Decreased expression of the DNA mismatch repair gene Mlh1 under hypoxic stress in mammalian cells

Abstract

The hypoxic tumor microenvironment has been shown to contribute to genetic instability. As one possible mechanism for this effect, we report that expression of the DNA mismatch repair (MMR) gene Mlh1 is specifically reduced in mammalian cells under hypoxia, whereas expression of other MMR genes, including Msh2, Msh6, and Pms2, is not altered at the mRNA level. However, levels of the PMS2 protein are reduced, consistent with destabilization of PMS2 in the absence of its heterodimer partner, MLH1. The hypoxia-induced reduction in Mlh1 mRNA was prevented by the histone deacetylase inhibitor trichostatin A, suggesting that hypoxia causes decreased Mlh1 transcription via histone deacetylation. In addition, treatment of cells with the iron chelator desferrioxamine also reduced MLH1 and PMS2 levels, in keeping with low oxygen tension being the stress signal that provokes the altered MMR gene expression. Functional MMR deficiency under hypoxia was detected as induced instability of a (CA)(29) dinucleotide repeat and by increased mutagenesis in a chromosomal reporter gene. These results identify a potential new pathway of genetic instability in cancer: hypoxia-induced reduction in the expression of key MMR proteins. In addition, this stress-induced genetic instability may represent a conceptual parallel to the pathway of stationary-phase mutagenesis seen in bacteria.

FIG. 1.

Decreased levels of MLH1 and PMS2 in mouse and human cells exposed to hypoxia. Western blot analyses were performed to determine the expression of the MMR proteins MLH1, PMS2, MSH2, and MSH6 in cells under normoxic (lanes N) or hypoxic (lanes H) conditions, as indicated. The period of time that the cells were maintained under hypoxic conditions (24 or 48 h) is given. Tubulin expression is presented to confirm equal sample loading. (A) Mouse 3340 cells. (B) HeLa cells. Expression of the hypoxia-inducible factor HIF-1α is also shown for comparison to verify that physiologically relevant levels of hypoxia were present in the treated cells.

FIG. 2.

Decreased levels of Mlh1 mRNA in mouse and human cells exposed to hypoxia. Northern blot analyses were performed on RNA samples obtained from cells grown under normoxic (lanes N) and hypoxic (lanes H) conditions for 24 h. (A) Mlh1, Pms2, Msh2, and Msh6 mRNA levels in mouse 3340 cells. To confirm equal sample loading, mRNA levels for transketolase (Tkt) were also determined. (B) Mlh1 mRNA levels in human HeLa cells under normoxic or hypoxic conditions. (C) Mlh1 and Pms2 mRNA levels in mouse EMT6 cells under normoxic or hypoxic conditions. Equal sample loading was verified in the cases of the HeLa and EMT6 cell blots by analysis of ethidium bromide-stained gels prior to transfer (not shown). Expression levels were quantified by phosphorimager analysis, and the ratio of expression under hypoxia to that under normoxia is listed to the right of each panel.

FIG. 3.

Decreased MMR gene expression in mouse and human cells exposed to the hypoxia-mimetic drug DFX. (A) Western blot analyses of the expression of the MMR proteins MLH1, PMS2, MSH2, and MSH6 in human HeLa, mouse 3340, and mouse EMT6 cells treated with DFX. Tubulin expression is presented to confirm equal sample loading. Expression of the hypoxia-inducible factor HIF-1α in the HeLa cells is also shown for comparison to demonstrate that the DFX treatment provoked a hypoxia-like response in the cells. (B) Northern blot analyses of Mlh1 and Pms2 mRNA levels in mouse 3340 cells with and without DFX treatment. (C) Northern blot analysis of Mlh1 and Pms2 mRNA levels in mouse EMT6 cells with and without DFX treatment. (D) Northern blot analysis of Mlh1 mRNA levels in HeLa cells with and without DFX treatment.

FIG. 4.

The histone deacetylase inhibitor TSA prevents the down-regulation of Mlh1 expression by either hypoxia or treatment with the hypoxia mimetic drug DFX. (A) Northern blot analyses of Mlh1 expression levels were carried out on RNA samples from 3340, EMT6, and HeLa cells incubated under the following conditions as indicated: normoxia, hypoxia for 24 h, hypoxia plus the cytosine methylation inhibitor azaC, hypoxia plus TSA, or hypoxia plus both azaC and TSA. Equal sample loading was verified by analysis of ethidium bromide-stained gels prior to transfer (not shown). (B) Northern blot analyses of Mlh1 expression levels in 3340, EMT6, and HeLa cells under normoxic or hypoxic conditions and with or without TSA, as indicated. (C) Northern blot analyses of Mlh1 expression levels in 3340, EMT6, and HeLa cells with and without DFX treatment and with or without TSA, as indicated.

FIG. 5.

Reversibility of the hypoxia-induced down-regulation of Mlh1 after replacement of cells in normoxic conditions. Northern blot analysis of Mlh1 mRNA levels in 3340 mouse cells grown under normoxia (lane N), grown under hypoxia for 48 h (lane H), or grown under hypoxia for 48 h and then returned to normoxia for the indicated times is shown.

FIG. 6.

Induction of mutagenesis by hypoxia and association with MLH1 levels. (A) Frequencies of mutations in the chromosomal supFG1 reporter gene in 3340 cells exposed to hypoxia. Cells were maintained under normoxic or hypoxic conditions for 24 h. After 3 days of subsequent growth of the cells under normal culture conditions, mutagenesis in the supFG1 reporter gene was assayed by rescue of the chromosomally integrated λsupFG1 shuttle vector. Error bars indicate standard errors. (B) Restoration of β-galactosidase (beta-gal) activity via frameshift mutagenesis in a lacZ reporter gene construct in HeLa and EMT6 cells under normoxia or hypoxia. Cells were transfected with an episomal, replicative vector, pCAR-OF (33), containing the β-galactosidase gene interrupted by a 58-bp out-of-frame poly (CA)₂₉ insertion tract at the 5′ end of its coding region. Restoration of the proper reading frame to generate a functional enzyme occurs when replication slippage errors within the repeat sequence tract are not corrected by MMR. This experiment was performed three times, and the relative β-galactosidase values were normalized to a value of 1 for the normoxic cells in each case. Standard errors are indicated.

FIG. 7.

Model illustrating the pathway of hypoxia-induced genetic instability via down-regulation of the MMR gene Mlh1. The present data support a mechanism in which hypoxia causes histone deacetylation, leading to repression of Mlh1 transcription. Decreased MLH1 protein secondarily leads to destabilization of PMS2. In the setting of reduced levels of the MLH1/PMS2 heterodimer, cellular MMR activity is suboptimal, resulting in increased genetic instability.