the comparison of antiproliferation effect between paroxetine and fluoxetine in multiple cell lines.
To explore the beneficial therapeutic strategies for cancer patients with depression, we used antidepressants as a breakthrough to evaluate the therapeutic potential of different drugs. The antiproliferative activities of paroxetine and fluoxetine in multiple cell types were assessed by CCK8. For NSCLC cell lines (H460 and A549), the results showed that both paroxetine and fluoxetine inhibited the cell viability in a concentration-dependent way (Fig. 1a, b, f, g). Then, we compared the cytotoxicity of drugs in the human hepatoma cell line (Huh7). The data showed that paroxetine inhibited the growth of Huh7, while fluoxetine had a slightly inhibitory effect (Fig. 1d, i). The results of normal lung epithelial cells (BEAS-2B) and normal liver cell line (L02) indicated that paroxetine had a greater effect on normal cells than fluoxetine (Fig. 1c, e, h, j). Based on these data, we thought that the protective effect of fluoxetine on normal cells made it safer in clinical therapies, although the inhibitory effect of fluoxetine on tumor cells was not as significant as that of paroxetine, so we chose fluoxetine for further study.
Fluoxetine arrested cell cycle at G0/G1.
To clarify the inhibitory mechanism of fluoxetine, we then focused on the cell cycle. Flow cytometry was used to evaluate the distribution of the cell cycle. We observed that the percentage of the G0/G1 phase was increased with the drug concentration, these results indicated that fluoxetine could arrest the cell cycle in G0/G1 phase, and this ability was in a concentration-dependent manner (Fig. 2a, b). Meanwhile, we used western blot to evaluate the expression of cyclin-dependent kinases (CDKs) and p21, p27, the proteins related to the G1 phase. CDKs are the main regulator of the cell cycle, and CDK2 is an essential kinase for the G1/S transition (Du et al. 2016). p21 and p27 are well-known inhibitors of CDK2 (Levkau et al. 1998), which can cause cell cycle arrest at the G0/G1 phase. Our research found that fluoxetine could increase the expression of p21 and p27 and decrease CDK2 in a dose-dependent manner both in the H460 cells and A549 cells (Fig. 2c, d). These results were consistent with the flow cytometry analysis. Therefore, we speculated that fluoxetine might arrest the cell cycle. Altogether, the fluoxetine induced inhibitory influence on cell proliferation was demonstrated from these results.
Fluoxetine induced autophagy in a dose and time-dependent way.
Then, we shone a light on autophagy to discover the underlying mechanism of fluoxetine in regressing cell proliferation. Immunofluorescence was used to observe the autophagosome. We found that the intensity of the fluorescence increased with the drug concentration. The same phenomena could be detected in A549 cells as well (Fig. 3a). Then, the western blot was used to measure the expression of autophagy-related proteins. p62 and LC3 are the two markers of autophagy; therefore, these proteins can represent autophagy to some extent. We found that as the drug concentration increased, the expression of p62 and LC3B continued to increase, while the level of LC3A decreased (Fig. 3b). Another obvious phenomenon that could be observed was that the induction of autophagy flux gradually became more significant over time. Compared with control group, the level of LC3B changed from 3h and gradually became obvious, reaching the maximum at 24h (Fig. 3c). Therefore, we thought autophagy could be induced by fluoxetine in a dose-dependent and time-dependent fashion.
Fluoxetine had a connection with ER stress and the mTOR signaling pathway.
In order to investigate the specific mechanism of fluoxetine, we chose H460 cells that were more sensitive to the drug. RNA sequencing (RNA-seq) was used to study the transcriptome of cells treated with fluoxetine and analyze the difference with control treatment. In this result, 166 differentially expressed genes (P < 0.05, fold change ≥ 2) were identified, of which 12 genes were downregulated, and 154 genes were upregulated (Fig. 4a, b). We further analyzed the autophagy-related genes in the gene pool, and four genes proved to be significantly different. Interestingly, we found DDIT3, also known as C/EBP homologous protein (CHOP), was upregulated after fluoxetine treatment (Fig. 4c). CHOP is a characteristic biomarker of ER stress (Senft et al. 2015). A plethora of studies has been conducted to investigate the relationship between ER stress and autophagy (Song et al. 2018). Taking these into consideration, we wanted to explore the connection between autophagy and ER stress after fluoxetine treatment. Western blot was used to detect the ER stress-related markers. BIP expression was detected firstly, and it was found that fluoxetine induced an up-regulation of BIP levels. The activation of BIP can trigger the downstream sensors and the PERK pathway involved in autophagy is one of these sensors (Kouroku et al. 2007). The result of western blot showed that fluoxetine increased the content of PERK, activating transcription factor 4 (ATF4), and CHOP in a dose-dependent way (Fig. 4e). All these results indicated that fluoxetine had a close relationship with ER stress.
The analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) showed that the genes were enriched in 8 pathways which included biosynthesis, metabolism, mTOR signaling pathway, Lysosome, and AMPK signaling pathway (Fig. 4d). AKT/mTOR signaling pathway has been proved to play a critical role in regulating the process of autophagy (Heras-Sandoval et al. 2014). Hence, we detected the expression of AKT and mTOR. In our results, the fluoxetine could down-regulate the p-AKT/AKT and p-mTOR/mTOR in a dose-dependent way (Fig. 4f). Based on these results, we thought that the ER stress and AKT/mTOR signaling pathway might play a significant role in fluoxetine-treated cells.
AKT/mTOR signaling pathway was regulated through ATF4 in fluoxetine treatment.
A recent study has demonstrated that ER stress can regulate the mTOR signaling pathway to play a role (Yao et al. 2020). Thus, we further investigated the regulation of ER stress and AKT/mTOR signaling pathway induced by fluoxetine. The role of CHOP and ATF4 were identified firstly. The analysis of bioinformatics showed that ATF4 had a higher expression in lung cancer tissues (Fig. 5a), and the higher expression of ATF4 was significantly related to longer OS and DSS of NSCLC patients (Fig. 5b, c). The role of CHOP in expression and prognosis is not obvious (data not shown). Therefore, we chose ATF4 for further experimental validation. To clarify whether the down-regulation of AKT/mTOR signaling pathway was induced by ATF4, the cells were transfected with ATF4 siRNA and the western blot was used to examine the efficiency of siRNA (Fig. 5d). We next investigated whether fluoxetine-mediated inhibition of AKT/mTOR occurred through increased ATF4. The cells were transfected with ATF4 siRNA, and levels of p-AKT/AKT and p-mTOR/mTOR were assessed after fluoxetine treatment. Compared with NC-siRNA, the expression of p-AKT/AKT and p-mTOR/mTOR recovered in ATF4-knockdown cells, and the expression was reduced after fluoxetine treatment (Fig. 5e).
The fluoxetine-induced anticancer effect through the ATF4-AKT-mTOR signaling pathway.
And then to analyze whether the fluoxetine induced anticancer effect through the ATF4-AKT-mTOR signaling pathway. Cell proliferation, cell cycle, and autophagy were measured after ATF4 siRNA treatment. The CCK8 assay showed that the cell viability recovered after ATF4 knockdown (Fig. 6a). The analysis of cell cycle results found that the proportion of G0/G1 phase and the expression of G0/G1 related proteins was reduced after ATF4 knockdown (Fig. 6b, c, e). Moreover, the results of immunofluorescence and protein level of LC3B further confirmed that the treatment with ATF4 siRNA decreased the induction of autophagy after fluoxetine treatment (Fig. 6d, e). All these data indicated that the ATF4-AKT-mTOR signaling pathway exerted a significant role in the fluoxetine-induced anticancer effect.
Fluoxetine exerted anti-tumor effects while not damaging normal cells.
The previous result showed that fluoxetine had no inhibitory effect on normal lung epithelial cells. To confirm this result further, we used flow cytometry and immunofluorescence to analyze. The results demonstrated that fluoxetine was unacted on the cell cycle and autophagy (Fig. 7a, b, c). Western blot also revealed that the protein levels of cell cycle and autophagy were slightly changed after fluoxetine treatment (Fig. 7d, e).