Overexpression of MiR-155 Results in Enhanced Mutation Rate.

MiR-155 targets core components of the DNA mismatch repair (MMR) machinery, among other mutator pathways (15), suggesting that elevated levels of miR-155 might enhance the rate of spontaneous mutations. To measure the mutation rate, we took advantage of the hypoxanthine phosphoribosyltransferase (HPRT) locus, which is a well-established method for estimating mutation rate. The HPRT enzyme catalyzes the conversion of guanine into guanine monophosphate and hypoxanthine into inosine monophosphate in the purine salvage pathway (16, 17). The loss of HPRT function confers resistance to 6-thioguanine (6-TG), because 6-TG becomes cytotoxic only after phosphoribosylation by HPRT. This resistance can be used to identify cells that have acquired mutations at the HPRT locus (16, 17). Because the acquired mutations are thought to occur randomly, the HPRT gene can be used as a reporter gene, and the frequency of mutation at the HPRT locus can be used as an estimate of global genomic instability. To measure the effects of miR-155 on mutation rate, we first developed stable clones of SW620 colorectal adenocarcinoma cells and MDA-MB-231 breast adenocarcinoma cells expressing mature miR-155 under the control of the Tet-On inducible system. Incubation of SW620 clones 8A, 22C, and 23A with doxycycline increased miR-155 expression by 2.94 ± 0.23-, 5.58 ± 0.43-, and 8.10 ± 0.65-fold, respectively (mean ± SD) (Fig. S1 and Table S1). Similarly, doxycycline treatment increased miR-155 expression by 12.01 ± 2.34-, 26.42 ± 1.11-, and 32.07 ± 3.27-fold in MDA-MB-231 clones 9C, 2B, and 19B, respectively.

The cell growth-adjusted HPRT mutation rate, estimated based on a modified version of fluctuation analysis (18), increased with miR-155 levels in both SW620 and MDA-MB-231 cell clones (Fig. 1 A and B). Constant elevated expression of miR-155 the enhanced mutation rate by up to 3.39-fold in SW620 clones and up to 3.47-fold in MDA-MB-231 clones (Table S1). Moreover, HCT116 colorectal carcinoma cells transiently overexpressing miR-155 by 19.57 ± 0.62-fold under the control of Tet-On inducible system (Fig. S2) showed a 2.81-fold higher mutation rate (Table S1). These results establish a direct link between mutation rate and miR-155 levels.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

The average mutation rate in SW620 and MBA-MD-231 cells increases with the rate of miR-155 expression. (A and B) Average mutation rates of 6-TG–resistant colonies from SW620 (A) and MDA-MB-231 clones (B) stably expressing miR-155 and either mock-treated or treated with doxycycline, plotted against the fold upregulation of miR-155 values. Fold upregulation 1 value corresponds to mock-treated controls.

As expected, the basal spontaneous cell growth-adjusted mutation rates of SW620 and MDA-MB-231 cells was comparable (0.75 ± 0.27 × 10⁻⁷ and 1.28 × 10⁻⁷ mutations per cell, respectively) and were ~440-fold lower than the spontaneous mutation rate of HCT116 cells (560 × 10⁻⁷ mutations per cell) (Table S1), because HCT116 cells contain a deletion of the hMLH1 (human MutL homolog 1) MMR gene (19). The deletion of the hMLH1 MMR gene decreased the miR-155-induced mutator activity only partially, suggesting that miR-155–induced mutator activity is not very sensitive to the basal level of mutation rate (i.e., to the integrity of the DNA safeguarding machinery) and that miR-155 targets additional transcripts implicated in DNA repair and/or genome stability.

Inflammatory Stimuli Up-Regulate MiR-155 in Breast Cancer Cells.

Infection-driven inflammation has been associated with cancer (20). A number of studies have shown that miR-155 is a key component of the inflammatory response (4–7). We thus screened human cell lines for the effects of proinflammatory environment on miR-155 expression. We included colon cancer cell lines (SW480, SW620, HCT15, HCT116, and RKO), because a fraction of colorectal cancers appear linked to the inflammatory environment (20), breast (MDA-MB-231, T47D, 453, 436, and MCF7) and lung cancer (A459) cell lines, because miR-155 is up-regulated in these types of cancers (12), and four other cell lines for comparison. Cells were treated overnight with the supernatant of LPS-stimulated human THP-1 monocytic cells, namely LPS-stimulated macrophage-conditioned medium (LSMCM), which contains many inflammatory cytokines such as TNF, IL-6, IL-8, and IL1-β (1, 3). Based on quantitative RT-PCR (qRT-PCR) analyses, miR-155 expression in colon and lung cancer cell lines was affected only slightly by LSMCM (Fig. 2A). In contrast, miR-155 levels increased by 9-, 17-, and 21-fold in MDA-MB-231, BC-453, and T47D breast cancer cell lines, respectively. We also analyzed the expression of miR-146a, a microRNA that is up-regulated in certain tumors (10, 11) and in LPS-challenged THP-1 cells. It has been proposed that miR-146a controls the termination of the immune response (21). In sharp contrast with miR-155, the highest miR-146a levels were found in HCT15 and HCT116 colon cell lines (not shown), indicating that the up-regulation of miR-155 and miR-146a by inflammatory stimuli occurs independently in the above cancer cell lines and probably is tissue specific. Because both TNF (22) and LPS can induce miR-155 (6, 7), we analyzed the effects of these two molecules in MDA-MB-231 cells. As expected, stimulation with either TNF or LPS or with both increased miR-155 expression (Fig. 2B). We therefore used TNF/LPS to mimic the effects of a proinflammatory environment.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Proinflammatory environment increases the mutation rates in MDA-MB-231 cells and the frequency of mutant colonies in T47D cells. (A) The relative up-regulation of miR-155 in the indicated cell lines either mock treated or treated with LSMCM for 48 h were determined using qRT-PCR. The figure gives the ratios LPS-stimulated/unstimulated. Values represent mean ± SD (n = 3). (B) The relative up-regulation of miR-155 in MDA-MB-231 cells treated as indicated was determined using qRT-PCR. (C) The mutation rates (MR) in MDA-MB-231 cells were estimated by calculating the slopes of the curves following mock and TNF/LPS treatment (n = 4 estimations of mutation frequency). (D) Ratios of 6-TG–resistant colonies in T47D cells treated with LSMCM for 48 h versus control mock-treated cells.

Inflammatory Stimuli Enhance the Mutation Rate.

In MDA-MB-231 cells without stimulation, the mutation rate was calculated to be 0.69 × 10⁻⁷ mutations per cell per generation, a value similar to that previously found in SW480 cells (0.75 × 10⁻⁷) (23). The estimated mutation rate was based on the average mutant frequency and population doubling (23, 24). Of note, both MDA-MB-231 and SW480 cells have intact DNA-repair machinery, unlike HCT116 cells. Mutant frequencies already had became significantly different (P = 0.038) 3 d after treatment, with mutation rate increasing by 2.52-fold to 1.73 × 10⁻⁷ mutations per cell per generation (Fig. 2C) in TNF/LPS-treated MDA-MB-231 cells vs. untreated control cells. The treatment lowered the rate of cell proliferation, probably because TNF induces growth arrest in breast cancer cells (25). Accordingly, one or two LSMCM stimulations of T47D cells increased the frequency of 6-TG–resistant colonies by 50% and 150%, respectively (Fig. 2D). Thus, proinflammatory signals resulting in the up-regulation of miR-155 expression induce a significant, although moderate, mutator phenotype, which might be enhanced with chronic inflammation.

Characterization of HPRT Mutants.

We then analyzed HPRT mutations found in cDNAs prepared from RNA extracted from T47D, HCT116, and MDA-MB-231 6-TG–resistant colonies to determine the mutation signature (Table 1 and Table S2). As expected, HPRT mutations from doxycycline-treated HCT116 cells displayed single base deletions or insertions of the type generally found in DNA MMR-deficient cells. The majority displayed a frameshift, transition, and transversion mutation signature consistent with an MMR defect. There also was an increase in insertions and exon deletions with miR-155 overexpression; this increase generally has been ascribed to altered recombination repair (26). In contrast, deletion mutations consistent with recombination repair defects accounted for the majority of HPRT mutations in LSMCM-stimulated T47D cells and doxycycline-treated MDA-MB-231 cells, regardless of conditions (Table 1 and Table S2). These results are consistent with a role for recombination repair, exemplified by breast cancer 1 (BRCA1) and breast cancer 2 (BRCA2) mutations, in these breast tumor cell lines (27). Of note, these types of mutations have been found previously in several T-cell leukemic cell lines (16, 28). However, we noted a modest increase in transitions and transversions consistent with decreased MMR in these cells.

Overexpression of miR-155 Enhances Cell Proliferation.

Remarkably, miR-155 up-regulation increased the size of HCT116 (Fig. 3A) and MDA-MB-231 HPRT mutant colonies and allowed them to appear earlier during the selection process. Based on a forward scatter comparison, larger colonies of MDA-MB-231 clone 19B, that presented a 32-fold up-regulation of miR-155 after doxycycline treatment, did not arise from the presence of larger cells (not shown). In contrast, carboxyfluorescein succinimidyl ester (CFSE) staining suggested that these cells underwent at least one extra round of cell division within 4–5 d as compared with untreated cells (Fig. 3B). These results correlate with reports showing that miR-155 promotes proliferation in transgenic mice (13, 14). Thus, the larger size of HPRT mutant colonies overexpressing miR-155 probably arises from enhanced cell proliferation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Elevated miR-155 levels increase the rate of proliferation by targeting WEE1 transcripts. (A) Phenotypes of 6-TG–resistant HCT116 colonies following 48-h mock treatment (− doxycycline) or 48-h doxycycline treatment (+ doxycycline). HCT116 cells were transiently infected with pRetroX-Tight-Pur-miR-155 and Tet-On constructs before doxycycline treatment. (B) Cells from MDA-MB-231 clone 19B were stained with CFSE before induction of miR-155 expression by doxycycline treatment. The proliferation rate was analyzed by flow cytometry 4 and 5 d later. The experiment was repeated two times with similar results. (C) The levels of WEE1 in T47D cells transfected with premiR-Control, premiR-155, or antisense miR-155 inhibitory RNA (155-I) and subsequently either mock treated or treated with LSMCM for 48 h were determined by Western blotting. (D) The levels of Wee1 in primary B cells isolated from the spleen of Eμ-miR-155 transgenic mice after transfection with premiR-Control or antisense miR-155 inhibitory RNA (155-I).

MiR-155 and Inflammatory Environment Down-Regulate WEE1, a Cell-Cycle Inhibitor.

MiR-155 might enhance cell proliferation by targeting cell-cycle regulators. Indeed, in T47D cells, both LSMCM treatment and miR-155 overexpression reduced the levels of WEE1, a kinase that catalyzes the inhibitory tyrosine phosphorylation of Cdc2/cyclin B, blocking cell-cycle progression at the G₂/M phase (Fig. 3C) (29). In contrast, an antisense miR-155 inhibitory RNA (155-I) increased WEE1 accumulation. Both LSMCM and miR-155 overexpression also reduced the expression of a luciferase reporter construct containing the WEE1 3′ UTR (Fig. S3). Accordingly, 155-I increased Wee1 levels in primary B cells isolated from Eμ-miR-155 transgenic mice that overexpress miR-155 in B-cell lineage (13), thus confirming that Wee1 is a bona fide miR-155 target (Fig. 3D). Taken together, these results show that inflammatory stimuli down-regulate WEE1 through up-regulation of miR-155. Because WEE1 depletion rapidly induces DNA damage in newly replicated DNA (30), these results suggest that miR-155 overexpression may shorten the period required for selection of cancer-associated mutations. Furthermore, Affymetrix microarrays revealed that transcripts coding for several factors controlling cell cycle, DNA repair, and genome stability were affected by LSMCM in both T47D and MDA-MB-231 cell lines (Table 2). This result suggests that the ability of inflammatory stimuli to induce defective checkpoints and genomic instability, similar to miR-155, might contribute to tumorigenesis.