ATF3, an adaptive-response gene, enhances TGF{beta} signaling and cancer-initiating cell features in breast cancer cells

Abstract

The activating transcription factor 3 (ATF3) gene is induced by a variety of signals, including many of those encountered by cancer cells. We present evidence that ATF3 is induced by TGFβ in the MCF10CA1a breast cancer cells and plays an integral role for TGFβ to upregulate its target genes snail, slug and twist, and to enhance cell motility. Furthermore, ATF3 upregulates the expression of the TGFb gene itself, forming a positive-feedback loop for TGFβ signaling. Functionally, ectopic expression of ATF3 leads to morphological changes and alterations of markers consistent with epithelial-to-mesenchymal transition (EMT). It also leads to features associated with breast-cancer-initiating cells: increased CD24(low)-CD44(high) population of cells, mammosphere formation and tumorigenesis. Conversely, knockdown of ATF3 reduces EMT, CD24(low)-CD44(high) cells and mammosphere formation. Importantly, knocking down twist, a downstream target, reduces the ability of ATF3 to enhance mammosphere formation, indicating the functional significance of twist in ATF3 action. To our knowledge, this is the first report demonstrating the ability of ATF3 to enhance breast cancer-initiating cell features and to feedback on TGFβ. Because ATF3 is an adaptive-response gene and is induced by various stromal signals, these findings have significant implications for how the tumor microenvironment might affect cancer development.

Fig. 1.

ATF3 plays an integral role in the ability of TGFβ to regulate its target genes. (A) MCF10CA1a cells were transfected with control (siCtrl) or ATF3 (siATF3) siRNAs for 72 hours before TGFβ treatment for the indicated times. The mRNA levels of the indicated genes were determined by qRT-PCR and standardized against β-actin mRNA. The standardized level for each gene at time 0 was arbitrarily defined as 1. Shown is the average of three independent experiments (*P<0.05; ^#P<0.08). (B) Same as A, except that immunoblot for proteins using the indicated antibodies was shown. −, control siRNA; +, siATF3. A representative of two independent experiments is shown. (C) MCF10CA1a cells with ectopic expression of ATF3 or the control vector (Vec) were treated for the indicated time or untreated (0h) with TGFβ and analyzed by qRT-PCR as in A. *P<0.05. (D) MCF10CA1a cells with ectopic expression of ATF3 or the control vector (Vec) were analyzed by immunoblot using the indicated antibodies. A representative of two independent experiments is shown. (E) COS-1 cells were transfected with DNAs to express ATF3 and the indicated SMAD proteins (Flag-SMADs) followed by co-immunoprecipitation analyses (top). IP, immunoprecipitation. Five percent input is shown (bottom). (F) MCF10CA1a cells with (2h) or without (0h) TGFβ treatment were analyzed by co-immunoprecipitation as indicated. (G) MCF10CA1a cells with (2h) or without (0h) TGFβ treatment were analyzed by sequential ChIP using anti-SMAD2/3 antibody (S) followed by anti-ATF3 antibody (A) for immunoprecipitation. Rabbit IgG was used as a control.

Fig. 2.

ATF3 positively feeds back on TGFβ. (A) MCF10CA1a cells ectopically expressing ATF3 or the vector (Vec) were analyzed by qRT-PCR for the indicated mRNAs. Standardized signals in the control cells were arbitrarily defined as 1 and the average of three independent experiments is shown (*P<0.05). (B) ATF3 or vector (Vec) cells were analyzed by immunoblot for the indicated proteins. p-SMAD3, phospho-SMAD3; TGFβ, intracellular TGFβ. Shown is the representative of three experiments. (C) Conditioned medium was collected from MCF10CA1a cells ectopically expressing ATF3 or vector only (Vec) and incubated with the TGFβ reporter cells. The reporter activity was determined by luciferase assay and the number from the conditioned medium derived from Vec cells was arbitrarily defined as 1. The average of three independent experiments is shown (*P<0.05).

Fig. 3.

ATF3 promotes cell motility. (A,B) MCF10CA1a cells stably expressing control (shCtrl) or ATF3 (shATF3) shRNA were analyzed by the Boyden chamber migration assay in the presence or absence of 2.5 ng/ml TGFβ (A; *P<0.05, derived from three experiments). ATF3 protein levels were determined by immunoblot (B; a representative of three experiments). (C) MCF10CA1a cells ectopically expressing SMAD3 were transfected with control siRNAs (−) or siRNAs to knockdown ATF3 (+) for 72 hours, followed by a cell motility assay. Migration of cells without ectopic expression of SMAD3 was arbitrarily defined as 1 (*P<0.05). (D) As in C, except that ATF3-expressing cells were used and SMAD2/3 was knocked down (*P<0.05).

Fig. 4.

ATF3 enhances EMT. (A) Stable derivatives of MCF10CA1a cells with ATF3 knockdown (shATF3), ectopic expression of ATF3 or vector control (Vec) were photographed under a light microscope (at ×200 magnification). (B) Cells as indicated in A were analyzed by immunoblot for the indicated proteins. N-cad, N-cadherin; E-cad, E-cadherin; FN, fibronectin; Vim, vimentin. (C) Stable derivatives of MCF7 cells with ectopic expression of ATF3 or vector control (Vec) were photographed under a light microscope (at ×200 magnification). (D) Cells as indicated in C were analyzed by immunoblot for the indicated proteins. All results in this figure are representatives of at least three independent experiments.

Fig. 5.

ATF3 enhances cancer-initiating cell features. (A,B) Stable derivatives of MCF10CA1a cells with ATF3 knockdown (shATF3), ectopic expression of ATF3 or vector control (Vec) were analyzed by FACS for the CD24 and CD44 surface markers. A shows representative images and B the quantitation from eight repeated experiments (*P<0.05). Numbers in A indicate the percentage of the CD24^low–CD44^high population. (C–E) The indicated cells were seeded to form mammospheres and analyzed for mammosphere forming efficiency (C; *P<0.05, from four experiments), size (D; *P<0.05, from four experiments) and representative phase-contrast images (E). For C, only mammospheres with a diameter ≥70 μm were counted and the number of mammospheres derived from vector control cells was arbitrarily defined as 1 (*P<0.05, from four experiments). Scale bars: 40 μm in E. (F) Mammospheres derived from the indicated cells were analyzed for their IL6 mRNA levels by qRT-PCR. Standardized signal from vector control cells was arbitrarily defined as 1 (*P<0.05, from four experiments).

Fig. 6.

Knockdown of twist reduces the ability of ATF3 to enhance cancer-initiating cell features. Stable derivatives of MCF10CA1a cells with ectopic expression of ATF3 or vector control (Vec) were transfected with control (siCtrl) or twist (siTwist) siRNAs for 72 hours before being analyzed by immunoblot using the indicated antibodies (A), by FACS for the CD24 and CD44 surface markers (B) and for mammosphere formation (C). Only mammospheres with a diameter ≥70 μm were counted. A shows a representative image and B and C show the quantitation from three independent experiments (*P<0.05). For both panels, the number derived from vector control cells with control siRNA knockdown was arbitrarily defined as 1.