The zinc finger transcription factor ZXDC activates CCL2 gene expression by opposing BCL6-mediated repression

Abstract

The zinc finger X-linked duplicated (ZXD) family of transcription factors has been implicated in regulating transcription of major histocompatibility complex class II genes in antigen presenting cells; roles beyond this function are not yet known. The expression of one gene in this family, ZXD family zinc finger C (ZXDC), is enriched in myeloid lineages and therefore we hypothesized that ZXDC may regulate myeloid-specific gene expression. Here we demonstrate that ZXDC regulates genes involved in myeloid cell differentiation and inflammation. Overexpression of the larger isoform of ZXDC, ZXDC1, activates expression of monocyte-specific markers of differentiation and synergizes with phorbol 12-myristate 13-acetate (which causes differentiation) in the human leukemic monoblast cell line U937. To identify additional gene targets of ZXDC1, we performed gene expression profiling which revealed multiple inflammatory gene clusters regulated by ZXDC1. Using a combination of approaches we show that ZXDC1 activates transcription of a gene within one of the regulated clusters, chemokine (C-C motif) ligand 2 (CCL2; monocyte chemoattractant protein 1; MCP1) via a previously defined distal regulatory element. Further, ZXDC1-dependent up-regulation of the gene involves eviction of the transcriptional repressor B-cell CLL/lymphoma 6 (BCL6), a factor known to be important in resolving inflammatory responses, from this region of the promoter. Collectively, our data show that ZXDC1 is a regulator in the process of myeloid function and that ZXDC1 is responsible for Ccl2 gene de-repression by BCL6.

Keywords: Chemokine; Gene regulation; Inflammation; Zinc finger transcription factor.

Figure 1. Expression of ZXDC isoforms during PMA-induced differentiation of U937

A, Time course of PMA treatment of U937 shows that ZXDC isoforms are differentially expressed during differentiation as detected by Western blot. The bands labeled ZXDC1 represent the sumoylated (upper) and unmodified (lower) forms. B, Schematic of ZXDC gene and exons included in ZXDC1 and ZXDC2. ZXDC2 is derived from alternative transcriptional termination site usage within the sixth intron. C, Relative band intensities from western blots (as shown in A) for ZXDC1 (black circles) and ZXDC2 (grey bars) were measured by densitometry and normalized to intensities obtained for β-Actin which was used as a loading control. D, Time course of PMA treatment analyzed by qRT-PCR incorporating primers specific for ZXDC1 (black circles) and ZXDC2 (grey squares). Values reported represent the means (± s.d.) of three independent experiments. ‡P < 0.05.

Figure 2. Effect of ZXDC level manipulation on monocyte-specific surface marker expression in U937

U937 cells were transduced with lentiviral vectors expressing (A) ZXDC1, or (B) ZXDC2, or (C) shRNA targeting ZXDC1/2. Cells were treated with vehicle (dashed lines) or PMA (solid lines) for 72h and expression of indicated surface antigens was assessed by flow cytometry. Black traces: transduced with control lentivirus; Gray traces: transduced with expression lentivirus as indicated. Only transduced cells (GFP positive) were included in the analyses. Mean fluorescence intensities obtained for the indicated surface antigens from individual independent experiments were normalized to compare the effects of ZXDC1/2 overexpression (D; Black bars: transduction with control lentivirus; light grey bars: transduction with ZXDC1 overexpressing lentivirus; dark grey bars: transduction with ZXDC2 overexpressing lentivirus), and knockdown of ZXDC1/2 (E; Black bars: transduction with control shRNA lentivirus; grey bars: transduction with ZXDC1/2 specific shRNA lentivirus). Bars represent the means of n (indicated) biological replicates +/− s.d.. **P < 0.005, *P < 0.01, ‡P < 0.05.

Figure 3. Differential gene expression caused by knockdown of ZXDC in U937 cells treated with PMA for 72 hours as identified by microarray

A, Probes exhibiting significant expression changes upon PMA treatment. B, Detection of probes up-regulated by PMA whose expression is significantly altered by ZXDC knockdown. The solid line represents no change (unity); dashed lines represent ± three times the RMSD of all probes up-regulated by PMA. Probes in red represent those exhibiting > 4-fold PMA activation and significant impairment upon ZXDC knockdown. C, Genes whose activation by PMA was impaired by knockdown of ZXDC in U937 cells. Results are presented as the means of biological triplicates.

Figure 4. Ccl2 gene expression was activated by over-expression of ZXDC1 but not ZXDC2

A, U937 were transiently transduced with lentivirus expressing ZXDC1, ZXDC2 or control as indicated 96 hours prior to treatment with vehicle (black bars) or PMA (gray bars) for 72h and analyzed by qRT-PCR to measure Ccl2 expression. B, U937 were transiently transduced with lentiviral vectors expressing either shRNA against ZXDC1/2 (shZXDC, grey bars) or luciferase (shControl, black bars) for 96 hours prior to treatment with vehicle or PMA for 72h and analyzed by qRT-PCR to measure Ccl2 expression. Bars represent the means of three biological replicates +/− s.d. **P < 0.005, *P < 0.01, † < 0.02.

Figure 5. Identification of an overlapping ZXDC1 and BCL6 responsive element in the Ccl2 gene promoter

A, Schematic of Ccl2 Gaussia-luciferase reporter constructs used to generate stable U937 reporter cell lines. Coordinates are with respect to the transcriptional start site of Ccl2. B, U937 reporter cell lines were transduced with lentivirus expressing ZXDC1 (gray bars) or vector (black bars) followed by luciferase assay 48 hours later. C. U937 reporter cell lines were transduced with lentivirus expressing BCL6 (gray bars) or vector (black bars) and analyzed as in (B). D, U937 reporter cell line stable for the (−3158/+73) reporter construct were transduced with lentivirus expressing BCL6 alone or with increasing amounts of lentivirus expressing ZXDC, as indicated, and analyzed as in (B). Bars represent the means of three biological replicates +/− s.d. **P < 0.005, *P < 0.01.

Figure 6. ZXDC1 interacts with BCL6 via its zinc finger (ZF) domain region

A, Raji cell lysates were subjected to immunoprecipitation with the indicated antibodies followed by Western blot with anti-BCL6. The first lane contains cellular lysate. B, Schematic representation of BCL6 and domain deletion derivatives used in transient transfection experiments. C, HEK293 cells were transfected with the indicated expression plasmids, followed by immunoprecipitation with anti-FLAG (ZXDC) or anti-HA (BCL6). Western blot was performed on the immunoprecipitates as indicated; ns: non-specific band, hc: IgG heavy chain.

Figure 7. Loss of ZXDC function promotes increased BCL6 occupancy of the Ccl2 distal regulatory element

A, Inspection of the Ccl2 promoter distal response region (DRR) sequence revealed the presence of a putative BCL6 binding site (grey highlight) co-localized with an opposing, well established pair of NF-κB binding sites (black higlights) (Ping et al., 1999; Ping et al., 1996). Location of ChIP amplicons are indicated in relation to the putative BCL6 binding site. B, U937 cells stably expressing a shRNA targeting either ZXDC1/2 (shZXDC) or a control shRNA (shControl) were treated with vehicle or PMA for 72 hours. Chromatin was prepared and subjected to immunoprecipitation with anti-BCL6 or control IgG antibody followed by qRT-PCR on the isolated DNA with primers specific for amplification of either the “−3kb Enhancer” (grey bars), or a “Control Region” of genomic DNA, approximately 3.5kb usptream of the putative BLC6 binding site (black bars). Data is represented as “Fold IgG”; the amount of anti-BCL6 immunoprecipitated DNA (normalized to input) divided by the amount of normal IgG immunoprecipitated DNA (also normalized to input). C, Analysis of BCL6 occupancy of the −3kb enhancer throughout a 72 hour PMA treatment time course by ChIP. Data was analyzed as in B. D, Measurement of corresponding CCL2 gene expression by qRT-PCR from samples used for ChIP analysis in C. Data is represented as the mean of, at minimum,3 biological replicates +/− s.d. **P < 0.005, *P < 0.01, ‡P < 0.05).

Figure 8. Co-localization of ZXDC1 and BCL6 in the nuclei of U937 increases after PMA induction

U937 were treated with PMA for the indicated times and attached to slides via cytospin. Endogenous ZXDC1/2 and BCL6 proteins were stained with specific antibodies and visualized by confocal fluorescent microscopy.