The mucosal adjuvant cyclic di-GMP enhances antigen uptake and selectively activates pinocytosis-efficient cells in vivo

1) STING-dependency is shown only for immune correlates (for example; cytokines, antigen uptake, etc). Whether actual immunity to a pathogen (after immunization with CDN adjuvants) requires STING has not been tested in this study. It would therefore be interesting to challenge STING-deficient mice (both germline knockout and conditional CD11c-Cre) with S.pneumoniae following PspA+c-di-GMP vaccination (as done in Figure 1E with WT mice). The STING is indeed crucial to successful vaccination with CDNs, pathogen burden in STING knockouts should be similar to that seen in WT mice vaccinated with non-adjuvanted PspA. This experiment is important, as it would implicate STING in the actual outcome/effectiveness of CDN vaccination rather than in immune correlates of the immunization.

As requested, we now included the pneumococcal challenge studies done on CDG/PspA or PspA immunized WT vs Tmem173^-/- mice (Figure 7G) and WT vs Itgax^CreTmem173^Flox/Flox mice (Figure 9E). As expected, there is no significantly CFU difference between CDG/PspA and PspA immunized Tmem173^-/- mice. However, we found that PspA immunized Tmem173^-/- mice have significantly lower CFU than PspA immunize WT mice (Figure 7G). We further found that Tmem173^-/- mice, without PspA immunization, are more resistant to S. pneumoniae infection than the WT mice (see Author response image 1). This is a novel unexpected finding and points to a detrimental role of MPYS in host defense against S. pneumoniae infection. Currently, we are actively pursuing this novel and exciting discovery.

Author response image 1

Download asset Open asset

Similarly, CDG/PspA immunization did not lower bacterial burden in the Itgax^CreTmem173^Flox/Flox mice (Figure 9E). However, differently from the Tmem173^-/- mice, PspA immunized Tmem173^Flox/Flox mice had similar bacterial burden as the PspA immunized Itgax^CreTmem173^Flox/Flox mice (Figure 9E). We further found that Itgax^CreTmem173^Flox/Flox mice, without PspA immunization, were not resistant to S. pneumoniae infection (see Author response image 2). This suggested that MPYS expression in DCs and alveolar macrophages are not responsible for its detrimental role in host response to S. pneumoniae infection. Currently, we are actively pursuing this novel and exciting discovery.

Author response image 2

Download asset Open asset

2) The observation that the responsiveness of STING for CDNs varies between species raises some concern. C-di-GMP (the CDN used in this study) is far more potent in mouse cells in vitro compared to human cells (for example, Cavlar et al., EMBO J 2013, PMID 23604073, Figure 4). Therefore, whether c-di-GMP is superior to cGAMP as an adjuvant in humans is questionable. It is important that the authors discuss this point and amend the Abstract and other parts of the text to make it clear that c-di-GMP elicits stronger responses than cGAMP in mice. It is also noteworthy that different STING alleles found in humans encode proteins responding differentially to cGAMP isoforms (for example, Gao et al., Cell 2013, PMID 23910378).

We have amended the manuscript to emphasize that our observation was made in mice. We further added two paragraphs in the Discussion to address the issues raised here.

Minor comments:

1) Figure 4A: To help those readers who are unfamiliar with the DQ-OVA⁺OVA647 assay, could the authors include a cartoon of how these compounds work here?

Yes. We added Figure 4A and 4B to illustrate the mechanism of action of DQ-OVA and OVA-647. Thanks for the suggestion.

2) Figure 8F-J: Have the authors measured epithelial cell derived cytokines (as in Figure 3)? These should not be affected by knockout of STING in CD11C⁺ cells.

We added Figure 8K showing the IL-33 and TSLP production in lung from cyclic di-GMP treated Itgax^CreTmem173^Flox/Flox mice. The lung TSLP production was slightly lower in cyclic di-GMP treated CD11C^CreMPYS^flox/flox mice than in the Tmem173^Flox/Flox, but it was not statistically significant (Figure 8K). However, the lung IL-33 production was significantly lower in cyclic di-GMP treated Itgax^CreTmem173^Flox/Flox mice than in the Tmem173^Flox/Flox (Figure 8K).

MPYS expression in CD11C⁺ cells is responsible for the proinflammatory cytokines TNFα and IL-1β production, which can act on lung epithelial cells (Figure 8J). Furthermore IFN-λ receptors are uniquely expressed on lung epithelial cells and Itgax^CreTmem173^Flox/Flox mice lack cyclic di-GMP induced IFN-λ production (Figure 8I). We favored the idea that there is a crosstalk/communication between lung epithelial cells and CD11C⁺ cells during cyclic di-GMP induced immune response.