TRAF3 regulates the oncogenic proteins Pim2 and c-Myc to restrain survival in normal and malignant B cells

TRAF3 is a versatile intracellular adapter protein with multiple context-specific roles. Uniquely in B cells, TRAF3 deficiency enhances survival and increases the risk of transformation, as loss of TRAF3 is observed in several types of B cell cancers. Here, we report a new mechanism for TRAF3 in the restraint of B cell survival. We found that TRAF3 deficiency was associated with induction of the pro-survival kinase Pim2 in mouse primary B cells and human malignant B cell lines. The increase in Pim2 was independent of NF-κB2 activation but was ameliorated with inhibition of STAT3 expression or function. TRAF3 deficiency also led to a Pim2-dependent increase in c-Myc protein levels and was associated with reduced c-Myc ubiquitination. TRAF3-deficient primary B cells were less sensitive to cell death induced by the Pim inhibitors SGI-1776 and TP-3654. Interestingly, human malignant B cell lines with low expression of TRAF3 were more sensitive to Pim inhibition-induced cell death. Combination treatment of TRAF3-deficient B cells and B cell tumor lines with c-Myc inhibitors enhanced their sensitivity to Pim inhibition, suggesting a possible therapeutic strategy. TRAF3 thus suppresses a Pim2-mediated B cell survival axis, which can be a potential target for treatment of B cell malignancies.


TRAF3-mediated regulation of B cell Pim2 expression.
To determine the pathways involved in the hyper-survival of TRAF3-deficient B cells (TRAF3 −/− ), we performed a microarray analysis, comparing the gene expression of TRAF3 −/− mouse B cells compared to those of wild type littermate controls (WT) 17 . We also compared gene expression in TRAF3 −/− B cells versus TRAF3 −/− T cells from conditional knockout mice, as these T cells do not have enhanced survival 1,14 . Pim2 gene expression was increased in TRAF3 −/− B cells compared to either WT B cells or TRAF3 −/− T cells. Confirming microarray data, TRAF3 −/− B cells had 6-fold higher expression of Pim2 mRNA compared to WT B cells when examined by RT-PCR (Fig. 1a). Pim2 protein was also increased in TRAF3 −/− compared to WT B cells (Fig. 1b). Interestingly, TRAF3 deficiency specifically regulated the Pim2 isoform, as expression of Pim1 and Pim3 was unchanged (Supplemental Fig. 1).
Our observations in mouse primary B cells led us to predict that TRAF3 protein levels in B cell tumors would impact their relative levels of Pim2 protein. We examined 3 human MM-derived cell lines (OPM2, LP1, and RPMI8226) and observed an inverse correlation between their relative TRAF3 and Pim2 protein levels (Fig. 1c). In DLBCL-derived human cell lines, OCI-Ly7 cells had undetectable TRAF3 protein and increased Pim2 expression compared to TRAF3-positive BJAB cells (Fig. 1d). Figure 1c,d were previously presented in the doctoral dissertation of N.M. 23 . Although we expect that there are multiple gene alterations in tumor cells that could impact Pim2 expression, our results indicate that TRAF3 likely serves as an important regulator to restrain Pim2 expression at both the mRNA and protein levels in normal and malignant B cells. This conclusion is strengthened by the recent complementary finding described in the Introduction that human BCL cell lines expressing LMP1, which avidly binds and sequesters TRAF3, also display a TRAF3-deficient phenotype, including elevated Pim2 protein 10 .
Effect of loss of TRAF3 on Pim2 target phosphorylation. Phosphorylation of the pro-apoptotic BAD protein at serine-112 by Pim2 inhibits cell death 24 . The kinase p70-S6 (p70-S6K), S6 ribosomal protein (S6), and 4E-BP1, involved in protein translation, are also Pim2 phosphorylation targets, and contribute to regulation of cell survival 25,26 . Increased expression of Pim2, which is a constitutively active kinase, within TRAF3 −/− mouse B cells resulted in enhanced expression of its known targets BAD 27 , p70-S6K, 4E-BP1, and ribosomal protein S6 (Fig. 2a-e), as well as phosphorylated (active) forms of these proteins. In the case of 4EBP1 and S6, there was also a selective increase in the phosphorylated forms, above the increase in total amounts. These results indicate that the increased Pim2 protein is accompanied by increased Pim2 activity in TRAF3-deficient B cells.
Impact of TRAF3 and Pim2 upon c-Myc stability and levels. The transcription factor c-Myc has frequently been implicated in B cell oncogenesis 28,29 , and is a potential therapeutic target. Additionally, proviral tagging in myc transgenic mice revealed Pim2 as a collaborating oncogene driving lymphomagenesis 30 . Pim2 directly phosphorylates c-Myc on serine 329, leading to stabilization of c-Myc protein and enhancing cell transformation 31 . We thus investigated whether TRAF3-deficient B cells also display alterations in c-Myc expression and/or post-translational modification. TRAF3-deficient B cells had increased expression of c-Myc protein, without an increase in c-Myc mRNA (Fig. 3a,b). Inhibiting Pim2 expression with siRNA attenuated levels of c-Myc expression in TRAF3 −/− B cells (Fig. 3c). Although inter-mouse variability in absolute levels of both Pim2 and c-Myc resulted in differences that together did not reach statistical significance, the clear downward trend in c-Myc when Pim2 expression was reduced was consistent in all samples. Stability of c-Myc protein is regulated by K48-linked poly-ubiquitination 32 . In the absence of TRAF3, the elevated c-Myc protein levels were associated with reduction in K48-linked poly-ubiquitination of c-Myc ( Fig. 3d; previously presented in the doctoral www.nature.com/scientificreports www.nature.com/scientificreports/ dissertation of N.M. 23 ). These findings indicate that loss of TRAF3 in B cells results in increased cellular c-Myc protein, which requires the enhanced Pim2 expression downstream of TRAF3 deficiency and is associated with reduced c-Myc K48-linked ubiquitination.
To determine whether this increase in Pim2 was B cell-specific we examined T cell isolated from T cell-specific TRAF3 −/− mice 14 . There was no significant difference in Pim2 mRNA in TRAF3 −/− and WT T cells in microarray analysis 17 . Interestingly, Pim2 protein expression was increased in TRAF3 −/− T cells compared to WT T cells, although this difference was not statistically significant. There was no difference in c-Myc protein expression by T cells, related to TRAF3 status (Supplemental Fig. 2).

STAT3 dependence of increased Pim2 in TRAF3-deficient B cells. As mentioned above, TRAF3
promotes degradation of NIK. In the absence of TRAF3, NIK accumulates and promotes activation of the NF-κB2 pathway (reviewed in 1 ). B cell-specific deletion of NIK substantially decreases overall B cell survival and inhibits NF-κB2 activation 33 . Because BAFF-mediated induction of Pim2 is NF-κB2-dependent 34 , we tested the requirement for NIK in driving increased Pim2 expression of TRAF3-deficient B cells. Interestingly, loss of NIK in TRAF3-deficient B cells did not affect the elevated Pim2 (Supplemental Fig. 3; previously presented in the doctoral dissertation of N.M. 23 ). Previously, it has been shown that TRAF3-deficient B cells still show enhanced survival when NIK is deleted 35 , suggesting that there are NIK-independent, pro-survival pathways that are regulated by TRAF3.
STAT3 activation increases Pim2 expression in both hematological and solid tumors 21,36,37 . TRAF3 inhibits STAT3 activation downstream of the IL-6 receptor by recruiting the phosphatase PTPN22 38 . We thus investigated the relationship between increased expression of Pim2 in TRAF3 −/− B cells and levels of STAT3 activation. After www.nature.com/scientificreports www.nature.com/scientificreports/ treatment with the STAT3 inhibitor Cucurbatacin I 39 , Pim2 protein in TRAF3 −/− B cells decreased to WT levels ( Fig. 4a). Partial inhibition of STAT3 expression with siRNA in TRAF3 −/− B cells resulted in a similar decrease in Pim2 protein (Fig. 4b). Our earlier studies demonstrated that TRAF3 associates with the IL-6R following IL-6 binding, and recruits PTPN22 to this receptor to reduce IL-6-mediated STAT3 phosphorylation. However, IL-6 treatment did not further increase Pim2 levels in TRAF3 −/− B cells (data not shown), indicating that TRAF3 does not regulate Pim2-mediated homeostatic survival via the IL-6R pathway. Our results show that TRAF3 deficiency sustains increased Pim2 expression downstream of STAT3 activation and independently of NF-κB2 activation.

Impact of TRAF3 status on susceptibility of B cells to Pim inhibitors. Pim2 is important for B
cell survival, and has been suggested as a potential target for treatment of B cell malignancies 22,27 . To determine whether Pim2 inhibition induces death of TRAF3-deficient B cells, we tested B cell survival after 24 hr treatment with two different Pim kinase inhibitors: SGI-1776 40 and TP-3654 41 . While isoform-restricted Pim2 inhibitors are not currently available, Supp. Fig. 1 indicates that the expression of other Pim isoforms is not increased in Relative levels of p-BAD S112, p-p70S6K T389, p-4E-BP1 T37/46, or p-S6 ribosomal protein S235 were normalized first to the β-actin loading control (top row) or total BAD, p70S6K, 4E-BP1, or S6 (bottom row) and then to the average normalized WT value. Graphs depict mean values ± SEM. Unpaired t test was used in (b-d) and Mann Whitney test was used in (e) to evaluate differences for statistical significance (*p < 0.05, **p < 0.01; N = 4).
www.nature.com/scientificreports www.nature.com/scientificreports/ TRAF3-deficient B cells. Pim inhibition in both WT and TRAF3 −/− B cells led to increased cell death, consistent with the importance of Pim2 in B cell survival (Fig. 5a,b). TRAF3 −/− primary B cells were relatively more resistant to cell death mediated by both inhibitors, consistent with their elevated levels of Pim2 kinase.
We next wished to examine the relative sensitivity to Pim2 inhibition of human tumor-derived B cell lines with varying levels of TRAF3 protein (Fig. 1). While primary TRAF3 −/− B cells with elevated Pim2 resisted such inhibition, it is well-documented that tumors can develop dependence upon abnormally elevated proteins that promote growth and survival, a phenomenon known as 'oncogene addiction' 42 . Consistent with www.nature.com/scientificreports www.nature.com/scientificreports/ such dependence, the TRAF3 low RPMI8226 MM cell line was more susceptible to Pim2 inhibition compared to higher TRAF3-expressing LP1 and OPM2 MM cell lines, and at lower inhibitor concentrations (Fig. 5c,d). TRAF3-deficient OCI-Ly7 BCL cells were also more susceptible to SGI-1776-induced death than TRAF3 high BJAB BCL cells at intermediate inhibitor concentration (Fig. 5d).

Increased effect of combined inhibition of c-Myc and Pim2 on survival of TRAF3-deficient B cells.
Inhibiting c-Myc expression is a potential therapeutic strategy to treat c-Myc-driven malignancies; c-Myc is the most frequently dysregulated gene in MM 43,44 . In DLBCL, 5% to 14% of tumors have c-Myc rearrangements or gene amplifications 45 . The drug JQ1 inhibits bromodomain-containing proteins, as these can promote c-Myc gene expression, and has shown promise in treatment of human MM and DLBCL 46,47 . To test whether targeting Myc can decrease the enhanced survival of TRAF3-deficient B cells, we treated WT or TRAF3 −/− B cells with the c-Myc inhibitors JQ1 or 10058-F4, the latter being an inhibitor of c-Myc-mediated transcriptional activity 48 . Combination treatment with either of the c-Myc inhibitors enhanced killing by the Pim inhibitor SGI-1776 in both TRAF3 −/− mouse primary B cells (Fig. 6a,b) and in the TRAF3-deficient human MM cell line RPMI8226 (Fig. 6c,d). As use of Pim inhibitors has been limited by their cardiotoxic effects at higher doses 41 , these results suggest that combined targeting of Pim2 and c-Myc expression may be beneficial, and allow lower doses of Pim inhibitors to be efficacious.

Discussion
After it was revealed that TRAF3 plays a powerful B cell-specific role in restraint of homeostatic survival 2,4 , it has become evident that this TRAF3 role renders it an important tumor suppressor for B cell malignancies, particularly MM and BCL 9 . Because a striking component of the phenotype of all TRAF3-deficient cell types is constitutive activation of the NF-κB2 pathway 49 , it has often been assumed that enhanced NF-κB2 activation is both necessary and sufficient to explain how TRAF3 exerts its inhibition of B cell survival. However, a number of subsequent studies provide evidence that additional TRAF3-regulated pro-survival pathways, independent of NF-κB2, make important contributions to how TRAF3 restrains B cell survival (reviewed in 1, 2). We have thus worked to identify such pathways, as their characterization not only enhances understanding of basic TRAF3 functions, but also has important implications for choosing the most effective therapies for B cell malignancies.
Our results presented here show that loss of B cell TRAF3 led to induction of the constitutively active pro-survival kinase Pim2, and enhanced expression and activation of its downstream targets BAD, p-70S6K, ). An unpaired t test was used to evaluate differences for statistical significance (*p < 0.05, **p < 0.01). (b) TRAF3 −/− B cells were transfected with control or siRNA targeting STAT3. 24 hrs later, Pim2 and STAT3 expression in WCLs were analyzed by WB. Graphs depict STAT3 or Pim2 expression/β-actin loading control and normalized to the average of the siRNA control. A paired t test was used to evaluate differences for statistical significance (*p < 0.05, **p < 0.01, N = 5).
www.nature.com/scientificreports www.nature.com/scientificreports/ S6, and 4E-BP1 (Fig. 7). Interestingly, the Pim2 increase in TRAF3 −/− B cells is not attenuated with loss of NIK, suggesting that in these cells Pim2 levels are enhanced in TRAF3-deficient B cells independently of NF-κB2 activation. Our results also further illustrate multiple ways in which TRAF3 deficiency can enable and enhance pathogenesis of B cell malignancies. NF-κB2 is required for homeostasis of WT B cells 33 , but in TRAF3-deficient B cells certain oncogenic programming can be maintained independently of NF-κB2. This could potentially allow TRAF3-deficient B cells to serve as a reservoir for tumor recurrence and resistance to proliferation-targeting chemotherapy, as well as accumulation of additional oncogenic mutations.
TRAF3-deficient B cells maintained their elevated Pim2 expression in the absence of NIK, the initiating kinase of the NF-κB2 pathway, so we sought other upstream pathways that mediate the Pim2 increase resulting from TRAF3 deficiency. We observed here constitutive phosphorylation of STAT3 in TRAF3-deficient B cells and show that inhibition of STAT3 expression or function in TRAF3-deficient B cells decreased Pim2 elevation. Thus, TRAF3 in normal B cells inhibits Pim2 expression at least in part by restraining activation of STAT3, and by doing so inhibits B cell homeostatic survival, independent of direct engagement of the IL-6R. Interestingly, STAT3 is frequently constitutively active in bone marrow B cells from MM patients 50 , and TRAF3 deficiency is considered an early event in MM oncogenesis 51 .
Alternations in function of the oncogene c-Myc have long been implicated as playing an important role in the pathogenesis of BCL and MM 44 . Additionally, increased expression of c-Myc protein is associated with a poorer prognosis in patients with DLBCL, independent of other prognostic factors 45,52 . A substantial proportion (29% to 47%) of DLBCL have high protein expression of c-Myc, which, interestingly, is higher than the incidence of c-Myc gene translocations 45 . This suggests that in some cases of DLBCL, c-Myc protein is increased through alternative mechanisms such as post-translational modifications, and not purely from genetic alterations. c-Myc has many transcriptional targets and wide-ranging impacts on cell proliferation and growth. It is thus intriguing that the increase in Pim2 mediated by loss of TRAF3 also impacted B cell levels of c-Myc. Unlike the mRNA-directed increase in Pim2, this elevation in c-Myc was accomplished at the level of protein regulation, by decreased K48-linked poly-ubiquitination of c-Myc in TRAF3-deficient B cells, resulting in greater c-Myc protein www.nature.com/scientificreports www.nature.com/scientificreports/ stability. The present results are the first demonstration to our knowledge of Pim2-mediated enhancement of B cell c-Myc mediated at the post-translational level. In addition, c-Myc has pro-apoptotic functions 45 . Additional pro-survival pathways mediated by increased Pim2 expression, such as inhibition of activity of the pro-apoptotic  www.nature.com/scientificreports www.nature.com/scientificreports/ BAD protein, may help to counteract the pro-apoptotic functions of c-Myc in TRAF3-deficient B cells, allowing their markedly enhanced viability.
In contrast to TRAF3-deficient B cells, TRAF3-deficient T cells do not have increased viability and loss of TRAF3 has not been reported to be associated with T cell malignancies 5 . TRAF3 −/− T cells do have constitutive NF-κB2 activation, a feature we and others have seen in all TRAF3-deficient cell types examined to date 4,5,15 . Here, we found that TRAF3 −/− T cells had a modestly trending increase in Pim2 protein. TRAF3 −/− T cells do not have an increase in STAT3 expression or activation, suggesting another mechanism is responsible for the increased Pim2 in these T cells 53 . One possible explanation is that TRAF3 −/− mice have increased regulatory T cells (Tregs) compared to WT mice 54 and Tregs have increased Pim2 expression compared to conventional T cells, due to increased FoxP3 expression 55 . It would be interesting to explore this pathway further in future studies.
Our findings are summarized in a proposed model for TRAF3-mediated regulation of Pim2 and c-Myc in B cells (Fig. 7). These results have implications for selection of the most efficacious pathway inhibitors for treatment of TRAF3-deficient B cell malignancies, whether resulting from alterations in TRAF3 genes or post-translational protein reduction. Although TRAF3 levels inversely correlated with Pim2 expression in tumor-derived B cells as well as their primary cell counterparts, low expression of TRAF3 increased susceptibility to Pim inhibition in human malignant B cells, in contrast to resistance to Pim inhibitors in primary B cells. This is consistent with 'oncogene addiction' , wherein tumors can develop dependence upon abnormally elevated proteins that promote growth and survival 42 . Oncogene addiction has been demonstrated previously with the oncogenes c-Myc and BCR-ABL in leukemia 56 . TRAF3-deficient tumors may be more likely to become dependent on Pim2 activity than TRAF3-sufficient tumors. Thus, inhibiting Pim2 may be a strategy to selectively target TRAF3-deficient B cell tumors. The Pim inhibitor SGI-1776 failed in Phase I clinical trials due to a narrow therapeutic window caused by cardiac toxicity 19 . Here, we show that the addition of c-Myc inhibitors enhanced lower-dose SGI-1776-mediated killing of TRAF3 −/− B cells; in the case of 10058-F4, even the lowest dose of SGI-1776 effectively killed 70-90% of TRAF3-deficient B cells when the two drugs were combined. Combined targeting of these two proteins may thus have beneficial effect in treatment of B cell malignancies deficient in TRAF3.
In both mouse and human B cells, loss of TRAF3 leads to an enhanced survival phenotype. The frequent association of TRAF3 loss in MM and BCL, as well as mechanisms of post-translational loss of TRAF3 that may impact an additional group of B cell malignancies, suggest that this loss predisposes B cells to transformation, oncogenesis, and possibly resistance to conventional chemotherapy, because TRAF3 deficiency selectively enhances B cell survival without altering proliferation 4 . Thus, agents of conventional chemotherapy, many of which enhance DNA damage, will not kill TRAF3-deficient B cells, but could leave them with additional transforming mutations. It is thus important to fully understand the multiple ways in which TRAF3 regulates B cell survival and how these pathways may interact to most effectively treat B cell malignancies.
www.nature.com/scientificreports www.nature.com/scientificreports/ Viability assays. Splenic B cells were cultured at 0.5-1 × 10 6 per ml. Cells were stained with the DNA dye propidium iodide (PI) and the % live cells was determined by the % of the PI-negative population, analyzed by fluorescence detection, using the Accuri C6 Flow Cytometer (BD Bioscience, San Jose, CA). The data were analyzed with FlowJo software (FlowJo LLC, Ashland, OR).

Real-time PCR (RT-PCR). RNA from mouse primary B cells was extracted with an RNeasy extraction kit
(Qiagen, Gaithersburg, MD) and cDNA synthesized using SuperScript III polymerase (Invitrogen, Carlsbad, CA). RT-PCR was performed on an ABI PRISM 7900 Sequence Detection System (Applied Biosystems, Grand Island, NY). Primers were purchased from Applied Biosystems (Mm00454579_m1 for Pim2, Mm00487804_m1 for c-Myc, Mm00435712_m1 for Pim1, Mm01308047_g1for Pim3, and Mm99999915_g1 for GAPDH). Expression data were analyzed using the comparative Ct method 64 and normalized to GAPDH expression.
Immunoprecipitation. 20-30 × 10 6 cells/sample were lysed with IP Lite lysis buffer (0.5% Triton X, 40 mM Tris, 100 mM NaCl, 1 mM CaCl 2 ) containing EDTA-free Mini Complete protease inhibitor (Roche) and 100 ug/mL DNase I (Roche). For assaying K48 ubiquitination, lysis buffer was supplemented with 20 mM N-ethylmaleimide (Sigma). Lysates were incubated with an Ab against c-Myc overnight at 4 °C with constant agitation. The immune complex was precipitated with protein G Dynabeads (Life Technologies), washed and resuspended in SDS-PAGE loading buffer and heated to 95 °C for 10 min.
Transfection of mouse B cells with siRNA. B cells were transfected using electroporation following a protocol previously described for primary T cells 65 . The Pim2 and STAT3 siRNAs were purchased from Santa Cruz Biotechnology. 5-10 × 10 6 cells were resuspended in 100 μL 2S electroporation buffer (5 mM KCl, 15 mM MgCl 2 , 15 mM HEPES, 150 mM Na 2 HPO 4 /NaH 2 PO 4 pH 7.2 and 50 mM sodium succinate). 300 pmol of siRNA was added and cells were electroporated using the Nucleofector II electroporation system (Lonza, Basel, Switzerland), with the X-001 program. Cells were resuspended in 20% FBS-supplemented BCM without antibiotics and analyzed 48 or 24 hrs later by WB.
Statistical analysis. Unpaired, two-tailed student's t tests were used to evaluate differences for statistical significance for most experiments. A paired two-tailed student's t test was used to evaluate statistical significance for paired sets of samples. An unpaired t test with Welch's correction was to compare groups with unequal variances. The Mann-Whitney test was used for independent samples that are not normally distributed. A Wilcoxon signed rank test was used to assess statistical significance for paired samples that were not normally distributed. Statistical details are listed in the caption of each figure. Statistical analysis and graphs were prepared using GraphPad Prism (GraphPad Software, San Diego, CA). (NS = not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).