Inhibition of fibroblast growth by Notch1 signaling is mediated by induction of Wnt11-dependent WISP-1

Abstract

Fibroblasts are an integral component of stroma and important source of growth factors and extracellular matrix (ECM). They play a prominent role in maintaining tissue homeostasis and in wound healing and tumor growth. Notch signaling regulates biological function in a variety of cells. To elucidate the physiological function of Notch signaling in fibroblasts, we ablated Notch1 in mouse (Notch1(Flox/Flox)) embryonic fibroblasts (MEFs). Notch1-deficient (Notch1(-/-)) MEFs displayed faster growth and motility rate compared to Notch1(Flox/Flox) MEFs. Such phenotypic changes, however, were reversible by reconstitution of Notch1 activation via overexpression of the intracellular domain of Notch1 (NICD1) in Notch1-deficient MEFs. In contrast, constitutive activation of Notch1 signaling by introducing NICD1 into primary human dermal fibroblasts (FF2441), which caused pan-Notch activation, inhibited cell growth and motility, whereas cellular inhibition was relievable when the Notch activation was countered with dominant-negative mutant of Master-mind like 1 (DN-MAML-1). Functionally, "Notch-activated" stromal fibroblasts could inhibit tumor cell growth/invasion. Moreover, Notch activation induced expression of Wnt-induced secreted proteins-1 (WISP-1/CCN4) in FF2441 cells while deletion of Notch1 in MEFs resulted in an opposite effect. Notably, WISP-1 suppressed fibroblast proliferation, and was responsible for mediating Notch1's inhibitory effect since siRNA-mediated blockade of WISP-1 expression could relieve cell growth inhibition. Notch1-induced WISP-1 expression appeared to be Wnt11-dependent, but Wnt1-independent. Blockade of Wnt11 expression resulted in decreased WISP-1 expression and liberated Notch-induced cell growth inhibition. These findings indicated that inhibition of fibroblast proliferation by Notch pathway activation is mediated, at least in part, through regulating Wnt1-independent, but Wnt11-dependent WISP-1 expression.

Figure 1. Deletion of Notch1 relieves its suppressive effect on cell growth and motility of MEFs.

(A) Expression of full length Notch1 (250 Kd) and Trans-Membrane and Intra-Cellular domain (TMIC, 120 Kd) in Cre/lenti- versus GFP/Lenti-transduced Notch1^Flox/Flox MEFs. Knock-out Notch1 in Cre/lenti-transduced Notch1^Flox/Flox MEFs was analyzed by Western blotting assay. β-actin was used as loading control. (B) Rate of cell growth of MEFs in the presence 2% or absence of serum was measured by MTT assay. 5,000 cells/well were plated. Deletion of Notch1 promotes cell growth of MEFs. (C) Deletion of Notch1 increases cell motility of MEFs. The migration of 15 randomly selected cell pairs was tracked over time (20 hours) by time-lapse photography and velocity was calculated with software. (D) Left: Expression of endogenous full length Notch1 (250 Kd), TMIC (120 Kd) and exogenous NICD (119 Kd) in MEFs was shown. Expression of activated Notch1 (NICD) was detected by specific antibody recognizing Val1744. Right: Increased cell growth rate was reversed by reconstituted Notch1 activation in Notch1-deficint (Notch1 ^−/−) MEFs by NICD overexpression as demonstrated by immunoblotting (β-actin was used as loading control). Cell growth rate of MEFs in the presence 2% or absence of serum was measured by MTT assay. * P<0.05; ** P<0.01. Data are presented as mean ± SD of three independently performed experiments in (B), (C), and (D).

Figure 2. Activated Notch1 induces pan-Notch activation in human fibroblasts.

(A) FF2441 cells were transduced with either NICD1-GFP/lenti or GFP/lenti. Two days after transduction, transductants were harvested and subjected to Western blot analysis. Expression of endogenous full length Notch1 (250 Kd), TMIC (120 Kd) and exogenous NICD (119 Kd) in MEFs was shown. Expression of activated Notch1 (NICD) was detected by specific antibody recognizing Val1744. β-actin was used as control. (B) Constitutive activation of Notch1 results in self-propagated pan-Notch activation. Up-regulation of Notch receptors, ligands and target genes in FF2441-NICD1-GFP versus FF2441-GFP is observed by RT² PCR Array. The folds of increased gene expression in FF2441-NICD1-GFP compared to FF2441-GFP cells from two independently performed experiments were shown.

Figure 3. Effects of Notch activation and inhibition on human dermal fibroblast growth.

Notch activation slows down cell growth rate of fibroblasts. Proliferation of transduced FF2441 cells was determined by MTT assays (A). Activated Notch1 inhibited cell growth of fibroblasts. (B) Inhibition of Notch activation by ectopic expression of DN-MAML-1 (upper) in FF2441-NICD-GFP cells relieves cell growth inhibition as measured by MTT assay (lower). Data are presented as mean ± SD of three independently performed experiments in (A) and (B). (C) Notch activation does not induce cell apoptosis in fibroblasts. Cell apoptosis was determined by TUNEL assay. No obvious apoptotic cells were detectable in FF2441-NICD1-GFP cells. Serum starvation-induced cells were used as positive control (FBS-).

Figure 4. The Notch pathway is maintained in a less-activated or inactivated status in proliferating human dermal fibroblasts.

The increased expression of Hey-1 protein in quiescent fibroblasts compared to proliferating fibroblasts was demonstrated by immunoblotting analysis. β-actin was used as loading control.

Figure 5. Inhibitory effects of “Notch-activated” fibroblasts as stromal cells on melanoma cell growth and invasion in 3D model.

(A) Representative images of H&E staining of sections of 3D skin melanoma model. (B) Melanoma cell growth in 3D skin model. Significantly decreased melanoma cells were observed in 3D reconstructs in which FF2441-NICD1-GFP cells were embedded in collagen. (C) Decreased invasion into ‘dermis’ by melanoma cells in 3D skin reconstructs in which FF2441-NICD1-GFP cells were embedded. All data are calculated based on that from 5 randomly selected LPF/section and totally 10 sections per group.

Figure 6. Notch pathway activation induces expression of WISP-1/CCN4 in fibroblasts.

(A) Levels of mRNA of WISP-1/CCN4 were up-regulated in FF2441-NICD1-GFP compared to FF2441-GFP cells. Data are from RT² PCR Array and presented as fold changed in gene expression by setting levels of genes in FF2441-GFP as “1”. Data are presented as mean ± SD of three independently performed experiments. (B) Expression of WISP-1/CCN4 protein was up-regulated by Notch pathway activation in fibroblasts. Expression of WISP-1/CCN4 in FF2441-NICD1-GFP versus FF2441-GFP and untreated F2441 (−) cells was analyzed by Western blotting assay. β-actin was used as loading control. (C) Expression of WISP-1/CCN4 protein was down-regulated in Notch1 ^−/− MEFs. Expression of WISP-1/CCN4 in Notch1 ^−/− versus Notch1^Flox/Flox MEFs (both untreated (−) and GFP/lenti-transduced) was analyzed by Western blotting assay. β-actin was used as loading control. (D) Increased WISP-1 promotor-driven luciferase activity in FF2441-NICD1-GFP compared to FF2441-GFP cells. Data are presented as mean ± SD of three independently performed experiments.

Figure 7. Effect of WISP-1/CCN4 on cell growth of fibroblasts.

Supplementation of exogenous recombinant human WISP-1/CCN4 slowed down cell growth of FF2441 cells (A) and Cre/MEFs (B). Proliferation of cells was determined by MTT assays. 5,000 FF2441 cells/well and 2,000 Cre/MEFs/well were plated respectively. Data are presented as mean ± SD of three independently performed experiments. (C) Addition of 200 ng/mL of recombinant human WISP-1/CCN4 inhibited serum-induced phosphorylation of Erk1/2 in FF2441 cells. Total amount of Erk1/2 was used as control. Time course of phosphorylation of Erk1/2. Serum starved cells were stimulated with 10% serum in the presence of WISP-1/CCN4 for varying times. Autophotographs of Western blots were quantified by computerized densitometry. pERK1/2 signals were normalized to total ERK levels and unstimulated (time “0”) samples were set as “1”. Relative index of pErk1/2 compared to unstimulated samples from three independent Western blots was plotted. (D) A representative Western blot was shown.

Figure 8. WISP-1/CCN4 is partially responsible for mediating the inhibitory effect of Notch signaling on fibroblast proliferation.

(A) siRNA-mediated knocking-down WISP-1/CCN4 expression in human fibroblasts. Expression of WISP-1/CCN4 in siRNA-transfected versus non-targeting control siRNA-transfected or untransduced FF2441-NICD1-GFP cells was analyzed by Western blotting assay. β-actin was used as loading control. (B) Blockage of WISP-1/CCN4 expression relieved Notch's inhibitory effect on cell growth of fibroblasts. Proliferation of siRNA-transfected versus non-targeting control siRNA-transfected FF2441-NICD1-GFP cells was determined by MTT assays. 2,000 cells/well were plated. Data are presented as mean ± SD of three independently performed experiments.

Figure 9. Notch-induced WISP-1/CCN expression is Wnt11-dependent in the fibroblasts.

(A) Levels of mRNA of Wnt11 were up-regulated in FF2441-NICD1-GFP cells compared to FF2441-GFP cells. Data are from RT² PCR Array and presented as fold changed in gene expression by setting levels of genes in FF2441-GFP as “1”. (B) Expression of Wnt11 protein was up-regulated by Notch pathway activation in fibroblasts. Expression of Wnt11 in FF2441-NICD1-GFP versus FF2441-GFP and untreated FF2441 (−) cells was analyzed by Western blotting assay. β-actin was used as loading control. (C) Blockage of Wnt11 expression resulted in decreased expression of WISP-1/CCN expression. siRNA-mediated knocking-down Wnt11 expression in human fibroblasts. Expression of WISP-1/CCN4 in wnt11siRNA-transfected versus non-targeting control siRNA-transfected FF2441-NICD1-GFP cells was analyzed by Western blotting assay. β-actin was used as loading control. (D) Blockage of Wnt11 expression relieved Notch's inhibitory effect on cell growth of fibroblasts. Cell proliferation of wnt11 siRNA-transfected-, non-targeting control siRNA-transfected- and their “parental” FF2441-NICD1-GFP was determined by MTT assays. 2,000 cells/well were plated. Data are presented as mean ± SD of three independently performed experiments.