Introduction
According to the latest cancer statistics data, breast cancer (BC) is the most common malignant tumor in women, accounting for 31% of the total number of female cancers, and the incidence has gradually increased in recent years [
1,
2]. With continuous improvement in diagnosis and treatment of BC, early- stage patients have good prognosis, with an overall cure rate of 90% [
3]. However, the 5-year survival rate is significantly reduced for patients with advanced stage, poor tissue typing or resistance to combined therapy [
4]. Therefore, clarifying the specific molecular mechanism of the occurrence and development of BC to assist in early diagnosis, as well as to find new and more accurate targeted molecules, is crucial for improving the overall survival rate of BC patients.
In recent years, positive roles of a variety of non-coding RNAs (ncRNAs), including micro RNAs (miRNAs), long non-coding RNAs (lncRNAs), circular RNAs (circRNAs) and tRNA-derived small RNAs (tsRNAs), in cancer have received much attention and been widely reported; some have been included in clinical application, fully demonstrating the huge potential of ncRNAs in tumor diagnosis and treatment [
5‐
7]. TsRNAs are the products of tRNA or pre-tRNA cleavage during maturation. TsRNAs can be divided according to different fracture sites into tRNA-related fragments (tRFs) and tRNA halves (tiRNAs), which complement each other in terms of formation pathway, cell localization and function [
8,
9]. Although tsRNAs were initially thought to be produced by random degradation and have no special function, an increasing number of studies have shown that tsRNAs are not only widely expressed in tumors but also play important roles in the occurrence and development of tumors [
10]. In addition, tsRNAs are present in a large number of human body fluids, such as saliva, urine, and semen, providing a new target for early tumor diagnosis and precise treatment [
11].
Studies have shown that tsRNAs are expressed specifically in BC and widely exist in the serum of BC patients. TsRNAs differentially expressed in different subtypes of BC, including triple-negative breast cancer, have been mined through RNA sequencing and verified with clinical samples, suggesting the significance of tsRNAs in identifying tumor properties and tumor tissue types [
12‐
14]. In terms of mechanism research, tsRNAs are significantly enriched in hormone-dependent BC and are not only directly involved in regulating the occurrence and development of BC but also closely related to drug resistance and tumor recurrence in BC [
15‐
17]. Overall, tsRNAs are good detection markers and potential therapeutic targets for BC patients. Therefore, we summarized existing studies on BC and tsRNAs and used meta-analysis to explore the roles and values of tsRNAs in the prognosis, diagnosis and clinicopathological features of BC patients to provide evidence for future applications of tsRNAs.
Methods
Publication search
To collect relevant research literature for meta-analysis, the Boolean logic method was used to search PubMed, Web of Science, Cochrane Library, Embase and other databases as of March 1, 2023. Specific search terms included “breast cancer”, “breast carcinoma”, “tsRNA”, “tRNA-derived small RNA”, “tRF”, “tRNA-derived fragment”, “tiRNA” and “tRNA-derived stress-induced RNA”.
Inclusion and exclusion criteria
Inclusion criteria: (1) the content included the relationship between tsRNAs and BC; (2) all cases passed the gold standard, that is, the histopathological diagnosis was clear; and (3) data related to prognosis, diagnosis or clinicopathological features could be extracted directly or indirectly. Exclusion criteria: (1) the content was not related to tsRNAs or BC; (2) reviews, meta-analyses, letters, case reports, conference abstracts; (3) non-English and nonhuman studies; and (4) failure to report or extract important indicators and data. Two researchers independently read the title and abstract of the retrieved literature for preliminary screening; in cases in which the abstract results were not clear, the full text was read to determine inclusion eligibility. If the screening results were inconsistent, the third researcher decided.
The general data extracted from each article included the following: title, first author, tsRNAs, year, country, expression, cut-off, sample size, data source, detected sample, among others. Data related to prognosis included follow-up time, survival outcome (progression-free survival (PFS), recurrence-free survival (RFS), disease-free survival (DFS), overall survival (OS)), survival analysis, hazard ratio (HR), and 95% confidence interval (95%CI), among others. Data related to diagnosis included case, control, area under the receiver operating characteristic curve (AUC), sensitivity (SEN), specificity (SPE), true positive (TP), false positive (FP), true negative (TN), false negative (FN), positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and cancer type, among others. Data related to clinicopathologic features included age, TNM stage, and lymphatic metastasis, among others. All records were approved by two independent researchers.
Quality assessment
As the quality of studies may influence the results of meta-analysis, each included study was evaluated using the Newcastle‒Ottawa score (NOS) tables in the Cochrane manual and Quality Assessment for Studies of Diagnostic Accuracy 2 (QUADAS 2). After all evaluations, RevMan 5.3 software was used to output the evaluation results. NOS scores ≥ 7 or QUADAS 2 scores ≥ 4 are generally considered to be high quality [
18,
19].
Data synthesis and analysis
Statistical analysis of the data was performed with Stata software (version 15.1). HRs, odds ratios (ORs), and 95%CIs were used to evaluate the effect of tsRNAs on prognosis and clinicopathologic features in BC patients. To evaluate the diagnostic value of tsRNAs, summary receiver operating characteristic (SROC) curves were drawn, and AUCs were calculated, as were Fagan nomograms and scatter plots. For studies in which Kaplan–Meier (KM) and ROC curve data could not be extracted directly, we used Engage Digitizer and GetData Graph Digitizer and calculated HR and the corresponding 95%CI based on the method of Tierney et al. [
20‐
22]. The Cochran-Q test and
I2 statistics were used to assess heterogeneity. When the heterogeneity was small (
P > 0.10 or
I2 < 50%), a meta-analysis was conducted using the fixed effects model. Otherwise, random-effects models were employed to combine effect sizes, and further subgroup analysis and meta-regression (MetaDiSc software) were applied to explore possible sources of interstudy heterogeneity. Sensitivity analysis was carried out by excluding the included studies one by one. Finally, Begg’s funnel plot, Deeks’ funnel plot and Egger’s test were used to evaluate publication bias, and
P > 0.05 was considered to indicate no publication bias. All statistical tests were bilateral, and P < 0.05 was considered statistically significant.
Discussion
BC is the malignant tumor with the highest incidence in the female population at present [
35,
36]. Its causes are still unclear, and some types lack effective intervention targets, resulting in a high fatality rate. Traditional BC detection methods, such as CEA, CA153, B-ultrasound, and molybdenum targeted therapy, among others, are relatively less traumatic than other methods but have low sensitivity and specificity for early diagnosis, and pathology, the gold standard of diagnosis, is difficult to popularize in the population without disease [
37‐
40]. Therefore, exploration of more effective markers for early diagnosis and precise treatment is crucial for improving the poor prognosis of BC patients.
As emerging ncRNAs, abnormal expression of tsRNAs has been confirmed in a variety of diseases, including cancers, and plays an important biological role and function [
41,
42]. As tRNA fragments, tsRNAs are characterized by low molecular weight and stable expression. In addition, tsRNAs are widely expressed and specific in human tissues and body fluids. These characteristics all render tsRNAs important measures in cancer diagnosis and prognosis prediction [
43,
44]. Since Hani et al. discovered that tRFs derived from tRNA-Glu, tRNA-Asp, tRNA-Gly, and tRNA-Tyr competitively bind to the endogenous oncogene YBX1 in BC cells, inhibiting cell growth by interfering with oncogene transcription stability, the role and mechanism of tsRNAs in BC occurrence and development have been extensively explored [
15]. Wang et al. conducted tsRNA expression sequencing on plasma samples from 8 BC patients and 4 healthy women, and found that the expression level of tRF-Glu-CTC-003 in BC patients' plasma was lower compared to that in healthy women, and the expression level of tRF-Glu-CTC-003 in TNBC patients' plasma was lower than in other subtypes [
31]. These differential results suggest that tsRNA may serve as a potential biomarker for BC. Detecting tsRNA levels in plasma can aid in early BC diagnosis and patient prognosis evaluation. Mechanistically, tsRNAs primarily regulate protein expression at different stages by binding to other molecules, thereby affecting protein biosynthesis through transcription or post-transcriptional processes, and playing a regulatory role in BC. Maurizio et al. found significantly lower expression levels of tRF3E derived from mature tRNA-Glu in BC tissues compared to normal tissues, which could inhibit BC progression by binding to the RNA-binding protein NCL [
45]. Zhu et al. identified high expression of tRF-Lys-CTT-010 in TNBC, demonstrating its ability to promote cell proliferation and migration, participate in metabolic pathways, and regulate cell survival and proliferation by manipulating lactic acid production and glycogen consumption [
46]. Mo et al. discovered the inhibitory effects of tRF-17-79MP9PP on BC cell invasion and metastasis via the THBS1/TGF-β1/Smad3 axis [
30]. These findings collectively indicate that tsRNA plays a crucial role in BC occurrence and development, offering potential avenues for precise BC treatment.
To our knowledge, this is the first meta-analysis exploring the value of tsRNAs in BC, with a view to providing evidence-based medical evidence for future clinical applications of tsRNAs. According to our inclusion and exclusion criteria, 13 studies on the correlation between tsRNAs and BC were identified as eligible for meta-analysis, and no studies that did not meet the criteria were found after quality assessment. In the study on the correlation between tsRNAs and BC patient prognosis, we included 7 studies involving a total of 5257 patients. The results showed that tsRNA expression correlated positively with prognostic indicators (PFS, RFS, DFS, OS), which was consistent with the study on the mechanism of differentially expressed tsRNAs playing a role in promoting or suppressing BC. It should be noted that the cut-off method for classifying tsRNA expression was not unified among the included studies, and there is still a lack of quantitative analysis standards for tsRNAs, posing challenges for practical clinical application of tsRNAs in the future. In addition, the vast majority of studies focused on Asia, and more solid research is needed to confirm whether our results are biased by race and/or region. Wang et al. found that the AUC of the combination of six tsRNAs in the plasma of early-stage BC patients was 0.844, superior to any single identified tsRNA [
32]. In addition, Zhang et al. reported that when three tsRNAs were combined with the traditional tumor markers CEA, CA125 and CA153, the AUC increased to 0.801 [
33]. These results suggest that tsRNAs have reliable potential for BC diagnosis. Therefore, we systematically reviewed studies on the diagnostic value of tsRNAs in BC, aiming to identify biomarkers that can be used for diagnosis. Eventually, 7 studies involving 21 different tsRNAs were included in our analysis. After meta-analysis of all tsRNAs, the combined SEN was 72% (95%CI 68–76%), the combined SEP was 64% (95%CI 61–67%), and the AUC of the SROC curve was 0.72 (95%CI 0.68–0.75), indicating that tsRNAs have good diagnostic efficacy. As an independent prognostic indicator, the DOR value can indicate the degree of association between diagnosis results and diseases, whereby a higher value indicates a more reliable diagnosis. In diagnostic assessment of tsRNAs, the combined DOR reached 4.62 (95 CI 3.76–5.68), further indicating reliable accuracy of tsRNA diagnosis in BC. Finally, Fagan's nomogram was used to analyze the clinical value of tsRNAs. The results showed that when the pretest probability was set to 20%, the probability of BC in positive tsRNA test results increased to 30% and that in negative tsRNA test results decreased to 10%. Regarding clinicopathological studies, 3 studies were included in total. The results showed that low expression of tsRNAs correlated significantly with age, TNM stage and lymphatic metastasis, suggesting a role in inhibiting BC progression. In addition, we found a certain publication bias in the included studies, which affected the accuracy of the results obtained; the reason may be that there were few studies and that some of them had small sample sizes. For example, in the study of Wang et al. ts-32-Q99P9P9NH57SJ and ts-17-79MP9PP showed low expression in BC, though the trend in lymphatic metastasis was the opposite, which may also be related to the small number of samples (
n = 16) [
28]. In conclusion, we believe that tsRNAs may become a new target in BC treatment and provide new ideas for BC treatment. However, more literature needs to be examined to obtain reliable data support.
At present, research on tsRNAs is still in the initial stage, and the specific molecular mechanism of the role of a large number of differentially expressed tsRNAs in tumors, especially BC, has not been clarified. Nevertheless, we hope that our meta-analysis will be helpful for future clinical applications of tsRNAs. Of course, there are several limitations of this study that should not be ignored. First, as the patients included were mostly from China, the conclusion may not be applicable to different regions or populations. Second, some data and their 95%CIs were obtained indirectly through software, which may lead to deviation from the real data, resulting in certain publication bias. In addition, there are few related studies on different subtypes, especially triple-negative breast cancer; thus, the positive role of tsRNAs in the early diagnosis and prognosis of these patients should be explored. Finally, comprehensive treatment tolerance is one of the main factors affecting the survival of BC patients at present, but there are few studies focusing on such factors. Future studies with more and larger samples should be designed and performed to guide individualized clinical therapy (Additional file
6: Figure S6).
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.