Abstract
Background/Aim: The incidence of cancer continues to rise, highlighting the urgent need for reliable, non-invasive biomarkers to support early diagnosis and improve treatment outcomes. In addition to circulating microRNAs, circulating PIWI-interacting RNAs (cir-piRNAs) have emerged as promising candidates. Although the biological functions of piRNAs are not yet fully understood, they are known to suppress transposable elements and may regulate structural genes involved in tumorigenesis. In this two-phase study, we aimed to identify serum piRNAs with diagnostic potential for clear cell renal cell carcinoma (ccRCC).
Materials and Methods: A total of 238 serum samples from ccRCC patients and 208 healthy controls were analyzed. In the exploratory phase, next-generation sequencing (NGS) was performed on pooled samples representing different clinical stages of ccRCC and healthy controls.
Results: We identified 35 piRNAs with significantly different expression (p<0.01) between groups. Based on statistical significance, read abundance, and fold change, six piRNAs were selected for the training phase and subsequently, three piRNAs, piR-24672, piR-27140, and piR-28876, were selected for validation. Validation by RT-qPCR in an independent cohort confirmed significantly reduced levels of all three piRNAs in ccRCC patients. ROC analysis demonstrated that piR-28876 is a superior diagnostic biomarker, reaching AUC=0.787, with 85.0% specificity and 66.3% sensitivity in distinguishing ccRCC patients from healthy controls.
Conclusion: Circulating piRNAs, particularly piR-24672, piR-27140, and piR-28876, may serve as potential non-invasive biomarkers for the early detection of ccRCC.
Introduction
Cancer remains the second most common cause of death worldwide, particularly in economically developed countries (1). In the Czech Republic, it is the second most common cause of mortality after cardiovascular diseases (2), with renal cell carcinoma (RCC) among the most frequently diagnosed and deadliest malignancies (3). The Czech, along with the Baltic states report one of the highest incidences of RCC globally (4-6) with growing tendencies, while mortality rates have levelled off or are slightly decreasing since early 2000s (5-7). RCC comprises several histological subtypes, of which clear cell RCC is the most prevalent (70-75%), followed by papillary (15%), chromophobe RCC (5%), and multiple rare subtypes (6, 8).
RCC is often detected incidentally and at a late stage, because renal masses remain asymptomatic until late disease stages, and the classical diagnostic triad of hematuria, abdominal pain, and a palpable mass is rare (6). Many patients are diagnosed in stage III or IV, where five-year overall survival (OS) rates drop to 78% and 27%, respectively (9, 10). In contrast, early-stage diagnosis (stage I or II) is associated with a significantly improved prognosis, with five-year OS rates of 90,4% and 83,4% (10, 11). These statistics underscore the critical need for reliable, preferably non-invasive biomarkers for early RCC detection.
The field of non-coding RNAs (ncRNAs) has experienced tremendous growth over the past two decades – not only in the volume of publications but also in the diversity of identified molecules and their biological relevance (12). MicroRNAs (miRNAs) remain the most extensively studied class of regulatory ncRNAs, established as useful small non-coding RNA biomarkers across diverse cancer types, including the molecular discrimination of brain metastases origins (13). Nevertheless, other classes such as circular RNAs, long non-coding RNAs (lncRNAs), and PIWI-interacting RNAs (piRNAs) have emerged as functionally important. However, these later-discovered classes still lag behind microRNAs in terms of research coverage (14).
piRNAs are small ncRNAs (24-31 nucleotides) originally identified for their role in silencing transposable elements in the germline via interaction with PIWI proteins – members of the Argonaute family, which are also essential in the mechanism of miRNA-induced silencing. However, unlike miRNAs, piRNAs are processed independently of Dicer and rely on PIWI-mediated biogenesis (15). In addition to their canonical role in transposon repression through transcript cleavage, methyltransferase regulation, and chromatin remodeling (16), piRNAs have also been implicated in the regulation of structural genes not only in germinal but also in somatic animal cells (17), although the full scope of their regulatory potential in somatic tissues remains to be elucidated.
Emerging evidence points to the relevance of piRNAs in cancer. Aberrant expression of tissue piRNAs has been reported in multiple myeloma (18), breast (19, 20), gastric (21-23), liver (24), lung (25), head and neck (21), bladder (26), and renal cancers (21, 27). In RCC specifically, altered levels of several piRNAs, such as piR-32051, piR-38894, and piR-43607, have been associated with metastasis, tumor stage, and patient survival (28), while reduced expression of piR-38756, piR-57125, and piR-30924 correlated with unfavorable prognosis (29). The downregulation of piR-823 in tumor tissue was linked to prolonged survival, yet paradoxically, elevated piR-823 levels were observed in patient serum without any clear association with clinicopathological parameters (27).
Despite these intriguing findings, studies focused specifically on piRNAs in RCC remain scarce (19, 27, 30-32). To date, the only study investigating circulating piRNAs in RCC patients examined piR-823 across several specimen types, including blood serum and urine (27). In contrast, the utility of circulating piRNAs in other cancers, particularly colorectal cancer, is better documented and shows considerable promise for their use as minimally invasive biomarkers (33-37). The most recent study in colorectal cancer demonstrated that specific circulating piRNAs can carry prognostic information (34), reinforcing the potential of piRNAs in liquid biopsy approaches (38).
Given the increasing evidence of piRNA dysregulation in cancer and the urgent need for non-invasive biomarkers in RCC, we undertook a comprehensive analysis of piRNA expression profiles in the plasma of RCC patients. Our study aims to expand current knowledge on circulating piRNAs and evaluate their potential as biomarkers for RCC diagnosis and prognosis.
Materials and Methods
Patient cohorts and study design. The study was conducted in three phases as illustrated in the Figure 1. In the first (discovery) phase, we created four groups of twelve patients each, representing different clinical stages of clear cell renal cell carcinoma (ccRCC). These groups included 32 men and 16 women, with a mean age of 64 years and a median age of 65 years (range=47-79) at the time of diagnosis. From each patient, 200 μl of serum were available for analysis. The control group for the discovery phase consisted of 48 healthy individuals who had undergone routine preventive examinations at the Masaryk Memorial Cancer Institute and the University Hospital Brno. In the training and validation phases, we analyzed samples from 190 ccRCC patients (mean age 64 years, median 65; range=26-86) and 160 age- and sex-matched control subjects. Cancer patients were divided into two independent cohorts, training (N=40) and validation (N=150). All patients were treated at the Masaryk Memorial Cancer Institute in Brno. Informed consent was obtained from all participants, and the study was approved by the local ethics committee of the Masaryk Memorial Cancer Institute. Clinical and pathological data for the cancer patients are summarized in Table I.
A study design flowchart. Created in BioRender. Slaby, O. (2025). https://BioRender.com/lp6614h.
Characteristics of the study population.
Sample handling and RNA isolation. Serum samples were collected and processed according to standard operating procedures at the time of the study. Aliquots were stored at −80°C within 2 h of collection and thawed only once before RNA extraction. Total RNA was isolated from blood serum using the miRNeasy Serum/Plasma Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol with following modifications: a total of 220 μl of serum per sample was centrifuged at 4°C to remove cellular debris; for lysis, 1 ml of QIAzol Lysis Reagent was added to each sample, 1.25 μl of MS2 RNA carrier (0.8 μg/μl; Roche, Basel, Switzerland) was included to improve RNA yield; and lastly, RNA was eluted in 20 μl of preheated DEPC-treated water.
RNA concentration and quality were assessed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).
Due to the low RNA concentration in individual samples, pooling was necessary. For each clinical stage of ccRCC (Stages I-IV), 12 patient samples were selected (total n=48), along with 48 sex- and age-matched non-tumor controls. Four pooled RNA samples were created from the RCC cohort, with each pool representing one clinical stage (12 samples per pool). Likewise, four pooled RNA samples were generated from the corresponding control individuals (12 samples per pool). Total RNA extracted from 12 patients in each group was pooled in equimolar amounts before sequencing.
Small RNA library sequencing. Small RNA libraries were prepared using the TruSeq Small RNA Sample Preparation Kit (Illumina, San Diego, CA, USA). For each sample, 1 ng of total RNA was used, and the protocol was followed according to the manufacturer’s instructions.
Adapters were ligated to the 3′ and 5′ ends of the RNA, followed by reverse transcription and amplification through 12 cycles of PCR using indexed primers specific to each sample. The resulting amplicons were separated by 6% polyacrylamide gel electrophoresis, and only fragments corresponding to 145-160 base pairs were selected for sequencing.
The selected gel bands were excised, eluted, and purified by ethanol precipitation. The resulting RNA pellet was dissolved in 8 μl of nuclease-free water and quantified using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).
An equimolar amount of each of the four libraries was pooled for sequencing. The pooled libraries were loaded onto a flow cell and sequenced using the MiSeq platform (Illumina, San Diego, CA, USA).
RT-qPCR validation phase. Selected piRNAs were validated using reverse transcription quantitative PCR (RT-qPCR). Complementary DNA (cDNA) was synthesized from 7.5 μL of total RNA using the TaqMan MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA).
Primers and probes were designed based on piRNA sequences retrieved from the piRBase database. For reverse transcription, 2.5 μl of total RNA at a concentration of 2 ng/μl was used per reaction.
qPCR reactions were performed in a final volume of 10 μl, containing 9.3 μl of TaqMan (No UNG) Universal PCR Master Mix and 0.665 μl of sample. Amplification was carried out in 384-well plates using the QuantStudio™ 12K Flex Real-Time PCR System (Thermo Fisher Scientific Inc.). The thermal cycling conditions were as follows: 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min.
Quantification was performed using a calibration curve generated from synthetic oligoribonucleotides (IDT, Coralville, IA, USA).
Statistical analysis. All sequencing reads were mapped against the piRBase database using standard alignment protocols optimized for small RNA analysis. Further analyses were performed using R and Bioconductor packages. Read counts were pre-normalized using scaling factors with the edgeR package, followed by between-sample normalization with the LIMMA package. Differential expression analysis was performed using linear model fitting and empirical Bayes moderation. Resulting p-values were adjusted for multiple testing using the Benjamini-Hochberg method.
For RT-qPCR data, a manual threshold of 0.2 was set for Ct value calculation. All PCR reactions were performed in duplicate. Statistical significance of piRNA expression differences was assessed using GraphPad Prism version 5 (GraphPad Software, La Jolla, CA, USA). A two-tailed non-parametric Mann–Whitney test was used to compare two groups. A Kruskal–Wallis test was used for comparisons across more than two groups. Receiver operating characteristic (ROC) curve analysis was used to evaluate the ability of piRNAs to distinguish RCC patients from healthy controls. Kaplan–Meier survival analysis was performed to assess the association between piRNA expression levels and patient survival. Ct values in validation phase were normalized using absolute quantification with standard curves. In the training phase, the expressions were not normalized due to high Ct value variability of selected endogenous controls. Ct values were transformed using 40-Ct method.
Results
In the first phase of the study, we analyzed 48 serum samples from patients with clear cell renal cell carcinoma (ccRCC) and 48 samples from healthy individuals. From these, a total of eight small RNA libraries were prepared for sequencing on the MiSeq platform (Illumina) – four from ccRCC patients and four from healthy controls.
Following bioinformatic analysis, 327 unique piRNAs were detected. Among these, 35 piRNAs showed statistically significant differential expression between ccRCC patients and healthy controls (p<0.01) (Figure 2). Based on statistical significance, read counts, and fold changes, six piRNAs were selected for training phase (Table II), and after initial analysis of diagnostic and prognostic potential, we selected three piRNAs for further validation: piR-24672 (DQ594453), piR-27140 (DQ597403), and piR-28876 (DQ598676).
A heatmap showing statistically significant (p<0.01) differences in expression of 35 piRNAs between the control group (yellow) and the group of patients with RCC (blue).
Relative expression of piRNAs in serum samples.
As endogenous controls, we initially selected piRNAs that showed no significant difference between patients and controls in the discovery phase – piR-12790, piR-28374, and piR-32161. However, in the independent validation cohort, these piRNAs showed greater variability. Therefore, we proceeded with absolute quantification using standard curves.
Both in the training and the validation phase, all three piRNAs were significantly downregulated in ccRCC patients compared to healthy individuals (Table II, Figure 3).
Circulating levels of three piRNAs in extended validation cohort of RCC patients and healthy controls. Levels of all three piRNAs were decreased in serum of RCC patients. RCC: Renal cell carcinoma.
Receiver operating characteristic (ROC) curve analysis showed (Figure 4) piR-24672: AUC=0.752, with 82.5% specificity and 62.2% sensitivity, piR-27140: AUC=0.744, with 79.8% specificity and 65.8% sensitivity, and piR-28876: AUC=0.787, with 85.0% specificity and 66.3% sensitivity. When all three piRNAs were combined in a ROC analysis, the model distinguished ccRCC patients from healthy controls with 91.8% specificity and 54.8% sensitivity, but the AUC of 0.748 didn’t exceed the AUC of piR-28876 alone. Combinations of two piRNAs did not bring any significant improvement in diagnostic value (data not shown).
ROC curves of individual piRNAs and their diagnostic potential and their combination in a diagnostic model. PiR-28876 alone achieved the best performance as a diagnostic biomarker. ROC: Receiver operating characteristic.
Besides diagnostic potential, we assessed the relationship between piRNA expression levels and clinical–pathological characteristics of the patients. Unfortunately, no significant associations were found between individual piRNA levels and tumor stage, grade, or overall survival (Table II). We observed statistically significant association with grade and overall survival of patients, and a trend toward significant difference (piR-28131) in training phase, but the effect diminished in the validation cohort. No correlation was observed with progression-free survival (PFS).
Discussion
In recent years, accumulating evidence has highlighted the involvement of PIWI-interacting RNAs (piRNAs) in cancer development and progression. piRNAs have been implicated in multiple aspects of carcinogenesis, including regulation of gene expression, suppression of transposable elements, and modulation of cellular stress responses. They have also shown potential as prognostic and diagnostic biomarkers across several tumor types.
In this study, we investigated the expression of circulating piRNAs in the serum of patients with clear cell renal cell carcinoma (ccRCC). We identified three piRNAs – piR-24672, piR-27140, and piR-28876 – that were significantly downregulated in ccRCC patients compared to healthy controls. These piRNAs showed moderate diagnostic potential in ROC analysis, with the highest specificity observed for piR-28876.
piR-24672 is likely derived from a Glu tRNA, as its sequence overlaps with the 5′ end of this molecule (39). Fragments of tRNAs, known as tRNA-derived fragments (tRFs) or tiRNAs, are typically produced in response to cellular stress and are conserved across eukaryotic species (40). The stress-responsive ribonuclease angiogenin is responsible for cleaving tRNAs, generating 30-40 nucleotide-long tiRNAs (41). These 5′ tRNA fragments can inhibit protein translation directly and promote the formation of stress granules, which contribute to translational repression during stress (42).
In addition to its tRNA origin, piR-24672 shares substantial complementarity with TFCP2L1 (CP2-like 1 mRNA transcription factor). This factor plays a key role in maintaining the pluripotency of embryonic stem cells and is part of a core transcriptional network alongside OCT4, SOX2, and NANOG (43). Reduced expression of piR-24672 may enable the tumor to both relieve translational inhibition and support stem cell–like properties by enhancing TFCP2L1 activity. Another predicted target of piR-24672 is GRIP2, a gene with limited functional characterization in cancer, although its involvement in colorectal cancer and potential impact on prognosis has been reported (44).
piR-27140 shows partial complementarity to REC8, a meiotic recombination protein, and PHRF1, both of which are considered tumor suppressors in gastric and non-small cell lung cancers, respectively (45, 46). It is plausible that, similar to the dual roles observed with TGF-β signaling, these genes may have context-dependent functions in renal cell carcinoma (47). Notably, we observed a trend suggesting that higher expression of piR-27140 may be associated with shorter overall survival, although this was not statistically significant.
piR-28876 was predicted to partially target Enoyl-CoA delta isomerase 2 (ECI2), an enzyme involved in lipid metabolism. ECI2 has been shown to promote cancer cell survival in prostate tumors by supporting metabolic adaptation (48). Its downregulation leads to reduced glucose utilization, fatty acid accumulation, and suppression of cell cycle–related genes. Thus, the observed decrease in circulating piR-28876 in ccRCC patients may reflect alterations in tumor metabolic regulation.
To clarify the roles of piR-24672, piR-27140, and piR-28876 in RCC pathogenesis and biomarker utility, further studies involving larger, independent patient cohorts and functional assays in vitro are warranted. In the present study, we employed an RNA pooling strategy, which is a cost-effective approach for exploratory biomarker discovery. Pooling is particularly useful for low-input specimens such as body fluids, as it enables retrieval of meaningful information from otherwise challenging material while averaging out inter-individual variability. However, this design also poses important limitations: individual variation cannot be assessed, patient-specific differences may be masked, and outliers could disproportionately influence the pooled signal. Thus, while our results should be interpreted as exploratory rather than definitive at the patient level, they provide a valuable first step in identifying candidate piRNAs for subsequent validation in independent cohorts. Nonetheless, our findings support the hypothesis that circulating piRNAs, similar to microRNAs, hold promise as minimally invasive biomarkers in oncology.
Although our data focus on circulating piRNA profiles in RCC, this emerging field of non-coding RNAs as biomarkers is being reinforced by complementary studies: for instance, miRNA signatures have been used to classify tumor origin in brain metastases (13), and epigenetic regulators like ZNF844 have been identified as independent prognostic markers in ccRCC (49). Together, these findings underscore the broader promise of both small RNA molecules and transcriptional networks in the biomarker-driven characterization of RCC.
Conclusion
This study identified three circulating piRNAs, piR-24672, piR-27140, and piR-28876, as significantly decreased in the serum of patients with clear cell renal cell carcinoma. Their consistent dysregulation and diagnostic performance in both phases of independent validation support their potential as non-invasive biomarkers for RCC detection. While further studies are necessary to explore their functional roles and further investigate their prognostic value which was just hinted in our training cohort, our findings contribute to the growing body of evidence that circulating piRNAs could serve as tools in liquid biopsy-based cancer diagnostics.
Acknowledgements
Core Facility Bioinformatics of CEITEC Masaryk University is gratefully acknowledged for the obtaining of the scientific data presented in this paper.
Footnotes
Conflicts of Interest
The Authors have no conflicts of interest to declare.
Authors’ Contributions
Formal analysis – L.R. and P.V. Visualization – R.I. and S.R. Investigation – R.I., S.R., P.V and M.K. Methodology and Conceptualization – L.R., O.S., M.F., M.S., and J.D. Writing – Original draft – R.I. and P.V. Writing – Review and editing – J.B. and M.K. Funding acquisition – O.S., M.F., J.D., and M.S.
Funding
This work was created with support of projects EXCELES – NÚVR (LX22NPO5102) and CZECHRIN (LM2015090) from the financial resources of the Ministry of Education, Youth and Sports of the Czech Republic.
Artificial Intelligence (AI) Disclosure
During the preparation of this manuscript, a large language model ChatGPT ver. 4.5 (OpenAI) was used solely for language editing and stylistic improvements in select paragraphs. No sections involving the generation, analysis, or interpretation of research data were produced by generative AI. All scientific content was created and verified by the authors. Furthermore, no figures or visual data were generated or modified using generative AI or machine learning–based image enhancement tools.
- Received June 27, 2025.
- Revision received August 26, 2025.
- Accepted September 10, 2025.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.