Background
Tuberculosis (TB), caused by infection with
Mycobacterium tuberculosis (
Mtb), remains a major disease worldwide and is the leading infectious disease in terms of mortality, being responsible for an estimated 1.3 million deaths globally in 2016. Moreover, in the same year, there were an estimated 10.4 million new cases of active TB worldwide.
Mycobacterium bovis bacillus Calmette–Guérin (BCG) is the only TB vaccine for humans in current use, but its efficacy is insufficient to prevent pulmonary TB in adults and reactivation of latent
Mtb infection [
1]. BCG vaccination mainly induces effector, rather than central, memory T cells, which are maintained for a shorter period, explaining the limited duration of protection afforded [
2,
3].
CD4
+ T-cell responses and the production of interferon gamma (IFN-γ) are particularly important to the containment of
Mtb infection [
4,
5]. Dendritic cells (DCs) represent the bridge between the innate and adaptive immune responses and specifically strengthen the cellular immune response against mycobacterial infections [
6,
7]. Thus, the mechanisms involved in major histocompatibility complex class II (MHC II) antigen (Ag) processing and presentation, which are required for CD4
+ T-cell activation, are crucial for controlling
Mtb infection [
8]. Much research has indicated that DCs play an important role in anti-mycobacterial immune responses in the early stages of infection, but little is known of the kinetics of Ag presentation by these cells soon after
M. bovis BCG exposure. Indeed, efforts to understand the basis of protective immunity against
Mtb have led us the examinntion of even earlier infection time points. We previously investigated the Ag-presenting cell (APC) functions of murine DCs during the first 2 weeks following intravenous administration of recombinant BCG (rBCG) expressing the
Escherichia coli MalE protein as a reporter Ag [
9]. However, this process has not yet been directly examined in lymph node (LN) DCs using an endogenous
M. bovis BCG Ag.
In the present study, we evaluated the in vivo dynamics of early Ag presentation by murine inguinal LN DCs in response to M. bovis BCG. The results showed that the early Ag-presenting activity of murine DCs induced by M. bovis BCG Ag85A protein in vivo was transient and that the inhibition of Ag processing due to the decreased production of Ag85A peptide is the primary reason for the rapid loss of Ag85A peptide-MHC II complexes.
Discussion
CD4
+ T-cell responses and the production of IFN-γ are particularly important to the containment of
Mtb infection. In mice, between 1 and 3 weeks after initial infection,
Mtb-specific T cells appear in the lungs, IFN-γ is expressed, and the bacterial burden is controlled [
10]. Production of IFN-γ by splenocytes in response to Ag restimulation is observed within 6 days after i.v.
Mtb infection [
11]. To determine when the T-cell response is initiated, we obtained splenocytes and inguinal LN cells from mice 3, 6, and 9 days after s.c. BCG injection. Inguinal LN cells collected 6 days after infection produced IFN-γ in response to Ag85A restimulation. Thus, the T-cell immune response appears to have been initiated in the inguinal LN day 6 following BCG infection.
The mechanisms involved in MHC class II Ag processing and presentation, which are required for CD4
+ T-cell activation, are crucial for controlling
Mtb infection. Previous research investigated the kinetics of Ag-presenting activity by harvesting spleens following i.v. administration of rBCG expressing the
E. coli MalE protein as a reporter Ag. The formation of MalE peptide-MHC complexes in splenic DCs was detected at 2, 4, and 12 h after rBCG infection, while MalE was barely detectable at 48 h [
9]. However, this process has not yet been directly examined in LN DCs and by using an endogenous
M. bovis BCG Ag. To investigate the dynamics of LN DCs Ag presentation, we harvested and sorted these cells from inguinal LNs at several time points after s.c. injection of mice with Ag85A protein or BCG and tested their capacity to stimulate DE10 T-cell hybridomas, which are specific for an immunodominant Ag85A peptide. In this manner, in vivo formation of Ag85A peptide-MHC complexes on DCs from BCG-injected mice was detected by measuring IL-2 production in DE10 T-cell hybridoma culture supernatants ex vivo. Ag85A peptide-MHC complexes on LN DCs appeared rapidly after inoculation, with IL-2 production being detected in response to DCs collected 4 h after BCG infection and the highest production in response to those harvested at 12 h. By contrast, IL-2 levels following exposure to DCs harvested 72 h after infection were barely detectable. Together, these results indicate that the MHC II presentation of mycobacteria-derived peptides by inguinal LN DCs is only transient, with Ag85A peptide-MHC II complexes on the surfaces of inguinal LN DCs disappearing rapidly. Some reports have shown that peptide-MHC complexes have a half-life of 25 h [
12]. Thus, it can be concluded that the synthesis of Ag85A peptide-MHC II complexes on inguinal LN DCs was interfered.
Several reports have shown that
Mtb and
M. bovis inhibit intracellular processes associated with Ag presentation, including Ag processing, MHC class II expression, the trafficking of MHC class II molecules, and peptide-MHC class II binding [
13,
14]. CIITA is the master transcriptional regulator of MHC class II molecules [
15]. The transcription of CIITA itself is regulated by the three unique promoters pI, pIII, and pIV, which drive the expression of CIITA types I, III, and IV, respectively. pI is constitutively active in DCs [
14]. In the current investigation, LN DCs exhibited the up-regulation of cell-surface MHC II molecules from 4 to 96 h following infection. Using real-time PCR, we analyzed the transcription levels of MHC II, total CIITA, and CIITA type I in DCs in response to BCG. Expression of MHC II was found to be induced by BCG infection relatively slowly, while total CIITA transcription was rapidly induced. This indicates that the transcription and expression of MHC II proteins on the cell surface did not declined following BCG infection, suggesting that the expression and trafficking of MHC class II molecules may be not associated with the rapid loss of Ag85A peptide-MHC II complexes. As a result, it can be deduced that LN DCs do not provide a continuing source of mycobacterial Ag85A peptides for the formation of peptide-MHC II complexes.
Considerable evidence shows that DCs can phagocytose mycobacteria and may be the first cells to encounter such pathogens, therefore, DCs are likely to be responsible for initiating the subsequent immune response. The survival of mycobacteria within DCs has been assessed previously in vitro using the BCG vaccine strain and virulent
M. bovis, both of which were shown to be phagocytosed by DCs after 24 h of infection [
16].
Mtb cells disseminate to draining LNs within 8 days following respiratory infection [
17]. Approximately 2% of the splenic DC population (CD11c
+ cells) was found to contain BCG at 4 h following i.v. infection [
9]. In the present study, the presence of rBCG-GFP bacilli in inguinal LN DCs following s.c. inoculation of mice was monitored by FACS. As expected, the percentage of infected DCs increased to 2% after 96 h of infection. We then examined whether mycobacteria survive and multiply within DCs during infection. Following s.c. administration of BCG to mice, CFUs appeared at 4 h, increased significantly by 12 h, remaining elevated until the last time point. These results suggest that BCG survives within the inguinal LN DC pool, representing a continuing source of mycobacterial Ag85A protein with which LN DCs can form Ag85A peptide-MHCII complexes in vivo. Some reports have shown that live
Mtb can alter phagosome maturation and decrease Ag processing, providing a mechanism for
Mtb to evade immune surveillance and enhance its survival within the host [
18‐
20]. Based on our findings, we conclude that the inhibition of Ag processing due to the reduced production of Ag85A peptide is the primary reason for the rapid loss of Ag85A peptide-MHC II complexes.
Methods
Experimental animals
Six-week-old female C57BL/6 mice were purchased from Vital River (Beijing, China). The mice were housed, handled, and immunized at our animal biosafety facilities, and all procedures were approved by the Institutional Animal Experimental Committee of Yangzhou University. All experiments were performed according to the national guidelines for animal welfare. The mice were euthanized by cervical dislocation under isoflurane, and spleens and inguinal LNs were collected for analysis.
Bacterial strains and culture conditions
M. bovis BCG Pasteur 1173P2 and rBCG expressing GFP (rBCG-GFP) were kindly provided by Dr. Xiaoming Zhang (Institut Pasteur of Shanghai, Chinese Academy of Sciences, Shanghai, China). Both strains were grown with gentle agitation (80 rpm) in Middlebrook 7H9 medium (Difco, Detroit, MI, USA) supplemented with 0.05% Tween 80 and 10% albumin-dextrose-catalase (ADC) enrichment or on solid Middlebrook 7H10 medium (Difco) supplemented with 0.05% Tween 80 and 10% oleic-ADC enrichment.
T-cell hybridoma and Ags
MHC II-restricted DE10 T-cell hybridomas specific for the
Mtb Ag85A peptide comprising amino acids 241 to 260 [
21] were kindly provided by Dr. Claude Leclerc (Institut Pasteur, Paris, France). The Ag85A protein was constructed and expressed in our laboratory, and the Ag85A peptide (amino acids 241–260) was synthesized by SciLight Biotechnology (Beijing, China).
Detection of IFN-γ production following BCG infection
C57BL/6 mice were s.c. vaccinated with 1 × 108 CFU BCG and sacrificed 3, 6, and 9 days later, at which point, spleens and inguinal LNs were removed aseptically and transferred to complete RPMI-1640 medium for preparation of single-cell suspensions. The mononuclear cells, isolated using Histopaque 1083 (Sigma, St. Louis, MO, USA), were seeded at 1 × 106 cells/well in 96-well plates containing complete RPMI-1640 medium. They were subsequently stimulated with 10 μg/ml Ag85A peptide, 10 μg/ml Ag85A protein, or 5 μg/ml bovine PPD (Prionics, Schlieren, Switzerland) and incubated at 37 °C in an atmosphere of 5% CO2 in air. Supernatants were then harvested at 48 h post-stimulation, frozen, and later tested for IFN-γ concentration by sandwich enzyme-linked immunosorbent assay (ELISA, BD Biosciences, Franklin Lakes, NJ, USA).
Ag presentation assay
C57BL/6 mice were s.c. injected with 1 × 108 CFU BCG or heat-killed BCG in 200 μl PBS or with PBS alone. Mice were sacrificed at various time points, and their inguinal LNs removed and perfused with 400 U/ml collagenase type IV (Invitrogen, Carlsbad, CA, USA) containing 50 μg/ml DNase I (Invitrogen). Single LN-cell suspensions were prepared, and DCs were sorted with an autoMACS separator (Miltenyi Biotec, Bergisch Gladbach, Germany) using CD11c as a cell marker. Specifically, LN cells were first incubated with anti-CD11c MicroBeads (Miltenyi Biotec) before autoMACS separation, resulting in a population of CD11chigh cells (DCs). The purity of these murine LN DCs was then analyzed using a FACSCalibur instrument (BD Biosciences). For the ex vivo Ag presentation assay itself, the purified LN DCs were transferred to 96-well microplates and serially diluted in complete RPMI-1640 medium. DE10 T-cell hybridomas at a density of 1 × 105/well were then added, and after incubation for 24 h, supernatants were collected, frozen, and later tested for IL-2 content by sandwich ELISA (BD Biosciences).
Cell phenotype analysis
C57BL/6 mice were s.c. injected with 1 × 108 CFU BCG in 200 μl PBS or with PBS alone. The mice were sacrificed after various periods for the preparation of single inguinal LN-cell suspensions and sorting of DCs by autoMACS. FITC-conjugated anti-CD11c, and biotinylated anti-I-Ad, anti-CD40, anti-CD54, anti-CD80, and anti-CD86 antibodies were used to label cells. Allophycocyanin-conjugated streptavidin was employed to visualize biotin conjugates. A FACSCalibur and FlowJo software (FlowJo LLC, Ashland, OR, USA) were then used for multicolor staining analysis of the labeled cells. DCs were sorted using the autoMACS system before being pelleted and resuspended in lysis buffer. Cellular RNA was purified with an RNeasy kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions, and total RNA was reverse-transcribed into cDNA using SuperScript reverse transcriptase (Thermo Fisher Scientific, Waltham, MA, USA).
In vivo infection assay
C57BL/6 mice were s.c. injected with 1 × 108 CFU BCG or rBCG-GFP in 200 μl PBS or with PBS alone. Mice were sacrificed at various time points, and single inguinal LNs were removed aseptically and transferred to complete RPMI-1640 medium. Single-cell suspensions were prepared, and DCs were sorted with an autoMACS separator as above. The percentage of DCs infected with rBCG-GFP was analyzed using a FACSCalibur instrument and FlowJo software. BCG-infected DCs were pelleted and resuspended in lysis buffer. Ten-fold serial dilutions of these suspensions were then plated on solid Middlebrook 7H10 medium, and colonies were counted after incubation at 37 °C for 2–3 weeks.
Statistical analysis
All data are expressed as means ± SE. Statistical analysis was performed by Student’s t-test using GraphPad Prism software. P values < 0.05 were considered statistically significant.
Acknowledgements
The authors are grateful to Dr. Claude Leclerc (Institut Pasteur, Paris, France) and Dr. Xiaoming Zhang (Institut Pasteur of Shanghai, Chinese Academy of Sciences, Shanghai, China).