tRFs and tiRNAs sequencing profiles of GC tissues
We performed tsRNAs sequencing to distinguish the differentially expressed tsRNAs (DETs) and not differentially expressed tsRNAs (NDETs), and tRF-Val-CAC-016 was finally selected based on the profiles. And then, we applied the hierarchical clustering to classify the DETs, a total of 69 up-regulated and 42 down-regulated DETs were presented with hierarchical clustering heatmap (Fig. 1a). We used the volcano plot to exhibit DETs of great significance, which indicated that five down-regulated and six up-regulated tsRNAs were presented in the plot based on the foldchange (FC) of the profiles (log2FC>=3 or log2FC<=-3, p<0.05) (Fig. 1b). Obviously, we selected tRF-Val-CAC-016 for further research considering the feasibility and statistical significance. Meanwhile, we accomplished the heatmap of the correlation coefficient to elaborate on the similarity of the GC samples using the R gplots package (Fig. 1c). The distribution and frequency of tsRNA subtypes were presented in Supplementary Fig. 1a-b and Supplementary Fig. 1c-d, respectively. The classification of tsRNA isodecoders is shown in Supplementary Fig. 1e-f.
tRF-Val-CAC-016 was significantly low-expressed in GC tissues.
We applied Gel-electrophoresis to verify the PCR product of tRF-Val-CAC-016, demonstrate the feasibility of PCR primers, and verify the authenticity of tsRNA sequencing (Fig. 2a). As shown in Fig. 2b, the length of tRF-Val-CAC-016 ranged between 50 bp and 100 bp, and we undertook the Sanger sequencing to verify the results. Then Fluorescent In Situ Hybridization (FISH) assay indicated that tRF-Val-CAC-016 is located in both nuclei and cytoplasm, but mainly in the cytoplasm (Fig. 2c). Afterward, the expression level of tRF-Val-CAC-016 was tested in GC cell lines, NCI-N87 and HGC-27 were finally selected (Fig. 2d). Finally, to confirm the efficiency of tRF-Val-CAC-016 mimics, we conducted the transfection, and the result was in line with our expectations (Fig. 2e). Analogously, the low expression level was confirmed in 40 pairs of GC tissues (Fig. 2f). And the expression of tRF-Val-CAC-016 was significantly associated with tumor size and histology in the aspect of clinicopathological features (Table 1). Furthermore, the prognostic outcome of tRF-Val-CAC-016 was calculated after the follow-up for the patients after gastrectomies, but the result was not significant (Fig. 2g). The sequences of tRF-Val-CAC-016 and related primers are listed in Supplementary Table 2.
tRF-Val-CAC-016 suppressed the proliferation of GC.
As shown in Fig. 3a-b, tRF-Val-CAC-016 significantly suppressed the proliferation of GC cells in CCK-8 assays, achieving a similar performance of oxaliplatin. Besides, tRF-Val-CAC-016 inhibitor promoted the proliferation of GC compared with the inhibitor control. The ethynyl-2'-deoxyuridine (EdU) assays demonstrated that tRF-Val-CAC-016 could suppress the cell replication activity of GC, but tRF-Val-CAC-016 inhibitor enhanced the replication activity (Fig. 3c-d). Meanwhile, we found that tRF-Val-CAC-016 could regulate the checkpoints of the cell cycle in GC. As presented in Fig. 3e-g, oxaliplatin mainly handled the G1 phase in GC. However, tRF-Val-CAC-016 adjusted the S phase significantly both in NCI-N87 and HGC-27. In colony formation assays, tRF-Val-CAC-016 could inhibit the viability of GC cells, slightly weaker than oxaliplatin. tRF-Val-CAC-016 inhibitor enhanced the ability of colony formation in GC cells (Fig. 3h-i). Consistently, these phenomena were rigorously explained in the immunoblotting assays, as indicated in Fig. 3j-k, tRF-Val-CAC-016 mimics obviously declined the protein expression of CyclinD1, CyclinB, c-myc. tRF-Val-CAC-016 inhibitor increased the protein expression of CyclinD1, CyclinB, c-myc. On the other hand, oxaliplatin decreased the expression of CyclinD1, c-myc compared with the control.
Bioinformatics analysis
Target genes of the down-regulated or the up-regulated tsRNAs were then enriched in the GO and KEGG analysis. In the down-regulated group, GO analysis was presented in Fig. 4a-c, andwe found that the MAPK signaling pathways were enriched significantly (Fig. 4d). In the up-regulated group, GO analysis was shown in Fig. 4e-g, andwe found that the Wnt signaling pathway was quite prominent (Fig. 4h). We then compared the bioinformatics data in the present study with GEO (Supplementary Fig. 2) and TCGA databases (Supplementary Fig. 3), which suggested that proliferation-related pathways were frequently enriched (Supplementary Fig. 2f and Supplementary Fig. 3f), and the Calcium signaling pathway was uncovered (Supplementary Fig. 3i), consistent with the function of CACNA1d.
CACNA1d was verified up-regulated in GC tissues and was selected as the potential target of tRF-Val-CAC-016.
We introduced the TCGA and GEO databases to predict the possible target to conduct further research. As shown in Fig. 5a, the Venn diagram took the overlap of MAPK components and the predicted target genes of tRF-Val-CAC-016. Then we analyzed the GEO database (GSE65801) and TCGA-STAD database and found that CACNA1d, PLA2G4A and TNF in GSE65801 (Fig. 5b), CACNA1d and PLA2G4A inTCGA-STAD were significantly up-regulated (Fig. 5c). Furthermore, we then applied the Kaplan-Meier plotter website and discovered that CACNA1d, TNF, TGFBR1, PDGFC, GADD45B were significantly and oppositely related to the prognosis of GC (Fig. 5d-l). Analysis above reminded us the vital role of CACNA1d as the possible downstream target of tRF-Val-CAC-016. Subsequently, we obtained the tissue microarray (TMA) with 90 pairs of GC specimens, including detailed follow-up data (Fig. 5m). IHC results were presented in Fig. 5n. Through the analysis of the follow-up data, we found that the expression of CACNA1d was not significantly related to the prognosis of GC (p= 0.1805) (Fig. 5o). The representative IHC images of GC and NATs are shown in Fig. 5p.However, the protein levels of CACNA1d in GC tissues were up-regulated compared with corresponding NATs (Fig. 5q). Hence, we selected CACNA1d as the target gene based on the intersection of the comprehensive analysis. To confirm the expression and function of CACNA1d, we purchased siRNAs for CACNA1d and selected si-CACNA1d-1 as the better inhibitory effect compared with si-CACNA1d-2 and si-CACNA1d-3 (Fig. 5r). On the contrary, pcDNA-CACNA1d couldsignificantly enhance the expression of CACNA1d (Fig. 5s). Moreover, the tRF-Val-CAC-016 inhibitor was able to reverse the suppressive function of si-CACNA1d on GC cells to some extent (Fig. 5t). Analogously, the effect of pcDNA-CACNA1d on GC was partly relieved by tRF-Val-CAC-016 mimics (Fig. 5u). The sequences of CACNA1d primers and siRNAs are listed in Supplementary Table 2.
tRF-Val-CAC-016 was immunoprecipitated by Argonaute-2 and could modulate the proliferation of GC by targeting CACNA1d.
The sequencing profile has elucidated the possible binding relation between tRF-Val-CAC-016 and CACNA1d mRNA (Fig. 6a), and the tRF-Val-CAC-016 mimics could significantly reduce the expression of CACNA1d in RT-PCR (Fig. 6b). Theprotein level of CACNA1d was decreased by tRF-Val-CAC-016 mimics and promoted by tRF-Val-CAC-016 inhibitor (Fig. 6c). Subsequently, we found that WT-CACNA1d-3’UTR plus tRF-Val-CAC-016 mimics group could significantly reduce the luciferase ratio compared with other groups in the Dual-luciferase reporter assay (Fig. 6d). Then we introduced RIP (RNA-binding protein immunoprecipitation) assay and found that tRF-Val-CAC-016 was significantly immunoprecipitated by Argonaute-2 compared with the IgG group (Fig. 6e). Gel-electrophoresis further verified the PCR product of RIP assays (Fig. 6f). And the immunoblotting confirmed the integrity of the process to wash the magnetic beads (Fig. 6g).
CACNA1d strengthened the proliferation of GC and was modulated by tRF-Val-CAC-016.
Rescue assays were undertaken to elucidate the connection between tRF-Val-CAC-016 and CACNA1d further. In the CCK-8 assays, si-CACNA1d could significantly decline the proliferation of GC cells, which was partially recovered by tRF-Val-CAC-016 inhibitor (Fig. 7a-b). In the EdU assays,the tRF-Val-CAC-016 inhibitor could rescue the inhibitory effect of si-CACNA1d in terms of cell replication activity (Fig. 7c-d). Interestingly, si-CACNA1d resulted in the G1 phase arrest in HGC-27 but S phase arrest in NCI-N87, and could both be rescued by tRF-Val-CAC-016 inhibitor (Fig. 7e-g). Similarly, the rescue effect of the tRF-Val-CAC-016 inhibitor on si-CACNA1d was also confirmed in the colony formation assays (Fig. 7h-i). The suppressive function of si-CACNA1d on the protein expression of CACNA1d, CyclinD1, CyclinB, c-myc was partly resumed by tRF-Val-CAC-016 inhibitor (Fig. 7j-k). Analogously, the reverse verification for the function of CACNA1d was then performed. pcDNA-CACNA1d and pcDNA-CACNA1d plus tRF-Val-CAC-016 mimics were transfected, respectively. As shown in Fig. 8a-i, tRF-Val-CAC-016 mimics could significantly reverse the reinforced effect of pcDNA-CACNA1d on the proliferation of GC cells. Moreover, pcDNA-CACNA1d significantly lifted the protein levels of CACNA1d, CyclinD1, CyclinB, c-myc, but tRF-Val-CAC-016 mimics could decline this function (Fig. 8j-k).
Inhibitor of MAPK signaling pathway (p38 MAPK-IN) significantly reversed the enhancement of tRF-Val-CAC-016 inhibitor on GC proliferation.
To demonstrate the role of the MAPK signaling pathway, we took p38 MAPK-IN to rescue the enhancement of tRF-Val-CAC-016 inhibitor on GC proliferation. It indicated that p38 MAPK-IN was able to block the pathway conduction partly as expected, which was manifested in the aspect of the deterioration to the influence of tRF-Val-CAC-016 inhibitor on GC. As shown in Fig. 9a-b, tRF-Val-CAC-016 inhibitor enhanced the GC proliferation in the CCK-8 assay, which p38 MAPK-IN could partly reverse. Meanwhile, tRF-Val-CAC-016 inhibitor promoted the cell replication activity, and p38 MAPK-IN could also decline this activity (Fig. 9c-d). The reversal phenomena also happened in the flow cytometry assays for cell cycle (Fig. 9e-g), and immunoblotting assays for CACNA1d, CyclinD1, CyclinB, c-myc (Fig. 9h-i).
tRF-Val-CAC-016 suppressed tumor proliferation in NCI-N87 xenografts.
Subcutaneous xenograft experiments were undertaken to discuss the function of tRF-Val-CAC-016 in vivo. The results indicated that tRF-Val-CAC-016 declined the capacity of tumor growth in mice in the aspect of the bodyweight of the xenografts and the tumor volumes (Fig. 10a-d). The PCR results of these resected tumors showed that tRF-Val-CAC-016 agomir significantly suppressed the proliferation compared with the agomir control group and normal saline (NS) group (Fig. 10e), which also happened in the immunoblotting assays for these mice tumors (Fig. 10f). Besides, the immunofluorescence staining assays for ki-67 and CACNA1d suggested that tRF-Val-CAC-016 reduced the proliferative capacity of GC and further confirmed the expression and location of CACNA1d protein (Fig. 10g-h). Meanwhile, theImmunohistochemistry (IHC) assays also demonstrated the low expression of CACNA1d in the tRF-Val-CAC-016 agomir group (Fig. 10i).
tRNA derivative tRF-Val-CAC-016 modulates the canonical MAPK signaling pathways by targeting CACNA1d.
As presented in Fig. 11a, the mechanism diagram clearly explained the pattern that tRF-Val-CAC-016 exerted its influence on the MAPK signaling pathway. Overview of MAPK signaling pathway was shown in Supplementary Fig. 4. In the immunoblotting assays of NCI-N87, tRF-Val-CAC-016 mimics suppressed the expression of CACNA1d (Cav1.3), ERK, p-ERK, and p-p38. tRF-Val-CAC-016 inhibitor enhanced the expression of CACNA1d (Cav1.3), ERK, p-JNK, p38, cyclinB, cyclinD1, c-myc (Fig. 11b). si-CACNA1d could inhibit the expression levels of CACNA1d (Cav1.3), JNK, p-p38, cyclinD1, and c-myc, which was able to be reversed by tRF-Val-CAC-016 inhibitor (Fig. 11c). pcDNA-CACNA1d increased the expression of CACNA1d (Cav1.3), ERK, p-ERK, p-JNK, p38, p-p38, cyclinB, cyclinD1, and c-myc, which was rescued by tRF-Val-CAC-016 mimics (Fig. 11d). In HGC-27, tRF-Val-CAC-016 mimics declined, but tRF-Val-CAC-016 inhibitor promoted the protein levels of CACNA1d (Cav1.3), ERK, p-ERK, JNK, p-JNK, p38, p-p38, cyclinB, cyclinD1, c-myc (Fig. 11e). And CACNA1d (Cav1.3), ERK, JNK, p-JNK, p38, p-p38, cyclinB, and cyclinD1 were down-regulated by si-CACNA1d, and this effect was reversed by tRF-Val-CAC-016 inhibitor (Fig. 11f). CACNA1d (Cav1.3), ERK, p-ERK, JNK, p-JNK, p38, p-p38, cyclinB, cyclinD1, and c-myc were enhanced by pcDNA-CACNA1d, which could be deteriorated by tRF-Val-CAC-016 mimics (Fig. 11g).