B Cell Activation Results in IKK-Dependent, but Not c-Rel- or RelA-Dependent, Decreases in Transcription of the B Cell Tolerance-Inducing Gene Ets1

Humoral immunity involves B cell differentiation into Ab-secreting plasma cells. This process is tightly regulated because impaired differentiation of B cells to plasma cells leads to immunodeficiency, whereas excessive differentiation of B cells to plasma cells results in secretion of autoantibodies (1, 2). A number of different transcription factors are involved in regulating the transition of B cells into plasma cells, including the transcription factor Ets1. Mice lacking Ets1 have increased percentages of Ab-secreting plasma cells, leading to increased levels of serum IgM, IgG1, and IgE (3, 4). Ets1 knockout (KO) mice have high titers of autoantibodies to dsDNA and other autoantigens (4, 5). These phenotypes are in part B cell intrinsic, because loss of Ets1 in B cells results in increased plasma cell differentiation and increased autoantibody secretion (6, 7). This B cell–intrinsic function of Ets1 is also demonstrated by the fact that purified Ets1-deficient B cells differentiate more readily into plasma cells in vitro when exposed to TLR stimulation (5, 6). Mechanistically, Ets1 functions to block B cell differentiation into plasma cells by antagonizing the function of the plasma cell transcription factor Blimp1 (8, 9). In addition to its B cell–intrinsic functions, Ets1 also regulates the differentiation of T cells and the absence of Ets1 leads to enhanced CD4 T cell activation and to increased T follicular helper cell generation, stimulating B cell responses (10). In this study, we further examine B cell–intrinsic pathways that regulate Ets1 expression.

Transmission of signals from the BCR depends on assembly of a variety of signaling molecules in the vicinity of the BCR, which results in phosphorylation of substrate proteins and induction of calcium flux. One pathway activated by BCR signaling is the canonical NF-κB pathway, which induces nuclear translocation and activation of NF-κB family transcription factors to regulate gene expression in the nucleus. This pathway is activated in response to BCR or TLR signaling in B cells and results in activation of the IKK complex, composed of the kinases IKK1 and IKK2 and the regulatory subunit NEMO. The IKK complex then phosphorylates IκB proteins, which are cytoplasmically located proteins that sequester NF-κB family transcription factors and prevent their nuclear entry. Phosphorylation of IκB proteins results in their degradation and the release of NF-κB proteins to enter the nucleus. The NF-κB factors activated by the canonical NF-κB pathway are heterodimers of the p50 subunit complexed with either RelA (p65) or c-Rel.

IKK signaling downstream of the BCR is dependent on activation of the kinase Btk (11, 12). PI3K signaling can cooperate with Btk signaling in the induction of NF-κB activity, because PI3K signaling upregulates expression of c-Rel, one of the NF-κB proteins (13). NF-κB signaling is an important contributor to B cell immune responses as shown by the phenotypes of B cells lacking components of these pathways (14–17). For instance, B cell–specific KO of subunits of the IKK kinase complex, either IKK2 or NEMO, results in defects in B cell proliferation and survival and reduced humoral immune responses (14, 15, 17). Additionally, although deletion of individual NF-κB subunits c-Rel or RelA does not result in a defect in B cell generation or maintenance, their combined deletion results in reduced peripheral B cell survival (15). Furthermore, c-Rel is required for formation of germinal center B cells and plasma cells in response to immunization with T-dependent Ag and for Ag-specific Ab production in response to T cell–independent Ag. In contrast, RelA is not required for germinal center responses and plays only a modest role in T-independent responses (15).

We showed previously that inhibition of PI3K, Btk, or IKK2 blocked the ability of BCR or TLR signaling to downregulate the expression of the transcription factor Ets1 (6). A constitutively active version of IKK2 can mediate downregulation of Ets1 in A20 B lymphoma cells in the absence of BCR signaling, whereas in primary B cells constitutively active IKK2 cooperates with BCR signaling to downregulate Ets1 more efficiently (6). Although it is clear that IKK is important for downregulation of Ets1 in B cells, how it might do so remains unclear. We show in this study that IKK specifically regulates the transcription of the Ets1 gene. Neither RelA nor c-Rel is individually required for BCR-induced downregulation of Ets1. Furthermore, Foxo3a, another transcription factor regulated by IKK, is also not required for Ets1 downregulation. Although IKK signaling can also activate the mTORC2 pathway, we found that inhibiting mTORC2 activity did not abrogate the downregulation of Ets1 upon BCR crosslinking. Therefore, an IKK-dependent pathway other than those tested appears to be required for Ets1 downregulation in B cells in response to BCR or TLR signaling.

The following mouse strains were used in this study: CD19-Cre (18), Rela^fl (19), and Rel^fl (19), all on a C57BL/6 genetic background, and Foxo3a^−/− mice (20) on an FVB background. Wild-type C57BL/6 and FVB littermate controls were also used. All mice were housed at the Roswell Park Comprehensive Cancer Center Laboratory Animal Shared Resource, except mice lacking Foxo3a and FVB/N littermate controls, which were housed in the UT Southwestern Medical Center Animal Facility. All animal experiments were performed in accordance with Institutional Animal Care and Use Committee protocols for the relevant institutes.

A20 B cell lymphoma cells were maintained in complete media (RPMI 1640 + 10% FBS, 1% GlutaMAX, 1% penicillin/streptomycin, and 50 µM 2-ME). Cells were counted and aliquoted in 12-well plates and then stimulated or not with 10 μg/ml goat anti-IgG F(ab′)₂ crosslinking Ab (Jackson ImmunoResearch Laboratories, West Grove, PA). Mouse B cells were purified from spleen using the EasySep mouse B cell isolation kit (STEMCELL Technologies) and were subsequently rested in a tissue culture incubator for 30 min in complete media to allow recovery from the stress of isolation. Cells were then either left unstimulated or stimulated with 10 μg/ml goat anti-IgM F(ab′)₂ crosslinking Ab (Jackson ImmunoResearch Laboratories), 5 μg/ml LPS (Sigma-Aldrich, St. Louis, MO) or 5 μg/ml CpG oligonucleotide (CpG ODN 1826, InvivoGen, San Diego, CA) for 2 or 4 h and then harvested for mRNA or protein, respectively. For inhibitor studies, rested B cells were treated prior to stimulation with 100 µM cycloheximide (CHX) for 15 min, 5 µM IKKβ inhibitor IV (Calbiochem) for 1 h, or 1 μM AZD8055 mTOR inhibitor for 30 min.

Whole B cell lysates were prepared by boiling in a SDS lysis buffer and samples were subjected to Western blotting. The following Abs were used: rabbit monoclonal anti-Ets1 (D8O8A, Cell Signaling Technology), rabbit monoclonal anti-p65 (D14E12, Cell Signaling Technology), rabbit monoclonal anti-c-Rel (D3B8S, Cell Signaling Technology), mouse monoclonal anti-GAPDH (6C5, MilliporeSigma), rabbit monoclonal anti-Foxo3a (D19A7, Cell Signaling Technology), rabbit monoclonal anti-Akt phospho-Ser⁴⁷³ (D9E, Cell Signaling Technology), and rabbit monoclonal anti-S6 ribosomal protein phospho-Ser^235/236 (D57.2.2E, Cell Signaling Technology).

B cells were harvested in TRIzol reagent (Invitrogen) and total RNA was isolated using the Direct-zol RNA miniprep kit (Zymo Research). The QuantiNova reverse transcription kit (Qiagen) or iScript cDNA synthesis kit (Bio-Rad) was then used to generate cDNA. Quantitative PCR (qPCR) was performed using iQ SYBR Green supermix (Bio-Rad) with the following primers: Ets1 mature mRNA (forward, 5′-AGTCTTGTCAGTCCTTTATCAGC-3′; reverse, 5′-TTTTCCTCTTTCCCCATCTCC-3′); Ets1 premRNA set A (forward, 5′-TCGATCTCAAGCCGACTCTC-3′; reverse, 5′-GTCTTGGGCCACCAACAGTC-3′); Ets1 premRNA set B (forward, 5′-CTTCTCCCAGCCTGACCTAC-3′; reverse, 5′-ACAGTGCTGCTCCACTCATAC-3′); Actb (forward, 5′-GCAGCTCCTTCGTTGCCGGTC-3′; reverse, 5′-TTTGCACATGCCGGAGCCGTTG-3′). Expression levels were normalized to Actb using Bio-Rad CFX Manager software and then normalized to control groups, which were arbitrarily set to a value of 1.

Primary mouse B cells were purified and stimulated with 10 μg/ml goat anti-IgM F(ab′)₂ crosslinking Ab (Jackson ImmunoResearch Laboratories). Two hours after addition of the stimulation, cells were crosslinked by adding formaldehyde to the media to a final concentration of 0.25%. Formaldehyde treatment was continued for 8 min at room temperature, and then fixed cells were isolated and washed. Fixed B cells were lysed using a Bioruptor sonicator (Diagenode). Chromatin was sonicated to yield fragments of an average size ∼200–700 bp and immunoprecipitated using the Active Motif ChIP-IT High Sensitivity chromatin immunoprecipitation kit using a mouse monoclonal anti-polymerase II (Pol II) Ab (clone 4H8, Active Motif) that has been validated to work in chromatin immunoprecipitation reactions. Chromatin immunoprecipitated DNA was quantitated using qPCR to detect binding to the Ets1 promoter (forward, 5′-AATCTTAGCAGCGGTGGGC-3′; reverse, 5′-AAAAGCGAAAGGAAGGGCGG-3′) and as a control to the Egr3 promoter (forward, 5′-CTTTCCCGGGGCTCAGATAA-3′; reverse, 5′-CTAGCTCACTGCTGCCCAAA-3′). Fold enrichment was calculated by normalizing to the nontargeting control and then to a control region of chromosome 9 (forward, 5′-CAGTCCCGATGCCTCCTTTT-3′; reverse, 5′-ACCTCTTACTCTGGCGGTCT-3′).

Statistics were calculated using GraphPad Prism software. One-way or two-way ANOVA with multiple comparisons were used to determine significance. A significant difference was indicated by a p value of <0.05.

We previously showed that stimulation via either the BCR or TLRs such as TLR4 or TLR9 results in downregulation of both Ets1 protein and mRNA (5). Those studies did not distinguish whether the downregulation occurred at the level of Ets1 gene transcription or at a posttranscriptional stage such as mRNA or protein stability. To determine whether transcription of the Ets1 gene is altered by BCR signaling, we designed primers that amplify newly transcribed premRNA of Ets1. Because the premRNA is a very short-lived product that is rapidly processed by removal of introns, quantifying premRNA levels is a good surrogate for quantifying gene transcription. We designed two sets of primers for the premRNA transcript, targeting exon-intron junctions, and one set of primers for the mature processed Ets1 transcript, targeting an exon-exon junction (Fig. 1A). Mouse splenic B cells were stimulated via the BCR, TLR4, or TLR9, or left unstimulated. qPCR showed that activation of BCR and TLR downregulates both the mature mRNA and the premRNA to a similar extent (Fig. 1B). Similarly, activation of BCR and TLR in the A20 mouse B cell lymphoma line, which expresses a BCR of the IgG isotype, downregulates Ets1 mRNA and premRNA (Fig. 1C). Western blotting confirmed downregulation of Ets1 protein in splenic B cells and A20 cells (Fig. 1D–(E). These data indicate that the BCR- and TLR-induced downregulation of Ets1 is mediated by a decrease in Ets1 gene transcription.

To further confirm that transcription of the Ets1 gene is affected by BCR crosslinking, we assessed binding of RNA Pol II at the Ets1 promoter using chromatin immunoprecipitation. As a control, we also assessed recruitment of Pol II to the Egr3 promoter, which has previously been shown to have increased Pol II binding upon BCR stimulation (21). Pol II binding is found at the Ets1 promoter in unstimulated B cells, in accordance with the high expression of Ets1 in naive B cells (Supplemental Fig. 1A). Upon BCR crosslinking, Pol II binding to the Ets1 promoter is decreased, whereas it is increased at the Egr3 promoter as expected. This was further confirmed by examining existing Pol II chromatin immunoprecipitation sequencing datasets in resting and BCR-stimulated B cells (Supplemental Fig. 1B) (15). Visualization of Pol II binding at the Ets1 promoter shows a marked decrease in stimulated B cells. Reduced levels of Pol II at the Ets1 promoter upon stimulation are consistent with reduced transcription of the Ets1 gene.

We next determined that the downregulation of Ets1 at the transcriptional level is dependent on IKK activity. Primary mouse B cells were pretreated with an inhibitor of IKK2 activity, then stimulated with either anti-IgM Ab or with LPS or CpG containing oligodeoxynucleotide. As shown in (Fig. 2A, incubation with an IKK2 inhibitor prevented the stimulation-induced downregulation of both Ets1 mature mRNA and premRNA. In sum, stimulation of BCR and TLR functions to downregulate the transcription level of the Ets1 gene and this is dependent on IKK signaling.

Changes in Ets1 gene transcription in BCR-stimulated B cells might be due to synthesis of a repressor protein, which could then bind to regulatory elements in the Ets1 locus and suppress expression of the gene. If so, then BCR-induced downregulation of Ets1 would require new protein synthesis. To test this possibility, we first used mouse splenic B cells treated or not with CHX to stop new protein synthesis. Cells were subsequently stimulated by BCR crosslinking Ab and Ets1 mature mRNA, and premRNA was quantitated by qPCR. As shown in (Fig. 2A, CHX pretreatment did not prevent downregulation of Ets1 transcription upon BCR crosslinking. We similarly tested the effects of CHX on the A20 B cell lymphoma cell line and showed that CHX also did not affect BCR-induced downregulation of Ets1 in this IgG-expressing cell line (Fig. 2B), indicating that de novo transcription of a repressor is not responsible for downregulating Ets1 in response to BCR signaling and implicating changes in the expression or activity of existing transcription factors. Interestingly, treatment with CHX also tended to increase the basal levels of Ets1 RNA in unstimulated B cells, although this effect was only statistically significant for the mature Ets1 mRNA transcript in A20 cells (Fig. 2B, 2C). The effects of CHX in unstimulated B cells suggest that there may be a labile protein that promotes degradation of Ets1 mRNA.

The most well-studied target of IKK2 downstream of BCR signaling is the canonical NF-κB pathway in which the IKK complex phosphorylates IκBα and triggers its degradation, releasing NF-κB complexes containing heterodimers of RelA:p50 or c-Rel:p50. To test roles for RelA and c-Rel in regulating Ets1, we generated mice lacking either RelA or c-Rel in B cells by crossing Rela^fl/fl mice or Rel^fl/fl mice to CD19-Cre mice. Western blotting confirmed that RelA expression was abrogated in B cells isolated from CD19-Cre Rela^fl/fl mice (Fig. 3B). Stimulating Rela^fl/fl B cells via the BCR or TLRs downregulated Ets1 expression efficiently at both the mRNA and protein level (Fig. 3) and to a similar extent as wild-type B cells (data not shown). B cells lacking RelA (CD19-Cre Rela^fl/fl, denoted as RelA KO) downregulated Ets1 premRNA, mature mRNA, and protein just as efficiently (Fig. 3). Similarly, loss of c-Rel (CD19-Cre Rel^fl/fl, designated c-Rel KO) did not prevent downregulation of Ets1 mRNA or premRNA by BCR crosslinking or TLR engagement (Fig. 4A). Interestingly, TLR stimulation of c-Rel KO B cells did not downregulate Ets1 protein as efficiently as did Rel^fl/fl controls (Fig. 4B), suggesting that although c-Rel does not regulate Ets1 gene transcription, it may regulate a modulator of Ets1 protein stability. Overall, neither RelA nor c-Rel is required for BCR- or TLR-induced downregulation of Ets1 gene transcription, although c-Rel does seem to play a role in regulating some aspect of Ets1 protein stability.

Our results above implicate IKK, but not canonical NF-κB signaling, in downregulation of Ets1 in BCR-stimulated B cells. Several additional targets for IKK have been described (22). One such target is the transcription factor Foxo3a (23). IKK-mediated phosphorylation of Foxo3a leads to its nuclear export and degradation. Foxo3a might serve as an activator of Ets1 expression in the basal state and, if so, IKK signaling could downregulate Ets1 expression by triggering degradation of Foxo3a. To test this possibility, we examined B cells isolated from Foxo3a^−/− mice and found that the basal levels of Ets1 mature mRNA and premRNA are unaffected by loss of Foxo3a (Fig. 5A), suggesting that Foxo3a is not responsible for maintaining Ets1 expression in the resting state. Foxo3a was also not required for BCR-induced downregulation of Ets1 RNA or protein (Fig. 5). We did note that basal levels of Ets1 protein tended to be lower in unstimulated Foxo3a^−/− B cells, although this effect was not statistically significant (Fig. 5B). Although B cells from both Foxo3a^−/− mice and wild-type littermates downregulated Ets1 similarly upon BCR crosslinking, we noted that this effect was not as robust as the downregulation seen in B cells shown in the prior figures. All mice used in those earlier experiments were on a C57BL/6 genetic background, whereas Foxo3a^−/− mice and their wild-type littermate controls were on an FVB/N background. C57BL/6 mice carry the IgMb allotype, whereas FVB/N mice carry the IgMa allotype. To determine whether allotype differences in IgM might explain the less robust response to BCR crosslinking (anti-IgM) Ab, we compared the ability of wild-type B cells from IgMb⁺ mice (C57BL/6) and IgMa⁺ mice (BALB/c and FVB/N) to downregulate Ets1 in response to BCR crosslinking. As shown in Supplemental Fig. 2, both C57BL/6 and BALB/c B cells downregulated Ets1 to a similar extent upon BCR crosslinking, whereas FVB/N B cells showed a less efficient downregulation. This implies that the allotype of the BCR is not the reason for weaker downregulation in FVB/N B cells and that some other as yet unknown factor in this strain accounts for the weaker response to signaling.

Another target of IKK that has been described is Rictor, a required scaffolding component of the mTORC2 complex (24). The phosphorylation of Rictor by IKK is thought to activate mTORC2 signaling. mTOR activity has been shown to regulate Ets1 levels in corneal epithelial cells and vascular smooth muscle cells, although in these cases mTOR activation upregulates Ets1 expression (25, 26). To test whether IKK-dependent activation of the mTOR pathway might be involved in downregulating Ets1 in B cells, we isolated splenic B cells and treated them with an inhibitor (AZD8055) that blocks activation of both mTORC1 and mTORC2 or with vehicle control, and then stimulated them with BCR crosslinking Ab. The inhibitor blocked phosphorylation of mTORC1 target S6 kinase (Thr³⁸⁹) induced by BCR crosslinking as well as phosphorylation of the S6 kinase target S6 ribosomal protein (S6RP) (Fig. 6A). The AZD8055 inhibitor was also able to block the phosphorylation of Akt on Ser⁴⁷³, which is known to be mediated by mTORC2. Despite the ability of AZD8055 to inhibit both the mTORC1 and mTORC2 pathways, it did not affect the downregulation of Ets1 mRNA or protein by BCR crosslinking (Fig. 6). However, similar to the effect we saw with CHX pretreatment, inhibition of mTOR increased basal levels of Ets1 mRNA in unstimulated B cells (Fig. 6B). In sum, IKK-dependent phosphorylation of Rictor to control mTORC2 activity is unlikely to play a role in the downregulation of Ets1, although it may play a role in restraining basal expression of Ets1 mRNA.

Kinases such as Btk and IKK2 are important in transmitting signals from the BCR or TLR to downstream targets that help to mediate functional outcomes in B cells. Transcription factor Ets1 is a key regulator of B cell differentiation into plasma cells and its levels are controlled by BCR and TLR signaling. Ets1 levels are high in naive B cells but are rapidly downregulated upon stimulation through either the BCR or TLRs. This downregulation of Ets1 takes place soon after B cell activation within 2–4 h of stimulation, and Ets1 levels remain low for at least 12 h poststimulation (6, 27, 28). We have previously demonstrated that downregulation of Ets1 by BCR or TLR stimuli is dependent on signaling via a Btk→IKK kinase pathway (6).

In the current study, we demonstrate that BCR signaling and TLR signaling both regulate the levels of transcription of the Ets1 gene as shown by their effects on newly transcribed Ets1 premRNA and on recruitment of RNA Pol II to the Ets1 promoter. In B cells, we found that Ets1 mRNA levels are generally well correlated with levels of Ets1 protein, suggesting that the main control of Ets1 is at the level of gene transcription. As anticipated by our prior studies, blocking IKK activity abrogates the ability of BCR or TLR stimulation to downregulate Ets1 gene transcription.

Reduced transcription of the Ets1 gene upon B cell activation could be caused by a number of different processes, including reduced binding of activator proteins or increased binding of repressor proteins. Transcriptional repressor proteins, such as Blimp1 and Bcl6, can be induced in B cells by BCR or TLR signaling (27). However, induction of repressor protein expression is not required to downregulate Ets1, because CHX treatment failed to block the downregulation. That implies that the activity or localization of a pre-existing transcription factor is changed by the stimulation conditions. Unexpectedly, we found that resting B cells treated with an mTOR inhibitor had increased Ets1 mRNA levels. This is due to effects of the inhibitor on Ets1 gene transcription, as levels of mature mRNA and premRNA were both elevated. mTOR kinase is known to regulate transcription through interactions with transcriptional machinery and transcription factors (29), and it can even be recruited directly to DNA regulatory regions (30) to influence gene expression. Therefore, inhibition of mTOR might lead to changes in Ets1 gene transcription.

IKK signaling is required to downregulate Ets1 in response to either BCR or TLR stimulation. The best known pathway activated by IKK signaling is the canonical NF-κB pathway involving formation of transcription factor heterodimers of p50:RelA or p50:c-Rel. This canonical NF-κB pathway does not seem to be required for Ets1 downregulation, as deletion of either c-Rel or RelA had no effect on Ets1 levels under basal or stimulated conditions. The p50 can form homodimers, and we considered the possibility that p50 homodimers might regulate Ets1 expression. p50 homodimers are present at high levels in resting B cells (31), where Ets1 is also highly expressed. p50 homodimers typically function as repressors of gene transcription, as p50 lacks a transcriptional activation domain. Thus, it seems unlikely that p50:p50 is involved in maintaining basal levels of Ets1 in resting B cells. Upon stimulation, heterodimeric p50:RelA or p50:c-Rel complexes are increased, whereas p50 homodimers remain similar to those found in unstimulated cells (32). For this reason, p50 homodimers are not likely to be involved in stimulation-dependent downregulation of Ets1. It remains possible that c-Rel and RelA may cooperatively regulate expression of Ets1, but we were unable to test this in our system, because combined deletion of both c-Rel and RelA using CD19-Cre leads to a significant decrease in mature B cell numbers, with many of the remaining B cells failing to delete efficiently (15). Thus, addressing this question will require generation of an inducible B cell–specific deletion of RelA and c-Rel.

A similar dependence on IKK signaling, but not NF-κB, has been described in autophagy. IKK was required to induce the autophagic response, but RelA was not required and an IκB superrepressor protein, which prevents the activation of either c-Rel or RelA, also did not affect autophagy (33). Similarly, IKK signaling, but not RelA, is required for bipolar spindle formation and proper chromosomal segregation during mitosis (34) and for mast cell degranulation (35). In epidermal development, IKK1 is required for terminal differentiation of keratinocytes, but NF-κB is not required as determined by expression of the IκB superrepressor protein (36). Thus, there are numerous biological phenomena that depend on IKK signaling, but not the activation of the classical NF-κB pathway, and this also seems to be the case for downregulation of Ets1 in B cells.

IKK kinases are known to phosphorylate and regulate the activation of a number of non–NF-κB targets. The transcription factor Foxo3a can be directly phosphorylated by IKK kinases, which results in its nuclear export and degradation. Foxo3a is expressed in B cells and regulates cell survival and differentiation (37, 38). Thus, it was a reasonable candidate to regulate expression of Ets1 in B cells. However, we found that Foxo3a was not involved in maintaining basal levels of Ets1 or in BCR-mediated downregulation of Ets1 RNA levels. IKK kinases can also phosphorylate Rictor, a subunit of the mTORC2 complex, and this is associated with increased mTORC2 activity and in phosphorylation of Ser⁴⁷³ of Akt (24). It was possible that IKK might stimulate mTORC2 activation, and mTORC2 could then directly or indirectly modulate the activity of a transcription factor regulating Ets1 levels. However, we found that an inhibitor of mTORC2 did not interfere with BCR-mediated downregulation of Ets1. Because the inhibitor we used also blocks mTORC1 activity, we can also conclude that mTORC1 signaling is not required for downregulation of Ets1 by BCR signaling.

Although the main control of Ets1 levels in B cells is at the level of gene transcription, Ets1 levels can also be affected by posttranscriptional events. Published studies have described roles for microRNAs in regulating Ets1 mRNA translation (39–49). Ets1 protein stability can also be regulated by signaling pathways that control its ubiquitination and degradation (50, 51). These posttranscriptional events may help in fine-tuning Ets1 levels in B cells. We found that Ets1 protein (but not mRNA) levels tended to be reduced in resting Foxo3a^−/− B cells, although this effect did not reach statistical significance. This implies that Foxo3a may regulate Ets1 in a non–transcription-dependent manner. Further studies are required to determine the exact mechanisms underlying this phenomenon. c-Rel appears to play a role in the posttranslational regulation of Ets1 in response to TLR signaling. In the absence of c-Rel, we observed that TLR stimulation is not as efficient at decreasing Ets1 protein despite efficient downregulation of Ets1 premRNA and mature mRNA. Ets1 protein stability has been shown to be influenced by various posttranslational modifications, including ubiquitination and ribosylation (52, 53). It is possible that c-Rel regulates the expression of enzymes responsible for the destabilizing posttranslational modification of Ets1.

In summary, we have demonstrated that BCR and TLR signaling regulate the transcription of the Ets1 gene in an IKK-dependent manner, but that several well-known targets of IKK, including the canonical NF-κB proteins, Foxo3a, and the mTOR pathway, are not required for the effects of IKK on Ets1.

We thank Kirsten Smalley for technical assistance and Dr. Satrajit Sinha for helpful discussions.

The authors have no financial conflicts of interest.