The question

Since the discovery of oxygen-sensing proteins in the early 2000s, the role of oxygen in cell biology has been increasingly explored1. Low blood oxygen levels (hypoxemia) are commonly found in disease processes that involve damage to the lungs and in severe illness, when oxygen delivery to tissues and organs can be compromised. Acute respiratory distress syndrome (ARDS)2, which is a feature of critical SARS-CoV-2 infection3, is a condition defined by the presence of hypoxemia, in which morbidity and mortality are both associated with increasing degrees of systemic hypoxia. We questioned whether the presence of systemic hypoxia could, in itself, directly alter innate immune responses, thereby contributing to the host-mediated injury associated with poor clinical outcomes in patients with ARDS.

The observation

We characterized the number and phenotype of circulating monocytes in a cohort of patients with ARDS who had substantial and sustained levels of hypoxemia despite ventilatory support, and observed that these patients with ARDS had early monocytopenia relative to healthy controls (Fig. 1a).

Fig. 1: Hypoxemia alters bone marrow responses to inflammation.
figure 1

a, Patients with ARDS have an early reduction in levels of circulating monocytes compared with healthy controls. b,c, In mouse models of hypoxic ALI, hypoxemia alters hematopoiesis and reduces monocyte egress from the bone marrow, as measured by proportions of colony forming unit-erythroids (CFU-Es) (b) and BrdU+ monocytes (c) after lipopolysaccharide (LPS) challenge, resulting in monocytopenia and failed accumulation of monocyte-derived macrophages (MDMs) in the lung. d, Administration of CSF-1 rescues hypoxia-induced monocytopenia, which enables the accumulation of MDMs within the lung and promotes the resolution of inflammation. © 2022, Mirchandani, et al.

We were able to replicate this phenomenon in mouse models of hypoxic experimental acute lung injury (ALI), in which we combined challenge with lipopolysaccharide (LPS), Streptococcus pneumoniae or influenza A virus with exposure to reduced inspired oxygen levels. In these mice, we show that the reduction in circulating monocyte numbers is secondary to changes in hematopoiesis (Fig. 1b) and a reduction in type I interferon signaling, which result in impaired monocyte egress from the bone marrow (Fig. 1c). These deficits, in turn, led to a failure to accumulate CD64brightSiglecFLy6C+ monocyte-derived macrophages in the inflamed lung and to impaired inflammation resolution. Furthermore, circulating monocytes from hypoxemic mice were phenotypically distinct from those seen in normoxic mice, with altered expression of several surface adhesion molecules and chemokine receptors that are important for monocyte effector function. These observations paralleled the phenotypic switch seen in blood monocytes from patients with ARDS.

To understand whether we could overcome the pathophysiological effect of hypoxia and enable the resolution of lung inflammation in our model, we expanded monocyte production in the bone marrow by treating the mice with colony-stimulating factor 1 (CSF-1). CSF-1 treatment not only expanded the number of blood monocytes but also normalized the expression of genes that were upregulated or downregulated by hypoxic exposure. Importantly, these changes led to the accumulation of CD64brightSiglecF macrophages in the lung, along with the appearance of a LYVE1+ subpopulation that is phenotypically identical to pro-reparative macrophages in the lung4. Together, these effects of CSF-1 treatment resulted in improved physiological outcomes, enhanced inflammation resolution and reduced lung injury (Fig. 1d).

The implications

Our data demonstrate that hypoxemia is not merely a bystander in severe, acute respiratory illness, but is an active participant that shapes the immune response and alters disease outcomes. Furthermore, we show that by overcoming the direct effects of hypoxemia on monocyte dynamics, we can improve outcomes in experimental models of hypoxic ALI.

Mouse models provide an important tool that enables the investigation of the direct contribution of hypoxia to the origin, accumulation and phenotype of monocyte-derived macrophages in the lung. However, whether the findings obtained in mice will directly translate to alterations within the human lung immune landscape warrants further investigation. Furthermore, studies of human tissue samples and cohorts of patients with ARDS will be required to confirm whether CSF-1 can be used as a treatment for ARDS.

Our future work will focus on further interrogating the aberrant hypoxia-induced immune changes we have identified, and on broadening our studies of CSF-1 as a therapeutic agent both in patient-derived tissue and in human participants.

Ananda S. Mirchandani and Sarah R. Walmsley,

University of Edinburgh, Edinburgh, UK.

Expert opinion

“This work seeks to understand how hypoxia affects the resolution of lung inflammation in the setting of ARDS, and therefore touches on an important area. Investigations into type 1 interferon signaling and the role of CSF-1 in this context are new and represent strengths of the paper. Taken as a whole, these findings are expected to be of high impact and should advance the field.” William Janssen, National Jewish Health, Denver, USA.

Behind the paper

When we set out to investigate the effect of ARDS-induced hypoxemia on circulating monocytes, we anticipated seeing phenotypic differences but the presence of monocytopenia was unexpected. However, we replicated this result in our hypoxic mouse model of ALI, which confirmed that monocytopenia was a direct effect of hypoxemia. This finding led us to investigate the consequences of hypoxemia for both reprogramming the immune landscape of the lung and inflammation outcomes. The experiments tracking monocyte egress from the bone marrow were a definite ‘eureka!’ moment, as they confirmed that hypoxemia had far-reaching effects beyond the lung.

The emergence of the SARS-CoV-2 pandemic led to many delays in completing this work, most notably because we were redeployed from research into front-line duties in the respiratory clinic. The morbidity and mortality associated with critical SARS-CoV-2 infections also highlight the clinical importance of continued investigation into how hypoxemia alters disease outcomes. A.S.M. and S.R.W.

From the editor

“The paper makes the important point that oxygen therapy does not rescue tissue hypoxia in patients with ARDS and that tissue hypoxia is an important driver of hematopoietic adaptations that occur in the bone marrow, which have important implications for the resolution of lung inflammation. Another important observation is that treatments that aim to rescue monocyte hematopoiesis during hypoxia also seem to be effective in the resolution of lung inflammation.” Ioana Visan, Senior Editor, Nature Immunology.