A pleiotropically acting microRNA, miR-31, inhibits breast cancer metastasis

Abstract

MicroRNAs are well suited to regulate tumor metastasis because of their capacity to coordinately repress numerous target genes, thereby potentially enabling their intervention at multiple steps of the invasion-metastasis cascade. We identify a microRNA exemplifying these attributes, miR-31, whose expression correlates inversely with metastasis in human breast cancer patients. Overexpression of miR-31 in otherwise-aggressive breast tumor cells suppresses metastasis. We deploy a stable microRNA sponge strategy to inhibit miR-31 in vivo; this allows otherwise-nonaggressive breast cancer cells to metastasize. These phenotypes do not involve confounding influences on primary tumor development and are specifically attributable to miR-31-mediated inhibition of several steps of metastasis, including local invasion, extravasation or initial survival at a distant site, and metastatic colonization. Such pleiotropy is achieved via coordinate repression of a cohort of metastasis-promoting genes, including RhoA. Indeed, RhoA re-expression partially reverses miR-31-imposed metastasis suppression. These findings indicate that miR-31 uses multiple mechanisms to oppose metastasis.

Figure 1. miR-31 Levels Correlate Inversely With Metastatic Ability in Breast Cell Lines

(A) RT-PCR for miR-31 in seven human breast cell lines. 5S rRNA was a loading control. NTC: no template control. n = 3. (B) miR-31 RT-PCR in eight murine mammary cell lines. 5S rRNA was a loading control. n = 3. (C) In situ hybridization for miR-31 (green) in animal-matched 4T1 cell primary mammary tumors and lung metastases; DAPI counterstain (blue). n = 4. (D) Hematoxylin and eosin (H&E) stain of a 4T1 cell primary mammary tumor (top); box: invasive front. miR-31 in situ hybridization in 4T1 cells located near the invasive front or the interior of the primary tumors (bottom). n = 3.

Figure 2. miR-31 Expression Inhibits Metastasis

(A) Invasion and motility assays after transfection of MDA-MB-231 (231) cells with the indicated constructs. n = 3. (B) Anoikis assays using 231 cells infected as indicated. n = 3. (C) Primary tumor growth upon orthotopic injection of 1.0 x 10⁶ GFP-labeled 231 cells infected as indicated. The experiment was terminated after 13 weeks due to primary tumor burden. n = 5 per group per timepoint. (D) H&E stain of 231 primary tumors 62 days after orthotopic injection. (E) H&E stain of tissue adjacent to the indicated 231 primary mammary tumors 62 days after injection. Arrows: disseminated tumor cells in normal fat (a, b), muscle (c, d), and subcutis (e, f). (F) Images of murine lungs to visualize GFP-labeled 231 cells 62 days after orthotopic implantation (left). H&E stain of lungs from animals bearing the indicated tumors (right); arrows: metastatic foci. n = 5. (G) Images of murine lungs to detect GFP-labeled 231 cells 88 days after tail vein injection (left). H&E stain of lungs (right); arrows: metastatic foci. Asterisks: P >0.66. n = 5, except for 10 min and two hrs (n = 4).

Figure 3. Inhibition of miR-31 Promotes Metastasis

(A) Invasion and motility assays using MCF7-Ras cells transfected with the indicated transient miR-31 inhibitors. n = 3. (B) Anoikis assays with MCF7-Ras cells stably expressing the indicated sponge. n = 3. (C) Primary tumor growth upon orthotopic implantation of 5.0 x 10⁵ GFP-labeled MCF7-Ras cells infected as indicated. The experiment was terminated after 16 weeks due to primary tumor burden. n = 5 per group per timepoint. (D) H&E stain of MCF7-Ras primary tumors 47 days after orthotopic injection. Arrows: regions of poor encapsulation. (E) H&E stain of tissue adjacent to the indicated MCF7-Ras primary tumors 47 days post-injection. Arrows: disseminated tumor cells in normal fat (a, c) and muscle (b, d). (F) Images of murine lungs to visualize GFP-labeled MCF7-Ras cells 113 days after orthotopic injection (left). H&E stain of lungs from animals bearing the indicated tumors (middle); arrows: metastatic foci. n = 5. (G) Images of murine lungs to detect GFP-labeled MCF7-Ras cells 122 days after tail vein injection (left). H&E stain of lungs (middle); arrow: metastasis. n = 4, except for one day (n = 3).

Figure 4. miR-31 Directly Regulates a Cohort of Pro-Metastatic Genes

(A) Luciferase activity in 231 cells infected with miR-31 or control vector after transfection of the indicated 3’ UTR-driven reporter constructs. n = 3. (B) Luciferase activity in the indicated 231 cells upon transfection of miR-31 site mutant 3’ UTR-driven reporter constructs. wt: wild type; site 1: the miR-31 motif at nt 145-151 of the RhoA 3’ UTR; site 2: the motif spanning nt 303-309. Asterisks: P >0.80 relative to mutant-UTR + vector controls. n = 3. (C) Immunoblots for endogenous Fzd3, ITGA5, MMP16, RDX, and RhoA in the indicated 231 cells. β-actin was a loading control. Repression: protein levels in miR-31-expressing cells relative to vector controls. (D) RT-PCR for endogenous CXCL12, Fzd3, ITGA5, M-RIP, MMP16, RDX, and RhoA. GAPDH was a loading control. Asterisks: P <0.03 relative to vector controls. n = 3.

Figure 5. Repression of Fzd3, ITGA5, RDX, and RhoA Underlies miR-31-Dependent Phenotypes in vitro

(A) Kaplan-Meier curves for 295 human primary breast tumors depicting metastasis-free survival, stratified based on expression of the six-gene miR-31 target signature. P-value based on a logrank test. (B) Kaplan-Meier five-year survival curves for 295 breast cancer patients, stratified based on miR-31 target signature expression in their primary tumors. P-value based on a logrank test. (C) Invasion assays with miR-31-expressing or control 231 cells transfected as indicated. Asterisks: P >0.19 relative to vector + siControl cells. n = 3. (D) Anoikis assays using 231 cells transfected with the indicated siRNAs. Asterisks: P >0.80 relative to vector + siControl cells. n = 3. (E) Invasion assays using the indicated 231 cells transfected with miRNA-resistant expression constructs. Asterisks: P >0.61 relative to miR-31 + mock cells. n = 3. (F) Anoikis assays with the indicated 231 cells transfected as noted. Asterisks: P >0.11 relative to miR-31 + mock cells. n = 3.

Figure 6. Re-Expression of RhoA Partially Reverses miR-31-Imposed Metastasis Defects in vivo

(A) Primary tumor growth upon orthotopic injection of 5.0 x 10⁵ GFP-labeled 231 cells. The experiment was terminated after 11 weeks due to primary tumor burden. Asterisks: P <0.02. n = 5 per group per timepoint. (B) H&E stain of 231 primary tumors 60 days after orthotopic injection. (C) H&E stain of tissue adjacent to the indicated 231 primary mammary tumors 60 days after injection. Arrows: disseminated tumor cells in normal muscle (a, c, e, g) and fat (b, d, f, h). (D) Images of murine lungs to visualize GFP-labeled 231 cells 60 days after orthotopic injection (left). H&E stain of lungs from animals bearing the indicated tumors (right); arrows: metastatic foci. n = 5. (E) Images of murine lungs to detect GFP-labeled 231 cells 86 days after tail vein injection (left); arrows: micrometastatic lesions. Asterisks: P >0.87 relative to vector + vector controls. n = 4, except for 2 weeks (n = 3).

Figure 7. miR-31 Levels Correlate Inversely With Metastasis in Human Breast Tumors

(A) miR-31 RT-PCR in 54 primary breast tumors. Normal: tissue from non-diseased individuals; metastasis-positive and -free: tumors of the indicated distant metastasis outcome. 5S rRNA was a loading control. n = 4 (normal); n = 14 (metastasis-positive); n = 40 (metastasis-free). (B) Kaplan-Meier distant metastasis-free survival curves for 54 breast cancer patients, stratified based on miR-31 levels in their primary tumors. P-value based on a chi-square test. (C) In situ hybridization for miR-31 (green) in patient-matched primary breast tumors and distant metastases (patient 1 = lung; 2 = pleura); DAPI counterstain (blue). n = 8 fields. (D) Immunohistochemical detection of ITGA5, RDX, and RhoA in patient-matched primary breast tumors and distant metastases (patient 1 = lung; 2 = pleura). n = 8 fields.