Transcription factor nuclear factor erythroid-2 mediates expression of the cytokine interleukin 8, a known predictor of inferior outcome in patients with myeloproliferative neoplasms

Abstract

The transcription factor nuclear factor erythroid-2 is over-expressed in patients with myeloproliferative neoplasms irrespective of the presence of the JAK2(V617F) mutation. Our transgenic mouse model over-expressing nuclear factor erythroid-2, which recapitulates many features of myeloproliferative neoplasms including transformation to acute myeloid leukemia, clearly implicates this transcription factor in the pathophysiology of myeloproliferative neoplasms. Because the targets mediating nuclear factor erythroid-2 effects are not well characterized, we conducted microarray analysis of CD34(+) cells lentivirally transduced to over-express nuclear factor erythroid-2 or to silence this transcription factor via shRNA, in order to identify novel target genes. Here, we report that the cytokine interleukin 8 is a novel target gene. Nuclear factor erythroid-2 directly binds the interleukin 8 promoter in vivo, and these binding sites are required for promoter activity. Serum levels of interleukin 8 are known to be elevated in both polycythemia vera and primary myelofibrosis patients. Recently, increased interleukin 8 levels have been shown to be predictive of inferior survival in primary myelofibrosis patients in multivariate analysis. Therefore, one of the mechanisms by which nuclear factor erythroid-2 contributes to myeloproliferative neoplasm pathology may be increased interleukin 8 expression.

Figure 1.

Identification of novel NF-E2 target genes. (A) Experimental Design. Purified peripheral blood CD34⁺-cells from healthy donors, were transduced with either the empty lentiviral vector, pLeGo-iG-hU6 (top), or pLEGO-iG-hU6 engineered to express human NF-E2 cDNA (middle) or a shRNA silencing NF-E2 expression (bottom). 84 h following transduction, GFP⁺-cells were FACS-sorted, RNA extracted, and used for microarray analysis on an Affymetrix Human Exon Array. Four separate pools of CD34⁺ cells were transduced in parallel with each of the three vectors, yielding a total of 12 microarrays for analysis, four independent arrays for each vector. (B) NF-E2 mRNA expression in lentivirally transduced CD34⁺ cells. The RNA used for microarray analysis [see (A) above] was quantitated for NF-E2 expression by qRT-PCR. Results are reported as copies NF-E2 expressed per 10³ copies of the beta-2-microglobulin housekeeping gene; n = 4 each; *P<0.05. (C) IL-8 mRNA expression in microarray analysis. (Top) Following data analysis, mRNA expression values are normalized, assigned to a color on a green-red scale and displayed as a color field. (Bottom) IL-8 expression for each of the four independent microarrays in the three conditions is depicted. (D) IL-8 mRNA expression by qRT-PCR. mRNA expression of IL-8 and the housekeeping gene beta-2-microglobulin (β-2-M) was quantified in CD34⁺ cells by qRT-PCR. IL-8 expression was normalized to β-2-M expression and the expression in empty virus transduced CD34⁺ cells set at 1. Values are reported as fold change over empty virus transduced cells and depict mean and SD of 4 individual experiments. *P≤0.05 by paired t-test.

Figure 2.

IL-8 protein expression following NF-E2 expression in U937 cells. U937 cells were transduced either with the empty lentivirus or with a virus expressing the NF-E2 cDNA. (A) Total cell extracts were interrogated with an antibody against NF-E2 (top). Equal loading was assured by reprobing with an antibody against GAPDH (bottom). (B-C) Lentivirally transduced U937 cells were cultured for 4 h in medium containing brefeldin A (10 ng/mL) to block the secretion of IL-8. Subsequently, intracellular IL-8 was stained and analyzed by FACS. (B) Representative histogram of n=5, depicting intracellular IL-8 staining of untreated cells (control) as well as cells infected with empty vector (empty) or with NF-E2 expressing virus (NF-E2), as indicated. (C) Mean and SD of the mean fluorescence intensity (MFI) of IL-8 staining in n=5 independent experiments are displayed, *P≤0.05 by one-way ANOVA with post-hoc Tukey’s Multiple Comparison Test. (D) Lentivirally transduced U937 cells were cultivated for 24 h and the supernatant subsequently removed for detection of IL-8 protein secretion by ELISA. Bars depict the mean and SD IL-8 concentration of 3 independent experiments from untreated cells (control) or from cells transduced with empty virus (empty) or with virus expressing the NF-E2 cDNA.

Figure 3.

IL-8 protein expression following knockdown of NF-E2 in MPN cell lines. (A) UKE1 cells (B) SET2 cells and (C) HEL cells were transduced with the empty lentiviral vector as well as a virus expressing a shRNA against NF-E2 (shRNA-NF-E2). 24 h after transduction, the supernatant was removed and IL-8 concentration determined by ELISA (top). Bars depict the mean and SD IL-8 concentration of 3 independent experiments measured in duplicate. *P≤ 0.05 by t-test. Total cell extracts were interrogated with an antibody against NF-E2 (top). Equal loading was assured by reprobing with an antibody against GAPDH (bottom).

Figure 4.

NF-E2 transactivation of the IL-8 promoter. (A) Schematic representation of the IL-8 promoter/enhancer reporter gene construct. The construct includes a 1 kb upstream enhancer region, from −5.5kb to −4.5kb, and the proximal 262 bp of the IL 8 promoter. NF-E2 sites predicted by in silico analysis are indicated by open circles. (B) The IL-8 promoter/enhancer luciferase vector was transfected into 293T cells together with plasmids encoding NF-E2 and/or MafG as indicated. Bar graphs represent the mean and SD of at least 5 independent experiments, each performed in duplicate. Data are shown normalized to the IL-8 reporter construct co-transfected with MafG alone, which was set at 1. ***P<0.001 by one-way ANOVA with post-hoc Tukey’s Multiple Comparison Test. (C) The potential NF-E2 binding sites were altered by site-directed mutagenesis. Presence of the mutations is indicated by crosses in the circles. Sequences of the wt and the mutated sites are depicted below (mutated bases in bold). The IL-8 reporter constructs shown were co-transfected into 293T cells together with expression plasmids for NF-E2 and MafG. Bar graphs represent the mean and SD of 4 independent experiments, each performed in duplicate. Data are normalized to the wt IL-8 reporter construct transfected with MafG alone, which was set at 1. ***P<0.001, **P<0.01, *P<0.05 by one-way ANOVA with post-hoc Tukey’s Multiple Comparison Test.

Figure 5.

Chromatin immunoprecipitation (ChIP) analysis of NF-E2 binding sites on the IL-8 promoter and enhancer. (A) Schematic representation of the primers used to ChIP the two NF-E2 binding sites. (B) HEL cell lysates were chromatin immunoprecipitated (ChIPed) either with an antibody to NF-E2 or with an unrelated IgG control, as indicated. ChIPed DNA was amplified by PCR using either primers covering the NF-E2 binding sites in the IL-8 promoter and enhancer or with control primers from other regions within the IL8 locus and from the myogenin promoter, as indicated. In lane 3 (Input), a 1:50 dilution of the input DNA was used, lane 4 (Control) shows control PCR reactions without DNA.