Protective immunity to Mycobacterium tuberculosis infection by chemokine and cytokine conditioned CFP-10 differentiated dendritic cells

Abstract

Background: Dendritic cells (DCs) play major roles in mediating immune responses to mycobacteria. A crucial aspect of this is the priming of T cells via chemokines and cytokines. In this study we investigated the roles of chemokines RANTES and IP-10 in regulating protective responses from Mycobacterium tuberculosis (M. tb) 10 kDa Culture Filtrate Protein-10 (CFP-10) differentiated DCs (CFP10-DCs).

Methods and findings: Infection of CFP10-DCs with mycobacteria down-modulated RANTES and IP-10 levels. Pathway specific microarray analyses showed that in addition to RANTES and IP-10, mycobacteria infected CFP10-DCs showed reduced expression of many Th1 promoting chemokines and chemokine receptors. Importantly, T cells co-cultured with RANTES and IP-10 conditioned CFP10-DCs mediated killing of mycobacteria from infected macrophages. Similarly, T cells recruited by RANTES and IP-10 conditioned CFP10-DCs mediated significant killing of mycobacteria from infected macrophages. IFN-gamma treatment of CFP10-DCs restored RANTES and IP-10 levels and T cells activated by these DCs mediated significant killing of virulent M. tb inside macrophages. Adoptive transfer of either RANTES and IP-10 or IL-12 and IFN-gamma conditioned CFP10-DCs cleared an established M. tb infection in mice. The extent of clearance was similar to that obtained with drug treatment.

Conclusions: These results indicate that chemokine and cytokine secretion by DCs differentiated by M. tb antigens such as CFP-10 play major roles in regulating protective immune responses at sites of infection.

Figure 1. CFP10-DCs downregulate RANTES and IP-10 levels following BCG infection.

GM-CSF-DCs or CFP10-DCs were infected with 1 MOI BCG for 24h. RANTES and IP-10 levels in culture supernatants were measured by ELISA. Data from one of five experiments is shown.

Figure 2. RANTES and IP-10 conditioned CFP10-DCs induce pro-inflammatory in vivo primary immune responses to BCG.

For Panel A GM-CSF-DCs or CFP10-DCs were infected in vitro with 1 MOI BCG for 24h. DCs were extensively washed to remove extracellular bacteria. 5×10⁶ DCs were adoptively transferred into naïve mice. Seven days later the inguinal lymph nodes were removed and 5×10⁶ cells/ml were cultured in 10% RPMI1640 medium with 10% FCS for 48h. Levels of indicated cytokines were measured in culture supernatants. For B, prior to infection with BCG, GM-CSF-DCs were either untreated or incubated with neutralizing monoclonal antibody to RANTES or IP-10 for 4h, while CFP10-DCs were either untreated or incubated with 25 ng/ml RANTES or IP-10 for 12h. DCs were extensively washed and then processed as in Panel A. Data from one of three experiments are shown.

Figure 3. RANTES and IP-10 elevate intracellular calcium in BCG stimulated CFP10-DCs.

FLUO-3-AM labeled GM-CSF-DCs (profile a–c) or CFP10-DCs (profile d–f) were stimulated with 1 MOI BCG and real-time increase in intracellular calcium was monitored over a period of 5 min. For profiles b and c GM-CSF-DCs were incubated with 5 μg/ml neutralizing monoclonal antibody to RANTES and IP-10, respectively, for 2h, prior to BCG stimulation. For profiles e and f, CFP10-DCs were incubated with 5 ng/ml recombinant RANTES and IP-10, respectively, for 12h, prior to BCG stimulation. DCs were extensively washed prior to flow cytometry. Data from one of three experiments are shown.

Figure 4. BCG infected CFP10-DCs show reduced expression of pro-inflammatory chemokines.

Total RNA was enriched from either GM-CSF-DCs (upper panels) or CFP10-DCs (lower panels) following infection with 1 MOI BCG (upper right and lower right panels) for 24h. 1 μg RNA was processed for microarray analyses using the GEArray Q Series mouse chemokine and receptor gene array from SuperArray strictly following the manufacturer's instructions. Rows A–I contains genes for various chemokines and their receptors while rows M and N contain house-keeping genes. Rows J–L represents negative controls. The Table below the figure depicts genes, their position and densitometric units of spots from blots of BCG infected GM-CSF-DCs and BCG infected CFP10-DCs. Data from one of two experiments are shown.

Figure 5. RANTES and IP-10 conditioned CFP10-DCs induce Th1 responses.

Either GM-CSF-DCs (GMCSF) or CFP10-DCs (CFP10) were infected with 1 MOI BCG for 24h and co-cultured for 48h with T cells enriched from BCG immunized mice. For some groups GM-CSF-DCs were incubated with 5 μg/ml neutralizing monoclonal antibody to RANTES and IP-10, respectively, for 2h, prior to BCG stimulation. CFP10-DCs were incubated with 5 ng/ml recombinant RANTES or IP-10 for 12h, prior to BCG stimulation. Data from one of three independent experiments are shown.

Figure 6. T cells co-cultured with or recruited by RANTES or IP-10 conditioned CFP10-DCs mediate clearance of BCG from macrophages.

For A, BCG infected CFP10-DCs (CFP10), conditioned or not with either 25 ng/ml RANTES or IP-10 were co-cultured for 48h with BCG primed T cells. From this, T cells were enriched and cultured with BCG infected macrophages (Mph) for 48h. A separate group wherein macrophages conditioned with 2 ng/ml IFN-γ (IFN-g) for 4h prior to infection with BCG was also included as a control. For B, T cells migrated into the lower chamber of a transwell apparatus in response to supernatants of BCG infected CFP-DCs, conditioned or not with 25 ng/ml RANTES or IP-10, were cultured with BCG infected macrophages (Mph) for 48h. A separate group wherein macrophages conditioned with 2 ng/ml IFN-γ (IFN-g) for 4h prior to infection with BCG was also included as a control. Cells from both Panels were lysed and plated in serial dilutions on 7H11 agar plates. CFU were counted 2–3 week later. Data are the mean±s.d. of three experiments. For A, P<0.002 (Mph vs CFP10+BCG+RANTES), P<0.004 (Mph vs CFP10+BCG+IP10). For B, P<0.02 (Mph vs CFP10+BCG+RANTES), P<0.01 (Mph vs CFP10+BCG+IP10).

Figure 7. RANTES and IP-10 conditioned CFP10-DCs mediate clearance of M. tb infection in vivo in mice.

Groups of mice were infected with 1×10⁶ M. tb H37Ra intravenously via the tail vein. Seven days post-infection, 10×10⁶ unconditioned CFP10-DCs or CFP10-DCs conditioned with both RANTES and IP-10 together (25 ng/ml each) for 12h (conditioned CFP10-DCs) were injected into the tail vein of mice. A repeat injection was given 7 days following the first transfer. Seven days following the 2^nd transfer, mice were sacrificed and lung and spleen homogenates were plated onto 7H11 agar plates in serial dilutions for CFU monitoring. In parallel, infected mice were injected with anti-TB drugs as given in Materials and Methods . Control represents mice infected with M. tb H37Ra followed by intravenous injection of PBS. Data are the mean±s.d. of three experiments. For lungs, P<0.02 (Control vs Conditioned CFP10-DCs), P<0.009 (CFP10-DCs vs conditioned CFP10-DCs). For spleen, P<0.006 (Control vs Conditioned CFP10-DCs), P<0.01 (CFP10-DCs vs conditioned CFP10-DCs).

Figure 8. Treatment with IFN-γ restores RANTES and IP-10 levels from BCG infected CFP10-DCs.

Either GM-CSF-DCs or CFP10-DCs were incubated with 2 ng/ml IFN-γ for 1h and washed. DCs were infected with 1 MOI BCG for 24h. Expression of RANTES and IP-10 was measured in culture supernatants. Data from one of three experiments are shown.

Figure 9. T cells co-cultured with IFN-γ or IL-12 conditioned CFP10-DCs mediate clearance of M. tb H37Rv from macrophages.

BCG infected CFP10-DCs (CFP10), conditioned or not with either 2 ng/ml IFN-γ (IFNg) or IL-12 or transformed with retroviruses expressing IFN-γ (retroIFNg) or IL-12p40 (retroIL-12), were co-cultured for 48h with BCG primed T cells. From this, T cells were enriched and cultured with M. tb H37Rv infected macrophages (Mph) for 48h. A separate group wherein macrophages conditioned with 2 ng/ml IFN-γ, prior to infection with M. tb H37Rv was also included as a control. Cells from all the groups were lysed and plated in serial dilutions on 7H11 agar plates for CFU monitoring 2–3 week later. Data are the mean±s.d. of three experiments. P<0.007 (Mph vs CFP10+IFNg), P<0.007 (Mph vs CFP10+IL-12); P<0.005 (Mph+IFNg vs CFP10+IFNg), P<0.005 (Mph+IFNg vs CFP10+IL-12); P<0.03 (CFP10 vs CFP10+IFNg), P<0.05 (CFP10 vs CFP10+IL-12).

Figure 10. IFN-γ or IL-12 conditioned CFP10-DCs mediate clearance of M. tb infection in vivo in mice.

Groups of mice were infected with 1×10⁶ M. tb H37Ra intravenously via the tail vein. Seven days post-infection, 10×10⁶ unconditioned CFP10-DCs or CFP10-DCs treated with both IFN-γ and IL-12 together (2 ng/ml each) for 2h (conditioned CFP10-DCs) were injected into the tail vein of mice. A repeat injection with the same number of DCs was given 7 days following the first transfer. ‘Control’ represents mice infected with M. tb H37Ra followed by intravenous injection of PBS. Seven days following the 2^nd transfer, mice were sacrificed and lung and spleen homogenates were plated onto 7H11 agar plates in serial dilutions for CFU monitoring. In parallel, infected mice were injected with anti-TB drugs as given in Materials and Methods . Data are the mean±s.d. of three experiments. For lungs, P<0.04 (Control vs Conditioned CFP10-DCs), P<0.03 (CFP10-DCs vs conditioned CFP10-DCs). For spleen, P<0.02 (Control vs Conditioned CFP10-DCs), P<0.02 (CFP10-DCs vs conditioned CFP10-DCs).

Figure 11. Conditioning CFP10-DCs with chemokines or cytokines induces protective responses.

Following infection by M. tb and their sequestration inside alveolar macrophages, antigens such as CFP-10 are secreted. CFP-10 induces the differentiation of DCs. CFP10-DCs secrete high levels of RANTES and IP-10. Following their interactions with mycobacteria CFP10-DCs reverse their phenotype and secrete low levels of RANTES, IP-10 and IL-12p40 and display poor calcium influx. This leads to increased survival of mycobacteria and the induction of suppressor T cell responses. Box, conditioning CFP10-DCs with either RANTES and IP-10 or IL-12 and IFN-γ induced pro-inflammatory T cell responses and T cells activated by these DCs mediate killing of M. tb inside infected macrophages. In addition, adoptive transfer of chemokine and cytokine conditioned CFP10-DCs into mice harboring an active M. tb infection results in clearance of infection that is commensurate with drug treatment.