The Influence of MHC Class II on B Cell Defects Induced by Invariant Chain/CD74 N-Terminal Fragments

The invariant chain, also referred to as li or CD74, is a central protein in APCs that is primarily known for its role as a chaperone and trafficking adaptor of MHC class II (MHCII) complexes (1). However, beyond that, CD74 has a regulatory impact on a variety of intracellular processes, including signal transduction, endosomal trafficking, and cell migration (1).

Upon delivery of MHCII to Ag-processing endosomal compartments, CD74 is subjected to a sequential proteolytic turnover. This commences with a step-wise trimming of the luminal domain that is mediated by different endosomal proteases (1). Of particular importance is the cleavage N-terminal of the class II–associated li chain peptide (CLIP) segment that occupies the peptide-binding cleft within the MHCII dimer. This cleavage is a prerequisite for peptide loading of MHCII, as well as efficient surface delivery. In B cells and dendritic cells, cathepsin S (CatS) is of critical importance for this step, as documented by an impairment of this cleavage in the absence of this protease (2, 3). In general, other endosomal cysteine proteases like cathepsin L and cathepsin F are similarly capable to perform this cleavage and are critically involved in this in thymic epithelial cells and macrophages, respectively (4, 5). At last, degradation of CD74 is completed by the intramembrane protease signal peptide peptidase-like 2a (SPPL2a), which processes the remaining membrane-bound N-terminal fragment (NTF) (6–8). Thereby, the CD74 intracellular domain is released into the cytosol; it can enter the nucleus and may exert downstream effects on transcriptional regulation (9).

We (6) and other investigators (7, 8) recently showed that this final step of CD74 turnover by SPPL2a is of critical importance for humoral immunity. B cells from SPPL2a-deficient mice accumulate massive amounts of an uncleaved CD74 NTF that causes a major B cell–maturation defect in these mice. This phenotype is characterized by a significant depletion of all peripheral B cell populations beyond the transitional stage 1 (T1), which is the last largely preserved B cell subset (6–8). Based on a significant amelioration of this phenotype upon additional genetic ablation of CD74, the build-up of the CD74 fragment could be unambiguously identified as a major causative factor (6).

Although the precise molecular mechanisms linking the CD74 NTF with this phenotype is not fully resolved, several changes in membrane trafficking and signal-transduction pathways in SPPL2a^−/− B cells have been described that are likely to be involved (10). SPPL2a-deficient B cells exhibit prominent accumulation of endosome-derived vacuoles (6), which is associated with a delayed degradation of fluid-phase endocytic cargo (10). Possibly related to this disturbance in membrane trafficking, the subcellular distribution of the BCR is shifted toward intracellular compartments in these cells. Thus, BCR surface levels were found to be reduced in SPPL2a^−/− B cells compared with wild-type (wt) cells (10). Beyond B cell activation upon Ag recognition, signal transduction downstream of the BCR is critical for the differentiation of developing B cells and the survival of mature B cells (11–13). A basal, so-called “tonic” BCR signaling is required to promote maturation of B cells according to current models. Although several signaling axes are triggered by the BCR, the PI3K/Akt pathway is particularly critical for the survival-promoting effect (14). FOXO transcription factors (e.g., FOXO1) are major downstream targets and are regulated by the kinase Akt via this pathway (15, 16). We found that, in SPPL2a-deficient B cells, BCR signaling is disturbed by the accumulating CD74 NTF (10). In particular, activation of the critical PI3K/Akt/FOXO axis is impaired upon BCR cross-linking. Importantly, this is also reflected in a reduced activity of this pathway under basal conditions (10). In light of the well-documented requirement of this axis to promote B cell maturation and survival, it seems likely that this defect in BCR signal transduction represents a major cause of the B cell phenotype in SPPL2a^−/− mice (10).

Although the CD74 NTF accumulating in SPPL2a-deficient B cells has lost the major interaction site with MHCII, the CLIP segment, MHCII homeostasis in these cells was found to be altered. T1 SPPL2a^−/− B cells from SPPL2a^−/− mice exhibit increased total MHCII levels, as revealed by flow cytometric analyses of permeabilized cells (6). Similarly, levels at the cell surface were also enhanced (6). The consequences of this altered MHCII homeostasis with regard to the B cell–maturation defect, as well as the effects of the CD74 NTF on trafficking and signaling pathways, are unclear. Beyond its role in peptide presentation, MHCII was found to exhibit nonclassical functions (17, 18). Binding of MHCII to physiological ligands or to certain Abs triggers different signaling cascades in MHCII-expressing cells (17, 18). Depending on cell type and physiological context, the observed effects were diverse, ranging from an activation of the cell to the initiation of cell death pathways (17, 18). Endosomal MHCII is important for activation of the kinase Btk and TLR responses in dendritic cells (19). In several mouse models, an altered intracellular homeostasis of MHCII is responsible for a concurrent loss of B cells. These include the transgenic overexpression of MHCII (20) and the knockout of CD74 (21) or MHC α-chains (22). Particularly, unpaired MHCII β-chains are toxic to B cells (22). Based on this, a causative involvement of MHCII in the development of the B cell phenotype of SPPL2a^−/− mice seems conceivable.

In this study, we aimed to systematically address how the presence and/or association with MHCII influences the impact of an accumulating CD74 NTF on B cells. Therefore, we analyzed B cells from CatS-deficient mice, which accumulate a longer CD74 NTF that is still associated with MHCII via the CLIP segment. Furthermore, we bred our SPPL2a^−/− mice onto an MHCII-deficient background and examined B cell maturation, as well as BCR signaling responses, in these SPPL2a-MHCII double-deficient mice.

CatS-deficient (23) and SPPL2a-deficient (6) mice were described previously. SPPL2a^−/− mice were backcrossed for 10 generations on a C57BL/6N Crl background and bred with MHCII-deficient mice (B6.129S2-H2^dlAb1-Ea/J; obtained from The Jackson Laboratory) to generate SPPL2a-MHCII double-deficient animals. All experiments were performed with littermates and/or controls matched for sex and age. Care and handling of the animals were performed in accordance 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 B cells were isolated by positive selection using B220 or IgM MicroBeads, as indicated, and LS columns of a 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 was found to be between 80 and 90%.

For treatment with inhibitors, freshly isolated wt splenocytes were cultivated overnight at 37°C in RPMI 1640 with l-glutamine (Life Technologies) supplemented with 10% FCS (Biochrom), 50 mM 2-ME (Life Technologies), 100 U/ml penicillin, and 100 μg/ml streptomycin (both from Sigma-Aldrich) (RPMI complete) in the presence of the indicated concentrations of (Z-LL)₂-ketone (PeptaNova) and inhibitor X (EMD Millipore) or E64-d (Enzo Life Sciences). Subsequently, IgM⁺ B cells were isolated by magnetic cell sorting and subjected to Western blot analysis for CD74.

For the electron microscopic analysis of mature splenic B cells, B220⁺ cells were isolated by MACS from wt and CatS^−/− splenocytes. Following staining with anti-CD21, anti-CD24, and anti-CD45R (B220), these cells were subjected to FACS using a FACSAria (BD Biosciences) to isolate B220⁺CD21^lowCD24^low mature B cells corresponding to the gating scheme depicted in Fig. 1E. Fixation and processing for ultrastructural analysis were performed as described below.

Analysis of BCR signal transduction was performed as reported before (10). Activation of the BCR was achieved by incubating cells in the presence of 10 μg/ml goat anti-mouse IgG/IgM F(ab′)₂ (Dianova). For flow cytometric analysis of Syk and Akt activation, freshly isolated splenocytes were stimulated in 500 μl of PBS for 5 min at 37°C. Alternatively, isolated splenic B220⁺ B cells were stimulated in 500 μl of RPMI complete medium for 5 min at 37°C and subjected to Western blot analysis.

Single-cell suspensions from spleens, thymus, and lymph nodes were obtained by passing the tissues through a 100-μm cell strainer suspended in FACS buffer (PBS including 2% [v/v] FBS, 0.1% [w/v] NaN₃, and 2 mM EDTA [pH 7.4]). Bone marrow cells were isolated from the murine femur by flushing the bones with FACS buffer and dispersing the cells via a 100-μm cell strainer. For isolation of peritoneal cells, FACS buffer was injected into the peritoneal cavity, and the cells were suspended by gentle massage. Subsequently, the fluid was recollected. To induce lysis of RBCs, samples were incubated in 155 mM NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA for 12 min at room temperature.

For the analysis of surface markers, the cells were incubated for 30 min at 4°C with the following FITC-, PE-, PE/Cy7-, or allophycocyanin-labeled Abs diluted in FACS buffer: anti-CD21/35 (8D9), anti-CD24 (30-F1), anti-CD43 (eBioR2/60), anti-CD45R (RA3-6B2), anti-IgM (II/41) (all from eBioscience), anti-CD23 (B3B4), anti-MHCII (M5/114.15.2) (both from BioLegend), and anti-CD19 (PeCa1; ImmunoTools). For detection of total cellular IgM, MHCII, or CD74, cells were fixed and permeabilized using the BD Cytofix/Cytoperm Fixation/Permeabilization Solution Kit (BD Biosciences) prior to staining with anti-IgM, anti-MHCII, or FITC-conjugated anti-CD74 (In-1; BD Biosciences). To allow gating on individual B cell subsets, colabeling with anti-CD21/35, anti-CD45R, and anti-CD24 (M1/69; BioLegend) was performed.

To analyze the phosphorylation status of the kinases Syk and Akt as described (10), stimulated cells were rapidly cooled down prior to fixation with paraformaldehyde at a final concentration of 3.75% (w/v) for 20 min. Subsequently, cells were permeabilized in 90% methanol for 10 min on ice and incubated in PBS supplemented with 0.5% BSA for 10 min prior to Ab labeling. Rabbit mAbs against p-Akt (Ser⁴⁷³, D9E), Akt (C67E7), pSyk (Tyr525/526, C87C1), or Syk (D3Z1E), (both from Cell Signaling Technology), together with anti-CD21/CD35, anti-CD24 (M1/69), and anti-CD45R, were applied for 1 h on ice in the same buffer. For detection of the anti–p-Akt and anti-Akt primary Abs, cells were incubated in the presence of Alexa Fluor 488 goat anti-rabbit IgG secondary Ab (Life Technologies) for 30 min on ice after washing. All samples were analyzed using a FACSCanto or FACSCanto II flow cytometer (BD Biosciences), followed by data evaluation with FlowJo (TreeStar) software.

Flow cytometric assays of the endocytic system were performed as described previously (10). To quantify the degradation of cargo taken up by fluid-phase endocytosis, splenocytes were pulsed with OVA-FITC (Life Technologies) for 30 min at 37°C in RPMI complete. After washing, cells were chased for different periods between 0 min and 6 h at 37°C before being chilled immediately by the addition of ice-cold PBS. Cells were stained with anti-CD21/35–PE (8D9), anti-CD45R–PE–Cy7 (RA3-6B2), and anti-CD24–allophycocyanin (M1/69), followed by erythrocyte lysis and flow cytometric analysis. The median fluorescence intensity (MFI) of the FITC channel was determined for the respective B cell subsets and corrected for unspecific binding of OVA-FITC based on samples that had been kept on ice under the same conditions. The obtained ΔMFIs were normalized to those of the 0-min chase time point directly after the pulse.

To follow internalization of the BCR, splenocytes were labeled with biotin-conjugated anti-mouse IgM (II/41; eBioscience) for 30 min on ice in MACS buffer. Following two washing steps, cells were incubated between 0 and 30 min at 37°C to allow endocytosis of the BCR complex and then immediately cooled down. To label the residual noninternalized BCR, cells were stained with streptavidin–allophycocyanin in conjunction with anti-CD21/35–PE (8D9), anti-CD24–FITC (30-F1), and anti-CD45R–PE–Cy7 (RA3-6B2). Erythrocytes were lysed prior to flow cytometric analysis. The MFI of the allophycocyanin channel was measured for the respective cell population and normalized to that of the 0-min chase time point of each sample.

IgM⁺ B cells were isolated from splenocytes by magnetic cell sorting, as described above. Cells were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) in suspension for 1 h at 22°C and overnight at 4°C. Subsequently, the cells were sedimented (220 × g, 5 min), incorporated in BSA, which was cross-linked by the addition of glutaraldehyde and paraformaldehyde as described (6, 24), and stored in 2.5% glutaraldehyde in phosphate buffer. Postfixation of BSA blocks was performed in 2% OsO₄, and embedding in araldite was conducted according to routine procedures. Ultrathin sections were stained with uranyl acetate and lead citrate. Per specimen derived from one mouse, 50 B cells were photographed at a magnification of 7000×, and vacuoles with a diameter ≥ 250 nm were counted on the computer screen. From this, the mean number of vacuoles per cell was determined for each specimen.

To further characterize the vacuoles of CatS^−/− B cells, we performed immunoelectron microscopy, as reported previously (10). In brief, splenocytes from CatS^−/− mice were incubated in the presence of IgM MicroBeads (Miltenyi Biotec) and isolated using LS columns strictly at 4°C. After warming up, cells were allowed to internalize the MicroBeads bound to the BCR for 30 min at 37°C. Cells were fixed immediately in 4% PFA and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) in suspension for 90 min at room temperature and then stored in 2% PFA in the same buffer at 4°C. Sample preparation, sectioning (ultrathin sections, 70 nm), and postembedding immunogold labeling were performed as described (10, 25). CD74 was detected with a rabbit polyclonal Ab directed against an N-terminal epitope (10). For labeling of MHCII, a rabbit antiserum against the MHCII β-chain (kind gift of Prof. W. Stoorvogel) was used. Both primary Abs were visualized with 15-nm large gold probes coupled to protein A (G. Posthuma, University Medical Center Utrecht). Specimens were examined using an EM902 transmission electron microscope (Zeiss) equipped with a MegaView III digital camera (A. Tröndle).

Cells were washed twice with PBS, sedimented, and resuspended in lysis buffer (50 mM Tris/HCl [pH 7.4], 150 mM NaCl, 1.0% [w/v] Triton X-100, 0.1% [w/v] SDS, and 4 mM EDTA) for detection of CD74 or 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) for the analysis of phosphorylated and total Akt and FOXO1. Both buffers were supplemented with protease inhibitors and, where applicable, PhosSTOP Phosphatase Inhibitor Cocktail (Roche). Samples were incubated on ice for 1 h and within this period were subjected to sonication (level 4 for 20 s) in a Branson Sonifier 450 (Emerson Industrial Automation) at 4°C, as described (26). Cell debris was removed by centrifugation (15,000 × g for 10 min), and protein concentration in the lysates was determined by bicinchoninic acid protein assay (Thermo Fisher Scientific). Protein separation was performed by SDS-PAGE using a Tris-tricine buffer system (27), when detection of CD74 was intended, or a standard Tris-glycine system, according to Laemmli (28), for all other analyzed Ags, which was followed by a semidry transfer to nitrocellulose (26).

For immunodetection, rabbit Abs against p-Akt (Ser⁴⁷³, D9E), Akt (C67E7), p-FOXO1 (Ser²⁵⁶), and FOXO1 (C29H4) were obtained from Cell Signaling Technology. The CD74 full-length protein, as well as the different NTFs, were visualized with rat monoclonal In-1 Ab (BD Biosciences), which recognizes an N-terminal epitope of the protein. MHCII dimers were detected with the monoclonal M5/114.15.2 Ab (BioLegend). In this case, samples were prepared in standard SDS-PAGE sample buffer. However, the heat-denaturation step was omitted, and samples were kept on ice prior to electrophoresis. To demonstrate equal protein loading, anti-actin (Sigma-Aldrich) and anti-GAPDH (Santa Cruz) were used. After application of Lumigen ECL Ultra (TMA-6; Lumigen), chemoluminescence was recorded using an ImageQuant LAS 4000 (GE Healthcare).

Data are represented as mean ± SD. Statistical significance was assessed using an unpaired two-tailed Student t test or one-way ANOVA, followed by the Newman–Keuls post hoc test. Significance levels of *p < 0.05, **p < 0.01, and ***p < 0.001 were applied.

CD74 undergoes sequential proteolytic processing in endosomal compartments via several C-terminally truncated degradation intermediates (Fig. 1A). In wt B cells, these are not detected under steady-state conditions using an Ab directed against the cytoplasmic N-terminal domain of the protein (Fig. 1B). However, when any of these cleavage events are blocked or impaired, as in CatS- or SPPL2a-deficient B cells, CD74 degradation intermediates are stabilized, which will be referred to as NTF1 (accumulation in CatS^−/− cells) and NTF2 (accumulation in SPPL2a^−/− cells) (Fig. 1A, 1B). Thus, similar to SPPL2a-deficient B cells, CatS^−/− B cells also accumulate a significant amount of CD74 fragments.

In the original description of CatS-deficient mice (23), normal ratios of B220⁺ cells in spleens and lymph nodes were reported. However, no detailed information about different B cell subpopulations was available in the literature that would be required to rule out subtle impairments in B cell maturation in these mice caused by the accumulating CD74 NTFs. Therefore, we performed a flow cytometric analysis of lymphatic tissues from CatS^−/− mice with a particular focus on the intermediate populations of B cell differentiation in bone marrow and spleen (Fig. 1C–F, Table I). Bone marrow pro-B, pre-B, immature B, and recirculating B cell populations were distinguished based on B220, CD43, and IgM expression (Fig. 1C, 1D, Table I). In the spleen, transitional B cell populations were identified based on B220, CD21, and CD24 surface levels, as indicated (Fig. 1E, 1F, Table I). Altogether, no major reductions in any of the assessed populations were seen, which excludes a relevant defect in B cell differentiation in the bone marrow or spleen. Interestingly, the proportion (percentage) of transitional stage 2 (T2) cells among splenocytes of CatS^−/− mice was statistically significantly reduced compared with wt mice (Table I). However, when comparing mean values for absolute numbers of T2 cells from wt and CatS^−/− spleens, this reduction did not meet the criteria for statistical significance because of the higher variance of these values (Fig. 1F, Table I). This minor effect was not reflected in a reduction in the mature B2 B cell pool in the spleen (Fig. 1F) and other lymphatic tissues (Table I). Similarly, the abundance of marginal zone B cells in the spleen and B1 B cells in the peritoneal cavity was not negatively affected by CatS deficiency (Table I). Thus, although CatS^−/− B cells accumulate significant amounts of a CD74 NTF, their maturation is not disturbed to a physiologically relevant degree.

Having identified alterations in membrane trafficking and signal transduction associated with the accumulating CD74 NTF in SPPL2a^−/− B cells, we specifically examined CatS-deficient B cells for these cellular phenotypes. Therefore, we compared the ultrastructure of wt and CatS^−/− B cells (Fig. 2A). Similar to SPPL2a-deficient B cells, CatS^−/− B cells contained numerous electron-lucent vacuoles (Fig. 2B). Vacuoles with a similar morphological appearance were also observed to some extent in wt cells. However, their abundance per cell was significantly increased in CatS^−/− B cells (Fig. 2C), indicating a relevant impairment in endocytic membrane trafficking. When directly comparing CatS^−/− and SPPL2a^−/− B cells, it became apparent that the vacuolization, assessed as vacuoles per cell, was much more heterogenous in the analyzed CatS^−/− cell population (Fig. 2D). In SPPL2a^−/− B cells, the distribution scattered around a maximum of four to five, which agrees well with the mean number of vacuoles per cell that we observed in previous analyses (6, 29). In contrast, 59% of CatS^−/− B cells exhibit two vacuoles or less, similar to wt B cells, and only 41% of the analyzed cells seem to exhibit more vacuoles than the majority of wt B cells. In SPPL2a^−/− B cells, the latter was observed for 76% of the cells. To further characterize the vacuoles, we performed immunogold labeling of splenic CatS^−/− B cells, which had been allowed to internalize IgM MicroBeads for 30 min (Fig. 2E). We found vacuoles that were accessible to endocytosed BCR, as documented by the presence of internalized IgM MicroBeads in their lumen (Fig. 2E, black arrows). Furthermore, those compartments were positive for CD74, as detected with an Ab against the N terminus, and MHCII (Fig. 2E, white arrows), indicating that they contain the CD74 NTF, presumably in complex with MHCII.

We investigated the putative impact of the significant morphological changes on the functionality of the endocytic system in CatS-deficient B cells. Therefore, we followed the degradation kinetics of the fluid-phase endocytic cargo OVA-FITC (OVA) in a pulse-chase experiment (Fig. 2F). As reported previously (10), the decline in OVA-associated fluorescence was delayed in SPPL2a-deficient T1 B cells compared with wt cells, which was reflected in a significantly prolonged half-life (Fig. 2G). In contrast, CatS^−/− T1 B cells degraded internalized OVA with a very similar kinetics as did wt T1 cells. Because T2 and mature B cells are present in CatS^−/− mice, we also analyzed OVA turnover in these B cell subsets in the same experiment. Although T2 cells were also unaffected by CatS deficiency (data not shown), a slight delay in OVA degradation was detected in the mature B cells (Fig. 2H). Thus, despite the prominent vacuolization of the endocytic system, functionality of the endosomal pathway with regard to fluid-phase cargo degradation was not modulated (T1, T2 stage) or was modulated only to a minor degree (mature) by CatS deficiency in B cells.

In light of the differential impairment of cargo degradation in mature versus T1 CatS-deficient B cells, we hypothesized that the described heterogeneity of vacuolization (Fig. 2D) may reflect differences between the individual B cell subsets that could not be distinguished in our analysis of total IgM⁺ B cells. Therefore, we isolated mature B cells (B220⁺CD21^lowCD24^low) from spleens of CatS-deficient mice by a combination of magnetic and flow cytometric sorting and performed electron microscopy (Fig. 2I). However, the heterogeneity that we observed in this population was very similar to the distribution of vacuoles in the total pool of IgM⁺ CatS^−/− B cells. Based on this, we consider it unlikely that the heterogenous vacuolization can be explained by a differential involvement of certain subpopulations.

Because we had previously found that SPPL2a^−/− T1 B cells exhibit an accelerated endocytosis of the BCR (10), we assessed trafficking of the BCR complex in CatS-deficient B cells (Fig. 2J). In the setup used, the Ab–BCR complex was internalized at a similar rate as in wt cells, as depicted for the T1 subset. Similar results were obtained for T2 and mature B cells (data not shown). Based on this, we conclude that there is no enhanced endocytic recovery of the BCR from the cell surface like in SPPL2a^−/− T1 B cells (Fig. 2J) (10).

Nevertheless, we analyzed surface and total IgM levels in different splenic B cell subsets from CatS^−/− mice by flow cytometry (Fig. 3A, 3B). Total IgM levels were similar in wt and CatS^−/− T1, T2, and mature B cells. This was also the case for the abundance of IgM at the plasma membrane of T1 and T2 cells. However, CatS^−/− mature B cells exhibited a statistically significant reduction in surface IgM to ∼70% of the level in wt cells. Because the overall IgM expression in permeabilized cells was not reduced to a similar degree, this argues for a shift in IgM from the plasma membrane toward intracellular compartments.

We continued to assess signal transduction downstream of the BCR, focusing particularly on the PI3K/Akt/Foxo axis, which is disturbed in SPPL2a-deficient B cells. Total B220⁺ splenic B cells were stimulated by cross-linking the BCR with anti-IgM/IgG F(ab′)₂ fragments and analyzed for p-Akt by Western blotting (Fig. 3C). Activation of this kinase after BCR ligation was comparable between CatS^−/− and wt B cells. Furthermore, the basal activation of Akt in unstimulated cells, which reflects tonic BCR signaling, was not negatively affected by CatS deficiency. We also analyzed whether this would be reflected in a comparable phosphorylation of the transcription factor FOXO1 as a downstream target (Fig. 3D). In line with the similar activation of Akt, basal and BCR-induced phosphorylation of FOXO1 were indistinguishable between wt and CatS-deficient B cells. Thus, no impairment of BCR signaling with regard to activation of the PI3K/Akt pathway was evident in total splenic B cells from CatS^−/− mice. In SPPL2a-deficient mice, the impairment of BCR signaling was characterized in T1 B cells, because these cells represented the most advanced population that was largely preserved. We wanted to exclude any differential impact of the CD74 NTF on BCR signaling in individual B cell subsets that may have masked possible effects in the previous experiments in which total B220⁺ splenic B cells had been analyzed.

Therefore, we specifically assessed Syk (Fig. 3E, 3F) and Akt (Fig. 3G, 3H) activation in the T1 B cell population of CatS^−/− mice by flow cytometry (Fig. 3G, 3H). Again, no statistically significant differences between wt and CatS-deficient cells were detected; however, tendencies toward a slightly reduced basal and BCR-triggered Akt phosphorylation associated with CatS deficiency were observed in this cell population. Similar results for Syk and Akt activation were obtained when T2 and mature CatS-deficient B cells were analyzed (data not shown). Altogether, despite morphologically detectable changes in the endocytic system of CatS-deficient B cells that were reminiscent of those in SPPL2a^−/− B cells, no major impact of CD74 NTF accumulation on BCR trafficking and signal transduction could be identified in these cells.

The described findings reveal a major phenotypic discrepancy between SPPL2a- and CatS-deficient B cells, although both accumulate significant amounts of CD74 NTFs. To clarify this, we assessed CD74 levels in permeabilized splenic B cells from both mouse strains by flow cytometry using a mAb against an N-terminal epitope, thus detecting the full-length protein and the NTFs (Fig. 4A). We directly analyzed CD74 levels in T1 B cells, because more mature B cell stages are depleted to a large extent in SPPL2a^−/− mice, impeding a comparison of these populations.

The flow cytometric signals for CD74 that were obtained from CatS^−/− T1 B cells were significantly higher than those from wt cells. Because steady-state levels of the CD74 full-length protein were not changed in these cells, this reflects the NTF accumulation as observed by Western blotting (Fig. 1B). In SPPL2a-deficient T1 B cells, the determined normalized CD74 MFI was ∼2.8-fold higher than in CatS^−/− cells (Fig. 4B). Thus, SPPL2a^−/− B cells accumulate significantly more CD74 NTFs than do CatS-knockout cells. In addition, we performed Western blot analyses (Fig. 4C), which confirmed the results from the flow cytometric quantitation. However, in addition to their abundance, CD74 NTFs from CatS^−/− and SPPL2a^−/− B cells differ with regard to size. As expected based on the current view of CD74 processing (Fig. 1A), the fragments accumulating in CatS^−/− B cells, which still contain the CLIP segment, are longer than the ones from SPPL2a-deficient cells. Because T2 and mature B cells are preserved in CatS^−/− mice, we analyzed putative differences between these subsets with regard to CD74 NTF accumulation (Fig. 4D). As revealed by flow cytometry, CatS-deficient T2 and mature B cells exhibited significantly higher CD74 NTF levels than the corresponding T1 cells.

We aimed to provide insight into why the levels of the accumulating NTF1 and NTF2 in CatS^−/− and SPPL2a^−/− B cells, respectively, are so different even though the deficiency in both proteases leads to a detectable block in CD74 processing. Therefore, we cultivated wt and CatS-deficient splenic B cells in the presence or absence of the SPP/SPPL inhibitors (Z-LL)₂-ketone and inhibitor X, followed by isolation of B cells and Western blot analysis of CD74 processing (Fig. 4E). As reported previously, this treatment stabilized NTF2 in wt B cells, mimicking the situation in SPPL2a^−/− B cells. This effect was observed to a similar degree in CatS-deficient B cells. Thus, the CD74 NTF2 is also generated in the absence of CatS. Under steady-state conditions, this is continuously processed by SPPL2a. Based on the similar degree of NTF2 accumulation in wt and CatS^−/− B cells, the overall rate of CD74 degradation seems to be quite comparable. This suggests that the role of CatS in this process in B cells is not unique and can be compensated for, to a large extent, by other proteases. Therefore, under steady-state conditions, only limited amounts of NTF1 accumulate in CatS-deficient B cells, because a significant portion of this fragment can still be cleaved to NTF2 and then subjected to intramembrane proteolysis by SPPL2a.

In addition to the different abundance of the accumulating NTF1 and NTF2, we considered alternative explanations for the phenotypic differences in CatS- and SPPL2a-deficient B cells. As mentioned above, the major MHCII binding site is still preserved in the CatS^−/− NTF1. When assessing MHCII in nonheat-denatured samples from CatS-deficient B cells, we observed a band at ∼ 75 kDa that was not present in wt cells and presumably represents an MHCII–NTF1 complex (Fig. 4F), because it was also revealed with the N-terminal CD74 Ab. Persistence of the complex in the presence of SDS indicates a tight association. Such a complex was not seen in SPPL2a^−/− B cells (data not shown), indicating that any putative interaction of NTF2 with MHCII via the transmembrane segments is much weaker. In CatS^−/− B cells, the CD74 NTF–MHCII association leads to an intracellular retention of MHCII (Fig. 4G, 4H). In agreement with previous reports (3), MHCII surface levels were modulated only mildly or not at all. However, total cellular MHCII levels, which were determined by flow cytometry in permeabilized cells, were significantly increased in all analyzed B cell subsets.

We reported previously that SPPL2a^−/− B cells also exhibit an altered MHCII homeostasis that is characterized by an increase in cellular MHCII levels (6). In light of several examples in the literature, in which alterations in MHCII complex formation and trafficking have been linked with B cell defects (20–22), this may be considered part of the sequence leading to the maturation defect of these cells. Possibly, the tight MHCII association of CatS^−/− CD74 NTF1 might prevent downstream effects exerted by an MHCII accumulation that is triggered by the “free” MHCII in SPPL2a^−/− B cells.

Following this hypothesis, we aimed to clarify whether MHCII has any causative role in the pathogenesis of the B cell defect in SPPL2a-deficient mice. Therefore, we generated SPPL2a-MHCII double-deficient mice by crossing our SPPL2a^−/− mice (6) with a mouse line in which all conventional MHCII genes have been deleted (30). We determined mature B cells and their respective precursor stages in lymphatic tissues from these mice. The additional MHCII deficiency did not rescue the reduction in total B220⁺ B cells seen in bone marrow, spleen, and lymph nodes of SPPL2a^−/− mice (Fig. 5A, Table II). We also resolved B cell precursor populations in the bone marrow (Fig. 5B) and spleen (Fig. 5C) of these mice. In general, SPPL2a^−/− and SPPL2a-MHCII double-deficient mice exhibited highly comparable B cell phenotypes. Thus, in the absence of MHCII, no recovery, but also no major worsening, of the splenic B cell–maturation block induced by the loss of SPPL2a was seen.

Having documented that the B cell phenotype of SPPL2a^−/− is independent of MHCII, we wanted to confirm that the so-far identified underlying mechanisms are also not affected by the MHCII deficiency. First, we tested whether the loss of MHCII influences the stability of the CD74 NTF. Therefore, we performed intracellular flow cytometric analysis of CD74, as above, and compared CD74 levels in SPPL2a^−/− and SPPL2a-MHCII double-deficient T1 B cells (Fig. 6A). No significant difference was observed, indicating that the presence of MHCII does not influence the accumulation of this fragment. Similarly, SPPL2a^−/− MHCII^−/− B cells exhibited vacuolation of endosomal compartments (data not shown) like SPPL2a single-deficient cells. Furthermore, the intracellular retention of the BCR leading to reduced surface levels (Fig. 6B), as well as the reduced basal and ligand-induced Akt activation (Figs. 6C, 7), was not affected by the presence or absence of MHCII. Therefore, we conclude that, in SPPL2a^−/− B cells, MHCII is dispensable for the capacity of the CD74 NTFs to modulate membrane trafficking and signaling pathways.

In this study, we systematically compared the B cell phenotypes associated with the knockout of the two proteases CatS and SPPL2a that catalyze distinct steps in the degradation of CD74. Based on the major B cell–maturation phenotype of SPPL2a^−/− mice, which is caused by uncleaved CD74 NTF2 (6, 7), we aimed to assess the effects of a different CD74-degradation intermediate that is accumulating in CatS-deficient B cells. Furthermore, we assessed the role of MHCII in the development of the phenotype in SPPL2a^−/− mice by generating and analyzing SPPL2a-MHCII double-deficient mice. The findings from the analysis of the respective mouse lines are summarized in Fig. 7.

Based on our results, we can clearly exclude an active role for free MHCII in mediating the B cell phenotype of SPPL2a^−/− mice. As discussed above, several examples have been reported whereby altered MHCII homeostasis or unpaired MHCII chains negatively affect B cell maturation (20–22). The most prominent example is the B cell phenotype of CD74^−/− mice, which can be reversed by additional ablation of MHCII (21). In addition to the CLIP segment, other MHCII interaction sites within the transmembrane segments of CD74 have been reported (31). This may explain why SPPL2a^−/− B cells exhibit increased levels of MHCII (6) and could suggest that the accumulated CD74 fragment induces cross-linking of MHCII, which might trigger any MHCII-associated signaling pathways (17, 18). However, analysis of the SPPL2a-MHCII double-deficient mice in this study revealed that changes in trafficking or signal-transduction pathways, as well as the B cell–maturation phenotype, of SPPL2a^−/− mice are not reverted by additional ablation of MHCII. Therefore, we conclude that these defects are independent of MHCII and represent intrinsic consequences of the accumulating CD74 NTF, which is supported by the major phenotypic rescue in SPPL2a-CD74 double-deficient mice (6, 10). However, at the current time, additional CD74-independent functions of SPPL2a in B cells with an impact on BCR signaling and B cell maturation may not be fully excluded. Of course, this could also have implications with regard to the apparent phenotypic differences between SPPL2a- and CatS-deficient mice discussed below.

In agreement with previous reports (23), we found that the absence of CatS does not lead to a reduction in mature B cells in lymphatic tissues, which argues against a major impairment of B cell differentiation. Nevertheless, we observed a tendency toward a minor, isolated reduction in the T2 B cell population. Conspicuously, this is the first population in which the B cell phenotype of SPPL2a^−/− mice becomes fully evident, as demonstrated by its near-complete depletion (Table II) (6–8). Apparently, the described reduction in the T2 population in CatS mice is not reflected in corresponding changes in the pool of mature B2 B cells in the spleen (Fig. 1D) or other lymphatic tissues (Table I). However, T2 B cells are intermediate short-lived cells that develop into mature B cells within 24 h (32) and are characterized by a high turnover compared with the latter, which was demonstrated in BrdU-labeling studies (33). Therefore, it seems likely that subtle negative effects on the survival of developing and mature B cells would be seen first in this transient rapidly cycling population and might still be compensated for in the long-lived mature population. Thus, although CatS deficiency does not impair B cell maturation to a relevant degree, some minor changes can be detected that are reminiscent of the phenotype of SPPL2a^−/− mice in a very subtle form.

In agreement with the absence of a major maturation defect, basal and ligand-induced signal transduction of the BCR in CatS^−/− B cells was similar to wt cells. This is in contrast to SPPL2a^−/− B cells, in which these pathways are significantly impaired, and that we proposed as a mechanism underlying the described phenotype (10). In this context, the role and importance of the morphological changes in the endocytic system are still unclear. Electron microscopic analysis of CatS-deficient LPS blasts has been reported (34) demonstrating the accumulation of endosome-derived vacuoles. We confirm in this study that these are also present in CatS-deficient primary unstimulated IgM⁺ B cells. In general, this substantiates the capability of the CD74 N terminus to modulate endosomal membrane trafficking, which was proposed based on CD74-overexpressing cells that exhibited enlarged endosomes and a delayed endosomal maturation (1, 35–37). The vacuoles that we observed in CatS^−/− B cells were similar to those in SPPL2a^−/− B cells (6, 29) with regard to size, shape, and appearance. Only their abundance, with a mean number of approximately three per cellular profile, was slightly lower than the approximately four or five vacuoles that we had detected in SPPL2a^−/− B cells in the preceding analyses (6, 29); also, the distribution of the detected vacuoles within the analyzed cell population was more heterogenous.

In sharp contrast to SPPL2a^−/− B cells, functionality of the endosomal pathway was unaffected in CatS^−/− T1 B cells with regard to the performed assays. Similarly, the impairment of BCR signaling and the B cell–maturation defect associated with the loss of SPPL2a were not detected in CatS-deficient B cells, or they only became evident as subtle tendencies. This could indicate that the accumulation of endosomal vacuoles is an independent phenomenon that is not directly linked to the other cellular effects of the CD74 NTF and to the failure of the B cells to undergo further maturation. However, it is possible that these different cellular phenotypes represent a continuum of effects with different dose-response curves that are based on the same molecular mechanism triggered by a CD74 NTF. In such a model, endosomal vacuolation might be the first manifestation triggered by the CD74 NTF that precedes a delay in endosomal maturation and, finally, the downregulation of BCR signaling. Such a model, at least with regard to the connection between morphological and functional changes in the endocytic system, might be supported by the slightly retarded kinetics of OVA degradation in mature CatS^−/− B cells. Interestingly, these CatS-deficient mature B cells exhibit higher CD74 NTF levels than do the respective T1 cells. Following this model, a more severe vacuolization of mature B cells compared with the total pool of splenic B cells also containing T1 and T2 may be expected; however, this was not observed (Fig. 2I). Altogether, these data may indicate that there is no direct correlation between the number of vacuoles and the functional impairment of endosomal cargo degradation, as well as BCR trafficking and signaling. Based on the current data, it cannot be excluded that different pathways and protein interactions of CD74 NTF are involved in these different effects, which could respond differentially to increasing levels of CD74 NTF. A molecular interaction between the N terminus of CD74 and the heat shock protein 70 (Hsc70) was reported to be critical for the morphological effects of CD74 on the endocytic system (38). Based on our findings, it may be questioned whether this CD74–Hsc70 interaction is also responsible for the delayed CD74 NTF–induced endocytic cargo degradation, as well as the impairments in BCR signaling and B cell maturation.

We have demonstrated that block of CD74 processing in CatS-deficient B cells is incomplete and that a certain amount of NTF1 bypasses this block, clearly arguing for some functional redundancy among endosomal proteases. It is well established that this proteolytic step, which separates the MHCII-bound CLIP segment from CD74, can be conducted by other cysteine cathepsins. Cathepsins L (4) and V (human) (39) and cathepsin F (5) perform this cleavage event in thymic epithelial cells and macrophages, respectively, thus the transition from NTF1 to NTF2. Therefore, these enzymes would be primary candidates for compensating for the loss of CatS. Cathepsins B, H, and D were unable to process CD74 at this specific site (2); however, it is possible that other endosomal proteases not yet linked to the processing of CD74 are involved in this process.

An interesting question in the context of CD74 processing is the physiological relevance of the CD74 intracellular domain (ICD), which is released into the cytosol (9, 40). This is still not fully understood (1). Suggested functions include the activation of NF-κB signaling, as well as other signaling pathways (40–44). Because generation of CD74 NTF2 is only partially blocked in CatS^−/− B cells, cytosolic liberation of the CD74 ICD will be preserved in these cells to a major extent, in contrast to SPPL2a^−/− B cells, in which this would be abandoned completely. Thus, any putative functions of the CD74 ICD would remain active, despite CatS deficiency. This represents an additional difference between the two model systems. However, because the phenotype of SPPL2a^−/− mice is driven by the membrane-bound CD74 NTF (6, 7, 10), we consider it unlikely that this difference in ICD production directly contributes to the phenotypic differences between CatS^−/− and SPPL2a^−/− mice. Nevertheless, it should be kept in mind when analyzing these mice.

The described incomplete cleavage block in CatS^−/− B cells is a likely explanation for why the NTF1 steady-state levels in these cells do not reach those of NTF2 in SPPL2a^−/− B cells and instead remain much lower. Furthermore, this significant discrepancy between the fragment levels in both model systems could account for the described phenotypic differences. However, we do not know whether these two fragments are functionally equivalent, particularly with regard to their impact on signaling and B cell maturation. In light of the highly divergent NTF1 and NTF2 levels, a comparison of SPPL2a- and CatS-deficient B cells can neither prove nor disprove any potentially different properties of these two fragments, because comparable abundance would be a prerequisite. It is possible that mice lacking multiple CD74-processing cysteine cathepsins, which might exhibit greater NTF1 accumulation, could be useful for testing this hypothesis. In this context, it should be emphasized that the longer NTF1 is still tightly bound to MHCII, which could sterically influence the spectrum and/or strength of protein interactions compared with NTF2. Therefore, it seems conceivable that this association with MHCII mitigates its downstream effects. In addition to the lower abundance, this also could explain why CatS^−/− B cells are not compromised by the accumulation of NTF1, which forms a complex with MHCII, whereas SPPL2a^−/− B cells are severely affected by free NTF2. The generation and analysis of CatS-MHCII double-deficient mice might be pursued in the future to test this hypothesis.

Altogether, we believe that our results further support a model in which successful CD74 degradation represents a checkpoint for B cells to undergo further differentiation and to reach maturity. From the cell’s perspective, degradation of CD74 might be an indicator of the functionality of its Ag-presentation machinery, and preventing cells with a defect in this system from further differentiation intuitively makes sense. Our findings support a dose dependence of cellular responses to accumulating CD74 fragments. Apparently, minor alterations and incomplete blockages of CD74 degradation are tolerated; only above a certain threshold BCR signaling is significantly impaired, thereby preventing further maturation of the affected B cell. We can now rule out an involvement of MHCII in this pathway. Therefore, further work will focus on defining the intrinsic molecular interactions of CD74 NTF that mediate these effects.

We thank Sebastian Held, Marlies Rusch, and Dagmar Niemeier for excellent technical assistance. Some images were produced using the UCSF Chimera package from the Computer Graphics Laboratory, University of California, San Francisco (supported by National Institutes of Health Grant P41 RR-01081). We also thank Sandra Ussat (Institute of Immunology, Christian-Albrechts-University of Kiel) for excellent conduction of FACS sorting. We are grateful to Prof. Willem Stoorvogel (University of Utrecht) for providing anti-MHCII β-chain Ab.

The authors have no financial conflicts of interest.