. Intriguingly, Cer–RUB nanomicelle treatments restored p53-dependent tumor suppression and sensitivity to cisplatin in OVCAR-3 ovarian cancer cells and xenograft tumors carrying p53 R248Q mutation. Moreover, Cer–RUB nanomicelles showed no signs of significant nonspecific toxicity to noncancerous cells or normal tissues, including bone marrow. Furthermore, Cer–RUB nanomicelles restored p53 phosphorylated protein and downstream function to wild-type levels in p53 R172H/+ transgenic mice. Altogether, this study, for the first time, indicates that natural Cer–RUB nanomicelles offer a feasible approach for efficaciously and safely targeting cancers carrying p53 missense mutations.
Ceramide (Cer) is an active sphingolipid metabolite that can function as a tumor-suppressing cellular signal in combatting cancers (1). Cellular Cer, which includes an array of distinct species, is predominantly synthesized by the de novo synthesis pathway from serine and palmitoyl-CoA, and is also produced from sphingomyelin breakdown. Cer glycosylation mainly converts Cer into glucosylceramide, catalyzed by glucosylceramide synthase (GCS), and further serial glycosylations generate other glycosphingolipids. In the opposite direction, ceramidase hydrolyzes a Cer to sphingosine, and the latter can further be phosphorylated to sphingosine-1-phosphate (S1P). Increased levels of cellular Cer, attained either by activating Cer synthases in the de novo synthesis pathway and sphingomyelinases in sphingomyelin breakdown, or by inhibiting GCS and ceramidases, can result in cell proliferation-arrest, apoptosis, and autophagy, thus suppressing tumor growth (1, 2). It has been proven that Cer-induced cell death occurs upon treatments with anticancer drugs, or with ionizing radiation, coincidently contributing to therapeutic efficacy (1, 2). Inhibition of enzymes involved in Cer metabolism, such as GCS or ceramidases, significantly increases cellular Cer levels, resulting in apoptosis of cancer cells, even drug-resistant ones (3–6). Furthermore, supplying exogenous Cer provides a direct approach for combatting cancers (7–9). Synthetic Cers with short chains are cell-permeable and more efficacious than physiologically prevalent ones with long-chains (i.e., C18-C24 Cers) in pushing cells into apoptosis (4, 10). It has been reported that Cer-incorporating liposomes could effectively solubilize short-chain Cer, conferring increased bioavailability and anticancer efficacies in tumors (9, 11–13). However, natural Cer nanoparticles that not only increase Cer in vivo bioavailability, but also exhibit minimal adverse effects caused by polymeric or other particle constituents, would be highly desirable and have yet to be reported.
Cer cross-talks with tumor-suppressor p53 in inducing cancer cell death (14–16). The p53 protein, encoded by the TP53 gene, functions as a key tumor suppressor that stabilizes the genome with respect to propensity for tumorigenesis and cancer progression. As one essential and powerful transcription factor in cells, p53 activates the expression of p53-responsive genes, including p21, Bax, and Puma, and these proteins induce cell death or proliferation-arrest in response to genotoxic stress (17). p53 can upregulate expression of Cer synthase 6 in the de novo synthesis pathway, or of neutral sphingomyelinase in the sphingomyelin breakdown route, to directly cause accumulation of cellular Cer (18, 19). In response to genotoxic stress, p53 can also activate the transcription of genes, including members of the Bcl-2 family (Bax, Noxa, Puma), effecting Cer-driven apoptosis of cells carrying wild-type p53 (wt p53; ref. 20). On the other hand, Cer (C16-Cer, for example) can elevate p53 levels, via binding within the p53 DNA-binding domain and disrupting its complex with E3 ligase MDM2, thus decreasing p53 degradation (21). Furthermore, Cer is able to induce apoptosis in WTK1 lymphoblastoid cells that carry p53 M237I mutation by a Cer-dependent pathway, rather than a p53-dependent pathway (22). Our recent studies showed that interventions to increase cellular Cer sensitized cells to anticancer drugs in refractory NCI/ADR-RES (del-21 bp in p53 exon-5) and OVCAR-8 (del-18 bp in p53 exon-5) ovarian cancer cell lines carrying a p53 deletion-mutation (4, 14).
It has been reported that the TP53 gene is mutated in approximately 42% of cases in almost all types of cancers; among these mutations, about 75% are point-mutations that can encode full-length missense proteins (23). TP53 mutants are often observed in metastatic tumors or in recurred cancers of ovaries and colon (24, 25). Point-mutations at codons 175, 248, and 273 constitute approximately 19% of all TP53 genetic alterations, and these codons are referred to as mutation hotspots, base substitutions at which are more commonly reported in ovaries, colon, and lungs (ref. 23; http://p53.free.fr/Database/p53_cancer/all_cancer.html). Missense p53-mutant proteins not only lack the tumor-suppression activity of wt p53, but also often exhibit oncogenic gain-of-function (17). Transgenic mice having knocked-in TP53 R172H or R270H mutations (corresponding to R175H or R273H in the human versions) present a broader tumor spectrum with metastases, as compared with mice carrying wt TP53 (26, 27). Patients with Li-Fraumeni syndrome (LFS) with TP53 missense mutation develop cancer earlier, and have more tumors and shorter tumor-free survival, than patients with LFS without p53 (28). Furthermore, recent studies indicated that an HSP 90 inhibitor (17AAG) could suppress mutant p53 levels, and in turn, tumor growth, in colons of mice carrying TP53 R248Q/+ (29, 30).
Rubusoside (RUB), a natural steviol glycoside, has successfully served to solubilize various hydrophobic compounds. RUB enabled dispersions of several lipophilic compounds, including paclitaxel and curcumin for in vivo studies (31, 32). We report herein that we have generated Cer–RUB nanomicelles for delivering C6-Cer, and assessed bioavailability and effects on xenograft tumors generated from cells carrying a p53 missense mutation.
Materials and Methods
Cer–RUB preparation and characterization
The nanomicelles were prepared via a solvent evaporation method, as described previously (32, 33). Briefly, C6-Cer (or NBD C6-Cer) and RUB (2:100 ratios in weight) were added and mixed in ethanol (at 1,000-fold w/v dilution) for optimal solubility. After filtration, the resulting ethanoic solution was allowed to stand at room temperature (∼22°C) for 30 minutes, followed by evaporation under reduced pressure at 50°C with agitation in a RAPIDVAP evaporation system until dry powder was obtained (Labconco). This Cer–RUB complex powder was stored at −20°C until needed.
For use, Cer–RUB powder was redissolved in deionized water (5 mg/mL), and was further diluted for in vitro and in vivo studies. N-Hexanoyl-d-erythro-sphingosine (C6-Cer) and N-[6-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]hexanoyl]-d-erythro-sphingosine (NBD C6-Cer) were purchased from Avanti Polar Lipids. RUB was prepared and characterized as described previously (32).
The morphology of the Cer–RUB complex was characterized by transmission electron microscopy (TEM), as described previously (32, 33). Briefly, an aliquot of Cer–RUB (10%, w/v) was placed on a 400-mesh carbon-coated copper grid and visualized on a JEOL 100-CX TEM instrument (JEOL Inc.), with an accelerating voltage at 80 kV. Average particle sizes were calculated for each of three samples, and further corroborated by dynamic light scattering (DLS) using a Zetasizer Nano ZS90 (Malvern Panalytical; see Supplementary Materials).
Ovarian cancer cell lines A2780 (wt TP53) and OVCAR-3 (missense mutation TP53 R248Q+/+) were purchased from the ATCC. Cells were cultured in RPMI-1640 medium (Invitrogen) supplemented with 10% of FBS for A2780 or 20% of FBS with 0.01 mg/mL bovine insulin for OVCAR-3, and 100 U/mL penicillin, 100 μg/mL streptomycin, and 584 mg/liter l-glutamine. Human MCF-12A mammary epithelial cells were purchased from the ATCC, and cultured in DMEM-F12 (1:1) with 5% horse serum, insulin (5 μg/mL), hydrocortisone (500 ng/mL), human epidermal growth factor (20 ng/mL), and cholera toxin (100 ng/mL). Cells were maintained in an incubator humidified with 95% air and 5% CO2 at 37°C. All cell lines were re-authenticated by short tandem repeat profiling within 3 months of the study, as described previously (33). Cell lines after thawing were cultured with antibiotic-antimycotic (#15240062, ThermoFisher) for 2 passages to prevent Mycoplasma contamination; and no testing was performed for Mycoplasma species. For all lines, cells in passages 4–15 were used in the described experiments.
Lipid extraction and HPLC analysis of NBD C6-Cer
Lipid extraction and Cer analysis were performed as described previously (33). Briefly, OVCAR-3 or A2780 cells (1 × 106 cells/100-mm petri dish) were grown in 10% FBS RPMI-1640 medium for 24 hours. After harvest, cells (106 cells/reaction) were incubated with 1.0 μmol/L of NBD Cer–BSA or 1.0 μmol/L NBD Cer–RUB in 1% BSA RPMI-1640 medium (200 μL) at 37°C for indicated periods. Following rinsing and collection, cells suspended in ice-cold acidic methanol (acetic acid: methanol, 1:50, v/v) were transferred into glass vials and mixed with chloroform and water (1:1:1, v/v/v) to extract cellular lipids. After centrifugation, the organic lower phase was collected and evaporated to dryness, and the resulting residue was then stored at −20°C until further analysis.
For HPLC analysis, the extracted lipids were dissolved in chloroform/methanol/ortho-phosphoric acid solvent (80:20:0.1, v/v/v; 200 fluorescence units in 100 μL). Each sample (5 μL) was loaded onto a normal-phase silica column (5 μm ZORBAX Rx-SIL, 4.6 × 250 mm) for separation and analysis. The HPLC system (Agilent 1220 Infinity LC Gradient System VL) was equipped with an Agilent 1260 fluorescence detector (Agilent). NBD sphingolipids were eluted by linear gradient (0–14 minutes, 1 mL/min) formed with solvent system A (chloroform/methanol/ortho-phosphoric acid; 80:20:0.1, v/v/v) and solvent system B (chloroform/methanol/H2O/ortho-phosphoric acid; 60:34:6:0.1, v/v/v/v). NBD fluorescence was detected at λexcitation 470 nm and λemission 530 nm, respectively. NBD C6-Cer in samples was quantitated against a standard calibration curve. Each sample was analyzed in triplicate, and reported Cer levels were as normalized against total cellular or tissue proteins.
Cell viability assay
Cell viability was determined by using the CellTiter-Glo luminescent cell viability assay kit (Promega), as described previously (4, 14). Briefly, cells (4,000 cells/well; 2,500 cells/well for MCF-12A) were grown in 96-well plates overnight and then switched to 5% FBS medium containing respective drugs for 72 hours treatments. For combination treatment, cells were cultured in 5% FBS medium containing Cer–BSA or Cer–RUB (1 μmol/L) and cisplatin (CDDP; 0∼15 μmol/L) for 72 hours. Cell viability was assessed in a Synergy HT microplate reader (BioTek), following incubation with CellTiter-Glo reagent. NBD C6-ceramide complexed to BSA (NBD Cer–BSA) was purchased from Invitrogen and cisplatin (CDDP) was procured from Sigma-Aldrich.
Western blot analysis
Western blotting was carried out as described previously (14, 34). Briefly, cells or tissue homogenates were lysed in NP40 cell lysis buffer (Biosource) to extract total cellular proteins following these treatments. Equal amounts of proteins (50 μg/lane) were resolved by using 4% to 20% gradient SDS-PAGE (Life Technologies). The nitrocellulose-membrane blots transferred were blocked in 5% fat-free milk in 0.05% Tween-20, 20 mmol/L phosphate-buffered saline, pH 7.4 (PBST), and then incubated with each one of the primary antibodies (1:500 or 1:5,000 dilution), at 4°C overnight. These blots were incubated with corresponding horseradish peroxidase–conjugated secondary antibodies (1:5,000 dilutions) and detected using SuperSignal West Femto substrate (Thermo Fisher Scientific). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control for cellular protein. Antibodies against human p53 phosphorylated at Ser15 were purchased from Cell Signaling Technology. Antibodies for Puma, p21Waf1/Cip1, Bax, p53, and GAPDH were obtained from Santa Cruz Biotechnology. Relative protein levels were calculated from optical density values for each band, normalized against those for GAPDH.
Animal studies in tumor-bearing mice and transgenic mice
All animal experiments were approved by the Institutional Animal Care and Use Committee, University of Louisiana at Monroe (ULM), and were carried out in strict accordance with good animal practice as defined by NIH guidelines. Athymic nude mice (Foxn1nu/Foxn1+, 4–5 weeks, female) were purchased from Harlan and maintained in the vivarium at ULM. Animal studies were conducted as described previously (3, 4, 34). Briefly, cell suspension of A2780 or OVCAR-3 (5–7 passages, 1 × 106 cells in 20 μL/mouse) was subcutaneously injected into the left flank of the mice. Mice with tumors (∼2 mm in diameter) were randomly allotted to different groups (5 mice/group) for treatments. CDDP (1.5 mg/kg, once every 6 days) or Cer–RUB nanomicelles (1 mg/kg of C6-Cer, once every 3 days) was administered intraperitoneally (i.p.) alone or in combination for 24 days. Mice were monitored by measuring tumor sizes, body weights, and clinical observations, twice a week. Tumor volume was approximated by the formula, L/2 × W2 (where L is the length and W is the width). Bone marrow cells were extracted from the femurs of mice and counted on a hemocytometer, as described previously (35). Bone marrow cell counts for each mouse were normalized against the body weight.
Heterozygous 129S4-Trp53tm2.1Tyj transgenic mice (HTZ, p53 R172H/+) that have oncogenesis potential (26, 36) and corresponding wt mice (WT) were generously provided by the NCI Mouse Repository. Mouse genotypes were identified by amplification of tail DNA using the Platinum Blue PCR SuperMix (Invitrogen) with primers (5′AGC CTG CCT AGC TTC CTC AGG 3′; 5′ CTT GGA GAC ATA GCC ACA CTG 3′ synthesized by Invitrogen; denaturation, 94°C, 1 minute; annealing 60°C, 2 minutes; extension 72°C, 2 minutes for 30 cycles), as previously described (26). Cer–RUB nanomicelles (1 mg/kg, i.p., twice in 6 days) or saline were administered to WT and HTZ mice (5 cases/group). Tumors and other tissues were examined and characterized following hematoxylin and eosin staining of tissue microsections (5 μm).
Pharmacokinetics assessments were conducted in mice bearing tumors (A2780 or OVCAR-3; ∼5 mm in diameter) generated as described above and previously (9, 37). NBD Cer–RUB nanomicelles (1 mg/kg) or NBD Cer–BSA complex (1 mg/kg in medium) was intraperitoneally injected into mice. Blood, tumor, and other organs were collected at chosen time-points. Lipids were extracted and analyzed as described above. The pharmacokinetic parameters for noncompartmental analysis were calculated using software PK Solver 2.0.
Flow cytometry analysis
Flow cytometry analysis was carried out as described previously (34, 38). Briefly, the resected tumors (∼60 mg) were immediately dispersed in medium with collagenase IV (500 U/mL) at 37°C for 120 minutes with shaking. Cell suspensions were incubated with CD24 Alexa-Fluor 488–conjugated antibody (5 μL/106 cells; mouse IgG1 2F10, R&D Systems) and CD133/2 APC–conjugated antibody (5 μL/106 cells; mouse IgG2b 293C3, Miltenyi Biotec) in 1% BSA PBS at 4°C for 45 minutes. Tumor cells were resuspended in 1% BSA PBS (1 mL) following washing, and analyzed using a FACSCalibur instrument (BD Biosciences). Samples (10,000 events per each) were counted in triplicate. Tumor cells further were smeared on tissue slides, and images captured (×200 magnifications) using an EVOS FL cell imaging system with color CCD camera (Life Technologies). The flow cytometry data were further analyzed by using the FlowJo program (v10; FlowJo).
To observe the NBD Cer accumulation, cells (10,000 cells/chamber) grew in a 4-chamber slide with 5% FBS culture medium containing indicated agents for 24 hours, and were then switched to 1% BSA medium containing NBD Cer–BSA (1 μmol/L) or NBD Cer–RUB and incubated for 30 and 240 minutes. After washing with ice-cold PBS and methanol fixation, cells were counterstained with DAPI (4′,6-diamidino-2-phenylindole) in mounting solution (Vector Laboratories). Cells or tissues were observed and images (×200 magnifications) captured using an EVOS FL cell imaging system.
For immunostaining of tissue microsections, antigens were retrieved in steaming sodium citrate buffer (10 mmol/L, 0.05% Tween-20, pH 6.0). After blocking with 5% goat serum in PBST, slides were immunostained with primary antibodies against ceramide (1:500) and p21 (1:500), and then Alexa Fluor 488- or 555-conjugated secondary antibodies. Cell nuclei were counterstained with DAPI in mounting solution. Mouse monoclonal antibody against Cer (MID 15B4) was purchased from Sigma-Aldrich, and p21 antibody from Santa Cruz Biotechnology.
All experiments were repeated 2 or 3 times. Data are expressed as mean ± SD. Two-tailed Student t tests and ANOVA tests were used to compare the continuous variables in groups, using the Prism v7 program (GraphPad Software). All P < 0.05 comparisons were regarded as statistically significant.
Nanomicelles of Cer–RUB solubilize sphingolipid C6-Cer in aqueous solution
To enhance the bioavailability of C6-Cer (Fig. 1A), we encapsulated this sphingolipid with a natural polyglycoside, RUB, that has been applied in prior instances to enhance solubility and bioavailability of drugs (31, 32). Because of its pronounced lipophilicity, C6-Cer is poorly soluble in aqueous solutions, such as RPMI-1640 medium (0.1 mg/mL); however, Cer–RUB complex is soluble, generating a clear, transparent dispersion in medium at a relatively high concentration of 2 mg/mL C6-Cer (Fig. 1B). On the basis of saturation concentrations in medium and photometry, the solubility of C6-Cer was increased by more than 1,000-fold (5 vs. <0.005 mg/mL). With TEM analysis, we found the average diameter of the Cer–RUB nanomicelles to be 32 nm (Fig. 1C). The mean zeta potential and particle size of the Cer–RUB nanomicelles in PBS (pH 7.4, 37°C) was −15 mV and 32 nm, respectively, when measured by DLS. Significant changes in particle size and zeta potential of the nanomicelles were not observed upon storage in PBS for 5 days (Supplementary Fig. S1A and S1B). These results indicate that the Cer–RUB nanomicelles are sufficiently stable as to enable valid experimentation under physiological conditions.
Cer–RUB nanomicelles substantially enhance Cer uptake and bioavailability in cancer cells and in tumors
To examine the impact of Cer–RUB nanomicelles on bioavailability, we first assessed cellular uptake of Cer–RUB nanomicelles by cancer cells, comparing with Cer–BSA. In OVCAR-3 cells that carry a p53 R248Q missense mutation, Cer (green) levels in cytoplasm increased in cells incubated with Cer–RUB nanomicelles (1 μmol/L, 240 minutes) to an approximately 3-fold greater extent than with Cer–BSA (Fig. 2A and B). HPLC analysis further corroborated that Cer–RUB nanomicelles significantly increased Cer levels in OVCAR-3 and A2780 cells (Fig. 2B–D). Upon incubation with Cer–BSA complex, cellular Cer levels quickly increased, peaking at 60 minutes, to approximately 530 and 300 fmol/mg in A2780 and OVCAR-3 cells, respectively, and then decreased to 157 and 184 fmol/mg at 240 minutes. Interestingly, the Cer–RUB nanomicelles gradually and continuously increased cellular Cer during incubations, attaining levels 5-fold (778 vs. 157 fmol/mg, P < 0.001) and 4-fold (665 vs. 184 fmol/mg, P < 0.001) in A2780 and OVCAR-3 cells, respectively, those attained with Cer–BSA complex, after 240-minutes incubation (Fig. 2C and D). The delayed increases of cellular Cer in cells exposed to Cer–RUB are possibly attributable to a protracted release of Cer from endocytosed nanomicelles. Accumulated cellular Cer (AUC) was enhanced by 34% in A2780 and 73% in OVCAR-3 cells incubated with Cer–RUB nanomicelles, as compared with Cer–BSA complex.
We further assessed Cer bioavailability in mice after intraperitoneal administrations of NBD Cer–RUB, comparing with NBD Cer–BSA. At 3 hours after administrations, Cer levels were substantially increased, by approximately 28-fold in tumors, 26-fold in kidney, 83-fold in lung, 95-fold in liver, and 55-fold in serum of mice administered Cer–RUB nanomicelles (1 mg/kg; Fig. 3A), comparing with Cer–BSA complex. The tissue Cer levels of mice administered Cer–BSA were barely detectable with HPLC assays, for instance in tumors (Fig. 3A and B). In contrast, administration of Cer–RUB nanomicelles (1 mg/kg C6-Cer) can clearly provide high levels of Cer in most tissues, including xenograft tumors (Fig. 3A and B). Furthermore, we found that Cer–RUB nanomicelles equally delivered Cer to tumors and other tissues of mice either bearing OVCAR-3 tumors or bearing A2780 tumors (Fig. 3C), except that the Cer levels in brain were much lower (∼0.01 ng/mg) than in other tissues (Fig. 3C). Furthermore, our pharmacokinetic studies showed that tissue Cer concentrations reached maximum (Cmax) in lungs, serum, tumors, and colons of mice about 3 hours after administration (Fig. 3D). The elimination half-life of NBD Cer (t1/2) following the delivery was approximately 6.8 hours, and Cer levels in tumors and in serum remained relatively higher after peaking than in lung and colon (Fig. 3D, Table 1).
|Tissues .||tmax (h) .||Cmax (μg/mL or ng/mg) .||AUC (μg·h/mL) .||t1/2 (h) .|
|Tissues .||tmax (h) .||Cmax (μg/mL or ng/mg) .||AUC (μg·h/mL) .||t1/2 (h) .|
Abbreviations: AUC, area under the concentration-time course; Cmax, maximum concentration of Cer; t1/2, elimination half-life; tmax, time to reach.
Cer–RUB nanomicelles sensitize cancer cells carrying p53 R248Q missense mutation to cisplatin
We further assessed the effects of Cer–RUB nanomicelles on cancer drug resistance. Previous reports showed that OVCAR-3 ovarian cancer cells (p53 R248Q+/+) were resistant to CDDP and other anticancer drugs (39, 40). In the present study, the mean IC50 value for CDDP in OVCAR-3 cells was 4-fold greater (3.8 vs. 0.84 μmol/L, P < 0.001) than for A2780 cells that carry only wt p53 (Fig. 4A and B; Supplementary Fig. S2A). OVCAR-3 cells also displayed resistance to either Cer–BSA or Cer–RUB: the IC50 values were increased 7-fold (6.2 vs. 0.9 μmol/L, P < 0.001) for Cer–BSA, and 3-fold (3.8 vs. 1.1 μmol/L, P < 0.001) for Cer–RUB, as compared with A2780 cells, respectively (Fig. 4A and B; Supplementary Fig. S2A). Furthermore, noncancerous MCF-12A cells (wt p53) displayed relative resistance to insult from treatments with CDDP, Cer–BSA or Cer–RUB alone (Supplementary Figs. S2A, S2B, and 4B). As well, RUB itself did not show any cytotoxic effects on any of these cell lines, even at much higher concentrations (>15 μmol/L, Fig. 4A; Supplementary Fig. S2). Interestingly, though, Cer (1 μmol/L) significantly sensitized OVCAR-3 cells to CDDP, as evidenced by reduced IC50 values for CDDP in combination with Cer–BSA (2.3 vs. 3.8 μmol/L, P < 0.001) or with Cer–RUB (1.1 vs. 3.8 μmol/L, P < 0.001), respectively (Fig. 4B). However, combination of CDDP with either Cer–BSA or Cer–RUB did not have any significant effect on the IC50 values of CDDP in non-cancerous MCF-12A cells (Supplementary Fig. S2). These results indicate that the re-sensitizing effect of Cer on OVCAR-3 occurs in conjunction with p53 mutation.
Furthermore, we examined the effects of Cer–RUB on tumor growth in mice bearing xenograft tumors of OVCAR-3 or A2780 cancer cells. OVCAR-3 tumors were resistant to CDDP treatments (1.5 mg/kg), and the mean tumor volume was 60% larger (1,650 vs. 1,020 mm3, P < 0.001) than what was seen for A2780 tumors (Fig. 4C and D). Cer–RUB alone treatments (1 mg/kg) significantly decreased tumor volumes (1,253 vs. 1,650 mm3, P < 0.001) in mice bearing OVCAR-3 tumors (Fig. 4D), but did not significantly reduce A2780 tumors. Furthermore, Cer–RUB nanomicelles sensitized tumors to CDDP inhibition of growth: The mean OVCAR-3 tumor volume decreased upon combination treatment, to 48% of that seen with CDDP alone (805 vs. 1,650 mm3, P < 0.001) after 24 days of treatments (Fig. 4D). Flow cytometry analyses of ovarian cancer stem cells (CSC) in tumors indicated that the fractional CSC population (CD24+/CD133+) in OVCAR-3 tumors was 2-fold the fraction (0.86% vs. 0.4%, P < 0.001) found in A2780 tumors of mice treated with CDDP alone (Fig. 4E and F). Cer–RUB combined with CDDP significantly decreased CSC fraction in A2780 tumors of mice, by 5-fold (0.08% vs. 0.4%, P < 0.001), and in OVCAR-3 tumors, by 12-fold (0.07 vs. 0.86%, P < 0.001; Fig. 4E). These effects of combination treatments on CSCs were also correspondingly observed in tumor sections upon immunohistochemical staining (Fig. 4F). During animal studies, we did not observe any severe side-effects, and the body weights of treated mice did not present any significant alteration (Supplementary Fig. S3). Combination treatments with Cer–RUB and CDDP significantly increased bone marrow cells, by approximately 40% (8.7 vs. 6.2 × 104 cells/g, P < 0.05) as compared with treatments with CDDP alone, in mice bearing either OVCAR-3 or A2780 tumors (Fig. 4G). This observation is interesting given that CDDP often causes bone marrow suppression (41). Treatment with Cer–RUB alone slightly enhanced bone marrow cell numbers, but without significance, in mice bearing OVCAR-3 tumors (Fig. 4G).
Effects of Cer–RUB nanomicelles on the protein expression of TP53 and p53-responsive genes
To explore how Cer–RUB nanomicelles sensitize OVCAR-3 cells to anticancer drugs and suppress growth of tumors generated therefrom, we assessed treatment effects on the protein expression of TP53 and p53-responsive genes. The codon 248 missense mutation is among the most common p53 mutants, appearing in approximately 10% of human cancers (IARC TP53 Database version R17). OVCAR-3 cells carry R248Q missense mutation (Arg→Gln) in exon 7 (39, 40, 42). The wt p53 protein, which can be phosphorylated at serine 15, is barely detected in OVCAR-3 cells or in xenograft tumors generated therewith in mice, in contrast to the situation with A2780 cells or tumors (Fig. 5A–C). Interestingly, we found that Cer, either in Cer–BSA complex or as Cer–RUB, significantly increased the levels of pp53, by 28- or 40-fold, respectively in OVCAR-3 cells, as compared with vehicle treatment (Fig. 5A and B). Concordantly, Cer significantly increased the protein levels of p21 and Bax in OVCAR-3 cells treated with Cer–BSA (2-fold and 3-fold) or Cer–RUB (3-fold and 3-fold). However, Cer did not show such effects on p53 or p53-responsive gene expression in A2780 cells treated with Cer–BSA or Cer–RUB (Fig. 5A and B).
In accord with the enhanced in vivo bioavailability of Cer (Fig. 3A), Cer–RUB treatment significantly increased pp53 levels in OVCAR-3 tumors, in comparison with Cer–BSA treatment (Fig. 5C). Consistent with its tumor-suppressing effects (Fig. 4B), treatments with Cer–RUB alone or combined with CDDP significantly increased the levels of pp53, by 4-fold or 6-fold, respectively, as compared with treatment with CDDP alone, in OVCAR-3 tumors (Fig. 5D).
The effects of Cer–RUB on restoration of p53 were further examined in HTZ transgenic mice that carry p53 R172H/+ (corresponding to human R175H; identified by genotyping with PCR, bottom in Fig. 5E). In colons of HTZ mice, protein levels of pp53 and p21 were very much lower than in WT mice (Fig. 5E). We found that Cer–RUB treatments (1 mg/kg) significantly increased levels of pp53 and p21 in colons of HTZ mice, by 4-fold and 2-fold, respectively, as compared with saline treatments (Fig. 5E). However, Cer–RUB treatments only modestly increased p21 levels, and did not appreciably affect pp53, in the colons of WT mice. Immunohistochemical staining also corroborated the increased p21 expression (red) upon Cer uptake (green) in the cytoplasm of colon cells of HTZ mice treated with Cer–RUB (Fig. 5F). Altogether, these results indicate that Cer–RUB can effectively restore the protein expression of wt p53 and p53-responsive gene in cancer cells or transgenic mice carrying p53 missense mutations.
Cer–RUB nanomicelles, in which natural RUB encapsulates Cer, succeeded in overcoming hydrophobicity and substantially increasing Cer bioavailability for in vivo applications. We did not detect any significant nonspecific toxicity of RUB in cell culture or animal studies, even at significantly higher concentrations or amounts than with the RUB-encapsulated Cer at the highest doses used. By comparison with particles formed with synthetic polymers that have been used to incorporate Cer (9, 37, 43), the novel Cer–RUB nanomicelles are smaller (32 nm), even though encapsulation efficiency is relative lower (Cer:RUB, 1:50, w/w), and exhibit high water-solubility (>2 mg/mL). Cer efficacy is directly correlated with bioavailability—more specifically, the intracellular or intra-organ Cer levels obtained upon treatments. Cer–BSA complex, which is widely used in cell culture studies, delivered levels of Cer to cancer cells sufficient to reinstate CDDP lethality thereto (Fig. 2B and C; Fig. 4A and B); however, Cer–BSA could not appreciably deliver Cer into tissues after administration (Fig. 3A and B). By contrast, Cer–RUB not only increased delivery of Cer into cultured cells (>3-fold), but also overcame tissue barriers and provided substantial levels into tissues (Fig. 3A). Pharmacokinetics studies further indicate that Cer–RUB nanomicelles can efficiently deliver Cer into tumors and other tissues for various prospective therapeutic purposes (Fig. 3D, Table 1).
In the present study, Cer–RUB nanomicelles specifically sensitized cancer cells carrying p53 missense mutation to anticancer agents. OVCAR-3 cancer cells homozygously carry p53 R248Q and are resistant to many drugs, including CDDP (39, 40, 42). Cer–RUB comparably delivered Cer into OVCAR-3 and A2780 cells, or xenograft tumors generated therefrom (Figs. 2D and 3C); however, this nanomicellular agent proved much more efficacious, upon concomitant treatment with CDDP, in killing OVCAR-3 cells, and inhibiting the growth of the generated tumors (Fig. 4), than was the case with wt p53-carrying A2780 cells or tumors. Upon further assessment, the re-sensitizing effects of Cer–RUB correlate with restored protein expression of wt p53 and of p53-responsive genes (p21, Bax; Fig. 5A–D). Cer–RUB also restored the expression levels of pp53 and p21 in transgenic mice carrying the p53 R172H/+ missense mutation (Fig. 5E).
In addition to many cancer cell lines and tumors that present low levels of Cer and are resistant to chemotherapy, enhanced levels of endogenous Cer species (C16∼C20 carbon chains) have also been observed in patients afflicted with high-grade ovarian serous carcinoma, and in certain patients with breast cancer—circumstances associated with cancer aggressiveness (44, 45). In light of a “sphingolipid rheostat,” these apparently conflicting findings can be understood from the standpoint that a balance between pro-apoptotic Cer and pro-survival S1P determines cancer cell fate (46). Cancer tissues can break down endogenous Cer via ceramidases to sphingosine which can then be converted via sphingosine kinase into S1P, elevated levels of which are associated with propensity for drug resistance, aberrant proliferation, angiogenesis, and metastasis (1, 2). In this regard, Cer chain-length clearly matters: Not only can exogenously delivered short-chain Cer increase levels of endogenous Cers (47), but it also displays greater potency in inducing cell death, and is slower to be glycosylated, than are longer-chain Cer species (33, 48, 49). Cer–RUB, shown to effectively deliver Cer for in vivo targeting of cancer cells, may improve treatments of at least some cancers, regardless of altered Cer-glycosylation or of the sphingolipid rheostat; however, testing this proposition first demands further preclinical studies.
Emerging evidence documents that Cers are involved in restoring tumor-suppression activity in cancer cells carrying p53 mutations. Prior studies showed that Cer modulates pre-mRNA alternative splicing so as to restore p53 protein expression and tumor suppression in ovarian cancer cells carrying a p53 deletion-mutation (4, 14). Suppression of Cer glycosylation can restore p53-attributable tumor suppression in conjunction with a decrease of N6-methyladenosine formation in cancer cells carrying the p53 R273H/+ mutation (34, 50). Most likely, Cer modulates posttranscriptional processes to re-sensitize OVCAR-3 cells carrying the missense R248Q+/+ mutation, though the mechanisms still need to be further detailed. Even so, the present study, for the first time, has demonstrated that natural Cer–RUB nanomicelles can efficaciously deliver Cer in vivo so as to restore p53-dependent tumor suppression in cases of p53 missense mutations, while sparing normal tissues. The evident potential of Cer–RUB nanomicelles in cancer treatments clearly warrants further investigation.
Disclosure of Potential Conflicts of Interest
R.A. Hill was an expert panel member at Cosmetic Ingredient Review. No potential conflicts of interest were disclosed by the other authors.
Conception and design: S.K. Khiste, S. Nazzal, R.A. Hill, Y.-Y. Liu
Development of methodology: S.K. Khiste, Z. Liu, S.B. Hosain, Y.-Y. Liu
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.K. Khiste, Y.-Y. Liu
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.K. Khiste, K.R. Roy, M.B. Uddin, X. Gu, Y.-Y. Liu
Writing, review, and/or revision of the manuscript: Z. Liu, S. Nazzal, R.A. Hill, Y.-Y. Liu
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Nazzal, Y.-Y. Liu
Study supervision: Y.-Y. Liu
This work was supported by National Institutes of Health grants P20 GM103424-11 from the National Institute of General Medical Sciences, and R15CA167476 from the National Cancer Institute (to Y.-Y. Liu). This work was also partially supported by funds from the Board of Regents Louisiana (LEQSF-EPS2012-PFUND-299), the Louisiana Campuses Research Initiative (LaCRI-UL/Liu), and the Mizutani Foundation for Glycoscience (120068; to Y.-Y. Liu). The authors thank Dr. Zhangwei Hu (Department of Medicine, Brigham and Women's Hospital, Harvard Medical School) for his assistance in the dynamic light scattering analysis.
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.