Induction of epidermal growth factor receptor expression by Epstein-Barr virus latent membrane protein 1 C-terminal-activating region 1 is mediated by NF-kappaB p50 homodimer/Bcl-3 complexes

Abstract

The Epstein-Barr virus (EBV) is associated with the development of numerous malignancies, including the epithelial malignancy nasopharyngeal carcinoma (NPC). The viral oncoprotein latent membrane protein 1 (LMP1) is expressed in almost all EBV-associated malignancies and has profound effects on gene expression. LMP1 acts as a constitutively active tumor necrosis factor receptor and activates multiple forms of the NF-kappaB family of transcription factors. LMP1 has two domains that both activate NF-kappaB. In epithelial cells, LMP1 C-terminal activating region 1 (CTAR1) uniquely activates p50/p50-, p50/p52-, and p65-containing complexes while CTAR2 activates canonical p50/p65 complexes. CTAR1 also uniquely upregulates the epidermal growth factor receptor (EGFR). In NPC, NF-kappaB p50/p50 homodimers and the transactivator Bcl-3 were detected on the EGFR promoter. In this study, the role of NF-kappaB p50 and Bcl-3 in LMP1-mediated upregulation of EGFR was analyzed. In LMP1-CTAR1-expressing cells, chromatin immunoprecipitation detected p50 and Bcl-3 on the NF-kappaB consensus sites within the egfr promoter. Transient overexpression of p50 and Bcl-3 increased EGFR expression, confirming the regulation of EGFR by these factors. Treatment with p105/p50 siRNA effectively reduced p105/p50 levels but unexpectedly increased Bcl-3 expression and levels of p50/Bcl-3 complexes, resulting in increased EGFR expression. These data suggest that induction of p50/p50/Bcl-3 complexes by LMP1 CTAR1 mediates LMP1-induced EGFR upregulation and that formation of the p50/p50/Bcl-3 complex is negatively regulated by the p105 precursor. The distinct forms of NF-kappaB that are induced by LMP1 CTAR1 likely activate distinct cellular genes.

FIG. 1.

NF-κB p50 and Bcl-3 immunoprecipitate on the EGFR promoter. (A) EGFR and LMP1 expression was examined by immunoblotting with C33A cells stably expressing Myc-tagged LMP1 or deletion mutant 1-231 or Δ187-351 with anti-c-myc and anti-EGFR. The bottom panel is a loading control. (B) A chromatin immunoprecipitation of p50, Bcl-3, p100/p52, p65, and RelB on the egfr promoter was performed with C33A cells stably expressing LMP1 or deletion mutant 1-231 or Δ187-351. Lanes 1 and 16 are DNA markers, lanes 2 and 17 are water PCR controls, lanes 3 and 18 are genomic DNA from C33A cells, lanes 4 to 7 are immunoprecipitated in the absence of antibody, lanes 8 to 11 are immunoprecipitated with anti-p105/p50 (α-p105/p50), lanes 12 to 15 are immunoprecipitated with anti-Bcl-3 (α-Bcl-3), lanes 19 to 22 are input DNA from all four stable cell lines, lanes 23 to 26 are immunoprecipitated with anti-p100/p52 (α-p100/p52), lanes 27 to 30 are immunoprecipitated with anti-p65 (α-p65), and lanes 31 to 34 are immunoprecipitated with anti-RelB (α-RelB).

FIG. 2.

Transient expression of p50 and Bcl-3 increases EGFR protein. C33A cells were transiently transfected with p50 and Bcl-3, and levels of EGFR were examined by immunoblotting. Equal loading was confirmed with anti-β-actin. EGFR levels were measured by densitometry of the immunoblot in the top panel. The Image J 1.32j program was used to calculate the intensity of each band from one Western blot. Pixel densities are listed above their corresponding bands. The blot is representative of five independent experiments.

FIG. 3.

Increased EGFR expression requires NF-κB activation. Stable cell lines expressing the pCDNA vector control or 1-231 were transiently transfected with the vector control, DN NIK, IKKα, or IKKβ. RNA was harvested and used for quantitative reverse transcription-PCR using primer sets for β-actin and EGFR. The graph shows the n-fold change from the level for pCDNA3 stable cells transfected with the vector control and represents the average from separate experiments. All samples are normalized to vector control results, and EGFR levels were additionally normalized to the sample's corresponding β-actin level. The bars represent the n-fold change in EGFR levels from levels for pCDNA3 cells transfected with vector control. Each sample was performed in triplicate, and the standard error is indicated. The graph shown is representative of three independent experiments.

FIG. 4.

DN IKKβ decreases nuclear p50 homodimers and EGFR protein. C33A cells transiently transfected with pCDNA3 vector control, wild type IKKβ (WT IKKβ), or double-point-mutant (SS 177, 181 AA) DN IKKβ (DN IKKβ) were tested for nuclear NF-κB complexes by EMSA. Levels of EGFR, LMP1, and IKKβ were examined by immunoblotting. The LMP1 expression panel is a composite from two different gels. Equal loading was confirmed by using a loading control (bottom panel). Lane 1 is a probe-alone control for the EMSA. Lanes 2 to 4 are pCDNA3 stable cells transiently transfected with pCDNA3, wt IKKβ, or DN IKKβ; lanes 5 to 7 are LMP1 stable cells transfected with pCDNA3, wt IKKβ, or DN IKKβ; lanes 8 to 10 are 1-231 stable cells transfected with pCDNA3, wt IKKβ, or DN IKKβ, and lanes 11 to 13 are Δ187-351 stable cells transfected with pCDNA3, wt IKKβ, or DN IKKβ. Arrows identify the complexes. The EMSA and Western blots are each representative of four independent experiments.

FIG. 5.

NF-κB1 siRNA negatively regulates p105/p50 expression and increases Bcl-3 and EGFR expression. (A) C33A cells were transfected with 100 pmol irrelevant siRNA, 100 pmol NF-κB1 siRNA, 100 pmol Bcl-3 siRNA, or 50 pmol NF-κB1 and Bcl-3 siRNAs. Knockdown was confirmed by immunoblotting. (B) C33A cells stably expressing pCDNA3 vector control or 1-231 were transfected with 100 pmol irrelevant siRNA, 100 pmol NF-κB1 siRNA, 100 pmol Bcl-3 siRNA, or 50 pmol NF-κB1 and Bcl-3 siRNAs. NF-κB and EGFR protein levels were examined by immunoblotting. The Image J 1.32j computer program was used to calculate the intensity of each EGFR band from one Western blot. Pixel density values are listed above the corresponding bands. Units are arbitrary. The blot is representative of five independent experiments.

FIG. 6.

NF-κB1 siRNA increases EGFR RNA. C33A cells stably expressing vector control or 1-231 were transfected with 100 pmol irrelevant siRNA (+−), 100 pmol NF-κB1 siRNA (NF-κB1), 100 pmol Bcl-3 siRNA (Bcl-3), or 50 pmol NF-κB1 and Bcl-3 siRNAs (NF-κB1+Bcl-3). β-Actin and EGFR levels were examined with quantitative reverse transcription-PCR. All samples were normalized to results for 1-231 stable cells transfected with irrelevant siRNA, and each EGFR PCR was further normalized to the β-actin level in its corresponding sample. The graph shows n-fold change in EGFR RNA level over that for 1-231 with irrelevant siRNA and is a representative result from five independent experiments.

FIG. 7.

Increased Bcl-3 coimmunoprecipitates with NF-κB1 after transfection with NF-κB1 siRNA. (A) Coimmunoprecipitation of p105/p50 with p100/p52 and Bcl-3 was measured after knockdown with 100 pmol irrelevant siRNA, 100 pmol NF-κB1 siRNA, 100 pmol Bcl-3 siRNA, or 50 pmol NF-κB1 and Bcl-3 in C33A cells stably expressing pCDNA3 or 1-231. Stable cells were transfected with the siRNA indicated above the immunoblot. Lanes 1 and 2 are a direct load of protein lysates from pCDNA3 and 1-231 stable cells. Lanes 3 to 7 are immunoprecipitated with rabbit immunoglobulin G (IgG) isotype control, lanes 8 to 12 are immunoprecipitated with anti-p100/p52 (α-p100/p52), and lanes 13 to 17 are immunoprecipitated with anti-Bcl-3 (α-Bcl-3). All immunoprecipitates were immunoblotted with anti-p50. (B) Coimmunoprecipitation of Bcl-3 with anti-p105/p50 (α-p105/p50) and anti-p100/p52 was measured after knockdown with siRNAs as described for panel A with C33A cells stably expressing pCDNA3 or 1-231. Stable cells were transfected with the siRNA indicated above the immunoblot. Lanes 1 and 2 are direct loads of protein lysates from pCDNA3 and 1-231 stable cells. Lanes 3 to 7 are immunoprecipitated with goat IgG isotype control, lanes 8 to 12 are immunoprecipitated with anti-p105/p50, and lanes 13 to 17 are immunoprecipitated with anti-p100/p52. The complexes were immunoblotted with anti-Bcl-3. (C) Coimmunoprecipitation of anti-p100/p52 with anti-p105/p50 and Bcl-3 was measured after siRNA knockdown as described for panel A in pCDNA3 and 1-231 stable C33A cells. Stable cells were transfected with the siRNA indicated above the immunoblot. Lanes 1 and 2 are a direct load of protein lysates from pCDNA3 and 1-231 stable cells. Lanes 3 to 7 are immunoprecipitated with mouse IgG isotype control, lanes 8 to 12 are immunoprecipitated with anti-p105/p50, and lanes 13 to 17 are immunoprecipitated with anti-Bcl-3. The immunoprecipitates were immunoblotted with anti-p52.