ABSTRACT
Accumulation of Tropheryma whipplei-stuffed macrophages in the duodenum, impaired T. whipplei-specific Th1 responses, and weak secretion of interleukin-12 (IL-12) are hallmarks of classical Whipple's disease (CWD). This study addresses dendritic cell (DC) functionality during CWD. We documented composition, distribution, and functionality of DC ex vivo or after in vitro maturation by fluorescence-activated cell sorting (FACS) and by immunohistochemistry in situ. A decrease in peripheral DC of untreated CWD patients compared to healthy donors was due to reduced CD11chigh myeloid DC (M-DC). Decreased maturation markers CD83, CD86, and CCR7, as well as low IL-12 production in response to stimulation, disclosed an immature M-DC phenotype. In vitro-generated monocyte-derived DC from CWD patients showed normal maturation and T cell-stimulatory capacity under proinflammatory conditions but produced less IL-12 and failed to activate T. whipplei-specific Th1 cells. In duodenal and lymphoid tissues, T. whipplei was found within immature DC-SIGN+ DC. DC and proliferating lymphocytes were reduced in lymph nodes of CWD patients compared to levels in controls. Our results indicate that dysfunctional IL-12 production by DC provides suboptimal conditions for priming of T. whipplei-specific T cells during CWD and that immature DC carrying T. whipplei contribute to the dissemination of the bacterium.
INTRODUCTION
Classical Whipple's disease (CWD) is a chronic multisystemic infection with Tropheryma whipplei (1). Self-limiting asymptomatic or moderate infections with this agent are common (2, 3) and result in a protective humoral and cellular immunity in most individuals (4, 5). In contrast, subtle immune defects that occur only rarely seem to predispose for CWD: an HLA association (6) and immunological aberrations point at problems in the axis of antigen processing/presentation/T cell activation (4, 7–12). These differences include low peripheral interleukin-12 (IL-12) and gamma interferon (IFN-γ) (4, 10–12), enhanced anti-inflammatory cytokines and regulatory T cells (Treg) (7, 12), and the absence of a T. whipplei-specific Th1 response (4). T. whipplei itself shapes the immune response, creating an anti-inflammatory milieu (7, 12, 13).
Dendritic cells (DC) are critical for the initiation of protective Th1 responses to ward off pathogenic microorganisms (14). IL-12 may be produced by DC and is the key cytokine in this process (15). Various DC subtypes in humans, such as myeloid DC (M-DC), plasmacytoid DC (P-DC), and the Langerhans cells, play an important role in the priming of T cell responses (16). In addition, 6-sulfo LacNAc DCs (slanDCs) were described as a source of IL-12 (17). Impaired T cell-stimulatory capacity has been demonstrated for DC infected with Mycobacterium tuberculosis (18), and alterations in the distribution and functionality of DC subsets have been described for various infections and chronic inflammatory diseases (19–28).
Addressing the so-far largely unknown role of DC in the pathogenesis of CWD, we examined the DC population in CWD patients in comparison to healthy control subjects in vivo in terms of distribution within different tissues and regarding its composition, phenotype, and response to pathogenic signals within the peripheral blood. In addition, intrinsic functional aberrations of DC and their consequences for the interaction with T. whipplei were investigated using monocyte-derived DC (Mo-DC) as the most appropriate in vitro model to study the functionality of DC.
MATERIALS AND METHODS
Patients and control subjects.Samples from 91 patients with CWD (confirmed by at least two tests [1]) and 99 control subjects without clinical signs of T. whipplei infection were studied (Table 1). Gastrointestinal symptoms were present in 85 CWD patients; three patients had isolated neurological symptoms, and three had only articular manifestation. The majority of patients were treated for 2 weeks with ceftriaxone, followed by either 12 or 3 months of trimethoprim-sulfamethoxazole; five patients received alternative treatments. Three patients were treated additionally with IFN-γ. Total remission was achieved in 84 patients, four died, and three had persisting problems.
Investigated samples from CWD patients and control subjects
Blood was collected in heparinized tubes (Vacutainer; BD Biosciences, Heidelberg, Germany) and processed within 24 h. Due to the high quantity of blood required and the need to schedule the experiments, Mo-DC were prepared only from treated CWD patients at the time of the projected clinical checkup. Tissue specimens (Table 1) were fixed in 4% paraformaldehyde (PFA; Sigma-Aldrich, Taufkirchen, Germany). Lymph node specimens were collected for differential diagnosis of 16 CWD patients with adenopathy (seven mesenteric, three axillary, three cervical, two inguinal, and one bronchial lymph node), from 11 subjects without clinical findings (eight mesenteric and three parotideal LN), eight subjects with tuberculosis (four cervical, two mediastinal, and one each of axillary and mesenteric lymph node), and 6 subjects with sarcoidosis (three cervical and one each of bronchial, mediastinal, and pulmonary lymph node).
Experiments were conducted in accordance with the Declaration of Helsinki. The study was approved by the Clinical Ethics Committee of the Charité (approval no. EA4-01-122-10), and all adult subjects provided written informed consent.
Bacterial strains and preparations.T. whipplei strain Twist Marseille (CNCM I-2202) was cultured in axenic medium (29) and used as viable bacteria or as a heat-killed lysate (4).
Evaluation of DC activation in whole-blood specimens.Fresh heparinized blood (500 μl) was incubated with 10 μg/ml lipopolysaccharide (LPS; Sigma-Aldrich) or T. whipplei lysate (107 bacteria/ml) or without any stimulus (negative control) for 6 h at 37°C in a humidified 5% CO2 atmosphere. Some samples were preincubated for 1 h with 2.5 μg/ml rat anti-IL-10 (JES3-9D7; eBioscience, Frankfurt, Germany), 0.5 μg/ml mouse anti-transforming growth factor beta 1, 2, 3 (TGFβ1,2,3; 1D11; R&D Systems, Wiesbaden, Germany), or the corresponding isotype controls (rat IgG1 was from eBioscience, and mouse IgG1 was from R&D Systems) in equivalent concentrations. Other samples were supplemented with 10 ng/ml IFN-γ (R&D Systems) at the start of the 6-h incubation period. Surface markers were stained directly in whole blood. For the detection of IL-12, IL-23, and IL-10, brefeldin A (10 μg/ml; Sigma-Aldrich) was added for the last 3 h of stimulation, and intracellular staining and flow cytometric analysis was performed as described previously (12).
Isolation of monocytes and T lymphocytes from peripheral blood.Peripheral blood mononuclear cells (PBMC) were prepared as described previously (12) from healthy controls and treated CWD patients. Monocytes were isolated from PBMC by using CD14 MicroBeads (1:10) according to the manufacturer's protocol (Miltenyi Biotec, Bergisch Gladbach, Germany), with a yield of 0.4 × 107 to 6.1 × 107 CD14+ cells from 108 PBMC and a purity of 92% ± 3% CD14+ cells. T cells were obtained from the CD14− cell population by the depletion of HLA-DR+ cells using HLA-DR MicroBeads (1:10; Miltenyi Biotec), with a yield of 1.5 × 107 to 8.5 × 107 CD14− HLA-DR− cells from 108 PBMC and <1% contaminating HLA-DR+ cells. These T cell preparations were stored at −80°C.
Mo-DC.Freshly isolated monocytes (1 × 106 cells/ml) in R10 medium consisting of RPMI 1640 medium with GlutaMAX (Life Technologies, Darmstadt, Germany), 10% fetal calf serum (FCS), 25 mM HEPES (PAA Laboratories), 100 U/ml penicillin, and 100 μg/ml streptomycin (both from Biochrom, Berlin, Germany) were differentiated into immature Mo-DC in the presence of 103 U/ml granulocyte-macrophage colony-stimulating factor (GM-CSF) (Berlex Laboratories, Richmond, VA) and 2.5 ng/ml IL-4 (R&D Systems) within 36 h according to a modified FastDC protocol published by Dauer et al. (30) that was demonstrated to be equivalent to longer protocols (data not shown). To induce maturation, 10 ng/ml tumor necrosis factor alpha (TNF-α), 20 ng/ml IL-1β, and 3.3 ng/ml IL-6 (all from R&D Systems) and 1 μM prostaglandin E2 (PGE2; Sigma-Aldrich) were added for an additional 24 h. Mo-DC were loaded with either T. whipplei lysate (107 bacteria/ml) or 1.5 μg/ml cytomegalovirus antigen (CMV Ag; AD169; Dunn Labortechnik, Asbach, Germany) added to the cells together with the maturation stimuli. DC supernatants were stored at −80°C.
Assessment of antigen-presenting and T cell-stimulatory capacity of Mo-DC.CFSE-labeled (12) T cells (2 × 105) were mixed with autologous mature antigen-loaded Mo-DC at ratios of 1:80, 1:40, 1:20, 1:10, and 1:5 in a final volume of 200 μl/well in 96-well flat-bottom plates (Nunc, Wiesbaden, Germany) for 4 days in R10 medium. Preparations with unloaded Mo-DC served as negative controls, and preparations with unloaded Mo-DC and 2 μg/ml of staphylococcal enterotoxin B (SEB; Sigma-Aldrich) served as positive controls. Brefeldin A (10 μg/ml) was added for the last 4 h before analysis. Values for the negative control were subtracted from values for the corresponding antigen-specific T-cell activation, and values below the background level were defined as 0.
Flow-cytometric analysis.The following fluorochrome-conjugated monoclonal antibodies (Abs) were used: Lin-1 fluorescein isothiocyanate (FITC) cocktail, mouse anti-HLA-DR peridinin chlorophyll protein (PerCP) (L243 G46-6), CD11c (S-HCL-3), CD123 (9F5), CD83 (HB15e), CD86 [2331(FUN-1)], CD19 (HIB19), CD14 (MϕP9), DC-SIGN (DCN46), CD40 (5C3), CCR2 (48607), CCR5 (3A9), IL-12 (C11.5.14), IL-10 (JES3-19F1), CD3 (SK7), IFN-γ (B27), and CCR7 (3D12) from BD Biosciences; PD-L1 (MIH1), IL-23p19 (23dcdp), and TLR4 (HTA125) from eBioscience; CD141 (BDCA-3; AD5-14H12), CD1c (BDCA-1; AD5-8E7), and M-DC8 (DD1) from Miltenyi Biotec; CD4 (MT310; DakoCytomation, Glostrup, Denmark); and CD80 (MEM-233; EXBIO, Prague, Czech Republic). Surface staining of DC from fresh whole blood and cultured Mo-DC was performed as described previously (4, 12), eventually followed by fixation and intracellular staining (12). Data were acquired on a FACSCalibur with CellQuest (BD Biosciences) and analyzed using the FlowJo software (TreeStar, Ashland, OR).
Analysis of cytokine secretion.Cytokines were quantified in supernatants of Mo-DC cultures by using the IL-12p70 enzyme-linked immunosorbent assay (ELISA) (U-CyTech, Utrecht, Netherlands), IL-12p40 OptEIA (BD Biosciences), and IL-23 and IL-10 Ready-SET-Go! (both from eBioscience) according to the manufacturers' protocols.
Evaluation of T. whipplei uptake by Mo-DC.Immature Mo-DC (2 × 106) were incubated with viable 5 × 107 T. whipplei in 1 ml R10 medium without antibiotics for 30 min at 37°C, washed, and resuspended in phosphate-buffered saline (PBS). Cytospins were prepared from 5 × 105 Mo-DC with a Cytospin 2 Cytofuge (Thermo Shandon, Frankfurt, Germany) at 100 × g for 7 min, air dried, fixed in 1% PFA solution for 60 min at 4°C, and stored at −80°C.
Immunohistochemistry and immunofluorescence staining.Immunostaining on paraffin sections was performed as previously described (7, 8, 31). We identified DC populations in tissues by the expression of DC-SIGN (32) and S-100 (31). The following primary Abs were used: mouse anti-DC-SIGN (111H2.02; Dendritics, Lyon, France), Ki-67 (MIB-1; Dako), DC-LAMP (104.G4; Immunotech, Marseille, France), HLA-DR (TAL1B5; Dako), rat anti-FOXP3 (PCH101; eBioscience), and rabbit anti-S-100 (Dako). Rabbit anti-T. whipplei was a kind gift from D. Raoult (33). For double stainings, T. whipplei first was visualized using biotinylated anti-rabbit, streptavidin horseradish peroxidase (HRP; Sigma-Aldrich) and 3,3′-diaminobenzidine (DAB; Sigma-Aldrich). A streptavidin/biotin block (Vector Laboratories, Burlingame, CA) and a serum-free protein block (Dako) were applied for 15 and 10 min, respectively, before detection of DC-SIGN, S-100, Ki-67, or FOXP3. Nuclei were counterstained with Meyer's hematoxylin (Dako), and the cytoskeleton (F-actin) was visualized using Alexa 647-labeled phalloidin (Life Technologies, Glasgow, Scotland). Positive cells were quantified as the averaged numbers of 10 to 20 high-powered fields (1 hpf; 0.237 mm2) in a blinded manner.
For immunofluorescence labeling, specimens were incubated with mouse anti-DC-SIGN (111H2.02; Dendritics), followed by biotinylated anti-mouse and indocarbocyanine (Cy3)-conjugated streptavidin (Dianova, Hamburg, Germany), washed in Tris-buffered saline (TBS), and incubated with rabbit anti-T. whipplei followed by Alexa Fluor 488-conjugated anti-rabbit (Life Technologies). Cytospins of Mo-DC were pretreated for 20 min with PBS containing 10% donkey serum (tebu-bio, Offenbach, Germany) and 0.1% saponin, and all Abs were applied in PBS containing 3% donkey serum and 0.1% saponin. DC-SIGN, DC-LAMP, and HLA-DR were visualized using Cy3-conjugated anti-mouse (Dianova). Cytospins were thoroughly rinsed with PBS before T. whipplei was stained as described above. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Roche, Mannheim, Germany), and slides were mounted in Fluoromount-G (Southern Biotech, Birmingham, AL). Images were acquired using a confocal laser scanning microscope (510 META) and processed with the LSM 5 Image Browser software (both from Carl Zeiss, Jena, Germany). Negative controls were performed by omitting the primary Ab or the biotinylated secondary Ab.
Statistical analysis.Quantitative parameters are presented either as individual data points with the median, box plots with the median, both extreme values and the 25% and 75% quartiles, or mean values with standard errors of the means (SEM). Data were analyzed using GraphPad Prism 5 software (GraphPad, La Jolla, CA). Statistical differences between independent groups were calculated by means of the two-tailed Mann-Whitney U test. The two-tailed Wilcoxon signed-rank test was used to compare data of CWD patients before and after antibiotic treatment. Correlation between data sets was evaluated using Spearman's correlation analysis. P values of ≤0.05 were considered significant.
RESULTS
Reduced percentage of DC in peripheral blood of CWD patients.The percentages of total DC and the DC subpopulations M-DC, M-DC1, MDC2, P-DC (Fig. 1A), and slanDC within the mononuclear cell population were determined in blood samples from CWD patients before and after antibiotic treatment and compared to healthy control subjects.
Subpopulations and phenotype of peripheral blood DC from CWD patients. Whole blood and PBMC from CWD patients before and after treatment and healthy control subjects were used to determine the percentages of M-DC and P-DC subsets within the mononuclear cell (MNC) population and to analyze the maturation status within the subpopulation of M-DC by FACS. (A) DC in whole-blood specimens were defined as Lin-1− HLA-DR+ and subdivided into M-DC and P-DC subsets according to the expression of CD11c and CD123; M-DC1 and M-DC2 were determined in PBMC preparations as CD19− CD14− live cells according to the expression of CD1c and CD141. SSC, side scatter; FSC, forward scatter. (B) Percentages of total DC among all MNC in whole blood. (C) Percentages of M-DC and P-DC in whole blood and M-DC1 and M-DC2 among MNC in PBMC preparations. (D) The portion of CD83-expressing cells and the expression level of CD86 and CCR7, represented by the mean fluorescence intensity (MFI), are indicated for the M-DC population. (B to D) Dot plots show all data points, and the median value is indicated as a horizontal line. P < 0.01 (**) and P < 0.05 (*) according to Mann-Whitney U test.
Circulating HLA-DRhigh Lin-1− DC were significantly reduced in the peripheral blood from CWD patients compared to the levels in control subjects (Fig. 1B). The subpopulations of M-DC8+ CD14− slanDC (data not shown) and CD123high P-DC were not altered (Fig. 1C). Thus, the overall decrease in DC of untreated CWD patients was due mainly to significantly decreased percentages of CD11c+ M-DC (Fig. 1C). Percentages of CD1c+ M-DC1 were similar in PBMC from CWD patients and cells from controls. In contrast, the proportion of CD141+ M-DC2 was significantly reduced in untreated CWD patients compared to that in control subjects (Fig. 1C), although the proportion of M-DC2 was low in both healthy controls and CWD patients.
To test if this loss of M-DC was paralleled by differences in maturation, M-DC were stained for markers of maturation and LN homing. As in healthy control subjects, less than 20% of the M-DC subpopulation in CWD expressed CD83 and could be considered mature DC (Fig. 1D). The expression of CD80, CD40, and HLA-DR and the tissue-homing-mediating chemokine receptors CCR2 and CCR5 (data not shown), as well as of the costimulatory molecule CD86 and the chemokine receptor CCR7 (Fig. 1D), did not differ between M-DC within the three groups.
Impaired TLR4-mediated maturation of circulating M-DC from CWD patients.To determine whether circulating immature M-DC in CWD patients mature properly in response to bacterial stimuli, whole blood was incubated with the TLR4 ligand LPS or T. whipplei lysate.
The expression of TLR4 on M-DC did not differ between untreated and treated CWD patients and healthy control subjects (see Fig. S1A in the supplemental material), and LPS enhanced the expression of CD83, CD86, and CCR7 on Lin-1− HLA-DRhigh CD11c+ M-DC from CWD patients and healthy control subjects, while no differences were observed for CCR7 (Fig. 2A). The LPS-induced increase of CD83 and CD86 expression, however, was lower for M-DC from CWD patients than for cells from control subjects (Fig. 2A). Compared to control subjects, the percentage of M-DC that produced IL-12 in response to LPS was significantly reduced in the blood of CWD patients irrespective of their treatment status (Fig. 2B). The percentage of cells producing IL-23p19 upon stimulation with LPS was low in M-DC from control subjects and untreated CWD patients but was significantly increased in M-DC from CWD patients after treatment (Fig. 2B). The percentage of IL-10-producing M-DC was reduced upon LPS treatment in all three groups (Fig. 2B). Pointing to a weak induction of maturation, T. whipplei lysate induced the expression of CD83 and CD86 but failed to enhance the expression of CCR7 and of cytokines, i.e., IL-12, IL-23, or IL-10, in M-DC from CWD patients and control subjects (Fig. 2A and B).
Response of M-DC to bacterial stimuli in blood from CWD patients. (A and B) Whole-blood specimens from untreated CWD (n = 8 [A], 14 [B, left], 4 [B, middle], and 9 [B, right]) and treated CWD patients (n = 10 [A], 44 [B, left], 12 [B, middle], and 23 [B, right]), as well as healthy control subjects (n = 16 [A], 27 [B, left], 11 [B, middle], and 23 [B, right]), were incubated with 10 μg/ml LPS or 107 CFU/ml T. whipplei lysate or without stimulus. The percentages of CD83+ cells, the expression of CD86 and CCR7 (A), and the proportions of cells expressing IL-12, IL-23p19, or IL-10 (B) were determined within the Lin-1− HLA-DR+ CD11c+ M-DC population by flow cytometry (gated as demonstrated for Fig. 1A). (C) Whole-blood specimens from treated CWD patients (n = 4) or healthy control subjects (n = 5) were preincubated with 2.5 μg/ml neutralizing anti-IL-10 or control rat IgG1 for 1 h, followed by 6 h in the presence of 10 μg/ml LPS or without stimulus. IL-12-producing Lin-1− HLA-DR+ CD11c+ M-DC were assessed by flow cytometry (gated as demonstrated for Fig. 1A). Box plots show median, 25%, and 75% quartiles and both extreme values. Statistical differences were calculated using Mann-Whitney U test (***, P < 0.001; **, P < 0.01; *, P < 0.05) and Wilcoxon signed-rank test for paired data (black indicates with stimulus versus without stimulus; gray indicates significance at P < 0.001 [###], P < 0.01 [##], and P < 0.05 [#]).
Blocking of IL-10 (Fig. 2C) or transforming growth factor beta (TGF-β) (see Fig. S1B in the supplemental material) that both are enhanced in the serum and duodenum of CWD patients (7, 12) did not enhance the percentage of IL-12-expressing M-DC in the presence of LPS in whole blood from CWD patients. In samples from healthy controls, however, the percentage of IL-12-producing M-DC increased following blockade of IL-10 (Fig. 2C). Although adding IFN-γ increased the portion of IL-12-expressing M-DC in the presence of LPS in the blood of treated CWD patients, the percentage of IL-12-expressing M-DC in this group did not reach the level of that observed in control subjects (see Fig. S1C in the supplemental material).
Altered cytokine production of in vitro-generated Mo-DC from CWD patients.We addressed intrinsic defects in DC of myeloid origin from CWD patients by studying the effect of T. whipplei lysate on DC of Mo-DC as an in vitro model.
Immature and mature Mo-DC from treated CWD patients and control subjects showed similar percentages of CD83+ mature cells and a comparable expression of the costimulatory molecules CD86, CD80, and CD40, of the inhibitory ligand PD-L1, and of HLA-DR, CCR7, and CD11c (Fig. 3A and data not shown). About 70% of Mo-DC without maturation stimulus and 40% of Mo-DC with maturation stimulus from CWD patients (immature Mo-DC median fluorescence intensity [MFI], 70; range, 35 to 161; mature Mo-DC MFI, 40; range, 14 to 100) and healthy control subjects (immature Mo-DC MFI, 69; range, 41 to 139; mature Mo-DC MFI, 44; range, 11 to 90) expressed DC-SIGN, and less than 1% of immature or mature Mo-DC were CD14+ (data not shown). The supplementing of T. whipplei lysate augmented the expression of CD83 and CD86 of Mo-DC from control subjects compared to mature Mo-DC (Fig. 3A). In addition, the expression of CD83 and PD-L1 was significantly enhanced in Mo-DC from control subjects compared to that from CWD patients in the presence of T. whipplei lysate and maturation stimuli.
Surface (phenotype) marker expression, cytokine production, and T cell-stimulatory potential of Mo-DC from CWD patients. Monocytes (106 cells/ml) from treated CWD patients and healthy control subjects (A, n = 16; B, n = 5 [left], 10 [middle], and 16 [right]; C, n = 5 [left] and 8 [right]) were differentiated into immature Mo-DC, and maturation was induced by TNF-α, IL-1β, PGE2, and IL-6. Some samples were incubated additionally with heat-inactivated T. whipplei (A, B, and C, right) or CMV antigens (C, left) during the maturation period. (A) The percentage of CD83-expressing Mo-DC and the expression of CD86, HLA-DR, and PD-L1 (from left to right) were analyzed by FACS (a gate was set on the whole population of Mo-DC). (B) The concentrations of IL-12p70 (left), IL-12p40 (middle), and IL-23 (right) were determined in culture supernatants by ELISA. (C) Antigen-stimulated or unstimulated Mo-DC were cocultured with autologous CFSE-labeled T cells to assess the portion of proliferating (CFSElow) and IFN-γ-producing cells within the CD4+ CD3+ T cell population. To identify specific effects, the T cell activation induced by unstimulated Mo-DC was subtracted. (A and B) Box plots show median, 25%, and 75% quartiles and both extreme values. (C) Mean values and SEM are indicated. Statistical differences were calculated using Mann-Whitney U test (*, P < 0.05) and Wilcoxon signed-rank test for paired data (black symbols indicate P < 0.001 [###], P < 0.01 [##], and P < 0.05 [#] for immature versus mature Mo-DC, and gray symbols indicate mature versus T. whipplei-stimulated Mo-DC).
Matured Mo-DC from CWD patients did not respond to the pathogen (Fig. 3A). Upon maturation in the presence or absence of T. whipplei, IL-12p70 secretion of Mo-DC from CWD patients was below the detection level of the ELISA used, whereas Mo-DC from control subjects secreted significantly more, but still small, amounts of IL-12p70 (Fig. 3B). However, the secretion of IL-12p40 and IL-23 did not differ between mature Mo-DC from CWD patients and control subjects. The simultaneous addition of T. whipplei lysate to the maturation stimulus augmented the secretion of IL-23 by Mo-DC from control subjects and significantly reduced the secretion of IL-12p40 by Mo-DC from CWD patients (Fig. 3B).
To study the overall potential of Mo-DC to induce recall T-helper cell responses, Mo-DC loaded with CMV antigens or T. whipplei lysate were cocultured with autologous CFSE-labeled T cells. Although Mo-DC from CWD patients were as potent as stimulators of CMV-specific recall responses as Mo-DC from control subjects, they were not able to induce equivalent T. whipplei-specific CD4+ T cell responses in CWD patients and healthy controls (Fig. 3C).
DC are present in tissues and ingest T. whipplei in vitro and in situ.As DC play an important role in the uptake and presentation of antigens at mucosal surfaces, we studied DC-SIGN+ immature and S-100+ mature DC in the duodenal mucosa and other tissues infected with T. whipplei. We also investigated the uptake of T. whipplei by DC in duodenal biopsy specimens in situ and in in vitro-generated immature Mo-DC.
Within the duodenal lamina propria, the numbers and distribution of DC-SIGN+ and S-100+ DC did not differ between control subjects and CWD patients before and after treatment (Fig. 4A). T. whipplei bacteria were found preferentially within DC-SIGN+ cells in the duodenal mucosa of untreated CWD patients (Fig. 4B).
DC populations in the duodenal mucosa from CWD patients and uptake of T. whipplei by DC in situ and in vitro. (A) DC were detected immunohistochemically in duodenal biopsy specimens. Representative images show DC-SIGN- or S-100-expressing cells (in red; indicated by black arrowheads). (B and C) Immunofluorescence staining to analyze the uptake of T. whipplei by DC in duodenal biopsy specimens from untreated CWD patients (B) and in cytospins of Mo-DC from a treated CWD patient (C). (B) DC-SIGN+ cell (red) with ingested T. whipplei (green). Nuclei were stained with DAPI (blue), and the cytoskeleton (F-actin; gray) was visualized using Alexa 647-labeled phalloidin. (C) Immature Mo-DC from a CWD patient were incubated with live T. whipplei and stained for the bacterium (green) in DC-SIGN+ cells (red; left) or DC-LAMP for intracellular compartments (red; middle and right). Cell borders were drawn as white lines from the corresponding phase-contrast image. Colocalization (yellow), as analyzed by confocal laser scanning microscopy, is indicated by white arrows.
If incubated with viable T. whipplei, immature DC-SIGN+ Mo-DC from healthy subjects (not shown) or treated CWD patients (Fig. 4C) exhibited T. whipplei bacteria partly colocalizing with the major histocompatibility complex class II-loading compartments, as characterized by HLA-DR and DC-LAMP. However, T. whipplei did not colocalize with the pattern recognition receptor DC-SIGN in Mo-DC (Fig. 4C).
DC populations and lymphocyte proliferation in T. whipplei-infected LN.Since DC prime naive T cells in LN, LN integrity is crucial for the initiation of adaptive immune responses. Therefore, we determined in situ the distribution and number of DC and Treg, as well as local lymphocyte proliferation.
In control LN, DC-SIGN+ cells commonly appeared in dense networks around medullary cords and subcapsular areas and sporadically in paracortical areas (Fig. 5A). S-100+ DC were dispersed mainly throughout the T cell areas of the paracortex but also were found in smaller numbers within germinal centers (Fig. 5B). In LN from CWD patients, the chronic infection with T. whipplei resulted in lymphadenopathy and in profound histoarchitectural alterations, such as lymphangiectasia, mostly in mesenteric LN, or the formation of granulomas predominantly in peripheral LN (Fig. 5). Lymph nodes with severe T. whipplei infestation, particularly mesenteric LN, showed reduced numbers of DC-SIGN+ cells and S-100+ cells (Fig. 5A and B; also see Fig. S2A in the supplemental material) but also of CD1a+ cells and DC-LAMP+ cells (data not shown) compared to control LN. Aside from the typical T. whipplei-stuffed enlarged tissue macrophages (7), DC in LN from CWD patients (primarily DC-SIGN+ immature DC and, to a lesser extent, S-100+ mature DC) carried intracellular T. whipplei bacteria (Fig. 5A and B).
DC populations and lymphocyte proliferation in T. whipplei-infected LN. DC and proliferating cells were detected immunohistochemically in mesenteric and peripheral LN from untreated CWD patients and control subjects without pathological findings. Representative examples of the costaining of DC-SIGN (red) (A), S-100 (red) (B), or Ki-67 (red) (C) with T. whipplei (brown) are shown. GC, germinal centers; L, lymphangiectasia; G, granuloma.
Lymph nodes from CWD patients exhibited a significantly reduced proliferation of B cells in germinal centers and of T cells in the paracortical areas compared to control LN (Fig. 5C). Although the lymphocyte proliferation correlated positively with the generation of Treg in LN from CWD patients, the presence of FOXP3+ Treg did not differ significantly between LN from CWD patients and control subjects (see Fig. S2B in the supplemental material). Subjects with tuberculosis or sarcoidosis also exhibited reduced numbers of DC-SIGN+ and S-100+ DC (Fig. S2A in the supplemental material), as well as of Ki-67+ proliferating cells within LN, but showed numbers of FOXP3+ Treg similar to those of control subjects without pathological findings (see Fig. S2B in the supplemental material).
DISCUSSION
The courses of a number of infections, e.g., with Mycobacterium tuberculosis (21, 24, 26), HIV (19), or hepatitis C virus (22), are associated with a modulation of DC functions that are suspected to be involved in the pathogenesis and persistence of the agent, and primary immunodeficiencies can be triggered by the genetically determined absence of DC subpopulations (28). Here, we characterized for the first time the potential of peripheral DC populations and of defined DC populations in situ from patients with CWD.
DC circulating in the peripheral blood of patients with untreated CWD, in particular CD11c+ M-DC, were reduced, as previously described for patients with tuberculosis (21, 24) or chronic HIV or hepatitis C infection (19, 22), as well as for inflammatory bowel disease (20) or other chronic inflammatory disorders (23). We did not observe, however, a selective depletion of CD1c+ M-DC-1, as described for primary immunodeficiencies caused by mutations in the interferon regulatory factor 8 gene (28). The decrease of M-DC might be induced by migration of mature DC to the gut, the major site of T. whipplei infection, as described for Crohn's disease (34). However, apoptosis following chronic activation by high levels of circulating inflammatory mediators (23) or by a T. whipplei infection, as shown for macrophages (35), also might be responsible for reduced numbers of circulating DC, since elevated apoptosis markers have been described for serum of untreated CWD patients (36). During chronic disorders, circulating M-DC may show an enhanced maturation (20), and, especially during inflammatory bowel disease, this might be triggered by microbial translocation through a leaky intestinal barrier. However, M-DC from CWD patients were mainly immature, as observed under steady-state conditions in healthy individuals (37). The poor maturation of peripheral M-DC of CWD patients might be explained by impaired TLR-mediated maturation that was demonstrated here. This impaired maturation of DC from CWD patients does not result from reduced expression of TLR4 and does not seem to be an intrinsic defect, since in vitro-generated Mo-DC from CWD patients and control subjects were phenotypically similar and stimulated recall responses of CMV-specific CD4+ T cells. However, Mo-DC from CWD patients were not able to activate T. whipplei-specific CD4+ T cells. This inability of Mo-DC to stimulate T. whipplei responses can be explained by the absence of a CD4+ T cell memory for T. whipplei (4) but also could be due to an impaired functionality of Mo-DC. In fact, a significant decrease in expression of CD83 and PD-L1 of Mo-DC from CWD patients compared to control subjects upon maturation in the presence of T. whipplei lysate indicated impaired maturation as described for DC during, e.g., hepatitis C virus infection (38). More importantly, there seems to be an intrinsic dysfunction in the production of IL-12 by DC from CWD patients, since the proportion of IL-12-synthesizing M-DC was significantly reduced in CWD patients compared to that of controls. The secretion of IL-12 also was reduced in Mo-DC cultures of CWD patients compared to Mo-DC from healthy controls, although the overall amount of IL-12 secreted was low. This comparably low IL-12 secretion by Mo-DC cultures might originate from insufficient CD40/CD40L ligation during Mo-DC cultivation (39, 40). IL-12 production in M-DC could not be restored by the blockage of TGF-β or IL-10, both of which are enhanced in the serum and duodenal mucosa of CWD patients (7, 12) and may decrease IL-12 production (15), or by the addition of IFN-γ, which may stimulate the release of IL-12 (15). Interestingly, the blockade of IL-10 resulted in an enhanced proportion of IL-12-producing M-DC in healthy subjects, which might be relevant in other pathophysiological situations when IL-10 is secreted (25, 40).
M. tuberculosis and Bordetella pertussis, as well as hepatitis C virus, have been reported to modulate DC maturation and cytokine expression and to support the generation of a regulatory milieu (27, 38, 41, 42). Similarly, T. whipplei lysate was a weak activation stimulus only for M-DC and Mo-DC from control subjects in terms of the enhanced expression of costimulatory molecules and of IL-23. However, additional signals, such as that from IFN-γ (43), or the interaction of CD40 and CD40 ligand (39) might contribute to the induction of CCR7, IL-12, and IL-23 after contact with T. whipplei in vivo. T. whipplei lysate further impaired the secretion of IL-12p40 in in vitro-generated Mo-DC from CWD patients, which supports the idea of an immunomodulation by T. whipplei (7, 44). An immunomodulation by T. whipplei itself is probable, since some of the immunological deficiencies of CWD patients, like the percentage of M-DC and IL23p19+ M-DC after stimulation with LPS, as demonstrated here, or the Th1-reactivity to SEB (12), are restored after successful treatment, while others persist despite the eradication of the agent (4, 7, 10). Although the analyses of cytokine secretion and DC activation in this study were limited by relatively low numbers of specimens (due to the rarity of CWD) and a high intra-assay variability due to the natural heterogeneity of the individuals included, collectively the data demonstrate specific deficiencies in DC functionality in CWD.
In tissue from CWD patients, we identified DC populations by the expression of DC-SIGN (32) and S-100 (31), as there is no single marker to label human DC (32). DC-SIGN has been described as a marker of immature phagocytosing DC due to their low expression of HLA-DR and costimulatory molecules (32), and S-100 defines mature DC (31). In fact, DC-SIGN+ cells seem to be composed of a heterogeneous population of antigen-presenting cells that include immature/semimature DC and macrophages (34). However, DC and macrophages were distinguishable, since T. whipplei-stuffed macrophages in the duodenum of CWD patients were greatly enlarged (7) and showed a CD163+ CD68+ CD11c+ HLA-DR+ phenotype (data not shown) and only a weak granular staining of DC-SIGN and S-100 in the cytoplasm. In contrast, DC were characterized by a strong, diffuse expression of DC-SIGN and S-100.
In contrast to Crohn's disease, which is associated with an accumulation of DC and T cells (34), we observed only steady-state numbers of DC-SIGN+ and S-100+ DC in the lamina propria of untreated and treated CWD patients. Thus, in CWD only a very weak inflammatory response is induced despite a massive bacterial load (1, 7). To maintain the DC population in the duodenal mucosa, continuous replenishment is required. Consequently, in CWD patients, an increased portion of DC-SIGN+ mononuclear cells in peripheral blood, the enhanced expression of the intestinal homing molecule integrin β7 on circulating monocytes, and the increased expression of CCR2 ligand in the duodenal mucosa point to a recruitment of monocytes to the intestine (7 and unpublished data from our group).
While acute HIV infection is attended by an enhanced migration of mature DC to the LN (45), numbers of DC-SIGN+ and S-100+ DC were markedly reduced in the draining mesenteric LN of CWD patients. The reduction was similar to that observed in LN from patients with tuberculosis or sarcoidosis. The massive infestation by T. whipplei may provoke the destruction of germinal centers. The low numbers of DC coincided with a reduced proliferation of lymphocytes in germinal centers and T cell areas, indicating an impaired activation of B and T cells. While the numbers of FOXP3+ Treg in LN from CWD patients and control groups were similar, the extent of proliferation correlated with the number of FOXP3+ Treg within LN of CWD patients, which was in line with the impairment of lymphocyte activation. Given that increased numbers of FOXP3+ Treg previously were detected by us in the duodenal mucosa of untreated CWD patients (12), we conclude that freshly generated Treg continuously migrate from the LN into the tissue (46).
In addition to their impaired functions as antigen-presenting cells in CWD patients, immature DC might directly contribute to the systemic spread of T. whipplei, as single intracellular bacteria were detected within DC-SIGN+ DC not only in the duodenum but also in lymph nodes from different locations. Thus, T. whipplei shows similarities to M. tuberculosis, which preferentially infects M-DC in the lungs and thereby is transported to the LN (26). Similarly, T. whipplei may be ingested by newly recruited or differentiated tissue DC that might transport T. whipplei to the draining LN (47). The short lifetime of DC (48) suggests that this is a continuous process. This hypothesis is supported by the fact that in LN from untreated CWD patients, mainly single T. whipplei bacteria were detected, but no enlarged periodic acid-Schiff (PAS)-positive macrophages were found (33).
In conclusion, our results demonstrate that CWD patient DC reveal deficient IL-12 production and maturation paralleled by reduced numbers of M-DC in the peripheral blood and DC in the LN. While the impaired maturation of DC seems to be caused by the overall regulatory milieu in the CWD patients, the reduced ability to produce IL-12 is an intrinsic defect that provides only suboptimal conditions for the priming of T. whipplei-specific CD4+ T cell responses. In addition, peripheral immature DC or myeloid cells seem to ingest the bacterium and subsequently differentiate and migrate, supporting the systemic distribution of T. whipplei and feeding the chronic infection.
ACKNOWLEDGMENTS
We thank Diana Bösel and Martina Seipel for excellent technical assistance. We are grateful to all patients, control subjects, and staff of the referring hospitals for help in obtaining samples.
We thank the 5th Framework Program of the European Commission: QLG1-CT-2002-01049, Deutsche Forschungsgemeinschaft (KFO 104 and SFB633), for financial support.
None of the sponsors had any influence on the planning of the study, experimental setup, or interpretation of data.
We have no commercial or financial conflict of interest to declare.
FOOTNOTES
- Received 7 August 2014.
- Returned for modification 30 September 2014.
- Accepted 5 November 2014.
- Accepted manuscript posted online 10 November 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.02463-14.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
