p100 Deficiency is insufficient for full activation of the alternative NF-κB pathway: TNF cooperates with p52-RelB in target gene transcription

Abstract

Background: Constitutive activation of the alternative NF-κB pathway leads to marginal zone B cell expansion and disorganized spleen microarchitecture. Furthermore, uncontrolled alternative NF-κB signaling may result in the development and progression of cancer. Here, we focused on the question how does the constitutive alternative NF-κB signaling exert its effects in these malignant processes.

Methodology/principal findings: To explore the consequences of unrestricted alternative NF-κB activation on genome-wide transcription, we compared gene expression profiles of wild-type and NF-κB2/p100-deficient (p100(-/-)) primary mouse embryonic fibroblasts (MEFs) and spleens. Microarray experiments revealed only 73 differentially regulated genes in p100(-/-) vs. wild-type MEFs. Chromatin immunoprecipitation (ChIP) assays showed in p100(-/-) MEFs direct binding of p52 and RelB to the promoter of the Enpp2 gene encoding ENPP2/Autotaxin, a protein with an important role in lymphocyte homing and cell migration. Gene ontology analysis revealed upregulation of genes with anti-apoptotic/proliferative activity (Enpp2/Atx, Serpina3g, Traf1, Rrad), chemotactic/locomotory activity (Enpp2/Atx, Ccl8), and lymphocyte homing activity (Enpp2/Atx, Cd34). Most importantly, biochemical and gene expression analyses of MEFs and spleen, respectively, indicated a marked crosstalk between classical and alternative NF-κB pathways.

Conclusions/significance: Our results show that p100 deficiency alone was insufficient for full induction of genes regulated by the alternative NF-κB pathway. Moreover, alternative NF-κB signaling strongly synergized both in vitro and in vivo with classical NF-κB activation, thereby extending the number of genes under the control of the p100 inhibitor of the alternative NF-κB signaling pathway.

Figure 1. Expression profiling of wild-type and p100^−/− MEFs.

(A) Western blot analysis of cytoplasmic and nuclear protein extracts (20 µg/sample) were analyzed for the presence of NF-κB family members p100/p52, RelB, and RelA in wild-type (+/+) and in p100^−/− (−/−) MEFs. As cytoplasmic loading control S6 ribosomal protein and as nuclear loading control RNA Pol II was assayed. (B) Increased κB DNA-binding activity in nuclear extracts from p100^−/− (−/−) compared to wild-type MEFs (+/+). Five µg protein extract per cell line were incubated with a radioactively labeled Igκ oligo and analyzed by EMSA. Supershift analysis was performed using pre-immune serum (p.i.), anti-RelA (α-RelA), anti-RelB (α-RelB), and anti-p52 antibodies (α-p52). Super-shifted RelB and p52 complexes are indicated by arrow and arrowhead, respectively. (C) Heatmap displaying fold change values observed in p100^−/− vs. wild-type cells. The color code indicates the fold change values between +6 fold up- (red) and −20 fold downregulation (green). Each horizontal line on the heatmap corresponds to one gene. Genes labeled with blue boxes on the left were verified by qRT-PCR. Gene symbols and abbreviations of GO terms are displayed on the right. CA, Cytokine activity; GPCRB, G-protein-coupled receptor binding; IGFB, insulin-like growth factor binding; ER, extracellular region; ESO, extracellular structure organization and biogenesis; D/M, developmental process; B/ND, forebrain development and nervous system development; CG/S, regulation of cell size; IR/RES, immune response and response to external stimulus; M/L/T, locomotory behavior. (D) Significantly regulated Gene Ontology terms with respective gene numbers.

Figure 2. Lack of p100 results in enhanced p52 and RelB binding to NF-κB target sites in the Enpp2/Atx promoter.

(A) Schematic view of the Enpp2/Atx promoter. In silico analysis of the Enpp2/Atx promoter (−2485 bp upstream of the ATG codon) with MatInspector and TFSearch softwares revealed four potential NF-κB sites: ATX4 at position −1162 (CGGGGGCTTC), ATX3.2 at position −596 (GGAAGCTCCC), ATX3.1 at position −529 (AGGGTCATTCC), and ATX1 at position −375 (GGGAAATTCT). ATX3.2 and ATX1 κB target site have been previously described using the software Footer . (B) In vitro binding of p52 and RelB to κB target sites in the Enpp2/Atx promoter was determined by the TransAM Flexi NF-κB Family Transcription Factor Assay (Active Motif). Three independent experiments were performed. Data are presented as mean values ± SD. Statistically significant differences are indicated by * (Student's t-test, P≤0.05). (C) In vivo binding of p52 and RelB to κB target sites in the Enpp2/Atx promoter. An unrelated site (ATXneg) served as a negative control. For ChIP experiments, the Express Magnetic Chromatin Immunoprecipitation Kit (Active Motif) was employed according to the manufacturer's instructions. ChIP assays for RNA polymerase II binding to the Gapdh promoter were used for normalization as input control. Data are presented as mean values ± SD, n = 3. Statistically significant differences are indicated by * (Student's t-test, P≤0.05).

Figure 3. Gene expression analysis of wild-type and p100^−/− spleens.

(A) Changes in mRNA levels of selected genes were analyzed by qRT-PCR using RNA samples isolated from spleens of four wild-type and four p100^−/− animals. Data are expressed as mean values ± SD. Differences were analyzed by Welch tests. P≤0.05 was considered significant. All genes shown were significantly differentially regulated in wild-type compared to mutant mice. (B) Western blot analysis of total protein extracts from wild-type and p100^−/− spleens (30 µg/sample; protein extracts from three independent spleens are shown for each genotype) for the presence of p100/p52, ENPP2/ATX, and TRAF1. p52 wild-type protein and p52 protein resulting from the knock-in of the stop codon migrated with different speed due to a few amino acids difference in length. The asterisk indicates an unspecific signal. The membrane was probed for β-actin as a loading control.

Figure 4. TNF synergizes with the lack of p100 in the induction of target gene expression.

(A) Western blot of NF-κB family members in cytoplasmic and nuclear protein extracts (15 µg/sample) from unstimulated (0 h) and TNF-stimulated (6 and 24 h, 20 ng/ml recombinant murine TNF) wild-type and p100^−/− MEFs. As cytoplasmic and nuclear loading controls β-actin and RNA Pol II were assayed, respectively. (B) Changes in mRNA levels of selected genes were analyzed by qRT-PCR. From 16 analyzed genes, 12 responded synergistically to TNF and the lack of p100 whereas only four genes responded similarly to TNF treatment of wild-type (grey squares) and p100^−/− MEFs (black circles; see also Figure S4). qRT-PCR data represent n = 3 independent TNF stimulation experiments and are expressed as mean values ± SD. Differences between wild-type and p100^−/− MEFs at each time-point were analyzed by Welch tests. P≤0.05 was considered significant (*).

Figure 5. Lack of p100 cooperates with TNF to increase binding of p52 and RelB to the κB elements in the Enpp2/Atx promoter.

Wild-type and p100^−/− primary MEFs were stimulated with recombinant murine TNF (20 ng/ml) for 6 and 24 h or were left untreated and subsequently nuclear proteins were isolated. In vitro binding of NF-κB subunits p52 (A) and RelB (B) to the κB target sites in the Enpp2/Atx promoter was determined as in Figure 2. Data are presented as mean values ± SD from n = 3 independent experiments. Significant differences (P≤0.05) in p52 and RelB binding were calculated by Student's t-test and are indicated (*).

Figure 6. Synergistic regulation of gene expression by the classical (TNF) and the alternative (p100^−/−) NF-κB pathway in vivo.

Changes in mRNA levels of selected genes contributing to chemotaxis, lymphocyte homing, and cell adhesion were analyzed by qRT-PCR using RNA samples isolated from spleens of control (wild-type for the Nfkb2 locus and heterozygous for the TnfLta locus), p100^−/− (deficient for the p100/Nfkb2 and wild-type for the TnfLta locus), and p100^−/− Tnf^−/− (deficient for both the p100/Nfkb2 and the TnfLta locus) animals. Four independent experiments were performed representing n = 6 control mice, n = 5 p100^−/− mice, and n = 4 p100^−/− Tnf^−/− mice. Data are presented as mean values ± SD. Significant differences (P≤0.05) were calculated by Student's t-test and are indicated (*). * P≤0.05; ** P≤0.01; n.s. = not significant.