In the paper “Sustained innate interferon is an essential inducer of tertiary lymphoid structures” Calvanese et al., investigate which cell types and signals are involved in the development of tertiary lymphoid structures (TLS), using in vivo lung TLS development models based on instillation of p(I:C) plus ovalbumin (Ova) intranasally in several genetic backgrounds to dissect contribution of distinct signalling pathways in TLS formation. The Author identify type I IFN signalling as an essential requirement for TLS induction, providing insight into the molecular events required for TLS formation in mice.
Overall, the manuscript is clearly written, and the data is well presented. Nevertheless, some control experiments are missing to robustly support some of the author conclusions and would benefit the quality and interpretation of the findings.
Specific comments:
• In Figure 1SB the percentages of TLS and mature TLS in the 6x group are higher compared to what was observed in Figure 1C at 21 days, while the percentages seems similar between the 1x poly (I:C) groups in Fig.1SB and the mice treat 6 times in the main Figure 1C. Could the authors comment on this difference? Could this be a matter of kinetics that differ between 1x and 6 x poly(I:C) exposure? Time course experiments also with the 1x group would be helpful.
• Along these lines, also in Figure 2C the TLS numbers with IFNAR blockade resemble more what was seen in Figure 1SB with 1x poly (I:C). Would this suggest that IFNAR blockade induces a delay in the kinetics of induction, mimicking lower intensity signalling?
• In Figure 2E authors show the IFN is equally produced in 1x and 6x poly(I:C) exposed mice. So, what makes the different between the two treatment groups in terms of TLS density then? Or is the GFP too stable to appreciate differences? How would the GFP mRNA levels be in these experimental groups?
• It is unclear if each Figure represents only one experiment with several animals or if animals from independent experiments have been pooled. This information would be important to better clarify reproducibility across experiments and help address/clarify variability across Figures. This also impacts robustness and validity of statistical tests applied across the study.
• In Figure S2, data demonstrating effective depletion of FAP+ cells and alveolar macrophages is lacking. Also, controls are lacking to compare +/- tamoxifen and WT animals. Moreover, animals constitutively lacking alveolar macrophages seem to have more mature TLS compared to the inducible strains. Overall, data presented here does not allow to clearly exclude a potential role of the macrophages in TLS formation and maturation.

The strict requirement of type I IFN is not entirely convincing based on the present data:
• RT-PCR of IFN genes only could be misleading as very few transcripts can be sufficient to trigger responses. What about other ISG?
• Figure S3, what is the effect of anti-IFNAR on myeloid and lymphoid cells in the Mx-GFP mouse model?
• Figure 4D, Cxcl13 expression in lower but not null in IFNAR1 KO mice, suggestion additional drivers of this expression besides type I IFN.
• Figure 4F, Ifnlr-/- KO mice do respond to poly(I:C) in terms of Ccl21, differently from the WT controls and what stated by the authors in the conclusion. More generally speaking, the way this data is presented renders comparisons difficult and a bit confusing. Overall, concluding on dispensability of type III IFN seems premature based only on gene expression data in this paragraph.
• Indeed, in Figure 5, the density of TLS is reduced by 50% in Ifnlr-/- animals with one outlier that contributes to lack of significance.
• Figure 5F is unclear, could the signal be quantified? Also, it is not mentioned in the text.