Overexpression of MUC1 Induces Non-Canonical TGF-β Signaling in Pancreatic Ductal Adenocarcinoma

Abstract

Pancreatic ductal adenocarcinoma (PDA) is one of the most lethal human cancers. Transforming Growth Factor Beta (TGF-β) is a cytokine that switches from a tumor-suppressor at early stages to a tumor promoter in the late stages of tumor development, by yet unknown mechanisms. Tumor associated MUC1 is aberrantly glycosylated and overexpressed in >80% of PDAs and is associated with poor prognosis. MUC1 expression is found in the early stages of PDA development with subsequent increase in later stages. Analysis of human PDA samples from TCGA database showed significant differences in gene expression and survival profiles between low and high MUC1 samples. Further, high MUC1 expression was found to positively correlate to TGF-βRII expression and negatively correlate to TGF-βRI expression in PDA cell lines. We hypothesized that MUC1 overexpression induces TGF-β mediated non-canonical signaling pathways which is known to be associated with poor prognosis. In this study, we report that MUC1 overexpression in PDA cells directly activates the JNK pathway in response to TGF-β, and leads to increased cell viability via up-regulation and stabilization of c-Myc. Conversely, in low MUC1 expressing PDA cells, TGF-β preserves its tumor-suppressive function and inhibits phosphorylation of JNK and stabilization of c-Myc. Knockdown of MUC1 in PDA cells also results in decreased phosphorylation of JNK and c-Myc in response to TGF-β treatment. Taken together, the results indicate that overexpression of MUC1 plays a significant role in switching the TGF-β function from a tumor-suppressor to a tumor promoter by directly activating JNK. Lastly, we report that high-MUC1 PDA tumors respond to TGF-β neutralizing antibody in vivo showing significantly reduced tumor growth while low-MUC1 tumors do not respond to TGF-β neutralizing antibody further confirming our hypothesis.

Keywords: JNK (c-Jun N-terminal kinase); MUC1—mucin 1; TGF-beta; non-canonical pathways; pancreatic ductal adenocarcinoma.

FIGURE 1

Heatmap showing top 30 differentially expressed genes in high/moderate vs low MUC1 PDA samples from TCGA. (A) Top panel shows the color key for MUC1 expression in the 29 PDA samples. Right hand side shows the color key histogram for expression levels of each gene named on the right. Left hand side color key shows the genes associated with each of the three pathways in pink (TGF-β), green (MAPK) and peach (BMP). Genes with a false discovery rate adjusted p < 0.05 are shown. (B) Kaplan-Meier curve for overall survival (OS) in the 29 PDA patients from TCGA in low (blue) vs high/moderate (red) groups are shown.

FIGURE 2

High MUC1 expression in PDA cells positively correlates to TGF-βRII and negatively correlates to TGF-βRI levels. (A) The expression of MUC1-CT, TGF-βRI, TGF-βRII, and endogenous loading control β-actin in a panel of PDA cell lines, determined by Western blot. (B) Densitometric analysis of MUC1 expression versus TGF-βRI expression shows a negative correlation (Spearman’s correlation coefficient r = -0.2381, NS). (C) Densitometric analysis of MUC1 expression versus TGF-βRII expression shows a significantly positive correlation (Spearman’s correlation coefficient r = 0.8810, p = 0.0072).

FIGURE 3

Overexpression of MUC1 leads to increased phosphorylation of JNK and c-Myc and knockdown of MUC1 reduces phosphorylation of JNK and c-Myc. (A) Western blot expression of phosphorylation of JNK and c-Myc compared to total JNK and total c-Myc in MiaPaca2 vs MiaPaca2. MUC1 cells in response to 10 ng/ml of TGF-β at 10 min. (B) Western blot expression of phosphorylation of JNK and c-Myc compared to total JNK and total c-Myc in HPAFII cells treated with control siRNA vs MUC1 siRNA in response to 10 ng/ml of TGF-β at 10 min. (C) Densitometric analysis of fold change of expressions of pJNK/Total JNK and p-c-Myc/Total c-Myc normalized to endogenous β-actin is presented in MiaPaca2 cells. (D) Densitometric analysis of fold change of expressions of pJNK/Total JNK and p-c-Myc/Total c-Myc normalized to endogenous β-actin is presented in HPAFII cells. (E) Knockdown efficiency of MUC1 in HPAFII after 72 h of siRNA treatment. Data are presented as means ± SEM of n = 3; Unpaired Student’s t-test and one-way ANOVA were used to analyze the differences between treatment groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 4

TGF-β exposure increases viability in cells with high MUC1 and reduces viability in low MUC1 PDA cells. MTT cell viability assay on (A) MiaPaca2.Neo cells with 10 ng/ml of TGF-β for 48 h. (B) HPAFII and (C) MiaPaca2. MUC1 cells with 10 ng/ml of TGF-β for 72 h. (D) HPAFII treated with control or MUC1 siRNA for 72 h followed by treatment with 10ng/ml of TGF-β for 24 h. All data are shown as means ± SEM of n = 3. Unpaired t-test was performed to compare between treated and untreated cells for each one of experiments A-C and two-way ANOVA was used to compare between untreated and treated in HPAFII.controlsiRNA and HPAFII.MUC1siRNA. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 5

TGF-β neutralizing antibody treatment significantly reduced high-MUC1 (HPAFII) but not low MUC1 (MiaPaca2) tumor growth in vivo. (A) A schematic of the xenograft study showing the treatment with control IgG and anti-TGF-β antibody (20 ug/100 ul per mouse). (B) On the left: Tumor growth of HPAFII (n = 5 for TGF-β neutralizing Ab and n = 4 for IgG isotype) is shown. On the right: Tumor growth of MiaPaca2 (n = 6 for both groups) is shown. Tumor growth was determined biweekly by caliper measurements and tumor size in mm³ is plotted. (C) Wet weight of HPAFII tumors (left) and MiaPaca2 tumors (right) respectively are shown. Two-way ANOVA was used to compare between the different treatment groups. *p <0.05, NS: non-significant. (D) Immunohistochemistry showing expression of MUC1 in MiaPaca2 (left) and HPAFII (right) tumors.

FIGURE 6

Schematic diagram of the proposed mechanism of TGF-β signaling and functions in high versus low MUC1 PDA. Left panel shows activation of SMAD-dependent canonical pathway in low-MUC1 PDA cells. TGF-β ligands bind to the membranous TGF-β receptor (TGF-βRII) homodimers with high affinity. TGF-βRII binding allows dimerization with TGF-β type I receptor (TGF-βRI) homodimers, activation of the TGF-βRI kinase domain and signal transduction via phosphorylation of the C-terminus of receptor-regulated SMADs (R-SMAD), SMAD2 and SMAD3. The SMAD2/3 dimer then forms a heterotrimeric complex with SMAD4 which translocates in the nucleus (Massagué and Wotton, 2000; Ross and Hill, 2008). This leads to growth inhibition, cell cycle arrest and apoptosis of PDA cells, thus TGF-β acts as a tumor suppressor. Right panel shows activation of SMAD-independent non-canonical pathway in high-MUC1 PDA cells. In this pathway, binding of TGF-β mainly to TGF-β-RII most likely increases phosphorylation of c-SRC which in turn phosphorylates MAPK, followed by JNK and c-Myc (Bunda et al., 2014). This phosphorylation cascade activates the MAPK/JNK pathway and stabilizes c-Myc which translocates into the nucleus to increase transcription of oncogenic proteins and leads to increased growth, invasion and EMT of PDA cells (Fey et al., 2016). MUC1-CT also aids in the process by its oncogenic signaling. Thus, in high-MUC1 PDA cells TGF-β acts as a pro-tumorigenic cytokine. The schematic was created with BioRender.com.