Bronchus-associated macrophages efficiently capture and present soluble inhaled antigens and are capable of local Th2 cell activation

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "Bronchus-associated macrophages efficiently capture and present soluble inhaled antigens for local Th2 cell activation" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

The study here presented interesting evidence on the nature of a cell population responsible for antigen uptake and presentation in the airway and how this cell population could stimulate a Th2 response locally. Authors used various transgenic mice models to selectively deplete DC or macrophages and performed multi photon imaging of live tissue or 3D reconstruction of thick lung sections from these mouse models. However, all the reviewers concluded that the results presented here were too preliminary and lacking substantial argument on the novelty of the characterised population (overlap with lung IM? no RNA-seq characterisation) as well as the quantitative aspect of the data reported here by imaging and finally the claim on the local Th2 cell activation performed by this population. We all discussed these matters and estimated that too many additional experiments would be necessary to further address these very important questions. Please find below the comments of the reviewers that we hope that you will find helpful.

Reviewer #1:

The authors address the question of antigen uptake focusing on the bronchus, with the rational that it represents the main site of antigen deposit. They challenge the idea that contrariwise to the alveolar space a specialized subset of bronchus-associated macrophages BAM but not the DC are the main subset involved in Ag-uptake in the conducting airways. They re-evaluate the contribution of DC in the initiation of inflammation in an experimental model of allergic airway disease. The authors conduct a very nice set of experiments with different transgenic mice models to selectively deplete DC or macrophages and multi photon imaging of live tissue or 3D reconstruction of thick lung sections.

The authors bring evidence that BAM are potent phagocyte of soluble Ag in specific regions of the lung. Although they show that BAM mediates eosinophil recruitment in a model of allergic reaction, the experiment performed do not support convincingly their contribution to T cell activation.

I raised different points below.

1. One limitation of the study is that the phenotype of BAMs by flow cytometry overlaps with all IM of the lungs, vascular-, nerve associated and pleural macrophages. It is thus difficult to address specifically the function of BAM in the initiation of allergic inflammation as the depletion strategies are not specific of BAM.

The authors focus their imaging study at the branchpoint regions, in collagen rich structure where they observe accumulation of BAM, eosinophils and numerous other CSF1R+ cells.

It is unclear whether the author suggest that these regions represent a more specific route of Ag uptake compared to other regions of the bronchial epithelium. The images of OVA uptake are very convincing but lack quantification to address this point.

The authors could provide a quantification from their imaging study showing the relative proportion of EGFP+ cells with OVA uptake in different regions (branchpoints, airway epithelium and pleura).

The authors claim that soluble Ag is preferentially captured by BAMs but they excluded AM from the flow experiment quantification. As it is important to compare Ag uptake between DC and IM it is also important to see the contribution of AM. The authors should add in figure 4A B the contribution of siglecF+ AM.

2. Eosinophilia is used as the major read out of the macrophage depletion experiments at the beginning of the manuscript and the authors conclude on a non-necessary role of DC to drive allergic recall response and that other subsets may activate T cells. Eosinophil recruitment might be mediated by macrophages but T cell activation mediated by DC. A direct link between effector T cell activation and eosinophilia should be presented at the very beginning to draw such conclusion or maybe postpone this conclusion only after the OT2 experiments. However, the DC compartment is not fully depleted, considering the strong ability of DC to activate T cells compared to macrophages as showed Figure 6, even a small fraction of DC could maintain allergic response.

3. Figure 6C. Proliferation of OT2 cells can be easily reached in the presence of the OVA peptide even with nonprofessional APC. OVA peptide addition is a good control but the stimulation of OT2 cells by freshly purified APC from OVA-sensitized mice and/or with APC loaded with OVA protein is necessary.

4. During live imaging, many cell behaviors can be recorded, the authors need to provide proper quantification of cell interactions in order to identify an overall tendency. What is the proportion of OT2 interacting with EGFP+ cells compared to YFP+? Is the interaction OVA-dependent?

5. Percentages do not prove efficient depletion of cell in DTR system. Lower % could be biased by increased % of another subset. To compare the depletion of the AM in CD11c and CD169 DTR systems the authors should show the absolute number of AM as done for IM Figure 3.

6. Figure 4 There is a significant fraction of CD24+ CD64+ subset. In panel 3C the CD24dimerCD64+ subset is excluded from gating but in Figure 4A it is included in the macrophages. What is this subset exactly?

Reviewer #2:

In allergic asthma, parabronchial inflammation occurs within the lung upon inhalation of allergens and activation of Th2 leading to cytokine productions promoting pathogenesis. The study of Tang et al. aimed to identify the cell population responsible for antigen uptake and presentation in the airway and to characterize how this cell population can stimulate a Th2 response locally. The study is based in large parts on usage of two photon microscopy analysis of lungs from a collection of genetically engineered mice to follow myeloid cells at the level of airways in the steady state or after antigen inhalation. The authors found that a population of interstitial macrophages referred to as Bronchus-Associate Macrophages (BAMs) are very efficient for capture and presentation of soluble antigens. These BAMs are located in the steady state at airway bifurcation/branchpoint underneath the epithelium together with Dendritic cells (DCs essentially DC2) and collagen. Interestingly, this corresponds to location where inhaled antigens are likely to be deposited. Accordingly, the authors show that the BAMs exhibit a better ability to uptake inhaled antigens than the DCs. In addition, upon local sensitization, DCs but not BAMs, upregulate CCR7 and migrate to LN while BAMs remain lung resident where they can locally induce allergic inflammation.

The study is well conducted with some elegant experiments and represents a large body of work. The conclusions are of interest. The fact that this population has been missed before due to its intermediate expression of CD11c is very well documented.

I have two main concerns:

1) The authors claim that they have identified a new population of macrophages and try to compare it with the abundant literature on myeloid mouse subsets in their 10 page long discussion. I guess the best way to clearly define a new population for which they have a nice phenotype would be to purify these cells and RNAseq them. This would allow the authors and the readers to actually compare this population to the ones identified in previous studies. RNAseq in bulk is relatively cheap and does not need a lot of cells nowadays.

2) Many nice observations by two photon microscopy are provided without any quantification or performed with too few experimental points to be convincing. "Representative images" that are presented cannot be taken as results since some of the most important conclusions of the study rely on these images. For instance, in Figure 2 there is no quantification of the images of antigen uptake, OVA and HDM (Figure 2D and E). The same holds true for almost all the images presented. I understand that 2 photons microscopy is difficult and time consuming but for some experiments it should be done like the authors did in Figure 1G and H.

Reviewer #3:

In this report, Tang and colleagues have used two-photo microscopy to assess in detail the localization and the behavior of CX3CR1-GFP+CD11c+ MHC-II high interstitial macrophages (IM) in vivo. They convincingly identified areas rich in CX3CR1-GFP+ IMs and CD11c-YFP+ DCs in the steady-state lung, particularly in collagen-rich regions near some bifurcating airways. They called the CX3CR1-GFP+ IMs of those areas "Bronchus-Associated Macrophages (BAMs)", and importantly showed that BAMs have a dendritic morphology and can take up soluble antigens more importantly than DCs. BAMs remain lung-resident, are able to process and present soluble antigens, activate T cells in vitro and interact with T cells in an OVA-induced asthma model. These findings are consistent with the idea that BAMs may have previously been overlooked as antigen-presenting cells in the lung. Thus, the title claims that "Bronchus-associated macrophages efficiently capture and present soluble antigens", and such claim is fully supported by the data and well demonstrated. The experiments related to that part are sound, and the report is of great interest and adds to our understanding of lung interstitial macrophage biology by showing that MHC-II high IMs can act as antigen-presenting cells in vivo, as previously suggested in vitro (Chakarov et al., Science, 2019). This part of the work warrants publication in eLife.

The concern is related to the second claim of the title, "… for local Th2 cell activation". In my opinion, experimental evidence for that is very weak, and, more globally, the paradigm-shifting idea that BAMs, but not DCs, are necessary for "allergic recall responses in the lung" should, in my opinion, be removed from this report (data related to Figure 3). Indeed, unambiguously demonstrating this would require a huge amount of additional experiments and should be the scope of another entire paper. The main issues related to this part are summarized below:

– No evidence for local Th2 activation has been shown in vivo. Only some data showing that BAMs can trigger T cell proliferation and IL-13 are shown ex vivo, using Th2 polarized effector T cells. A previous report (Chakarov et al., Science, 2019) has shown that CX3CR1high MHChigh IM subpopulations could induce Foxp3+ Tregs ex vivo. Additional intracellular stainings for transcription factors (Foxp3, Gata-3, T-bet, Rorgt) and cytokines (IL-4, IFN-g, IL-17), co-culture experiments with unpolarized naïve, or polarized memory or effector T cells should help deciphering differences in the ability of cDCs vs. BAMs to regulate T cell responses ex vivo. However, additional in vivo proof would also be required.

– The authors employed DT-induced cell depletion in several transgenic animals, in combination with a mouse model of OVA/alum-induced eosinophilia, to support the claim that BAMs, rather than DCs, can serve as antigen-presenting cells to activate local Th2 cell responses in the lung. Weaknesses of this set of experiments are:

1) the paucity of read-outs: only BAL eosinophilia has been measured, while many cardinal features of allergic asthma should have been assessed: peribronchial and perivascular inflammation, mucus production, local Th2 cell recruitment and activation, antigen-specific humoral responses.

2) relevance of the model: most of the recent studies aiming at deciphering the contribution of cDCs in type 2-allergic asthma have used another more relevant model of allergic asthma based on exposure to house dust mite extracts (HDM), a major allergen source in humans. This model should have been used to validate the findings in the OVA/alum "artificial" model. Using a model based on HDM would also allow to decipher the contribution of cDCs vs. BAM in the sensitization phase vs. the recall response.

3) while it seems indeed that BAMs are depleted in CD11c-DTR mice, but not in zDC-DTR mice, while cDCs are depleted in both models, the unambiguous demonstration of the role of cDCs vs. BAM in triggering Th2-cell-mediated allergic responses (either during the sensitization or the challenge phase) would require adoptive transfer of cDCs vs. BAM in DT-treated CD11c-DTR or zDC-DTR animals, to see if that transfer can restore the phenotype affected by DT treatment.