Discussion
The next generation connectivity map (cMap) has been recently developed and should be acknowledged that the cMap methods and data are available without restriction to the research community (
Subramanian et al., 2017). As pointed out in the original paper, a future comprehensive cMap might expand in multiple dimensions, e.g., new cell types, patient-derived induced pluripotent stem cells and genome-edited isogenic cell lines (
Subramanian et al., 2017). Using this method, we found that the scores of the known drugs in treatment of GC (including ivermectin) were too low to indicate strong associations between these drugs and human GC gene expression signature, which was most likely due to the fact that the reference profile catalogue of cMap has been built to date on 12,328 genes of various cancer cell lines (including AGS which is a moderately differentiated human gastric adenocarcinoma hyperdiploid cell line) but not tumor tissues (
https://clue.io/connectopedia/l1000_gene_space and
https://clue.io/connectopedia/core_cmap_cell_panel).
In addition to the hypothesis generation approach by cMap, we further utilized data mining and pathway mining of knowledge-based datasets to identify the potential drugs in connection with a broad concept ranging from molecular entities (such as genes and proteins) to biological phenomena (such as molecular functions, pathways and phenotypes). Based on a better understanding of GC biology and signaling pathways, in the present study we focused on the WNT/β-catenin pathway by utilizing the algorithms of IPA which is built on a comprehensive, manually curated content of the QIAGEN Knowledge Base (over 57,000 publicly available datasets and continuously updated).
The results of the present study showed that the potential molecular targets and the mechanisms of action of ivermectin in GC differed from those in parasites in which ivermectin causes an influx of Cl-ions through the cell membrane of invertebrates by activation of specific ivermectin-sensitive ion channels (
Laing et al., 2017;
Chen and Kubo, 2018). In the present study, we identified ivermectin in connection with cell proliferation, particularly towards the genes (e.g., members of the adenosine triphosphate (ATP)-binding cassette (ABC) transporters). ABC are a superfamily of membrane proteins which play significant roles in transporting various exogenous and endogenous substances across membranes against concentration gradients through ATP hydrolysis, and many of these transporters are known as multidrug resistance proteins (MRPs) (
Mao et al., 2019). As showed in the present study, ivermectin also acted on the ABC and the signaling pathways, leading to inhibition of cell proliferation by deactivating LXR/RXR signaling (
Beltowski, 2011;
Saito-Hakoda et al., 2015).
It has been shown that activation of the WNT/β-catenin signaling pathway plays a pivotal role in many types of cancer (Clevers, 2006; Zhan et al., 2017). Previously, we and other research groups have demonstrated that the tumorigenesis of gastric cancer involves the WNT/β-catenin signaling pathway and the inhibition of the signaling pathways by means of denervation can suppress the tumorigenesis (Zhao et al., 2014; Chiurillo, 2015; Koushyar et al., 2020; Rabben et al., 2021). In the present study, we applied in silico modelling to show that ivermectin could inhibit the WNT/β-catenin signaling pathway including HIPPO signaling pathway, which is known to interact each other (Hayakawa et al., 2017; Li et al., 2019). We then employed in vitro and in vivoapproaches to show that ivermectin could inhibit cell proliferation and reduce tumor size, which was associated with the inhibition of the WNT/β-catenin signaling pathway. Thus, we may suggest that ivermectin could target the WNT/β-catenin singling pathway, leading to a reduced tumorigenesis. This was also in line with possible antitumor actions of ivermectin in other types of cancer cells, such as breast, colon, lung, prostate and bladder (
Melotti et al., 2014;
Diao et al., 2019;
Nappi et al., 2020).
Control of cell proliferation generally occurs during the G1 phase and multiple signals, ranging from growth factors to DNA damage to developmental cues, influence the decision to enter S phase, when DNA is replicated (
Duronio and Xiong, 2013). The results of the present study showed that ivermectin altered cell cycle in a concentration-dependent manner, which is consistent with a previous report showing accumulation of cells in the G1/S phases (
Zhang et al., 2019). In the present study, IC50-dose of ivermectin caused cell cycle arrest at G1 phase, whereas at higher doses, it caused S phase arrest. It has been suggested that WNT/β-catenin activation triggered cells in S phase, and HIPPO signaling might involve in G1 phase (
Benham-Pyle et al., 2016;
Kim et al., 2019). The evidence of possible link between the cell cycle arrest and inhibition of WNT/β-catenin and/or HIPPO singling pathways is needed to be further investigated, particularly in the context of ivermectin for GC.
There were several limitations of the present study. The cell proliferation and apoptosis in the
in vitroexperiment were not evaluated further by flow cytometry nor specific assays, e.g., annexin V staining or caspase activity. However, the gene expression profiling confirmed the association between the activities of networks of cell proliferation and cell death in mice, namely increased in cell proliferation and decrease in cell death in GC mice without treatment, and reversed activities in GC mice treated with ivermectin.
It should be noticed that the decrease in tumor size 2 months after ivermectin treatment was modest. As a matter of fact, in a separate experiment, we found that chemotherapy with 5-FU and oxaliplatin at the maximal dosage given to GC mice at the same age as ones in this study was without inhibition on the tumor size during 2 months of treatment (as same as in this study) (data not shown). However, the impacts of ivermectin treatment after a longer period of treatment alone and/or in combination with chemotherapy on resistance, migration and invasion could be worthwhile for future investigation. The results of the present study showed evidence of possible involvement of WNT/β-catenin signaling pathway in connection with the anti-cancer effect of ivermectin. For instance, prediction of ivermectin was successfully made by the WNT/β-catenin signaling pathway mining but not cMap. Validation of ivermectin was significant
in silico model of the WNT/β-catenin signaling pathway. Up-regulation of WNT/β-catenin signaling pathway took place in patients, human cell lines and mouse model of GC. Ivermectin treatment induced downregulation of the WNT/β-catenin signaling pathway in the mouse GC. However, additional evidence is needed to demonstrate that the effect of ivermectin is dependent on WNT/β-catenin signaling pathway. For instance, it would be worthwhile to further investigate how modulation of the WNT/β‐ catenin signaling pathway with specific inhibitors and activators will affect the response to ivermectin treatment
in vitro and
in vivo.