Themis2 Is Not Required for B Cell Development, Activation, and Antibody Responses

Themis-family proteins are defined by the presence of a cysteine-containing, all-β in Themis (CABIT) domain (1). The family can be traced down to cnidarians, but in mammals it is composed of five members: Themis1 (originally Themis, but in this article termed Themis1 for clarity), Themis2, Themis3, Garem, and Gareml. The three Themis proteins share a similar structure including two consecutive CABIT domains and a proline-rich region (PRR), whereas Garem and Gareml contain one CABIT domain, a PRR, and a SAM domain (1–3). Themis1 and Themis2 show high conservation between each other, and both have a putative nuclear localization signal (NLS) and conserved C-terminal tyrosine residues, which serve as SH2-binding sites upon phosphorylation (4–7).

Signals from the BCR and TCR play critical roles in the development, survival, and activation of B and T lymphocytes. Several studies showed that Themis1 is essential for normal T cell development (1, 4, 5, 8, 9). In the absence of Themis1, positive selection of CD4⁺CD8⁺ double-positive thymocytes is impaired. Themis1 is phosphorylated after TCR stimulation, suggesting that Themis1 participates in signaling from the TCR (4, 7, 10–12). A recent study showed that Themis1 dampens signaling in thymocytes after low-affinity TCR stimulation, which normally results in positive selection, but when Themis1 is absent, such low-affinity stimulation results in increased TCR signaling and hence negative selection (13).

In comparison, little is known about Themis2. Studies in cancer cell lines suggested a role for Themis2 in differentiation and proliferation (14). Subsequently, Themis2 was shown to be involved in LPS-induced TNF-α production in RAW cells and primary human macrophages where it interacts with Lyn, Grb2, and Vav1 (6). Themis2 also associates with Grb2 and Vav1 in B cells and is phosphorylated after BCR stimulation (7). Furthermore, expression of Themis2 in the T cell lineage rescues thymocyte development in Themis1-deficient mice, suggesting that Themis1 and Themis2 carry out similar functions (7).

Strikingly, Themis1 and Themis2 show mutually exclusive expression: Themis1 is expressed in T cells, whereas Themis2 is expressed in B cells, macrophages, and dendritic cells (http://www.immgen.org). It has been noted for a long time that there are strong similarities between BCR and TCR signaling pathways (15); often one protein will exert a certain function in T cells, whereas its paralogue performs a similar function in B cells. Given the important role for Themis1 in T cell development and the high degree of similarity between Themis1 and Themis2, we hypothesized that Themis2 may have a critical function in B cell development or activation.

In this article, we show that Themis2 is expressed in all subsets of developing and mature B cells, and that no other Themis-family member is expressed in the B cell lineage. Furthermore, we show that Themis2 expression is downregulated after B cell activation. However, despite this nonredundant expression pattern, we find that, surprisingly, in the absence of Themis2, B cell development, activation, and Ab responses are unaffected.

The embryonic stem cell line JM8.F6 with the Themis2^{tm1a(KOMP)Wtsi} targeted allele (http://www.komp.org) was used to generate C57BL/6J-Themis2^{tm1a(KOMP)Wtsi}/Nimr mice. These were crossed to C57BL/6J-Tg(ACTFLPe)9205Dym mice (16) and to C57BL/6J-Tg(Prm-cre)70Og mice (17) to delete the frt-flanked lacZ reporter and the loxP-flanked exon4 of Themis2 respectively, generating C57BL/6J-Themis2^{tm1d(KOMP)Wtsi}/Nimr mice (Themis2^KO). All experiments used age- and sex-matched littermate control animals. Other strains used: B6.SJL-Ptprc^a/Nimr (CD45.1⁺), C57BL/6J-Rag1^tm1Mom/Nimr (Rag1-deficient), and C57BL/6J-Tg(TcraTcrb)425Cbn/Nimr (OT-II) (18). All mice were bred under specific pathogen-free conditions at the National Institute for Medical Research. Radiation chimeras were generated as described previously (19).

Mice were injected i.p. with PBS containing 10⁹ SRBC (Innovative Research), 50 μg NP (4-hydroxy-3-nitrophenylacetyl)_(0.6)-LPS in PBS, 50 μg NP₍₄₀₎-AECM-Ficoll in PBS, 50 μg NP₍₂₁₎-CGG (all BioSearch Technologies) in Alum (Thermo Scientific) with PBS for primary immunizations, or 50 μg NP₍₂₁₎-CGG in PBS for rechallenge and were bled or sacrificed at various time points for analysis or cell sorting. For cholera toxin (CTX) immunization, food was withdrawn from mice 3 h before being orally gavaged first with 1.5% NaHCO₃ and then with 125 μg/kg CTX from Vibrio cholerae Inaba 569B (List Biological Laboratories) in PBS. Blood was withdrawn at the indicated time points. Mice were sacrificed 14 d after immunization to obtain fecal samples from the small intestine.

RBC-lysed single-cell suspensions were stained in ice-cold PBS containing LIVE/DEAD fixable near-IR dead cell stain (Life Technologies) and the appropriate pretitered Abs. Cell numbers in the bone marrow are quoted per leg (one femur and one tibia). B10 cells were purified using the Miltenyi Biotec Regulatory B cell isolation kit with 24-h in vitro stimulation followed by flow cytometric sorting for B220⁺CD19⁺IL-10⁺ cells. To isolate plasma cells and plasmablasts, we enriched organ suspensions from mice immunized 5 d earlier with SRBC for CD138⁺ cells by staining with anti–CD138-PE and then using anti-PE beads (Miltenyi Biotec). To isolate germinal center B cells, we depleted splenocytes from mice immunized 10 d earlier with SRBC using anti–CD43-biotin and anti–IgD-biotin followed by streptavidin-Dynabeads (Life Technologies) and then sorted for B220⁺PNA⁺GL7⁺Fas⁺CD38^low. Data were collected on a BD FACSCanto II or sorted on BD FACSAria II, BD Influx, or Beckman Coulter MoFlo XDP cell sorters. Data were analyzed using FlowJo 9.6 (TreeStar).

B cells were purified from RBC-lysed single-cell suspensions by magnetic negative depletion using biotinylated Abs against CD43, CD11c, CD11b, and Ly6G and streptavidin-Dynabeads. B cells were cultured in DMEM, 100 μM nonessential amino acids, 20 mM HEPES buffer, 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 100 μM 2-ME. Unless indicated otherwise, cells were stimulated with 10 μg/ml anti-IgM F(ab)₂ (Jackson Immunoresearch), 10 μg/ml LPS (Alexis), 1 μg/ml CD40L (R&D Systems), 100 ng/ml IL-4, 100 ng/ml BAFF (both Peprotech), 50 ng/ml PMA, or 500 ng/ml ionomycin (both Sigma-Aldrich).

B cells were isolated and cultured as described above but labeled with 5 μM CFSE in PBS for 10 min at 37°C before culturing for 72 h. Proliferation was assessed by CFSE dilution and number of live cells was measured by a LIVE/DEAD fixable near-IR dead cell stain and counted using PerCP-beads on a flow cytometer. To measure cytokine secretion, B cells were cultured for 48 or 72 h and cytokines in the supernatants were measured with the 16-plex Fluorokine MAP assay (R&D Systems) on a Luminex 100 machine.

Total RNA was extracted with RNEasy Mini or Micro Plus kits (Qiagen) followed by cDNA synthesis using the Superscript III First Strand Synthesis SuperMix for quantitative RT-PCR (qRT-PCR) kit (Life Technologies). Samples were analyzed on an ABI Fast 7900HT under standard conditions using the TaqMan Gene Expression Assays (Life Technologies). Data were normalized to HRPT1 and analyzed using the comparative threshold cycle method.

Splenic B cells were purified by depletion with biotinylated anti-CD43 Ab and streptavidin-Dynabeads. Cell lysates were prepared using radioimmunoprecipitation assay buffer containing protease inhibitors. Cell debris was removed by centrifugation. Proteins were separated in 7% SDS-PAGE gels and wet-blotted onto Immobilon-FL polyvinylidene difluoride membrane (Millipore) by standard techniques. Membranes were blocked in Odyssey Blocking Buffer (Li-Cor), and then probed with anti-Themis2 exon 4 (sc-160439; Santa Cruz), anti-Themis2 exon 6 (AP9910b; Abgent), or anti–α-tubulin mouse monoclonal (clone TAT-1), followed by appropriate secondary Abs. Signals were detected with an Odyssey Infrared Imager (Li-Cor) and analyzed with the manufacturer’s software.

Follicular B cells were sorted by flow cytometry. Unstimulated samples were lysed directly, and stimulated samples were cultured as described at 3 × 10⁶ cells/ml for 6 h with the indicated stimuli. RNA was isolated using TRIzol and cleaned up using RNeasy mini kit. Single-end, unstranded, poly-A–enriched libraries were made using the TruSeq RNA sample preparation kit (Illumina). Samples were analyzed with an Illumina HiSeq 2000, collecting 13.2–76.1 million reads of 75 bases per sample. Reads were aligned to mm10 (Ensembl version 72) using Tophat (version 2.0.9). Raw counts were determined using the union method in htseq-count (version 0.5.4p3), and mappings were filtered for a phred score >10. edgeR (version 3.2.4) was used for filtration of lowly expressed features, normalization, and statistical analysis. Statistical significance of differences in gene expression was determined using the exact binomial test, reporting differences with a false discovery rate <0.05 (20). For VDJ analysis, reads were mapped to NCBI37/mm10 including an update containing additional VDJ annotation derived from GenBank entries (Avadis NGS). Aligned reads were mapped to the RefSeq transcripts database and normalized by DESeq (21). Gene expression is presented as read density defined as normalized counts per kb of exon. All RNA sequencing (RNAseq) data are deposited in ArrayExpress (https://www.ebi.ac.uk/arrayexpress) with accession number E-MTAB-2499.

Splenic B cells were isolated using magnetic negative depletion with PE-conjugated Abs against CD43, CD11b, CD11c, and TCRβ and anti-PE beads. Splenic and lymph node CD4⁺ T cells from OT-II Rag1–deficient mice were isolated using depletion with PE-conjugated Abs against CD11b, CD11c, CD8, and B220. To activate B cells and deliver Ag to the BCR, we stained B cells with 2.5, 0.25, 0.025, or 0 μg/ml anti–IgM F(ab)₂-biotin and then with OVA Ag delivery reagent (Miltenyi Biotec). A total of 2 × 10⁵ B cells were cocultured with 10⁵ OT-II CD4⁺ T cells in 96-well plates for 72 h. T cell activation was assessed by IL-2 production and measured in the supernatants using the DuoSet mouse IL-2 ELISA (R&D Systems).

Splenocytes were stained on ice with 10 μg/ml anti-IgM F(ab)₂-biotin, incubated at 37°C for the indicated times, and BCR internalization stopped with PBS at 0°C. Remaining surface BCR was stained with streptavidin-PE and analyzed by flow cytometry.

Mice were infected intranasally with 30 μl PBS containing X31 influenza 50% tissue culture infective dose = 8 × 10³ or 8 × 10⁴ and monitored daily for weight loss and clinical symptoms for 15 d and bled at day 10. Mice that reached 75% of their starting weight were euthanized.

Total serum Ig levels were quantified by ELISA with the SBA Clonotyping System-B6/C57J-HRP (Southern Biotech) according to the manufacturer’s instructions and were revealed with 3,3′,5,5′-tetramethylbenzidine. Signals were quantified using OD at 450 nm of serial dilutions, taking values in the linear portion of the response curve to calculate Ab levels. NP-specific and CTX-specific ELISAs were performed analogously, but plates were coated with 5 μg/ml NP₁₈-BSA (BioSearch Technologies) or 1 μg/ml CTX, respectively. For determination of anti-X31 HA-IgG, plates were coated with 2.5 μg/ml of purified X31 HA (John Skehel, National Institute for Medical Research). Serial dilutions of heat-inactivated sera samples were detected with HRP-conjugated goat anti-mouse IgG (Bio-Rad) and 3,3′,5,5′-tetramethylbenzidine.

Statistical analysis was carried out using a Mann–Whitney U test.

We first measured expression of Themis-family genes in the B cell lineage. RNAseq analysis of mRNA isolated from mouse splenic follicular B cells showed expression of Themis2, but no detectable expression of either of the closely related Themis1 or Themis3 genes, or of Garem and Gareml (Fig. 1A). To extend this analysis, we examined expression of the Themis-family genes in developing subsets of B lineage cells in the bone marrow (pro-B, pre-B, immature B, and mature recirculating B cells) and the spleen (transitional type 1 [T1], T2, and T3 B cells), and in mature B cell subsets in the spleen (follicular and marginal zone B cells), lymph nodes (follicular B cells), and peritoneal cavity (B1 and B2 cells). As a positive control for expression of Themis1, we used mRNA isolated from the thymus and from splenic and lymph node T cells, and for Themis3 we used mRNA from the small intestine. We found that Themis2 is expressed in all B cell subsets, both developing and mature, with very little or no detectable expression of either Themis1 or Themis3 (Fig. 1B). In contrast, thymocytes and T cells express Themis1, but not Themis2 or Themis3, and the small intestine expresses exclusively Themis3. Finally, we examined expression of Themis-family genes in Ag-activated B cells. Once again we found expression of Themis2, but not Themis1 or Themis3, in germinal center B cells, IL-10–secreting B cells (B10 cells), plasmablasts, and plasma cells in the spleen and in plasma cells in the bone marrow (Fig. 1C). However, we noted that expression of Themis2 was significantly lower in germinal center B cells and B10 cells compared with follicular B cells (p = 0.017 and 0.006, respectively). Thus, of the Themis-family, only Themis2 is expressed in the B cell lineage, with expression detectable in all B cell subsets, although the levels are reduced in Ag-activated B cells.

To study the function of Themis2 in B cells, we generated Themis2-deficient mice on a C57BL/6 background carrying the Themis2^{tm1d(KOMP)Wtsi} (Themis2^KO) allele in which exon 4 is deleted. Themis2^KO/KO mice were viable and produced at expected ratios from heterozygous parents. Exon 4 is the largest exon in the Themis2 gene and encodes the C terminus of the first CABIT domain, the whole second CABIT domain, NLS, and PRR. Deletion of exon 4 leads to a frameshift between exons 3 and 5, and is thus predicted to eliminate expression of the C terminus of the protein, including Y660, the proposed SH2 domain-binding site.

Analysis of Themis2 mRNA in splenic B cells from Themis2^KO/KO mice showed, as expected, no detectable mRNA spanning exons 3–4 and exons 4–5 (Fig. 2A). In contrast, mutant B cells contained normal levels of mRNA spanning exons 2–3 and exons 5–6, and showed splicing from exons 3–5, which was not detectable in wild-type (WT) B cells (Fig. 2A, Supplemental Fig. 2A). We conclude that the effect of the Themis2^KO mutation is to generate a truncated transcript in which exon 3 is spliced to exon 5. Immunoblot analysis of Themis2^KO/KO splenic B cell extracts probed with Abs to epitopes in exons 4 and 6 showed no detectable full-length or truncated Themis2 protein (Fig. 2B, 2C). Thus, the mutant allele has eliminated expression of full-length Themis2. However, in the absence of Abs to epitopes in exons 1–3, we cannot exclude that a short N-terminal fragment of Themis2 is still generated, although we note that such a hypothetical protein would contain only part of the first CABIT domain, and no other domain of Themis2.

To investigate whether Themis2 is required for B cell development, we used flow cytometric analysis to quantitate the numbers of developing and mature B cells in lymphoid tissues of Themis2^KO/KO mice. We found that the numbers of pro-B, pre-B, immature B, and mature B cells in the bone marrow, and T1, T2, and T3 cells in the spleen were unaltered in the mutant mice (Fig. 3A, Supplemental Fig. 1). Furthermore, there were also no changes in the numbers of mature follicular and marginal zone B cells in the spleen, B1a, B1b, and B2 cells in the peritoneal cavity, and B cells in blood, Peyer’s patches, and peripheral and mesenteric lymph nodes. Finally, we examined Ag-experienced B cells and found that loss of Themis2 did not change the number of plasma cells in the bone marrow, and numbers of germinal center B cells, B10 cells, plasmablasts, and plasma cells in the spleen. We also saw no effect on cell-surface levels of marker proteins on any of these cell types (Supplemental Fig. 1). To extend this analysis, we generated radiation chimeras reconstituted with mixtures of Themis2-deficient and WT bone marrow. Such mixed chimeras provide a competitive environment in which small defects in B cell development are more likely to be apparent. However, these studies once again showed no defect in the ability of Themis2-deficient progenitors to contribute to developing or mature B cell subsets, irrespective of whether the mutant cells made up 50 or 20% of the hematopoietic system (Supplemental Fig. 2B).

To analyze the development of Themis2-deficient B cells further, we used RNAseq to compare genome-wide gene expression between WT and mutant follicular B cells. This analysis showed only three genes that were differentially regulated >2-fold between the two genotypes (Fig. 3B, Supplemental Table 1A). We used these data to also examine usage of variable (V), diversity (D), joining (J), and constant (C) region genes for both heavy (H) and light (κ and λ) Ig chains. Alterations in usage of the V, D, or J region genes could indicate changes in positive or negative selection processes during B cell development. However, we found that usage of V_H, D_H, J_H, V_κ, J_κ, V_λ, and J_λ genes was largely identical between WT and Themis2-deficient B cells, as was usage of C_H, C_κ, and C_λ genes (Fig. 3C, Supplemental Fig. 3). Taken together, these results indicate that B cell development proceeds normally in Themis2^KO/KO mice.

To evaluate a potential role for Themis2 during B cell activation, we first analyzed expression of Themis-family genes in B cells after stimulation in vitro with anti-IgM, LPS, or CD40L and IL-4. We found that the expression of Themis2 decreased rapidly in response to all three stimuli, reaching a low point after around 6 h, after which the expression increased again (Fig. 4A, 4B). The levels of all other Themis-family genes (Themis1, Themis3, Garem, and Gareml) remained very low or undetectable. Furthermore, RNAseq analysis of Themis2^KO/KO B cells activated by anti-IgM, LPS, or with CD40L and IL-4, showed that no other Themis-family gene was upregulated to compensate for loss of Themis2 (Fig. 4B). Thus, in activated B cells, Themis2 is the only Themis-family gene expressed, and even in the absence of Themis2, no other family member is expressed.

Next, we examined the upregulation of cell-surface markers on WT and Themis2^KO/KO B cells after activation by anti-IgM, LPS, or with CD40L and IL-4. We found that in response to a range of different doses of these stimuli, the upregulation of CD69, CD23, CD86, and MHC class II was no different in Themis2-deficient B cells compared with WT B cells (Fig. 4C). Furthermore, lack of Themis2 also did not affect either the proliferation or survival of B cells in response to these stimuli (Fig. 4D). BAFFR, a receptor for the cytokine BAFF, transduces signals critical for B cell survival in part via the BCR (19, 22), and thus if Themis2 is involved in BCR signaling, it may also be required for BAFFR signaling. We investigated this possibility by treating WT or Themis2-deficient B cells with BAFF, but found that loss of Themis2 did not affect upregulation of CD86 or MHC class II, or BAFF-induced cell survival (Fig. 4C, 4D).

To extend this analysis, we measured secretion of 16 different cytokines (CCL2, CXCL1, CXCL2, GM-CSF, IFN-γ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, IL-13, IL-17, TNF-α, and VEGF) after in vitro stimulation of B cells with anti-IgM, LPS, or with CD40L and IL-4. Depending on the stimulus, we were able to detect substantial secretion of only four of these cytokines (IL-6, IL-10, TNF-α, and VEGF), but in all cases, loss of Themis2 did not affect the secretion of these proteins (Fig. 5A). RNAseq studies to measure gene expression in mutant or WT B cells stimulated with anti-IgM, LPS, or CD40L and IL-4 showed that in these 3 conditions, only 3, 8, and 17 genes, respectively, were differentially expressed >2-fold between Themis2-deficient and control B cells (Fig. 5B, Supplemental Table 1B–D). In conclusion, Themis2^KO/KO follicular B cells are transcriptionally very similar to WT B cells, both before and after activation.

Finally, we examined the ability of Themis2-deficient B cells to internalize and present Ag. Stimulation of splenic B cells with anti-IgM to cross-link the BCR led to internalization of the receptor at the same rate in both WT and Themis2-deficient cells (Fig. 5C). To measure the ability of Themis2^KO/KO B cells to present Ag to T cells, we treated B cells with an anti–IgM-OVA conjugate and then incubated them with OT-II CD4⁺ T cells bearing a monoclonal TCR specific for an OVA peptide (18) and assayed IL-2 secretion from the T cells as a readout of T cell activation. We found that at three different OVA doses, Themis2-deficient B cells were able to present Ag just as efficiently as WT B cells (Fig. 5D). Taken together, these results show that loss of Themis2 does not perturb B cell activation in vitro or their ability to present Ag to T cells.

Next, we investigated whether Themis2 was required for in vivo Ab responses. Themis2-deficient mice had normal levels of serum Ig isotypes (Fig. 6A). Immunization of mice with the T-independent type I Ag NP-LPS showed that WT and Themis2-deficient mice induce very similar levels of NP-specific IgM, IgG1, and IgG3 (Fig. 6B). Similarly, the Ab response of Themis2-deficient mice to the T-independent type II Ag NP-Ficoll was also normal (Fig. 6C). Analysis of mice immunized with NP-CGG (chicken γ-globulin) in alum, a T-dependent Ag, showed that Themis2^KO/KO mice mount a normal anti-NP IgM and IgG1 response (Fig. 6D). The same mice were rechallenged on day 63 with NP-CGG in PBS, and both mutant and control mice showed robust Ag-specific secondary responses (Fig. 6D). Finally, we evaluated the ability of Themis2-deficient mice to mount a mucosal Ab response to orally delivered CTX. Once again, Themis2^KO/KO mice showed a normal CTX-specific Ab response both in the serum and in feces (Fig. 6E, 6F). Taken together, these data clearly demonstrate that Themis2 is dispensable for Ab responses, at least to model Ags.

Finally, we assessed the ability of Themis2-deficient mice to mount an immune response to a pathogen. Control and Themis2^KO/KO mice were challenged intranasally with influenza virus, using either a low dose of virus from which most C57BL/6 mice survive, or a high dose, which is usually lethal. With both doses, Themis2-deficient mice showed weight loss and mortality rates that were similar to WT mice, and both genotypes of mice generated similar levels of serum IgG specific to influenza HA protein (Fig. 7A–C). Thus, Themis2 does not appear to play an important role in the immune response to infection by influenza virus. Because Themis2 is expressed in myeloid cells such as macrophages and dendritic cells, in addition to B cells, and these cell types play important roles in an antiviral immune response (23), we conclude that Themis2 plays no critical role in the antiviral immune response, both in B cells and in myeloid cells.

Previous studies demonstrating a critical role for Themis1 in T cell development and activation, and the reported reciprocal mutually exclusive expression of Themis1 in T cells and Themis2 in B cells, led us to hypothesize that Themis2 may play an analogous important role in B cell development, activation, or function. Such a possible related function was further supported by the report that ectopic expression of Themis2 in the T cell lineage rescues thymocyte development in Themis1-deficient mice (7). We show in this article that the only Themis-family member expressed in B lineage cells is Themis2, with no expression of either Themis1 or Themis3, or of the single CABIT domain proteins Garem and Gareml. Furthermore, we show that this exclusive expression is seen in all subsets of B cells examined, including developing B cells in the marrow, mature B cells in the spleen and lymph nodes, B1 cells in the peritoneal cavity, and in activated B cell subsets. Comparison of Themis2 expression showed that germinal center B cells and B10 cells have the lowest level of expression, suggesting that activation of B cells results in a decrease in Themis2 expression. In agreement with this, we found that activation of B cells in vitro with anti-IgM, LPS, and CD40L+IL-4 results in a rapid decline of Themis2 expression, with no compensating upregulation of expression of any other Themis-family gene. Similarly, previous studies had shown that expression of Themis2 in macrophages is reduced in mice undergoing infection or inflammation (6).

To investigate the function of Themis2, we generated a mouse strain deficient in Themis2 by deleting exon 4 of Themis2. The resulting Themis2 allele is still transcribed, resulting in an mRNA with exon 3 spliced to exon 5, which is predicted to result in a frameshift and subsequent termination of translation in exon 6. As expected, immunoblot analysis of B cell extracts from these mutant mice showed that no full-length or truncated Themis2 protein was detectable using Abs recognizing epitopes in exons 4 and 6. It remains possible that an N-terminal fragment of Themis2 is still made in this strain, although we were unable to test for this in the absence of Abs to epitopes encoded by exons 1, 2, or 3 of Themis2. However, such a hypothetical truncated Themis2 protein would include only the N-terminal part of the first CABIT domain of Themis2. It would not contain the rest of this domain, the second CABIT domain, NLS, PRR, or Y660.

It is unlikely that such a truncated protein would be functional for several reasons. First, the PRR and Y660 of Themis2 are required in macrophages for binding of Themis2 to Grb2 and Lyn, respectively (6). Second, mutation of either the PRR or Y660 impairs the ability of Themis2 to promote LPS-induced release of TNF-α. Third, the PRR of Themis1 is required for binding to Grb2 in T cells and for the recruitment of Themis1 to the LAT adapter protein, and hence the immunological synapse (12). Fourth, the CABIT domains, NLS and PRR of Themis1, are required for normal T cell development (11). Finally, a point mutation in the C-terminal end of the second CABIT domain of Themis1 leading to a premature stop codon also abrogates thymocyte development, demonstrating that a Themis1 protein containing only the first CABIT domain is not sufficient to maintain normal function (1). Taken together, these results suggest that if a truncated Themis2 protein is generated from the Themis2^KO allele, it is unlikely to be functional.

Analysis of Themis2^KO/KO mice showed that, surprisingly, Themis2 is not required for normal B cell development. No developmental defect was seen even under competitive conditions, where small deficits are more likely to be visible. Analysis of V, D, and J gene usage for both H and L chain Ig genes showed no significant changes in Themis2-deficient mice, indicating that loss of Themis2 did not measurably affect B cell selection. Similarly, loss of Themis2 did not affect in vitro activation of B cells in response to stimulation through the BCR, TLR4, CD40, or IL4R, and did not perturb BCR internalization or Ag presentation to T cells. Finally, Themis2 was not required for Ab responses either to T-independent or T-dependent model Ags, or to influenza virus during the immune response to infection.

This lack of phenotype could potentially be because of functional redundancy between Themis2 and other related proteins. However, RNAseq analysis showed that Themis2-deficient B cells did not upregulate expression of other Themis-family genes. Thus, B lineage cells in Themis2^KO/KO mice do not express any Themis-family member, making it unlikely that the lack of phenotype in Themis2^KO/KO mice is due to redundancy with Themis-family proteins. Another possibility for the lack of phenotype is that the Themis2 gene is nonfunctional. This is unlikely for two reasons. First, the Themis2 gene is found in all vertebrates, indicating that Themis2 is likely to have an important function (1). Second, it is possible that the Themis2 gene in laboratory mice has become nonfunctional. However, this is also unlikely because ectopic expression of mouse Themis2 in the T cell lineage can rescue T cell development in Themis1-deficient mice (7). Taken together, these observations suggest that Themis2 has an important physiological role, but not in the processes examined in this study. We note that Themis2 is expressed in myeloid cells, and it may be that Themis2 is more important in these cells.

In summary, we have shown that Themis2 is expressed in all subsets of developing, mature and activated B cells. Surprisingly, however, loss of Themis2 does not perturb B cell development, activation, or Ab responses. Further studies are needed to identify the physiological function of this conserved protein.

We thank the National Institute for Medical Research Flow Cytometry, High Throughput Sequencing and Biological Services; Nikolai Nikolov for RNAseq analysis; Matteo Villa for help with CTX immunizations; and David Tarlinton and Victor Peperzak for advice.

The authors have no financial conflicts of interest.