Discussion
This study defines the oncogenic properties of LMP1 in promoting B cell lymphomagenesis. LMP1 transgenic mice have a higher incidence of lymphoma [26] and the progression to lymphoma correlates with higher expression levels of LMP1 (Figure 1A and 1B), suggesting that LMP1 is directly involved in tumor development. Table 1 summarizes the biological and molecular properties that were identified in wild-type and LMP1 transgenic lymphomas. Although many of the molecular properties studied were similar between wild-type and LMP1 transgenic lymphomas, there were distinguishing biological properties, namely the ability of LMP1 transgenic lymphomas to induce higher levels of survival and proliferation. Interestingly, although LMP1 transgenic mice develop lymphomas in the same B-1a cell type as spontaneous wild-type lymphomas (Figure 2), some signaling effects induced by LMP1 may explain the enhanced promotion to lymphomagenesis. Since CD40-deficient mice have decreased numbers of IgMhighIgDlow cells, a phenotype associated with B-1, marginal zone, and memory B cells, the mimicry of CD40 signaling by LMP1 could possibly contribute to the expansion of B-1 cells [20]. It is noteworthy that expression of LMP1 in transgenic mice has been shown to inhibit the formation of GCs [23,49], preventing typical B-2 cells from antigen-driven selection and expansion. The lack of GC reactions may contribute to the bias of LMP1 transgenic mice towards B-1 cell lymphomas. Interestingly, LMP2 signaling also favors development of B-1 cells, but this occurs in the absence of transformation. These results suggest that the mimicry of B cell receptor signaling by LMP2 promotes B-1 cell differentiation but not transformation [24,50,51]. This promotion of B-1 differentiation may account for the ability of LMP2 to exacerbate autoimmunity and bypass anergy induction [52,53]. In contrast, the preponderance of tumors of B-1a origin does not reflect effects of LMP1 signaling on B cell differentiation, as splenic B cells from healthy LMP1 transgenic mice contain similar numbers of B-1 and B-2 cells as wild-type mice. In support of this lack of effect, the differentiation of B-1 versus B-2 cells is thought to be independent of CD40 signaling [54]. 
B-1 cells constitute the predominant lymphocyte population in the peritoneal and pleuropericardial cavities, while B-2 cells are mainly found in the spleen, lymph node, and peripheral blood. B-1 cells produce the main source of IgM and IgA antibodies in serum, which are involved in T cell-independent responses to common microbial antigens. Importantly, B-1 cells have the unique capacity to self replenish and are also predisposed to transformation [28,29]. Clonal expansion of B-1 cells can be detected in aging mice above 18 mo of age, and B-1 cells are thought to be the murine progenitor of B cell chronic lymphocytic leukemia [30]. The data presented in this study indicate that although LMP1 is expressed in all B lymphocytes in the transgenic mice, malignancy develops in this specific subset of B cells. The elevated expression of LMP1 in B-1a cells and the activation of specific pathways apparently induce malignant growth. These same pathways can also become sporadically activated in aged mice and also result in lymphoma development. This is similar to EBV-associated cancers in vivo, where pathways that are activated by LMP1 are also activated in the less prevalent EBV-negative forms of the cancers [40,55-58]. Thus, the contribution of EBV and LMP1 to tumor development is apparently the continuous activation of pathways that can also be sporadically activated and contribute to tumor development. 
The lymphomas were marked by the upregulation of IL10, constitutive activation of Stat3 signaling, and a requirement for activation of Akt, NFkappaB, and Stat3 pathways (Figures 5 and 7). Induction of IL10 is associated with the transformation of B-1 lymphomas in mice [59,60] and is frequently associated with EBV-positive B cell malignancies acting as a B cell growth factor [35-38]. In addition, LMP1 has been shown to stimulate IL10 expression in Burkitt lymphoma cell lines [61,62]. This suggests that although Stat3 is constitutively activated in the lymphoma cells, the induction of IL10 may further enhance Stat3 activation or may contribute to other IL10-responsive signaling pathways. 
LMP1 activates both the canonical and non-canonical pathways of NFkappaB signaling [14,63-65], and inhibition of NFkappaB blocked the survival of LMP1 transgenic lymphocytes and LMP1-positive and -negative lymphoma cells. NFkappaB and PI3K signaling are crucial for CD40-induced proliferation, and mice deficient for cRel or the p85 regulatory subunit of PI3K are unresponsive to mitogenic stimuli, including CD40 ligation [66-68]. We have previously shown that cRel is specifically activated in both wild-type and LMP1 transgenic lymphomas, suggesting that activation of cRel is associated with B cell transformation [27]. Our observations suggest that similar to CD40-induced proliferation, LMP1 induces proliferation through PI3K-mediated activation of Akt and activation of NFkappaB components such as cRel. CD40 also induces downregulation of the cell cycle inhibitor p27 through a PI3K-dependent manner, and the LMP1 lymphoma cells also had decreased levels of p27 with phosphorylation of Rb and increased Cdk2 (Figure 6C) [66]. Although LMP1 has been shown to deregulate the Rb pathway in epithelial cells [69], to our knowledge this is the first demonstration of this property in B lymphocytes. 
The requirement for Akt activation was confirmed by the striking inhibition of lymphoma viability by triciribine, an Akt inhibitor. However, the activated pAkt did not lead to phosphorylation and inactivation of the downstream target GSK3 (Figure 6A). This effect has also been described in EBV-positive HD biopsies [40]. In contrast, rapamycin, U0126, and SB202190 did not affect the survival of LMP1 transgenic lymphocytes or the wild-type and LMP1 transgenic lymphoma cells (Figure 7A, 7C, and 7E). This lack of effect by rapamycin confirmed the absence of activated p-mTOR levels (Figure 6B). These findings suggest that other Akt targets contribute to malignant progression. One key target is likely the inhibition of the Fox01 transcription factors. Repression of the pro-apoptotic transcription factor FoxO1 in a PI3K-dependent manner can inhibit expression of bcl6, a transcription factor necessary for GC formation [49,70]. It has been shown that overstimulation of CD40 signaling with agonistic antibodies inhibits GC formation [71]. Similarly, due to mimicry of CD40 signaling, transgenic LMP1 mice are also defective in GC formation [23,49]. The constitutive signaling by LMP1 likely blocks GC formation through downregulation of bcl6. Interestingly, clinical studies indicate that expression of LMP1 and bcl6 are mutually exclusive in non-HD and classical HD [72,73]. Thus, the LMP1 transgenic lymphomas mirror aspects of EBV-induced HD. Although the activation of Akt and the lack of Fox01 in the lymphoma cells suggest that LMP1 affects bcl6 and GC formation through this pathway, regulation of other Forkhead targets involved in cell cycle progression, such as p27 and CyclinD2, likely contribute to malignant transformation. Indeed, loss of FoxO1 expression in lymphoma cells correlated with a loss of p27 (Figure 6B and 6C). CyclinD2 has also been shown to be upregulated by LMP1 through release of FoxO1-mediated repression [70]. 
In summary, in this transgenic model of lymphomagenesis, LMP1 promotes malignancy in B-1a cells, a population that is predisposed to clonal expansion with age. The malignant lymphocytes were distinguished by constitutively active Stat3 signaling, decreased p27, and activated Akt and NFkappaB pathways, properties that are associated with promoting the growth and survival of B lymphocytes. Importantly, Akt, NFkappaB, and Stat3 pathways were critically required for the growth and survival of malignant lymphocytes as well as healthy LMP1 transgenic lymphocytes. The growth of EBV-transformed lymphocytes also requires activation of NFkappaB, and these studies provide insight into how LMP1 contributes to EBV-associated transformation. The transgenic lymphomas mirror multiple aspects of EBV-induced tumors and suggest that in vivo these properties of LMP1 are major factors in the development of cancer. 
