A lipidated peptide of Mycobacterium tuberculosis resuscitates the protective efficacy of BCG vaccine by evoking memory T cell immunity

Female BALB/c mice (6–8 week) were procured from the Experimental Animal Facility, CSIR-Institute of Microbial Technology, Chandigarh, India. Mice were immunized subcutaneously (sc) at the base of tail with the Danish strain of BCG (10⁶ CFU/mouse). Twenty-one days later, BCG-primed mice were boosted twice with L91, at an interval of 14 days apart (BCG-L91). Control groups were immunized with BCG alone (BCG), placebo (PBS) or an antigenically irrelevant lipidated influenza virus hemagglutinin peptide (abbreviated as LH or Pam2Cys). Mice were aerosol challenged with Mtb (~100 CFU/mouse), 90 days after the last booster. Subsequently, animals were sacrificed after 90 days of Mtb challenge. Later, immunological (ex vivo), protection and histopathology studies were performed. To monitor the antigen specific T cell response, mice were sacrificed 30 days after Mtb infection, and cellular responses were examined following in vitro stimulation with L91, Pam2Cys and short term culture filtrate of H37Rv (ST-CF). In all the experiments, changes in the response on vaccination were compared among BCG-L91 and control BCG and placebo (PBS) groups or otherwise indicated.

Lipidated synthetic peptides used in the study were produced by solid phase synthesis method, as described elsewhere [12]. The lipidated promiscuous peptide of sequence SEFAYGSFVRTVSLPVGADE was from the Acr1 antigen of Mtb (L91). The control, non-mycobacterial, lipidated peptide (LH) sequence ALNNRFQIKGVELKS was from influenza virus hemagglutinin light chain and was shown to be active in BALB/c mice [13].

Mtb H37Rv strain was cultured in 7H9 medium containing Tween-80 (0.05%), supplemented with albumin (10%), dextrose and catalase (ADC). Glycerol stocks of H37Rv were prepared and stored at −80 °C, and later used for infection studies. BCG vaccine (TUBERVAC) used for immunization was purchased from Serum Institute of India, Pune, India. TUBERVAC (Bacillus Calmette-Guerin Vaccine I.P.) is a live freeze-dried vaccine derived from an attenuated strain of Mycobacterium bovis and meets the requirements of WHO and I.P. when tested by the methods outlined in WHO, TRS. 745 (1987), 771 (1988) and I.P.

Chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO). Anti-mouse or anti-human fluorochrome labeled antibodies (Abs): CD4-PB, CD62L-APC, CD44-PerCP-Cy5.5, CD127-PE, FoxP3-FITC, Tim3-PE, PD1-PECy7, IFN-γ-PECy7, TNFα-PerCPCy5.5, IL-17-PerCPCy5.5, CD25APC-Cy7, CD45RA-PE, CD45RO-APC, and Abs for ELISA were procured from BD Pharmingen (San Diego, CA) or otherwise mentioned. RPMI-1640 and FBS were purchased from GIBCO (Grand Island, NY) for cell culture. For culturing of cells, tissue culture grade plastic-wares were purchased from BD Biosciences (Bedford, MA). Ab against iNOS used in Western blot was procured from (Abcam, Cambridge, United Kingdom).

Spleens and LNs obtained from the immunized mice and exposed to Mtb, were pooled and single cell suspension was prepared by gently pressing through frosted slides. Lungs were perfused with chilled PBS and small pieces were prepared and digested with collagenase (2 mg/ml) for 30 min/37 °C. Later, cells were passed through a sieve (70 μM). Viability was checked by trypan blue dye-exclusion method and cells (2 × 10⁵/well) were added to 96 well U-bottom culture plates and cultured with L91 (1 nmol), Pam2Cys (50 ng/ml), ST-CF (25 μg/ml) and medium for 72 h.

The cells (2 × 10⁷ cells) were incubated with carboxyfluoresceinsuccinimidyl ester (CFSE) dye in PBS (1 μM, 4 ml) at 37 °C. Free CFSE was quenched with 2 ml of FCS and excess was removed by washing with RPMI-FCS-10%. CFSE-labeled cells were cultured with either L91 (1 nmol), Pam2Cys (50 ng/ml), ST-CF (25 μg/ml) or medium for 72 h. The proliferation of CFSE-labeled cells was analyzed by flow cytometry.

The cells were stimulated as mentioned in the proliferation assay and then stimulated with PMA (50 ng/ml) and ionomycin (10 μM) for 4 h followed by incubation with brefeldin A (5 mg/ml) for an additional 2 h. The cells were then harvested, washed twice with buffer (PBS-FCS-2%) and fixed with paraformaldehyde (1X) at 4 °C for 30 min. Fixed cells were perforated with saponin (0.2%) and incubated with fluorochrome tagged anti-IFN-γ, IL-17A and TNF-α Abs at 4 °C for 90 min. The cells were washed with saponin (0.2%), followed by wash buffer. For surface staining, the cells were incubated with either fluorochrome labeled Abs or biotinylated Abs/streptavidin-fluorochrome conjugates. Standard protocols of washing/incubation were followed at each stage.

Flow cytometry was carried out using a FACS-Aria III and data was analyzed using the BD FACS DIVA software package (BD Biosciences, San Jose, CA). The gating strategies for FoxP3, PD-1, Tim-3, IFN-γ/TNF-α, IFN-γ/IL-17, CD62L/CD44, CD127 (Additional file 1: Figure S1) and CCR6/CXCR3 expression on IL-17/IFN-γ double positive cells (Additional file 2: Figure S2) have been shown in their respective figures.

The cultures were set as mentioned in T cell proliferation assay. Later, culture supernatants (SNs) were collected and cytokine concentrations were determined using a standard sandwich ELISA [14].

Monocytes were isolated from the femurs and tibia of the mice. The cells (2 × 10⁶/well) were cultured in the presence of granulocyte macrophage colony-stimulating factor (GM-CSF; 2 ng/ml) and interleukin-4 (IL-4; 4 ng/ml) for the generation of DCs. On day 3, the cultures were replenished with fresh medium and supplemented with GM-CSF and IL-4. On day 6, the cells were harvested, washed and added to 24 well plates (2 × 10⁵ cells/well). Bone marrow derived dendritic cells (BMDCs) were stimulated with either L91 (3 nmol), F91 (3 nmol), Pam2Cys (50 ng/ml) or LPS (4 μg/ml) for 16 h. SNs were collected for cytokines estimation by ELISA and the expression of surface markers was assessed by flow cytometry.

As mentioned above for the cultures of DCs and macrophages, BMDCs were stimulated with L91, F91, Pam2Cys and LPS for 16 h in 12 well plate (10⁶ cells/well). The cells were harvested and lysed in radioimmunoprecipitation assay buffer (RIPA) containing a protease inhibitor cocktail. Protein was estimated and equal amounts (30 μg) were subjected to SDS-PAGE (10%) followed by transfer to PVDF membrane. The non-specific sites were blocked with BSA (5%) and the blot was probed with anti-iNOS Abs (1:200) (Ab3523) or actin as a control. Later, the blot was probed with anti-rabbit-HRP Ab and finally developed using an enhanced chemiluminescence method (Lumigen, Inc. Southfield, MI). The blot was finally scanned using Image Quant LAS 4000 (GE Healthcare, Pittsburgh, PA).

BMDCs were stimulated with L91 (9 nmol), F91 (9 nmol), Pam2Cys (150 ng/ml) or LPS (4 μg/ml) for 30 min in 12 well plates (10⁶ cells/well). The cells were harvested and nuclear extract was prepared. Nuclear extracts of each sample were incubated with [P³²] labeled oligonucleotides containing the binding site for NF-κB at 37 °C for 20 min to allow the formation of DNA–protein complexes, which was then resolved by native gel electrophoresis using a 6% gel. Later, the gel was dried and exposed to a blank screen at room temperature for 6–10 h and scanned by a phosphorimager (Fujifilm, Tokyo, Japan).

Blood was obtained from BCG vaccinated healthy volunteers in sterile vacutainers. Blood was diluted with PBS in 1:1 ratio and overlaid on histopaque. Peripheral blood mononuclear cells (PBMCs) were separated by centrifugation at 400 g for 30 min at 25 °C. PBMCs were washed 3 times with PBS + 2% FCS and in vitro cultured with L91 (1 nmol), F91 (1 nmol) or Pam2Cys (50 ng/ml) for 96 h. During culture, IL-2 (100U) was added after 24 h.

The data are presented as mean ± standard error. Statistical analysis was performed using Graph Pad Prism employing an unpaired Student’s ‘t’test.

One of the main reasons associated with BCG failure is the generation of regulatory T cells (Tregs) following vaccination, which hampers its protective efficacy [5]. Consequently, we examined ex vivo generation of regulatory T cells (Tregs) in the cells isolated from the lungs of the immunized mice after 90 days of Mtb infection. We observed significantly (p ≤ 0.005) higher percentage of FoxP3⁺ CD4 Tregs from the BCG vaccinated mice, as compared to the control (placebo) group (Fig. 1a, b). On the other hand, a significant reduction of the FoxP3⁺ CD4 Tregs (p ≤ 0.0001) was observed in the group of mice that was primed with BCG followed boosting with L91 (BCG-L91), as compared to BCG. Thus, indicating that L91 boosting restricted BCG-mediated induction of Tregs and thereby may be augmenting immunity against Mtb.

Mycobacterium tuberculosis is known to induce exhaustion of T cells [11]. Recently, we have demonstrated that signaling via TLR-2 rescues CD4 T cells from exhaustion [11]. L91 comprises of TLR-2 agonist Pam2Cys and a MHC-II binding promiscuous peptide of Acr1 [10]. Therefore, we monitored whether BCG-L91 vaccination could rescue CD4 T cells from exhaustion against Mtb infection. The results showed a substantial reduction of the exhaustion markers PD-1 (p ≤ 0.05) (Fig. 1c, d) and Tim-3 (p ≤ 0.05) (Fig. 1e, f) on the CD4 gated T cells obtained from BCG-L91 group, compared to those of BCG group (Fig. 1c–f). There was no difference in the expression of these two markers on the cells isolated from BCG and the placebo group. These results demonstrate that L91-boosting prevented T cell exhaustion in Mtb infected mice.

Th1 cells and Th17 cells are two major arms of adaptive immunity that play a cardinal role in safeguarding against Mtb [15]. Th1 cells are known to mediate protection against Mtb through the release of IFN-γ and TNF-α [7, 10, 16]. Th17 cells confer protection by chemokine mediated recruitment of the cells of innate and adaptive immunity [17, 18]. BCG has shown to elicit a mixed Th1 and Th2 response and also a weak Th17 response [19, 20]. Our results show that L91 booster significantly augmented Th1 and Th17 immunity, as evidenced by a greater release of IFN-γ following in vitro stimulation with L91 (p ≤ 0.005) or ST-CF (p ≤ 0.05) and IL-17A (in vitro stimulation with L91: p ≤ 0.005; ST-CF: p ≤ 0.0005) by lung cells (Fig. 2a, b) than the control groups. Thus, BCG-primed and antigen-boost strategy generated better immunity than BCG alone against Mtb [17, 21].

Polyfunctional Th1 cells and Th17 cells are considered better in protecting against Mtb, compared to their counter parts secreting single cytokine [22, 23]. In our study here, the cells isolated from the lungs of the BCG-L91 group showed significantly greater expansion in the percentage of multifunctional Th1 cells (IFN-γ⁺TNF-α⁺) (L91: p ≤ 0.005, ST-CF: p ≤ 0.005) and Th17 cells (IL-17⁺IFN-γ⁺) (L91: p ≤ 0.005, ST-CF: p ≤ 0.05) following in vitro stimulation with either L91 or ST-CF, compared to the BCG group (Fig. 2c–f).

Multifunctional Th17 cells (IFN-γ⁺IL-17A⁺) co-expressing CXCR3/CCR6 are known to upregulate the ligands for CXCR3 (CXCL9/CXCL10/CXCL11) on the lung parenchymal cells, which helps to recruit Th1 cells to the site of infection [17, 18]. The results of our study showed that significantly more CXCR3⁺CCR6⁺ co-expressing polyfunctional Th17 cells (IFN-γ⁺IL-17A⁺) had been observed in the lungs of the BCG-L91 group following culturing with L91 (p ≤ 0.0005) and ST-CF (p ≤ 0.05) (Fig. 2g, h).

One of the main reasons for BCG’s inability to evoke long-term protection against Mtb is its failure to generate enduring memory T cells [5, 10]. The results of this study show that L91 induced significantly greater expansion in the pool of both central (CD44^hiCD62L^hi; L91: p ≤ 0.05; ST-CF: p ≤ 0.005) and effector memory (CD44^hiCD62L^lo; L91: p ≤ 0.05; ST-CF: p ≤ 0.05) CD4 T cells in the BCG-L91 administered mice than the control group following in vivo stimulation of L91 and ST-CF (Fig. 3a–c). These results were further substantiated by significantly (p ≤ 0.005) higher display of CD127, another marker for memory CD4 T cells (Fig. 3d, e). No discernible change was detected in the case of BCG group. The CD127 has been established to be an important marker for the sustenance of memory CD4 T cells [24].

The Acr1 antigen of Mtb is known to tolerize dendritic cells by upregulating the inhibitory molecule Tim-3 and consequently suppresses the host immunity [25]. L91 is derived from Acr1. Hence, it was worth to examine the influence of L91 on Tim-3 expression on dendritic cells. Our results show that in vitro stimulation with L91 significantly (p ≤ 0.0005) downregulated the expression of Tim-3 compared to unstimulated cultures (Fig. 4a). L91 comprises of the components that are capable of inducing both innate and adaptive immunity [10]. Innate immunity is the first line of defense to combat against any pathogen [26]. Nitric oxide (NO) mediated killing of Mtb is one of the well-established protection mechanism [27]. Hence, we thought it would be worth to check the impact of L91 on DCs in inducing the release of NO. Interestingly, we observed a dose-dependent increase in iNOS, the enzyme responsible for NO synthesis (Fig. 4b) upon stimulation with L91. Additionally, L91 considerably (p ≤ 0.0005) augmented TNF-α secretion (Fig. 4c). TNF-α is a major factor in restricting the intracellular survival of Mtb [7, 25]. Further, the augmentation in the expression of NF-κB was also detected with L91 stimulation (Fig. 4d). NF-κB is a key transcription factor responsible for the propagation of cells and pro-inflammatory responses [28]. Besides L91, considerably higher NF-κB expression was detected following stimulation with the Pam2Cys control but not un-lipidated peptide (F91). These results together demonstrate the role of Pam2Cys as a component of L91 in promoting the release of molecules of innate immunity responsible for protection.

After establishing the role of L91 in inducing the long-lasting memory T cell response, we next evaluated the protective efficacy of BCG-L91 against Mtb. To generate bona fide memory T cell response, BCG-L91 vaccinated mice were aerosol challenged with Mtb 90 days after the last immunization. Bacterial burden in the lungs and spleen was enumerated 90 days following the challenge. The results show that L91 boosting led to a greater reduction in the mycobacterial burden in the lungs (p ≤ 0.005) and spleen (p ≤ 0.005), compared to BCG-vaccinated group, and also the Pam2Cys and placebo control (Fig. 5a, b). The decrease in the CFU in the spleen also signifies the importance of BCG-L91 in restricting the dissemination of Mtb.

Further, we substantiated CFU data by histopathological study of lungs. Animals immunized with BCG-L91 exhibited reduced pathology in the lungs, manifested by decreased macrophage infiltration and lesser and smaller size of granulomas, fewer scattered peribronchiolar lymphocyte-rich areas and better preserved alveolar spaces, when compared to the BCG group, placebo and Pam2Cys controls. The control animals revealed Mtb induced severe pathology that was characterized by the larger size and number of lymphocyte rich compact granulomas with signs of irregular lung architecture due to excessive inflammation. Peribronchiolar cuffs and bigger confluent areas of consolidated mixed lymphocytes-histiocytes were also apparent (Fig. 5c). The quantitative assessment of the granulomas in lung sections revealed a decreased Mtb burden in BCG-L91 immunized animals (Fig. 5d).

L91 efficiently amplified BCG induced immunity in the experimental model of TB. Hence, we were next curious to know whether L91 was able to activate human CD4 T cells obtained from BCG vaccinated healthy volunteers. PBMCs cultured with L91 exhibited significantly higher (p ≤ 0.0005) proliferation and upregulation of the activation marker CD25 (p ≤ 0.0005) on CD4 T cells, as compared to those cultured with free peptide or Pam2Cys control (Fig. 6a–c).

We next observed that L91 stimulation improved the generation of Th1 cells, as it significantly (p ≤ 0.05) expanded the percentage of IFN-γ⁺ CD4 T cells. Further, we observed increased frequency (p ≤ 0.005) of polyfunctional Th1 cells expressing both IFN-γ and TNF-α (Fig. 7a–d). The results were obtained using blood from the non-TB endemic Australian population. Furthermore, we also monitored the efficacy of L91 using PBMCs from BCG-vaccinated healthy volunteers of the TB-endemic Indian population. It is worth to mention here that the BCG has failed to protect Indian population from Mtb [4]. We observed a significant augmentation in Th1 and Th17 immunity, as documented by the increased percentage of IFN-γ⁺ (p ≤ 0.0005) and IL-17A⁺ (p ≤ 0.005) CD4 T cells, respectively (Fig. 8a–c). Further, enhanced frequency (p ≤ 0.005) of IFN-γ⁺ and IL-17A⁺ polyfunctional Th17 cells was detected (Fig. 8d). Furthermore, expansion (p ≤ 0.005) of the pool of the memory precursors of CD4 T cells (CD45RA⁺CD45RO⁺) was noted (Fig. 8e). These results suggest that L91 efficiently activated polyfunctional human Th1 cells and Th17 cells and expanded the pool of memory CD4 T cells. Thus, has a great potential to reinvigorate the efficacy of BCG vaccine in TB endemic and non-endemic population.