1. Mucosal adjuvant activity of flagellin is dependent on Tlr5 signaling and antigen specific CD4+T cells.
To determine whether Tlr5 signaling is required for flagellin’s mucosal adjuvant activity, mice were immunized on day 0 and day 21 with ovalbumin and flagellin that is defective in Tlr5 signaling. We found that antigen specific CD4 T cells that reside in the lung draining is essential to mediate flagellin's adjuvanticity (Fig. 1).
2. Lung epithelial cells (LECs) sense flagellin via Tlr5 to release GM-CSF.
In previous work, we and others found that lung structural cells are primary targets that mediate flagellin’s mucosal adjuvant effect. We observed that the tracheal epithelial cells that were grown, in vitro, as air-liquid interface (ALI) cultures respond to flagellin and upregulate several genes associated with innate immune signaling. We observed that Gm-csf mRNA is significantly upregulated (~10 fold) at 2 hour post-treatment while genes for other DC modulating cytokines such as TSLP, IL-33 and IL-25 were not impacted. Consistent with our results using ALI culture, GM-CSF was detected in the bronchio-alveolar lavage of mice that were given intra-nasal instillation with flagellin (Fig. 2).
3. In-vitro generated type 2 conventional dendritic cells respond to GM-CSF released by flagellin stimulated LECs.
While lung resident CD11b+ dendritic cells (cDC2) are indispensable for flagellin’s adjuvant effect, how LECs cross-talk with cDC2 is not known. We used Flt3L to generate cDCs from the bone marrow of Tlr5-/- mice which closely resemble the cDC population found in murine lungs. We observed that neutralizing GM-CSF from flagellin activated LECs with the use of anti-GMCSF antibody significantly reduced the frequency of activated cDC2 (Fig. 3).
4. Flagellin dependent GM-CSF production in the lung enhance the number of pulmonary cDC2 in the draining lymph nodes.
To extent our in vitro observation we evaluated whether LECs derived GM-CSF can drive the accumulation of lung DCs in the lung draining lymph nodes. Intranasal instillation of flagellin increased the proportion of cDC2 among total CD11c+MHCII+ DCs by 35% compared to naïve animals. Concomitantly, we found significant reduction in the fraction of cDC2 among total DCs in animals that received anti-GM-CSF antibody along with flagellin. Furthermore, the frequency of CD80+CD86+ cDC2 were significantly reduced by neutralizing flagellin induced GM-CSF production in the lung. In addition, the activated cDC2 were less activated exhibited by the significant reduction in the MFI of CD80 following GM-CSF antibody treatment but not with isotype control (Fig. 4).
5. LECs derived GM-CSF improves ability of cDC2 to prime antigen specific CD4+ T cells.
To address the contribution of GM-CSF primed cDC2 to interact with CD4+ T cells, we loaded LEC supernatant treated Flt3L derived DCs with MHCII restricted OVA peptide before incubating with CTV stained CD4+ T cells from OTII mice. Flagellin stimulated LEC supernatant was sufficient to activate DCs thereby increasing the frequency of proliferated CD4+ T cells. However only those DCs that were exposed to flagellin activated LEC supernatant had robust expansion of CD4+ T cells. Importantly, blocking the activity of GM-CSF in flagellin stimulated LEC supernatant abrogate DC dependent T cell proliferation (Fig. 5).
6. In vivo neutralization of GM-CSF abrogates the mucosal adjuvant activity of flagellin.
In order to determine the role of GM-CSF primed DCs in promoting the adjuvant effect of flagellin we administred neutralizing GM-CSF antibody concomitant with flagellin and OVA treatements. Animals that were given anti GM-CSF antibody with flagellin and OVA antigen had 723.25% less antigen specific CD4+ T cells compared to isotype treated controls. Futhermore, we observed a significant decrease in serum as well as BAL OVA specific IgG response in mice that received GM-CSF antibody (Fig. 6).
***The manuscript is in preparation for submission to the journal ''Mucosal Immunology''
