High Intratumoral i-tRF-GlyGCC Expression Predicts Short-Term Relapse and Poor Overall Survival of Colorectal Cancer Patients, Independent of the TNM Stage

Colorectal cancer (CRC), one of the most prevalent types of cancer, requires the discovery of new tumor biomarkers for accurate patient prognosis. In this work, the prognostic value of the tRNA fragment i-tRF-GlyGCC in CRC was examined. Total RNA extraction from 211 CRC patient cancer tissue specimens and 83 adjacent normal tissues was conducted. Each RNA extract was subjected to in vitro polyadenylation and reverse transcription. A real-time quantitative PCR assay was used to quantify i-tRF-GlyGCC in all samples. Extensive biostatics analysis showed that i-tRF-GlyGCC levels in CRC tissues were significantly lower than in matched normal colorectal tissues. Additionally, the disease-free survival (DFS) and overall survival (OS) time intervals were considerably shorter in CRC patients with high i-tRF-GlyGCC expression. i-tRF-GlyGCC expression maintained its prognostic value independently of other established prognostic factors, as shown by the multivariate Cox regression analysis. Additionally, survival analysis after TNM stage stratification revealed that higher i-tRF-GlyGCC levels were linked to shorter DFS time intervals in patients with TNM stage II tumors, as well as an increased probability of having a worse OS for patients in TNM stage II. In conclusion, i-tRF-GlyGCC has the potential to be a useful molecular tissue biomarker in CRC, independent of other clinicopathological variables.


Introduction
Colorectal cancer (CRC) is a significant issue for public health, accounting for the third most prevalent cancer diagnosis and the leading cause of cancer-related deaths globally [1]. Less than half of cases are diagnosed when the cancer is locally advanced. Nowadays, most cases are often discovered at a later stage due to the extensive use of semi-invasive endoscopic techniques and fecal blood testing, both of which have subpar diagnostic accuracy [1,2]. Despite progress in understanding the molecular and cellular underpinnings of CRC, early identification remains difficult due to the lack of symptoms in the early stages, even though the prevalence of early onset colorectal cancer, or colon cancer diagnosed in patients under the age of 50, has been rising [3,4]. Moreover, the general global incidence of CRC is anticipated to increase in the upcoming decade despite improvements in diagnostic tools and treatment approaches, highlighting the urgent need for new prognostic markers and customized therapy approaches based on molecular biomarkers [5].
Biomedicines 2023, 11, 1945 2 of 13 The most accurate prognostic indicator for CRC patients continues to be the TNM staging system, which is based on the extent of tumor invasion depth (T), lymph node infiltration (N), and presence or absence of distant metastasis (M) [6]. Tumors that are confined to the innermost layers of the colon or rectum and have not spread to the surrounding tissues are referred to as stage I CRC [7]. Because the cancer is confined and has not yet moved past the initial site, stage I CRC has a good prognosis [8]. The migration of the tumor into adjacent tissues through the colon or rectum wall is a hallmark of stage II CRC. Due to the possibility of microscopic cancer cell spread, stage II CRC has a higher chance of recurrence than stage I [9]. Stage III CRC denotes the absence of distant metastases but the presence of local lymph node involvement. In comparison to earlier stages, stage III CRC has a higher chance of recurrence and a worse prognosis [7,9].
However, the usefulness of TNM staging in actual clinical practice, particularly in detecting high-risk stage II patients, is constrained [10]. Moreover, precise biomarkers capable of discriminating between stage II and stage III cancers and accurately predicting patient relapse are severely lacking [11,12]. Furthermore, it is critical to accurately predict patient relapse in order to inform postsurgical therapy choices and surveillance strategies. Current prognostic indicators have some predictive value but fall short of accurately predicting the chance of relapse [13]. A very promising area with several benefits is realtime prognosis using liquid samples, such as circulating tumor cells or circulating free DNA in the blood [14,15]. Nonetheless, given the especially heterogeneous nature of CRC among cancer patients, there is an urgent need for additional research to identify novel prognostic, diagnostic, and predictive biomarkers [16].
Recently, tRNA-derived RNA fragments (tRFs) have also emerged as a novel frontier with potential diagnostic and prognostic importance among the numerous classes of ncRNAs [43,44]. These fragments are endogenous single-stranded ncRNAs ranging in length from 14 to 40 nucleotides. Notably, tRFs appear as significant regulators of gene expression acting similarly to other members of ncRNAs class, such as miRNAs. Furthermore, they post-transcriptionally regulate stability or translation of carcinogenic transcripts, leading to tumor suppression [45,46]. tRFs have been shown to significantly regulate cancer, hematologic malignancies, disorders of metabolism, inflammation, infections from viruses, and diseases of the nervous system [44]. Their expression is deregulated in cancer and hematological malignancies; this fact, along with their abundant presence in bodily fluids, renders them as molecular biomarkers in CRC [43,47]. Recent evidence suggests that particular tRFs are deregulated in this malignancy, impacting important pathways involved in cancer development. For instance, tRF-3022b regulates colorectal cell apoptosis and M2 macrophage polarization by binding to cytokines, and tRF3008A inhibits colorectal cancer development and metastasis by disrupting FOXK1 in an AGO-dependent way [48,49]. An intriguing tRF that derives from the internal part of the mature tRNA (internal tRF, i-tRF), bearing the glycine "GCC" anticodon, is i-tRF-Gly GCC . The oncogenic role of this tRF has already been established in ovarian cancer, where it is associated with early progression and poor overall survival, and in chronic lymphocytic leukemia, where its significance as an independent unfavorable biomarker was uncovered [50,51]. The aforementioned evidence prompted us to assess the potential of i-tRF-Gly GCC expression in colorectal tumors as a prognostic molecular biomarker. To achieve this goal, we applied a real-time quantitative polymerase chain reaction (qPCR) assay for the relative quantification of i-tRF-Gly GCC levels in colorectal cancer tissue specimens and adjacent normal colorectal tissues, using the comparative C t method for calculations.

Collection of Colorectal Tissue Samples
The present study included tissue specimens from 211 patients with primary CRC, operated at the University General Hospital "Attikon", from 2009 to 2019. All tissue samples were histologically evaluated by a pathologist and immediately frozen in liquid nitrogen. Normal colorectal tissue samples were acquired from 83 cases.
This study was approved by the Ethics Committee of the University General Hospital "Attikon" (number of approval: 13; date 29 January 2009), according to the guidelines of the Declaration of Helsinki. All patients were informed about the scope of the research and provided their consent.

Clinical Characteristics of CRC Patients
This study included 211 tissue samples of primary CRC and 83 adjacent normal colorectal tissue samples. In total, 108 male and 103 female CRC patients were included in this study. Patients had a median age of 66 years (interquartile range: 57-72 years) at the time of diagnosis. The clinical features of CRC patients shown in Table 1 include tumor size, histological grade, and TNM stage. According to the revised TNM classification system, patients are classified by taking into account the invasion of tumor (T), the infiltration of regional lymph nodes (N), and the potential presence of distant metastases (M) [52]. Moreover, information about treatment of CRC patients is presented in Table 1. Survival data were available for all patients included in the current study; however, 28 of them presented distant metastasis (M1) at the time of surgery and were hence excluded from DFS analysis. The follow-up information included the date of disease recurrence diagnosis; the date and cause of death were also recorded for those patients who succumbed to their disease during the follow-up period.

Total RNA Extraction and Polyadenylation, Followed by First-Strand cDNA Synthesis
The DLD-1 colorectal adenocarcinoma cell line was purchased from the American Type Culture Collection (ATCC ® ) and cultured according to the ATCC guidelines. Colorectal tissue homogenization followed, and total RNA extraction was performed from DLD-1 cells and each tissue specimen using TRI Reagent ® (Molecular Research Center, Inc., Cincinnati, OH, USA). RNA was diluted in Storage Solution (Life Technologies Ltd., Carlsbad, CA, USA), and its concentration and purity were assessed spectrophotometrically at 260 and 280 nm with a BioSpec-nano microvolume UV-Vis spectrophotometer (Shimadju, Kyoto, Japan). All total RNA extracts were stored at −80 • C prior to their polyadenylation with E. coli poly(A) polymerase and reverse transcription into first-strand cDNA starting next to an oligo-dT adapter primer [53].

SYBR Green Based Real-Time Quantitative PCR (qPCR)
A real-time quantitative PCR method, based on SYBR Green chemistry, was applied as previously described, to perform relative quantification of i-tRF-Gly GCC . The comparative Ct (2 −∆∆Ct ) method was applied for all calculations to determine the tissue levels of this small ncRNA in each tissue sample [54]. SNORD43 and SNORD48 were used as internal reference genes to normalize the i-tRF-Gly GCC expression levels; the DLD-1 cell line extract served as a calibrator in the real-time qPCR. All primers that were used were gene-specific, as previously described [51,55]. Normalized expression values of this tRF were expressed in relative quantification units (RQU).

Extended Biostatistics, including Disease-Free and Overall Survival Analyses
Non-parametric statistical tests were used in the biostatistics analysis. In particular, the Wilcoxon signed-rank test was used to assess the statistical significance of difference of i-tRF-Gly GCC expression levels in pairs of CRC and normal adjacent tissues; differences of i-tRF-Gly GCC expression levels among subgroups of patients (based on each clinicopathological factor) were checked with the Jonckheere-Terpstra test.
A receiver operating characteristic (ROC) curve was built by plotting sensitivity versus (1-specificity), and the area under the curve (AUC) was calculated. Logistic regression analysis was also performed to assess the potential of i-tRF-Gly GCC expression to predict CRC occurrence.
In order to assess the prognostic value of i-tRF-Gly GCC expression in CRC, we constructed Kaplan-Meier disease-free survival (DFS) and overall survival (OS) curves; for this purpose, this continuous variable was split at the median value. Stratified Kaplan-Meier survival analyses were also conducted. The differences between the curves were evaluated with the Mantel-Cox (log-rank) test. To evaluate the prognostic potential of i-tRF-Gly GCC expression and determine the hazard ratio (HR) for patients' relapse and disease-related death, bootstrapped Cox regression analyses were carried out with 1000 bootstrap samples. The bootstrap bias-corrected and accelerated (BCa) method was implemented to calculate bootstrap p values and 95% confidence intervals (CIs) for each estimated HR. Furthermore, multivariate prognostic models were built and adjusted for the most important clinicopathological characteristics and type of treatment each patient received. Only p values lower than 0.050 (p < 0.050) were considered as statistically significant in each statistical test.

i-tRF-Gly GCC Expression Is Downregulated in CRC Tissues, Compared to Adjacent Normal Colorectal Tissues
i-tRF-Gly GCC levels in CRC specimens ranged from 0.001 to 3.4 RQU with a mean ± SEM of 0.44 ± 0.036, and from 0.002 to 1.1 RQU with a mean ± SEM of 0.36 ± 0.031 in noncancerous specimens ( Table 2). Although the distribution in the two cohorts is quite similar, i-tRF-Gly GCC levels were downregulated in the vast majority (58 out of 83) of the malignant tumors, compared to their matched normal tissue specimens, unraveling the utility of i-tRF-Gly GCC expression levels for screening purposes (Figure 1). strap samples. The bootstrap bias-corrected and accelerated (BCa) method was implemented to calculate bootstrap p values and 95% confidence intervals (CIs) for each estimated HR. Furthermore, multivariate prognostic models were built and adjusted for the most important clinicopathological characteristics and type of treatment each patient received. Only p values lower than 0.050 (p < 0.050) were considered as statistically significant in each statistical test.

i-tRF-Gly GCC Expression Is Downregulated in CRC Tissues, Compared to Adjacent Normal Colorectal Tissues
i-tRF-Gly GCC levels in CRC specimens ranged from 0.001 to 3.4 RQU with a mean ± SEM of 0.44 ± 0.036, and from 0.002 to 1.1 RQU with a mean ± SEM of 0.36 ± 0.031 in noncancerous specimens (Table 2). Although the distribution in the two cohorts is quite similar, i-tRF-Gly GCC levels were downregulated in the vast majority (58 out of 83) of the malignant tumors, compared to their matched normal tissue specimens, unraveling the utility of i-tRF-Gly GCC expression levels for screening purposes (Figure 1).

i-tRF-Gly GCC Overexpression Represents a Reliable Indicator of Poor Prognosis in CRC
A total of 61 (33.3%) out of the 183 patients who were included in the disease-free survival (DFS) analysis exhibited tumor recurrence observed during the accrual follow-up period. Similarly, 94 (44.5%) deaths associated with CRC occurred during the follow-up period. The median follow-up time was 93 months.
In order to assess the prognostic value of i-tRF-Gly GCC , we categorized the CRC patients into two groups, those with low i-tRF-Gly GCC levels and those with higher ones, by splitting at the median the i-tRF-Gly GCC expression in cancerous samples (cut-off point: 0.50 RQU). Kaplan-Meier survival analysis revealed that patients with high i-tRF-Gly GCC levels have significantly shorter DFS (p < 0.001) and OS (p = 0.007) intervals, compared to patients with low i-tRF-Gly GCC levels ( Figure 2). These results were also confirmed by univariate Cox regression analysis, in which a hazard ratio (HR) of 2.39 (p < 0.001) was calculated for disease recurrence in patients with high i-tRF-Gly GCC expression (Table 3), with an HR of 1.79 (p = 0.003) for CRC-related death in the same group of patients (Table 4), compared to CRC patients presenting with low intratumoral i-tRF-Gly GCC expression.
A total of 61 (33.3%) out of the 183 patients who were included in the disease-free survival (DFS) analysis exhibited tumor recurrence observed during the accrual followup period. Similarly, 94 (44.5%) deaths associated with CRC occurred during the followup period. The median follow-up time was 93 months.
In order to assess the prognostic value of i-tRF-Gly GCC , we categorized the CRC patients into two groups, those with low i-tRF-Gly GCC levels and those with higher ones, by splitting at the median the i-tRF-Gly GCC expression in cancerous samples (cut-off point: 0.50 RQU). Kaplan-Meier survival analysis revealed that patients with high i-tRF-Gly GCC levels have significantly shorter DFS (p < 0.001) and OS (p = 0.007) intervals, compared to patients with low i-tRF-Gly GCC levels ( Figure 2). These results were also confirmed by univariate Cox regression analysis, in which a hazard ratio (HR) of 2.39 (p < 0.001) was calculated for disease recurrence in patients with high i-tRF-Gly GCC expression (Table 3), with an HR of 1.79 (p = 0.003) for CRC-related death in the same group of patients (Table 4), compared to CRC patients presenting with low intratumoral i-tRF-Gly GCC expression.

The Prognostic Signficance of i-tRF-Gly GCC Expression Is Independent of Other Classical Prognostic Factors Applied in CRC
In the multivariate Cox regression analysis, the importance of the i-tRF-Gly GCC expression status in the prognosis of the patients' DFS remained unaffected (HR = 2.64; p = 0.004), even when combined with the tumor size, histological grade, depth of tumor invasion, regional lymph node status, and treatment with radiotherapy and/or chemotherapy (Table 3). i-tRF-Gly GCC retained its prognostic significance regarding OS, as well (HR = 1.56; p = 0.046), when combined with the aforementioned classical prognosticators plus the presence or absence of distant metastases (Table 4).

i-tRF-Gly GCC Overexpression Predicts Tumor Recurence and Poor Prognostic Outcome in CRC Patients in TNM Stage II
After stratification, according to the most important prognostic factor used for CRC prognosis, namely, the TNM stage, patients with TNM stage II colorectal tumors overexpressing i-tRF-Gly GCC showed remarkably shorter DFS intervals (p = 0.003) compared to patients in the same TNM stage and with low i-tRF-Gly GCC levels ( Figure 3A).

The Prognostic Signficance of i-tRF-Gly GCC Expression Is Independent of Other Classical Prognostic Factors Applied in CRC
In the multivariate Cox regression analysis, the importance of the i-tRF-Gly GCC expression status in the prognosis of the patients DFS remained unaffected (HR = 2.64; p = 0.004), even when combined with the tumor size, histological grade, depth of tumor invasion, regional lymph node status, and treatment with radiotherapy and/or chemotherapy (Table 3). i-tRF-Gly GCC retained its prognostic significance regarding OS, as well (HR = 1.56; p = 0.046), when combined with the aforementioned classical prognosticators plus the presence or absence of distant metastases (Table 4).

i-tRF-Gly GCC Overexpression Predicts Tumor Recurence and Poor Prognostic Outcome in CRC Patients in TNM Stage II
After stratification, according to the most important prognostic factor used for CRC prognosis, namely, the TNM stage, patients with TNM stage II colorectal tumors overexpressing i-tRF-Gly GCC showed remarkably shorter DFS intervals (p = 0.003) compared to patients in the same TNM stage and with low i-tRF-Gly GCC levels ( Figure 3A).  Furthermore, patients of TNM stage II with increased i-tRF-Gly GCC levels showed an elevated probability of a poorer OS (p < 0.001), in comparison with patients of the same TNM stage and low i-tRF-Gly GCC expression ( Figure 3B). However, no statistically significant results were obtained for patients in other TNM stages.

Discussion
Given that stage I colon cancer patients have a high five-year relative survival rate, early identification of CRC is essential for patient survival. According to the American Cancer Society, this ratio lowers significantly for later-stage CRC patients. Moreover, a lot of therapeutic efforts have been directed towards developing a better and more precise classification of patients due to the variable survival outcomes and treatment responses [56]. During the last 5 years, four consensus molecular CRC subtypes (CMS) were characterized by an international collaboration of expert groups, demonstrating considerable interconnectedness between six independent classification systems [57]. Additionally, clinical researchers have focused on identifying molecular biomarkers that could potentially be employed in clinical practice for early diagnosis and reliability in CRC prognosis.
Despite being in its infancy, research on tRNA derivatives represents a research subject that has attracted scientific interest since the advent of small RNA sequencing and other innovative methods. Their involvement in numerous molecular and cellular processes that lead to the cancerous phenotype and include the dysfunctionality of transcription, cell proliferation, and differentiation has emphasized their role as potential biomarkers and therapeutic targets [58]. Furthermore, several studies have sought to link their abnormally expressed levels in cancer with the malignant phenotype [59].
It is hypothesized that tRFs, like other small ncRNAs, constitute key players in the epigenetic regulation of the protein-coding genes [60]. This observation has been proposed by several studies that showed a close association between particular tRFs and AGO or PIWI proteins in the time when gene silencing occurred in a variety of types of human specimens [61][62][63]. As a result, there is growing curiosity about their biomarker utility in solid tumors and hematological malignancies [43,51,[64][65][66][67]. High-throughput genomic studies that specifically examined CRC have found significant changes in tRNA-derived small RNA levels between malignant and benign colorectal cancers [68,69]. In one of the aforementioned works, the authors discovered two distinct tRF signatures that may distinguish between colorectal adenomas and adenocarcinomas. Another study noted that 5 -tiRNA-Pro TGG is associated with poor survival of CRC patients and disease recurrence [70].
The tRNA molecules bearing the glycine "GCC" anticodon represent the origin of the internal tRF investigated in this original study. The biomarker utility of the i-tRF-Gly GCC molecule has been previously proposed for chronic lymphocytic leukemia and multiple myeloma [51,67], while 5 -tRF-Gly GCC , which derives from the same tRNAs, has been proposed as a putative discriminatory biomarker for CRC [71]. Prompted by these, in this study, we explored the biomarker potential of i-tRF-Gly GCC in CRC, since it is a malignancy for which the identification of reliable biomarkers would improve patients' prognosis and individualized treatment options [72].
Our findings showed that i-tRF-Gly GCC levels were considerably lower in CRC tissues compared to paired noncancerous tissues. Considering that the majority of malignant tumors had lower levels of i-tRF-Gly GCC , this discovery implies that it could be used as a discriminatory biomarker for CRC. Furthermore, high i-tRF-Gly GCC levels were related to a poor prognosis in CRC patients. When compared to patients with lower i-tRF-Gly GCC levels, those with greater i-tRF-Gly GCC levels had significantly shorter disease-free survival (DFS) and overall survival (OS) intervals. These data suggest that i-tRF-Gly GCC could be used as a valid prognostic biomarker for CRC, assisting in the identification of patients at increased risk of disease recurrence and mortality from CRC.
Additionally, we investigated the prognostic utility of i-tRF-Gly GCC in CRC patients according to their TNM stage. After grouping the patients based on TNM stage, we found that patients in TNM stage II had significantly shorter DFS intervals when the intratumoral i-tRF-Gly GCC levels were high. Similar to this, patients with higher i-tRF-Gly GCC levels and TNM stage II were more likely to have inferior OS, compared to those with TNM stage II tumors with low levels of this tRF. These results imply that i-tRF-Gly GCC may be an effective prognostic marker for the course of the disease, particularly in specific stages of CRC. Patients with CRC at the TNM II stage frequently have considerably different survival probabilities. In order to develop stratification systems that more effectively define the clinical stratification and predict the course of the disease, studies have concentrated on the molecular pathways underlying CRC. CMS was developed as a first step towards the development of a new stratification system, as a recently developed stratification technique that is based on the biological characteristics of CRC patients [57]. CMS has recently been proposed by Purcell et al. since it improves the prognosis for CRC patients at the TNM II stage [73]. These data highlight the necessity for novel biomarkers, able to further stratify CRC patients, especially in groups with high heterogeneity [74].
Moreover, our findings provide significant insights both for the scientific community and regarding the clinical utility of i-tRF-Gly GCC . Firstly, these results add to the expanding body of evidence supporting the importance of tRNA-derived fragments as possible biomarkers in cancer. We provide valuable insights into the discriminatory and prognostic importance of the i-tRF-Gly GCC fragment in CRC by particularly examining its expression levels in CRC and paired normal samples, as well as its prognostic value. This contributes to a better understanding of the complicated molecular pathways that underpin CRC formation and progression. In terms of clinical application, i-tRF-Gly GCC levels in CRC patients might be measured as part of normal discriminatory procedures. The downregulation of i-tRF-Gly GCC in CRC tissues implies that it could be used as a noninvasive biomarker for CRC identification and screening [75]. Furthermore, the predictive value of i-tRF-Gly GCC emphasizes its usefulness in stratifying patients based on their risk of disease recurrence and CRC-related death. More specifically, incorporating i-tRF-Gly GCC into existing prognostic models may improve their accuracy and allow for more precise risk categorization of CRC patients, allowing for tailored treatment methods and surveillance approaches.
Despite these promising findings, it is important to note several limitations of this study. First, although this study reveals the link between i-tRF-Gly GCC levels and CRC, it does not provide mechanistic insights into the functional involvement of i-tRF-Gly GCC in CRC development and progression. Understanding the underlying biological mechanisms and molecular pathways impacted by i-tRF-Gly GCC would support its usage as a biomarker. In order to determine the functional significance of i-tRF-Gly GCC in CRC, more functional investigations, such as in vivo tests, are required. Secondly, this study's findings would benefit from independent confirmation in other CRC patient groups. Cross-validation across several populations and centers is critical for determining the robustness and reproducibility of the results. Furthermore, validation in cohorts with varying stages of CRC and clinical features would give a more comprehensive assessment of the biomarker's performance. Lastly, this study focused solely on the discriminatory and prognostic potential of i-tRF-Gly GCC in CRC. While the findings are encouraging, it is crucial to note that CRC is a complicated disease with various molecular alterations. A single biomarker may not reflect the whole complexities of CRC biology, limiting its clinical utility. To improve the accuracy and reliability of CRC diagnosis and prognosis, future research should consider incorporating numerous biomarkers or a panel of biomarkers.
Overall, our findings provide novel evidence for the utility of i-tRF-Gly GCC as a discriminatory and prognostic biomarker in CRC. The downregulation of i-tRF-Gly GCC in CRC tissues, as well as its association with a poor prognosis, emphasizes its clinical importance. Further studies should focus on verifying these findings in bigger cohorts, explaining the underlying processes, and investigating the clinical value of i-tRF-Gly GCC . In this challenging disease, incorporating i-tRF-Gly GCC evaluation into normal CRC management has the potential to improve patient outcomes and better individualized treatment regimens.

Conclusions
According to our study, high i-tRF-Gly GCC expression is associated with poor survival and higher relapse rates in CRC patients. Therefore, independent of other clinicopathological variables, the expression status of this tRF may be utilized to evaluate the prognosis of CRC patients in addition to TNM staging. Informed Consent Statement: Written informed consent was obtained from all patients.

Data Availability Statement:
The data presented in this study are available on reasonable request from the corresponding authors.