Mechanisms of tRNA-derived fragments and tRNA halves in cancer treatment resistance

The tRNA-derived fragments (tRFs) and tRNA halves (tiRNAs) are newly discovered noncoding RNAs in recent years. They are derived from specific cleavage of mature and pre-tRNAs and expressed in various cancers. They enhance cell proliferation and metastasis or inhibit cancer progression. Many studies have investigated their roles in the diagnosis, progression, metastasis, and prognosis of various cancers, but the mechanisms through which they are involved in resistance to cancer treatment are unclear. This review outlines the classification of tRFs and tiRNAs and their mechanisms in cancer drug resistance, thus providing new ideas for cancer treatment.


Background
Transfer RNAs (tRNAs) have long been regarded as classic noncoding RNAs and are involved in protein translation [1]. Recently, many researchers have discovered new noncoding RNAs that are derived from specific cleavage of pre-and mature tRNA. Noncoding RNAs derived from tRNA are grouped into two categories: tRNA-derived fragments (tRFs) and tRNA halves (tiR-NAs) [1]. The tRFs originate from mature or pre-tRNAs, and they are approximately 14-30 nucleotides (nt) in length. tiRNAs are 29-50 nt in length and originate from specific cleavage of mature tRNA anticodon loop under stress [2,3]. The tRFs are conservative and widespread in nature [4]. They were initially discovered as random tRNA degradation fragments and later found to be generated by conservative and specific tRNA cleavage [5]. The tRFs are associated with cancer, inherited metabolic diseases, viral infections, and neurodegenerative diseases [6]. Here, we introduce the classification of tRFs and tiRNAs, discuss their roles in cancers, summarize their mechanisms of drug resistance in cancer treatment, and describe techniques for studying tRFs and tiRNAs.
Nutritional deficiency, hypoxia, heat shock, and oxidative stress can stimulate angiogenin activity and increase tRNA cleavage [11].
The tRFs and tiRNAs regulate ribosome and protein biogenesis; DNA damage response; gene expression; RNA interference; immune mediation; and cell proliferation, apoptosis, and migration [4,8,57]. They are expressed and dysregulated in various cancers [58] (Table 1). The heterogeneity and stability of tRFs make it suitable biomarkers for cancer diagnosis and prognosis [19]. Acquired treatment resistance is the main dilemma of cancer treatment [59]. Many studies have systematically discussed the roles of tRFs and tiRNAs in the diagnosis, progression, metastasis, and prognosis of various cancers [1,5], but few studies have investigated their roles in resistance to cancer treatment. Herein, we summarize their mechanisms in cancer drug resistance ( Table 2).

Mechanisms of tRFs and tiRNAs in cancer drug resistance
Drug resistance occurs when the therapeutic efficacy of a drug decreases. Drug resistance is an obstacle to cancer treatment and patient survival. The tRFs and tiR-NAs can replace eukaryotic translation initiation factor 4G (eIF4G) that binds to mRNA, thus inhibiting protein translation [41]. The downregulation of eIF4G increases doxorubicin sensitivity by inhibiting the expression of adenosine triphosphate-binding cassette (ABC) transporter in breast cancer cells [60]. ABC transporter can efflux anticancer drugs across cell membranes, which is associated with drug resistance of many solid tumors [13,61]. The inhibition of eIF4F complex, including eIF4E, eIF4G, and eIF4A, enhances the sensitivity of various anticancer drugs, such as cisplatin sensitivity in non-small cell lung cancer (NSCLC) [62], trastuzumab and tamoxifen sensitivity in breast cancer [63,64], and enzalutamide and bicalutamide sensitivity in castration-resistant prostate cancer (CRPA) [65]. In addition, tRFs and tiRNAs can promote the assembly of stress granules (SGs) under stress conditions [40]. SGs allow cells to recruit and protect mRNAs during stress [66]. SGs are related to drug resistance [67,68]. The elimination of SG sequestration is associated with acquired drug resistance [69]. The assembly of SGs makes glioblastoma resistant to bortezomib, and thus, the drug is unable to inhibit angiogenesis [66].
Sun et al. investigated tRFs related to trastuzumab resistance in human epidermal growth factor receptor-2 (HER2)-positive breast cancer [12]. tRF-30-JZOYJE22 RR33 and tRF-27-ZDXPHO53KSN were overexpressed in trastuzumab-resistant patients. These two tRFs may lead to trastuzumab resistance by regulating the expression product of target genes or competing with mRNAs for binding to RNA-binding proteins [12]. They are intervention targets for predicting trastuzumab resistance in the treatment of breast cancer.
In addition, Corce M et al. reported that ts-46 and ts-47 were upregulated by PIK3CA and KRAS mutations, respectively. These two mutations were involved in the resistance of breast cancer cells to lapatinib [26].
Chemoresistance causes high mortality of patients with breast cancer. The mechanisms by which tRFs induce chemoresistance in breast cancer can be applied to develop therapies to overcome drug resistance and prolong survival.

Mechanisms of lung cancer resistance related to tRFs and tiRNAs
Lung cancer has the highest mortality rate among all tumors. The 5-year survival rate is 19% [70]. Among all lung cancers, NSCLC accounts for more than 80% [107]. Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) usually produce incomplete response within 9-12 months of treatment and then develop resistance. Shao et al. found that tRF-Leu-CAG was significantly upregulated in advanced NSCLC and promoted cell proliferation. Knockdown of tRF-Leu-CAG downregulated Aurora-A kinase (AURKA) in H1299 cells, indicating that tRF-Leu-CAG may be related to AURKA [27]. AURKA is a member of the Aurora kinase family of serine/threonine kinases and is involved in cell mitosis [108]. AURKA phosphorylates p53 at Ser315, resulting in its ubiquitination by Mdm2 and proteolysis [109]. In HCC827 cells, AURKA overexpression induced gefitinib resistance by downregulating the p53 signaling pathway [54]. The combination of EGFR-TKIs and AURKA inhibitors can suppress resistance [110]. Furthermore, in vitro data showed that AURKA overexpression was related to cisplatin resistance [55]. In addition to targeted therapy, the AURKA/ NF-ĸB pathway is related to radio-resistance in docetaxelresistant lung adenocarcinoma [111]. These data indicate that tRF-Leu-CAG may become a potential therapeutic target to reverse drug resistance in NSCLC. (Fig. 3).
The P53 pathway mediates cisplatin resistance in NSCL C [129,130]. These data indicate that ts-101, ts-46, and ts-47 may be involved in drug resistance of lung cancer cells. (Fig. 3).

Mechanisms of pancreatic cancer resistance related to tRFs and tiRNAs
Pancreatic cancer (PC) has a high mortality rate, and the 5-year survival rate of patients with PC is only 8% [131]. Jin et al. reported that tRF-1391 and AS-tDR-000064 were associated with PC [28]. The KEGG pathway analysis showed that target genes of tRF-1391 were mainly concentrated in the PI3K/Akt pathway and 18 target genes of AS-tDR-000064 were related to resistance to EGFR-TKIs [28]. Metformin mediates PI3K/Akt/mTOR signaling, thereby enhancing the sensitivity of gemcitabine treatment in PC [132]. Furthermore, the PI3K/AKT signaling pathway is related to doxorubicin resistance in PC [56]. The roles of tRFs in drug resistance of PC need further research.

Mechanisms of colorectal cancer resistance related to tRFs and tiRNAs
Colorectal cancer (CRC) ranks third in morbidity and mortality among all tumors [70]. Chemoresistance leads to a low response rate to chemotherapy in advanced CRC [133]. Huang et al. reported that tRF/miR-1280 expression was reduced in CRC. tRF/miR-1280 was derived from tRNALeu and pre-miRNA. In addition, the miR-200b level decreased and Jagged-2 (JAG2), Gata1, Gata3, and zinc finger E-box binding homeobox 1 (Zeb1) levels increased in CRC [29]. JAG2 is a Notch ligand and is involved in tumor initiation and maintenance [134]. Gata factors are zinc finger DNA-binding proteins that control tissue development through the activation or repression of transcription [135]. tRF/miR-1280 binds to 3′-UTR of JAG2, thus inhibiting Notch/Gata and miR-200b signaling. The inactivation of Notch signal mediated by tRF/miR-1280 inhibits CSC phenotype through transcriptional repressing of Gata1/3 and miR-200b and further inhibits cell proliferation and metastasis in CRC [29]. Ectopic expression of JAG2 mediates drug resistance in CRC. JAG2 regulates the sensitivity of CRC cells to chemotherapeutic drugs through p21 [136]. Notch signaling regulates cell proliferation, differentiation, and apoptosis [137] and is related to chemoresistance of colorectal CSCs [138]. Notch-1 mediates regorafenib resistance [137]. Reversing Notch signaling sensitizes CRC cells to 5-fluorouracil (5-FU) and irinotecan [139]. In addition, tRF/miR-1280 downregulated the expression of Fig. 3 tRF-Leu-CAG, ts-101, ts-46, and ts-47 are involved in drug resistance of lung cancer cells. Knockdown of tRF-Leu-CAG downregulates AURKA. AURKA overexpression induces gefitinib resistance by downregulating p53. The AURKA/NF-ĸB pathway is related to radio-resistance. ts-46 is associated with excessive activation of ILK, integrin, and mTOR signaling. Both ts-46 and ts-101 are correlated with the inhibition of PTEN signaling. RNA silencing of ILK enhances cisplatin sensitivity by regulating p-GSK3β, p-AKT, AP-1, β-catenin, cyclin D1, and MMP-9. Integrin activates Akt, which determines resistance to cisplatin. Integrin mediates EGFR-TKIs resistance through EMT, the KRAS-RalB-NF-κB pathway, and the Src/Akt pathway. Integrin/Src/Akt in NSCLC cells confers resistance to cixutumumab. PI3K/AKT/mTOR signaling promotes chemoresistance. ts-46 and ts-47 can affect the p53 pathway, which mediates cisplatin resistance in NSCLC Gata1 and Gata3 through the Notch pathway. Gata binds to miR-200b to inhibit transcription initiation. Inhibition of Gata1 increases miR-200b expression and reduces EMT [29]. EMT and miR-200b are involved in chemoresistance of CRC [140,141]. Furthermore, Zeb1 and miR-200b regulate each other negatively. The genes ubiquitin-specific peptidase 17, chromodomain helicase DNA-binding protein 1-like, and double homeobox 4 are associated with DNA damage response. The inhibitory effect of Zeb1 on these three genes induces drug resistance [142]. In summary, tRF/miR-1280 may be involved in chemotherapy resistance of CRC. Drugs that targeting tRF/miR-1280 may reverse drug resistance in future treatment and prolong survival.
Wang et al. reported that tiRNA-Tyr-GTA and tRF-Gln-CTG were upregulated, while tRF-Leu-TAG was downregulated in CRC [30]. The tiRNA-Tyr-GTA, tRF-Gln-CTG, and tRF-Leu-TAG are tiRNA-5, tRF-5c, and tRF-5a with lengths of 30, 29, and 16 nt, respectively. The GO and KEGG pathway analysis indicated that the functions of tiRNA-Tyr-GTA targets are mainly negative regulation of epithelial cell apoptotic and peroxisome proliferator activated-receptor (PPAR) pathway [30]. PPAR is a ligand-activated transcription factor and is involved in regulating cancer progression [143]. The PPAR pathway is related to radiation resistance in CRC [144]. Inhibition of PPARα expression confers resistance to hydroxycamptothecin [145]. PPARδ confers resistance to PPARγ-induced apoptosis in CRC through mediation of survivin and caspase-3 [146]. In addition, tRF-Gln-CTG enhances the negative regulation of c-Jun N-terminal kinase (JNK) cascade and choline metabolism, while tRF-Leu-TAG enhances MET function [30]. Silencing of the JNK1 gene can dephosphorylate c-Jun and reduce transport of the G2 subfamily of ABC transporters (ABCG2). The JNK1/c-jun pathway is related to ABCG2mediated multidrug resistance in CRC [147]. JNK activation can confer resistance to 5-FU in CRC patients with p53 mutation by inducing Bcl-2 phosphorylation [148]. Taken together, tiRNA-Tyr-GTA, tRF-Gln-CTG, and tRF-Leu-TAG may be involved in chemoresistance of CRC.

Mechanisms of ovarian cancer resistance related to tRFs and tiRNAs
Ovarian cancer (OC) is the fourth most common gynecological malignant tumor with high drug resistance [149]. High grade serous ovarian cancer (HGSOC) accounts for 75% of OC cases. Zhang et al. identified 27 differentially expressed tRFs, such as tRF-03357 and tRF-03358, from serum samples of patients with HGSOC and healthy controls. These 27 tRFs are involved in the MAPK, FoxO, and Wnt pathways [40]. Most patients with advanced OC develop treatment resistance [150]. MAPK is associated with acquired chemoresistance of OC [151]. The C-KIT/MAPK/ MEK pathway establishes a link between platinum resistance and CSC phenotype in OC [152]. The Ras-MAPK/Erk-ETS1-ELK1/CFTR1 axis confers resistance to cisplatin [153]. The ERK/MAPK pathway promotes EMT phenotype, which is accompanied by increased cisplatin resistance [154]. The p38 MAPK activation upregulates p-glycoprotein expression, which enhances drug efflux, thereby inducing OC cell resistance to paclitaxel [155]. In addition, the activation of Wnt and its downstream molecules Wnt5a, β-catenin, c-Myc, and cyclin D1 upregulates the glycolysis level, thereby inducing cisplatin resistance in OC [156]. Fukumoto et al. showed that N6-methyladenosine of Frizzled family receptor 10 mRNA promoted resistance to poly (ADP)-ribose polymerase inhibitors (PARPi) by upregulating the Wnt/ β-catenin pathway in BRCA-mutated epithelial ovarian cancer (EOC) [157]. FoxO proteins have been reported to affect the effectiveness of anticancer drug treatment. Reduced FoxO1 expression is related to resistance to cisplatin [158]. To summarize, these differentially expressed tRFs affect the sensitivity of OC to chemotherapy, especially to cisplatin.
In addition, Zhou et al. reported that tRF5-Glu was expressed in OC. The tRF5-Glu binds to breast cancer antiestrogen resistance 3 (BCAR3) and downregulates its expression. BCAR3 and tRF5-Glu contribute to heterogeneity of OC [39]. Overexpression of BCAR induces antiestrogen resistance in breast cancer cells; however, it remains unclear whether BCAR has a similar effect in OC resistance [159,160].

Mechanisms of chronic lymphocytic leukemia resistance related to tRFs and tiRNAs
Balatti et al. found that ts-101, ts-53, ts-46, and ts-47 were downregulated in chronic lymphocytic leukemia (CLL) [25]. As mentioned before, ts-46 is associated with integrin and ceramide signaling. ts-101 is correlated with ceramide signaling inhibition [25]. The combination of fibronectin and α4β1 integrin induces fludarabine resistance in B-cell CLL by upregulating Bcl-xL [161]. Ceramide metabolism is related to fludarabine resistance in CLL [162]. In addition, the downregulation of ts-53 increases the TCL1 expression level, leading to the progression of CLL [42]. However, the role of TCL1 in chemoresistance of CLL remains unclear.

Conclusion
It is very common for tumors to acquire resistance to anticancer drugs. There is an urgent need to elucidate molecular mechanisms of cancer chemoresistance. This review outlines the classification of tRFs and tiRNAs, their roles in cancers, the mechanisms by which they play a role in cancer drug resistance, and the techniques for studying them. Their high equivalent and stability in body fluids, and differential expression between cancer patients and healthy controls open up broad prospects to develop screening, diagnostic, and prognostic biomarkers and targeted anticancer drugs that enhance the sensitivity of chemotherapeutic drugs. As the field is still in its infancy, it is a major challenge to fully understand the mechanism network of tRFs and tiRNAs in drug resistance. Overall, tRFs and tiRNAs give hope for cancer treatment, but further studies are required before clinical application. More work needs to be done to study broader and deeper mechanisms of tRFs and tiRNAs in drug resistance and the crosstalk between various signaling pathways.