Genes & Development



Supplemental Information

Survival Factor NFIL3 Restricts FOXO-induced Gene Expression in Cancer

Megan Keniry1, Maira M. Pires1,2, Sarah Mense2, Celine Lefebvre1,3, Boyi Gan4, Karen Justiano1, Ying-Ka Ingar Lau1, Ben Hopkins1, Cindy Hodakoski1,2, Susan Koujak1, Joseph Toole1, Franklyn Fenton1, Ashley Calahan1, Andrea Califano1,3, Ronald A. DePinho5, Matt Maurer1, 6, and Ramon Parsons2*

1Institute for Cancer Genetics and Herbert Irving Comprehensive Cancer Center, Columbia University, 1130 St. Nicholas Avenue, NY, NY 10032, USA. 2Department of Oncological Sciences, Icahn School of Medicine at Mount Sinai, 1470 Madison Ave HCSM 6-117, New York, NY 10029, USA. 3Department of Biomedical Informatics Columbia University, 1130 St Nicholas Ave, ICRC, New York, NY, 10032, USA. 4Department of Experimental Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA. 5Department of Cancer Biology, University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA. 6Department of Medicine, Columbia University Medical Center 630 W. 168th Street, NY, NY 10032, USA. Street, NY, NY 10032, USA.

*Correspondence: ramon.parsons@mssm.edu (R.P.)

*Contact: Ramon Parsons

email: ramon.parsons@mssm.edu

Itemized List of Supplemental Information

Supplemental Figures 1-8:

• Supplemental Figure 1 accompanies main text Figure 1.

• Supplemental Figure 2 accompanies main text Figure 2.

• Supplemental Figure 3 accompanies main text Figure 3.

• Supplemental Figure 4 accompanies main text Figure 4.

• Supplemental Figure 5 accompanies main text Figure 5.

• Supplemental Figure 6 accompanies main text Figure 6.

• Supplemental Figure 7 accompanies main text Figure 7.

• Supplemental Figure 8 accompanies main text Figure 8.

Supplemental Table 1 Legend (Accompanying Word File):

• Supplemental Table 1 accompanies main text Figure 4.

Supplemental Tables 2-6:

• Supplemental Table 2 accompanies main text Figure 4.

• Supplemental Table 3 accompanies main text Figure 4.

• Supplemental Table 4 accompanies main text Figure 4.

• Supplemental Table 5 accompanies main text Figure 4.

• Supplemental Table 6 accompanies main text Figure 6.

• Supplemental Table 7 accompanies main text Figure 8.

Supplemental Table 8 Legend (Accompanying Word File)

• Supplemental Table 8 is included as a Word file.

Supplemental Materials and Methods

• This section includes additional Materials and Methods.

Supplemental References

[pic]

Supplemental Figure 1. Genes Examined in Reporter Screen and Validation of the TRAIL Reporter as an Indicator of PI3K/PTEN Output.

(A) Graph depicts the differential expression of genes in the presence of exogenous PTEN from previously published microarray studies; more detailed descriptions of these previously published studies (cell lines utilized and experimental designs) can be found in the Supplemental Experimental Procedures (Hong et al. 2000; Matsushima-Nishiu et al. 2001; Simpson et al. 2001; Stolarov et al. 2001; Unoki and Nakamura 2001). Genes from these microarray studies that were differentially expressed by at least two-fold and had available cDNAs in the human Orfeome Library 1.1 were chosen as candidates for our screen (in blue on the graph). (B) Side by side comparison of PTEN-induction of six FOXO-regulated luciferase reporters in HEK293 cells, * significantly induced by exogenous PTEN. The TRAIL reporter was most strongly induced. (C) TRAIL reporter assays with exogenous PTEN, with or without DN-FOXO1, *significantly induced, **significantly lower than control vectors. PTEN induced reporter activity, except in the presence of DN-FOXO1. (D) TRAIL reporter assays were performed with 100 nM wortmannin (an inhibitor of PI3K), added 8 hours prior to assays, *significantly induced. (E) TRAIL reporter assays performed with exogenous Myr-AKT, which encodes constitutively membrane localized AKT, *significantly lower than control. (F) TRAIL eporter assays with exogenous NFIL3 and/or exogenous FOXO1, ** significantly less induced than exogenous FOXO1 alone sample. (G) U87MG cells with control vector or a retroviral PTEN vector were treated with control or NFIL3 shRNA; endogenous TRAIL expression was measured by qRT-PCR, *significantly different than control shRNA, ** significantly different than induction observed with both control vectors and control vector with exogenous PTEN. (H) ChIP analysis with FOXO1 antibody. DNA was subjected to quantitative PCR for proximal region of INSR promoter or control β-actin; FOXO1 associated with the INSR promoter. Data are means ± SEM of three experiments.

[pic]

Supplemental Figure 2. Sixty Base Pairs of the TRAIL Promoter are Sufficient for PTEN and NFIL3 Regulation

(A) Luciferase assays were performed using the reporters TRAIL-165 and TRAIL-60 that contain 165 and 60 base pairs of the TRAIL promoter, respectively, *significantly different than control, ** significantly less well induced by exogenous PTEN. Both of these reporters were regulated by PTEN and NFIL3. (B) Predicted NFIL3 and FOXO binding sites (underlined) based on published consensus sequences and mutant luciferase reporter sequences (used in Fig. 2C-D) are shown; the bold letters for the NFIL3 consensus are given a higher weight (Brunet et al. 1999; Macgillavry et al. 2011). (C) Putative FOXO and NFIL3 binding sites were mutated in TRAIL reporter as indicated and were tested for regulation by FOXO1 and NFIL3; FOXO1 induction was reduced with the mutant FOXO site, *significantly less induced by exogenous FOXO1 in comparison to the wildtype TRAIL reporter. Deletion of the NFIL3 site led to a significant (*) reduction in regulation by NFIL3. Data are means ± SEM of three experiments.

[pic]

Supplemental Figure 3. Role of HDACs in NFIL3/FOXO1 Transcriptional Regulation

(A-B) NFIL3 and HDAC2 physically associate. Co-immunoprecipitations were performed with HEK293 cell extracts that expressed FLAG-HDAC2 and V5-NFIL3. Faint background bands are present in input samples due to pre-clearing reactions with normal mouse IgG. (C) TRAIL reporter assays with samples that were treated with or without 2 μM SAHA for 24 hours. HDAC inhibition reduced the NFIL3-mediated repression of the TRAIL reporter in the presence of exogenous PTEN. (D) TRAIL reporter assays were performed with the FOXO1 acetylation deficient mutant (6KR) or the acetylation mimetic mutant (6KQ) with or without exogenous NFIL3. The acetylation mutants had no effect on NFIL3-mediated regulation of reporter activity, *significantly different than control vector sample, **significantly less well induced than with exogenous FOXO1 or 6KR mutant, respectively. Data are means ± SEM of three experiments.

[pic]

Supplemental Figure 4. GSEA with Microarray Data from BT549 cells with NFIL3 shRNA and HDAC Inhibition in the Burkitt’s Lymphoma Cell Line (RJ225)

(A-C) BT549 cells with lentiviral GFP or Nfil3 (not targetable by human NFIL3 shRNA) were treated with human NFIL3-targeting shRNA or control; q-RT-PCR analysis was performed for GADD45α, TRAIL and NFIL3 expression,*significantly different than control shRNA, ** significantly different than induction observed with GFP. (D) GSEA was performed with microarray data obtained from BT549 samples that had NFIL3 expression diminished by shRNA. Gene sets induced by exogenous FOXO1, treatment with LY294002, and exogenous PTEN were significantly enriched in the NFIL3 knockdown samples. (E) Gene Set Enrichment Analysis (GSEA) was performed with data from the Burkitt’s lymphoma cell line RJ225 that was treated with the HDAC inhibitor Trichostatin A (TSA) and PTEN pathway gene sets. Gene sets induced by the PI3K inhibitor LY294002, by exogenous FOXO1 and by exogenous PTEN were up-regulated by HDAC inhibition in a Burkitt’s lymphoma cell line. Abbreviations: ES= enrichment score and NES= normalized enrichment score. Data are means ± SEM of three experiments.

[pic]

Supplemental Figure 5. FOXO Subcellular Localization in Cell Lines and MEFs

(A) Subcellular fractionations with HEK293, MDA-MB-468 (labeled as 468) and BT549 cell lines probed with additional FOXO1 (C29H4) and FOXO3 (75D8) antibodies. MAX and β−tubulin are markers for the nuclear (N) and cytoplasmic (C) fractions respectively; the MAX and β−tubulin panels for HEK293 and MDA-MB-468 cells are same as those depicted in the main text Figure 5A. FOXO1 and FOXO3 were in the nucleus and cytoplasm. (B) Western blot analysis of subcellular fractionations with Large T antigen-immortalized MEFs (floxed/floxed, Rosa26CreERT2, FoxO1, FoxO3 and FoxO4 that were treated with Ad-GFP or Ad-Cre); samples were probed for FOXO1, FOXO3, and control antibodies. The detected FoxO bands were diminished in Ad-Cre treated samples showing the specificity of these antibodies. (C) Subcellular fractionations with control or Pten -/- MEFs; residual FoxO is detected in the nuclear fraction of Pten -/- MEFs. (D-E) The indicated FOXO transcription factor was targeted with siRNA in HEK293 cells; western blots are shown, indicating the specificity of each antibody (molecular weights in kD are indicated). (F) The expression of NFIL3 was diminished with shRNA hairpins (KD1 or KD2) in BT549 cells; nuclear and cytoplasmic fractions were prepared. Fractions were probed for indicated antibodies. NFIL3 diminishment did not alter FOXO localization.

[pic]

Supplemental Figure 6. NFIL3 transcription is induced by H2O2 and NFIL3 Regulates Transcription through FOXO.

(A) NFIL3 expression was assessed with qRT-PCR from BT549 extracts that were either treated with 250μM H2O2 for 18 hours or water; NFIL3 was induced by H2O2. (B) Primary MEFs were treated with Ad-GFP or Ad-Cre; qRT-PCR was performed to confirm the loss of FoxO expression in Cre treated samples; the PCR only detects recombinants. (C) Primary MEFs (Rosa26CreERT2, floxed/floxed FoxO1, FoxO3 and FoxO4) of the indicated genotype were infected with lentivirally delivered control or Nfil3-targetting shRNA; qRT-PCR was performed with samples. Data are means ± SEM of three experiments.

[pic]

Supplemental Figure 7. NFIL3 and HDAC2 are Expressed in Poor Prognosis Cancers and a Series of FOXO1 Targets are Associated with Poor Breast Cancer Prognosis.

The Oncomine database was queried for NFIL3 expression (Rhodes et al. 2007). (A) NFIL3 and HDAC2 expression were elevated in basal-like breast cancer (Farmer et al. 2005) and GBM (Shai et al. 2003). (B-D) Kaplan Meier analysis with NFIL3/FOXO target genes in breast cancer using NKI 295 data from luminal A and basal subtypes; NFIL3 expression is associated with poor prognosis. (E-H) Kaplan Meier analysis was done with previously published FOXO targets using NKI data. Twenty-seven known FOXO targets that showed no evidence of NFIL3 regulation in our microarray studies were tested for their relationship to breast cancer prognosis and 11 trended towards poor prognosis and 4 were significantly associated to poor prognosis with p value < 0.05.  (I) The ratio of expression of indicated FOXO target genes plus or minus the standard deviation of NFIL3 KD samples versus control shRNA samples is shown. Data are extracted from 293 microarray data from Figure 4, (left column) and ChIP for FOXO1 with the promoters of these genes in HEK293 cells (right column), *significantly different than rabbit IgG control binding, ^ChIPs were performed in BT549 cells for this gene.

[pic]

Supplemental Figure 8. HDAC2 and NFIL3 Regulate Cell Viability in Basal-like Breast Cell Lines.

(A) HDAC2 was diminished in MCF10A cells using shRNA (KD1 or KD2). A western blot is shown. (B) The MCF10A samples with and without HDAC2 knockdown were subjected to apoptosis analysis as described in the Supplemental Materials and Methods. HDAC2 shRNA induced cell death. (C-D) Cell lines with NFIL3 reduced by shRNA were analyzed by western analysis and these cells were examined for cell death/apoptosis as described in the Supplemental Materials and Methods. The percent of apoptotic cells equals the propidium iodide and Annexin V single and double positive cells divided by the total number of cells. (E) MCF10A cells with either lenti-viral GFP or mouse Nfil3 (not targetable by human NFIL3 shRNA) were treated with control or human NFIL3-targeting shRNA. Cells remaining after treatment were stained with crystal violet, plates were scanned with UMAX PowerLook 1100 scanner and Image J was utilized to quantify cell density. Mouse Nfil3 partially rescued the cell death induced by human NFIL3 shRNA. (F-G) qRT-PCR with MCF10A samples treated with NFIL3 shRNA or control, *significantly different than control shRNA, ** significantly different than induction observed with GFP, *** significantly lower detected expression than samples with lentivirally delivered mouse Nfil3; note that the control samples only have human NFIL3. (H) Colony assays in presence of NFIL3 shRNA with or without TRAIL shRNA). The TRAIL KD1 had 41% of endogenous TRAIL by q-RT-PCR with 1.5% error and the KD2 for TRAIL had 63% of endogenous TRAIL with 1.0% error; both TRAIL KDs were significant. Data are means ± SEM of three experiments.

Supplemental Table 1 Legend (See accompanying Microsoft Word file):

Supplemental Table 1. Differentially Expressed Genes in NFIL3 Knockdown Samples

Gene expression profiling was performed with NFIL3 knockdown samples and scramble shRNA controls (HEK293 cells). One-Way ANOVA analysis was used to detect 399 differentially expressed genes with an FDR=0.05 and at least a two-fold change in gene expression. 289 of these genes were up regulated.

[pic]

[pic]

Supplemental Table 4. Gene Set Enrichment Analysis with Curated Broad Institute Molecular Signatures Database Sets and PTEN Pathway Sets

Gene set enrichment analysis was performed with 1,679 gene sets that are curated by the Broad Institute Molecular Signatures Database and the NFIL3 knockdown microarray data from HEK293 cells. The experimentally derived PTEN-pathway gene sets were included in this analysis for comparison. Of these gene sets tested, 1403 were enriched to some degree in the NFIL3 knockdown samples. Enriched gene sets were ranked based on the normalized enrichment score (NES). The PTEN pathway gene sets and corresponding rankings are shown; three of these are in the top tenth percentile of enriched gene sets.

|Gene Set Name |SIZE |ES |NES |p-value |NES Rank (out of 1403 gene |

| | | | | |sets) |

|TERRAGNI- GENE SET INDUCED BY LY294002 |21 |0.64 |1.70 |0.03 |49 |

|RAMASWAMY- FOXO1 INDUCED SET [CLASS 2A] |31 |0.46 |1.60 | ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download