ATRA was purchased from Sigma-Aldrich; Merck KGaA. SKI II, a pan-SphK inhibitor, and SKI 5C, a selective inhibitor of SphK1, were obtained from Santa Cruz Biotechnology, Inc. The RARE-tk-Luc plasmid was kindly provided by Professor Michael J. Spinella (Dartmouth Medical School, Lebanon, NH, USA) (31). The pCMV-Blank vector was purchased from the Beyotime Institute of Biotechnology.

Extraction of asterosaponins from starfish. Asterias amurensis were collected from the sea area of Yantai (Shandong, China) in January 2018, stored in seawater containing 3.75% (w/v) MgCl₂ for 30 min and then frozen at −20°C. The frozen starfish were subsequently crushed, and extraction was performed with 65% (v/v) ethyl alcohol heat reflux; the extract was concentrated under low pressure. The concentrate was dispersed in water and extracted with petroleum ether three times. The aqueous phase was extracted with n-butanol three times, and the extract was concentrated under low pressure. The precipitate was dried to obtain a light brown powder. A small amount of the powder was examined and confirmed to be asterosaponins by a color reaction of concentrated sulfuric acid and hemolysis (32).

The K562 CML cell line was purchased from The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences. The K562 cell line was selected due to the presence of the BCR-ABL fusion protein, which is associated with SphK1 expression (27,28). The cells were cultured in RPMI-1640 medium (HyClone; Cytiva) supplemented with 10% (v/v) fetal bovine serum (HyClone; Cytiva), 100 µg/ml streptomycin, 100 U/ml penicillin and 2 mM glutamine (Sangon Biotech Co., Ltd.) at 37°C in a 5% (v/v) CO₂ humidified incubator.

Negative control (siNC) and siRNAs targeting SphK1 or SphK2 were synthesized by Shanghai Gene Pharma, Co., Ltd. siRNAs were transfected into K562 cells (1×10⁵ cells/ml) at 37°C using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) at a final concentration of 20 nM and incubated for 8 h. At 72 h post-transfection, the effects of the gene silencing were detected by western blotting. The siRNA sequences were as follows: SphK1 forward, 5′-GGCUGAAAUCUCCUUCACGTT-3′ and reverse, 5′-CGUGAAGGAGAUUUCAGCCTC-3′; SphK2 forward, 5′-GGGUAGUGCCUGAUCAAUGTT-3′ and reverse, 5′-CAUUGAUCAGGCACUACCCTC-3′; and siNC forward, 5′-UUCUCCGAACGGUCACGUTT-3′ and reverse, 5′-ACGUGACACGUUCGGAGAATT-3′.

K562 cells (2×10³ cells per well) were cultured in a 96-well plate. ATRA (1 µM), SKI II (0.5 or 5 µM), SKI 5C (0.5 or 5 µM) or asterosaponins (10 or 50 µg/ml) were added and incubated for 72 h at 37°C as indicated. In order to study the association between the expression of SphKs and cell sensitivity to ATRA, the cells were seeded into a 96-well microplate at a density of 2×10³ cells in 200 µl per well and treated with ATRA at 48 h post-transfection with SphK1/SphK2 siRNA. The control group was treated with 0.2% (v/v) DMSO. After 3 days, a Cell Counting Kit-8 (CCK-8) assay (Beijing Solarbio Science & Technology Co., Ltd.) was used to determine the effects of the drugs on cell viability. Three independent experiments were conducted in triplicate.

RT-PCR analysis was performed to analyze the expression levels of CYP26A1, SphK1 and SphK2 in K562 cells. K562 cells were cultured in 6-well plate (6×10⁵ cells/well) and were treated with ATRA (1 µM), SKI II (0.5 or 5 µM), SKI 5C (0.5 or 5 µM) or asterosaponins (10, 50 or 100 µg/ml) for 24 h at 37°C. Subsequently, total RNA was extracted from the cells by a Total RNA Extraction kit (Sangon Biotech Co., Ltd.) according to the manufacturer's instructions. For the cells transfected with siRNA, following transfection with SphK1/SphK2 siRNA or siNC, K562 cells (6×10⁵ cells/well) were treated with ATRA (1 µM) for 24 h at 37°C prior to RNA extraction. RNA was reverse-transcribed into single-stranded cDNA using a ReverTra Ace^® qPCR RT Kit (Toyobo Life Science). For conventional PCR, the RT product (1.5 µl) was used with the Quick Taq™ HS DyeMix (Toyobo Life Science). The PCR primer pairs were as follows: CYP26A1 forward, 5′-CCAAGGCAGCTCTACACTCAC-3′ and reverse, 5′-AGATGGCCAACATGAGCACA-3′; SphK1 forward, 5′-TCCTGGCACTGCTGCACTC-3′ and reverse, 5′-TAACCATCAATTCCCCATCCAC-3′; Sphk2 forward, 5′-CCAAGGCAGCTCTACACTCAC-3′, and reverse, 5′-AGATGGCCAACATGAGCACA-3′); and β-actin (reference gene) forward, 5′-GTCACCAACTGGGACGACA-3′ and reverse, 5′-TGGCCATCTCTTGCTCGAA-3′. The specific fragments amplified were as follows: 200 bp for CYP26A1, 260 bp for Sphk1, 178 bp for Sphk2 and 459 bp for β-actin. The PCR products were resolved on a 2% (w/v) agarose gel containing 0.1 µl/ml Gelview (BioTeke Corporation). Images of the gel were captured under an ultraviolet transmittance instrument. The relative light intensity was analyzed by AlphaEaseFC software (version 6.0.0; Alpha Innotech Corporation). Three experiments were carried out in three samples.

Prior to transfection, K562 cells were cultured in a 6-well plate at 4×10⁵ cells per well. The RARE-tk-Luc vector (1.2 µg) and β-galactosidase expression vector (80 ng) were transfected into the cells using Lipofectamine^® 2000 at 37°C for 8 h. The total amount of DNA per well was adjusted to 1.5 µg using the pCMV-Blank empty vector (Beyotime Institute of Biotechnology). At 24 h post-transfection, ATRA (1 µM), SKI II (0.5 µM) and SKI 5C (0.5 µM) were added into the medium at the indicated concentrations at 37°C. After 24 h, the cells were centrifuged at 1,000 × g for 5 min at 4°C, washed twice with RPMI-1640 medium, resuspended in lysis buffer (BioVision, Inc.) and incubated at 4°C for 20 min. The cell lysate was centrifuged at 10,000 × g for 10 min at 4°C, and the luciferase activity of the supernatant was measured using a luciferase assay kit (BioVision, Inc.). The luciferase activity was normalized to that of β-galactosidase. Three experiments were performed in three samples.

K562 cells were centrifuged at 1,000 × g for 5 min at 4°C and lysed with RIPA buffer (Beyotime Institute of Biotechnology). The total lysates (20 µg) were then separated by 8% SDS-PAGE and transferred to nitrocellulose membranes. The membranes were blocked for 10 h at 4°C in Tris-buffered saline with 0.05% (v/v) Tween-20 solution (Beyotime Institute of Biotechnology) containing 5% (w/v) non-fat milk. Subsequently, the membranes were incubated with primary antibodies against GAPDH (1:2,000; cat. no. 2118s; Cell Signaling Technology, Inc.), Sphk1 and Sphk2 (1:1,000) at 24°C for 1 h, followed by incubation with a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:50; cat. no. a0208; Beyotime Institute of Biotechnology) for 1 h at room temperature. An enhanced chemiluminescence system (Cytiva) was used to visualize the blots according to the manufacturer's protocol. The polyclonal antibodies anti-SphK2 (cat. no. sc-366378) and anti-SphK1 (cat. no. sc-48825) were purchased from Santa Cruz Biotechnology.

SPSS version 17.0 (SPSS, Inc.) was used for statistical analysis. Data are presented as the mean ± SD of at least three independent experiments and were analyzed by one-way ANOVA. Multiple comparisons were performed using the Tukey's studentized range test. P<0.05 was considered to indicate a statistically significant difference.

In order to investigate the effects of SphK enzyme activity on the intracellular function of ATRA, K562 cell viability was analyzed following treatments with ATRA, SKI II and SKI 5C. The effects of ATRA, SKI II or SKI 5C alone on the viability of K562 cells were first determined (Fig. 1). ATRA (1 µM) reduced the viability of K562 cells by 14.4%. SKI II slightly inhibited the viability of K562 cells at 0.5 and 5 µM, with inhibition rates of 19.1 and 23.5%, respectively (P<0.01 vs. control). The inhibition rates of SKI 5C on K562 cells at 0.5 and 5 µM were 24.8 and 33.3%, respectively, compared with the control group (P<0.01).

When 1 µM ATRA was combined with 0.5 µM or 5 µM SKI II, the inhibition rate of K562 cells was 34.8 and 41.7%, respectively (P<0.01 vs. 1 µM ATRA alone). When ATRA was co-administered with SKI 5C, the inhibition rate of K562 cells significantly increased (1 µM ATRA + 0.5 µM SKI 5C, 32.1%; 1 µM ATRA + 5 µM SKI 5C, 47.5%; both P<0.01 vs. 1 µM ATRA alone). These results suggested that the combination of SphK inhibitors and ATRA effectively inhibited the viability of CML cells.

In promyelocytic leukemia cells, CYP26 has been described as a retinoid-inducible gene (32). Therefore, the CYP26A1 expression levels were examined in K562 cells to determine the effects of SphK on ATRA activity. As presented in Fig. 2A, K562 cells expressed low levels of CYP26A1 mRNA, which significantly increased following ATRA treatment. Neither SKI II nor SKI 5C affected the expression levels of CYP26A1. When combined with SKI II or SKI 5C, the mRNA levels of CYP26A1 induced by ATRA were significantly higher compared with those observed in cells treated with ATRA alone (26 and 46%, respectively; P<0.01 vs. 1 µM ATRA). Thus, the increase in CYP26A1 transcription induced by SKI II and SKI 5C may be associated with the enhancement of ATRA-induced cytotoxicity in K562 cells.

Luciferase reporter assay was used to test this hypothesis. The results revealed that the luciferase activity of K562 cells transfected with the RARE-tk-Luc reporter plasmid was 3-fold higher compared with that of the control cells following ATRA treatment (P<0.01 vs. control; Fig. 2B). When ATRA was combined with SphK inhibitors, the luciferase activity of K562 cells was further increased (by 45.7% with 0.5 µM SKI II and by 19.8% with 0.5 µM SKI 5C; both P<0.01 vs. ATRA alone). These results further suggested that the enzymatic activity of SphK may affect the role of ATRA in promoting the target gene transcription.

In order to determine the role of SphKs in the retinoic acid resistance of CML cells, siRNAs were used to inhibit the expression of SphK1 or SphK2 in K562 cells. The results demonstrated that K562 cells expressed high protein levels of SphK1 and SphK2 (Fig. 3A). The protein levels of SphKs in K562 cells significantly decreased (by 56.3% for SphK1 and by 43.6% for SphK2; P<0.01 vs. siNC) compared with those in the siNC group at 48 h post-transfection with the corresponding siRNA, and remained inhibited for ≥72 h (data not shown). Subsequently, the transfected K562 cells were treated with ATRA, and the levels of CYP26A1 mRNA were analyzed. ATRA increased the mRNA levels of CYP26A1 by 19.1% (P<0.05 vs. siNC) in siNC-transfected K562 cells (Fig. 3B). When cells transfected with siSphK1 or siSphK2 were treated with ATRA, the expression levels of CYP26A1 increased (by 55.1% for siSphK2 and by 31.1% for siSphK1; P<0.01 vs. siNC), suggesting that ATRA significantly enhanced the expression levels of CYP26A1 following knockdown of SphKs.

The viability of the transfected cells was further analyzed by the CCK-8 assay. As presented in Fig. 3C, 1 µM ATRA slightly inhibited the proliferation of K562 cells transfected with siNC (by 14.4%; P<0.01 vs. vehicle control). Knockdown of SphK1 or SphK2 exerted no notable effects on the proliferation of K562 cells. When treated with ATRA, cells transfected with siSphK1 exhibited a higher rate of inhibition (19.1%) compared with the siNC group. Knockdown SphK2 also sensitized K562 cells to ATRA, and the inhibition rate of K562 cells treated with 1 µM ATRA was 24.8% (P<0.01 vs. siNC).

Experiments were performed to determine the effects of starfish extracts on the activity of SphKs, and it was observed that the expression levels of SphK2 mRNA were downregulated by asterosaponins (10 µg/ml, 17.1%, P<0.05; 50 µg/ml, 52.7%, P<0.01; and 100 µg/ml, 57.7%, P<0.01, respectively, vs. control), whereas SphK1 mRNA levels were not significantly affected compared with those in the control group (Fig. 4A). These results were validated by western blotting (SphK2: 10 µg/ml, 15.1%, P<0.05; and 50 µg/ml, 55.2%, P<0.01, respectively, vs. control; Fig. 4B). Since asterosaponins were demonstrated to inhibit the expression of SphK2, it was hypothesized that they may enhance the inhibitory effects of ATRA on the proliferation of K562 cells, similar to SphK inhibitors. K562 cells were treated with ATRA and/or asterosaponins, and the cell viability was analyzed. K562 cell proliferation was not notably inhibited by ATRA, with an inhibition rate of 6.2% at 1 µM (P>0.05 vs. control) (Fig. 4C). Asterosaponins inhibited the proliferation of K562 cells (10 µg/ml, 18.1%, P<0.05; and 50 µg/ml, 39.7%, P<0.01, respectively, vs. control). When 1 µM ATRA was used together with asterosaponins, the viability rate of K562 cells was further reduced (10 µg/ml, 22.4%; and 50 µg/ml, 31.7%; both P<0.01 vs. asterosaponins alone).

The present study further investigated whether the effects of asterosaponins on ATRA activity occurred due to the enhanced expression of ATRA-targeted genes in cancer cells. The results of RT-PCR analysis demonstrated that K562 cells expressed CYP26A1 at low levels. Following treatment with ATRA, the levels of CYP26A1 in K562 cells significantly increased by 61.4% compared with those in the control group (P<0.01; Fig. 4D). The effects of 50 µg/ml asterosaponins alone on the expression levels of CYP26A1 were not significant; however, the mRNA levels of CYP26A1 in K562 cells were significantly enhanced when the cells were co-treated with asterosaponins and ATRA (increased by 27.9%; P<0.01 vs. 1 µM ATRA alone). Therefore, the increase in the transcription levels of CYP26A1 induced by asterosaponins may be the mechanism underlying the enhanced ATRA-induced cell proliferation inhibition in K562 cells.