top of page

Alveolar macrophages strictly rely on GM-CSF from alveolar epithelial type 2 cells before and after birth

Gschwend J. et al. (BioRxiv) DOI: 10.1101/2021.04.01.438051

Alveolar macrophages strictly rely on GM-CSF from alveolar epithelial type 2 cells before and after birth


  • Alveolar macrophages

  • GM-CSF

  • Epithelial

Main Findings

In this pre-print manuscript by Gschwend et al., the authors used a combination of lineage-tracing for GM-CSF-producing cells in mouse models and high-dimensional flow cytometry to identify key cellular players that modulate alveolar macrophage (AM) occupation and maintenance in the pulmonary niche via GM-CSF signalling.

First, to address the first and foremost question of which cell types act as the main reservoirs of GM-CSF that modulate AM abundance in the alveoli, the authors generated Csf2fl/fl (encoding GM-CSF) reporter mice, in which GM-CSF-producing cells also express tdTomato. Of note, the authors surveyed both the hematopoietic and non-hematopoietic sources of GM-CSF in the lungs of reporter mice to identify specific subsets of immune and epithelial/stromal cells that significantly express GM-CSF. Type 2 innate lymphoid cells (ILC2s), T cells, and epithelial cells were the predominant producers of GM-CSF. And, to assess the importance of these different GM-CSF reservoirs for the AMs, the authors developed Cre-flox mice to remove GM-CSF production in specific target cell types. The authors used Vav1Cre-Csf2fl/fl to target hematopoietic cells that express GM-CSF and disable their expression of Csf2. No quantifiable difference in AM abundance was observed, suggesting that hematopoietic-derived GM-CSF is dispensable for the development of AMs in the neonatal lung.

Still, the authors provided additional experimentation that interrogated the basophil compartment in the lungs to address a previous report by Cohen et al. (Cell, 2018). In this report, the authors performed single-cell RNA-sequencing (scRNA-seq) and ligand-receptor analyses to identify immune cell interactions that might modulate the AM population. Cohen et al. reported that basophils engage the Csf2:Csf2rb and Il33:Ilr1 signalling axes, and these components are needed to properly induce the canonical AM transcriptional program in bone marrow-derived macrophages. Interestingly, Cohen et al. reported a slight loss of AMs and a decline in phagocytic capacity upon antibody-mediated depletion of basophils, though Gschwend et al. did not observe this loss of AM quantity and quality, despite using the same diphtheria toxin depletion model of basophils (Mcpt8Cre-R26DTA/+ mice). Collectively, these findings from Gschwend et al. and Cohen et al. could suggest that basophils are dispensable for AM occupation of the alveolar niche and self-maintenance but still play an important role in determining the full maturation and overall “quality” of AMs.

The authors then interrogated the non-hematopoietic compartment. Using flow cytometry to separate epithelial, endothelial, and mesenchymal cells, the authors concluded that EpCAM+ CD104- SP-C+ NaPi-IIb+ cells (alveolar type 2 epithelial cells; AT2s) are the primary non-hematopoietic cell type that produces GM-CSF. To determine whether this compartment controls the establishment of an AM population in the developing lungs, the authors generated SpcCre-Csf2fl/fl mice, which delete Csf2 expression in only AT2s, and SpcCreERT2/+-Csf2fl mice, which permit tamoxifen-inducible deletion of the Csf2 locus in AT2s. Using these mice and by comparing the abundance of AMs in these mice to those of Csf2+/+ or Csf2fl/fl mice, the authors were able to conclude that specific production of GM-CSF by AT2 cells serves a crucial and non-redundant function in (1) the differentiation of fetal liver monocytes into AMs across the different stages of embryonic and post-natal development and (2) the maintenance of this population through adulthood.


Gschwend et al. documents a significant amount of flow cytometric and imaging data that are accompanied by thorough quantification. However, several conclusions can be viewed as over-reaching. Namely, the experiments analysing lungs of mice at different stages of development to evaluate the impact of AT2-specific loss of GM-CSF production on the differentiation of fetal monocytes into AMs relied solely on flow cytometry. The intermediate cell states between the fetal monocyte and AM stages were separated according to degree of expression of Ly6C and F4/80. This is not a particularly robust method of characterizing the impact on differentiation. Qualitative sequencing data may have been more appropriate and more informative for this task. Overall, this is what the pre-print lacks: a qualitative comparison (e.g., transcriptomic or epigenetic) between, for example, AMs in wild-type (or Csf2fl/fl) mice and the few AMs that remain in SpcCre-Csf2fl/fl or SpcCreERT2/+-Csf2fl mice to ascertain what molecular imprints can be detected in AMs that were barred from any GM-CSF signalling.


Prior to the release of the pre-print manuscript, the field benefited from nearly two decades of research on the biology and function of embryonic alveolar macrophages in the pulmonary immune ecosystem. Extensive fate-mapping studies demonstrated that these cells are derived from fetal liver monocytes that migrate to the lungs during the first week of life and differentiate into alveolar macrophages, upon exposure to tissue-specific cues. Among these, granulocyte monocyte-colony stimulating factor (GM-CSF, encoded by Csf2) is required to maintain the embryonic pool of alveolar macrophages in the alveolar niche.

Studies preceding these findings had already demonstrated that pulmonary epithelial cells, namely alveolar type 2 epithelial cells (AT2s), produce GM-CSF and can correct alveolar proteinosis in GM-CSF-deficient mice. Subsequent investigations revealed additional characteristics about these tissue-resident cells. Of note, these cells are epigenetically programmed by their microenvironment (Lavin et al., Cell 2014). And, in the case of alveolar macrophages, they react to cues from the surrounding tissue and become programmed to engage traditional macrophage functions, including apoptotic cell clearance (Roberts et al., Immunity 2017). Most recently, lineage-tracing and scRNA-seq profiling of the myeloid compartment in lung lesions revealed that hematopoietic-derived macrophages and embryonic alveolar macrophages contribute to the anti-tumor response differently (Lavin et al., Cell 2017; Lohyer and Hamon et al., JEM 2018; Casanova-Acebes et al., Nature 2021). Even in settings of lung infection, embryonic and monocyte-derived macrophages are transcriptionally distinct and interact with their epithelial and stromal environments differently.

Perhaps most importantly, remodelling of the pulmonary tissue architecture – whether it be due to neoplastic conversion of AT2 cells into lung adenocarcinoma or physiological aging – can impact the availability, maintenance, and function of the epithelial cell compartment. Therefore, an improved understanding of how the non-hematopoietic compartment of the pulmonary environment engages alveolar macrophages at steady state was needed. This pre-print provides convincing experimental evidence that suggests that AT2 cells, since embryogenesis and through adulthood, are responsible for the establishment and maintenance of an AM population in the lungs of mice. This information offers a more precise explanation of how disruption of the pulmonary niche can result in the loss or exclusion of AMs, seen in a variety of lung pathologies, including pulmonary fibrosis, COVID-19, and lung cancer.


Reviewed by Matthew Park as part of the cross-institutional journal club of the Immunology Institute of the Icahn School of Medicine, Mount Sinai and the Kennedy Institute of Rheumatology, University of Oxford. Follow him on Twitter.

bottom of page