ARF and p53 coordinate tumor suppression of an oncogenic IFN-β-STAT1-ISG15 signaling axis

Abstract

The ARF and p53 tumor suppressors are thought to act in a linear pathway to prevent cellular transformation in response to various oncogenic signals. Here, we show that loss of p53 leads to an increase in ARF protein levels, which function to limit the proliferation and tumorigenicity of p53-deficient cells by inhibiting an IFN-β-STAT1-ISG15 signaling axis. Human triple-negative breast cancer (TNBC) tumor samples with coinactivation of p53 and ARF exhibit high expression of both STAT1 and ISG15, and TNBC cell lines are sensitive to STAT1 depletion. We propose that loss of p53 function and subsequent ARF induction creates a selective pressure to inactivate ARF and propose that tumors harboring coinactivation of ARF and p53 would benefit from therapies targeted against STAT1 and ISG15 activation.

Figure 1. Acute loss of p53 induces functional ARF

(A) Western blot analysis of cell lysates from p53^flox/flox MEFs infected with Ad-LacZ (L) or Ad-Cre (C) harvested at the indicated time points. Fold change of ARF levels are relative to Ad-LacZ control. (B) qRT-PCR analysis of p53 and ARF mRNA levels from p53^flox/flox MEFs infected with Ad-LacZ or Ad-Cre. mRNA levels were normalized to β-Actin and fold changes are relative to Ad-LacZ controls. Error bars represent s.d. for n=3 from three independent experiments. (C) Proliferation assay performed with cells described in (A andB). (D) Ad-LacZ or Ad-Cre infected p53^flox/flox MEFs pulsed with BrdU for 4 hours. BrdU and DAPI positive nuclei were visualized using immunofluorescence, and data represents percent BrdU positive nuclei from three independent experiments. (E) Representative image of foci assay with Ad-LacZ or Ad-Cre infected p53^flox/flox MEFs. (F) Western blot analysis of dp53 MEFs infected with shSCR or shARF. (G) Equal numbers of dp53 MEFs infected with shSCR or shARF were plated and manually counted on the indicated days. (H) Representative image of foci assay performed with dp53 MEFs expressing shSCR or shARF. See also Figure S2.

Figure 2. Endogenous ARF limits the tumorigenicity of p53-deficient cells

(A) Western blot analysis of dp53 MEFs expressing Ras^V12 (dp53R), and infected with shSCR or shARF. (B) Representative images of dp53R-shSCR or dp53R-shARF MEFs growing in soft agar. Macroscopic colonies were quantified in (C). Error bars represent s.d. of n=3. (D) Proliferation assay of dp53 MEFs expressing empty vector or Ras^V12 and infected with shARF or shSCR control. (E) Percent BrdU positive nuclei of cells described in (D) following 4-hour pulse with BrdU. Error bars represent s.d. from three independent measurements of 100 nuclei. (F) Representative image of foci assay performed with dp53R MEFs expressing shSCR or shARF. (G) Images of tumor-bearing mice and excised tumors from allograft experiments using dp53R-shARF or shSCR MEFs. (H) Tumor size was measured using calipers on the indicated days post-injection. Tumor size (volume) was calculated as described in the Methods section. Error bars represent s.d. of n=5.

Figure 3. ARF inhibits an Interferon-sensitive gene signature induced upon p53-loss

(A) Western blot verifying overexpression of Ras^V12 and knockdown of ARF in dp53 MEFs. RNA from three independent experiments was submitted for microarray analysis. (B) Heat map showing significantly altered genes (>2-fold change), and pathway analysis of significantly altered genes in dataset. (C) Validation of ISGs with qRT-PCR. Levels were normalized to histone 3.3 mRNA and are relative to shSCR controls. Error bars represent s.d. from three independent experiments. (D) qRT-PCR analysis of p53^flox/flox MEFs infected with Ad-LacZ or Ad-Cre from the indicated time points post-infection. Cells were all infected with shSCR(−) or shARF(+) 1 day post Cre- infection as indicated. mRNA levels are relative to Ad-LacZ-shSCR controls and represent averages of three independent experiments. (E) Western blot analysis of cells described in (D). See also Figures S1, S2, and S3.

Figure 4. IFN-β signaling is necessary and sufficient for enhanced tumorigenicity in dp53R-shARF MEFs

(A) qRT-PCR analysis of IFN-β mRNA levels in dp53R-shARF MEFs. Levels are normalized to histone 3.3 mRNA and relative to shSCR controls. (B) Extracellular IFN-β concentration measured by ELISA in dp53R-shARF MEFs. Values are fold changes relative to shSCR control. Error bars represent s.d. of three independent experiments. (C) qRT-PCR analysis of dp53R-shSCR or -shARF MEFs infected with two specific shRNAs targeting IFN-β. Relative mRNA expression was obtained by normalizing to Histone 3.3 mRNA. Error bars represent s.d. of three independent measurements. (D) Western blot analysis of cells described in (C) for the indicated proteins. (E) Representative image of foci assay performed with dp53R-shARF or shSCR MEFs infected with two IFN-β-specific shRNAs. Quantification of three independent measurements is shown (right). (F) qRT-PCR analysis of dp53R-shSCR or –shARF cells treated with the indicated concentration of IFN-β. Error bars represent s.d. of values from three independent measurements. (G) Representative image of foci assay performed with dp53R-shARF or shSCR MEFs treated with the indicated concentration of recombinant IFN-β. Quantification of three independent measurements (right). *=P<0.01

Figure 5. STAT1 activation is required for increased tumorigenicity in dp53R-shARF MEFs

(A) Western blot analysis of dp53R-shARF or shSCR MEFs showing STAT1 activation. (B) qRT-PCR analysis of total STAT1 mRNA levels in dp53R-shARF MEFs. mRNA levels are relative to shSCR controls and normalized to histone 3.3. (C) Western blot analysis of dp53R-shSCR or shARF MEFs infected with two different STAT1 shRNAs. (D) qRT-PCR analysis of dp53R-shSCR or shARF MEFs infected with control or two different STAT1 shRNAs. (E) Proliferation assay of dp53R MEFs expressing the indicated shRNAs. (F) Representative images of foci assays with dp53R MEFs expressing indicated shRNAs. (G) Soft agar quantification of STAT1 depleted dp53R-shARF MEFs. All error bars represent s.d. for n=3. *=P<0.0004, **=P<0.009. See also Figure S4.

Figure 6. ISG15 is required for increased tumorigenicity in dp53R-shARF MEFs

(A) Western blot analysis of ISG15 expression in dp53R-shARF MEFs. Free and conjugated forms are indicated. (B) Western blot analysis of dp53R-shSCR or shARF MEFs expressing an shRNA specifically targeting ISG15. (C) Quantification of macroscopic soft agar colony number with cells described in (B). (D) Representative image of foci experiment from dp53R MEFs infected with the indicated shRNAs. (E) Proliferation assay for dp53R MEFs infected with the indicated shRNAs. All error bars represent s.d. of n=3

Figure 7. Analysis of TNBC patient samples and cell lines

(A) Statistics from immunohistochemistry staining of human breast cancer tissue array. (B) Representative images from IHC displaying a section with high ARF staining (TNBC-1) and a section with low/no ARF and high ISG15/STAT1 (TNBC-2). (C) Proliferation assays of the indicated triple negative breast cancer cell lines infected with two different STAT1 shRNAs. (D) Western blot analysis showing STAT1 depletion with shRNAs in various TNBC lines. See also Figures S5, S6, and Table S1.