Elevated expression of miR-21 has been reported in PCa tissue (Damodaran et al., 2021), plasma (Bryant et al., 2012), and urine (Samsonov et al., 2016). miR-21 is known to regulate the proliferation and apoptosis of cancer cells through the PTEN/PI3K/AKT pathway which makes it a good potential novel target for anti-cancer therapies (Liu et al., 2019). In Samsonov et al. (2016) the analysis of PCa–related miRNAs using lectin-induced agglutination method for isolation of in urinary-exosomes revealed significant upregulation of miR-21-5p in PCa patients with PSA ranging from 10 to 20 as compared to healthy control. Although the study reported a notable diagnostic significance of miR-21-5p along with miRNA-574-3p and miR-141-5p, the sensitivity of miR-21-5p (area under the curve (AUC) and receiver operating characteristic curve (ROC) 0.65; 95%, CI = 0.477–0.814, sensitivity = 0.46, p < 0.05) was the lowest among the three. Another RNAseq study performed by Koppers-Lalic et al. (2016) on PCa patients (PSA = 8.5–9.5) observed a significant upregulation of miR-21 expression (p = 0.02) along with miR-375, and miR-204 as compared to healthy controlKoppers-Lalic et al. (2016). When the expression of these miRNAs was validated by RT-qPCR, the results regarding miR-21 were not significant; however, further studies showed later that those differences were associated with the presence of different miR-21-5p isomeric forms. This study proposed that the detection of isomiRs in combination with PSA levels could significantly increase the probability of detecting PCa as compared to mature miRNA or PSA alone (Koppers-Lalic et al., 2016).

Many studies have proposed the establishment and analysis of batteries of miRNAs, instead of individual miRNA, in order to achieve higher sensitivity and specificity in the diagnosis and stratification of PCa patients (Bryzgunova et al., 2019; Fredsøe et al., 2019; Davey et al., 2020). For instance, a study conducted by Bryzgunova et al. (2019) in uEVs highlighted the relevance of miR-21-5p in the discrimination of PCa and BPH patients when they applied an algorithm-based classifier, which analyzed the expression of five different pairs of diagnostically significant miRNAs: miR-21-5p+miR-29b, miR-30a+miR-125b, miR-425+miR-331, miR-191+miR-200a and miR-331+miR-106b.

To corroborate the association of miR-21-5p with PCa, in this review we have cross-checked the group of miR-21-5p target genes with the PCa-associated genes retrieved from GeneCards, resulting in a list of common genes shown in Table 4. The top 12 miR-21-5p target genes showing a score >95 are PCa-specific genes. This result, together with the aforementioned findings, highlights the relevant role of miR-21-5p in urine-based PCa diagnosis and underlines the necessity of conducting functional studies that help us identify the efficacy and suitability of miR-21-5 and its isomiR, alone or in combination with PSA or other miRNAs, in PCa diagnosis.

miR-141 is reported to affect angiogenesis, proliferation, and metastasis in colon cancer and nasopharyngeal carcinoma (Dong et al., 2019; Tong et al., 2021). Although its role in PCa is yet to be defined, studies have reported miR-141 to be commonly present in the blood of PCa patients (Yaman Agaoglu et al., 2011). In regard to studies performed in urine samples, Xu et al. (2017) did not find any significant differences in the expression of uEV miR-141-5p between BPH and PCa but its presence in urinary cells of PCa patients was observed to be significantly higher in the disease group. In contrast, other authors like Samsonov et al. (2016) found upregulated expression of uEV miR-141-5p in PCa patients when compared with healthy donors.

When we cross-checked the miRNA-141-5p target genes with the PCa-specific genes, we observed that the top five common genes are mainly involved in cell survival which does not shed enough light on the role of this miRNA in PCa disease progression (Supplementary Data; Table 4).

To summarize, the role of miR-41-5p as a PCa-specific biomarker remains questionable due to the low statistical significance of the results obtained in the referred studies; research involving larger case-control cohorts is required to obtain concluding results.

There exist contradicting findings in regard to this miRNA. Published data shows increased miR-375 levels in the blood of PCa patients with metastatic disease (Bhagirath et al., 2020). Additionally, a Next-Generation Sequencing (NGS)-based investigation revealed that miR-375-3p is significantly downregulated in PCa patients and combination of miR-375 and miR-451a has the potential to serve as a specific and sensitive molecular marker to differentiate PCa patients from BPH patients, and that miR-375 alone can distinguish localized PCa patients from metastatic PCa (Li et al., 2021). However, there is still disagreement about the putative discriminating function of miR-375 in PCa. In a study conducted by Foj et al. (2017), it was demonstrated that both miR-375 and miR-21 are significantly upregulated in intermediate and high risk PCa patients as copare to low risk and healthy population and could discriminate between PCa and non-PCa participants that had shown high PSA levels. Although, another study observed downregulation of miR-375-5p in cancer cases with NGS data from uEVs, the mature versions of miR-375, -141, and -21 cannot effectively distinguish between PCa and healthy individuals unlike their isomeric forms (Koppers-Lalic et al., 2016). In summary, the presence of miR-375 in uEVs has notably been observed in studies, but the significance of mature miR-375 as an overexpressed biomarker is debatable.

NGS of uEVs has revealed miR-10b-5p and other miRNAs such as let-7b-5p, miR-30a-5p, miR-10a-5p, and let-7a-5p, as the most abundant in urinary exosomes (Cheng et al., 2014; Koppers-Lalic et al., 2016; Rodríguez et al., 2017). Their presence does not seem to be associated with the disease state and, although further investigation is required, it has been suggested that these miRNAs could be applied as uEVs-specific normalizers. In fact, one of the studies used an average of three miRNAs/small RNAs for normalization (miR-10b-5p, let-7b-5p, and U6 snRNA) as miR-10b-5p and let-7b-5p were expressed similarly in the control and patient groups according to the NGS study (Rodríguez et al., 2017).

miR-331-3p has been reported to affect long-term survival in hepatocellular cancer (Chang et al., 2014). It is also a mediator in the epithelial-to-mesenchymal transition in PCa (Fujii et al., 2016) which has led to its consideration as PCa prognostic marker as it is also supported by its identified target genes in Table 4.

miR-331-3p has already been included in large cohort-based case-control studies as part of a predictor panel for biochemical recurrence and aggressive PCa risk (Fredsøe et al., 2019). Also, one study involving the use of RT-qPCR to quantify miRNAs in urine supernatant observed that miR-331-3p together with miR-92a allowed the differentiation of BPH and PCa cases from healthy controls with a 52% sensitivity (p = 0.025) (Bryzgunova et al., 2019). The same study also analyzed the expression values of miR-331-3p in uEVs which were paired up with miR-425; in that case, the algorithm could differentiate PCa from healthy and BPH cases with a 40% sensitivity (p = 0.002). The combination of miR331-3p and miR-106b in urine supernatant and microvesicles was also evaluated and could differentiate PCa from healthy and BPH cases with a 30% sensitivity (p = 0.036). This study concluded that a five-miRNA pairs panel could be used to develop an expression-based algorithm that allows us to classify PCa, BPH, and healthy with 100% specificity and 97.5% accuracy (Martin et al., 2018).

The presence of miR-19b in uEVs has been proposed as an auxiliary PCa differentiating criterion with a 100% specificity and a 93% sensitivity (Bryzgunova et al., 2016). High levels of this miRNA have been found in both blood serum of PCa patients and mice prostate tumors tissue which supports a potential diagnostic and a prognostic value (Duca et al., 2021). One study conducted by Konoshenko et al. (2020) included miR-19b in a battery of twelve miRNAs that were investigated in 31 combination ratios. Eight of those miRNAs in six ratios, including miR-19b/miR-92a in uEVs could discriminate PCa, BPH, and healthy donors with 100% specificity, and 100% sensitivity. Overall, studies so far suggest a key role of miR-19b, alone or in combination with other miRNAs, in the differentiation of healthy vs. PCa as well as BPH vs. PCa. Moreover, in this review, miR-19b has shown the highest number of prostate-associated target genes with a total of 84 genes (Table 4).

miR-204 is known to be an NF-kB pathway inhibitor and therefore hinders various tumor progression-related phenomena including metastasis (Todorova et al., 2016; Wa et al., 2019). The relevance of miR-204 in PCa has been confirmed in studies using RT-qPCR (Fredsøe et al., 2019) and NGS (Koppers-Lalic et al., 2016). Fredsøe et al. (2019) performed a novel logistic regression model comprising PSA and five urine miRNAs, one of them miR-204, which could significantly predict biochemical recurrence and clinical risk stratification of PCa. But further research on this miRNA has demonstrated that not only the copy number variation of miR-204 in uEVs but also the presence of its different isoforms (isomiRs) with 3′-end modifications have the ability to discriminate PCa cases from the healthy subjects. In this sense, a study conducted by Koppers-Lalic et al. (2016) observed that standard mapping protocols involving expression levels of matured miRNAs failed to robustly discriminate disease status whereas isomiRs of miR-21, miR-204, and miR-375 could detect PCa with 72.9% sensitivity and 88% specificity, with an AUC of 0.866.

The miRNA-30 (miR-30) family is a group of tumor suppressor miRNAs consisting of six mature miRNA molecules (miR-30a, miR-30b, miR-30c-1, miR-30c-2, miR-30d, and miR-30e) (Mao et al., 2018). miR-30a has been frequently reported to be downregulated in PCa and its expression inhibits androgen-independent growth in PCa (Zhao et al., 2019). In this sense, deep sequencing analysis conducted by Rodríguez et al. (2017) described miR-30a-5p as one of the most abundant miRNAs in uEVs slightly, but not significantly, downregulated in PCa cases compared to healthy males. In the aforementioned study involving the application of a five-miRNA pair-based algorithm, miR-30a was one of the miRNAs that could discriminate between PCa and BPH patients and healthy samples with 100% specificity and 97.5% accuracy (Bryzgunova et al., 2019). Since miR-30b is also among the most stably expressed miRNAs in uEVs along with miR-200b-3p, and miR-27b-3p, it has been used as an endogenous control in RT-qPCR experiments in some studies (Fredsøe et al., 2019) which questions its validity as PCa biomarker. Nevertheless, in another study, miR-30b-3p along with miR-126-3p was observed to be overexpressed in uEVs of PCa patients as compared to subjects showing cancer-negative biopsy results. This study proposed a PSA and age-adjusted logistic regression analysis for the prediction of PCa with 46.4% sensitivity and 88.0% specificity (Matsuzaki et al., 2021). miR30b along with miR-30e also showed up as members of the most diagnostically significant miRNA pairs in an algorithm created to differentiate PCa patients from a healthy population. This study used median ∆Ct difference, and confidence interval as statistical criteria for determining reliable markers of PCa (Lekchnov et al., 2018).

Finally, there are other miRNAs that have been reported to show significantly low expression levels in PCa compared to a healthy population and these are miR-196a-5p, miR-34a-5p, miR-143-3p, miR-501-3p, and miR-92a-1-5p in uEVs (Rodríguez et al., 2017). According to one study, miR-2909 can be used to monitor PCa aggression (Wani et al., 2017). Another study observed elevated levels of miR-888 cluster in patients with high-grade prostate cancer as compared to lower grades of the disease (Hasegawa et al., 2018). Three miRNAs from uEVS - miR-26a-5p, miR-532-5p, and miR-99b-3p, are suggested to be used in the prediction of biochemical recurrence of PCa (Kim et al., 2021). Using a machine learning method, the two miRNAs hsa-miR-532-3p and hsa-miR-6749-5p could distinguish BPH patients from those with PCa with 80% specificity and 66.7% sensitivity (Holdmann et al., 2022). miR-423-5p and miR-193-3p were downregulated in uEVS of high-risk patients’ post-digital rectal examination but their diagnostic significance is not known (Ruiz-Plazas et al., 2021).

We believe that the disparity in the results could be influenced by the variations in the selected control groups; whilst ten of the studies included in this review involved healthy donors as controls (Bryant et al., 2012; Cheng et al., 2014; Bryzgunova et al., 2016; Fujii et al., 2016; Samsonov et al., 2016; Wani et al., 2017; Bryzgunova et al., 2019; Dong et al., 2019; Fredsøe et al., 2019; Liu et al., 2019), three studies compared PCa with BPH samples (Foj et al., 2017; Damodaran et al., 2021; Kim et al., 2021), two studies looked at both BPH and healthy controls (Koppers-Lalic et al., 2016; Davey et al., 2020), and one study used non-recurrent PCa as control (Hasegawa et al., 2018) (Figure 1; Table 3).

Studies involving healthy subjects as control groups were subject to diversity of the population characteristics or previous history of infections which, together, may bias the miRNA-level differences between groups. With the exception of a few studies where controls were mentioned to be age-matched (Bryant et al., 2012; Cheng et al., 2014; Bryzgunova et al., 2016; Fujii et al., 2016; Koppers-Lalic et al., 2016; Liu et al., 2019), most of the studies did not mention age or risk factors in the matched controls. Furthermore, the inclusion of BPH patients as a control population adds an extra layer of diversity to the global picture.

In regard to the case group, all studies rightly disclosed detailed information on the patient population analyzed including PSA values, Gleason scores, and clinical presentation (Table 2).

The collection of urine samples for prostate-related analysis can take place before/pre-DRE or after/post-DRE. Prostate stimulation during a DRE generates the release of cells and vesicles to the urethra, thus urine collection post-DRE is advised for the identification of PCa biomarkers (Pellegrini et al., 2017). Five publications in our review (Koppers-Lalic et al., 2016; Foj et al., 2017; Davey et al., 2020; Matsuzaki et al., 2021; Ruiz-Plazas et al., 2021) highlighted post-DRE urine sample collection whereas three have done so pre-DRE (Samsonov et al., 2016; Fredsøe et al., 2019; Li et al., 2021) and the others have not specified this criterion. Barreiro et al. (2021) have demonstrated that storage time, temperature, and format impact the quantity and quality of the uEV miRNAs yielded. Most of the studies reviewed did not detail the sample storage conditions and those that did, indicated a storage temperature of −80° (Koppers-Lalic et al., 2016; Samsonov et al., 2016; Xu et al., 2017; Bryzgunova et al., 2019; Fredsøe et al., 2019; Davey et al., 2020; Kim et al., 2021; Li et al., 2021; Matsuzaki et al., 2021); other authors preferred to use fresh samples in order to avoid miRNAs degradation or loss. We believe that these aspects should be taken into account in future experiments and their standardization would help decrease inconsistencies in the results obtained across studies.

Another limitation identified in this review is the variability in the criteria used for the initial selection of the miRNAs of study; only few studies employed initial microarrays (Bryzgunova et al., 2019; Matsuzaki et al., 2021) or deep sequencing techniques for their further validation in case-control populations (Rodríguez et al., 2017; Kim et al., 2021; Li et al., 2021; Holdmann et al., 2022) whereas the rest of the projects based their miRNA selection on previously published data reporting a correlation between their high levels of expression in serum/blood/tissue and PCa.

In regard to the endogenous controls used in the miRNAs quantification, some studies utilized spike-in miRNA levels like cel-miR-39 (Foj et al., 2017; Xu et al., 2017; Konoshenko et al., 2020; Li et al., 2021), and others used more than one endogenous miRNA or their arithmetic mean (Rodríguez et al., 2017; Fredsøe et al., 2019; Ruiz-Plazas et al., 2021) and others utilized the pair ratio method (Lekchnov et al., 2018; Bryzgunova et al., 2019). In this sense, this field lacks validated internal controls and, given that the exact mechanisms of action of many miRNAs are yet to be elucidated, we may be using controls whose function is associated with PCa disease which could provide unreliable results.

The statistical approaches used in the analysis of the expression of uEV miRNAs evaluated by RT-qPCR also differed across studies. For instance, early studies in the field used the Mann–Whitney test and T-test (Bryzgunova et al., 2016; Samsonov et al., 2016; Foj et al., 2017; Wani et al., 2017; Xu et al., 2017; Hasegawa et al., 2018; Lekchnov et al., 2018) to determine the significance of the identified miRNAs in PCa diagnosis whereas most recent studies applied uni- or multivariate regression-based methods (Fredsøe et al., 2019; Davey et al., 2020; Li et al., 2021).