2.2. SZY-200 is membrane transport system independent
The differential expressions of nucleoside transport hENT1 mRNA in different cells were analyzed by qRT-PCR. The data showed that the expression levels were basically consistent with the sensitivities of the cells to the three drugs (Fig. 3A, 2A and S1). Subsequently, in order to study the role of hENT1 for these three drugs, differences in sensitivity to the drugs in the presence of hENT1 inhibitor NBMPR were determined. NBMPR decreased sensitivity to GEM in bladder cancer cells, while sensitivities to SZY-200 and CP-4126 were increased (Fig. 3B-3D). Taken together, different from GEM, SZY-200 and CP-4126 were independent of the membrane transport system. That is, they do not rely on hENT1 to help them enter into bladder cancer cells.
Fig. 3 The expressions of hENT1 at the transcriptional level in different cells and the effects of the nucleoside transport inhibitor NBMPR on the cytotoxic activities of SZY-200, GEM and CP-4126 in bladder cancer cells. (A) The expression of hENT1 in different cells, ****P < 0.0001 vs. the normal cells QSG-7701. Cell viability was measured at 48h after T24 cells (B), UM-UC-3 cells (C) and 5637 cells (D) were treated with SZY-200, GEM and CP-4126 (6.25, 12.5 and 25 nM) with or without 100 µM NBMPR. Data were mean ± S.D. (n = 3)
2.3. SZY-200 could arrest the cell cycle, induce apoptosis and significantly inhibit bladder cancer cells migration
Flow cytometric analysis was performed to determine whether SZY-200 affected the cell cycle distribution in bladder cancer cells. As shown in Fig. 4A and 4B, similar to GEM and CP-4126, SZY-200 significantly induced G1-phase cell cycle arrest in UM-UC-3 cells. In T24 and 5637 cells, SZY-200 also was similar to GEM and CP-4126 and caused the cell cycle arrest at G2 phase (Fig. S2A and S2B).
Then, we quantified SZY-200-induced apoptosis of cells by flow cytometry. The percentage of total apoptotic cells was significantly increased in UM-UC-3 (Fig. 4C and 4D), T24 and 5637 (Fig. S2C and S2D) cells after treated by SZY-200, GEM and CP-4126. Meanwhile, Hoechst 33258 staining showed changes in cellular morphology in UM-UC-3, T24 and 5637 cells after treated by SZY-200, GEM and CP-4126 (Fig. 4E, S2E). The untreated control cells were regular in shape, with abundant cytoplasm and also well attached to the membrane, the nuclei were round and stained homogeneously. On the contrary, the cells treated with three drugs were irregular in shape, the cell boundary was not smooth and the cells were more loosely attached to the membrane, there was lysis of some cells and showed evidence of blebbing. Moreover, apoptotic bladder cancer cells showed condensed and marked fragmented nuclei (white arrows) in a dose-dependent manner (Fig. 4E, S2E). Western blot analysis showed that the ratio of apoptosis-related proteins Bax and Bcl-2 were increased in UM-UC-3 cells after treated by SZY-200, CP-4126 and GEM (Fig. 4F and 4G). The similar data were obtained in T24 cells except no change of the ratio of Bax and Bcl-2 after treated by GEM (Fig. S2F and S2G). Taken together, these data suggested that SZY-200 induced apoptosis of bladder cancer cells.
Subsequently, cell motility was assessed by light microscopy. As shown in Fig. 4H and 4I, three drugs could significantly inhibit the migration of UM-UC-3 cells in a dose-dependent manner. The similar data were observed in both T24 and 5637 cells (Fig. S3A and S3B).
Fig. 4 Effects of SZY-200 on cell cycle progression and the apoptosis rate in UM-UC-3 cells and scratch tests. (A) After incubation, the cell cycle of UM-UC-3 cells was analyzed by flow cytometry. (B) Statistical analysis of the percentage of cells at the G0/G1, S and G2/M phases of cell cycle. (C) Apoptotic analysis was performed by flow cytometry using Annexin V-FITC/PI double staining method. (D) Statistical analysis of the percentages of apoptotic cells. (E) Apoptosis in UM-UC-3 cells was observed by fluorescence microscope using Hoechst 33258 staining (Scale bar = 50 µm). (F) Cell apoptosis-related protein expression in UM-UC-3 cells was detected by western blotting. (G) Represents the ratio of Bax/Bcl-2 protein expression levels. ****P༜0.0001 vs control. (H) Representative wound-healing assay pictures for UM-UC-3 cells treated with DMSO or with three drugs (12.5 and 25 nM) are shown (vertical lines indicate wound edges); (400×, original magnification). (I) Histograms represent quantitative analyses of cell migration. Values are expressed as the mean ± S.D. from three independent experiments. ****P༜0.0001 vs DMSO
2.4. SZY-200 down-regulates the expression of peroxisome proliferator-activated receptor γ (PPARG), cyclooxygenase-2 (COX‐2, also known as PTGS2) in bladder cancer cells
Finally, BATMAN-TCM database was used to predict the target genes of GEM and LA. As shown in Fig. 5A, there are 47 target genes for GEM, 11 for LA and no intersection between them. Disease enrichment analysis from TTD database showed that among the target genes of LA, PPARG and PTGS2, were related to the occurrence and development of bladder cancer (as shown in the black box). Moreover, qRT-PCR analysis showed that the mRNA expression of PPARG and PTGS2 were upregulated in two bladder cancer cell lines in comparison to the normal cells SV-HUC-1 (Fig. 5B). Importantly, SZY-200, CP-4126 and GEM could downregulate the expression of the two genes in bladder cancer cells (Fig. 5C).
Fig. 5 Prediction and analysis of target genes of GEM and LA. (A) Target genes from BATMAN-TCM database and disease enrichment analysis from TTD database. (B) Differential expressions of PPARG and PTGS2 in different cells. **P༜0.01, ****P༜0.0001 vs SV-HUC-1. (C) Effects of three drugs on the expressions of PPARG and PTGS2 in bladder cancer cells. Values are expressed as the mean ± SD from three independent experiments. ****P༜0.0001 vs DMSO
Previous studies have shown that CP-4126 was well tolerated with comparable toxicity profile to GEM in patients with some advanced solid tumors . In this study, SZY-200 and CP-4126 also showed similar cytotoxic effects in bladder cancer cells compared to GEM. The similar effect of SZY-200, CP-4126 and GEM demonstrated an effective conversion of the derivatives to GEM. CP-4126 was reported to traverse cell membranes by passive diffusion, followed by intracellular conversion to GEM by esterases in plasma and within tumor cells in order to be activated, and it is also dependent on dCK for phosphorylation . In addition, CP-4126 mainly localized in the membrane and cytosolic fraction, which leads to the long retention inside the cell . Compared with CP-4126, SZY-200 is assembled with LA not EA. Liu R et al.  once synthesized a drug (LA-Ara) based on the conjugation of cytarabine (Ara-C) with LA. The LA increased the lipophilicity of Ara-C and protected its NH2 group from the enzymatic attachment, thus markedly prolongs its plasma half-life. Furthermore, LA-Ara effectively improved anti-tumor activity compared with Ara-C. Importantly, LA-Ara could obviously decrease the incidence of toxic effects of Ara-C and is suitable to oral administration. From our data, clonogenic survival assay showed that LA could significantly inhibit the proliferation of bladder cancer cells in a dose-dependent manner, while EA has no inhibitory effect (Fig. 2B and 2C). Therefore, SZY-200 has the same or better anti-tumor activity as CP-4126.
Several previous studies highlighted that a significant correlation between the levels of hENT1 expression and the IC50 values for GEM in cancer cells [28, 29]. Neoptolemos et al.  demonstrated that patients with high hENT1 expression had a significantly higher median survival compared to patients with low hENT1 expression for those who received adjuvant GEM after undergoing surgical resection for pancreatic cancer. Galmarini et al.  found that the inhibition of hENT1 conferred resistance to GEM. In our study, the levels of hENT1 mRNA were found to basically be consistent with the sensitivities of the cells to GEM, CP-4126 and SZY-200. The hENT1 inhibitor NBMPR decreased sensitivity to GEM in bladder cancer cells and increased sensitivities to SZY-200 and CP-4126 (Fig. 3B-3D). Similarly, Bergman et al. (2011) reported that NBMPR and another nucleoside transport inhibitor dipyridamole decreased sensitivities to GEM 55 ~ 273-fold, while sensitivities to CP-4126 were only decreased 0 ~ 1.8-fold in THX human malignant melanoma cells and MOLT4 human T-cell leukemia cells. These studies revealed a nucleoside transporter independent transport over the cell membrane of CP-4126 and SZY-200. Moreover, hENT1 could operate as uptake as well as export transporter of GEM, the activity of SZY-200 and CP-4126 could be improved by cellular accumulation and prolonged retention of GEM inside the cell because of the inactivation of hENT1 . We therefore believe that SZY-200 could be more effective than GEM for bladder cancer patients with low hENT1 expression. For another, GEM was reported to inhibit the proliferation of bladder cancer cells by inducing cell cycle arrest and apoptosis [31, 32]. From our data, SZY-200 and CP-4126 exhibited the similar effects, which suggests that these GEM derivatives will release GEM in tumor cells to play its anticancer role. Interestingly, these drugs induced cell cycle arrest at G1-phase in UM-UC-3 cells, but G2-phase in T24 and 5637 cells (Fig. 4A, 4B, S2A and S2B). Several previous studies also showed similar results [33–35]. The underlying mechanism of this difference needs to be further explored in the future.
PPARG and PTGS2 are target genes of LA that we predicted by the BATMAN-TCM database. Previous studies have shown that they were related to the occurrence and development of bladder cancer. PPARG is a ligand-activated nuclear receptor and has been reported to interact with multiple signaling pathways, including Bcl-2, p53, p21, PTGS2 and cyclin D1 . The reduction of the activity of PPARG, whether through drug inhibition or gene ablation, would inhibit the proliferation of bladder cancer cells [37–39]. PTGS2 is a prostaglandin endogenous peroxide synthase. Recent studies have shown that human bladder cancer is associated with increased expression of PTGS2. Compared with normal bladder tissue, the level of PTGS2 in the bladder tissue of patients with cystitis or bladder cancer is increased . PTGS2 is usually an indicator of poor prognosis for patients . Cekanova et al.  suggested that the overexpression of PTGS2 could be a target for detection and treatment of bladder cancer. From our data, the mRNA expression of PPARG and PTGS2 were upregulated in two bladder cancer cell lines in comparison to the normal cells SV-HUC-1 (Fig. 5B). Importantly, SZY-200, CP-4126 and GEM could downregulate the expressions of the two genes in bladder cancer cells (Fig. 5C). As shown in Fig. S4 and S5, according to the KEGG biological pathway analyses, we found that these two genes (black arrows) are located downstream of the target genes of GEM. That's why these drugs could reduce the expression of these two genes.