The research team, led by Hasan Riza Aydin and Serap Ozer Yaman, stratified patients into low-to-intermediate-risk and high-risk groups according to clinical stage, Gleason score, and prostate-specific antigen levels. Using transrectal ultrasound-guided biopsy tissue, they quantified expression of miR-188-5p, GRP78, and CHOP by quantitative real-time PCR, with U6 small nuclear RNA serving as the reference for the microRNA and GAPDH for the two stress genes. In parallel, enzyme-linked immunosorbent assays measured circulating GRP78 and CHOP in serum, offering a potential window on the same biology without a needle in the prostate.

The tissue results were unambiguous. Mean miR-188-5p expression climbed from 2.19-fold in controls to 3.74-fold in low-to-intermediate-risk tumors and 7.11-fold in high-risk disease, while GRP78 followed the same trajectory from 2.30 to 4.20-fold. CHOP moved in the opposite direction, dropping from 0.83 in benign tissue to 0.45 in high-risk specimens. Serum measurements mirrored the tissue pattern: circulating GRP78 rose from roughly 33 to 78 nanograms per milliliter across the risk spectrum, whereas serum CHOP fell from about 67 to 21 nanograms per milliliter. All group differences were statistically significant at p<0.001.

Correlation analysis reinforced the picture of coordinated change. miR-188-5p was positively associated with GRP78 and inversely associated with CHOP, and GRP78 itself was negatively correlated with CHOP. The authors interpret these relationships cautiously. Because microRNAs typically repress their direct targets, the concordant rise of miR-188-5p and GRP78 does not fit a simple direct-repression model. Instead, the pattern may reflect shared dependence on tumor grade, common upstream stress signaling, compensatory feedback, or the mixed cellular composition of bulk biopsy tissue.

To probe possible molecular connections, the team turned to computational prediction. The miRWalk database, which integrates the machine-learning TarPmiR algorithm, identified predicted miR-188-5p binding sites in the transcripts of HSPA5, the gene encoding GRP78, and DDIT3, which encodes CHOP. Binding probability scores reached 0.85 for the 3-prime untranslated regions of both genes, and the predicted interaction probability for the HSPA5 5-prime untranslated region was calculated as 1. Folding analyses using the Sfold and STarMir tools revealed distinct thermodynamic profiles: CHOP interactions appeared predominantly in non-seed regions with favorable binding energies, while GRP78 showed both canonical seed-mediated and seedless interactions.

Independent support came from large public datasets. Using the GEPIA3 platform, which draws on The Cancer Genome Atlas and the Genotype-Tissue Expression project, the researchers found that HSPA5 expression was significantly elevated and DDIT3 significantly reduced in prostate adenocarcinoma compared with normal tissue. This population-scale expression pattern matched the direction of change observed in the Turkish biopsy cohort, lending external consistency to the clinical findings. The authors emphasize, however, that these computational results indicate predicted interactions only and do not establish direct binding or regulation.

Diagnostic performance was assessed with receiver operating characteristic analysis. Among the individual markers, miR-188-5p showed the strongest discrimination between cancer and benign tissue, followed by GRP78 and CHOP. When the three markers were combined into a composite model, the area under the curve reached 0.966, a figure the authors describe as excellent discriminatory power, although they stress that the model was developed and evaluated in the same small cohort without internal resampling or external validation, so the estimate may be optimistic. Serum biomarkers performed somewhat less well than tissue markers but still achieved clinically meaningful values, suggesting potential as non-invasive adjuncts.

The molecular signals also tracked with adverse pathology. Elevated miR-188-5p and serum GRP78 were significant positive predictors of cribriform morphology, an architectural pattern widely recognized as a high-risk feature linked to recurrence and metastasis, while higher tissue and serum CHOP levels were associated with reduced likelihood of cribriform growth. Exploratory univariate logistic regression confirmed associations between each marker and cancer presence, with miR-188-5p showing the largest expression change between cancer and control tissue at a log2 fold change of 1.28. The authors removed an initially fitted multivariable model because the microRNA estimate was wildly unstable, an odds ratio of 403 with a confidence interval spanning four orders of magnitude, clear evidence of overfitting in a cohort of this size.

The apparent contradiction with earlier literature deserves attention. Previous studies have described miR-188-5p as a tumor-suppressive microRNA in prostate and other cancers, where it inhibits proliferation and metastasis by repressing targets such as LAPTM4B, FGF5, and SIX1. The authors argue that the increase observed in high-grade tumors need not imply a direct oncogenic function. MicroRNA behavior is highly context-dependent, varying with disease stage, cellular composition, tumor microenvironment, and target availability, so the rise may represent a stage-associated alteration or a compensatory response to escalating stress rather than a causal driver of aggressiveness.

The study has clear limitations that the authors confront directly. Direct binding of miR-188-5p to HSPA5 or DDIT3 was not tested with dual-luciferase assays, and no gain- or loss-of-function experiments were performed, so causality remains unproven. GRP78 and CHOP proteins were measured in serum but not localized in prostate tissue by immunohistochemistry, leaving questions about intracellular accumulation and cell-surface redistribution open. The cohort was small and single-center, age differed between controls and high-risk cases, and residual confounding cannot be excluded. Nevertheless, the paired tissue-transcript and serum-protein data provide a clinically grounded foundation for future work. Larger multicenter studies with age-balanced cohorts, tissue-level protein validation, luciferase assays, and penalized regression with independent validation will be needed to determine whether this microRNA-stress signature can move from hypothesis to clinical tool.

Subject of Research: Stage-specific miR-188-5p and ER stress markers reflect tumor aggressiveness in prostate cancer

Article Title: Stage-specific miR-188-5p and ER stress markers reflect tumor aggressiveness in prostate cancer

Article References: Aydin, H. R., Sibal, I., Semerci, O., Aksoy, H. Z., Bedir, F., Çelik, V. E., Misir, S., & Ozer Yaman, S. (2026). Stage-specific miR-188-5p and ER stress markers reflect tumor aggressiveness in prostate cancer. Frontiers in Oncology, 16, Article 1886215. https://doi.org/10.3389/fonc.2026.1886215

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1886215

Keywords: Stage-specific, miR-188-5p, stress, markers, reflect, tumor, aggressiveness, prostate, cancer, scientific research