To the Editor,

Colorectal cancer (CRC) is a heterogeneous disease and is the second and third most common cancer in females and males, respectively [1]. In recent years, an increasing number of studies have shown that, when catalyzed by enzymes, tRNA can be cleaved into small molecular fragments called tRNA-derived fragments (tRFs), which participate in the central dogma independently as amino acid carriers. Evidence has shown that tRFs exert function by directly binding to proteins [2,3,4].

The workflow of this study is summarized in Fig. 1a, with more details provided in Additional file 2: Fig. S1. In this study, tissues were collected from 10 patients with CRC to transcriptome sequencing and small RNA sequencing, and plasma exosomes were obtained from 10 CRC patients and 10 healthy controls to small RNA sequencing (Additional file 3: Tables S1-1, S1-2). First, we analyzed mRNA expression in CRC tissues versus matching adjacent normal tissues and carried out Gene set enrichment analysis (GSEA). The results suggested that CRC tissues were enriched with signature pathways such as cytokine-cytokine receptor interaction, IL-17 signaling pathway, NOD-like receptor signaling pathway, and ribosome biogenesis in eukaryotes (Fig. 1b, c and Additional file 4: Tables S7, 8). Meanwhile, CRC tissues and plasma exosomes demonstrated distinct patterns of tRFs expressions (Fig. 1d, e and Additional file 4: Tables S9, 10). Some studies have reported that tRFs can trigger mRNA degradation or translational repression, similar to miRNAs [5]. Therefore, we used the miRanda algorithm to identify putative target mRNAs using the KEGG database. Multiple pathways, such as the PI3K-Akt, calcium, and Rap1 signaling pathways (Additional file 2: Fig. S2a, b) were observed. Next, we regrouped the tissue samples only based on the score of specific tRFs (tissues’ tRF or plasma exosomes’ tRF) and divided the samples into high tRF_group and low tRF_group (Additional file 5: Table S21-1, S21-2). Enrichment assays showed that IL-17 signaling pathway and ribosome biogenesis in eukaryotes were also associated with the high tRF_group and low tRF_group (Fig. 1f–i and Additional file 4: Tables S11–S14). We further found that tRF-3 was the most abundant tRF in tissues (50.70%) and plasma exosomes (46.84%) (Fig. 1j, k). The tRFs derived from tRNA-Gly were relatively high in exosomes and tissues (Additional file 2: Fig. S2c, d). Three tRFs (tRF-3022b, tRF-3030b, and tRF-5008b) had relatively higher expression levels than the controls (Additional file 2: Fig. S2e), and correlation analysis suggested that some plasma exosome tRF might be from tumor tissues (Fig. 1l). We then validated three tRFs were frequently upregulated in CRC tissues and plasma exosomes compared to their normal counterparts (Additional file 2: Fig. S3a–e and Additional file 3: Tables S2, S3). We next investigated the area under the receiver-operating characteristic ROC curve (AUC) values for each biomarker (Additional file 2: Fig. S4a–e). Altogether, these results suggest that these tRFs have potential diagnostic capabilities for patients with CRC. A previous study reported that tRFs are produced from pre-tRNAs or mature tRNAs via non-random cleavage events mediated by certain proteins and enzymes [6]. Likewise, it suggested that ALKBH3, DNMT2, AGO2 and ANG may be related to the biogenesis of tRFs in CRC (Additional file 2: Fig. S5a–d).

Fig. 1
figure 1

Screening and profiling of tRFs in the tissues and plasma exosomes from CRC patients and healthy controls. a Schematic presenting the experimental procedure of our study. b Volcano plot showing the gene expression profile of CRC tissues and matching adjacent normal tissues. Down-regulated genes are indicated in green; upregulated genes are indicated in red. c Gene set enrichment analysis of the differential gene in tissues. d A heatmap showing 33 differentially expressed tRFs in 10 paired CRC tissues and matching adjacent normal tissues. e A heatmap showing 25 differentially expressed tRFs between 10 plasma exosomes of CRC and healthy controls. f Volcano plot showing significant DEGs in tissues between the high tRF_group and low tRF_group. Down-regulated genes are indicated in green; upregulated genes are indicated in red. g GSEA plots for high tRF_group and low tRF_group in tissues. h Volcano plot showing significant DEGs in plasma exosomes between the high tRF_group and low tRF_group. Down-regulated genes are indicated in green; upregulated genes are indicated in red. i GSEA plots for high tRF_group and low tRF_group in plasma exosomes. j Percentage of each type of tRFs in tissues. k Percentage of each type of tRFs in plasma exosomes. l Correlation analysis of tRF-3022b, tRF-3030b and tRF-5008b between tissues and plasma exosomes. (***P < 0.001)

Full size image

To investigate the specific functions of tRFs, we performed flow cytometry assay to demonstrate that knockdown of tRF-3022b, tRF-3030b, and tRF-5008b arrested the G2/M phases and promoted cellular apoptosis (Additional file 2: Figs. S6a, b and S7a, b). Subcutaneous injection of LNA_3022b-transfected HCT116 cells into nude mice attenuated tumor growth in vivo (Fig. 2a–c). Considering the presence of tRFs in plasma exosomes, we analyzed expression level of tRFs in cell-derived exosomes of various cell lines (Fig. 2d, e and Additional file 2: Fig. S8a–d). Next, we established a co-culture system and illustrated knockdown of tRF-3022b became much easier to promote M2 macrophage polarization (Additional file 2: Fig. S9a–g). Consistent with this result, qRT-PCR and flow cytometry results showed that tRF-3022b effectively modulates M2 macrophage polarization in vitro (Fig. 2f–h and Additional file 2: Fig. S9h).

Fig. 2
figure 2

Suppression of tRF-3022b inhibits subcutaneous tumor growth and induces M2 polarization via LGALS1 and MIF. a Representative images of subcutaneous xenograft tumors (n = 4 mice per group). b Tumor volume at indicated time points was calculated and plotted (n = 4 mice per group). c Final weight of subcutaneous xenograft tumors was measured and quantified (n = 4 mice per group). d Transmission electron micrograph (TEM) of exosomes isolated from HCT116 cell (scale bar = 200 nm). e Western blot analysis of TSG101, CD9, and E-cadherin (negative control) in exosomes. Cellular lysates were used as positive loading controls. f mRNA levels of various cytokines and markers in the M0 macrophages were measured by qRT-PCR analysis. g mRNA levels of M2 macrophage markers were measured by qRT-PCR. h Flow cytometry was used to quantify the expression of CD206, an M2 macrophage marker. i Cluster heat map of the DEGs between tRF-3022b-knockdown and negative control HCT116 cells. j Silver SDS-PAGE gel image shows proteins immunoprecipitated by the tRF-3022b probe in HCT116 and RKO cells. k–l Western blot analysis of products from RNA pull-down assays using the tRF-3022b probe and a control probe suggested LGALS1 and MIF in HCT116 and RKO cells. m Protein levels of LGALS1 in CRC cells with tRF-3022b knockdown. n Protein levels of MIF in CRC cells with tRF-3022b knockdown. o Immunofluorescence images of tRF-3022b with LGALS1 and MIF co-staining. Cell nuclei were counterstained with Hoechst. Scale bar: 20 μm. p Diagram outlining discussed hypotheses about the roles of tRF-3022b in the CRC tumor microenvironment. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

Full size image

Next, RNA sequencing analysis identified the target genes regulated by tRF-3022b and bioinformatics analysis (KEGG and IPA analysis) demonstrated they were involved in tight interactions with cytokine-associated tumor microenvironment. (Fig. 2i, and Additional file 2: Fig. S10a–d and Additional file 4: Tables S15–S17). Moreover, we performed RNA pull-down assay to detect tRF-3022b-associated proteins, and galectin 1 (LGALS1) and macrophage migration inhibitory factor (MIF) attracted our attention (Fig. 2j and Additional file 4: Tables S18–S20). Then, western blot, RNA immunoprecipitation (RIP)-qPCR, RNA FISH and immunofluorescence results collectively evidenced that the association of tRF-3022b with MIF and LGALS1 (Fig. 2k–o and Additional file 2: Fig. S10e, f). Notably, western blot showed that tRF-3022b could precipitate MIF in THP-1 lysates (Additional file 2: Fig. S10g). Finally, we overexpressed LGALS1 and MIF in CRC cells and validated their oncogenic functions in CRC [7, 8] (Additional file 2: Fig. S11a–f).

In summary, we demonstrate, for the first time, tRF-3022b is specifically overexpressed in CRC and the oncogenic role in CRC progression through a regulatory pathway by binding to LGALS1 and MIF in CRC cells. In addition, tRF-3022b modulates M2 macrophage polarization by binding to MIF (Fig. 2p) (Additional file 1).