Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 May 8;8(1):7189.
doi: 10.1038/s41598-018-25320-z.

Cell-type Specific Expression of Oncogenic and Tumor Suppressive microRNAs in the Human Prostate and Prostate Cancer

Affiliations
Free PMC article

Cell-type Specific Expression of Oncogenic and Tumor Suppressive microRNAs in the Human Prostate and Prostate Cancer

Binod Kumar et al. Sci Rep. .
Free PMC article

Abstract

MiR-1 and miR-143 are frequently reduced in human prostate cancer (PCa), while miR-141 and miR-21 are frequently elevated. Consequently, these miRNAs have been studied as cell-autonomous tumor suppressors and oncogenes. However, the cell-type specificity of these miRNAs is not well defined in prostate tissue. Through two different microdissection techniques, and droplet digital RT-PCR, we quantified these miRNAs in the stroma and epithelium of radical prostatectomy specimens. In contrast to their purported roles as cell-autonomous tumor suppressors, we found miR-1 and miR-143 expression to be predominantly stromal. Conversely, miR-141 was predominantly epithelial. miR-21 was detected in both stroma and epithelium. Strikingly, the levels of miR-1 and miR-143 were significantly reduced in tumor-associated stroma, but not tumor epithelium. Gene expression analyses in human cell lines, tissues, and prostate-derived stromal cultures support the cell-type selective expression of miR-1, miR-141, and miR-143. Analyses of the PCa Genome Atlas (TCGA-PRAD) showed a strong positive correlation between stromal markers and miR-1 and miR-143, and a strong negative correlation between stromal markers and miR-141. In these tumors, loss of miR-1 and gain of miR-21 was highly associated with biochemical recurrence. These data shed new light on stromal and epithelial miRNA expression in the PCa tumor microenvironment.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Aberrantly expressed miRNAs in human prostate cancer. Dashed line separates miRNAs decreased in cancer versus increased in cancer. (A) Differential levels of miR-1, miR-143, miR-141, and miR-21 in macrodissected normal (N) and malignant (C) prostate tissue from the University of Washington (n = 15 per group) by RT-ddPCR. Copy numbers normalized per 10,000 RNU6B. (B) RT-ddPCR quantification of miRNAs in a separate set of macrodissected normal and malignant prostate tissues from Johns Hopkins radical prostatectomies (n = 4 per group). Copy numbers normalized per 10,000 copies cel-39 internal control. These sample are applied to xMD in Figs 2 and 4. Bars represent mean and standard error. Log Scale. P-values determined by Wilcox Rank Sum analysis. *p < 0.05.
Figure 2
Figure 2
Compartmentalized expression of miR-1, miR-141, and miR-143 in normal human prostate tissue. (A) Schematic of microdissection approaches applied to isolate prostatic stroma and epithelium from two separate sample sets. Johns Hopkins University (JHU). University of Illinois at Chicago (UIC). (B) Representative images of slides and membrane before and after α-SMA and AE1/AE3 xMD. Magnification, 2x or 10x. (C) RT-ddPCR quantification of miRNAs in xMD dissected prostate stroma (S) and epithelium (E) from JHU samples applied to Fig. 1B (n = 4). Dashed line separates miRNAs decreased in cancer versus increase in cancer. (D) RT-ddPCR quantification of miRNAs in LCM dissected prostate stroma (S) and epithelium (E) from normal prostate tissues microdissected at the UIC (n = 6). Blue symbols indicate patients who received Vitamin D. Bars represent mean and standard error; n.d., not detected. Log Scale. P-values determined by Wilcox Rank Sum analysis. *p < 0.05; **p < 0.01.
Figure 3
Figure 3
Cell and tissue type specific expression of miR-1, miR-143, miR-141, and miR-21. Expression levels from RNA sequencing data within the public Sequence Read Archive (SRA). Cell and tissue data were grouped according to tissue origin (Legend). Tissues ranked left to right from lower to higher expression levels per miRNA. (A) miR-1, (B) miR-141, (C) miR-143, (D) miR-21. Bars represent mean and standard error. Log scale.
Figure 4
Figure 4
Reduced miR-1 and miR-143 levels in PCa-associated tumor stroma. (A) Schematic of normal (N) versus cancer (C) comparison between miRNA levels in xMD microdissected stroma versus epithelium. Stroma was isolated by α-SMA xMD, and epithelium isolated by AE1/AE3 xMD, using JHU samples from Fig. 1B (n = 4). Levels of each miRNA quantified by RT-ddPCR and copy number normalized per 10,000 cel-39. (B) miR-1 and miR-143 levels in normal versus malignant epithelium and stroma. (C) miR-141 and miR-21 levels in normal versus malignant epithelium and stroma. Normal sample values presented from Fig. 2C for comparison. Bars represent mean and standard error. Log Scale. P-values determined by Wilcox Rank Sum analysis. *p < 0.05.
Figure 5
Figure 5
Cell-selective miRNA expression in prostate cell lines and stromal cultures. (A) RT-ddPCR quantification of miRNAs in epithelial prostate, BPH, and PCa cells lines and normal fibroblast cells. Results are mean normalized copies per 10,000 RNU6B, plus standard error from two independent experiments. n.d., not detected. (B) Characterization of stromal cultures derived from normal prostate (nPrSC-4c, nPrSC-7c, and nPrSC-8c) or from PCa (PrCSC-41c, PrCSC-44c, and PrCSC-40c) by WB. LNCaP PCa cells and Human Foreskin Fibroblasts (HFF) were included as reference controls for epithelial and stroma cells, respectively; n.s, non-specific. (C) RT-ddPCR quantification of miRNAs in stromal cultures derived from normal prostate or from PCa. (D) Comparison of miRNA copy number from normal stromal cultures (n = 3) versus PCa stromal cultures (n = 3). Bars represent mean plus standard error. No significant differences detected, Wilcox Rank Sum analysis. MiRNA copy numbers normalized per 10,000 copies RNU6B. Log scale.
Figure 6
Figure 6
Cell-type selective miRNA expression in prostate cancer tissue from the TCGA. (A) Relative miRNA expression levels in normal (N) prostate and PCa (C) tissue from TCGA-PRAD. Results are presented as mean and standard error. Log 10 scale. P-values determined by Wilcox Rank Sum analysis. n.s., not significant. ***p < 0.001. (B) Pearson’s correlation analysis of stromal markers MYH11 and ACT2 in TCGA-PRAD samples. p < 2.2 × 10−16. Legend indicates reported stromal content quadrants of adjacent sections. (C) Pearson’s correlation analysis of miR-1 and stromal MYH11. p < 2.2 × 10−16. (D) Pearson’s correlation analysis of miR-143 and stromal MYH11. p < 2.2 × 10−16. (E) Pearson’s correlation analysis of miR-141 and stromal MYH11. p < 2.2 × 10−16. (F) Pearson’s correlation analysis of miR-21 and stromal MYH11. p = 8.2 × 10−9. Pearson’s correlation coefficient determined by linear regression analysis and fitted using least-squares method.

Similar articles

See all similar articles

Cited by 8 articles

See all "Cited by" articles

References

    1. Lu J, et al. MicroRNA expression profiles classify human cancers. Nature. 2005;435:834–838. doi: 10.1038/nature03702. - DOI - PubMed
    1. Chivukula RR, et al. An essential mesenchymal function for miR-143/145 in intestinal epithelial regeneration. Cell. 2014;157:1104–1116. doi: 10.1016/j.cell.2014.03.055. - DOI - PMC - PubMed
    1. Dimitrova N, et al. Stromal Expression of miR-143/145 Promotes Neoangiogenesis in Lung Cancer Development. Cancer discovery. 2016;6:188–201. doi: 10.1158/2159-8290.CD-15-0854. - DOI - PMC - PubMed
    1. Kent OA, McCall MN, Cornish TC, Halushka MK. Lessons from miR-143/145: the importance of cell-type localization of miRNAs. Nucleic Acids Res. 2014;42:7528–7538. doi: 10.1093/nar/gku461. - DOI - PMC - PubMed
    1. Emmert-Buck MR, et al. Laser capture microdissection. Science. 1996;274:998–1001. doi: 10.1126/science.274.5289.998. - DOI - PubMed

Publication types

MeSH terms

Feedback