Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Oct 17;82(20):3687-3700.
doi: 10.1158/0008-5472.CAN-22-0240.

Pan-Cancer Analysis of Canonical and Modified miRNAs Enhances the Resolution of the Functional miRNAome in Cancer

Rosario Distefano  1 , Luisa Tomasello  1 , Gian Luca Rampioni Vinciguerra  1   2 , Pierluigi Gasparini  1   3   4 , Yujia Xiang  5 , Marina Bagnoli  6 , Gioacchino P Marceca  7 , Paolo Fadda  8 , Alessandro Laganà  9 , Mario Acunzo  10 , Qin Ma  5 , Giovanni Nigita #  1 , Carlo M Croce #  1
Affiliations

Pan-Cancer Analysis of Canonical and Modified miRNAs Enhances the Resolution of the Functional miRNAome in Cancer

Rosario Distefano et al. Cancer Res. .

Abstract

Epitranscriptomic studies of miRNAs have added a new layer of complexity to the cancer field. Although there is fast-growing interest in adenosine-to-inosine (A-to-I) miRNA editing and alternative cleavage that shifts miRNA isoforms, simultaneous evaluation of both modifications in cancer is still missing. Here, we concurrently profiled multiple miRNA modification types, including A-to-I miRNA editing and shifted miRNA isoforms, in >13,000 adult and pediatric tumor samples across 38 distinct cancer cohorts from The Cancer Genome Atlas and The Therapeutically Applicable Research to Generate Effective Treatments data sets. The differences between canonical miRNAs and the wider miRNAome in terms of expression, clustering, dysregulation, and prognostic standpoint were investigated. The combination of canonical miRNAs and modified miRNAs boosted the quality of clustering results, outlining unique clinicopathologic features among cohorts. Certain modified miRNAs showed opposite expression from their canonical counterparts in cancer, potentially impacting their targets and function. Finally, a shifted and edited miRNA isoform was experimentally validated to directly bind and suppress a unique target. These findings outline the importance of going beyond the well-established paradigm of one mature miRNA per miRNA arm to elucidate novel mechanisms related to cancer progression.

Significance: Modified miRNAs may act as cancer biomarkers and function as allies or antagonists of their canonical counterparts in gene regulation, suggesting the concurrent consideration of canonical and modified miRNAs can boost patient stratification.

PubMed Disclaimer

Figures

None
Graphical abstract
Figure 1. Data preprocessing workflow, miRNA isoforms classification, and miRNA modification types distribution. A–C, In-house data preprocessing workflow and data sources (A), examples of annotated miRNA isoforms (B), and distribution of expressed molecules (geometric mean >3 RPM) across cohorts and miRNA modification types (C). C reports, for each cohort, the number of molecules per miRNA modification type as a percentage (%), along with the median (M) of the percentiles computed using molecules’ expression average. Squares in green and red color indicate (column-wise) whether a cohort belongs to the lower (25th) or upper quartile (75th), respectively.
Figure 1.
Data preprocessing workflow, miRNA isoforms classification, and miRNA modification types distribution. A–C, In-house data preprocessing workflow and data sources (A), examples of annotated miRNA isoforms (B), and distribution of expressed molecules (geometric mean >3 RPM) across cohorts and miRNA modification types (C). Panel C reports, for each cohort, the number of molecules per miRNA modification type as a percentage (%), along with the median (M) of the percentiles computed using molecules’ expression average. Squares in green and red indicate (column-wise) whether a cohort belongs to the lower (25th) or upper quartile (75th), respectively.
Figure 2. miRNA isoform-based clustering better delineates clinicopathologic stratification. A–C, Type of miRNA isoforms used by the four benchmarked groups of molecules (A), workflow used to benchmark the four groups (B), and clustering results for the four groups. C reports quality scores, the number of identified clusters, and the number of unique significant (Chi-square P < 0.01) clinicopathologic features identified for each benchmarked group. See Supplementary Table S2 for the list of specified clinicopathologic features, with the most prominent and significant features highlighted in green. See Supplementary Information for more details.
Figure 2.
miRNA isoform-based clustering better delineates clinicopathologic stratification. A–C, Type of miRNA isoforms used by the four benchmarked groups of molecules (A), workflow used to benchmark the four groups (B), and clustering results for the four groups (C). Panel C reports quality scores, the number of identified clusters, and the number of unique significant (Chi-square P < 0.01) clinicopathologic features identified for each benchmarked group. See Supplementary Table S2 for the list of specified clinicopathologic features, with the most prominent and significant features highlighted in green. See Supplementary Information for more details.
Figure 3. miRNA isoforms dysregulation across cohorts and tissues. Distribution of dysregulated molecules per cohort/tissues and miRNA modification type. The figure reports the number of significant (adjusted P <0.05) upregulated (U; linear fold change >1.5) and downregulated (D; linear fold change <−1.5) molecules, including A-to-I edited miRNA isoforms.
Figure 3.
miRNA isoforms dysregulation across cohorts and tissues. Distribution of dysregulated molecules per cohort/tissues and miRNA modification type. The figure reports the number of significant (adjusted P <0.05) upregulated (U; linear fold change >1.5) and downregulated (D; linear fold change <−1.5) molecules, including A-to-I edited miRNA isoforms.
Figure 4. miRNA isoforms experimental gene targeting validation. A, C, E–G, miR-101-3p and (B, D, H–J) miR-101-3p (−1|−2) experimental targeting validation in lung cancer cells. Expression of miR-101-3p (A) and miR-101-3p (−1|−2; B) in normal and tumor samples in TCGA-LUAD cohort. PTGS2 (C) and DSC2 (D) expression in TCGA-LUAD samples in miR-101-3p/miR-101-3p (−1|−2) first (Q1) and third (Q3) quartile. Luciferase assay for psiCHECK-2-PTGS2 3′ UTR WT (E) and psiCHECK-2-DSC2 3′ UTR WT (H) constructs cotransfected with mirVana miRNA mimics for miR-101-3p, miR-101-3p (−1|−2), and negative scramble miRNA control (SCR) in HEK-293 cells performed 24 hours after the transfection. Western blotting depicts the downregulation of PTGS2 (F, G) or DSC2 (I, J) proteins in A549 and H1299 cells, respectively, after miR-101-3p and miR-101-3p (−1|−2) overexpression. Densitometric quantification of western blotting signals (F, G; I, J) was performed using ImageJ (NIH, https://imagej.nih.gov/ij/, 1997–2018). K, M, O–Q, miR-381-3p and (L, N, R–T) miR-381-3p_4_A_G experimental targeting validation in TNBC cells. Expression of both miR-381-3p miRNA isoforms in normal and breast cancer samples in TCGA-BRCA cohort (K, L). Luciferase assay for psiCHECK-2-UBE2C 3′ UTR WT (O) and psiCHECK-2-SYT13 3′ UTR WT (R) constructs cotransfected with mirVana miRNA mimics for miR-381-3p, miR-381-3p_4_A_G, and negative scramble miRNA control (SCR) in HEK-293 cells performed 24 hours after the transfection. Western blotting represents the downregulation of UBE2C (P, Q) and SYT13 (S, T) proteins in MDA-MB-231 and HCC70 cells after miR-381-3p and miR-381-3p_4_A_G upregulation via mirVana miRNA mimic transfection. The histogram reports densitometric quantification of western blotting signals (P, Q; S, T) performed using ImageJ (NIH, https://imagej.nih.gov/ij/, 1997–2018). Pictures are representative of at least three experiments. The fold of increase in the graphics is the mean value of 3 replicates. P < 0.05 was considered statistically significant. Annotations for *, 0.01 ≤ P < 0.05; **, 0.001 ≤ P <0.01; ***, P < 0.001 are provided accordingly. Error bars indicate the three biological replicates’ standard deviation (SD). The horizontal bar in each violin-like plot indicates the median. In western blot experiments, VCL indicates VINCULIN.
Figure 4.
miRNA isoforms experimental gene targeting validation. miR-101-3p (A, C, and E–G) and miR-101-3p (−1|−2) (B, D, and H–J) experimental targeting validation in lung cancer cells. Expression of miR-101-3p (A) and miR-101-3p (−1|−2) (B) in normal and tumor samples in TCGA-LUAD cohort. PTGS2 (C) and DSC2 (D) expression in TCGA-LUAD samples in miR-101-3p/miR-101-3p (−1|−2) first (Q1) and third (Q3) quartile. Luciferase assay for psiCHECK-2-PTGS2 3′ UTR WT (E) and psiCHECK-2-DSC2 3′ UTR WT (H) constructs cotransfected with mirVana miRNA mimics for miR-101-3p, miR-101-3p (−1|−2), and negative scramble miRNA control (SCR) in HEK293 cells performed 24 hours after the transfection. Western blotting depicts the downregulation of PTGS2 (F and G) or DSC2> (I and J) proteins in A549 and H1299 cells, respectively, after miR-101-3p and miR-101-3p (−1|−2) overexpression. Densitometric quantification of Western blotting signals (F, G, I and J) was performed using ImageJ (NIH; https://imagej.nih.gov/ij/, 1997–2018). miR-381-3p (K, M, and O–Q) and miR-381-3p_4_A_G (L, N, and R–T) experimental targeting validation in TNBC cells. Expression of both miR-381-3p miRNA isoforms in normal and breast cancer samples in TCGA-BRCA cohort (K and L). Luciferase assay for psiCHECK-2-UBE2C 3′ UTR WT (O) and psiCHECK-2-SYT13 3′ UTR WT (R) constructs cotransfected with mirVana miRNA mimics for miR-381-3p, miR-381-3p_4_A_G, and negative scramble miRNA control (SCR) in HEK293 cells performed 24 hours after the transfection. Western blotting represents the downregulation of UBE2C (P and Q) and SYT13 (S and T) proteins in MDA-MB-231 and HCC70 cells after miR-381-3p and miR-381-3p_4_A_G upregulation via mirVana miRNA mimic transfection. The histogram reports densitometric quantification of Western blotting signals (P, Q, S, and T) performed using ImageJ (NIH; https://imagej.nih.gov/ij/, 1997–2018). Pictures are representative of at least three experiments. The fold of increase in the graphics is the mean value of 3 replicates. P < 0.05 was considered statistically significant. Annotations for *, 0.01 ≤ P < 0.05; **, 0.001 ≤ P <0.01; ***, P < 0.001 are provided accordingly. Error bars indicate the three biological replicates’ SD. The horizontal bar in each violin-like plot indicates the median. In Western blot experiments. VCL, VINCULIN.
Figure 5. Overall and RFS risk score–based signatures. A and B, Overview of risk score–based signatures for OS and RFS (A) and results of statistical tests performed across groups of molecules of different sizes (B). A reports the number of identified signatures per cohort, the AUC, and the P (log-rank test). See Supplementary Fig. S3 and Supplementary Table S5 for more detailed information regarding the workflow used and the list of identified signatures.
Figure 5.
Overall and RFS risk score–based signatures. A and B, Overview of risk score–based signatures for OS and RFS (A) and results of statistical tests performed across groups of molecules of different sizes (B). Panel A reports the number of identified signatures per benchmarked group, the AUC, and the P (log-rank test). See Supplementary Fig. S3 and Supplementary Table S5 for more detailed information regarding the workflow used and the list of identified signatures.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004;116:281–97. - PubMed
    1. Mendell JT, Olson EN. MicroRNAs in stress signaling and human disease. Cell 2012;148:1172–87. - PMC - PubMed
    1. Croce CM. Causes and consequences of microRNA dysregulation in cancer. Nat Rev Genet 2009;10:704–14. - PMC - PubMed
    1. Lee LW, Zhang S, Etheridge A, Ma L, Martin D, Galas D, et al. . Complexity of the microRNA repertoire revealed by next-generation sequencing. RNA 2010;16:2170–80. - PMC - PubMed
    1. Kozomara A, Griffiths-Jones S. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucl Acids Res 2014;42:D68–73. - PMC - PubMed

Publication types