Renal Cell Tumors: Uncovering the Biomarker Potential of ncRNAs

Renal cell tumors (RCT) remain as one of the most common and lethal urological tumors worldwide. Discrimination between (1) benign and malignant disease, (2) indolent and aggressive tumors, and (3) patient responsiveness to a specific therapy is of major clinical importance, allowing for a more efficient patient management. Nonetheless, currently available tools provide limited information and novel strategies are needed. Over the years, a putative role of non-coding RNAs (ncRNAs) as disease biomarkers has gained relevance and is now one of the most prolific fields in biological sciences. Herein, we extensively sought the most significant reports on ncRNAs as potential RCTs’ diagnostic, prognostic, predictive, and monitoring biomarkers. We could conclude that ncRNAs, either alone or in combination with currently used clinical and pathological parameters, might represent key elements to improve patient management, potentiating the implementation of precision medicine. Nevertheless, most ncRNA biomarkers require large-scale validation studies, prior to clinical implementation.


Renal Cell Tumors
Renal cell tumors (RCT) rank 16th among the most common neoplasms in adults, representing more than 400,000 new cases yearly (2.2% of all cancer diagnosis) in both genders, with a mortality rate of 2.4/100,000, worldwide [1]. RCTs are a heterogenous group of tumors, spanning from benign to overtly malignant behavior and being highly diverse at the molecular, genomic/epigenomic, morphological, and clinical level [2]. Benign renal tumors correspond to 10-13% of all RCT, being oncocytomas the most prevalent, whereas clear cell renal cell carcinomas (ccRCC) are the most common and one of the most aggressive malignant RCT subtypes (70-75% of all cases), followed by papillary renal cell carcinomas (pRCC, 10-15%) and chromophobe renal cell carcinomas (chRCC, 5-10%) [3]. Since these four types of RCT represent the vast majority of renal tumors, they will represent the main focus of this review. Although, in recent years, mortality rate has dropped, incidence has increased, mainly due to incidental detection. Indeed, more than 50% of RCTs are incidentally detected after nonspecific musculoskeletal or gastrointestinal complaints entailing abdominal imaging [4]. Visible  (5-mC). Histone post-transcriptional modifications constitute another type of epigenetic-based gene regulatory mechanism. These reactions (e.g., acetylation and methylation) occur in the residues of histone tails and are extremely refined, being catalyzed by highly regulated enzymes. In addition, histones might also have variants, which alter nucleosome functionality. Chromatin remodeling complexes, such as SWItch/Sucrose Non-Fermentable (SWI/SNF), Imitation SWI (ISWI), INO80, and Nucleosome Remodeling Deacetylase (NuRD), alter chromatin architecture, through direct interaction with nucleosomes. Finally, non-coding RNAs, namely small non-coding RNAs (sncRNAs) and long non-coding RNAs (lncRNAs) post-transcriptionally regulate gene expression, both in the nucleus and cytoplasm. Created by BioRender.com (https://biorender.com/) Global increase in RCT incidence over the last decades and the concerns regarding the most suitable follow-up and treatment for each patient demand reliable biomarkers amenable to clinical use. Herein, we aimed to critically review and highlight the most scientifically relevant and clinically promising studies concerning ncRNA-based biomarkers for RCT detection, prognostication, prediction of response to therapy, and patient monitoring.

Evidence Acquisition
Bibliography was selected after a PubMed search up to 19 April 2020 using the keywords: Non-coding RNA, biomarkers, and renal cell tumor, which resulted in the analysis of more than 400 manuscripts. All articles' references were also examined for potentially useful studies. Furthermore, relevant articles were selected based on the following criteria: Written in English, the main topic is non-coding RNA, biomarkers, and renal cell carcinoma. Original reports were chosen based on the detail of analysis, mechanistic support of data, novelty, and potential clinical usefulness of the findings. After thorough analysis, 143 original articles were enrolled in the final version of this review.

Small Non-Coding RNAs (sncRNAs)
The classification of sncRNAs as epigenetic mechanism of gene expression control remains controversial. Several studies have pointed that sncRNAs' mechanism of action is post-transcriptional and should not be thus classified as epigenetic regulators, whereas others have a contrasting view. Nevertheless, this subclass of ncRNAs is biologically relevant. MicroRNAs (miRNAs) are the most well studied of these small molecules [35]. This class of small ncRNAs are 18-25 nucleotides in length [24] and regulate gene expression through RNA interference (RNAi) [23]. In the human genome, miRNAs are encoded by individual genes or clusters of few to several hundred different miRNAs genes [36]. The latter are then transcribed as polycistronic transcripts, which are ultimately processed into the individual mature miRNAs. In most cases, miRNAs are encoded by introns of non-coding or coding genes, but they can also be encoded by exonic regions [37]. Following transcription by RNA polymerase II, the primary miRNA (pri-miRNA) undergoes several steps of maturation, catalyzed by type III ribonucleases (RNases). First, in the nucleus, the Drosha complex cleaves the pri-miRNA, leading to the formation of the precursor miRNA (pre-miRNA). Then, after pre-miRNA transport to the cytoplasm, Dicer complex cleaves the molecule, generating a miRNA duplex, which is loaded into the pre-miRNA-inducing silencing complex (pre-miRISC), where the stable 5 end strand-guide strand-is selected, generating the mature miRISC complex, whereas the other strand-passenger strand-is rapidly degraded [38]. Together with GW182 family of proteins, miRISC binds to mRNA targets by base complementarity and, ultimately, leads to gene silencing. Another type of sncRNAs are the P-element Induced Wimpy testis (PIWI)-interacting RNAs (piRNAs). First discovered in the beginning of the 21st century [39], these 21-35 nucleotide-long molecules are involved in viral infection response, transposable elements silencing, and regulation of gene expression by (1) leading PIWI proteins to cleave target RNA, (2) promoting heterochromatin assembly, and (3) inducing DNA methylation [40,41]. One of the main features that distinguishes miRNAs from piRNAs is that the latter have single-strain RNA precursors and its processing requires PIWI proteins of the Argonaute/PIWI family, but does not need DICER complex [42]. Small interfering RNAs (SiRNAs) are also classified as sncRNAs [31], but this review solely focused on miRNAs and piRNAs, as these are the most well studied.

MiRNAs and piRNAs in RCTs
Deregulated miRNA expression in cancer was first reported in the early 2000s by Calin and colleagues [43]. Since then, various studies have demonstrated differential miRNA expression profiles in benign and malignant neoplasms compared to healthy individuals. Dysregulation of miRNA expression occurs in various steps of tumorigenesis and in several tumor models [23], including RCT [24,44]. MiRNAs possess the ability to target several mRNAs, and one mRNA might be targeted by many miRNAs [45]. Depending on the target, miRNAs are classified as tumor suppressor miRNAs (TSmiRs) or as oncogenic miRNAs (oncomiRs). TSmiRs are usually downregulated in cancer and act through transcriptional repression of oncogenes, whereas oncomiRs are normally upregulated, and act by targeting tumor suppressor genes, leading to mRNA decay and/or degradation [46]. Nonetheless, several reports have demonstrated that, depending on the cellular context and the tumor type, the same miRNA may exhibit oncogenic or tumor suppressor activity, such as let-7g, which is downregulated in lung cancer and upregulated in colorectal cancer [47,48].
Concerning piRNAs, most are not complementary to putative target mRNAs, indicating that piRNAs may be involved in epigenetic regulation rather than post-transcriptional regulation, controlling a variety of biological processes and being also implicated in cancer development [49]. Several studies aimed to disclose their biological role in different cancer types [50], including RCC [51][52][53]. However, Cancers 2020, 12, 2214 5 of 31 the specific molecular mechanism underlying piRNAs' deregulation in carcinogenesis is still poorly understood, and further investigation is needed.
Due to their tissue and cellular-specific functions and expression, the potential use of sncRNAs as diagnostic, prognostic, predictive, and monitoring cancer biomarkers has been extensively studied in the recent years. Here, we highlighted the most promising findings in RCT, both in tissue and liquid biopsies.

SncRNAs as Diagnostic Biomarkers
Tissue-Based Samples The increasing number of asymptomatic, incidentally detected renal masses constitutes a major clinical challenge, considering the need to define the potential threat to the life of the patient. Whether a biopsy is mandatory or not remains controversial, considering that histopathological and/or cytopathological assessment may not provide a definitive diagnosis in a sizeable proportion of cases. Thus, the ability of sncRNAs to discriminate between normal and benign/malignant tissue has been investigated. Wotschofsky and co-workers [54] measured the differential expression of several miRNAs in a series of 111 ccRCC and matched normal tissue (MNT) using quantitative real-time PCR (RT-qPCR). The combination of miR-141, miR-155, and miR-184 identified malignancy with 95% sensitivity, 100% specificity, corresponding to an area under curve (AUC) of 0.990 [54]. In another study, miR-141 or miR-200b levels discriminated RCC from normal renal tissue (NRT) with 99.2% sensitivity, 100% specificity, and an AUC of 0.991. Furthermore, the same panel distinguished ccRCC, pRCC, or chRCC from benign renal tumors with 85.6% sensitivity, 100% specificity, and an AUC of 0.914 [55]. In 2015, Busch and colleagues [51] reported that piR-30924, piR-57125, and piR-38756 were differentially expressed in ccRCC compared to NRT and the combination of these piRNAs identified malignant disease with 91% sensitivity, 86% specificity, and an AUC of 0.910. Notably, the combination of the duo piR-30924 and piR-57125 distinguished metastatic-ccRCC (mccRCC) from non-metastatic ccRCC (non-mccRCC) with 73.0 sensitivity, 74.0 specificity, and an AUC of 0.760 [51]. Several other studies have been published since, and are summarized in Table 1. Because tissue biopsies are seldom performed and might not represent the entire lesion, these biomarkers might assist in the correct classification of the tumor. Moreover, this is an invasive procedure, which submits patients to stress and pain, eventually associated with increased risk of metastization, especially in ccRCC. Hence, discovery and validation of non-invasive screening/diagnosis biomarkers, capable of accurately identifying the nature of renal masses, is urgently needed.

Liquid Biopsies
Recently, detection and characterization of circulating sncRNAs might represent a promising non-invasive technique to identify RCT [76]. SncRNAs are highly stable and abundant in plasma, serum, and other body fluids, being released from damaged or apoptotic normal cells, as well as from tumor cells. Numerous reports have proposed several RCT biomarkers in liquid biopsies. Serum samples were firstly used in a study by Wulfken and colleagues [77], which demonstrated that serum miR-1233 expression levels discriminated cancer patients from asymptomatic controls (AC) with 77.4% sensitivity and 37.6% specificity. The limited performance of this marker compared to tissue-based studies might be explained by technical limitations. Since then, methodology has improved, and miR-210 expression levels were found to discriminate ccRCC and AC in serum samples with 81.0% sensitivity and 79.4% specificity [78]. Recently, miR-1233 and miR-210 levels, in serum and in exosomes, discriminated ccRCC from healthy controls with 81.0/70.0% sensitivity and 76.0/62.2% specificity, respectively, with exosome-derived samples showing a better biomarker performance [79]. Moreover, plasma samples have also been tested. Specifically, miR-21 and miR-106a isolated from plasma (30 ccRCC and 30 AC) disclosed the ability to identify renal malignancy with 77.3% sensitivity and 96.4% specificity for the former miR and 86.7% sensitivity and 70.0% specificity for the latter [80]. Subsequently, Lou and colleagues [81] showed that miR-144-3p detected RCT with 87.1% sensitivity and 83.0% specificity. Notably, miR-144-3p was also able to distinguish ccRCC from benign mesenchymal tumors (angiomyolipomas) with 75.0% sensitivity and 71.7% specificity [81]. Finally, diagnostic biomarkers have also been tested in urine samples. In 2016, Butz and colleagues [82] reported that miR-126-3p and miR-34b-5p, isolated from urine exosomes, could discriminate ccRCC from healthy controls with 77.5% sensitivity and 72.4% specificity. Remarkably, both miRs also distinguished benign lesions from normal with 75.0% sensitivity and 82.8% [82]. Additionally, urinary miR-15a expression levels, evaluated in 67 RCT patients and 15 AC, detected malignancy with 98.1% sensitivity and 100% specificity [83]. A summary of these and other studies is depicted in Table 2.

. SncRNAs as Prognostic Biomarkers
Tissue-Based Samples Several sncRNAs have also been proposed as predictors of disease progression and outcome. Currently, RCT prognosis is mainly based on clinical stage and other clinical parameters at diagnosis. Nonetheless, specific sncRNAs might complement the currently used clinicopathological parameters, to improve patient management. In 2013, Wang and colleagues [96] reported that RCC patients disclosing higher miR-100 expression levels endured significantly shorter overall survival (OS), multiplying by a factor of three the risk of death comparing to those with low expression. Likewise, increased miR-630 expression levels independently predicted shorter OS, in multivariable analysis [97]. Importantly, sncRNAs have shown promise as predictors of disease-progression. Samaan and colleagues [98] divided their 258 ccRCC patient cohort into either miR-210 positive or negative expression groups. The first group of patients displayed markedly reduced disease-free survival (DFS) (hazard ratio (HR): 1.91; 95% confidence interval (CI): 1.10-3.310) compared to the negative expression group [98]. The same trend was observed in two subsequent studies, in which higher miR-210 expression associated with worse survival [99,100], whereas in another study increased miR-210 expression levels in ccRCC tissue associated with better survival [101]. Thus, multicentric studies with larger cohorts are needed to unveil the exact prognostic value of miR-210. Furthermore, high miR-27a-3p expression levels associated with shorter progression-free survival (PFS) [102], whereas low miR-155 expression entailed 5-fold increase risk to die from the disease. Notably, both miR-27a-3p and miR-155 expression levels were independent predictors of cancer-specific survival (CSS) in advanced ccRCC (stages III and IV) [103]. Table 3 summarizes these and other relevant findings concerning the prognostic value of miRNAs in RCC.  *-univariable analysis; HR-hazard ratio; 95% CI-95% confidence interval; OS-overall survival; DFS-disease-free survival; PFS-progression-free survival; MFS-metastasis-free survival; CSS-cancer-specific survival; RFS-recurrence-free survival.

Liquid Biopsies
Studies on sncRNAs as potential biomarkers for RCC progression and/or disease outcome in liquid biopsies are rather scarce. Let-7i-5p low expression in exosomes from plasma of 65 mRCC patients associated with shorter OS [133]. Fujii and colleagues [134] showed that higher plasma-derived exosomal miR-224 expression levels negatively associated with shorter OS, CSS, and recurrence-free survival (RFS). A subsequent analysis of 67 ccRCC serum samples demonstrated that increased miR-206 and miR-122-5p expression associated with increased risk of disease progression and mortality, although, in multivariable analysis, only miR-206 retained independent value as predictor of PFS [135]. Finally, Dias and colleagues reported that higher miR-210, miR-221, and miR-1233 plasma levels associated with shorter CSS [99]. Detailed information of all relevant studies may be found in Table 4.

SncRNAs as Predictive Biomarkers of Response to Therapy
Tissue-Based Samples Uncertainties concerning efficacy and deleterious side effects of RCC therapy negatively impact patient management [136][137][138]. Ideally, each patient should be prescribed the therapy most likely to specifically target and eliminate neoplastic cells, which sets the basis for precision medicine [139]. Considering their involvement in critical metabolic pathways, it is unsurprising that sncRNAs have been implicated in cancer therapy resistance [140][141][142]. Furthermore, sncRNAs have been proposed as predictors of response to therapy in RCC. Indeed, miR-141 expression levels were shown to predict response to sunitinib, as patients with low levels disclosed a significantly worse response [143]. In a different study, lower expression levels of both miR-155 (a well-known oncomiR) and miR-484 (with biological role yet to be fully understood) associated with increased time to progression (TTP) in a series of 63 mRCC patients (44 responders and 19 non-responders) treated with sunitinib [144]. Recently, Go and colleagues demonstrated that miR-421 was highly expressed in RCC tissues from patients who did not respond to VEGFR-TKI [145]. Table 5 provides additional information on the most relevant studies concerning the predictive value of sncRNAs in RCC. HR-hazard ratio; 95% CI-95% confidence interval; OS-overall survival; DSS-disease-specific survival; PFS-progression-free survival; CSS-cancer-specific survival.

Liquid Biopsies
One of the main drawbacks of tissue-based studies is the inability to capture the dynamic nature of sncRNAs' expression along time, either during disease progression or due to therapeutic intervention. The usage of liquid biopsies might circumvent this limitation, since it allows sample collection at several time points, i.e., prior to, during, and after treatment, enabling patient monitoring. In 2012, Gámez-Pozo and colleagues analyzed the expression of 287 miRs in 38 whole blood samples from patients with advanced RCC treated with sunitinib and constructed multiple models of poor and prolonged response to this TKI. Notably, miR-410, miR-1181, and miR-424 downregulation was associated with prolonged response, whereas low miR -192, miR-193a-3p, and miR-501-3p levels associated with limited response [146]. Additionally, serum miR-605 levels in mccRCC patients treated with vorinostat and bevacizumab were exclusively reduced in the responders' group, comparing to the disease progression group [156]. Additional studies are summarized in Table 5.

In Vitro Studies
Several studies have been performed in RCC in vitro models in the pursuit of both predictive biomarkers and insights on therapy-resistance mechanisms. Although in vitro models do not fully mimic biological conditions, they allow for the discovery of potential sncRNA-based biomarkers, which may be subsequently validated in clinical samples. Gao and colleagues [149] reported that transfecting mimic miR-200c into ccRCC cell lines resistant to imatinib and sorafenib re-sensitized cells to therapy. Moreover, miR-200c was downregulated in RCC cell lines, whereas one of its targets, CYP1B1, was overexpressed. Remarkably, increased CYP1B1 levels were associated with docetaxel resistance [152]. Additionally, miR-101 downregulation in ccRCC cell lines associated with high UHRF1 (a miR-101 target) levels and, ultimately, to sunitinib resistance [163]. Detailed information on these and other studies can be found in Table 5.

LncRNAs
As previously mentioned, lncRNAs are a class of ncRNAs with more than 200 base pairs in length [33,34]. These molecules can be further categorized into different groups, depending on their genome location, sequence, morphology, structure, and functional features [164,165]. Most lncRNAs are synthesized by the same biogenesis machinery as mRNAs and endure post-transcriptional modifications, such as 5 terminal methylguanosine cap (5 cap), are often spliced in a canonical manner, and some are 3 polyadenylated [165]. LncRNAs have a fine-tuned regulation by transcription factors and typically display a tissue-specific expression profile [164]. Functionally, lncRNAs have been implicated in a multitude of biological processes, such as nuclear organization through nucleosome remodeling [166], gene-to-gene interactions [167], and as regulators of miRNA expression [168], thus prompting the hypothesis that lncRNAs' differential expression might be associated with human disease. Indeed, several pathological conditions display aberrant lncRNAs' expression profile [169], including cancer [170].

LncRNAs in RCT
LncRNAs have been the focus of several research studies aiming at the discovery of novel biomarkers and understanding the biological mechanisms through which they influence the genesis and progression of RCT [171][172][173]. Compared to their protein-coding counterparts, lncRNAs are considerably less expressed, which might constitute a major pitfall for their use in clinical practice, since robust detection is quite challenging [174]. Nonetheless, the study of these molecules should be promoted, as technological advances might overcome the present limitations. Herein, we highlighted the most relevant studies reporting lncRNAs as potential diagnostic, prognostic, predictive, and monitoring biomarkers in RCT, both in tissue and liquid biopsies.

LncRNAs as Diagnostic Biomarkers
Tissue-Based Samples Contrarily to sncRNAs, published data concerning lncRNAs as RCT diagnostic biomarkers are limited. Over two decades ago, Thrash-Bingham and colleagues [175] reported, for the first time, that lncRNA expression was dissimilar among RCC subtypes. Using semiquantitative PCR, a marked increased expression of lncRNA antisense Hypoxia Inducible Factor (aHIF) in ccRCC, comparatively to pRCC, was disclosed [175]. Technology has evolved and these results were subsequently validated in 2011, when Bertozzi and colleagues [176] detected a differential expression of lncRNA aHIF between RCC and MNT, as well as between non-pRCC and pRCC tissue samples. In another study, comprising 102 ccRCC and 50 NRT, lncRNA CYP4A22-2/3 discriminated ccRCC from NRT with an AUC of 0.790 [177]. In 2016, the expression of lncRNA UC009YBY.1 and lncRNA ENST00000514034 was assessed in a set of 70 ccRCC and 70 MNT by Ren and colleagues [178]. These authors reported that the two lncRNAs could identify RCC tissue with 54.29% sensitivity and 82.86% specificity for the former, and 60.00% sensitivity and 67.14% specificity for the latter [178]. Finally, a recent study reported that lncRNA HOX Transcript Antisense RNA (HOTAIR) might also constitute a ccRCC diagnostic biomarker, disclosing an AUC of 0.9230 [71]. Table 6 summarizes the complete information on relevant published studies reporting lncRNAs as potential RCC diagnostic biomarkers.

Liquid Biopsies
Based on our literature search, only two relevant studies evaluating the potential of lncRNAs as RCC diagnostic biomarkers in liquid biopsies were found. Wu and colleagues [183] assessed the expression of five lncRNAs (lncRNA-low expression in tumor (LET), Plasmacytoma Variant Translocation 1 (PVT1), Promoter Of CDKN1A Antisense DNA Damage Activated RNA (PANDAR), Phosphatase and Tensin Homolog Pseudogene 1 (PTENP1), and long intergenic non-protein coding RNA 963 (linc00963)) in two sets of ccRCC and AC serum samples. When combined in a panel, these biomarkers identified malignancy with 79.2% sensitivity and 88.9% specificity in the training set (24 ccRCC and 27 AC) and with 67.6% sensitivity and 91.4% specificity in the testing set (37 ccRCC and 35 AC) [183]. Subsequently, serum expression of lncRNA GIHCG was assessed in a set of 46 ccRCC and 46 AC samples. GIHCG expression discriminated ccRCC from healthy donors with 87.0% sensitivity and 84.8% specificity. Remarkably, it could also distinguish early-stage ccRCC from AC (31 stage I ccRCC vs. 46 ACs) with 80.7% sensitivity and 84.8% specificity [184]. The complete information concerning these studies is provided in Table 6, together with tissue-based studies.

LncRNAs as Prognostic Biomarkers
Tissue-Based Samples As for sncRNAs, several reports on the potential of lncRNAs as RCC prognostic biomarkers in tissue samples have been published. In a series of 102 ccRCC, Ellinger and colleagues [177] showed that patients disclosing low lncRNA Zinc-Finger protein 180-2 (ZNF180-2) expression endured significantly shorter OS. Notably, lower expression of this lncRNA also correlated with shorter CSS and PFS [177]. In another study, expression levels of lncRNA regulator of Akt Signaling Associated With HCC And RCC (lncARSR) were assessed in a set of 205 ccRCC tissues from patients subdivided into high-and low-expression groups. Patients with lncARSR-high expression displayed significantly shorter RFS, doubling the risk of recurrence comparatively to the lncARSR-low expression group [185]. Furthermore, Bao and colleagues [186] reported that lncRNA PVT1 could serve as independent prognostic biomarker in ccRCC. Indeed, patients with high lncRNA PVT1 expression depicted 1.5 and 3.5 times higher risk of death or recurrence, respectively, compared to patients with low lncRNA PVT1 expression [186]. Recently, in a series of tissues from 204 ccRCC patients, Wang and colleagues [187] found that high lncRNA EGFR-antisense RNA 1 (EGFR-AS1) expression levels increased two-fold the risk of death. Table 7 depicts the complete information on these and other relevant studies assessing the potential of lncRNAs for RCC prognostication.

Liquid Biopsies
After literature revision, only two studies reporting on the potential of lncRNAs for RCC's prognostication in liquid biopsies were found. Qu and colleagues [204] assessed the expression of lncARSR in 71 ccRCC plasma samples and verified that higher expression levels were associated with worse PFS. In the other study, serum samples from 46 ccRCC patients were analyzed and increased GIHCG expression levels associated with lower OS [184]. However, both studies lack biomarker prognostic performance analysis, preventing a more robust evaluation of the potential of lncRNAs assessed in liquid biopsies. Table 7 provides the complete information on these two studies.

LncRNAs as Predictive Biomarkers of Response to Therapy
Considering the limited number of studies assessing the ability of lncRNAs to predict response to therapy in RCC, we decided not to subdivide this section according to sample source. Qu and colleagues analyzed 71 plasma samples from ccRCC patients treated with sunitinib, in which high lncARSR expression levels were found in patients with progressive disease and poor response [204]. In another study, lncRNA Sorafenib Resistance in Renal Cell Carcinoma Associated (lncSRLR) expression was determined in tissue samples from 53 ccRCC patients treated with sorafenib. Patients with poor or no response to sorafenib therapy showed significantly higher expression levels of that lncRNA, which correlated with shorter PFS. In vitro studies revealed that lncSRLR knockdown in sorafenib-resistant RCC cell lines resulted in increased sensitivity to the treatment [205]. Furthermore, studies on RCC cell lines demonstrated a potential positive feedback loop between sunitinib and lncRNA Suppressing Androgen Receptor in Renal Cell Carcinoma (SARCC). Upon treatment with this TKI, lncRNA SARCC expression levels increased, leading to decreased resistance to sunitinib therapy [206]. Finally, a recent study reported that patients with poor response to sorafenib therapy displayed lower lncRNA Growth arrest-specific 5 (GAS5) expression levels. In vitro and in vivo experiments disclosed that overexpression of lncRNA GAS5 resulted in increased sensitivity to sorafenib [207]. Complete information on these studies is summarized in Table 8. TargetedTherapy ↓ GAS5 in sorafenib non-responders; in vitro/ in vivo re-introduction ⇒ ↑ sensitivity [207] RCC-renal cell carcinoma; ccRCC-clear cell renal cell carcinoma; ↑-upregulated; ↓-downregulation.

Conclusions
In summary, the data presented herein clearly support the feasibility of using ncRNAs as RCT biomarkers, both in tissue and liquid biopsies. Importantly, liquid biopsy-based samples are easily collected, minimally invasive, and may help to overcome the limitations of tissue biopsies. Further studies on ncRNAs should focus on technical and clinical validation, preferably using large-scale multicenter cohorts, allowing to determine whether these novel biomarkers may improve the clinical management of RCT patients. Finally, Figure 2 summarizes and illustrates the most promising ncRNAs for RCTs' detection, diagnosis, prognostication, and prediction to therapy response included in this review.
In summary, the data presented herein clearly support the feasibility of using ncRNAs as RCT biomarkers, both in tissue and liquid biopsies. Importantly, liquid biopsy-based samples are easily collected, minimally invasive, and may help to overcome the limitations of tissue biopsies. Further studies on ncRNAs should focus on technical and clinical validation, preferably using large-scale multicenter cohorts, allowing to determine whether these novel biomarkers may improve the clinical management of RCT patients. Finally, Figure 2 summarizes and illustrates the most promising ncRNAs for RCTs' detection, diagnosis, prognostication, and prediction to therapy response included in this review. Figure 2. Summary of the most promising biomarker candidates for RCT. (A) Several ncRNAs have been proposed as ancillary tools for distinguishing tumor growth from normal tissue. Arrows represent each ncRNA expression level in RCT, comparing to normal samples ("red"-upregulated; "green"-downregulated). (B) Besides the potential use of ncRNAs to detect tumorigenesis, some were reported as capable to differentiate both malignant from benign lesions, and benign from Figure 2. Summary of the most promising biomarker candidates for RCT. (A) Several ncRNAs have been proposed as ancillary tools for distinguishing tumor growth from normal tissue. Arrows represent each ncRNA expression level in RCT, comparing to normal samples ("red"-upregulated; "green"-downregulated). (B) Besides the potential use of ncRNAs to detect tumorigenesis, some were reported as capable to differentiate both malignant from benign lesions, and benign from asymptomatic conditions. The arrows represent ncRNA expression level (upper) in RCC, comparing to RCT and (lower) renal benign tumors, comparing to normal samples ("red"-upregulated; "green"-downregulated). (C) The potential of ncRNAs to stratify high-risk patients has been highly studied in recent years. The arrows represent each ncRNA expression level in patients displaying worse survival, comparing to those with better survival ("red"-upregulated; "green"-downregulated). (D) Lastly, ncRNA expression level has been proposed as potential predictor of therapy response. The arrows show the biological status of each ncRNA for a better response to therapy ("red'-upregulated; "green"-downregulated).