Signal peptide peptidase–like 2a (SPPL2a) is an aspartyl intramembrane protease essential for degradation of the invariant chain CD74. In humans, absence of SPPL2a leads to Mendelian susceptibility to mycobacterial disease, which is attributed to a loss of the dendritic cell (DC) subset conventional DC2. In this study, we confirm depletion of conventional DC2 in lymphatic tissues of SPPL2a−/− mice and demonstrate dependence on CD74 using SPPL2a−/− CD74−/− mice. Upon contact with mycobacteria, SPPL2a−/− bone marrow–derived DCs show enhanced secretion of IL-1β, whereas production of IL-10 and IFN-β is reduced. These effects correlated with modulated responses upon selective stimulation of the pattern recognition receptors TLR4 and Dectin-1. In SPPL2a−/− bone marrow–derived DCs, Dectin-1 is redistributed to endosomal compartments. Thus, SPPL2a deficiency alters pattern recognition receptor pathways in a CD74-dependent way, shifting the balance from anti- to proinflammatory cytokines in antimycobacterial responses. We propose that in addition to the DC reduction, this altered DC functionality contributes to Mendelian susceptibility to mycobacterial disease upon SPPL2a deficiency.
This article is featured in Top Reads, p.1
The intramembrane protease signal peptide peptidase–like 2a (SPPL2a) is capable of cleaving type II–oriented membrane proteins within their hydrophobic transmembrane segment (1, 2). SPPL2a-mediated processing of the CD74 protein, also known as the invariant chain of the MHC class II (MHCII) complex, plays a central role in APCs (3). In mice, genetic ablation of SPPL2a leads to an arrest of B cell maturation at the transitional stage 1 and a reduction of dendritic cell (DC) numbers (4–6). As these phenotypes were significantly ameliorated in SPPL2a−/− CD74−/− mice (4, 5), the massive accumulation of uncleaved membrane-bound CD74 N-terminal fragments (NTFs) in the absence of SPPL2a could be identified as the underlying mechanism. This clearly highlights the requirement of SPPL2a for CD74 degradation as well as for homeostasis of B cells and DCs in mice. Induced by disturbed membrane traffic in SPPL2a−/− B cells, endosomal vacuoles accumulate in these cells, which goes along with a mislocalization and reduced surface exposure of receptors involved in B cell differentiation and survival like the B cell AgR (BCR) (7, 8). Consequently, activation of survival signaling pathways downstream of the BCR, in particular of the PI3K/Akt pathway, was reduced in these cells (7, 8). Although these findings provided important insights into how the accumulating CD74 NTF negatively affects B cell development and survival, the direct molecular link between the CD74 NTF and its impact on membrane trafficking and signal transduction remains to be unraveled.
Recently, three human individuals with genetic SPPL2a deficiency because of splice site mutations were identified (9). These patients presented clinically with a Mendelian susceptibility to mycobacterial disease (MSMD) manifesting after application of the bacillus Calmette–Guérin (BCG) tuberculosis vaccine (9). MSMD summarizes a rare group of innate immune defects characterized by a rather selective susceptibility to attenuated and environmental mycobacteria, typically in the absence of further overt immunological abnormalities (10, 11). In particular, mutations in genes involved in regulating action and/or production of IFN-γ, such as cytokines of the IL-12 family and their receptors, have been linked with MSMD (12, 13). Whereas B cell development was not significantly compromised in the three SPPL2a-deficient patients, they showed a reduction of conventional DCs (cDCs) similar to the phenotype of SPPL2a−/− mice (5, 6, 14). This cDC depletion affected primarily the cDC2 population (9). In humans, these cDC2 cells are known to be the main producers of IL-12 family cytokines, which are critically involved in priming TH1 immune responses (15). Therefore, cDC2 depletion was considered to represent a major cause of MSMD in the SPPL2a-deficient patients (9). The mycobacterial susceptibility phenotype could also be observed in SPPL2a-deficient mice. These mice exhibited impaired pathogen clearance upon i.v. infection with the attenuated tuberculosis vaccine strain BCG as well as following aerosol infection with Mycobacterium tuberculosis (9). In the latter experiment, they succumbed to the infection, whereas wild types survived the observation period of 28 wk. Similar to mice, human SPPL2a-deficient B cells (16) and DCs (9) accumulate major amounts of CD74 NTFs. Although strongly suggested by previous observations in SPPL2a-CD74 double-deficient mice, formal evidence that the phenotypes of the SPPL2a-deficient patients are caused by the CD74 fragments is difficult to obtain.
In this study, we employed our SPPL2a−/− mice as a model to further characterize the effect of SPPL2a deficiency on DC differentiation and function. We provide a comprehensive analysis of DC subsets in different lymphatic tissues of SPPL2a−/− mice, thereby significantly expanding previous analyses (6, 9), and analyze the CD74 dependence of the observed phenotypes. Upon interaction with mycobacteria, SPPL2a-deficient DCs exhibit a major shift in the secreted cytokine profile with a reduction of anti-inflammatory and enhanced secretion of proinflammatory cytokines. Our findings identify SPPL2a as a modulator of pattern recognition receptor (PRR) signaling pathways exemplified by TLR4 and Dectin-1. Altogether, we demonstrate that SPPL2a, beyond affecting DC homeostasis, also significantly modulates DC functionality, in particular with regard to antimycobacterial responses. This likely contributes to the pathophysiological mechanism underlying MSMD in SPPL2a-deficient mice and humans.
Materials and Methods
SPPL2a−/− (4), CD74−/− (17), and SPPL2a−/− CD74−/− (4) mice have been described before and were on a C57BL/6N Crl background. All experiments were performed with littermates and/or controls matched for sex and age. Breeding of mice has been approved by the Ministerium für Energiewende, Landwirtschaft, Umwelt und ländliche Räume of Schleswig-Holstein (V 242.7224.121-3) as well as the Landesdirektion Sachsen (DD24.1-5131/450/12). Care, handling, and euthanasia of the animals were conducted according to local and national guidelines.
Generation of bone marrow–derived DCs
Bone marrow–derived DCs (BMDCs) were generated according to the protocol described by Lutz et al. (18) with minor modifications. Soft tissue was removed from the hind legs of mice using sterile gauze swabs. Tibia and femur were cut open on both sides. Red bone marrow was flushed into a petri dish with 5–10 ml of BMDC medium consisting of RPMI 1640 medium supplemented with 10% FBS (Biochrom), 100 U/ml penicillin (Sigma-Aldrich), 100 μg/ml streptomycin (Sigma-Aldrich), and 50 μM 2-ME (Life Technologies) using a 27-gauge cannula. The obtained bone marrow was dissociated with a 23-gauge cannula, and the resulting suspension was passed through a 100-μm cell strainer (BD Biosciences). A total of 5 × 106 cells was seeded in 10 ml of BMDC medium supplemented with 20 ng/ml recombinant murine GM-CSF (Immunotools, Friesoythe, Germany) into a 10-cm petri dish (Sarstedt). Cells were cultured at 37°C and 95% air/5% CO2 under humidified air conditions. After 3 d, 10 ml of BMDC medium containing 20 ng/ml GM-CSF was added. After another 3 d, 10 ml of cell suspension was removed from each plate and centrifuged for 10 min at 210 × g. The sedimented cells were resuspended in 10 ml of BMDC medium supplemented with 10 ng/ml GM-CSF and added back to the culture dish. Cells were used either for direct analysis or further experiments after a total culture period of 8 d. Suspension and adherent cells, which were detached with Accutase (Thermo Fisher Scientific), were combined. For stimulation experiments, cells were seeded in BMDC medium supplemented with 10 ng/ml GM-CSF in a total volume of 1 ml at densities of 2 × 106, 1.5 × 106, or 0.5 × 106 cells/ml for protein isolation for signaling analysis, RNA isolation, or recovery of media for ELISA analysis, respectively. For stimulation and induction of cytokine responses, heat-killed M. tuberculosis (HKMT; Invivogen), depleted zymosan (dZym; Invivogen), and LPS from Escherichia coli O26:B06 (LPS; Sigma-Aldrich) were used as PRR ligands. If not stated otherwise, HKMT, dZym, and LPS were applied at final concentrations of 250 μg/ml, 50 μg/ml, and 500 ng/ml, respectively. For experiments with viable BCG mycobacteria, cells were seeded as described above but in medium devoid of antibiotics.
Quantification of DC subsets in lymphatic tissues
To analyze DC subsets in lymphatic tissues from different mouse strains, spleen, thymus, and lymph nodes were minced and dissociated with 75 IU/ml Collagenase IV and 50 μg/ml DNase I dissolved in 1× HBSS supplemented with 2% FCS for 30 min at 37°C. Bone marrow was flushed from tibia and femur with RPMI 1640 containing 2% FCS (RPMI+). Cell suspensions were passed through a 100-μm cell strainer. After sedimentation, cells were resuspended in 155 mM NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA (pH 7.4) and incubated for 5 min at room temperature (RT) if lysis of erythrocytes was necessary. The reaction was stopped by addition of RPMI+. In all cases, cells were rinsed through a 40-μm cell strainer (BD Bioscience) with RPMI+ and washed three times with FACS buffer (PBS + 2% FCS). Up to 1 × 106 cells were used for Ab labeling in FACS buffer. After blocking of Fcγ receptors with αFcγRIIB/RIII (clone 2.4G2) and αFcγRIV (clone 9E9), cells were labeled with Abs for CD3e (clone 17A2), CD11b (clone RM4-5), CD11c (clone HL3), CD19 (clone 1D3), CD45 (clone 30-F11), CD45R (B220; clone RA3-6B2), CD103 (clone 2E7), CD115 (M-CSFR; AFS98), CD161b/c (NK1.1, Ly-55; clone PK136), CD172A (signal regulatory protein α [SIRPα]; P84), CD317 (BST2/PDCA-1; 927), CD370 (Clec9A, DNGR1; 7H11), CX3CR1 (clone SA011F11), Ly-6C (clone HK1.4), Ly-6G (clone 1A8), MHCII (I-A & I-E; clone M5/114.15.2), and XCR1 (clone ZET). Subsequently, cells were washed twice after each staining step and resuspended in FACS buffer supplemented with DAPI prior to flow cytometric analysis with an LSR Fortessa SORP Cytometer (BD Bioscience).
Flow cytometric analysis of BMDCs
For flow cytometric analysis of BMDCs, suspension and adherent BMDCs were collected as described above. Cells were sedimented for 10 min at 210 × g and resuspended in MACS buffer (PBS, pH 7.4, with 0.5% [w/v] BSA [Roth] and 2 mM EDTA). A total of 1 × 105 cells were incubated with Ab dilutions in MACS buffer in the dark for 30 min on ice. After washing and resuspension in MACS buffer, cells were subjected to flow cytometric analysis. The following Abs for staining cell surface proteins of BMDCs were used: CD11c-FITC (N418; BD Biosciences), MHCII-APC (M5/114.15.2; BioLegend), CD11b-PE/Cy7 (M1/70; eBioscience), Dectin-1–PE (bg1fpj; eBioscience), TLR2-PE (CB225; BioLegend), and TLR4-PE (SA15-21; BioLegend). For detection of intracellular Ags like the N terminus of CD74, cells were fixed and permeabilized using the BD Cytofix/Cytoperm kit (BD Bioscience) according to the manufacturer’s protocol. The CD74-FITC Ab (In-1; BD Biosciences) was diluted in BD Perm/Wash Buffer from the fix/perm kit. BMDC populations were gated according to expression of CD11c, MHCII, and CD11b. Flow cytometric analysis was performed using a FACSCanto II or an LSR II flow cytometer (BD Bioscience). In all cases, flow cytometric data were evaluated with FlowJo software (Tree Star).
Cytokine determination by ELISA
ELISAs to determine IL-1β, IL-10, IL-12p40, IL-12p70, IL-23, IFN-β, and TNF-α were performed according to the manufacturer’s protocol (Duo Set; R&D systems). In brief, high binding ELISA plates (Sarstedt) were coated with capture Abs dissolved in PBS overnight at RT. Remaining binding sites were blocked with 1% (w/v) BSA in PBS for at least 1 h at RT after three washes with 0.05% (v/v) Tween in PBS. Conditioned media were either used directly or diluted in respective culture medium. After washing, biotin-conjugated detection Ab was applied in 1% BSA in PBS for 2 h at RT, which was finally detected with HRP-coupled streptavidin diluted in the same buffer. HRP activity was visualized with the BM Blue POD Substrate, soluble (Roche). After stopping the reaction with 1 M H2SO4, absorbance was measured at 450 and 690 nm as control.
Culture of Mycobacterium bovis BCG
Mycobacterium bovis BCG (Pasteur ATCC 35734) was grown in BCG medium containing 0.47% (w/v) Middelbrook-7H9-Bouillon (BD Biosciences), 0.045% (v/v) Tween-80, and 10% (v/v) Middlebrook OADC Growth Supplement (both Sigma-Aldrich). Bacteria were cultured in an upright-positioned 25-cm2 tissue culture flask with a ventilation cap at 37°C and shaking at 150–200 rpm for 2 wk. Cultures were split 1:5 every 2–3 d. To prepare the mycobacteria for BMDC coculture experiments, the bacteria were sedimented for 10 min at 3500 × g and 4°C and washed once with PBS (Sigma-Aldrich). Mycobacteria were resuspended in 1 ml of PBS with a 27-gauge cannula. For photometric determination of bacterial cell number, an aliquot of the bacterial suspension was diluted 1:10 in 4% paraformaldehyde dissolved in PBS and the OD measured at 595 nm. An OD of 0.1 corresponded to 5 × 106 bacteria/ml in this dilution. For the coculture experiments, the resuspended mycobacteria were diluted in BMDC medium without penicillin and streptomycin and added to BMDC cultures in a multiplicity of infection (MOI) of 10.
RNA from BMDCs was isolated with the NucleoSpin RNA Plus Kit (Machery-Nagel). Suspension cells were centrifuged for 10 min at 210 × g and RT. Adherent cells were lysed directly in the culture dish in the lysis buffer provided in the kit without prior detachment. The lysate from the adherent cells was used to resuspend and lyse the sedimented suspension cells. All further steps were performed according to the manufacturer’s instructions. RNA concentration was determined photometrically. RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) was used for generation of cDNA using Random Hexamer Primers. Up to 1 μg of RNA was applied for cDNA synthesis, with equal amounts being used within each experiment. The mRNA abundance of specific target genes was measured by quantitative real-time PCR (qRT-PCR) using the Universal Probe Library System Technology and a LightCycler 480 Instrument II from Roche. The following primer and probe combinations were used: IL-1β: 5′-TTGACGGACCCCAAAAGAT-3′, 5′-GAAGCTGGATGCTCTCATCTG-3′, probe 26; IL-10: 5′-CAGAGCCACATGCTCCTAGA-3′, 5′-TGTCCAGCTGGTCCTTTGTT-3′, probe 41; IL-12p19: 5′-TCCCTACTAGGACTCAGCCAAC-3′, 5′-AGAACTCAGGCTGGGCATC-3′, probe 19; IL-12p35: 5′-TCAGAATCACAACCATCAGCA-3′, 5′-CGCCATTATGATTCAGAGACTG-3′, probe 49; IL-12p40: 5′-TTGCTGGTGTCTCCACTCAT-3′, 5′-GGGAGTCCAGTCCACCTCTAC-3′, probe 78; IFN-β: 5′-CTGGCTTCCATCATGAACAA-3′, 5′-AGAGGGCTGTGGTGGAGAA-3′, probe 18; SDHa: 5′-TGTTCAGTTCCACCCCACA-3′, 5′-TCTCCACGACACCCTTCTG-3′, probe 71; TLR4: 5′-CTGATCCATGCATTGGTAGGT-3′, 5′-GGACTCTGATCATGGCACTG-3′, probe 2; TLR2: 5′-ACCGAAACCTCAGACAAAGC-3′, 5′-AGCGTTTGCTGAAGAGGACT-3′, probe 49; TNF: 5′-CTGTAGCCCACGTCGTAGC-3′, 5′-TTGAGATCCATGCCGTTG-3′, probe 25; Tub1a: 5′-CTGGAACCCACGGTCATC-3′, 5′-GTGGCCACGAGCATAGTTATT-3′, probe 88; Dectin-1: 5′-AGAGTGAAGGGCCATGGTT-3′, 5′-TGCATTTCTGACTTGAAACGA-3′, probe 88; and HPRT1: 5′-CCTCCTCAGACCGCTTTTT-3′, 5′-AACCTGGTTCATCATCGCTAA-3′, probe 95. A serial dilution of samples was performed for each gene of interest to evaluate primer efficiency. Calculation of mRNA abundance from the obtained Cp values has been described before (7). Values were normalized to at least one or a mean of the following housekeeping genes: tubulin (Tub1a), succinate dehydrogenase (SDH), and hypoxanthine–guanine phosphoribosyltransferase (HPRT).
Protein extraction and Western blot analysis
After recovery of suspension cells from the BMDC culture, the adherent cells were scraped off in PBS supplemented with Complete Protease Inhibitor (Roche). Suspension and adherent cells were combined and collected by centrifugation for 10 min at 210 × g and 4°C. Cells were washed once in PBS and then lysed by resuspension 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 supplemented with protease inhibitors as described (19)] and incubation on ice for 1 h. For detection of phosphorylated proteins, PhosStop phosphatase inhibitor mixture (Roche) as well as 20 mM β-glycerophosphate, 4 mM sodium orthovanadate and 20 mM sodium fluoride were added to the lysis buffer, and in these cases samples were incubated for 30 min only. During the incubation, samples were sonicated at level 4 for 20 s at 4°C using a Branson Sonifier 450 (Emerson Industrial Automation, Danbury, CT). Afterwards, lysates were cleared for 10 min at 17,000 × g and 4°C, and protein concentration in the lysates was measured by bicinchoninic acid protein assay (Thermo Fisher Scientific). Proteins were separated by SDS-PAGE. Depending on the protein to be detected, either a Tris-tricine (20) (CD74) or a standard Tris-glycine (21) buffer system (all other proteins) was employed. After semidry transfer to nitrocellulose membranes, immunodetection was performed as described (19). For detection of CD74, the monoclonal In-1 Ab (BD Biosciences) was employed. A polyclonal antiserum against SPPL2a has been described previously (22). Rabbit mAbs against TLR2 (E1J2W) and TLR4 (D8L5W) were obtained from Cell Signaling. For detection of murine Dectin-1, rabbits were immunized with a synthetic peptide from the N terminus of the protein (aa 2–20; KYHSHIENLDEDGYTQLDF) and the obtained serum affinity purified against the immobilized immunogen (Pineda Antikörper-Service, Berlin, Germany). Rabbit Abs against total and phosphorylated IRF3 (phospho-IRF3, S396: clone 4D4G; total IRF3: clone D83B9) and p38 (phospho-p38 Thr180/Tyr182: clone D3F9; total p38: polyclonal, no. 9212) were purchased from Cell Signaling. To confirm equal loading, anti-Actin (Sigma-Aldrich), anti-EEF2 (Abcam), or anti-Cofilin (D3F9; Cell Signaling) were used. HRP-conjugated secondary Abs were purchased from Dianova. HRP activity was visualized with Lumigen ECL Ultra (TMA-6; Lumigen), and chemoluminescent signals were documented using an ImageQuant LAS 4000 (GE Healthcare). Densitometric analysis of blot images was performed using ImageJ.
Biotinylation of cell surface proteins
Suspension cells were recovered from differentiated BMDC cultures. Adherent cells were detached from the plate with Accutase and then combined with the suspension cells for further processing. Cells were sedimented for 10 min at 210 × g and 4°C and washed twice with ice-cold PBS-CM (PBS, pH 8, containing 0.1 mM CaCl2 and 1 mM MgCl2). Cells were resuspended in 1 mg/ml Sulfo-NHS-SS-Biotin (Thermo Fisher Scientific) dissolved in PBS-CM and incubated for 30 min on ice. Controls without biotin were incubated in parallel.
Subsequently, cells were sedimented and resuspended in 50 mM Tris-HCl in PBS (pH 8) to stop the biotinylation reaction for 10 min on ice. Two washing steps were performed with PBS-CM. Cells were lysed in 600 μl of lysis buffer for 1 h on ice as described above. Equal protein amounts were used for the isolation of biotinylated proteins using High-Capacity Streptavidin Agarose Beads (Thermo Fisher Scientific). Pulldown was performed for 1 h at 4°C. After several washing steps, bound proteins were eluted in 1× SDS-PAGE loading buffer for 5 min at 95°C followed by 15 min at 37°C. The recovered proteins were then subjected to SDS-PAGE and Western blot analysis.
Coverslips were coated with 100 μg/ml poly-L-lysine (Sigma-Aldrich) in six-well plates overnight at RT. Afterwards, coverslips were rinsed with cell culture grade water (Sigma-Aldrich) three times for 1 h. BMDCs were collected as described above, and 1.5 × 105 cells were seeded per well in 1 ml of RPMI medium without any additives. Cells were allowed to adhere to the poly-L-lysine–coated coverslips for 2 h at 37°C and 95% air/5% CO2. Coverslips were then rinsed three times with PBS, and cells were fixed with 4% paraformaldehyde in PBS for 20 min at RT. After three washes with PBS, cells were incubated in permeabilization buffer (0.2% [w/v] saponin in PBS) for 5 min at RT. Free aldehyde groups were blocked with 0.12% (w/v) glycine in permeabilization buffer for 10 min at RT. After one washing step, unspecific binding sites were blocked with 10% (v/v) FBS in permeabilization buffer for at least 1 h at RT. Primary and secondary Abs were diluted in this solution. For detection of murine Dectin-1, the above-described newly generated polyclonal rabbit antiserum was employed. CD74 and LAMP-2 were visualized with the mAbs In-1 (BD Biosciences) and ABL93 (DSHB), respectively. The primary Abs were applied overnight at 4°C. Alexa-488– and -594–conjugated secondary Abs (Thermo Fisher Scientific) were used. Coverslips were mounted in Mowiol 4-88 supplemented with DABCO (1,4-Diazabicyclo[2.2.2]octan; Sigma-Aldrich) and DAPI (Sigma-Aldrich) on glass slides. Pictures were acquired with a Zeiss LSM 880 Airy confocal laser scanning microscope and the corresponding software. Colocalization was analyzed using FiJi with the JaCoP plugin (23). To exclude cells that were only partially covered by the images, a region of interest was drawn around completely imaged cells. The command “clear outside” was employed, and the background was subtracted from all channels using a “rolling ball radius” of 20 pixels. The JaCoP plugin was executed selecting the red channel as image A and the green channel as image B. A threshold was set automatically. For evaluation of colocalization, the Manders coefficient 2 (M2) based on the signals above the set threshold was calculated. From each specimen, three images were analyzed, and the mean was calculated.
Data are depicted as mean ± SD. Statistical significance was assessed using unpaired two-tailed Student t test and one-way or two-way ANOVA followed by Tukey, Bonferroni, or Sidak post hoc testing, as indicated. Significance levels of *p < 0.05, **p < 0.01, and ***p < 0.001 were applied.
SPPL2a−/− mice show a CD74-dependent cDC depletion
Based on the previous observation that CD11c+ MHCII+ cells are reduced in spleens and lymph nodes of SPPL2a−/− mice (14), we quantified DC subsets according to Supplemental Fig. 1 in spleen (Fig. 1A, Supplemental Fig. 2A), bone marrow (Fig. 1B), thymus (Fig. 1C, Supplemental Fig. 2B), and lymph nodes (Fig. 1D) of these animals. To determine CD74 dependency of DC phenotypes, SPPL2a-CD74 double-deficient (4) as well as CD74 single-deficient mice (17) were analyzed in comparison. The frequencies of plasmacytoid DCs were not relevantly affected in any of the tissues of SPPL2a-deficient mice. In contrast, the abundance of total cDCs was significantly reduced in spleen, bone marrow, and lymph nodes of SPPL2a−/− mice compared with wild type. In spleen, bone marrow, and thymus, exclusively, the cDC2 subset was significantly reduced, whereas the cDC1 population was unchanged in SPPL2a-deficient mice. In lymph nodes, this subset selectivity was not seen. There, also the abundance of cDC1 was significantly reduced. Although the reduction of the resident cDC1 and cDC2 populations was not statistically significant, there was a significant loss of both migratory cDC1 and cDC2 in lymph nodes, which may indicate an impact of SPPL2a deficiency on DC migration. In contrast, frequencies of migratory DC subsets in the thymus were not changed in SPPL2a−/− mice. Importantly, all described effects on resident and migratory DCs observed in spleen, thymus, and lymph nodes of SPPL2a−/− mice were rescued in the SPPL2a-CD74 double-deficient mice, demonstrating that these phenotypes are caused by the CD74 NTF accumulation. The normal frequencies of the cDC1 population in bone marrow and spleen indicate that differentiation of this DC subset is not affected. However, based on their depletion from lymph nodes, it seems likely that these cells are lost with a certain latency. This indicates a function of SPPL2a not only in cDC2, which are more severely affected, but also a relevant role in survival and/or functionality of the cDC1 population.
No relevant impairment of BMDC differentiation by SPPL2a deficiency
We compared the differentiation ability of wild-type and SPPL2a−/− bone marrow cells into BMDCs in the presence of GM-CSF in vitro (Fig. 2A). In addition to the overall yield of CD11c+MHCII+ cells, we also compared MHCint and MHChigh subpopulations discriminated by CD11b and MHCII surface levels according to Helft et al. (24). Similar cell numbers were obtained from cultures of both genotypes, and marker expression pattern of the populations was largely comparable. The MHChigh population in SPPL2a−/− BMDC cultures exhibited slightly reduced CD11c surface levels (Fig. 2A). Nevertheless, these data indicate that despite the major loss of DC populations in vivo, differentiation in this culture system is unaffected by SPPL2a deficiency. We confirmed that the SPPL2a-deficient BMDCs accumulate CD74 NTFs by Western blotting (Fig. 2B) and flow cytometry (Fig. 2C) using an Ab against an N-terminal epitope of the protein. We found that the overall abundance of CD74 compared between the different wild-type subpopulations and also the CD74 NTF accumulation compared between the corresponding wild-type and SPPL2a−/− subpopulations correlated with MHCII expression. Thus, highest levels were observed in the MHChigh population. Based on these results revealing BMDC populations with comparable differentiation marker phenotypes from both wild-type and SPPL2a−/− mice, we considered this system suitable to study the impact of SPPL2a on DC function. As the DC depletion is not complete in vivo, how the accumulating CD74 NTF modulates the functionality of the remaining DCs is a critical question.
SPPL2a−/− BMDCs show altered cytokine responses to mycobacteria
Based on the MSMD phenotype of SPPL2a-deficient patients and mice, we specifically sought to determine cytokine responses upon recognition of mycobacteria. Therefore, we stimulated wild-type and SPPL2a−/− BMDCs with HKMT. Because of their specific role in promoting TH1 polarization of T cells, which is critical for mycobacterial defense, we measured the IL-12 family cytokines IL-23 (Fig. 3A), IL-12p40 (Fig. 3B), and IL-12p70 (Fig. 3C). Secretion of these cytokines upon HKMT stimulation was comparable between wild-type and SPPL2a−/− BMDCs. However, the release of the anti-inflammatory cytokine IL-10 was significantly reduced in SPPL2a-deficient cells to ∼40% of the wild-type level (Fig. 3D). In contrast, these cells secreted a 2-fold higher amount of the proinflammatory IL-1β (Fig. 3E). We excluded that SPPL2a−/− BMDCs show increased cell death in our experimental set-up, as this might imitate an increased release of IL-1β, which follows an unconventional secretion pathway after activation of pro–IL-1β by the inflammasome (25) (Supplemental Fig. 3). Production of the proinflammatory TNF was also significantly increased, however, only to a minor degree (Fig. 3F). IL-10 and IL-1β secretion of SPPL2a−/− CD74−/− (double knock-out [DKO]) and also CD74−/− BMDCs was similar to that of wild-type cells, whereas TNF secretion even appeared to be slightly reduced in cells lacking CD74. Thus, complete reversion of the differential secretion of IL-10, IL-1β, and TNF by SPPL2a−/− BMDCs by additional depletion of CD74 demonstrates that the accumulating CD74 NTF is responsible for these phenotypes. In addition to quantification of the secreted cytokines by ELISA, we also analyzed induction of IL-10 (Fig. 3G), IL-1β (Fig. 3H), and TNF (Fig. 3I) mRNA upon HKMT stimulation of wild-type and SPPL2a−/− BMDCs by qRT-PCR. We observed a significant decrease of IL-10 transcripts but a significant increase of IL-1β and TNF mRNA corroborating the cytokine secretion data. These results suggest an impact of SPPL2a in this process upstream of the cytokine induction.
In the following, we aimed to determine whether the altered cytokine profiles upon treatment with HKMT also have implications for the responses toward viable mycobacteria. Therefore, we cocultured BMDCs from all four genotypes with BCG mycobacteria at an MOI of 10 and determined cytokine secretion and expression by ELISA and qRT-PCR, respectively (Fig. 4). Similar to HKMT, secretion of IL-12 family cytokines IL-23 and IL-12p70 was not significantly changed upon loss of SPPL2a (Fig. 4A–D). We observed a significant reduction of IL-12p40 mRNA but not of the secreted protein in the SPPL2a−/− cells upon induction by live BCG (Fig. 4E, 4F). TNF mRNA expression and secretion induced by live BCG showed similar profiles across the genotypes as observed upon HKMT stimulation but without statistically significant differences (Fig. 4G, 4H). IL-10 secretion by BMDCs upon BCG stimulation was, in general, low. Nevertheless, SPPL2a−/− BMDCs secreted significantly less IL-10 upon stimulation with BCG than wild-type cells (Fig. 4I, 4J) and showed significantly lower IL-10 mRNA levels. However, the reversal of this phenotype in the DKO cells was not as clear as seen upon stimulation with HKMT, likely reflecting the overall lower amounts of IL-10 upon stimulation with viable BCG. Very similar to effects induced by HKMT, SPPL2a-deficient BMDCs released ∼3-fold more IL-1β than wild-type cells, which, however, was reversed in DKO cells (Fig. 4K). Enhanced amounts of IL-1β secreted by SPPL2a−/− cells were reflected in an increased mRNA induction upon stimulation with live BCG (Fig. 4L). Altogether, these findings indicate that SPPL2a−/− BMDCs show altered cytokine responses toward mycobacteria, which were quite distinctive from wild-type cells. Results from BMDCs stimulated either with HKMT or viable BCG mycobacteria correlated well. Key findings upon both types of stimuli are downregulation of the anti-inflammatory cytokine IL-10 and upregulation of the proinflammatory IL-1β, which in sum can shift the immune balance toward inflammatory immune activation.
Dectin-1 and TLR4 responses are altered in SPPL2a−/− BMDCs
The altered cytokine induction in the absence of SPPL2a indicates that the CD74 NTF accumulation influences either recognition of the pathogens by PRRs or subsequent signal transduction cascades, leading to enhanced transcription. Based on the observed differential up- or downregulation of TNF, IL-1β, and IL-10, we hypothesized that more than one pathway may be affected. Therefore, we focused specifically on these three cytokines. With regard to recognition by PRRs, mycobacteria have been shown to interact with different members of the TLR and C-type lectin receptor (CTLR) families (26, 27). To analyze which pathways are altered in SPPL2a−/− versus wild-type BMDCs, we stimulated these cells with different established PRR ligands specific for single or families of receptors. Although many responses were unaffected, which confirms in general that SPPL2a−/− BMDCs are functional, we observed differential responses upon stimulation with dZym and LPS. dZym is a specific ligand for the CTLR Dectin-1 (28), which has been implicated in detection of mycobacteria (29). Dectin-1 stimulation induced mRNA expression as well as protein secretion of IL-10 (Fig. 5A, 5B), IL-1β (Fig. 5C, 5D), and TNF (Fig. 5E, 5F) in BMDCs. dZym-induced IL-10 secretion was significantly diminished in SPPL2a−/− BMDCs (Fig. 5A, 5B), which was fully reversed by additional CD74 knockout. In contrast, dZym induced comparable amounts of IL-1β and TNF in both wild-type and SPPL2a-deficient BMDCs (Fig. 5C–F). Upon 24 h of stimulation with the TLR4 ligand LPS, we were unable to detect IL-10 protein secretion in BMDCs by ELISA. However, using qRT-PCR induction of IL-10, mRNA expression was detected, which was only mildly downregulated in SPPL2a−/− BMDCs (Fig. 5G). LPS-induced secretion of TNF was not significantly different between wild-type and SPPL2a−/− BMDCs, although the induction of TNF was significantly increased in the latter (Fig. 5H, 5I). We observed a robust secretion of IL-1β (Fig. 5J, 5K), which was significantly enhanced in SPPL2a-deficient cells. Similar to the upregulation of IL-1β upon stimulation with either HKMT or BCG mycobacteria, the increase of LPS-induced IL-1β in SPPL2a−/− BMDCs was dependent on the CD74 NTF accumulation.
These findings demonstrate that in SPPL2a−/− BMDCs, both the Dectin-1 and TLR4 pathways are differentially altered as characterized by reduced responsiveness of the first and enhanced sensitivity of the latter. However, this does not seem to affect all responses induced by these receptors uniformly but rather selected cytokines, such as IL-10 and IL-1β. Altogether, the reduced IL-10 secretion following stimulation with dZym and the enhanced IL-1β pheno-copy the changes observed following interaction with heat-killed or viable mycobacteria. This strongly suggests that the altered responsiveness of the Dectin-1 and TLR4 pathways contributes to the altered antimycobacterial responses.
Reduced Dectin-1 surface levels in SPPL2a−/− BMDCs
We aimed to investigate the mechanisms underlying the modulation of PRR pathways in SPPL2a-deficient BMDCs. Because murine SPPL2a−/− B cells exhibit reduced surface levels of critical survival receptors (4, 7), we quantified the abundance of Dectin-1 by flow cytometry. In SPPL2a-deficient BMDCs, we observed significantly reduced Dectin-1 surface expression levels (Fig. 6A, 6B), which was rescued by additional CD74 deletion (Fig. 6B). We confirmed this observation by Western blot analysis of isolated plasma membrane proteins after labeling with a membrane-impermeable biotinylation reagent (Fig. 6C, 6D). Western blot analysis revealed significantly less Dectin-1 in the “bound” fraction recovered by streptavidin pulldown from SPPL2a−/− BMDCs compared with the other genotypes. Total Dectin-1 levels in cell lysates (“total”) were not significantly changed. In parallel with the reduction of the Dectin-1 full-length protein at the cell surface, SPPL2a−/− BMDCs accumulated a small Dectin-1 fragment in a CD74-dependent way. This fragment was also present at the cell surface and was detected only at very low levels in BMDCs of the other genotypes. The abundance of Dectin-1 mRNA was not affected by SPPL2a deficiency (Fig. 6E). Therefore, we analyzed the subcellular localization of Dectin-1 by indirect immunofluorescence. In wild-type BMDCs, the bulk of Dectin-1 labeling was observed at the plasma membrane (Fig. 6F). In contrast, Dectin-1 was redistributed to intracellular compartments in SPPL2a−/− BMDCS, where it colocalized with the late endosomal/lysosomal protein LAMP-2 (Fig. 6G). As suggested by the other approaches, the subcellular distribution of Dectin-1 in SPPL2a-CD74 double-deficient BMDCs was similar to that in wild-type cells (Fig. 6F). Furthermore, the intracellular Dectin-1 pool colocalized with CD74 recognized by an N-terminal epitope-specific Ab (Fig. 6H). In the SPPL2a-deficient cells, the CD74 signal was dominated by the endosomally localized CD74 NTF, whereas in wild-type cells, mainly full-length CD74 was detected, which had just reached endosomes prior to degradation or is in transit in the secretory pathway. In particular, the latter is presumably the reason why also wild-type BMDCs showed a certain degree of colocalization between Dectin-1 and CD74 (Fig. 6I), which was, however, significantly increased in the SPPL2a−/− cells. Altogether, these findings suggest that trafficking of Dectin-1 is significantly altered in SPPL2a−/− BMDCs by the CD74 NTF accumulation. In general, the reduced abundance and availability of the receptor at the plasma membrane may at least contribute to the observed hyporesponsiveness of this pathway with regard to IL-10 induction and secretion.
Induction of IFN-β by TLR4 is impaired
Changes observed in the Dectin-1 pathway at the level of the receptor prompted us to analyze TLR4 and, based on its role in mycobacterial recognition, also TLR2 expression (30–32). mRNAs for both receptors showed significant but minor changes in SPPL2a−/− BMDCs (Fig. 7A, 7B). However, total TLR2 and TLR4 protein levels were not changed in these cells (Fig. 7C, 7D). Furthermore, the accessibility of both receptors at the cell surface as determined by FACS was also not altered to a relevant degree in the SPPL2a-deficient cells (Fig. 7E, 7F). This argued against a major redistribution of TLR4 under steady-state conditions to explain the increased sensitivity of this pathway with regard to IL-1β production. TLR4 is known to undergo endocytosis following its activation, which can result in lysosomal degradation of the receptor and termination of its signaling (33). Furthermore, delivery to endosomes is critical for its ability to activate the TRAM–TRIF pathway leading to the induction of type I IFNs. Based on this notion, we tested if IFN-β production differed between SPPL2a−/− and wild-type BMDCs. Stimulation with HKMT resulted in a significantly reduced secretion of IFN-β by SPPL2a−/− BMDCs as compared with wild-type cells (Fig. 7G). Surprisingly, also CD74-deficient BMDCs showed a decreased IFN-β production, which was, however, much less pronounced than in the SPPL2a-deficient cells. Ablation of CD74 in SPPL2a−/− cells increased the IFN-β secretion to the level of the CD74−/− BMDCs but did not fully restore it to the wild-type level. This identifies a role for the CD74 NTF in suppressing the IFN-β response in SPPL2a−/− BMDCs but also points to a yet-undefined mechanism by which CD74 supports this process in wild-type cells. Because IFN-β concentrations in media from BCG coculture experiments were close to the detection limit of the available ELISA, we quantified induction of mRNA expression by qRT-PCR (Fig. 7H). Expression levels of IFN-β were significantly lower in SPPL2a-deficient BMDCs cocultured with BCG when compared with wild-type cells. The IFN-β upregulation in CD74 single-deficient and SPPL2a−/− CD74−/− BMDCs was intermediate between that of wild-type and SPPL2a−/− BMDCs, leading to similar conclusions regarding the role of CD74 and its NTF as obtained from the stimulation with HKMT. Finally, we analyzed IFN-β production upon stimulation with LPS. Also following specific activation of TLR4, BMDCs lacking SPPL2a produced significantly less IFN-β than wild-type controls, thereby pheno-copying the effects seen with mycobacterial stimulation (Fig. 7I). Interestingly, the LPS-triggered IFN-β production was fully reconstituted in SPPL2a-CD74 double-deficient BMDCs, being indistinguishable from that of wild-type cells. Altogether, these findings indicate that the hyperresponsiveness of the TLR4 pathway, as described for IL-1β production by SPPL2a−/− BMDCs, goes along with an impaired switch toward induction of IFN-β, which is secreted at reduced levels. To identify the underlying mechanism, we compared TLR4 signal transduction in wild-type and SPPL2a−/− BMDCs (Fig. 7J, 7K). Activation of p38 MAPK, which is considered to be initiated at the plasma membrane following TLR4 activation, was similar in both genotypes. The already mentioned TRAM–TRIF pathway triggered by internalized TLR4 from endosomes leads to the activation of the transcription factor IRF3 (34). There were no statistically significant differences with regard to LPS-induced IRF3 activation between wild-type and SPPL2a−/− BMDCs. However, we repeatedly observed a subtle tendency to slightly delayed kinetics of IRF3 activation, reaching the levels of wild-type cells after 60 min. Although this may indicate a slightly delayed delivery of activated TLR4 to endosomes, this effect lacked statistical significance and was rather minor in light of the alterations of cytokine secretion in SPPL2a−/− BMDCs. Altogether, we conclude that the receptor-proximal TLR4 signaling is intact in SPPL2a−/− BMDCs (Fig. 8).
Our results confirm the previously reported depletion of cDC2 in spleens of SPPL2a mutant and knockout mice (6, 9) and demonstrate that this phenotype is also present in bone marrow and lymph nodes and to a lesser extent in thymus. Importantly, we could prove that the loss of cDC2 is dependent on CD74 NTF accumulation because it was rescued in SPPL2a-CD74 double-deficient mice. In lymph nodes, particularly migratory DCs, including cDC1 cells, which were otherwise rather spared, were depleted. This demonstrates that the phenotype is not entirely subset selective and that, at least in the periphery, also cDC1 can be depleted. CD74 has been shown to slow down migration of DCs as indicated by faster migration of CD74−/− DCs (35) and reduced migration in DCs that accumulate CD74 NTFs because of deficiency of cathepsin S (CatS), a protease involved in processing the CD74 luminal domain (3). We recently compared CD74 NTF accumulation and cellular phenotypes in CatS−/− and SPPL2a−/− B cells and found both to be significantly less pronounced in CatS-deficient cells (8). The observed CD74-dependent depletion of migratory DCs in lymph nodes of SPPL2a−/− mice is a strong indication that also loss of SPPL2a impairs DC migration, very likely involving the same pathways affected in CatS-deficient cells (35).
The molecular mechanisms underlying the DC depletion and the more-severe impairment of cDC2 currently remain elusive. Treatment of human PBMCs with an SPPL2a inhibitor suggested a higher CD74 turnover in cDC2 compared with other subsets (9). Several examples of knockout mice with predominant or selective depletion of cDC2 cells have been described (15, 36). Many of the ablated genes are transcription factors such as IRF4 and Klf4. However, how these might be influenced by the membrane-bound CD74 NTF is not obvious. In this regard, an impact of this fragment on receptor proteins (e.g., on their subcellular trafficking) may be more likely. Deletion of the membrane proteins Notch2 (37), lymphotoxin β receptor (LTβR) (38, 39), or SIRPα (CD172a) (40) leads to predominant cDC2 depletion phenotypes in mice with certain similarities to the phenotype of SPPL2a−/− mice. Therefore, these proteins and the associated signaling pathways may be candidates to be analyzed in SPPL2a−/− DCs. It may seem surprising that yield and differentiation of BMDCs in vitro was not negatively affected by SPPL2a deficiency. However, similar discrepancies between normal BMDC generation in vitro and a major DC differentiation phenotype in vivo have also been described for LTβR- and SIRPα-deficient mice (40, 41). Differentiation of DCs, especially with regard to the specialization into subsets, is a complex, highly orchestrated process (36, 42). Therefore, pathways affected by CD74 in vivo may either not be relevant for BMDCs or may be activated in excess in vitro (e.g., by exogenously added GM-CSF) so that any inhibitory effect of CD74 is overrun.
In light of the major accumulation of the CD74 NTF and its effects on B cells (4), the functionality of SPPL2a−/− BMDCs with regard to cytokine secretion is surprisingly well preserved. The observed changes are rather selectively affecting individual PRR ligand–cytokine combinations. Importantly, these changes nonetheless significantly affect the responses of SPPL2a−/− BMDCs to mycobacteria, both heat killed or viable, raising the question of to what extent this may contribute to the MSMD phenotype of SPPL2a-deficient mice and patients. In this context, IL-12 family cytokines are critical for the induction of TH1 polarized immune responses (12, 13, 43, 44). We did not observe a major reduction in the secretion of IL-12 family cytokines. SPPL2a and its homolog SPPL2b have been previously linked with IL-12 secretion via releasing the intracellular domain of TNF (45). Our data from SPPL2a-single-deficient BMDCs do not readily support this model. Of course, in vivo, a reduction of IL-12–producing cDCs could account for an overall reduced availability of this cytokine, even if the production capacity by the individual cells is preserved. However, restimulation of lung homogenates from BCG-infected SPPL2a−/− mice led to significantly enhanced production of IL-12p40 and IFN-γ as compared with controls (9), which indicates that also in vivo, the IL-12–IFN-γ axis is not globally disturbed.
The major changes we observed in mycobacteria-stimulated BMDCs were an upregulation of IL-1β and a downregulation of IFN-β and IL-10 secretion (Fig. 8). IL-10 is produced by a variety of immune cells, including DCs, and has a broad anti-inflammatory activity (46). In mycobacterial infections, IL-10 can reduce antimycobacterial immune responses at various stages and thereby indirectly support pathogen growth but also limit immunopathology (47, 48). This notion is corroborated by studies in IL-10−/− mice on a C57BL/6J background subjected to a low-dose M. tuberculosis aerosol infection, which developed excessive pulmonary inflammation and disease progression after 6 mo despite enhanced TH1 immunity (49). In Card9−/− mice, which are compromised in fighting mycobacterial infections despite intact T cell responses, IL-10 was identified as part of a critical anti-inflammatory feedback loop preventing overshooting, nonproductive responses (50). In line with this report, IL-10 levels were positively correlating with the induction of local protective immunity by mucosal BCG vaccination in rhesus macaques (51).
A similarly dual role on the outcome of mycobacterial infections has been reported for type I IFNs like IFN-β (52, 53), which was produced in lower amounts by SPPL2a−/− BMDCs. Many studies showed that type I IFNs can exacerbate mycobacterial infections. However, in certain settings they can also have protective functions (52). Type I IFNs support the production of IL-10 and instead reduce the generation of IL-12 family members but also of IL-1β (52), a central proinflammatory cytokine. Host responses to M. tuberculosis were found to be severely compromised in IL-1β–deficient mice (54, 55). Furthermore, the IL-1 pathway was reported to be dispensable for controlling infection with the attenuated BCG strain (55). Despite being essential for the initiation of immune responses against mycobacteria, IL-1β has also been linked to immunopathology. In a mouse model with defective reactive oxygen species production, increased IL-1β production was responsible for neutrophil infiltration and lung pathology, which was significantly ameliorated by blocking this cytokine (56, 57). Several examples suggest that uncontrolled neutrophil recruitment, for which IL-1β is a key mediator, detrimentally leads to tuberculosis disease exacerbation (58, 59). Neutrophils are critical mediators of lung damage (50, 60–63) and at the same time fail to control mycobacterial growth (64). Thus, an overshooting innate immune response dominated by neutrophil recruitment can significantly contribute to a failed control of mycobacterial infections. Importantly, SPPL2a-deficient mice exhibited a significantly increased influx of neutrophils into the lung (9) following aerosol infection with M. tuberculosis, which indicates an imbalanced immune response. In addition to impaired T cell priming (9), this may also be an explanation for the MSMD phenotype associated with SPPL2a deficiency. The overactivation of neutrophils would be consistent with an increased activity of IL-1β (56, 57) as well as reduced IL-10 (50) and IFN-β, which can act as a negative regulator of IL-1β secretion (52, 54) as observed in this study in SPPL2a−/− BMDCs. DCs act as pathogen sensors and have a key role in setting the quality of immune responses. Therefore, the disturbed balance between anti- and proinflammatory cytokines as secreted by SPPL2a−/− versus wild-type BMDCs upon mycobacterial stimulation may be part of the failure to set up a protective immune response to M. tuberculosis in vivo.
Mycobacteria can activate DCs via different PRRs, including TLR1/2, TLR4, TLR9, Dectin-1, Mincle, and Dectin-2, as well as cytosolic sensors of the NOD and RIG-I–like family (26, 27). The spectrum of PRRs engaged by mycobacterial ligands can differ between strains of mycobacteria involved. Furthermore, pathways triggered by different receptors can be interconnected and act in synergy or modulatory to induce cytokine secretion as described for Dectin-1 and TLR2 (28, 30, 65). Using specific ligands, we present evidence that Dectin-1 and TLR4 pathways are disturbed in the absence of SPPL2a, altering cytokine secretion patterns upon mycobacterial stimulation. Different studies described a role of these receptors in mycobacterial recognition (29–32, 65–70). However, mice deficient in either one of these receptors in some models are still able to control the infection (71–73), which indicates a certain redundancy. It seems likely that the impairment of the Dectin-1 and TLR4 pathways is at least in part responsible for the altered cytokine responses to mycobacteria. However, we cannot exclude that additional PRRs and/or their associated signaling pathways contribute to the phenotype of SPPL2a−/− BMDCs.
All changes we found in SPPL2a−/− BMDCs were completely or at least in part dependent on the CD74 NTF accumulation. The mechanism of how the CD74 NTF impacts Dectin-1 and TLR4 signaling does not seem to be identical for both PRRs. Reduced Dectin-1 surface localization is reminiscent of the altered BCR trafficking in SPPL2a−/− B cells (7). This could indicate enhanced endocytosis as in the case of the BCR (7), but also transport of Dectin-1 to the plasma membrane may be affected. The capability of CD74 or its NTF to influence membrane trafficking in the endocytic system is well documented (3), although still not understood at the molecular level. Our results suggest that the induction of IL-10 is more sensitive to the reduced Dectin-1 surface levels than other cytokines. Dectin-1 is also a phagocytic receptor and by this means facilitates presentation of fungal Ags (74). Therefore, the reduced Dectin-1 surface levels may have implications for presentation of mycobacterial Ags based on reduced capacity to internalize these pathogens. However, reduction of Dectin-1, in this case, at the level of mRNA expression was also found in macrophages from mice deficient for the cytokine macrophage migration inhibitory factor (MIF) (75). This observation was suggested to be responsible for the susceptibility to mycobacterial infections of MIF−/− mice as well as of humans with low MIF expression (75). Conspicuously, CD74 is one of the receptors for this cytokine (76). However, the impact of SPPL2a-mediated cleavage of CD74 on MIF signal transduction is not known yet.
In contrast to Dectin-1, TLR4 surface levels were not significantly changed in SPPL2a−/− BMDCs. It may seem paradoxical that in these cells TLR4-mediated secretion of IL-1β and TNF was increased but that of IFN-β decreased. TLR4 has different signaling modes depending on whether it signals from the cell surface or from endosomal compartments, where it is delivered after engagement of its ligand at the cell surface (34, 77). Production of IFN-β is triggered primarily by the endosomal, MyD88-independent TLR signaling via the TRAM–TRIF pathway (34, 77, 78). We speculated that a delay in switching between these two signaling modes could explain the increase of IL-1β as well the decrease of IFN-β. In general, internalization of TLR4 followed by lysosomal degradation is critical for termination of TLR4 signaling (33). The observed tendency toward a slightly retarded activation of IRF3 could indicate a slower trafficking of TLR4 through the endosomal system in absence of SPPL2a. This would be consistent with previous findings in SPPL2a-deficient B cells, where we have observed a delay in the degradation of fluid phase endocytic cargo (7, 8). Furthermore, CD74 has been reported to delay endosomal maturation (79, 80). However, neither the overall MAPK nor IRF3 response was significantly different, which is why we conclude that proximal events following activation of TLR4 by LPS and the transition between the two signaling modes are not affected in a relevant way in the SPPL2a-deficient cells. Thus, in contrast to Dectin-1, disturbed trafficking of TLR4 does not seem to represent the underlying mechanisms for the altered cytokine responses. Instead, our results strongly suggest that these are caused by mechanisms acting further downstream at the level of regulating cytokine transcription. Thus, the mechanisms of how the accumulating CD74 NTF influences DC functionality and in particular antimycobacterial cytokine responses seem to be diverse and act at different levels. As discussed above in the context of differentiation, it does not seem obvious how the membrane-bound CD74 can influence IL-1β and IFN-β transcription. An interesting observation in this context was that the IFN-β production following stimulation with HKMT, but not with LPS, was negatively affected in CD74 single-deficient BMDCs. This indicates that IFN-β induction by HKMT does not exclusively depend on TLR4 and that other receptors are involved here. Why the presence of CD74 is important to support IFN-β production in HKMT-stimulated cells currently remains elusive. It could be connected to its roles as receptor for the cytokine macrophage MIF (3, 76) or as targeting adapter for delivering MHCII to endosomes because endosomal MHCII can influence PRR signaling and cytokine responses (81). Altogether, further mechanistic insights are needed to understand the impact of SPPL2a deficiency as well as of the accumulating CD74 NTF on DC differentiation and PRR responses to mycobacteria.
We thank Daniela Tenfelde, Sebastian Held, Marlies Rusch, and Petra Peche for excellent technical assistance. Acquisition of microscopic images was performed within the Core Facility Cellular Imaging of Technische Universität Dresden. Furthermore, we are grateful to Dr. Britta Schilling for help with administrative issues of the BCG coculture experiments as well as Suzan Leccese for help with conducting these experiments.
This work was supported by the Deutsche Forschungsgemeinschaft (SFB877 [project B7], SCHR 1284/1-1, and SCHR 1284/1-2 [to B.S.]; CRC1181-TPA7 [to D.D.]; Ex-Cluster, Inflammation at Interfaces [to M.A. and U.E.S. (Lysosomal Disorders and Bacteria-Induced Inflammation Subproject 3)]) and Interdisciplinary Centre for Clinical Research Intramural Grants IZKF A85 (to D.D.) and IZKF A87 to C.H.K.L.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
bone marrow–derived DC
conventional dendritic cell
heat-killed M. tuberculosis
MHC class II
migration inhibitory factor
multiplicity of infection
Mendelian susceptibility to mycobacterial disease
pattern recognition receptor
quantitative real-time PCR
RPMI 1640 containing 2% FCS
signal regulatory protein α
signal peptide peptidase–like 2a.
The authors have no financial conflicts of interest.