DUSP6 regulates Notch1 signalling in colorectal cancer

Abstract

Notch1 plays various roles in cancer development, and Notch1-induced transactivation is controlled by phosphorylation of its cleaved intracellular domain. However, it is unclear whether there are phosphatases capable of dephosphorylating the cleaved Notch1 transmembrane/intracellular region (NTM) to regulate its function. Here, we show that DUSP6 can function as a phosphatase for Notch1, thereby regulating NTM stability and transcriptional activity, thus influencing colorectal cancer (CRC) development. In human CRC cells, elevated DUSP6 expression correlates with increased NTM levels, leading to enhanced CRC cell proliferation both in vitro and in vivo. High tumoral DUSP6 protein expression is associated with poorer overall CRC patient survival. In mice, DUSP6 deficiency results in reduced CRC development. Mechanistically, DUSP6 dephosphorylates phospho-Y2116, which in turn reduces NTM ubiquitination, leading to increased NTM stability and transcriptional activity. As a result, the expression of Notch1-targeted proliferation genes is increased to promote tumour cell growth.

Fig. 1. DUSP6 positively regulates cellular proliferation and migration in CRC cells in vitro.

a, b Bar charts show the effect of DUSP6 expression on the rate of cell growth in Caco2 and DLD1 CRC cells with or without EGF treatment (n = 6 biologically independent samples). c, d Representative micrographs showing migration of Caco2 or DLD1 cells in wound healing assay. Bar charts under each micrograph panel show the percentage of wound closure/migration of cells compared to Time 0. Arrows denote the distance of each “wound” at the start of the experiment (Time = 0 h) and at the endpoint (Time = 24 h). e Western blot analysis of DUSP6 in vector and DUSP6 KO DLD1 cells transfected with a full-length cDNA of DUSP6. f Cell proliferation of DLD1 KO cells with or without DUSP6 overexpression was determined on Day 3 with or without EGF treatment. *, **, and *** denote P < 0.05, P < 0.01, and P < 0.001, respectively (two-tailed, nonparametric, Mann–Whitney test). Error bars = mean ± standard deviations. c, d and f (n = 8 biologically independent samples). Source data are provided as a Source Data file.

Fig. 2. DUSP6 negatively regulates the activation of MAPKs and positively regulates the level of Notch1 NTM.

a Western blot analysis showing changes in the levels of phosphorylated MAPKs (ERK1/2, P38, JNK) and the expression of target genes upon treatment with 100 ng/mL EGF in Caco2 cells overexpressing DUSP6 (DUSP6 OE) and DLD1 cells with DUSP6 knockout (DUSP6 KO) compared to the respective ppy-vector and sg-vector control cells. b Western blots showing the changes in the levels of NTM and components of γ-secretase after treatment with 100 ng/mL EGF. (Data shown in (a) and (b) is representative of three independent experiments). c Assessment of γ-secretase activity in Caco2 and DLD1 cells after treatment with 100 ng/mL EGF for 3 h. (n = 5 wells, representative of three independent experiments). d–f Effects of ERK inhibition on DLD1 DUSP6 KO cells. d Western blot analysis showing changes in the levels of NTM and ERK1/2 phosphorylation in DLD1 KO cells with or without ERK inhibition. The blots were analysed by densitometric analysis before the level of NTM relative to ACTIN is calculated and presented in the bar chart. The data shown is representative of two independent experiments. e Assessment of γ-secretase activity and (f) cell proliferation in vector and DUSP6 KO DLD1 cells with or without ERK inhibition (n = 4 biologically independent samples). * denotes P < 0.05 (two-tailed, nonparametric, Mann–Whitney test). Error bars = mean ± standard deviations. “OE” and “KO” denote overexpression and knockout, respectively. NTM cleaved Notch1 transmembrane/intracellular domain. Source data are provided as a Source Data file.

Fig. 3. DUSP6 expression is positively associated with CRC cell growth in vivo.

a Representative photographs of Caco2 xenograft tumours and bar chart showing the weight of Caco2 xenograft tumours with or without DUSP6 overexpression (n = 8 biologically independent samples). b, c Representative micrographs and bar charts showing the level of Ki67 and cleaved Notch1 staining in Caco2 xenografts, respectively (n = 8 biologically independent samples). d Representative photographs of DLD1 xenograft tumours and a bar chart show the weight of DLD1 xenograft tumours (n = 9 biologically independent samples). e, f Representative micrographs and bar charts showing level of Ki67 and cleaved Notch1 staining in DLD1 xenografts respectively (n = 9 biologically independent samples). Scale bar = 100 µm. Statistical analysis = two-tailed, nonparametric Mann–Whitney test. Error bars = mean ± standard deviations. “OE” and “KO” denote overexpression and knockout, respectively. The data shown is representative of two independent experiments. Source data are provided as a Source Data file.

Fig. 4. Induced expression of DUSP6 promotes CRC growth and loss of DUSP6 in mice resulted in reduced colonic tumour development.

a Caco2 cells transfected with Tet-on vector and full-length DUSP6 cDNA were subcutaneously injected into NSGS mice (5 × 10⁶ cells per mice). Mice (8–10 weeks old) were treated with doxycycline (Dox) (1 mg/kg) or saline twice a week for 4 weeks before assessment of tumour growth. The bar chart shows the weight of Caco2 xenograft tumours with or without Dox treatment (n = 4 biologically independent samples). b Representative micrographs of immunohistochemical staining and (c) bar charts showing the level of DUSP6 and cleaved Notch1 staining in the xenograft tumour tissues (n = 4 biologically independent samples). d Representative photograph of colonic tumours found in wildtype (WT) and DUSP6 knockout (KO) mice given AOM/DSS for induced tumour development. Bar charts show the total number of colonic tumours and the average size of tumour found in each mouse (WT n = 14, KO n = 11). e Representative micrographs showing Ki67 staining in the tumours from WT and DUSP6 KO mice (8–10 weeks old). Boxed areas are shown magnified in the bottom panels. The percentage of Ki67 positive cells per tumour is shown in the bar chart (WT n = 14, KO n = 11). f Representative micrographs showing cleaved Notch1 staining in the tumours from WT and DUSP6 KO mice (8–10 weeks old). Boxed areas are shown magnified in the bottom panels. The percentage of cleaved Notch1 positive cells per tumour is shown in the bar chart (WT n = 14, KO n = 11). Scale bar = 100 µm. Statistical analysis = two-tailed, nonparametric Mann–Whitney test. Error bars = mean ± standard deviations. Source data are provided as a Source Data file.

Fig. 5. DUSP6 binds and dephosphorylates cleaved Notch1, thereby regulating its ubiquitination and proteasomal degradation.

a Western blots showing endogenous cleaved Notch1 immunoprecipitated with Flag-DUSP6 in DLD1 cells. b Western blots showing endogenous DUSP6 immunoprecipitated with HA-NTM in DLD1 cells. c Western blots show dephosphorylation of NTM purified from DLD1 cells using immunoprecipitation. The blots were analysed by densitometric analysis before the level of total p-tyrosine on NTM is calculated and presented in the bar chart. d Western blots show results from sequential dephosphorylation of NTM and ubiquitination of the dephosphorylated NTM in vitro. NTM was purified from DLD1 cells by immunoprecipitation. Blots were analysed by densitometric analysis before the level of total p-tyrosine on NTM and total ubiquitinated NTM are calculated and presented in the bar charts. e Western blots show the changes in NTM level in DLD1 DUSP6 KO and sg-vector control cells after treatment with MG132 for 10 and 15 h to inhibit proteasomal degradation. Blots were analysed by densitometric analysis before the changes in NTM level are calculated and presented in the bar charts. f Western blots show the level of endogenous ubiquitinated NTM from the respective Caco2 and DLD1 cells. Blots were analysed by densitometric analysis before the ratio of ubiquitinated NTM are calculated and shown in the bar charts. Data shown in (a–c), (e, f) are representative of three independent experiments, (d) is representative of two independent experiments. Source data are provided as a Source Data file.

Fig. 6. Regulation of tumour cell proliferation by DUSP6 is Notch1 dependent.

a Bar chart shows the effect of DUSP6 expression and loss of NOTCH1 gene on the rate of Caco2 cell growth (n = 8 biologically independent samples). b Western blot analysis showing changes in protein expression of NTM and target genes after treatment of Caco2 cells with 100 ng/mL EGF for 3 h. Blots were analysed by densitometric analysis before the ratio of each target protein to ACTIN are calculated and shown in the bar charts. “OE”, “KO” and “Ctl” denote overexpression, knockout and control, respectively. The data shown is representative of two independent experiments. c Bar chart shows the level of cleaved Notch1-mediated Notch1 pathway activity using Notch1 reporter assay (n = 14 biologically independent samples). d Bar chart shows the growth of DUSP6 control and KO DLD1 cells with or without transfection of NTM WT or NTM Y2116A mutant (n = 8 biologically independent samples). ** and “ns” denote P < 0.01 and “not significant”, respectively (ANOVA Kruskal–Wallis, Dunn’s multiple comparisons test). Error bars = mean ± standard deviations. Source data are provided as a Source Data file.

Fig. 7. High DUSP6 expression in CRC tumours is associated with higher cleaved Notch1 levels and worse overall survival in CRC patients.

a Representative micrographs showing DUSP6 staining in tumours and normal adjacent tissues from CRC patients. Scale bar = 100 μM. Dot-plot shows the comparisons of the level of DUSP6 staining in normal and tumour tissues, Statistical test = two-tailed, nonparametric, Mann–Whitney test. Error bars = mean ± standard deviations. b Kaplan–Meier survival curve and log-rank (Mantel–Cox) test showing overall survival of CRC patients stratified by high or low DUSP6 expression in tumours cells. c Representative micrographs showing cleaved Notch1 staining in tumours and normal adjacent tissues from CRC patients. Scale bar = 100 μM. Dot-plot shows the comparisons of the level of cleaved Notch1 staining in normal and tumour tissues. Statistical test = two-tailed, nonparametric, Mann–Whitney test. Error bars = mean ± standard deviations. d Dot-plot shows the correlation between the levels of DUSP6 and cleaved Notch1 in tumours from patients. Normal samples n = 81; tumour samples n = 81. Source data are provided as a Source Data file.