Transcriptome Analysis of Cells Exposed to Actinomycin D and Nutlin-3a Reveals New Candidate p53-Target Genes and Indicates That CHIR-98014 Is an Important Inhibitor of p53 Activity

Abstract

Co-treatment with actinomycin D and nutlin-3a (A + N) strongly activates p53. Previously we reported that CHIR-98014 (GSK-3 kinase inhibitor), acting in cells exposed to A + N, prevents activation of TREM2-an innate immunity and p53-regulated gene associated with Alzheimer's disease. In order to find novel candidate p53-target genes and genes regulated by CHIR-98014, we performed RNA-Seq of control A549 cells and the cells exposed to A + N, A + N with CHIR-98014 or to CHIR-98014. We validated the data for selected genes using RT-PCR and/or Western blotting. Using CRISPR/Cas9 technology we generated p53-deficient cells. These tools enabled us to identify dozens of candidate p53-regulated genes. We confirmed that p53 participates in upregulation of BLNK, APOE and IRF1. BLNK assists in activation of immune cells, APOE codes for apolipoprotein associated with Alzheimer's disease and IRF1 is activated by interferon gamma and regulates expression of antiviral genes. CHIR-98014 prevented or inhibited the upregulation of a fraction of genes stimulated by A + N. Downregulation of GSK-3 did not mimic the activity of CHIR-98014. Our data generate the hypothesis, that an unidentified kinase inhibited by CHIR-98014, participates in modification of p53 and enables it to activate a subset of its target genes, e.g., the ones associated with innate immunity.

Keywords: GSK-3; RNA-Seq; innate immunity; kinase inhibitor; p53.

Figure 1

Results of RNA-seq analysis of actinomycin D (A), nutlin-3a (N) and CHIR treatment combination on lung adenocarcinoma cells (A549). (a) Venn diagram comparing lists of genes significantly upregulated in response to treatment with AN and with AN in the presence of CHIR-98014 (ANCHIR). (b–d) Top 10 (if available) overrepresented pathways with adjusted p value < 0.05, size < 200 and intersection to query size ratio >0.03 for: (b) down-regulated genes in A + N versus control comparison, (c) up-regulated genes in A + N versus control comparison and (d) up-regulated genes in A + N + CHIR versus control comparison (n = 1).

Figure 2

CHIR-98014 attenuates the expression of a subset of genes stimulated by treatment with actinomycin D + nutlin-3a (A + N). Heatmaps demonstrating the RNA-seq expression of genes belonging to selected Gene Ontology groups in control cells and the cells exposed as indicated. In order to generate heatmaps control sample measurements were averaged across three comparisons.

Figure 3

APOE is regulated in p53-dependent fashion. (a) Measurement of relative APOE mRNA level in A549 cells exposed to indicated substances or their combinations for 30 h. CPT-camptothecin, ActD-actinomycin D, Nut-nutlin-3a. The mean and standard deviation from three biological replicates performed in triplicate are presented, p values calculated by Brown-Forsythe and Welch’s ANOVA tests (* p < 0.05, ** p < 0.01) (b) Expression of indicated proteins (or p53 phosphorylated on Ser33, Ser37) in p53 knockdown (CRISP) A549 cells or in controls for knockdown. Cells were exposed to A + N or CPT for 30h. (c) Relative APOE mRNA levels in p53 knockdown cells (CRISP-p53) or the control cells for knockdown (CRISP-Con) exposed to A + N or CPT for 30 h. The results represent mean and standard deviation from three independent experiments in triplicate, p values calculated by Brown-Forsythe and Welch’s ANOVA tests (* p < 0.05). (d) The level of APOE mRNAs measured by semi-quantitative real-time PCR of samples isolated from mock-treated A549 cells (Con) or from cells incubated for 30 h as indicated. The mRNA level in the control cells was defined as 1 (means and standard deviations from three biological replicates). The concentration of CHIR-98014 was 1 µM. Statistical differences in expression were as calculated using unpaired t test with Welch’s correction (p = 0.5250).

Figure 4

The p53 protein activates expression of BLNK gene from response element in promoter region. (a) Changes in the levels of BLNK mRNA, measured by semi-quantitative real-time PCR of samples isolated from mock-treated A549 cells (Con) or from cells incubated for 30 h as indicated. The mRNA level in the mock-treated population was defined as 1. Results represent means and standard deviations from three biological replicates, p values calculated by one-way analysis of variance (ANOVA) with uncorrected Fisher’s LSD (* p < 0.05). (b) The expression of BLNK mRNA determined by semi-quantitative RT-PCR in mock-treated control cells (Con) and in the cells exposed to A + N or CPT for 30 h. The A549 cells with knocked-down p53 are marked as “CRISP-p53”, whereas the controls for knock-down are “CRISP-Con”. The results represent means and standard deviations from three biological replicates, p values were calculated by Kruskal-Wallis statistic test with uncorrected Dunn’s test (* p < 0.05). (c) Genome browser (IGV) views of p53 binding peak in the promoter of BLNK. Using ChIP-Atlas tool [15] we imported publically available coverage tracks from two ChIP-Seq experiments aimed at finding p53 binding sites in MCF-7 cell line exposed to ionizing radiation and nutlin (sample ID SRX2924018) or in MCF7 cells treated with nutlin alone (sample ID SRX2060922). The red bar above the tracks marks the promoter sequence cloned in the reporter plasmid. We marked the location of the 3Q p53 response element identified by Tebaldi et al. [16] by an arrow. The 3Q sequence is shown from right to left to match the direction of BLNK gene shown by the genome browser. Note that the 3Q element is located near the transcription start site. The position of nucleotide changes generated in mutant promoter are underlined. The nucleotides in the mutation are shown below. (d) The fold change of the normalized firefly luciferase activity (NFLA) in U-2 OS cells transfected with the reporter vector coding for firefly luciferase under the transcriptional control of BLNK promoter and either empty vector, expression plasmid coding for wild-type p53 or the expression plasmid coding for mutated p53 (Val143Ala). “Mut” prefix indicates the BLNK promoter version with the mutated p53 response element as indicated on panel c. The means and standard deviations from three biological replicates performed in triplicate are shown, p values calculated by Brown-Forsythe and Dunnett’s T3 multiple comparisons. (e) Measurement of relative BLNK mRNA level in NCI-H460 cells exposed to indicated substances or their combinations for 30 h, p values calculated by Brown-Forsythe and Welch’s ANOVA tests (* p < 0.05, ** p < 0.01).

Figure 5

CHIR-98014 promotes activation of executioner caspase of apoptosis in A549 cells exposed to A + N. (a) Western blot showing expression of indicated proteins (or p53 phosphorylated on Sr46 or Ser392) in A549 cells mock-treated (Con) or incubated for 48 h with actinomycin D and nutlin-3a (A + N) with or without addition of CHIR-98014 at indicated concentrations. The level of active caspase-3 is also indicated. (b) Expression of proteins in A549 cells exposed as indicated. The phosphorylation of GSK-3α on Ser21 was assessed using antibody anti-phospho GSK-3α (Ser21). The phosphorylation of GSK-3β on Ser9 was assessed using antibody anti-phospho GSK-3β (Ser9), which also detects phosphorylated GSK-3α (upper band). (c) The expression of cleaved forms of caspase-8 and caspase-9 (arrows) and the activated caspase-3 in A549 cells exposed as indicated. In all experiments from this figure cells were incubated with the substances for 48 h.

Figure 6

Flow cytometry analysis confirms ability of CHIR-98014 to induce apoptosis in cells exposed to A + N. Cytometric analysis of cell populations mock-treated (Con) or exposed to A + N, A + N with CHIR-98014 or CHIR-98014 alone for 48 h. Viable cells are 7-AAD (7-Amino-Actinomycin) negative and PE Annexin V negative; cells in early apoptosis are PE Annexin V positive and 7-AAD negative, whereas late apoptotic or dead cells are both PE Annexin V and 7-AAD positive. The graph below shows the frequency of indicated cell types calculated from three biological repeats (p values calculated by Brown-Forsythe and Welch’s ANOVA tests (** p < 0.01).

Figure 7

Genes response to the treatment with CHIR-98014 is not uniform. Changes in the levels of mRNAs of indicated genes, measured by semi-quantitative real-time PCR of samples isolated from mock-treated A549 cells (Con) or from cells incubated for 30 h as indicated. The mRNA level in the mock-treated population was defined as 1. The concentration of CHIR-98014 was 1 µM. Results represent means and standard deviations from three biological replicates. Statistical differences in expression were calculated using unpaired t test with Welch’s correction (* p < 0.05, ** p < 0.01, **** p < 0.0001).

Figure 8

CHIR-98014 promotes activation of caspase-3 and represses TREM2 in cell lines other than A549. (a,b) Two lung cancer cell lines NCI-H460 and NCI-H292 with wild-type p53 were treated as indicated. The expression of selected proteins or their forms phosphorylated on indicated amino acids was examined by Western blotting. (c,d) The expression of active caspase-3 in cells exposed to anticancer drugs acting alone or in combination with CHIR-98014. The following concentrations of drugs were used: camptothecin (CPT) 1 µM, paclitaxel (PTX)-5 µM, cisplatin (CisPt)–20 µM, etoposide (ETO)–20 µM, CHIR-98014 -1 µM. (e) The expression of active caspase-3 and phosphorylated forms of GSK-3α or GSK-3β in p53-null NCI-H1299 cells exposed as indicated. The concentration of CHIR-98014 was 2 µM.

Figure 9

Downregulation of GSK-3 does not mimic the effect of CHIR-98014 on TREM2 expression or caspase-3 activation. (a) The expression of selected proteins in cells with the expression of either GSK-3α, GSK-3β or p53 downregulated by CRISPR/Cas9 technology. Control cells for knockdown were also used. The cells were either mock-treated (U) or exposed to A + N for 48 h. (b) Expression of TREM2 and active caspase-3 in cells with double knockdown of GSK3A and GSK3B. The cells with GSK3B expression knocked-down by CRISPR/Cas9 technology or the controls for knockdown were transduced either with lentivirus producing the shRNAs directed against GSK3A mRNA or with the control lentivirus. We obtained four cell lines as demonstrated. One of the cell lines had strong downregulation of both GSK3A and GSK3B. The cells were either mock-treated (U) or exposed to A + N for 48 h.

Figure 10

In A549 cells p53 participates in the induction of IRF1 transcription factor. (a) Western blot demonstrating expression of p53, its form phosphorylated on Ser37 and of IRF1 protein in A549 cells treated as indicated for 48 h. (b) Measurement of relative IRF1 mRNA level in A549 cells exposed to indicated substances or their combinations for 30 h. Statistical differences in expression was calculated by Kruskal-Wallis statistic test with uncorrected Dunn’s test (* p < 0.05) (c) Relative IRF1 mRNA levels in p53 knockdown cells (CRISP-p53) or the control cells for knockdown (CRISP-Con) exposed to A + N or CPT for 30 h (p values were calculated by Kruskal-Wallis statistic test with uncorrected Dunn’s test (* p < 0.05, ** p < 0.01). (d) Expression of indicated proteins (or p53 phosphorylated on Ser33 or Ser37) in p53 knockdown (CRISP) A549 cells or in controls for knockdown. Cells were exposed to A + N or CPT for 30h. (e) Expression of p53, IRF1 and p21 in U-2 OS cells with p53 expression knocked-down by CRISPR/Cas9 method employed to knock down p53 in A549 cells.

Figure 11

A subset of innate immunity genes can be activated by A + N without the assistance of p53. (a) Measurement of relative IFIT2 mRNA level in p53 knockdown cells (CRISP-p53) or the control cells for knockdown (CRISP-Con) exposed to A + N or camptothecin (CPT) for 30 h. Statistical differences in expression were calculated by Kruskal-Wallis statistic test with uncorrected Dunn’s test. (b) Expression of indicated proteins (or p53 phosphorylated on Ser37) in p53 knockdown (CRISP) A549 cells or in controls for knockdown. Cells were exposed to A + N or CPT for 48 h. (c) The expression of indicated proteins in clones (#4 and #7) of p53 knockout cells from A549 cell line. The control clone for knockout (#1) as well as the parental A549 cell line were used as controls. The cells were mock-treated (C) or exposed to A + N for 48 h. We presented enlarged image of p53 blot to demonstrate that no protein versions with deletion or insertion are expressed.

Figure 12

CHIR-98014 attenuates the expression of innate immunity genes stimulated by A + N without the assistance of p53. (a) Relative expression of IFIT2 mRNA in cells exposed as demonstrated for 30 h (CHIR, 1µM). Statistical differences in expression were calculated using unpaired t test with Welch’s correction (** p < 0.01). (b) The expression of indicated proteins in A549 cells exposed to A + N or A + N with various concentrations of CHIR-98014 for 48 h. (c) The expression of p53, IFIT3 and IFIH1 proteins in cells with wild-type p53 or expressing no p53. The cells were exposed to A + N or camptothecin (CPT) for 48 h. (d) The expression of the innate immunity proteins in p53-null NCI-H1299 cell line exposed as indicated (CHIR, 2 µM).

Figure 13

A schematic representation of the two hypotheses (a,b) generated by the data presented in this paper. There are two groups of innate immunity genes activated by A + N − one group is stimulated by p53 and the other is activated independently from p53. The other group must be under the control of a transcription factor, which is currently unidentified (uTF). The outcome of activity of both transcription regulators (p53 and uTF) is modulated by a kinase inhibited by CHIR-98014. This kinase plays crucial role in gene regulation in stress conditions elicited by A + N. In principle, the kinase can act upstream (a) or downstream (b) from p53 or uTF.