This study integrates public datasets, paired-tissue validation, cell migration assays, target network prediction, and immune signature analysis to evaluate miR-192-5p in colorectal cancer. The findings highlight discordance between TCGA/GDC expression data and local qRT-PCR results and caution against overinterpretation of its diagnostic, prognostic, functional, and immune-related significance.
Research Article
Clinical Significance, Immune Infiltration Landscape, and Functional Network Analysis of miR-192-5p in Colorectal Cancer
* These authors contributed equally
In This Article
Summary
Abstract
Colorectal cancer (CRC) is a heterogeneous malignancy in which microRNAs may contribute to tumor progression and biomarker discovery. This study evaluated the clinical significance, immune associations, and functional network of miR-192-5p in CRC by integrating TCGA/GDC and GEO bioinformatic analyses, paired tissue qRT-PCR validation, target prediction, immune signature analysis, and HT29 wound-healing experiments. Reanalysis of TCGA/GDC miRNA-seq data included 616 tumor and 11 normal COADREAD samples and showed higher database-level hsa-mir-192 signal in tumor samples than in normal samples (Mann-Whitney p = 9.96 x 10-8). The same direction was observed in 11 paired TCGA cases (median tumor-normal difference = 3.13 log2[RPM + 1], Wilcoxon p = 0.0029), whereas local qRT-PCR in paired clinical tissues supported lower mature miR-192-5p expression in CRC tissues. TCGA-based ROC analysis showed strong tumor-normal separation, but balanced resampling and leave-one-normal-out analyses confirmed that this result should be interpreted cautiously due to the limited number of normal samples. Clinicopathological analysis suggested associations with N stage and age in the TCGA cohort, whereas T stage and M stage were not significant in the summarized table, and overall survival analysis did not show a significant prognostic association. Candidate hub-gene analyses using TCGA RNA-seq provided exploratory expression, correlation, and survival evidence but did not establish direct miR-192-5p targeting. Immune signature analysis showed heterogeneous, mostly weak correlations. Functionally, miR-192-5p mimic transfection reduced HT29 wound closure. These findings indicate that miR-192-5p is biologically relevant to CRC, but its expression direction and clinical interpretation are platform-dependent and require further experimental validation.
Introduction
Colorectal cancer (CRC) is one of the most common malignancies worldwide and remains a leading cause of cancer-related mortality1,2. Although colonoscopy, fecal immunochemical testing, SEPT9 methylation testing, circulating tumor markers, and molecular profiling have improved CRC detection and clinical management3,4, substantial heterogeneity remains in tumor biology, treatment response, and patient outcome5,6,7. Biomarkers that help characterize CRC biology may therefore contribute to risk stratification and to the development of more individualized therapeutic strategies8.
MicroRNAs (miRNAs) are small non-coding RNAs, approximately 19–24 nucleotides in length, that regulate gene expression mainly through post-transcriptional mechanisms9. Dysregulated miRNA expression can influence cancer cell proliferation, migration, invasion, apoptosis, immune interactions, and treatment response10. miR-192-5p has been reported to act in a context-dependent manner across different diseases and cancers, with either oncogenic or tumor-suppressive associations depending on tissue type, molecular background, and experimental system11,12,13,14,15,16. In CRC, previous studies have reported altered miRNA profiles in clinical specimens and functional regulation of malignant phenotypes in colon cancer cells12,14,15. However, public database analyses may not always show the same expression direction, and the clinical significance, target network, and immune associations of miR-192-5p in CRC require cautious evaluation.
The present study evaluated miR-192-5p in CRC by integrating TCGA/GDC and GEO bioinformatic analyses, qRT-PCR validation in paired clinical tissues, in vitro migration assays, candidate target-gene prediction, pathway enrichment, and immune signature correlation analysis. The study was based on the hypothesis that mature miR-192-5p may be downregulated in CRC tissues and may exert a tumor-suppressive effect. Because the reanalyzed TCGA/GDC miRNA-seq signal showed the opposite expression direction, the results are interpreted as exploratory and platform-dependent rather than as a single definitive expression pattern.
Protocol
This study was performed in accordance with the Declaration of Helsinki. Approval was granted by the Ethics Committee of Inner Mongolia Medical University on October 21, 2021. Written informed consent was obtained from all participants included in the study.
Clinical specimen collection
Paired CRC and adjacent noncancerous tissues were collected from 45 patients who underwent radical resection between December 2017 and February 2020. None of the patients received preoperative radiotherapy or chemotherapy, and patients with other malignancies were excluded. Adjacent noncancerous tissue was collected at least 5 cm from the tumor margin. Samples were frozen promptly after surgical removal and stored in liquid nitrogen until RNA extraction. The cohort included 27 male and 18 female patients; 13 patients were younger than 60 years, and 32 were 60 years or older. Clinicopathological variables included tumor diameter, TNM stage, differentiation, vascular tumor thrombus, lymph node metastasis, and preoperative CEA level.
Public data acquisition and preprocessing
Open-access miRNA-seq, RNA-seq, and clinical data for TCGA-COAD and TCGA-READ were obtained from the Genomic Data Commons portal. Primary tumors and normal solid tissue samples were identified according to GDC sample type. Metastatic and recurrent tumor samples were excluded from tumor-normal differential expression analysis. For TCGA miRNA-seq files, the database-level entry detected for this analysis was hsa-mir-192; expression values were analyzed as log2(RPM + 1), where RPM indicates reads per million mapped miRNA reads. Both unpaired and paired tumor-normal analyses were performed when matched samples were available. GEO datasets GSE89076 and GSE156355 were used for candidate target-gene screening. Differentially expressed genes from GEO were screened using adjusted p < 0.05 and absolute log2 fold-change ≥ 2. Dataset names were checked for consistency throughout the manuscript.
Diagnostic, clinicopathological, and survival analyses
Receiver operating characteristic (ROC) analysis was used to assess exploratory tumor-normal separation in TCGA-COAD, TCGA-READ, and the combined COADREAD cohort. Because the normal-sample number was small, additional sensitivity analyses were performed, including 5,000-iteration balanced tumor-normal resampling and leave-one-normal-out analysis. For clinicopathological association analyses, primary tumor samples were divided into high- and low-expression groups by the median TCGA miR-192 expression level. Analyses were repeated for the combined COADREAD cohort and for COAD and READ separately when sample size permitted. Overall survival was evaluated using Kaplan-Meier analysis with the log-rank test and univariable Cox regression. Diagnostic, clinicopathological, and survival analyses were reported separately to avoid overstating prognostic conclusions.
Candidate target-gene and hub-gene analyses
Candidate target genes were identified by integrating miRWalk prediction with differentially expressed genes from GSE89076 and GSE156355. The overlap among these three sources was retained for downstream exploratory analysis. A protein-protein interaction network was constructed using STRING, and hub genes were selected according to node degree and module connectivity using Cytoscape-based cytoHubba and MCODE analyses. The retained hub genes (KIF20A, TPX2, CDCA5, CCNB1, CDK1, PLP1, NRXN1, GRIK3, and KIF5C) were further evaluated using TCGA RNA-seq data. Tumor-normal expression differences, Spearman correlations with TCGA miR-192 expression, and overall-survival associations were summarized as supplementary exploratory analyses. These genes were considered candidate targets only; no direct binding, rescue effect, or downstream protein-level regulation was inferred from bioinformatic data.
GO and KEGG enrichment analysis
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed for the candidate gene set and significant network modules. GO terms were grouped into biological process, cellular component, and molecular function categories. Enrichment results with p < 0.05 were considered statistically notable and were interpreted as hypothesis-generating pathway signals rather than experimentally validated mechanisms.
RNA extraction and qRT-PCR analysis
Total RNA was extracted from 45 paired CRC and adjacent noncancerous tissue samples using TRIzol reagent according to the manufacturer's protocol. RNA purity and concentration were assessed spectrophotometrically, and samples with A260/A280 values between 1.8 and 2.1 were used for reverse transcription. qRT-PCR was performed using a miRNA reverse transcription system and SYBR Green-based miRNA qPCR reagents from TransGen Biotech. Reverse transcription and qPCR amplification were performed according to the reagent manufacturer's instructions. Each qRT-PCR reaction was run in technical triplicate. RNU6B was used as the internal reference. The miR-192-5p primers were 5'-GGGGCTGACCTATGAATTGA-3' (forward) and 5'-CAGTGCAGGGTCCGAGGT-3' (reverse). The RNU6B primers were 5'-ATTGGAACGATACAGAGAAGATT-3' (forward) and 5'-GGAACGCTTCACGAATTTG-3' (reverse). Melt-curve analysis was used to assess amplification specificity when supported by the qPCR platform. Relative expression was calculated using the 2-ΔΔCt method. Paired tumor and adjacent tissue results were analyzed as paired clinical measurements.
Cell culture and transfection
The human CRC cell line HT29 was obtained from the authors' laboratory cell stock and had been authenticated and confirmed to be free of mycoplasma contamination. HT29 cells were cultured in McCoy's 5A medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 oC in a 5% CO2 atmosphere. Cells were passaged when they reached approximately 80% confluence. miR-192-5p mimics and the negative-control oligonucleotide con238 were purchased from Sangon Biotech (Shanghai, China) and used for transient transfection. HT29 cells were seeded in 6-well plates and transfected when the cell density reached approximately 20–30%, using a final oligonucleotide concentration of 50 nM and a lipid-based transfection protocol performed according to the reagent manufacturer's instructions. The exact reagent-to-oligonucleotide ratio and reagent amount followed the manufacturer's recommended protocol used in the authors' laboratory records. Cells were incubated with transfection complexes under standard culture conditions, and downstream transfection imaging and wound-healing assays were performed after transfection according to the experimental schedule. Only HT29 cell data were included in the present study.
Wound-healing assay
Transfected HT29 cells were seeded in 6-well plates and cultured until a confluent monolayer formed. A linear scratch was generated using a sterile 200 µL pipette tip, and detached cells were removed by washing with PBS. The medium was then replaced with serum-free medium to reduce proliferation-related effects during migration assessment. Wound images were captured at 0 h, 24 h, and 48 h. The wound-healing assay was performed in three independent biological replicates, and multiple microscopic fields were measured for each group and time point. Wound width was quantified from microscopy images using image-analysis software; the exact software version used for the original measurements was not available in the source records. Wound-healing rate was calculated as [(wound width at 0 h - wound width at the indicated time point) / wound width at 0 h] x 100%. Because the available manuscript data do not include a separate proliferation-control experiment, the wound-healing results were interpreted as migration-related but not as a complete exclusion of proliferation effects.
Immune signature correlation analysis
The association between TCGA miR-192 expression and immune-related signatures was reassessed using primary tumor RNA-seq data. A total of 425 tumor samples with matched miRNA and RNA-seq data were included in the signature-score correlation analysis. miR-192 expression showed statistically significant but weak negative correlations with several immune signatures after false-discovery rate correction, including Tfh cells (rho = -0.205, FDR = 0.00049), activated dendritic cells (rho = -0.186, FDR = 0.00098), Treg cells (rho = -0.185, FDR = 0.00098), plasmacytoid dendritic cells (rho = -0.169, FDR = 0.0027), and macrophages (rho = -0.149, FDR = 0.0092). Positive correlations were observed mainly for eosinophil (rho = 0.139, FDR = 0.0155) and Th17-cell signatures (rho = 0.121, FDR = 0.0296). These findings indicate heterogeneous and generally weak immune associations rather than a uniform positive immune-infiltration pattern.
Statistical analysis
Statistical analyses were performed using GraphPad Prism 9.0, SPSS 25.0, and R software 3.6.3. Continuous variables were summarized as mean ± standard deviation or median with interquartile range, as appropriate. Paired tissue comparisons were analyzed using paired tests when tumor and adjacent tissue samples were matched. Nonparametric tests were used when normality assumptions were not met. ROC curves, balanced resampling, and leave-one-normal-out analysis were used for exploratory diagnostic discrimination; Chi-square tests were used for categorical clinicopathological associations; log-rank and Cox models were used for survival analyses; and Spearman correlation was used for immune and hub-gene analyses. Two-sided p < 0.05 was considered statistically significant, with false-discovery rate correction reported for immune signature correlations.
Results
Expression analysis in public datasets and paired clinical tissues
The expression pattern of miR-192 was reassessed using GDC-derived TCGA-COAD and TCGA-READ miRNA-seq data (Figure 1). In these files, the database-level entry was annotated as hsa-mir-192 rather than directly as the mature hsa-miR-192-5p species; therefore, this public-database result was interpreted as a TCGA miR-192 signal. In the combined COADREAD cohort, 616 primary tumors and 11 solid-tissue normal samples were analyzed. Tumor samples showed higher miR-192 expression than normal samples (median 16.05 vs 13.46 log2[RPM + 1], Mann-Whitney p = 9.96 x 10-8). The same higher-tumor direction was observed in COAD alone (455 tumors and 8 normal samples, p = 9.89 x 10-8) and READ alone (161 tumors and 3 normal samples, p = 1.94 x 10-5). Paired TCGA analysis also showed higher tumor expression in 11 matched cases (median tumor-normal difference = 3.13 log2[RPM + 1], Wilcoxon p = 0.0029). These results remained discordant with the local paired qRT-PCR result, which supported reduced mature miR-192-5p expression in CRC tissues (Figure 2).
Diagnostic performance and clinicopathological associations
TCGA-based ROC analysis was repeated using the definition of discrimination between tumor and normal samples described above (Figure 3). The AUC was 0.968 for the combined COADREAD cohort, 0.947 for COAD alone, and 0.994 for READ alone. Additional balanced resampling yielded median AUCs of 0.983 for COADREAD, 0.984 for COAD, and 1.000 for READ; however, the 95% confidence intervals were wide for COADREAD (0.868–1.000) and COAD (0.781–1.000), and READ included only three normal samples. Leave-one-out normal analysis also showed that the ROC estimate was sensitive to the small normal group. Therefore, the ROC result was retained only as exploratory evidence of tumor-normal separation, not as sufficient evidence for a clinical diagnostic biomarker. Clinicopathological association analysis based on median TCGA miR-192 expression showed significant associations with N stage (p = 0.019) and age (p = 0.004), whereas T stage (p = 0.110) and M stage (p = 0.932) were not significant (Table 1).
Survival analysis
Overall survival analysis did not support a significant prognostic association for TCGA miR-192 expression in CRC (Figure 3). In the combined COADREAD cohort, the median-cut Kaplan-Meier comparison was not significant (log-rank p = 0.857), and univariable Cox regression also showed no significant association (HR = 0.934 per log2 expression unit, p = 0.374). The same conclusion was observed in COAD (log-rank p = 0.889; Cox p = 0.493) and READ (log-rank p = 0.807; Cox p = 0.716). Thus, the data do not support a strong prognostic-biomarker claim for miR-192 in CRC.
Candidate target genes and functional enrichment
Candidate target-gene analysis was performed by integrating miRWalk prediction with differentially expressed genes from the GEO datasets GSE89076 and GSE156355. This screening identified 93 candidate genes for downstream network and enrichment analyses (Figure 4). Protein-interaction analysis retained KIF20A, TPX2, CDCA5, CCNB1, CDK1, PLP1, NRXN1, GRIK3, and KIF5C as exploratory hub genes (Figure 5). TCGA RNA-seq sensitivity analysis showed that KIF20A, TPX2, CDCA5, CCNB1, and CDK1 were higher in tumor tissue than in normal tissue, whereas PLP1, NRXN1, GRIK3, and KIF5C were lower in tumor tissue. Correlation analysis with TCGA miR-192 expression was mixed: CDCA5, CCNB1, and CDK1 showed positive correlations, whereas PLP1, NRXN1, and KIF5C showed negative correlations. Enrichment analysis highlighted representative terms related to drug metabolism by cytochrome P450, metabolism of xenobiotics by cytochrome P450, retinol metabolism, glutamatergic synapse-related terms, and presynapse organization (Figure 6). These findings support the biological relevance of the hub-gene network but do not establish direct miR-192-5p targeting.
Cell migration assay
HT29 cells were transfected with miR-192 mimics to assess the functional effects of miR-192 upregulation. Fluorescence microscopy confirmed successful transfection in the miR-192-5p mimic and negative-control transfection groups (Figure 7). In the wound-healing assay, the migration rate of miR-192-transfected HT29 cells was lower than that of untransfected HT29 cells and negative-control-transfected cells at both 24 h and 48 h (Figure 8). At 24 h, the migration rate was 9.75 +/- 2.43% in the miR-192-transfected group compared with 15.69 +/- 3.47% in the untransfected group and 15.45 +/- 3.92% in the negative-control group (P = 0.010). At 48 h, the migration rate was 20.04 +/- 2.54% in the miR-192-transfected group compared with 26.84 +/- 7.65% and 27.98 +/- 6.78% in the two control groups, respectively (P = 0.021). These results suggest that miR-192 upregulation suppresses HT29 cell migration in vitro.
Immune signature correlation analysis
The association between TCGA miR-192 expression and immune-related signatures was reassessed using primary tumor RNA-seq data (Figure 9). A total of 425 tumor samples with matched miRNA and RNA-seq data were included in the signature-score correlation analysis. miR-192 expression showed weak negative correlations with multiple immune cell signatures, including Tfh cells (rho = -0.205, p = 2.15e-05), activated dendritic cells (rho = -0.186, p = 0.00011), Treg cells (rho = -0.185, p = 0.00013), plasmacytoid dendritic cells (rho = -0.169, p = 0.00046), macrophages (rho = -0.149, p = 0.0020), Tcm cells (rho = -0.136, p = 0.0050), NK CD56bright cells (rho = -0.132, p = 0.0065), and Tem cells (rho = -0.129, p = 0.0078). Weak positive correlations were observed for eosinophil (rho = 0.139, p = 0.0040) and Th17-cell (rho = 0.121, p = 0.0129) signatures. Thus, the immune association of miR-192 in CRC appears heterogeneous and should not be described as uniformly positive across immune infiltrates.
Data Availability
The public datasets analyzed in this study are available from the GDC/TCGA portal and the Gene Expression Omnibus (GEO) database under accession numbers GSE89076 and GSE156355. The dry-lab analyses and reanalysis generated during this study are provided in Supplementary Table 1. The underlying experimental data supporting the qRT-PCR and wound-healing analyses, including raw qRT-PCR Ct values, wound-width measurements, original microscopy images, and related experimental data, are provided in Supplementary Folder 1.
Figure 1: Pan-cancer expression overview of miR-192 in TCGA. The x-axis shows tumor abbreviations, and the y-axis shows relative miR-192 expression. Blue indicates normal tissue, and red indicates tumor tissue. Statistical significance is indicated as follows: * = p < 0.05, ** = p < 0.01, *** = p < 0.001; NS = not significant. Please click here to view a larger version of this figure.
Figure 2: miR-192/miR-192-5p expression in CRC datasets and paired tissues. (A) TCGA/GDC miRNA-seq analysis showing higher database-level hsa-mir-192 signal in CRC tumor samples than in normal samples. (B) qRT-PCR analysis of paired local tissue samples showing lower mature miR-192-5p expression in CRC tissues than in adjacent noncancerous tissues. Please click here to view a larger version of this figure.
Figure 3: Exploratory diagnostic and survival analyses based on TCGA hsa-mir-192 expression. (A–C) Kaplan-Meier survival curves. Survival analysis did not support a significant prognostic association. (D) ROC curve for tumor-normal discrimination in the TCGA cohort. The AUC should be interpreted cautiously because of the small number of normal samples. Please click here to view a larger version of this figure.
Figure 4: Candidate target-gene screening and protein interaction analysis. (A) Venn diagram showing the overlap among miRNA-target prediction and differentially expressed genes from GSE89076 and GSE156355. (B) Protein-protein interaction network of the 93 candidate genes. Nodes represent genes, and edges represent predicted or curated gene associations. Please click here to view a larger version of this figure.
Figure 5: Hub-gene screening from the candidate target network. (A,B) Network modules identified by module screening. (C) Hub-gene ranking based on network connectivity. (D) Candidate hub targets are retained after integrating the module and degree-based screening. Please click here to view a larger version of this figure.
Figure 6: Selected GO and KEGG enrichment terms for candidate genes and significant network modules. The bubble plot summarizes representative enriched biological functions and pathways. Enrichment results are exploratory and require experimental validation. Please click here to view a larger version of this figure.
Figure 7: HT29 cell transfection images. (A) Representative HT29 cells. (B) Fluorescence image of HT29 cells transfected with miR-192-5p mimic. (C) Fluorescence image of HT29 cells transfected with negative control con238. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 8: Wound-healing assay in HT29 cells. (A) Representative scratch images at 0 h, 24 h, and 48 h in untransfected HT29 cells, miR-192-5p mimic-transfected HT29 cells, and con238-transfected HT29 cells. (B) Quantification of wound-healing rates at 24 h and 48 h, shown as percentages. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 9: Correlation between TCGA miR-192 expression and immune-related signature scores in CRC tumor samples. The analysis showed heterogeneous immune associations, with several weak negative correlations and, mainly for eosinophil and Th17-cell signatures, positive correlations. Please click here to view a larger version of this figure.
| Characteristics | Low expression of TCGA miR-192 | High expression of TCGA miR-192 | p value | Method |
| n | 308 | 308 | ||
| T stage, n (%) | 0.11 | Chisq.test | ||
| T1 | 9 (1.5%) | 11 (1.8%) | ||
| T2 | 44 (7.2%) | 59 (9.6%) | ||
| T3 | 209 (34.1%) | 211 (34.4%) | ||
| T4 | 43 (7%) | 27 (4.4%) | ||
| N stage, n (%) | 0.019 | Chisq.test | ||
| N0 | 156 (25.5%) | 190 (31%) | ||
| N1 | 88 (14.4%) | 63 (10.3%) | ||
| N2 | 61 (10%) | 54 (8.8%) | ||
| M stage, n (%) | 0.932 | Chisq.test | ||
| M0 | 214 (39.9%) | 234 (43.7%) | ||
| M1 | 41 (7.6%) | 47 (8.8%) | ||
| Age, median (IQR) | 66 (56, 75) | 69 (60, 77) | 0.004 | Wilcoxon |
Table 1: Association between TCGA miR-192 expression groups and clinicopathological characteristics in CRC. High and low expression groups were defined by the median TCGA miR-192 expression level. In the summarized table, N stage and age were significant, whereas T stage and M stage were not significant.
Supplementary Table 1: Dry-lab reanalysis results generated for this study. The workbook includes TCGA/GDC expression comparisons, balanced ROC sensitivity analyses, leave-one-normal-out analyses, clinicopathological association analyses, survival analyses, hub-gene expression and correlation summaries, and immune-signature correlation results.Please click here to download this file.
Supplementary Folder 1: Raw experimental data supporting the qRT-PCR, cell transfection, and wound-healing experiments. These files include original qRT-PCR Ct values, flow cytometry/transfection data, wound-width measurements, and representative microscopy images used for the analyses presented in the manuscript.Please click here to download this file.
Discussion
This study evaluated the clinical and biological relevance of miR-192-5p in CRC by combining public dataset analysis, local tissue validation, target network prediction, immune signature analysis, and in vitro migration experiments. The analyses used verified TCGA sample definitions, separated COAD and READ where possible, used log2(RPM + 1) miRNA values, distinguished paired from unpaired comparisons, tested ROC robustness, and added hub-gene RNA-seq sensitivity analyses. The central finding is that the public TCGA/GDC miR-192 signal and the local mature miR-192-5p qRT-PCR result point in opposite expression directions, emphasizing the need to distinguish database-level precursor annotations from mature miRNA measurements9,11.
This discordance should not be explained solely by unequal sample sizes. The TCGA normal group was small, but unpaired, paired, COAD-only, and READ-only analyses all showed the same higher-tumor direction in TCGA/GDC. Several factors may contribute, including differences in sample source, sequencing versus qRT-PCR platforms, database-level hsa-mir-192 annotation versus mature hsa-miR-192-5p measurement, RNA processing, tumor purity, adjacent-tissue definition, and population characteristics. Similar concerns about biomarker interpretation and tumor heterogeneity have been emphasized in CRC studies5,6,7,8. Therefore, the findings support a context- and platform-dependent interpretation of miR-192-5p in CRC rather than a single universal expression direction.
The diagnostic analysis should also be interpreted cautiously. Although the TCGA-based ROC analysis produced a high AUC, the comparison included many more tumor samples than normal samples, with only 11 normal samples in the combined COADREAD cohort. Balanced resampling still showed high median AUC values, but the intervals were wide because each balanced comparison was constrained by the small normal group. Thus, the ROC result supports exploratory tumor-normal separation within TCGA/GDC but is insufficient to establish miR-192-5p as a clinical diagnostic biomarker. This interpretation is consistent with the general distinction between exploratory biomarker signals and clinically validated diagnostic or prognostic biomarkers4,8.
Survival analysis did not show a significant association between miR-192 expression and overall survival in COADREAD, COAD, or READ. In the current analysis, miR-192 may be associated with selected clinicopathological variables, but the available data do not support a strong prognostic-biomarker conclusion. This distinction is important because prognostic biomarkers require evidence that marker status is associated with outcome independently and reproducibly across clinically relevant contexts8.
The in vitro wound-healing assay provides functional evidence that miR-192 upregulation can reduce HT29 cell migration. This result is consistent with a potential tumor-suppressive function in CRC cells and with the local qRT-PCR direction. Previous work has also supported roles for miRNAs, including miR-192-related pathways, in regulating epithelial-to-mesenchymal transition, cell fate, and migration-related phenotypes15,16,17. However, the functional data remain incomplete. Additional wet-lab validation would be needed to support a mechanistic tumor-suppressor model, including repeat transfections in at least one additional CRC cell line, proliferation controls for wound-healing assays, Transwell migration/invasion assays, dual-luciferase validation of selected targets, target mRNA and protein measurements, and rescue experiments.
The hub-gene analysis should be regarded as hypothesis-generating. TCGA RNA-seq supplementation showed strong tumor-normal expression differences for the retained hub genes, but their correlations with miR-192 were mixed. Several cell-cycle genes, including CDCA5, CCNB1, and CDK1, were positively rather than negatively correlated with miR-192, which does not fit a simple direct miRNA-target repression model. In contrast, PLP1, NRXN1, and KIF5C showed negative correlations. These results support the biological relevance of the network but indicate that direct targeting cannot be concluded without luciferase, protein-level, and rescue validation. This cautious interpretation is consistent with the exploratory nature of in silico target-network analyses10,11.
The immune analysis further supports cautious interpretation. Tumor RNA-seq immune signature scores showed that miR-192 expression was negatively correlated with several immune signatures and positively correlated mainly with eosinophil and Th17-cell signatures. Most correlations were weak, even after false-discovery rate correction. Because immune-signature methods infer pathway or cell-state enrichment from bulk expression data, these findings should be considered exploratory and require validation using orthogonal immune-profiling approaches13.
This study has several limitations. First, the local validation cohort was relatively small and came from a single center. Second, public-dataset analysis and local qRT-PCR validation used different platforms and measurement levels. Third, TCGA normal miRNA-seq samples were few, limiting diagnostic inference. Fourth, the hub-gene and immune analyses were computational and require experimental validation. Fifth, the available functional experiments were limited to HT29 wound-healing data. These limitations justify presenting the study as an integrated exploratory analysis rather than as definitive evidence that miR-192-5p is a validated diagnostic, prognostic, or mechanistic biomarker in CRC6,8.
In summary, miR-192-5p appears to be biologically relevant to CRC, but its expression pattern and clinical interpretation are complex. TCGA/GDC analysis showed higher miR-192 signal in CRC tumor samples, whereas qRT-PCR in a local paired-tissue cohort showed lower mature miR-192-5p expression. ROC analysis indicated exploratory tumor-normal separation but was limited by the small TCGA normal group. No significant association with overall survival was observed. Hub-gene and immune analyses provide hypothesis-generating evidence, and HT29 wound-healing data suggest that miR-192-5p upregulation may reduce cell migration. Further independent cohorts and targeted wet-lab experiments are required before miR-192-5p can be considered a validated diagnostic, prognostic, or mechanistic biomarker in CRC.
Disclosures
The authors have no conflict of interest to declare.
Acknowledgements
The authors thank the Molecular Diagnostic Laboratory of Tumor of Inner Mongolia Medical University for technical support. This work was supported by the Scientific Research Project of the Yancheng Municipal Health Commission (Grant No. YK2023101), which also funded the article processing charge.
Materials
| Name | Company | Catalog Number | Comments |
|---|---|---|---|
| 0.25%+2:60 trypsin-EDTA | Biological Industries | 03-054-1 | Cell digestion/passaging |
| 10% ammonium persulfate | Solarbio Science & Technology Co., Ltd. | A8090 | SDS-PAGE reagent |
| 10% sodium dodecyl sulfate | Solarbio Science & Technology Co., Ltd. | SIS1070 | SDS-PAGE reagent |
| 4x protein loading buffer | Solarbio Science & Technology Co., Ltd. | P1016 | Protein sample preparation |
| -80 C ultra-low temperature freezer | Qingdao Haier Biomedical Co., Ltd. | https://www.haiermedical.com/ | Sample storage |
| Anti-Myc mycoplasma removal reagent | Wuhan Procell Life Science & Technology Co., Ltd. | P-CMR-001 | Mycoplasma decontamination |
| ARL2, BATF2, MEN1, and beta-actin primer sets | Sangon Biotech Co., Ltd., Shanghai, China | Custom synthesized | mRNA qPCR primers |
| Automatic microplate reader | PerkinElmer, USA | VICTOR Nivo | Absorbance measurement |
| BCA Protein Assay Kit | Thermo Fisher Scientific | 23225 | Protein quantification |
| Biosafety cabinet / clean bench | Qingdao Haier Biomedical Co., Ltd. | https://www.haiermedical.com/ | Aseptic cell culture |
| Cell counting chamber / hemocytometer | Shanghai Qiujing Biochemical Reagent Instrument Co., Ltd. | http://www.shqiujing.net/ | Cell counting |
| Cell Counting Kit-8 (CCK-8) | Biosharp | BS350B | Cell proliferation/cytotoxicity assay |
| Cell culture flasks, culture plates, centrifuge tubes, and pipette tips | Corning | https://www.corning.com/in/en.html | General cell culture consumables |
| Cell Cycle and Apoptosis Detection Kit | Meilun Biotechnology Co., Ltd., Dalian, China | https://www.chemicalbook.com/ShowSupplierProductsList13366/0_EN.htm | Flow-cytometric cell-cycle/apoptosis analysis |
| CO2 cell incubator | Qingdao Haier Biomedical Co., Ltd. | https://www.haiermedical.com/ | Cell culture |
| con238 negative-control oligonucleotide | Sangon Biotech Co., Ltd., Shanghai, China | Custom synthesized | Transfection control |
| Cytoscape software | Cytoscape Consortium | https://cytoscape.org/ | Network module and hub-gene visualization |
| Digital thermostatic magnetic stirrer | Changzhou Saipu Experimental Instrument Factory | https://czxtyq.en.alibaba.com/ | Solution preparation |
| Electric thermostatic blast drying oven | Nanjing Wohuan Technology Industry Co., Ltd. | https://www.wtsensor.com/about-wt/aboutwt/ | Drying/sterilization support |
| Electrophoresis system | Bio-Rad Laboratories | Mini-PROTEAN 3 Cell | Gel electrophoresis |
| Electrotransfer system | Hoefer | TE77XP | Protein transfer |
| Fetal bovine serum | Biological Industries | 04-001-1ACS | Cell culture supplement |
| Flow cytometer | ACEA Biosciences | NovoCyte | Flow cytometry |
| Fluorescence inverted microscope | Ningbo Sunny Instruments Co., Ltd. | https://www.sunny-instrument.com/en | Cell and wound imaging |
| GDC/TCGA portal | National Cancer Institute | Open-access database | Public miRNA-seq, RNA-seq, and clinical data |
| GEO database | National Center for Biotechnology Information | GSE89076 and GSE156355 | Public gene-expression datasets |
| GraphPad Prism | GraphPad Software | Version 9 | Statistical analysis and graphing |
| High-speed refrigerated centrifuge | Shanghai Huxiangyi Centrifuge Instrument Co., Ltd. | https://www.bioridgecentrifuge.com/ | Sample preparation |
| HT29 cells | Authors' laboratory cell stock | Not applicable | Human colorectal cancer cell line |
| Image analysis software | ImageJ | Not applicable | Wound-width quantification |
| Imaging system | Tanon Science & Technology Co., Ltd. | Tanon-5200 | Gel/blot imaging |
| IncuCyte S3 dynamic cell observation and functional analysis system | Beijing Sensi Wantong Technology Co., Ltd. | IncuCyte S3 | Live-cell imaging/function analysis |
| Intelligent thermostatic water bath | Leica | HI1210 | Temperature-controlled incubation |
| Lentivirus and infection reagent | Shanghai GeneChem Co., Ltd. | https://synapse.patsnap.com/organization/04ffbf33ff2a6311226a8fa05498cec4 | Stable cell transduction |
| Manual pipettes | Gilson | PIPETMAN P | Liquid handling |
| McCoy's 5A basal medium | Wuhan Procell Life Science & Technology Co., Ltd. | PM150710 | HT29 cell culture medium |
| Medical centrifuge | Zhuhai manufacturer not specified | 3200 | Sample preparation |
| Medical refrigerator-freezer | Qingdao Haier Biomedical Co., Ltd. | https://www.haiermedical.com/ | Reagent/sample storage |
| Meilunbio Fixer Supersensitive ECL Chemiluminescence Substrate | Meilun Biotechnology Co., Ltd., Dalian, China | MA0186 | Immunoblot detection |
| Mini centrifuge | Beijing DLAB Scientific Co., Ltd. | https://www.dlabsci.com/ | Sample preparation |
| miR-192-5p and U6 primer sets | TIANGEN Biotech Co., Ltd., Beijing, China | Custom synthesized | miRNA qPCR primers |
| miR-192-5p mimic | Sangon Biotech Co., Ltd., Shanghai, China | Custom synthesized | Transient transfection oligonucleotide |
| miRcute Enhanced miRNA First-Strand cDNA Synthesis Kit | TIANGEN Biotech Co., Ltd., Beijing, China | KR211 | miRNA reverse transcription |
| miRcute Enhanced miRNA Fluorescence Quantitation Kit (SYBR Green FP) | TIANGEN Biotech Co., Ltd., Beijing, China | FP411 | miRNA qPCR |
| miRWalk | Public database | Web database | miRNA-target prediction database |
| Nitrocellulose membrane | Millipore | HATF00010 | Protein transfer membrane |
| Nonfat dry milk | Solarbio Science & Technology Co., Ltd. | D8340 | Immunoblot blocking reagent |
| Paired colorectal cancer and adjacent noncancerous tissues | Authors' institution | Not applicable | Clinical specimens |
| PCR tubes | Corning | PCR-02-C | PCR/qRT-PCR setup |
| Penicillin-streptomycin solution | Biological Industries | 03-031-1B | Cell culture antibiotic |
| PerfectStart Uni RT&qPCR Kit | TransGen Biotech Co., Ltd., Beijing, China | AUQ-01 | mRNA reverse transcription and qPCR |
| Phosphate-buffered saline (PBS) | Biological Industries | 02-023-1A | Cell washing buffer |
| Prestained protein marker | Fermentas | 26619 | Protein molecular-weight marker |
| Pure water preparation system | China Moer | Not available | Pure water preparation |
| Puromycin | Phygene | PH1143 | Selection of transduced cells |
| R software | R Foundation for Statistical Computing | Version 3.6.3 | Statistical analysis and bioinformatics |
| Rapid transfer buffer | Crisbio | Not applicable | Protein electrotransfer |
| Real-time fluorescence quantitative PCR system | Applied Biosystems (ABI), USA | Not available | qRT-PCR detection |
| RIPA lysis buffer | Solarbio Science & Technology Co., Ltd. | R0020 | Protein extraction |
| RNase-free microcentrifuge tubes | Corning | 3208 | RNA handling |
| Serum-free cell cryopreservation medium | New Cell & Molecular Biotech Co., Ltd. | C40100 | Cell cryopreservation |
| SPSS Statistics | IBM | Version 25 | Statistical analysis |
| Stabilized antibody diluent | Solarbio Science & Technology Co., Ltd. | Not applicable | Primary/secondary antibody dilution |
| Sterile pipette tips | Wuhan Servicebio Technology Co., Ltd. | TP-10-C | Wound-healing scratch tool |
| STRING database | STRING Consortium | Web database | Protein-interaction network analysis |
| TEMED | Solarbio Science & Technology Co., Ltd. | T8090 | SDS-PAGE reagent |
| TransZol Up | TransGen Biotech Co., Ltd., Beijing, China | ET111-01 | RNA extraction |
| Tris-HCl buffer, pH 6.8 | Solarbio Science & Technology Co., Ltd. | T1020 | SDS-PAGE stacking-gel buffer |
| Tris-HCl buffer, pH 8.8 | Solarbio Science & Technology Co., Ltd. | T1010 | SDS-PAGE resolving-gel buffer |
| Tween-20 | Solarbio Science & Technology Co., Ltd. | T8220 | Immunoblot washing reagent |
| Vertical pressure steam sterilizer | Shanghai Boxun Medical Biological Instrument Corp. | BXM-30R | Sterilization |
References
- Sung, H. et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 71 (3), 209-249 (2021).
- Siegel, R. L., Miller, K. D., Jemal, A. Cancer statistics, 2019. CA Cancer J Clin. 69 (1), 7-34 (2019).
- Issa, I. A., Noureddine, M. Colorectal cancer screening: an updated review of the available options. World J Gastroenterol. 23 (28), 5086-5096 (2017).
- Song, L., Jia, J., Peng, X., Xiao, W., Li, Y. The performance of the SEPT9 gene methylation assay and a comparison with other CRC screening tests: a meta-analysis. Sci Rep. 7 (1), 3032 (2017).
- Linnekamp, J. F., Wang, X., Medema, J. P., Vermeulen, L. Colorectal cancer heterogeneity and targeted therapy: a case for molecular disease subtypes. Cancer Res. 75 (2), 245-249 (2015).
- Sveen, A., Kopetz, S., Lothe, R. A. Biomarker-guided therapy for colorectal cancer: strength in complexity. Nat Rev Clin Oncol. 17 (1), 11-32 (2020).
- Whiffin, N. et al. Identification of susceptibility loci for colorectal cancer in a genome-wide meta-analysis. Hum Mol Genet. 23 (17), 4729-4737 (2014).
- Ballman, K. V. Biomarker: predictive or prognostic? J Clin Oncol. 33 (33), 3968-3971 (2015).
- Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 116 (2), 281-297 (2004).
- Forterre, A., Komuro, H., Aminova, S., Harada, M. A comprehensive review of cancer microRNA therapeutic delivery strategies. Cancers (Basel). 12 (7), 1852 (2020).
- Ren, F. J., Yao, Y., Cai, X. Y., Fang, G. Y. Emerging role of miR-192-5p in human diseases. Front Pharmacol. 12, 614068 (2021).
- Nagy, Z. B. et al. Colorectal adenoma and carcinoma specific miRNA profiles in biopsy and their expression in plasma specimens. Clin Epigenetics. 9, 22 (2017).
- Hanzelmann, S., Castelo, R., Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 14 (1), 7 (2013).
- Staiteieh, S. A. et al. Study of microRNA expression profiling as biomarkers for colorectal cancer patients in Lebanon. Mol Clin Oncol. 16 (2), 39 (2022).
- Przygodzka, P. et al. Regulation of miRNAs by Snail during epithelial-to-mesenchymal transition in HT29 colon cancer cells. Sci Rep. 9 (1), 2165 (2019).
- Sun, C. Y., Zhang, X. P., Wang, W. Coordination of miR-192 and miR-22 in p53-mediated cell fate decision. Int J Mol Sci. 20 (19), 4768 (2019).
- Zhao, L., Wang, B., Sun, L., Sun, B., Li, Y. Association of miR-192-5p with atherosclerosis and its effect on proliferation and migration of vascular smooth muscle cells. Mol Biotechnol. 63 (12), 1244-1251 (2021).