Processing of CD74 by the Intramembrane Protease SPPL2a Is Critical for B Cell Receptor Signaling in Transitional B Cells

The invariant chain (CD74) is an essential component of the MHC class II (MHCII) Ag presentation pathway facilitating assembly and endosomal targeting of the MHCII complexes (1). CD74 binds to the MHCII peptide-binding groove with its CLIP segment and prevents premature peptide binding. In Ag-processing compartments, the luminal domain of CD74 is sequentially degraded by endosomal proteases, thereby releasing MHCII. Interference with this process, for example by the absence of the cysteine protease cathepsin S, significantly impairs Ag presentation by MHCII (1). Beyond its canonical function as chaperone of MHCII, CD74 was reported to modulate endocytic trafficking and endosomal maturation in APCs (2–4). Furthermore, the regulation of dendritic cell migration (5) and also the cellular responses to the proinflammatory cytokine macrophage migration inhibitory factor are dependent on CD74 as cell surface receptor (6). Coreceptors such as CD44 (7) or CXCR4 (8) are required to initiate signal transduction following migration inhibitory factor binding to CD74. Interestingly, enhanced expression levels of CD74 have been observed in different forms of malignancies and are positively correlated with increased cellular migration and/or invasion (9–11). In this context, overexpression of CD74 in HEK293 cells activates central signaling pathways such as the PI3K/Akt and the MAPK/ERK cascades (12).

We recently found that clearance of the final membrane-bound N-terminal fragment (NTF) of CD74 requires activity of the GxGD-type intramembrane protease signal peptide peptidase-like (SPPL)2a (13), but not of its closely related homolog SPPL2b (14). This unique role of SPPL2a is also conserved in human B cells (15). SPPL2a is part of the SPP/SPPL protease family with specificity for transmembrane proteins in type II orientation (16). Within this family, SPPL2a is unique with its localization in lysosomes/late endosomes (14, 17). Additional substrates of SPPL2a include TNF-α, Fas ligand, Bri2, and TMEM106B (16, 18).

In SPPL2a-deficient B cells, unprocessed CD74 NTF accumulates in significant amounts, leading to an arrest of B cell maturation at the transitional stage 1 (T1) in mice (13). Furthermore, functionality of the remaining B cells is considerably impaired with regard to Ig production. Major phenotypic findings described in this study were additionally confirmed in two SPPL2a mutant mouse lines (19, 20). Additional ablation of CD74 in SPPL2a^−/− mice improved B cell maturation and function, indicating that the B cell phenotype of these mice can be mainly ascribed to the incomplete turnover of the CD74 NTF (13). In SPPL2a^−/− B cells, the unprocessed CD74 NTF accumulates in endosomal compartments where it disturbs membrane trafficking, leading to a highly increased abundance of endosome-derived vacuoles in these cells. Hence, control of CD74 NTF levels by the intramembrane protease SPPL2a is essential for B cell maturation and survival as well as the integrity of the endocytic system. In light of these findings, pharmacological inhibition of SPPL2a has been proposed as a putative therapeutic concept to achieve B cell–directed immunosuppression. However, the molecular details of how the accumulating CD74 NTF impairs B cell development remained unclear to date.

To address this issue, we analyzed the impact of SPPL2a deficiency and the subsequent CD74 NTF accumulation on signaling pathways involved in B cell survival and differentiation and identified impairment of the PI3K/Akt signaling pathway as a major cause of the B cell maturation defect of SPPL2a-deficient mice. Particularly, we show that tonic as well as BCR-induced activation of the kinase Akt is significantly compromised in SPPL2a^−/− B cells. Altered BCR trafficking causes a reduction of surface IgM in these cells and a subsequent impaired signal transmission via the BCR and the tyrosine kinase Syk. We reveal that disturbances of the PI3K/Akt pathway are not directly coupled to surface BCR depletion and impaired Syk activation. In agreement with an impaired tonic PI3K/Akt signaling, SPPL2a^−/− B cells show enhanced transcription of proapoptotic genes caused by dysregulation of the FOXO1 transcription factor. Thus, we demonstrate that SPPL2a-mediated processing of CD74 NTF is indispensable for BCR-mediated signal transduction especially via the PI3K/Akt pathway to support maturation and survival of B cells.

Generation of SPPL2a^−/− mice has been described previously (13). Mice used in this study were backcrossed for 10 generations on a C57BL/6N Crl background. CD74-deficient mice (B6.129S6-Cd74^tm1Liz/J) (21) were obtained from The Jackson Laboratory and crossed with SPPL2a^−/− mice for generation of SPPL2a-CD74 double-deficient mice as described before. Animal care and handling were performed in agreement with local and national guidelines.

Single-cell suspensions of splenocytes were obtained by cutting the spleens into pieces and passing them through a 100 μm cell strainer in MACS buffer (0.5% [w/v] BSA and 2 mM EDTA in PBS). Splenic B220⁺ B cells were isolated by positive selection using B220 MicroBeads and LS columns of the MACS cell separation system (Miltenyi Biotec) according to the manufacturer’s recommendations. Purity of the obtained cell suspension was assessed by flow cytometry using anti–CD45R-allophycocyanin (eBioscience) and found to be between 80 and 90%. For the analysis of T1 B cells by Western blotting and quantitative real-time PCR (qRT-PCR), splenic T1 B cells (B220⁺CD21^lowCD24^high) were obtained by staining of B220⁺ MACS-isolated splenic B cells with anti-CD21/35-PE (8D9, eBioscience), anti-CD24–allophycocyanin (M1/69, BioLegend), and anti-CD45R–PE-Cy7 (RA3, 6B2, eBioscience) (B220) and subsequent FACS using a FACSAria (BD Biosciences).

For treatment with SPPL inhibitors, freshly isolated wild-type splenocytes were cultivated overnight at 37°C in RPMI 1640 with l-glutamine (Sigma-Aldrich) supplemented with 10% FCS (Biochrom), 50 μM 2-ME (Life Technologies), 100 U/ml penicillin (Sigma-Aldrich), and 100 μg/ml streptomycin (Sigma-Aldrich) (RPMI complete) in the presence of 10 μM (Z-LL)₂-ketone (PeptaNova) and 1 μM inhibitor X (EMD Millipore).

To enrich for B220⁺ IgM⁺ immature B cells, bone marrow cells were cultured in IMDM supplemented with GlutaMAX (Life Technologies), 10% FCS, 50 μM 2-ME, 1× nonessential amino acids (Sigma-Aldrich), 100 U/ml penicillin, and 100 μg/ml streptomycin in the presence of 5 ng/ml IL-7 (BioLegend) for 4 d as described in Rowland et al. (22). Prior to further analysis of signaling pathways, cells were washed twice in PBS and cultured overnight in fresh complete IMDM without IL-7.

Activation of the BCR was achieved by incubating the cells in the presence of 10 μg/ml goat anti-mouse IgG/IgM F(ab′)₂ (Dianova). For analysis of signal transduction pathways, cells were stimulated in 500 μl RPMI 1640 medium (Sigma-Aldrich) or 500 μl PBS for 5 min at 37°C followed by Western blot or flow cytometric determination, respectively, of kinase activation (Syk, ERK1/2, Akt). Depending on the applied readout system, stimulation experiments were either performed with purified splenic B220⁺ B cells (Western blotting) or with unfractionated splenocytes [freshly isolated or cultivated overnight with (Z-LL)₂-ketone and inhibitor X] or enriched B220⁺IgM⁺ from bone marrow cultures (flow cytometry). To investigate the effect of BCR stimulation on the expression of downstream target genes via qRT-PCR, splenic B220⁺ B cells were stimulated for 3 h in RPMI complete medium.

For the analysis of surface IgM in splenic B cell subsets, splenocytes were stained for 30 min at 4°C with anti–CD21/35-PE (8D9), anti–CD45R-PE-Cy7 (RA3-6B2), and anti–CD24-FITC (30-F1) along with anti–IgM-allophycocyanin (II/41) (all eBioscience) in MACS buffer. For detection of total cellular IgM, anti-IgM staining was performed after fixation and permeabilization of cells using the BD Cytofix/Cytoperm fixation/permealization solution kit (BD Biosciences) according to the manufacturer’s instructions. Erythrocytes were lysed by incubation in FACS lysing solution as described before (13). Similarly, intracellular CD74 levels were determined in fixed and permeabilized splenocytes by labeling with anti–CD74-FITC (In-1, BD Biosciences) in combination with anti–CD21/35-PE (8D9), anti–CD45R-PE-Cy7 (RA3-6B2), and anti–CD24-allophycocyanin (M1/69).

To detect the phosphorylation level of the kinases Akt, ERK1/2, and Syk, cells stimulated in 500 μl PBS were immediately chilled on ice, fixed for 20 min by the addition of 1.5 ml 5% (w/v) paraformaldehyde (PFA) in PBS, and subsequently permeabilized in 90% methanol for 10 min on ice. Cells were incubated in PBS supplemented with 0.5% BSA for 10 min and labeled with rabbit mAbs against pERK1/2 (Thr²⁰²/Tyr²⁰⁴, D13.14.4E), ERK1/2 (137F5), pAkt (Ser⁴⁷³, D9E), Akt (C67E7), pSyk (Tyr⁵²⁵/⁵²⁶, C87C1), or Syk (D3Z1E) (all Cell Signaling Technology) together with anti–CD21/CD35-PE (8D9), anti–CD24-allophycocyanin (M1/69), and anti–CD45R-PE-Cy7 (RA3-6B2) for 1 h on ice in the same buffer. Activated and total levels of the kinases were then detected following incubation with the Alexa Fluor 488 goat anti-rabbit IgG secondary Ab (Life Technologies) for 30 min on ice and lysis of erythrocytes.

For the analysis of cellular phosphatidylinositol 3,4,5-trisphosphate (PIP₃) levels, cells were fixed with a final concentration of 1.5% PFA in PBS for 10 min similar to Anzelon et al. (23). Cells were then permeabilized for 10 min in PBS supplemented with 1% BSA and 0.2% saponin and stained with biotinylated anti-PIP₃ (Echelon Biosciences) for 30 min on ice in the same buffer. Detection of PIP₃ was achieved by incubation with streptavidin-allophycocyanin (eBioscience) for 30 min on ice along with anti–CD21/CD35-PE (8D9), anti–CD24-FITC (30-F1), and anti–CD45R-PE-Cy7 (RA3-6B2) to distinguish splenic B cell subsets.

In all cases labeled cells were analyzed using a FACSCanto flow cytometer (BD Biosciences) and data were evaluated with the FlowJo (Tree Star) software.

For the analysis of fluid-phase endocytosis, 2.4 × 10⁷ splenocytes were pulsed with 250 μg/ml OVA-FITC (Life Technologies) for 30 min at 37°C in RPMI complete medium. Subsequently, cells were washed three times in cold PBS, resuspended in prewarmed medium, and chased for 0 min to 24 h at 37°C. Subsequently, the chase was stopped by the addition of ice-cold PBS. To distinguish B cell subsets, cells were stained with anti–CD21/35-PE (8D9), anti–CD45R-PE-Cy7 (RA3-6B2), and anti–CD24-allophycocyanin (M1/69) prior to erythrocyte lysis and flow cytometric analysis. The median fluorescence intensity (MFI) of the FITC channel was determined for T1 B cells (B220⁺CD21^lowCD24^high) that were identified based on the costained markers. To control for signal resulting from unspecific OVA binding, the MFI of cells that had been incubated on ice in the presence of OVA under the same conditions was subtracted from the MFIs of the individual time points. The resulting ΔMFIs were then normalized for each genotype separately to the ΔMFI measured directly after the pulse (0 h chase). To block OVA-FITC degradation, lysosomal acidification or lysosomal proteases were inhibited by addition of 40 nM bafilomycin (Sigma-Aldrich) or a mix of 100 μM leupeptin (Carl Roth), 1 μg/ml pepstatin A (Biomol), and 50 μM E64-d (Cayman Chemical) to the assay following a 2 h preincubation of the cells in the presence of these compounds.

To determine the kinetics of BCR internalization, splenocytes were incubated with biotin-conjugated anti-mouse IgM (II/41, eBioscience) for 30 min on ice in MACS buffer. Cells were washed twice with PBS and resuspended in prewarmed RPMI complete medium. Internalization of the BCR complex was allowed for 0–30 min by incubation at 37°C. Endocytosis was stopped immediately by the addition of ice-cold PBS. Finally, cells were stained with streptavidin-allophycocyanin, anti–CD21/35-PE (8D9), anti–CD24-FITC (30-F1), and anti–CD45R-PE-Cy7 (RA3-6B2) and analyzed flow cytometrically after erythrocyte lysis. The MFI of the allophycocyanin channel was determined for T1 B cells normalized to the MFI of the 0 min time point of each genotype.

Splenocytes isolated from wild-type and SPPL2a-deficient mice were labeled and isolated by positive selection in MACS buffer using IgM MicroBeads and LS columns of the MACS cell separation system according to the manufacturer’s recommendations. Isolation was carried out strictly at 4°C to avoid any internalization of the IgM-labeled BCR complex. Cells were resuspended in prewarmed RPMI complete medium and internalization of the Microbead–BCR complex was allowed for 20 min at 37°C. Subsequently, cells were fixed in 4% PFA and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) in suspension for 90 min at room temperature, transferred to 2% PFA in 0.1 M phosphate buffer (pH 7.4), and stored at 4°C until further processing. For postembedding immunogold labeling, small pieces of cryoprotected B cell pellets (2.3 M sucrose) were mounted on specimen holders and immersed in liquid nitrogen. Ultrathin sections (70 nm) were cut and labeled according to Slot and Geuze (24). Briefly, they were collected on carbon-Formvar–coated nickel grids (Science Services) and incubated in a series of solutions as described (24). For detection of CD74, a rabbit polyclonal Ab was generated against a synthetic peptide comprising amino acids 2–27 of murine CD74 (Pineda Antikörper Service, Berlin, Germany). Anti-CD74 was applied (1:200) and detected with 15-nm large gold probes coupled to protein A (G. Posthuma, University Medical Center Utrecht). Sections were examined in an EM902 (Zeiss) and images were taken with a MegaView III digital camera (A. Tröndle).

Splenic B220⁺ B cells were adhered to poly-l-lysine–coated coverslips (13) for 30 min at 37°C in RPMI 1640 supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin and fixed with 4% PFA for 20 min. Immunofluorescent staining of cells was performed as described previously (13). The BCR was visualized with Alexa Fluor 594–conjugated goat anti-mouse IgM (Life Technologies). In parallel, rabbit polyclonal anti-MHCII β-chain (gift of Prof. Willem Stoorvogel) in combination with a goat anti-rabbit Alexa Fluor 488 conjugate (Life Technologies) was used.

For visualization of endocytosed BCR by indirect immunofluorescence, splenic B220⁺ B cells were adhered to poly-l-lysine–coated coverslips as described above. Cells were cooled down and the medium was replaced by ice-cold RPMI complete medium containing 10 μg/ml Alexa Flour 594–conjugated goat anti-mouse IgM (Life Technologies). Staining of the surface-localized BCR was performed for 30 min at 4°C followed by two washing steps with ice-cold PBS. Subsequently, cells were resuspended in prewarmed RPMI complete medium, and internalization of the IgM–BCR complex was allowed for 30 min at 37°C. Cells were fixed with 4% PFA for 20 min and further processed as described above. Images were acquired with a FV1000 confocal laser scanning microscope (Olympus).

For proximity ligation assay (PLA) probes against specific targets, the following unlabeled Abs were used: anti-CD74 (In-1, BD Biosciences), anti-Igα (HMK7/A9, Abcam), Syk (Syk-01, BioLegend), and anti-MHCII β-chain (W. Stoorvogel). Fab fragments were prepared with Pierce Fab Micro preparation kit (Thermo Scientific) using immobilized papain according to the manufacturer’s protocol. After desalting (Zeba spin desalting columns, Thermo Scientific), the Fab fragments were coupled with PLA Probemaker Plus or Minus oligonucleotides (Sigma-Aldrich) to generate Fab-PLA probes. Splenocytes were obtained from wild-type, SPPL2a^−/−, and SPPL2a^−/− CD74^−/− mice as described above. Naive B cells were enriched by MACS depletion using anti-CD43 MicroBeads (Miltyeni Biotec) according to the manufacturer’s protocol. Prior to stimulation, the enriched B cells were rested overnight in complete RPMI 1640 medium. For in situ PLA, B cells were settled on polytetrafluoroethylene slides (Thermo Fisher Scientific) for 30 min at 37°C. BCR stimulation was performed with 5 μg/ml goat anti-mouse IgM F(ab′)₂ (SouthernBiotech). Fixation, permeabilization, and PLA reactions were conducted according to a previously described protocol (25). Resulting samples were directly mounted on slides with DAPI Fluoromount-G (SouthernBiotech) to visualize the PLA signals in relationship to the nuclei. Microscopic images were acquired with a Zeiss 780 Meta confocal microscope (Carl Zeiss, Jena, Germany). For each experiment a minimum of 1000 cells from several images was analyzed using the BlobFinder software (Centre for Image Analysis, Uppsala University). Raw data were exported to Prism software (GraphPad Software). For each sample, the mean PLA signal count per cell was calculated from the corresponding images.

Cells were lysed in radioimmunoprecipitation assay buffer (25 mM Tris-HCl [pH 7.6], 150 mM NaCl, 1% [v/v] Nonidet P-40, 1% [w/v] sodium deoxycholate, 0.1% [w/v] SDS, 4 mM EDTA) supplemented with PhosSTOP phosphatase inhibitor cocktail (Roche) and protease inhibitors on ice for 1 h as described previously (26). Lysates were cleared by centrifugation and protein concentration was determined by bicinchoninic acid protein assay (Thermo Fisher Scientific). SDS-PAGE (27), semidry transfer onto nitrocellulose membrane, and immunodetection were performed as described before (26). Separation of proteins from T1 B cell lysates for the detection of p100/p52 was conducted using a Novex 4–12% Tris-glycine mini protein gel, 1.5 mm (Life Technologies). Abs against total and phosphorylated forms of ERK1/2, Akt, and Syk have been listed above. Furthermore, anti-pFOXO1 (Ser²⁵⁶), anti-FOXO1 (C29H4), anti-p100/p52 (all Cell Signaling Technology), anti–eukaryotic elongation factor 2 (EEF2) (ab33523, Abcam), and anti–β-tubulin (E7, DSHB) were used as primary and HRP-conjugated goat anti-rabbit and sheep anti-mouse (Dianova) as secondary Abs. Immunoblots were detected with Lumigen ECL Ultra (TMA-6, Lumigen) using an ImageQuant LAS 4000 (GE Healthcare).

Total RNA of B220⁺ splenic B cells was extracted using the NucleoSpin RNA kit (Macherey-Nagel) according to the manufacturer’s instructions. For total RNA isolation of FACS-sorted T1 B cells, the RNeasy micro kit (Qiagen) was used. For cDNA synthesis, 40–100 ng total RNA was transcribed using a first-strand cDNA synthesis kit (Thermo Scientific) and random hexamer oligonucleotides according to the manufacturer’s recommendations. Gene expression was quantified using the Universal ProbeLibrary technology (Roche) and the Universal ProbeLibrary Set mouse (04 683 641 001). Primers and probes were designed using the Assay Design Center (Roche) to amplify intron-spanning regions with an amplicon of <130 bp. Hydrolysis probes were labeled with the reporter dye fluorescein (5′ end) and a dark quencher dye (3′ end). The following oligonucleotides (Sigma-Aldrich) and hydrolysis probes were used: FOXO1 (forward, 5′-CTTCAAGGATAAGGGCGACA-3′, reverse, 5′-GACAGATTGTGGCGAATTGA-3′; probe no. 11), p21 (forward, 5′-TCCACAGCGATATCCAGACA-3′, reverse, 5′-GGACATCACCAGGATTGGAC-3′; probe no. 21), p27 (forward, 5′-GTTAGCGGAGCAGTGTCCA-3′, reverse, 5′-TCTGTTCTGTTGGCCCTTTT-3′; probe no. 62), Bim (forward, 5′-GGAGACGAGTTCAACGAAACTT-3′, reverse, 5′-AACAGTTGTAAGATAACCATTTGAGG-3′; probe no. 41), RAG1 (forward, 5′- AGGCCTGTGGAGCAAGGTA-3′, reverse, 5′-GCTCAGGGTAGACGGCAAG-3′; probe no. 46), RAG2 (forward, 5′-TGCCAAAATAAGAAAGAGTATTTCAC-3′, reverse, 5′-GGGACATTTTTGATTGTGAATAGG-3′; probe no. 4), EEF2 (forward, 5′-GTTGACGTCAGCGGTCTCTT-3′, reverse, 5′-GCACGGATCTGATCTACTGTGA-3′; probe no. 70), tubulin α (forward, 5′-CTGGAACCCACGGTCATC-3′, reverse, 5′-GTGGCCACGAGCATAGTTATT-3′, probe no. 88), succinate dehydrogenase α (forward, 5′-TGTTCAGTTCCACCCCACA-3′, reverse, 5′-TCTCCACGACACCCTTCTG-3′; probe no. 71). Samples containing 0.5 μl cDNA, 4.5 μl LightCycler 480 Probes Master (Roche), 0.1 μM hydrolysis probe, and 0.3 μM each oligonucleotide in a final volume of 10 μl were analyzed on a LightCycler 480 instrument II (Roche). All assays were performed in duplicate. With the exception of RAG1 and RAG2 as well as analyses of T1 B cells, primer efficiency (E) was calculated for each sample and gene: E = 10^−1/slope. The ΔCp (crossing point) was calculated by normalizing the Cp of each assay to the average Cp of the reference genes EEF2, tubulin α, and succinate dehydrogenase of the same assay. Finally, ΔΔCp was calculated (ΔΔCp = E^−ΔCp × 100) and normalized to unstimulated wild-type samples. Otherwise, E was adjusted to a value of 2.

Data are represented as means ± SD or ± SEM, as indicated. Statistical significance was assessed as indicated by using an unpaired two-tailed Student t test or one-way ANOVA followed by Newman–Keuls post hoc testing. Data from the PLA were evaluated with a two-tailed Mann–Whitney U test. A p value <0.05 was considered statistically significant.

In this study we aimed to provide deeper insights into the molecular mechanisms underlying the B cell maturation defect in SPPL2a^−/− mice and therefore specifically examined the activation status of signaling pathways implicated in B cell differentiation and survival. We had reported previously that SPPL2a-deficient B cells exhibited a reduced surface expression of the receptor for BAFF (BAFF-R) (13). Activation of this receptor stimulates the alternative NF-κB pathway by inducing processing of the NF-κB protein p100 into its transcriptionally active form p52 (28, 29). In light of the reduced BAFF-R in SPPL2a^−/− B cells, we analyzed p100 and p52 levels in these cells (Fig. 1A, 1B). We observed similar levels of the active p52 as compared with wild-type B220⁺ B cells. Thus, constitutive activation of the alternative NF-κB pathway was not significantly altered in SPPL2a^−/− B cells despite the lower BAFF-R levels. However, we found that the abundance of the uncleaved p100 precursor was significantly reduced in SPPL2a-deficient B cells. Both observations were confirmed when FACS-sorted T1 B cells (B220⁺CD21^lowCD24^high) from wild-type and SPPL2a^−/− mice were compared (Fig. 1C). BAFF-R expression (30) as well as the synthesis of the NF-κB pathway substrate p100 (31) are correlated and linked to tonic signaling of the BCR, which is essential for B cell differentiation and survival (32, 33). Therefore, we hypothesized that impairment of tonic BCR signaling provoked by SPPL2a deficiency represents the underlying cause of the B cell maturation defect and the reduced BAFF-R and p100 levels. Altogether, this prompted us to investigate the effect of SPPL2a deficiency on signal transduction pathways downstream of the BCR in more detail.

Upon BCR ligation, nonreceptor protein tyrosine kinases, primarily of the Src family, are rapidly activated, leading to phosphorylation of ITAMs in the cytoplasmic tail of the BCR subunits Igα and Igβ. The subsequent recruitment and activation of the spleen tyrosine kinase (Syk) by these phosphorylated ITAMs represents one of the initial steps of BCR signaling (34, 35). Therefore, we compared the levels of pSyk in wild-type and SPPL2a^−/− B cells that was induced by cross-linking of the BCR with anti–IgM/IgG F(ab′)₂ fragments (Fig. 2). In splenic B220⁺ cells from wild-type mice, BCR stimulation triggered an obvious increase of phosphorylated and thus activated Syk as determined by Western blotting (Fig. 2A). In comparison, the level of pSyk that was detected in B220⁺ cells from spleens of SPPL2a^−/− mice following BCR ligation was found to be lower in several experiments as depicted in Fig. 2A. However, this Western blot–based analysis of total splenic SPPL2a^−/− B cells revealed significant interexperimental variability (Fig. 2B). Due to the described B cell maturation defect of SPPL2a^−/− mice (13), the representation of different B cell subsets within the employed total B220⁺ splenic population differed between the two genotypes. To rule out any bias introduced by differences in the composition of the analyzed cell population, we repeated the experiment with flow cytometric analysis of Syk phosphorylation. Therefore, in addition to staining for phosphorylated or total Syk, cells were also labeled for B220, CD21, and CD24. Based on these markers, T1, T2, and mature B cells were identified according to the gating scheme depicted in Fig. 2C. Histograms from a representative experiment are depicted in Supplemental Fig. 1. Because T1 B cells had been shown to be the last stage prior to the maturation block that was present at a similar frequency in SPPL2a^−/− and wild-type mice, we compared Syk phosphorylation specifically for this cell population. In accordance with the tendency observed in B220⁺ B cells, pSyk levels following BCR stimulation were significantly lower in SPPL2a^−/− T1 B cells as compared with the respective population from wild-type mice (Fig. 2D).

We aimed to further assess the consequences of the decreased Syk activation in SPPL2a-deficient B cells on signal transmission via the BCR complex. Therefore, we analyzed effects on activation of the downstream MAPK and PI3K/Akt pathways. In SPPL2a^−/− splenic B220⁺ (Fig. 3A, 3B) as well as T1 (Fig. 3C) B cells, BCR-induced activation of the MAPKs ERK1/2 as determined by pERK1/2 was not significantly different from that in wild-type B cells. In contrast, we detected reduced levels of PIP₃ in stimulated SPPL2a^−/− B220⁺ (Fig. 3D) or T1 (Fig. 3E) B cells by flow cytometry, indicating impaired activation of PI3K. Histograms from a representative experiment are depicted in Supplemental Fig. 1. The altered generation of PIP₃ was reflected in a significant decrease of the phosphorylated and thus activated form of the kinase Akt in SPPL2a-deficient B cells. This applied both to total splenic B220⁺ (Fig. 3G) and T1 (Fig. 3H) B cells. Interestingly, in the T1 population already the basal Akt phosphorylation level in unstimulated cells was found to be significantly decreased in SPPL2a^−/− B cells. These findings suggest a significant, although differential, impairment of SPPL2a deficiency on pathways downstream of the BCR. The decreased basal pAkt level already detected without BCR stimulation points to a reduced tonic activation of the PI3K/Akt pathway in T1 B cells.

Having identified the accumulating CD74 NTF as a major driving force of the B cell maturation defect of SPPL2a^−/− mice (13), we clarified whether this also applies to the described disturbance of BCR-associated signaling pathways. We therefore specifically analyzed Syk and Akt phosphorylation at basal levels as well as following BCR ligation in double-deficient SPPL2a-CD74 in comparison with SPPL2a^−/− B cells (Fig. 4). For Syk, the activation induced by BCR stimulation in SPPL2a^−/− CD74^−/− B cells was not significantly different from that seen in wild-type B cells, irrespective of being analyzed in total splenic B cells (Fig. 4A, 4B) or the T1 population (Fig. 4C). Thus, the described impairment of signal transduction via Syk in SPPL2a^−/− B cells was reversed by additional ablation of CD74. A similar tendency was observed for ligand-induced activation of Akt in splenic B220⁺ cells analyzed by Western blotting (Fig. 4D, 4E). In T1 B cells, the BCR-induced PI3K/Akt signaling was significantly, although not entirely, recovered in SPPL2a^−/− CD74^−/− B cells (Fig. 4F). CD74 ablation did not lead to a major amelioration of the reduced basal PI3K/Akt activation in SPPL2a^−/− T1 B cells. This reflects that already CD74 single-deficient T1 B cells exhibited an impairment of basal and BCR-induced Akt phosphorylation in comparison with wild-type T1 B cells. Importantly, in all setups, activation of Syk and Akt in SPPL2a-CD74 double-deficient B cells was found to be comparable to CD74 single-deficient B cells, indicating that possible residual impairments of these signaling pathways in the double-deficient B cells may be caused by the absence of CD74. In conclusion, the described recovery of the impaired Syk and alleviation of Akt activation after BCR stimulation by additional CD74 ablation identify the CD74 NTF accumulation as an important link between SPPL2a deficiency and the disturbed BCR signaling.

Having demonstrated a critical role of the CD74 NTF, we went on to investigate how this protein fragment inhibits BCR signaling. Because CD74 influences endosomal maturation (2–4) and SPPL2a-deficient B cells exhibit major morphological changes in their endocytic system caused by the CD74 NTF (13), we considered a possible impact on trafficking of the BCR.

Therefore, we analyzed the subcellular distribution of the BCR in isolated splenic B220⁺ B cells from wild-type and SPPL2a^−/− mice by immunocytochemistry (Fig. 5A). In B cells of both genotypes, staining of the BCR was seen at the plasma membrane as well as in intracellular compartments where partial colocalization with MHCII was observed. Based on this microscopical analysis, the steady-state distribution of the BCR did not appear to be grossly altered by SPPL2a deficiency. We performed flow cytometric analysis of splenocytes from wild-type and SPPL2a^−/− mice with or without permeabilization prior to Ab staining to assess the distribution of the BCR between the cell surface and intracellular compartments more accurately (Fig. 5B, 5C). T1 B cells from SPPL2a^−/− mice exhibited a statistically significant reduction of cell surface IgM to ∼50% of the wild-type level (Fig. 5B, 5C). However, total cellular IgM levels determined after cell permeabilization were not different between wild-type and SPPL2a-deficient T1 B cells. In combination, these findings suggest a shifted subcellular distribution of IgM in SPPL2a^−/− cells characterized by an increased pool in intracellular compartments and fewer BCRs at the plasma membrane that are accessible for Ab-induced cross-linking. Therefore, this BCR redistribution may represent a major cause of the impaired signaling response induced by anti-IgM/IgG treatment. We also examined whether the reduction of IgM surface expression was associated with the CD74 NTF accumulation. Flow cytometric analysis revealed full restoration of IgM surface expression in SPPL2a-CD74 double-deficient B cells as compared with the SPPL2a single-deficient B cells (Fig. 5C). Interestingly, CD74^−/− and SPPL2a^−/− CD74^−/− T1 B cells exhibited significantly higher surface IgM levels than did wild-type B cells. These findings clearly indicate a major impact of the CD74 NTF on BCR trafficking in SPPL2a^−/− B cells.

Based on the described CD74 NTF-associated BCR redistribution and the impact of the CD74 NTF on endosomal morphology (13), we examined trafficking within the endocytic system of SPPL2a-deficient B cells in more detail (Fig. 6). Therefore, we analyzed uptake and degradation of fluid-phase cargo in wild-type and SPPL2a^−/− B cells. In the employed setup, cells were allowed to internalize FITC-conjugated OVA for 30 min (pulse) and then incubated in fresh medium for various time periods to allow lysosomal delivery and degradation of the cargo (chase), which was detected as decrease of the cell-associated FITC fluorescence. When inhibitors of lysosomal acidification (bafilomycin) or lysosomal proteases (leupeptin/peptstatin A/E64-d) were applied, the fluorescent signals were stabilized (Fig. 6A). Thus, the observed loss of cell-associated fluorescence represents a suitable readout for turnover of the FITC OVA in late endosomes/lysosomes. First, we compared the amount of the cargo that was internalized by wild-type and SPPL2a^−/− T1 B cells (Fig. 6B). No significant differences were observed. During the chase period, T1 B cells of both genotypes showed a time-dependent decrease in cell-associated fluorescence leading to similar endpoints after 24 h incubation (Fig. 6C). However, cargo degradation was significantly decelerated in SPPL2a^−/− T1 B cells. In contrast, the degradation rate in non–B cells (B220⁻) from spleens of wild-type and SPPL2a-deficient mice was similar (Fig. 6C), indicating that this phenotype was B cell specific at least among the major splenic cell populations. Because we did not observe any impact of SPPL2a deficiency on the acidification of lysosomes or the expression level and maturation of proteases in this compartment (not shown), the retarded cargo degradation in the SPPL2a^−/− B cells most likely reflects a delayed trafficking through the endocytic pathway.

To analyze whether this effect was dependent on the accumulating CD74 NTF, we included B cells from SPPL2a-CD74 double-deficient mice into the uptake assay (Fig. 6D). In SPPL2a^−/− CD74^−/− T1 B cells, the retardation of the degradation kinetics associated with SPPL2a deficiency was rescued to the level of CD74^−/− cells (Fig. 6D). This is illustrated by the calculated half-life of the endocytosed FITC-OVA (Fig. 6E). Whereas wild-type as well as CD74-deficient and SPPL2a-CD74 double-deficient B cells managed to degrade 50% of the cargo in a maximum of 2 h, SPPL2a^−/− T1 B cells required >4 h (Fig. 6E). This provides clear evidence that the CD74 NTF is a major cause of this trafficking impairment.

In addition to providing signals for B cell maturation and activation, a major function of the BCR is to deliver captured Ags to Ag-processing compartments. This requires internalization of the receptor after ligand binding (36–38). Therefore, we compared the rate and velocity of BCR internalization in SPPL2a^−/− and wild-type T1 B cells following BCR stimulation (Fig. 7A). In the SPPL2a-deficient T1 B cells, the kinetics of this ligand-induced BCR internalization appeared to be significantly accelerated. This enhanced and accelerated endocytosis may represent an underlying cause of the described diminished steady-state BCR surface levels of the SPPL2a^−/− B cells.

It was shown previously that especially activation of the PI3K/Akt pathway after BCR ligation, which is particularly disturbed in SPPL2a^−/− B cells, requires receptor internalization and delivery to endosomal compartments (38). Therefore, we analyzed the fate of the internalized BCR in wild-type and SPPL2a^−/− splenic B220⁺ B cells by microscopic imaging. After 30 min of internalization, BCR complexes were delivered into intracellular vesicles partially colocalizing with MHCII as revealed by indirect immunofluorescence for both wild-type and SPPL2a-deficient splenic B220⁺ B cells (Fig. 7B). We further analyzed the ultrastructure of the BCR-containing compartments by electron microscopy in splenic IgM⁺ B cells from wild-type and SPPL2a^−/− mice after endocytosis of IgM MicroBeads (Fig. 7C). The IgM beads were detected in endocytic compartments with a multivesiculated morphology that was found to be comparable between wild-type and SPPL2a-deficient B cells. Thus, we did not obtain any evidence that trafficking of the BCR after its internalization is altered in SPPL2a^−/− B cells so that it could preclude delivery to endosomes for PI3K/Akt activation according to Chaturvedi et al. (38).

A key question is whether the observed effects of the CD74 NTF on BCR trafficking and signal transduction involve a direct molecular interaction between both entities. As a prerequisite for that, we assessed colocalization between the endocytosed IgM MicroBeads and the CD74 NTF in SPPL2a^−/− B cells at the ultrastructural level (Fig. 7D). The latter was visualized by immunogold labeling with an Ab against the CD74 N terminus. The bulk of CD74 labeling was seen in abundant electron-lucent vacuoles as described previously (13), which, at least after 20 min, had not been accessed by the internalized BCR. Additionally, we found CD74 together with internalized anti–IgM MicroBeads in endocytic compartments. Therefore, based on this partial colocalization between the CD74 NTF and the stimulated BCR, a direct effect of the accumulating fragment on the receptor and/or the associated signal transduction machinery seems possible.

To further scrutinize such molecular interactions of the CD74 NTF, we performed a PLA on wild-type and SPPL2a^−/− B cells using Fab fragments recognizing intracellular epitopes of CD74, Syk, and the BCR complex subunit Igα (Fig. 8A, 8B). To define the background levels and to control the specificity of the detected interactions, SPPL2a-CD74 double-deficient B cells were included in the analyses. To confirm applicability of the system to CD74, a Fab-PLA for the established interaction between CD74 and MHCII was performed and found to generate signals in wild-type and SPPL2a^−/− B cells (Fig. 8C). In SPPL2a^−/− B cells, we observed specific signals in both the CD74/Igα (Fig. 8A) and the CD74/Syk (Fig. 8B) Fab-PLA that were significantly higher than in wild-type B cells. This strongly indicates that these signals were associated with the accumulation of the CD74 NTF. Interestingly, the signal intensity was independent of a preceding stimulation of the BCR with anti-IgM. Altogether, this suggests that in SPPL2a^−/− B cells the accumulating CD74 NTF and the BCR complex with its associated signaling machinery are not only in the same compartment, but they also exhibit close proximity as would be the case in a direct molecular interaction. Therefore, the described findings present an initial indication for a direct inhibitory effect of the CD74 fragment on BCR signal transduction as one element of the defects in SPPL2a^−/− B cells. However, to formally prove that, further experiments would be required because the detected proximity may also simply reflect the massive abundance of the CD74 NTF in the SPPL2a-deficient B cells. Interestingly, also in wild-type B cells a specific signal in the CD74/Igα Fab-PLA was detected that was, however, lost after BCR stimulation. In this case, this most likely indicates an interaction with the full-length CD74 protein. In addition to the increase of the IgM surface levels in CD74^−/− B cells that were described above (Fig. 5C), this may further advocate a previously unrecognized role of CD74 in the regulation of BCR trafficking already in wild-type B cells.

The splenic T1 B cell population is the last fully preserved stage prior to the maturation block in SPPL2a^−/− mice. We hypothesized that the significant alterations of BCR signaling and endocytic trafficking that we found in this population may represent an endpoint of a process initiated already in earlier developmental stages. Therefore, we analyzed immature B cells from bone marrow that precede the T1 stage to clarify the sequence of events behind this phenotype.

Because the amount of immature B cells (B220⁺IgM⁺) under steady-state conditions is low, we enriched immature B220⁺IgM⁺ B cells by culturing cells from the red bone marrow for 4 d in medium supplemented with IL-7 (30). Prior to analysis, cells were cultured for one additional night in medium without IL-7. After this time, cultures from wild-type and SPPL2a^−/− bone marrow contained 17–25% B220⁺IgM⁺ immature B cells (Fig. 9A). We analyzed CD74 expression in the immature B cells by flow cytometry after intracellular staining. The employed In-1 Ab is directed against the N terminus of CD74 and therefore detects presence of the CD74 full-length protein as well as of the CD74 NTF. We detected CD74 in immature B cells from wild-type mice and observed ∼3-fold enhanced signals in SPPL2a^−/− cells (Fig. 9B). Because SPPL2a deficiency has no impact on the abundance of CD74 full-length protein (13, 14), this signal increase most likely reflects the accumulation of the CD74 NTF. The signal increase from wild-type to SPPL2a-deficient immature B cells was in a similar range as that between T1 B cells of the respective genotypes. Thus, a significant CD74 NTF accumulation can already be detected in immature SPPL2a^−/− B cells. We analyzed BCR signal transduction in these cultured immature B cells. As shown in Fig. 9C, activation of the tyrosine kinase Syk following BCR stimulation was significantly compromised in immature SPPL2a^−/− B cells. However, the impairment of PI3K/Akt signaling was very minor and significantly less pronounced than in T1 B cells (Fig. 9C). We also analyzed IgM surface expression and observed that IgM surface levels of immature B220⁺ IgM⁺ B cells from SPPL2a^−/− mice were decreased in a statistically significant way (Fig. 9D). To validate this observation, we examined the effects of SPPL2a inhibition on Syk and Akt phosphorylation in wild-type T1 B cells after cultivating freshly isolated splenocytes overnight in the presence of the established SPP/SPPL inhibitors (Z-LL)₂-ketone and inhibitor X (16). As shown by flow cytometric analysis, application of the inhibitors clearly induced the accumulation of CD74 NTF in wild-type T1 B cells as compared with DMSO-treated control cells (Fig. 9E). Interestingly, the pool of surface IgM was significantly diminished (Fig. 9E) accompanied by a significant impairment of Syk activation upon BCR ligation (Fig. 9F). However, basal and BCR-induced phosphorylation of Akt was comparable in inhibitor-treated as well as control T1 B cells (Fig. 9F). These findings strongly suggest that diminished IgM surface levels and the reduced ability to activate Syk are interconnected processes, which are not immediately associated with defective PI3K/Akt signaling. This could mean that impairment of PI3K/Akt signaling simply requires a longer latency period to develop. However, this may also indicate that beyond preventing BCR surface exposure, SPPL2a deficiency and/or CD74 NTF accumulation exerts additional yet to be defined effects on the intracellular mechanisms leading to the activation of the PI3K/Akt in B cells.

Among several evaluated pathways only activation of PI3K/Akt signaling was shown to be capable of promoting survival of B cells that lack a functional BCR, demonstrating its critical role for transducing tonic BCR-associated survival signals (39). The kinase Akt phosphorylates several downstream targets, including the FOXO family of transcription factors (40, 41). In resting B lymphocytes, FOXO proteins reside in the nucleus and drive the transcription of cell cycle–regulating genes. Phosphorylation of FOXO by Akt leads to its stabilization, cytoplasmic sequestration, and subsequent inhibition of gene transcription. Among the FOXO family members, FOXO1 has been implicated in B cell differentiation with unique roles at the different maturation stages (42). Because SPPL2a^−/− B cells exhibited reduced Akt activation even under basal conditions, we analyzed whether this had any impact on FOXO1 phosphorylation. As revealed by Western blotting, phosphorylated FOXO1 levels extensively increased in wild-type splenic B220⁺ cells after BCR stimulation (Fig. 10A, 10B). In comparison, basal as well as induced phosphorylation of FOXO1 was significantly reduced in SPPL2a^−/− B220⁺ cells.

In addition to the described posttranslational control of FOXO1 activity by phosphorylation, BCR signaling has also been reported to regulate FOXO1 expression at the transcriptional level. It was found that BCR stimulation induces downregulation of FOXO1 mRNA levels, which is dependent on PI3K activity (43). To examine the impact of BCR signaling on FOXO1 expression in SPPL2a-deficient B220⁺ B cells, transcript levels were quantified by qRT-PCR under basal conditions and following a 3-h BCR stimulation (Fig. 10C). It was confirmed that the employed 3-h anti-IgG/IgM treatment did not negatively affect cell viability or induce apoptosis (not shown). FOXO1 mRNA levels in SPPL2a^−/− B cells were responsive to BCR signaling and decreased in a time-dependent manner after stimulation with similar kinetics as in wild-type B cells. However, SPPL2a^−/− B cells exhibited ∼2-fold increased FOXO1 expression. This was detected after BCR activation but most importantly also in unstimulated cells, presumably reflecting reduced tonic BCR signaling in these cells as suggested by the reduced Akt phosphorylation under basal conditions. The significantly enhanced FOXO1 mRNA levels of SPPL2a^−/− B cells were, however, not mirrored in an increase of FOXO1 protein (Fig. 10A), which might suggest that in these cells also the turnover of this protein is enhanced. Nevertheless, these findings indicate a considerable dysregulation of FOXO1 in SPPL2a-deficient B cells. Therefore, we sought to determine whether this had an impact on the activity of FOXO1 and/or other members of this transcription factor family that was reflected in the expression level of representative FOXO target genes. We determined transcript levels of CDKN1B, the gene encoding the cyclin-dependent kinase inhibitor 1B (p27, Kip1) (44–46) and BCL2L11, encoding the Bcl-2–interacting mediator of cell death (Bim) (46, 47). Both proteins can have proapoptotic functions and are associated with regulation of B cell proliferation and survival (48). In B220⁺ wild-type B cells, Bim mRNA levels were downregulated after 3 h of IgG/IgM stimulation as a response to BCR signaling (Fig. 10D). In contrast, SPPL2a^−/− B220⁺ B cells showed an extensively upregulated transcription of Bim in comparison with wild-type cells after initiation of BCR signaling, suggesting a failure to downregulate FOXO transcriptional activity in the SPPL2a-deficient B cells. Similarly, expression of p27 was found to be significantly enhanced in BCR-stimulated SPPL2a^−/− B220⁺ B cells versus wild-type B cells (Fig. 10E). In contrast to Bim, significant upregulation of p27 mRNA was already observed under steady-state conditions in unstimulated B cells (Fig. 10E). To rule out that the differences described for B220⁺ B cells were not due to the different proportions of B cell subsets in both genotypes, we also determined basal transcript levels of the proapoptotic genes p21, p27, and Bim as well as FOXO1 in FACS-sorted unstimulated T1 B cells from wild-type and SPPL2a^−/− mice (Fig. 10F). Except for p27, we observed an upregulation of these genes in the SPPL2a-deficient T1 B cells, confirming the conclusions from the total B220⁺ B cells.

In B cells, also the RAGs 1 and 2 (RAG1, RAG2) were reported to be targets of the transcription factor FOXO1 (42, 49). After successful Ig rearrangement and expression of a functional BCR, RAG expression is terminated at the immature B cell stage to prevent further rearrangements (32). Negative regulation of FOXO1 activity by the PI3K/Akt pathway is critically important for this RAG downregulation (50, 51). We compared RAG1/2 mRNA levels in wild-type and SPPL2a^−/− splenic B220⁺ B cells by qRT-PCR (Fig. 10G). Thereby, we observed 5- to 6-fold higher expression of RAG1 and RAG2 in the SPPL2a-deficient versus the wild-type B cells, indicating that downregulation of RAG expression in SPPL2a^−/− B cells is incomplete. Altogether, the described findings strongly confirm increased activity of FOXO transcription factors. This agrees well with the reduction of basal PI3K/Akt signaling that we detected in these cells (Fig. 3) and thereby corroborates that this impaired signaling leads to a relevant dysregulation of downstream targets of this pathway.

In the course of analyzing the molecular basis of the B cell maturation defect of SPPL2a-deficient mice, we could demonstrate that BCR signaling in SPPL2a^−/− B cells is impaired at various levels. After BCR stimulation, activation of Syk is reduced in these cells. Furthermore, ligand-induced but also basal activation of the PI3K/Akt pathway was found to be impaired in contrast to MAPK signaling, which appeared to be intact under the conditions tested. The reduced Akt signaling was reflected by an impaired inactivation of the transcription factor FOXO1 as revealed by increased transcription of FOXO target genes in SPPL2a^−/− B cells. Ablation of the BCR or its signaling capacity has been shown to prevent development of B cells as well as survival of mature B cells in the periphery (32, 52–54). Based on this, the essential role of tonic BCR signaling has been defined (32, 33). In light of this concept, it is well conceivable that the impairment of BCR signaling in SPPL2a^−/− B cells is a major cause of why these cells do not mature properly. This hypothesis is supported by the multiplicity of phenotypic similarities between SPPL2a^−/− mice and different knockout mouse models of those components of the BCR signal transduction machinery that are disturbed by SPPL2a deficiency.

Ablation of the kinase Syk already provokes an arrest of B cell development at the CD43^high pro–B cell stage in the bone marrow (55, 56), which is based on its requirement for pre-BCR signaling. However, absence of Syk also impaired further developmental progression of immature B cells (57). Apparently, pro– to pre–B cell transition is not negatively affected in SPPL2a^−/− mice because all populations up to the T1 stage are present in these mice in comparable amounts as compared with the wild-type situation (13). This most likely reflects that CD74 is not yet expressed in these early B cell stages in amounts that lead to an effective accumulation of the CD74 NTF.

Also, the genetic deletion of components of the PI3K/Akt pathway had a major impact on multiple stages of B cell development (58). Among the different regulatory and catalytic subunits of PI3Ks, especially the p85α and p110δ subunits are involved in BCR signal transmission (59–63). Common phenotypic features of p85α- and p110δ-deficient mice include an impairment of pro–B cell differentiation, a decrease of splenic B cell populations, and a deficiency of B1 B cells. In comparison, Akt1/2 double deficiency resulted in depletion of marginal zone and B1 B cells, but showed less pronounced effects on transitional and follicular B cells (64). However, Akt3 expression was still intact in these mice. A significant depletion of marginal zone and B1 B cells is also a characteristic feature of SPPL2a^−/− mice (13). In several of the above-mentioned models, the capability of the antiapoptotic protein Bcl-2 to rescue the respective B cell phenotypes upon transgenic overexpression has been evaluated (57, 64, 65). Although Bcl-2 expression was able to enhance survival of the knockout B cells, their failure to differentiate and to mature properly was not rescued accordingly, underlining the unique requirement of the individual pathways for B cell differentiation. Similarly, the maturation defect of SPPL2a^−/− B cells was not reverted by a Bcl-2 transgene (20). As mentioned above, the FOXO transcription factors are major effectors of the PI3K/Akt pathway (41). In murine primary B cells, FOXO overexpression has been shown to cause cell cycle arrest and increased apoptosis (66) by regulating target genes involved in these processes (67). This effect could be opposed by PI3K/Akt signaling, demonstrating that this pathway is involved in FOXO inactivation. In line with this, enhanced FOXO1 transcriptional activity in vivo significantly disturbs peripheral B cell homeostasis, as it was observed in mice deficient for the 14-3-3σ scaffolding protein that is involved in nuclear export of pFOXO1 (48). Spleens of these mice displayed a significant depletion of follicular and marginal zone B cells. However, the frequency of splenic B220⁺CD21⁻CD23⁻ B cells that had not yet undergone further maturation was not altered in the 14-3-3σ^−/− mice (48). This is very similar to SPPL2a^−/− mice where the B cell depletion spares the T1 stage, but affects all subsequent stages, including marginal zone B cells. The particular importance of PI3K signaling among the different pathways triggered by the BCR has been revealed by Srinivasan et al. (39). The authors reported that the survival of mature BCR⁻ B cells could be rescued either by expression of a constitutively active PI3K or knockout of the phosphatase PTEN, which is a negative regulator of the PI3K pathway. Also, knockout of FOXO1 significantly ameliorated survival of BCR-deleted B cells. In contrast, this was not achieved by enhancing NF-κB or MAPK signaling. In light of these different studies, the described disturbance of the PI3K/Akt/FOXO axis, which was evident already under basal conditions without stimulating the cells, very likely represents a major underlying cause of the impaired maturation and survival of SPPL2a^−/− B cells.

BCR-mediated activation of the PI3K/Akt pathway is generally considered to be mediated via Syk (58, 68, 69), explaining the concomitant impairment of Syk and Akt activation in SPPL2a^−/− B cells. Interestingly, SPPL2a^−/− B cells exhibited no detectable reduction of phosphorylated ERK levels following BCR stimulation despite the reduced Syk activation. In contrast to activation of the MAPK pathway, induction of PI3K/Akt signaling also depends on components of the BCR coreceptor complex such as CD19 (70). We observed a reduction of CD19 in SPPL2a-deficient B cells (not shown), which might explain the differential effects of pSyk reduction on ERK and Akt activation. Furthermore, Syk-independent mechanisms have been suggested to be involved in inducing PI3K/Akt signaling. In support of this, Yokozeki et al. reported that in a Syk-deficient A20 B cell line BCR-induced activation of downstream pathways including MAPK and PI3K/Akt signaling was at least partially preserved (71). Possibly, the degree of Syk dependency differs between the individual downstream pathways. In contrast to Akt, activation of Syk in SPPL2a^−/− B cells was only reduced after BCR stimulation but not under basal conditions, indicating that Syk- or even BCR-independent mechanisms may contribute to induce this basal Akt activation. Although BCR tonic signaling is considered to be the main inducer of basal PI3K/Akt signaling in B cells (39), also the BAFF/BAFF-R axis has been shown to be capable of stimulating Akt (72, 73). As we showed previously, SPPL2a^−/− B cells exhibit reduced BAFF-R surface levels. Thus, it is conceivable that this also contributes to the reduced basal activation of Akt in these cells. Importantly, the reduced basal Akt activation of SPPL2a^−/− B cells was not significantly rescued in the SPPL2a-CD74 double-deficient B cells and was also observed in CD74^−/− B cells. Apparently, also the loss of CD74 impairs pathways leading to PI3K/Akt activation. It was suggested that CD74 in its role as a receptor for the cytokine macrophage migration inhibitory factor can trigger Akt activation (74, 75), which could play a role here.

Interestingly, several of the cellular phenotypes we describe in the present study for SPPL2a^−/− B cells, including reduced IgM surface levels, intracellular IgM sequestration, enhanced IgM endocytosis, and suppression of BCR signal transduction, especially of the PI3K/Akt pathway, resemble characteristics that were described for anergic B cells (76–78). Anergy is a state of unresponsiveness that is induced in silenced peripheral autoreactive B cells when they have not successfully undergone receptor editing or been eliminated by apoptosis in the bone marrow (76). Thus, induction of anergy follows chronic antigenic stimulation of the BCR. In light of the intriguing phenotypic similarities between SPPL2a^−/− and anergic B cells, it may be speculated that similar pathways are involved in the cellular responses toward SPPL2a deficiency and chronic antigenic stimulation.

Studies in cultured cells have indicated that CD74 overexpression is capable of delaying endosomal maturation (2–4). In agreement with this, we found that SPPL2a^−/− B cells exhibited retarded degradation kinetics of endocytosed fluid-phase cargo that was associated with the CD74 NTF accumulation. However, in light of the severe ultrastructural changes in these cells (13), the general functionality of the endosomal–lysosomal system in these cells apart from the described trafficking delay appeared to be rather well preserved. Unexpectedly, we found that ligand-induced endocytosis of the BCR was enhanced in SPPL2a^−/− B cells. At the same time, steady-state BCR surface exposure was diminished in a CD74 NTF-dependent manner. It was postulated that an assembled (pre)BCR complex has to reach the cell surface to support signaling (79). Therefore, the observed BCR redistribution in SPPL2a^−/− B cells likely represents a major cause not only of the impaired signaling following BCR stimulation, but also of the reduced basal PI3K/Akt signaling. Interestingly, the relationship between BCR endocytosis and signaling has been controversial. Hou et al. (37) suggested that signaling is induced exclusively at the cell surface by noninternalized BCR. In contrast, Chaturvedi et al. (38) showed that both processes are functionally linked and that endocytosis of the BCR is required for activating the PI3K/Akt pathway, but not for inducing activation of Syk and MAPKs. We showed that SPPL2a^−/− B cells were capable to endocytose the BCR, ruling out that the impaired PI3K/Akt activation simply reflects defective BCR internalization. According to Chaturvedi et al. BCR-induced Akt activation occurs in endosomal compartments where the uncleaved CD74 NTF accumulates in SPPL2a^−/− B cells. Because activation of Akt is a membrane-bound process, it is well conceivable that the recruitment and activation of this kinase are disturbed by the CD74 NTF in these compartments. However, additional CD74 NTF-independent mechanisms on the PI3K/Akt pathway cannot be fully excluded. Because also loss of CD74 in T1 B cells from CD74^−/− mice was associated with an impairment of basal and BCR-induced PI3K/Akt signaling (Fig. 4F), a significant rescue of the impaired PI3K/Akt signaling in SPPL2a-CD74 double-deficient cells, which would formally prove a causative role of the CD74 NTF, could not be observed.

Apparently, activation of signaling pathways can also have a major impact on membrane trafficking in the endocytic system as has been reported for both Akt (80) and Syk (81). It is therefore conceivable that the CD74 NTF-induced interference with signaling pathways contributes mechanistically to the impairment of membrane trafficking. Multiple interconnected mechanisms may be part of the sequence of events that lead to the described alterations in SPPL2a^−/− B cells following accumulation of the CD74 NTF. In SPPL2a^−/− immature B cells as well as SPPL inhibitor–treated wild-type T1 B cells, we observed a reduction of IgM surface levels and at the same time a reduced capacity to activate Syk upon stimulation, strongly suggesting that both are mechanistically associated. However, the PI3K/Akt pathway was only disturbed to a very minor degree in both experimental setups. This could indicate that this cellular phenotype is at least in part independent from the BCR mistrafficking and impairment of Syk activation. Possibly, additional effects of the CD74 NTF and/or SPPL2a deficiency are required to affect also the PI3K/Akt pathway, which may become effective only after a certain latency following further accumulation of the CD74 NTF. The precise molecular details how the CD74 NTF perturbs membrane trafficking and influences kinase activation remain unclear to date. An interaction of CD74 with the uncoating ATPase Hsc70 has been reported (82). The impact of overexpressed CD74 on endosomal size and maturation was shown to depend on the net negative charge of the cytosolic N-terminal segment of the protein (83) and could be abrogated by single amino acid exchanges that abolish this net charge. Therefore, it may be speculated that the CD74 NTF, if present in significant amounts as in SPPL2a^−/− B cells, is capable of interfering with electrostatic interactions that are involved in recruiting certain effector proteins to endosomal membranes.

Conceptually, it may be asked why a molecule such as CD74, which plays a major role in MHCII Ag presentation, can have such a distinct impact on central signaling pathways involved in supporting B cell maturation. For a mature B cell the capacity to process Ags and to present derived peptides on MHCII complexes are essential to be able to receive coactivating signals from primed CD4⁺ T helper cells. A specific role of CD74 in influencing B cell maturation has been suggested previously (84). It may be speculated that the capability of CD74 to influence BCR signaling and to impact on B cell maturation provide a safeguard mechanism ensuring that only B cells capable of degrading CD74 and therefore presumably also of processing and presenting Ags reach full maturity and functionality. In this regard, the observed interaction between CD74 and the BCR may deserve further attention with regard to a potential role of CD74 influencing BCR trafficking. In general, our findings strongly indicate that the molecular functions of CD74 in B cells may not have been fully explored yet.

With respect to the phenotype of SPPL2a^−/− mice, we have proposed that pharmacological inhibition of this enzyme may represent a novel small molecule–based strategy to achieve B cell depletion in autoimmune disorders. We could confirm recently that the essential role of SPPL2a for turnover of the CD74 NTF is preserved in human B cells (15). However, so far the functional consequences of SPPL2a deficiency on human B cell development and functionality are not known. Furthermore, potent and specific SPPL2a inhibitors are not yet available (16). Once such compounds have been obtained, it will be interesting to assess their potential to complement strategies that aim to interfere with signaling downstream of the BCR, especially Syk and PI3K, that are currently being evaluated in the treatment of B cell neoplasms (35).

We thank Sebastian Held, Marlies Rusch, and Emanuela Szpotowicz for excellent technical assistance and Dr. Roberta Pelanda, University of Colorado, for advice and very helpful discussions on bone marrow cultures. We also thank Sandra Ussat, Institute of Immunology, Christian Albrechts University of Kiel, for excellent conduction of T1 B cell sorting. Furthermore, we are grateful to Prof. Willem Stoorvogel, University of Utrecht, for providing anti-MHCII β-chain Ab.

The authors have no conflict of interest.