Down-regulation of MutS homolog 3 by hypoxia in human colorectal cancer

Abstract

Down-regulation of hMSH3 is associated with elevated microsatellite alterations at selected tetranucleotide repeats and low levels of microsatellite instability in colorectal cancer (CRC). However, the mechanism that down-regulates hMSH3 in CRC is not known. In this study, a significant association between over-expression of glucose transporter 1, a marker for hypoxia, and down-regulation of hMSH3 in CRC tissues was observed. Therefore, we examined the effect of hypoxia on the expression of hMSH3 in human cell lines. When cells with wild type p53 (wt-p53) were exposed to hypoxia, rapid down-regulation of both hMSH2 and hMSH3 occurred. In contrast, when null or mutated p53 (null/mut-p53) cells were exposed to hypoxia, only hMSH3 was down-regulated, and at slower rate than wt-p53 cells. Using a reporter assay, we found that disruption of the two putative hypoxia response elements (HREs) located within the promoter region of the hMSH3 abrogated the suppressive effect of hypoxia on reporter activity regardless of p53 status. In an EMSA, two different forms of HIF-1α complexes that specifically bind to these HREs were detected. A larger complex containing HIF-1α predominantly bound to the HREs in hypoxic null/mut-p53 cells whereas a smaller complex predominated in wt-p53 cells. Finally, HIF-1α knockdown by siRNA significantly inhibited down-regulation of hMSH3 by hypoxia in both wt-p53 and mut-p53 cells. Taken together, our results suggest that the binding of HIF-1α complexes to HRE sites is necessary for down-regulation of hMSH3 in both wt-p53 and mut-p53 cells.

Fig. 1

The expression of GLUT1 and hMSH3 in CRC tissues. A: Microsatellite stable CRC tissue positive for nuclear hMSH3 expression. B: Microsatellite stable CRC tissue negative for cytoplasmic GLUT1 expression. C: An EMAST-positive CRC tissue negative for nuclear hMSH3 expression. D: An EMAST-positive CRC tissue over expressing cytoplasmic GLUT1. Despite the presence of non-specific MSH3 signals in the cytoplasm of CRC cells, there was a clear difference between the stained and unstained nucleus by MSH3 IHC (×200, x 630 (insets)).

Fig. 2

p53-dependent and p53-independent down-regulation of hMSH3 and hMSH2 proteins. A: hMSH3 expressed by normoxic and hypoxic mut-p53 SW480 cells. B: hMSH2 expressed by normoxic and hypoxic mut-p53 SW480 cells C: hMSH3 expressed by normoxic and hypoxic null-p53 HCT116+5 cells. D: hMSH2 expressed by normoxic and hypoxic null-p53 HCT116+5 cells. E: hMSH3 expressed by normoxic and hypoxic wt-p53 HCT116+5 cells. F: hMSH2 expressed by normoxic and hypoxic wt-p53 HCT116+5 cells. A significant reduction in hMSH3 and hMSH2 proteins was observed within 3 days after hypoxia in cells with wt-p53 but not in cells with null- or mut-p53. A significant reduction in hMSH3 but not in hMSH2 was observed in cells with null- or mut-p53 at 6 days after hypoxia.

Fig. 3

p53-independent down-regulation of hMSH3 mRNA. A: hMSH3 mRNA expressed by normoxic and hypoxic mut-p53 SW480 cells. B: hMSH3 mRNA expressed by normoxic and hypoxic null-p53 HCT116 +5 and wt-p53 HCT116+5 cells. Slow but similar levels of reduction in hMSH3 mRNA was observed after 6 days of hypoxia in null- or mut-p53 cells compared to wt-p53 cells.

Fig. 4

Characterization of the hMSH3 promoter region. A: Reporter constructs containing different regions of the hMSH3 5′ sequences and their reporter activities in SW620 or HeLa S3 cells. The activity of pGL3-promoter was defined as 1 and the relative luciferase activity was determined. A difference in reporter activity was evaluated between the pGL3-MSH3-2658-20 and the other constructs using the Student t-test. *, P<0.01; **, P<0.05. B: The effect of DPE on transcriptional activity. Reporter constructs with and without DPE sequences were compared in SW620 cells. *, P<0.01; **, P<0.05. C: Consensus sequences for TATA-Box (TATAWAAR), Int (YYANWYY), and DPE (RGWYVT) within the core promoter region of hMSH3. A nucleotide sequence from −31 to +39 of the core promoter region of hMSH3 (lower line) and the positions of each consensus sequence are shown. The transcription initiation site is expressed as 1.

Fig. 5

The role of the putative HRE in down-regulation of reporter activity by hypoxia. A: A schematic map of two putative HREs in the 5′-upstream of hMSH3. The HRE1 (GCGTG) located at −2525 to −2521 and the HRE2 (ACGTG) located at −1639 to −1635. The arrow indicates the transcription initiation site. The HRE1 sequence was replaced by GCATA and the HRE2 site was replaced by TTTTG for mutation experiments in 5 C. B: The effect of HRE deletion on reporter activity in response to hypoxia. SW620 or HCT116+5 were transfected with the reporter plasmids, pGL3-MS3-2658-20, pGL3-MSH3-2480-20 and pGL3-MS3-1267-20. After transfection, cells were cultured for another 24 hrs in normoxia (open bars) and hypoxia (closed bars). *, P<0.01. C: The effect of mutations in HRE on reporter activity in response to hypoxia. Three reporter plasmids having mutations at HRE1, at HRE2, and at both HRE1 and HRE2, designated as pGL3-MS3-HRE1M, pGL3-MS3-HRE2M, and pGL3-MS3-WM were constructed from pGL3-MS3-2658-20. SW620 was transfected with these plasmids and cultured for another 24 hrs in normoxia (open bars) and hypoxia (closed bars). *, P<0.01.

Fig. 6

Role of HIF-1α for down-regulation of hMSH3 by hypoxia. A: Stabilization of the HIF-1α protein in SW620 and HCT116+5 cells by hypoxia. Cells were cultured for 0, 1, 2, and 3 days in hypoxia and the amount of HIF-1α was monitored by Western blot analysis with anti HIF-1α antibody. α-Tubulin was used as a loading control. B: HIF-1α knockdown by siRNA. Cells were transfected with a negative control (ctr) oligomer or an HIF-1α siRNA (HIF) oligomer and cultured under hypoxia for the indicated days. The amount of HIF-1α was compared between hypoxic cells transfected with a negative control (ctr) oligomer and the HIF-1α siRNA (HIF) oligomer after normalization by α-tubulin signals. C: Effect of HIF-1α knockdown on hypoxia-induced down-regulation of hMSH3 mRNA. HCT116+5 and SW620 transfected with a negative control (ctr) oligomer (open bars) and an HIF-1α specific siRNA oligomer (closed bars) were cultured in hypoxia for the indicated number of days. The amount of hMSH3 mRNA was compared between the cells transfected with ctr and HIF siRNA. *, P<0.01; **, P<0.05. D: Effect of HIF-1α knockdown on hypoxia-induced down-regulation of hMSH3 protein. HCT116+5 and SW620 transfected with a negative control (ctr) oligomer (open bars) and an HIF-1α specific siRNA oligomer (closed bars) were cultured in hypoxia for the indicated number of days. The amount of hMSH3 was compared between hypoxic cells transfected with a negative control (ctr) oligomer and the HIF-1α siRNA (HIF) oligomer after normalization by α-tubulin signals.

Fig. 7

Binding of HIF-1α to the HRE1 site within the hMSH3 promoter region in hypoxic wt-p53 HCT116+5 and mut-p53 SW620 cells. A and D: Two binding products, “a” and “b”, are formed between the protein complexes containing HIF-1α, A and B respectively, and the radio-labeled (hot) oligonucleotide probe containing the HRE1 site and its flanking sequences (WT-HRE1) in hypoxic wt-p53 HCT116+5 (A) or mut-p53 SW620 (D). –: nuclear extracts were not added (lane 1), NE: nuclear extracts, N: normoxic cell nuclear extracts were added (lanes 2), H: hypoxic cell nuclear extracts were added (lanes 3–8). Left panel: a non-radio-labeled (cold) WT-HRE1 probe was not added (–, lanes 2, 3) or was added 25-fold excess (25, lane 4) or 50-fold excess (50, lane 5) of hot probe. Right panel: anti HIF-1α antibody was either not added (–, lane 6) or added (anti HIF-1α antibody, lane 7), or control mouse IgG (IgG, lane 8) was added to the reaction mixture. F: free probe. Bands “a” and “b” were diminished by the anti HIF-1α antibody. B and E: Inhibition of binding of A or B to hot WT-HRE1 probe by 8-fold excess of cold-WT-HRE1: (diminished bands “a” and “b” in lane 3) in hypoxic wt-p53 HCT116+5 (B) or mut-p53 SW620 (E). No inhibition of binding of A or B to hot WT-HRE1 probe by 8-fold excess of cold-MT1-HRE1 probe that contain exact the same DNA sequence as WT-HRE1 probe except HRE1 site was replaced by oligonucleotides, TTTTG, (lane 5). Inhibition of binding of A but not B to hot WT-HRE1 probe by 8-fold excess of cold-MT2-HRE1 probe containing replaced oligonucleotides, GCATA: (diminished band “a” but not “b” in lane 4). NE: nuclear extracts, H: hypoxic cell nuclear extracts, (+): added to the reaction mixture, (–): not added, F: free probe. These results indicate that the protein complexes A and B specifically bound to the HRE1 site but A has an additional binding specificity to another sequences compared to B. C and F: Inhibition of binding of A or B to hot WT-HRE1 probe by 100-fold excess of cold-HRE2: (diminished bands “a” and “b” in lane 2) in hypoxic wt-p53 HCT116+5 (C) or mut-p53 SW620 (F). No inhibition of binding of A or B to hot WT-HRE1 probe by 100-fold excess of cold-MT1-HRE2 probe that contain exact same sequence as WT-HRE2 probe except HRE2 site was replaced by oligonucleotides, TTTTG, (lane 3). NE: nuclear extracts, H: hypoxic cell nuclear extracts, (–): not added, F: free probe. The results indicate that A and B bound to the HRE2 sites of WT-HRE2 probe.

Fig. 8

Bindings of protein complexes containing HIF-1α to the HRE1 and HRE2 sites in hypoxic null-p53 HCT116 +5. A: A hot-WT-HRE1 probe was mixed with nuclear extracts (NE) from normoxic cells (N) or hypoxic cells (H) and subjected to electrophoresis. Five species of protein complexes were induced by hypoxia (lane 2) compared to normoxic cell (lane 1). Among them bands “a” and “b” were diminished by anti HIF-1α antibody (lane 4) compared to control, nothing added (–) (lane 3) or mouse IgG added (lane 5). F: free probe. Compared to wt-p53 HCT116+5, the intensity of band “a” becomes stronger, indicating that the binding of protein complex A to the HRE1 site becomes dominant over complex B where the case is the same in hypoxic mut-p53 SW620. B: A hot-WT-HRE2 probe was mixed with nuclear extracts (NE) from normoxic cells (N) or hypoxic cells (H) and subjected to electrophoresis. Five species of protein complexes were induced by hypoxia (lane 2) compared to normoxic cell (lane 1). Among them bands “c” and “d” were diminished by anti HIF-1α antibody (lane 4) compared to control, nothing added (–) (lane 3) or mouse IgG added (lane 5). F: free probe. Compared to wt-p53 HCT116+5 (Supplementary Figs. 4A and 4B), the intensity of band “c” becomes stronger, indicating that the binding of protein complex A to the HRE2 site becomes dominant over complex B where the case is the same hypoxic mut-p53 SW620 (Supplementary Figs. 4C and 4D).