Linkage between STAT regulation and Epstein-Barr virus gene expression in tumors

Abstract

Epstein-Barr virus (EBV) latency gene expression in lymphoblastoid cell lines is regulated by EBNA2. However, the factors regulating viral expression in EBV-associated tumors that do not express EBNA2 are poorly understood. In EBV-associated tumors, EBNA1 and frequently LMP1 are synthesized. We found that an alternative latent membrane protein 1 (LMP1) promoter, L1-TR, located within the terminal repeats is active in both nasopharyngeal carcinoma and Hodgkin's disease tissues. Examination of the L1-TR and the standard ED-L1 LMP1 promoters in electrophoretic mobility shift assays revealed that both promoters contain functional STAT binding sites. Further, both LMP1 promoters responded in reporter assays to activation of JAK-STAT signaling. Cotransfection of JAK1 or v-Src or treatment of cells with the cytokine interleukin-6 upregulated expression from ED-L1 and L1-TR reporter plasmids. Cotransfection of a dominant negative STAT3 beta revealed that STAT3 is likely to be the biologically relevant STAT for EBNA1 Qp and LMP1 L1-TR promoter regulation. In contrast, LMP1 expression from ED-L1 was not abrogated by STAT3 beta, indicating that the two LMP1 promoters are regulated by different STAT family members. Taken together with the previous demonstration of JAK-STAT activation of Qp driven EBNA1 expression, this places two of the EBV genes most commonly expressed in tumors under the control of the same signal transduction pathway. Immunohistochemical analyses of nasopharyngeal carcinoma tumors revealed that STAT3, STAT5, and STAT1 are constitutively activated in these tumors while STAT3 is constitutively activated in the malignant cells of Hodgkin's disease. We hypothesize that chronic or aberrant STAT activation may be both a necessary and predisposing event for EBV-driven tumorigenesis in immunocompetent individuals.

FIG. 1

LMP1 promoter usage in EBV-associated tumors. (a) Diagram of the LMP1 gene, showing the exon structure and the relative positions of the standard promoter (pED-L1) and the terminal repeat (TR) promoter previously identified in NPC tissue samples (pL1-TR) (45,53). The locations of potential STAT binding sites and of the PCR primers (P1 and P2) used to detect pL1-TR initiated mRNAs are also indicated. (b) Northern blot analysis of LMP1 mRNA isolated from two NPC tissue biopsy specimens. A 3.5-kb RNA indicative of pL1-TR usage was detected in both samples. A 2.8-kb RNA initiating from pED-L1 was also seen in NPC2. (c) Southern blot analysis of RT-PCR products generated using the P1 and P2 primers and RNA isolated from a case of EBV-positive Hodgkin's disease. The blot was incubated with an LMP1-specific ³²P-labeled oligonucleotide probe. The 749-bp product is diagnostic for an LMP1 mRNA initiating from pL1-TR. The 903-bp product was also detected in the absence of the RT reaction and was generated from EBV genomic DNA.

FIG. 2

The ED-L1 and L1-TR LMP1 promoters each contain STAT binding sites. (a) EMSA showing binding of purified, activated STAT1 and STAT4 to ³²P-labeled oligonucleotide probes containing the potential STAT binding sites from the Qp, ED-L1, and L1-TR promoters. The control STAT1 and STAT4 probes contain consensus STAT binding sites. Qp promotes EBNA1 expression in tumors. (b) Competition and supershift EMSAs illustrating the specificity of STAT4 binding to the L1-TR promoter probe. The L1-TR(mt) competitor contains a mutated STAT binding site. The Flag competitor oligonucleotide and antibody were used as controls for nonspecific effects.

FIG. 3

The L1-TR LMP1 promoter is responsive to the STAT activators JAK1 and IL-6 and has reduced activity in JAK mutant cell lines. (a) Reporter assay showing induction of CAT expression from L1-TRp–CAT in HeLa cells cotransfected with a JAK1 expression vector or treated for 48 h with human IL-6 (100 ng/ml). The results shown are an average of three experiments, with the standard deviation indicated. (b) Reporter assay comparing CAT expression from L1-TRp–CAT and L1-TR(mt)p–CAT in parental (2fTGH) versus TYK2 (U1A) and JAK1 (U4A) mutant cells. The STAT binding site is mutated in L1-TR(mt)p–CAT. The promoter in TK-CAT is the non-STAT-regulated herpes simplex thymidine kinase promoter. The results shown are an average of three experiments, with the standard deviation indicated.

FIG. 4

The endogenous L1-TR promoter in B95-8 cells is activated by IL-6. Southern blot analyses of RT-PCR products generated from B95-8 lymphoblastoid cells using the P1 and P2 primers for L1-TRp initiated mRNA (top) or primers for the polymerase III EBER1 RNAs (bottom) are shown. cDNAs were detected using specific ³²P-labeled oligonucleotide probes. Growth of B95-8 cells in medium containing IL-6 increased the amount of the 749-bp L1-TR-initiated LMP1 mRNA but did not affect EBER1 RNA levels. The larger PCR product was also generated in the absence of the RT reaction.

FIG. 5

The EBNA1 Qp and both LMP1 promoters are activated by v-Src-induced signaling and differentially repressed by interference with STAT3 function. Reporter assay in HeLa cells cotransfected with the EBV latency Qp-CAT, ED-L1–CAT, and L1-TR–CAT constructions and either control vector DNA (vector), an expression plasmid for v-Src, or v-Src plus a dominant negative STAT3 inhibitor (STAT3β). The results shown are an average of three experiments, with the standard deviation indicated.

FIG. 6

NPC and Hodgkin's disease Reed-Sternberg cells contain activated nuclear STATs. Immunohistochemical analyses of STAT localization in NPC and Hodgkin's disease tissue samples are shown. (b and d) STAT3 staining in EBV-positive (b) and EBV-negative (d) Hodgkin's disease tissue. The malignant Reed-Sternberg cells are indicated by arrowheads. (a and c) STAT1 staining in EBV-positive (a) and EBV-negative (c) Hodgkin's disease tissue. (e and f) NPC tissue stained for STAT1 (e) and STAT3 (f). STATs were detected using anti-STAT1 and anti-STAT3 primary antibodies (Santa Cruz), and reactive complexes were visualized using StrpABComplex/horseradish peroxidase (Dako). Tissue was counterstained with hematoxylin.

FIG. 7

Further evaluation of the intracellular localization of STATs in NPC tissues. Immunohistochemical staining was performed as described in the legend to Fig. 6. STAT3 and STAT5 staining is visible within tumor cell nuclei. In contrast, STAT4 staining is restricted to the cytoplasm. (a) STAT3 (magnification, ×400). (b) STAT3 (magnification, ×600). (c) STAT5 (magnification, ×600). (d) STAT4 (magnification, ×600).

FIG. 8

Model for in vivo EBV gene regulation and tumorigenesis. In primary infection, EBNA2 regulates the expression of the nuclear EBNAs and the LMP genes including LMP1 and modulates cellular gene expression. The strong immune response to the immunogeneic EBNAs limits the occurrence of EBNA2-expressing tumors to immunocompromised individuals. During in vivo latency, in the absence of EBNA2, the EBNA1 and LMP1 genes are regulated by STATs. Chronic activation of STATs through a natural cytokine signaling event such as inflammation or through aberrant oncogene-activated signaling may upregulate EBV EBNA1 and LMP1 expression and predispose the cell to EBV-driven tumorigenesis.

FIG. 9

Potentiation of STAT signaling by LMP1. LMP1 may contribute to a self-sustaining cycle of STAT activation and continued LMP1 synthesis. LMP1 upregulates expression of the cytokine IL-6 and EGFR, which mediate tyrosine phosphorylation of STAT1 and STAT3. LMP1 is also able to activate JAK3, whose targets include STAT5. In addition, LMP1 increases the activity of JNK, a kinase involved in serine phosphorylation of the STAT protein transcriptional activation domain. Thus, STAT-induced LMP1 expression may lead to a state of constitutive STAT activation that can be maintained independently of ongoing external signaling.