Decoding fibrosis in nonresolvable COVID-19: a role for myeloid-specific HIF1A deletion

Abstract

A complication of viral lung infections is the development of pulmonary fibrosis. This phenomenon is most evident in patients with COVID-19, where in its most aggressive form patients developed nonresolvable (NR) COVID-19 requiring lung transplantation. NR-COVID-19 was characterized by the presentation of a fulminant fibrotic lung injury that progressed rapidly, even in patients with limited comorbidities. However, the mechanisms that led to this rapidly progressing form of fibrosis are not fully understood. A common clinical manifestation in the most severe cases of COVID-19 was the presence of “silent” hypoxemia. Thus, we hypothesized that a dysfunctional hypoxic response may result in exacerbated lung injury seen in patients with severe forms of COVID-19. Our results demonstrate that despite increased expression of hypoxia-inducible factor 1A (HIF1A) and its downstream mediator adenosine A_2B receptor (ADORA2B), reduced macrophage HIF1A was observed in patients with severe COVID-19, including NR-COVID-19. Utilizing mice lacking HIF1A in myeloid cells using the lysozyme M Cre promoter, we demonstrate that these mice present with increased lung inflammation and pulmonary fibrosis following chronic low-dose bleomycin treatment. The augmented lung injury was associated with reduced markers for alternatively activated macrophages also observed in NR-COVID-19 lungs. These results point to reduced myeloid HIF1A as a mechanism that can lead to exacerbated lung injury in mice, which parallels the rapid fibrotic response observed in NR-COVID-19. Collectively, our results point to using HIF1A stabilizers as a potential avenue to prevent the development of rapidly progressing postviral lung fibrosis. However, special care is necessary since chronic HIF1A activation is also linked to fibrotic outcomes.

NEW & NOTEWORTHY

Postviral-induced lung fibrosis represents a severe and potentially fatal outcome. This was most evident during the COVID-19 pandemic, where a subset of individuals presented with fulminant lung fibrosis requiring lung transplantation. The mechanisms that promote this exacerbated lung injury are not fully known. Herein, we demonstrate that mice lacking myeloid hypoxia-inducible factor 1A (HIF1A) develop an exacerbated lung injury response to bleomycin that was consistent with reduced macrophage HIF1A expression in nonresolvable (NR)-COVID-19, characterized by extensive lung fibrosis.

INTRODUCTION

Postviral-induced lung fibrosis following acute respiratory distress syndrome (ARDS) is common (1–3). Whereas this fibrotic response is normally self-resolvable or progresses slowly, there are some reported cases of severe fulminant lung fibrosis following viral infection, most notably after severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) infection (4). Patients with SARS-CoV2 infection developed a fulminant nonresolvable (NR) COVID-19 fibrotic lung injury requiring lung transplantation (4). Importantly, the fibrotic injury in NR-COVID-19 presented with a unique profile compared to idiopathic pulmonary fibrosis (IPF), the most common diagnosis for lung transplantation (5). The fibrotic response in NR-COVID-19 was dominated by increased expression of periostin (POSTN) and collagen (COL) 3A1 compared to IPF (4). Intriguingly, in NR-COVID-19, there was also an upregulation of transitional stem cells expressing cytokeratin (KRT)-5 or KRT8 in the lung (4), which have been associated with deficits in lung regeneration (5). These observations point to a mechanism resulting in accelerated lung fibrosis following SARS-COV2 infection. Thus, we reasoned that comparing IPF (a chronic, progressive fibrotic disease), NR-COVID-19 (an acute, fulminant fibrotic outcome), and the bleomycin (BLM) model (which recapitulates the transition from acute injury to fibrosis) would allow us to dissect how hypoxia/hypoxia-inducible factor 1A (HIF1A) signaling in macrophages influences disease trajectory. This comparative framework provides both a chronic reference (IPF) and an acute-to-fibrotic model (bleomycin) against which to interpret the unique features of NR-COVID-19.

A unique feature that was observed in many COVID-19 patients was the presence of “silent” or “happy” hypoxia. Low oxygen saturation levels characterize this form of hypoxia without individuals experiencing breathing difficulty (6). Epidemiological studies have identified that between 5% and 65% of COVID-19 patients presented with silent hypoxia, which correlated strongly with male sex; older age; a history of diabetes, smoking, or chronic lung diseases; and high d-dimer (7). Hypoxia leads to the stabilization of hypoxia-inducible factor 1A (HIF1A), which is known to modulate protective responses in acute lung injury (8) and is regulated by adenosine (9), among other mediators. Several studies linked elevated HIF-1A levels with worsening infection and inflammatory responses (10–12). Yet, pharmacological activation of HIF1A was postulated as an early therapeutic approach (13), with a seminal study demonstrating how pharmacological activation of HIF1A reduced SARS-COV2 infection (14). The discrepancy in these studies may stem from the multicellular and pleiotropic effects of hypoxia/HIF-1A (15, 16). HIF-1A expression in macrophages has been shown to modulate alveolar macrophage (AM) stemness and proinflammatory effects (17), which may be linked with the initial inflammatory response in COVID-19 (18). Intriguingly, the identification of an inactive HIF-1A stabilization in severe disease particularly in macrophages has been postulated (19). In the context of lung fibrosis, several studies have demonstrated that elevated HIF1A expression promotes the development of pulmonary fibrosis (20–22). Furthermore, deletion of HIF1A once lung fibrosis is established was able to halt the progression of further fibrotic injury (20). In these instances, the fibrotic response was associated with an increase in alternatively activated macrophages (AAM) that was attenuated following inhibition of HIF1A (20). Thus, we hypothesized that dysregulation of myeloid-specific HIF1A may provide a mechanism leading to an accelerated fibrotic response post-SARS-COV2 infection.

Although our data demonstrate elevated expression of HIF1A in NR-COVID-19 samples, we report reduced HIF1A signals in macrophages. Subsequent studies in bleomycin-treated mice lacking HIF1A in constitutively expressed lysozyme M Cre expressing cells (HIF1A^f/fLysM^Cre) resulted in an augmented fibrotic and inflammatory response that was greater than in bleomycin-treated HIF1A competent LysM^Cre mice. Bleomycin-treated HIF1A^f/fLysM^Cre mice were devoid of markers for AAM and presented with elevated inflammation. This observation was consistent with reduced signals for AAM in NR-COVID-19 lung samples compared to IPF lungs. Our results suggest that turning off HIF1A expression in macrophages may result in an augmented profibrotic response that may explain the accelerated fibrotic injury seen in NR-COVID-19.

METHODS

Human Subjects

Explanted lung tissue samples from discarded donor lungs that served as controls were deidentified and obtained from LifeGift Texas (Houston, TX). NR-COVID-19 and IPF patients were obtained from the University of Texas Health Science Center, Houston (UTHealth Houston) Pulmonary Center of Excellence Biobank (Table 1). Collection of these explanted lungs was performed adhering to the Declaration of Helsinki and was reviewed by Institutional Review Boards from UTHealth Houston (HSC-MS-08–0354 and HSC-MS-15–1049).

Table 1.

Patient demographics

Diagnosis	Application	n	Average Age (yr)	Sex	Ethnicity
Discarded donor (control)	WB/IHC	6	39	M (6); F (0)	H (6)
NR-COVID-19	WB/IHC	6	44.7	M (5); F (1)	B (1), H (4), W (1)
IPF	WB/IHC	5	60.4	M (5); F (0)	B (1); H (2), W (2)
Discarded donor (control)	RT-qPCR	6	39	M (6); F (0)	H (6)
NR-COVID-19	RT-qPCR	23	46.7	M (19); F (4)	B (2), H (13), W (5), ND (3)
IPF	RT-qPCR	12	53.6	M (10); F (2)	B (2); H (2), W (6), ND (2)

Patient demographics for samples used for Western blot (WB), immunohistochemistry (IHC), or quantitative RT-PCR (RT-qPCR). B, Black; F, female; H, Hispanic; IPF, idiopathic pulmonary fibrosis; M, male; ND, not disclosed; NR, nonresolving; W, White.

Generation and Genotyping of Mouse Lines

All experiments were performed in male, 5- to 6-wk-old mice on the C57Bl6 background. LysMCre and HIF1A^f/fLysM^Cre were a kind gift from Holger K. Eltzschig and were back-crossed at least six generations on the C57Bl6 background. The number (n) used was between 4 and 12 mice as denoted in each figure legend. Study designs were approved by the Institutional Animal Care and Use Committee review board from UTHealth Houston (AWC-15–0143). Hif1A was conditionally deleted using LysMCre, a widely used driver with predominant activity in neutrophils and monocyte/macrophage lineages. We note that Cre activity outside classical myeloid compartments can occur in some contexts; therefore, our primary readouts emphasized myeloid-linked biology [bronchoalveolar lavage (BAL) differentials, neutrophil influx, and macrophage/AAM markers] to interpret phenotypes.

Intraperitoneal Bleomycin Mouse Model

Male, 5–6 wk old, LysM^Cre and HIF1A^f/fLysM^Cre mice were treated with eight intraperitoneal bleomycin (BLM) doses (0.035 U/g; Teva Pharmaceutical, Irvine, CA) or PBS injections (vehicle), twice a week over a 4-wk period as described previously (23). Mice were euthanized and samples collected at the following time points after BLM or PBS treatment: 14, 17, 20, 21, 25, and 33 days.

Assessment of Arterial Oxygen Saturation

Mice were shaved along the neck and awake for arterial oxygen saturation measurement on day 33 using the Pulse MouseOx software (STARR Life Sciences Corp., Oakmont, PA).

Plasma, Bronchoalveolar Lavage Fluid, Cellular Differentials, and Histology

Once mice were anesthetized with Avertin, blood was collected and centrifuged to collect plasma. Lungs were lavaged four times with 0.3 mL PBS to collect BAL fluid (BALF). A hemocytometer was used to determine total cell counts. Aliquots of BAL cells spun onto microscope slides and stained with Diff-Quick (Dade Behring, Deerfield, IL) were used to assess cellular differentials. The remaining BALF was centrifuged, and the cell pellet and supernatant were stored for further analysis. Once BALF was collected, lungs were pressure inflated, fixed in 10% formalin, and embedded in paraffin. Then, 5-μm lung sections were used for immunostaining and further histological analysis. Day 7 postbleomycin was selected for assessment of acute lung injury markers [albumin leakage, myeloperoxidase (MPO), and C-C motif chemokine ligand 2 (CCL2)], as this time point precedes the fibrotic phase and corresponds to peak barrier dysfunction and neutrophil infiltration, enabling a clearer evaluation of early injury mechanisms. Although CCL2 was elevated, neutrophil recruitment is more typically driven by CXCL1/2, and we did not perform neutralization studies to directly test CCL2’s role in this setting. Therefore, we interpret CCL2 elevation as part of the overall inflammatory response, not as a direct mechanism for neutrophil recruitment.

Immunohistochemistry and Immunofluorescence

Immunohistochemistry was performed on 5-μm sections from formalin-fixed, paraffin-embedded lungs. Antigen retrieval was performed on rehydrated sections using Tris-based Antigen Unmasking Solution (Vector Laboratories) followed by blocking of endogenous avidin and biotin using the Biotin-Blocking system (Vector Laboratories, Burlingame, CA), Bloxall (Vector Laboratories) for 30 min, and normal horse serum (Vector Laboratories) for 1 h at room temperature. Nuclear permeabilization with 1% Triton-X 100 was also carried out for 10 min after antigen retrieval and before blocking. Slides for immunohistochemistry were incubated with anti-smooth muscle actin (anti-SMA; 1:1,000 mouse monoclonal 1A4 clone; Sigma-Aldrich cat. no. A2547; Research Resource Identifier (RRID):AB_476701], anti-CD68 (1:500 mouse recombinant monoclonal; Abcam cat. no. ab955; RRID: AB_307338), and anti-HIF1A (1:250; Novus cat. no. NB100–105; RRID:AB_10001154) overnight at 4°C and developed the following day with the ImmPRESSTM-AP Anti-Rabbit IgG (Alkaline Phosphatase) Polymer Detection Kit (Vector Laboratories) before dehydration and mounting with coverslip. For immunofluorescence staining of alpha-smooth muscle actin (SMA), slides were incubated with Cy3-conjugated mouse monoclonal anti-SMA (1:1,000; Sigma-Aldrich cat. no. C6198; RRID:AB_476856) overnight at 4°C before mounting with Vectashield Antifade Mounting Medium with DAPI (Vector Laboratories). Dual immunohistochemistry (IHC) for HIF1A and CD14 in human samples was performed using anti-HIF1A (1:100; Cell Signaling Technology cat. no. 48085; RRID:AB_3698324) and anti-CD14 1:250 (Thermo Fisher Scientific cat. no. 14-0149-82; RRID:AB_467129) overnight at 4°C and developed the following day with the ImmPRESSTM-AP Anti-Rabbit IgG (Hydrogen Peroxide or Alkaline Phosphatase) Polymer Detection Kits (Vector Laboratories).

Assessment of Fibrosis

Masson’s trichrome collagen staining was performed according to the manufacturer’s instructions on rehydrated lung sections (EM Science, Gibbstown, NJ). A modified Ashcroft scoring method was used to quantify fibrosis among stained slides (23). Slides were blinded to reviewers with respect to the treatment group. Whole lung RNA was used to quantify collagen transcripts, and whole lung protein lysates were utilized for determining fibronectin expression.

Western Blot Analysis

Mouse lung tissues were pulverized, or cells from bone marrow-derived macrophages in culture were lysed in RIPA lysis buffer (50 mM Tris·HCl pH 7.4, 150 mM NaCl, and 1% NP-40) containing a protease inhibitor cocktail to extract protein (Thermo Fisher Scientific, Waltham, MA). Equal amounts of protein were separated on SDS-PAGE and transferred to PVDF membranes (GE Healthcare, Life Sciences, Pittsburgh, PA). Membranes were blocked for 1 h at room temperature with 5% nonfat milk and then washed with Tris-buffered saline-Tween-20. Mouse samples were incubated with primary HIF1A (1:500; GeneTex cat. no. GTX127309; RRID: AB_2616089); adenosine A_2B receptor [ADORA2B; 1:1,000 rabbit polyclonal; LSBio (LifeSpan) cat. no. LS-C20310–50; RRID: AB_896237]; CD206 (1:250 goat polyclonal; R&D Systems cat. no. AF2535; RRID:AB_2063012); fibronectin (1:1,000 rabbit polyclonal; Sigma-Aldrich cat. no. SAB4500974; RRID:AB_10744833); collagen I (1:1,000 rabbit polyclonal; Abcam cat. no. ab21286; RRID:AB_446161); and β-actin (1:1,000 rabbit polyclonal; Cell Signaling Technology cat. no. 4967; RRID:AB_330288) overnight at 4°C. Human samples were incubated with HIF1A (1:1,000; GeneTex cat. no. GTX127309; RRID:AB_2616089); ADORA2B (1:1,000 rabbit polyclonal; Abcam cat. no. ab229671; RRID:AB_3661889); neutrophil elastase (1:1,000; Cell Signaling Technology cat. no. 89241; RRID:AB_2800136); CD206 [1:1,000;Abcam cat. ab64693–1001)]; MPO (1:500; Thermo Fisher Scientific cat. no. MA5–15480; RRID:AB_10987565), inducible nitric oxide synthase (iNOS; 1:500; Abcam cat. no. ab15323; RRID:AB_301857), CD73 (1:2,000; Abcam cat. no. ab133582; RRID:AB_3674653), and GAPDH (1:2,000; Cell Signaling Technology cat. no. 2118; RRID:AB_561053) overnight at 4°C. Washed membranes were incubated with corresponding secondary antibodies conjugated to horseradish peroxidase (Cell Signaling Technology, Danvers, MA) for 1 h at room temperature and developed with Amersham ECL Prime Western Blotting Direction Agent (GE Healthcare, Life Sciences) or Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific).

Analysis of Whole Lung and BAL Cell Pellet RNA

RNA was isolated from frozen lung tissue, bronchoalveolar lavage cell pellet, or macrophage cell lysate using TRIzol (Invitrogen) as per the manufacturer’s instructions. cDNA was synthesized using the iScriptTM cDNA synthesis kit (Bio-Rad), and equal amounts were analyzed for transcript levels of COL1A1, COL1A2, ARG-1, ADORA2B, nitric oxide synthase 2 (NOS2), IL-6, and CXCL1 with normalization to beta actin or 18S rRNA. Primers for COL1A1 were forward 5′-GCTCCTCTTAGGGGCCACT-3′ and reverse 5′-CCACGTCTCACCATTGGGG-3′. Primers for COL1A2 were forward 5′-AAGGGTGCTACTGGACTCCC-3′ and reverse 5′-TTGTTACCGGATTCTCCTTTGG-3′. Primers for ARG-1 were forward 5′-CAGAAGAATGGAAGAGTCAG-3′ and reverse 5′-CAGATATGCAGGGAGTCACC-3′. Primers for ADORA2B were forward 5′-GCGTCCCGCTCAGGTATAAAG-3′ and reverse 5′-CGGAGTCAATCCAATGCCAAAG-3′. Primers for NOS2 were forward 5′-GTTCTCAGCCCAACAATACAAGA-3′ and reverse 5′-GTGGACGGGTCGATGTCA-3′. Primers for IL-6 were forward 5′-TAGTCCTTCCTACCCCAATTTCC-3′ and reverse 5′-TTGGTCCTTAGCCACTCCTTC-3′. Primers for CXCL1 were forward 5′-CTGCACCCAAACCGAAGTC-3′ and reverse 5′-AGCTTCAGGGTCAAGGCAAG-3′. Primers for CCL2 were forward 5′-TCAGCCAGATGCAGTTAACGC-3′ and reverse 5′-TGATCCTCTTGTAGCTCTCCAGC-3′. Next, the comparative cycle threshold (Ct) method was used to present data as mean normalized transcript levels.

Quantification of Protein in Bronchoalveolar Lavage Fluid

IL-6 levels in BALF from C57BL/6 and ADORA2Bf/fLysMCre mice after bleomycin or PBS treatment were quantified using the murine IL-6 ultrasensitive kit (Meso Scale Discovery) and read on the Sector Imager 2400 (Meso Scale Discovery) per manufacturer’s protocol. The Mouse ELISA Kit from Immunology Consultants Laboratory was used to quantify albumin levels, and Hycult Biotech ELISA Kit was used to assess MPO levels in BALF from HIF1Af/fLysMCre and LysMCre mice after PBS or BLM exposure on days 3 and 7.

Statistical Analysis

Experiments were repeated at least two times with biological and technical duplicates at a minimum. One-way ANOVA was used for comparisons among multiple groups. An unpaired, two-tailed Student’s t test was used for comparisons between two groups. A P value of less than or equal to 0.05 was considered statistically significant. All statistical analyses were performed with GraphPad Prism 6.0 software (La Jolla, CA). Sample sizes (n) are reported in each figure legend. For mouse studies, n refers to individual mice per group per time point, and results were typically reproduced in at least two independent experiments. For bone marrow-derived macrophage (BMDM) assays, n refers to independent primary macrophage preparations derived from different mice; results were confirmed ingreater or less than three independent experiments. For human studies, n corresponds to individual donor lungs analyzed. Technical replicates were averaged to yield a single value per biological replicate.

RESULTS

Elevated Hypoxic-Adenosinergic Axis and Inflammatory Cell Markers in NR-COVID-19 versus IPF

Using lung samples from control, NR-COVID-19, and IPF, we first assessed whether the hypoxic-adenosinergic response was upregulated in NR-COVID-19. These studies revealed increased expression of HIF1A, ADORA2B, and CD73, which contains a hypoxia response element (HRE) (24) in NR-COVID-19 and IPF compared to control samples (Fig. 1, A–D). Comparisons between NR-COVID-19 versus IPF revealed increased HIF-1A and ADORA2B in NR-COVID-19 versus IPF but no changes in CD73 (Fig. 1, A–D). Next, neutrophil elastase (N.E.), a marker for neutrophil presence in the lungs, demonstrated reduced expression in IPF lung samples versus control, with a variable expression in NR-COVID-19 samples (Fig. 1, A and D). Similarly, myeloperoxidase (MPO) levels, a marker for neutrophil activation, revealed reduced expression in both NR-COVID-19 and IPF versus control samples (Fig. 1, A and E). Markers for alternatively activated macrophages (AAM) (CD206) or classically activated macrophages [inducible nitric oxide synthase (iNOS)] did not show differences among control, NR-COVID-19, or IPF samples (Fig. 1, A, F, and G). These results demonstrate that the hypoxic-adenosinergic response is elevated in both NR-COVID-19 and IPF. Yet inflammatory cell recruitment is mixed, with reduced markers for neutrophil activation with no discernible difference in the expression of CD206 or iNOS.

Figure 1.

Elevated hypoxia-inducible factor 1A (HIF1A) and adenosine A_2B receptor (ADORA2B) in nonresolvable (NR)-COVID-19. A–H: Western blot for HIF1A, ADORA2B, CD73, neutrophil elastase (N.E.), myeloperoxidase (MPO), CD206, inducible nitric oxide synthase (iNOS), and GAPDH from control (white squares), NR-COVID-19 (black triangles), and idiopathic pulmonary fibrosis (IPF; black circles) lung samples (A) and densitometries for HIF1A (B), ADORA2B (C), CD73 (D), N.E. (E), MPO (F), CD206 (G), and iNOS (H). All data were checked for morality using the Shapiro-Wilk test. Significance levels ***P < 0.001 and *P < 0.05 refer to ANOVA comparisons with the Benjamini, Krieger, and Yekutieli multiple comparison test. M denotes a molecular marker. Lane 1 (*) was excluded due to death from anoxia, which causes artificial HIF1A stabilization and subsequent increased ADORA2B and CD73 expression. Control: n = 3; NR-COVID-19: n = 6; IPF: n = 4; a.u., arbitrary units; nd, not determined; ns, not significant.

Localizing HIF-1A Expression

We next aimed to investigate the regional expression of HIF-1A in the lungs of NR-COVID-19 individuals. In line with previous studies in IPF, we demonstrate increased signals for HIF1A in areas rich in smooth muscle actin (SMA), denoting the presence of activated fibroblasts, myofibroblasts in NR-COVID-19 lungs (Fig. 2A). Yet, dual staining for CD14 (a marker for macrophages) and HIF1A revealed a lack of dual IHC signals for HIF1A and CD14 in NR-COVID-19 patients who developed fulminant lung fibrosis, yet dual HIF1A/CD14 signals were detected in lungs from ARDS-COVID-19 individuals (Fig. 2, B–D). CD14 was selected since this marker is associated with circulating monocytes/macrophages (25), which provide a better link to our myeloid studies in vivo. These results demonstrate that despite the increased expression of both HIF1A and ADORA2B in NR-COVID-19, HIF1A expression was not observed in CD14-expressing cells in COVID-19 associated with a fulminant presentation of lung fibrosis, namely NR-COVID-19.

Figure 2.

Hypoxia-inducible factor 1A (HIF1A) localization in nonresolvable (NR)-COVID-19 acute respiratory distress syndrome (ARDS)-COVID-19. A: dual immunohistochemistry (IHC) for smooth muscle actin (SMA; magenta signals) and HIF1A (blue signals) from 3 representative NR-COVID-19 lung sections. B–D: dual IHC for the macrophage marker: CD14 (magenta signals) and HIF1A (red signals) from 3 representative NR-COVID-19 lungs (B) or 2 ARDS-COVID-19 lungs (C) and corresponding morphometric evaluation denoting a ratio of dual positive CD14/HIF1A cells relative to CD14 positive cells only (D). Scale bar = 50 μm. All data were checked for normality using the Shapiro-Wilks test. Significance levels **0.001 < P < 0.01 refer to a Mann-Whitney comparison between NR-COVID-19 (n = 7) vs. ARDS-COVID-19 samples (n = 5).

Myeloid-Specific HIF1A Deletion Alters Temporal Changes in Cellular Inflammation and Increases Neutrophil Infiltration after BLM Exposure

To investigate the effects of HF1A deletion from myeloid cells that include macrophages, we generated HIF1A^f/fLysM^Cre mice. These mice and control LysM^Cre mice were treated with intraperitoneal bleomycin (BLM) over 33 days, with samples collected at days 3, 7, 14, and 33 post-BLM or -PBS treatment. BLM-treated LysMCre mice presented with increased total inflammatory cell infiltration starting on day 14 of BLM exposure, remaining elevated on days 20 and 33 (Fig. 3A). HIF1A^f/fLysM^Cre mice presented with reduced total cell counts compared to BLM-LysM^Cre mice on day 20 of BLM treatment and no differences were seen among PBS-treated mice (Fig. 3A). Differential cell counts revealed increased macrophages in BLM-treated Lysm^Cre mice compared to the PBS Lysm^Cre group starting on day 14 that persisted on days 20 and 33 (Fig. 3B). Deletion of HIF1A in myeloid cells reduced macrophages on days 14 and 20 in BLM-treated mice (Fig. 3B). No differences among PBS-treated mice were observed (Fig. 3B). Assessment of neutrophils in BLM-treated LysMCre mice revealed increased numbers on day 14 after BLM (Fig. 3C). Deletion of HIF1A from myeloid cells promoted neutrophil infiltration on day 7, yet by day 14, no differences were seen among BLM-treated groups (Fig. 3C). Lymphocyte cell numbers revealed cell infiltration on day 14 in BLM-treated Lysm^Cre compared to PBS Lysm^Cre mice (Fig. 3D). No differences were seen in comparisons with myeloid HIF1A deletion among BLM or PBS groups (Fig. 3C). These results demonstrate that deletion of myeloid HIF1A resulted in a reduction of inflammatory cell infiltration dominated by a reduction in macrophages in bronchoalveolar lavage fluid (BALF) in BLM-treated mice. These BALF-focused readouts anchor our interpretation to myeloid compartments, consistent with the LysMCre targeting.

Figure 3.

Temporal changes in cellular infiltration after bleomycin (BLM) exposure. A–D: total cell counts (A), macrophages (B), neutrophils (C), and lymphocytes (D) present in bronchoalveolar lavage fluid (BALF) from hypoxia-inducible factor 1A (HIF1A)^f/f/LysM^Cre and LysM^Cre mice after intraperitoneal PBS or bleomycin treatment for days (D) 3, 7, 14, 20, or 33. Results are presented as mean ± SE. All data were checked for normality using the Shapiro-Wilks test. Significance levels ***P < 0.001, **0.001 < P < 0.01, and *P < 0.05 refer to ANOVA followed by a Bonferroni post hoc test. Comparison between LysM^Cre + PBS and LysM^Cre + bleomycin: ####P < 0.001, ##0.001 < P < 0.01, and #P < 0.05 by ANOVA followed by a Bonferroni post hoc test comparison between LysM^Cre + bleomycin and HIF1A^f/fLysM^Cre + bleomycin. LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 5; LysM^Cre + BLM: n = 4; HIF1A^f/f LysM^Cre + BLM: n = 6.

To further identify alterations in neutrophil infiltration on day 7, lung sections were stained for Ly.6b expression, a surface marker found on murine neutrophils (Supplemental Fig. S1A) (26). Semi-quantification of Ly.6b-positive cells shows increased neutrophils in BLM-treated HIF1A^f/fLysM^Cre mice compared to LysM^Cre mice (Supplemental Fig. S1B).

HIF1A Deletion on Myeloid Cells Exacerbates Acute Lung Injury after BLM Exposure

Macrophage activation during early lung injury could release proinflammatory cytokines and cause epithelial cell damage. We next determined how the loss of myeloid HIF1A expression altered pulmonary barrier function and the development of pulmonary edema. Herein, we evaluated albumin levels in the BALF on day 7 and found significant increases in HIF1A^f/fLysM^Cre mice after bleomycin exposure as compared to control mice with bleomycin treatment (Fig. 4A). To further evaluate chemokines involved in neutrophil recruitment in the setting of acute lung injury, we evaluated mRNA levels of C-C motif chemokine ligand 2 [CCL2; also known as monocyte chemoattractant protein-1 (MCP-1)] and found increases in control bleomycin-treated LysM^Cre mice which were robustly increased after bleomycin treatment in mice lacking HIF1A on myeloid cells (Fig. 4B) (27–29). Similarly, another indicator of acute lung injury, myeloperoxidase (MPO), a major component of neutrophil cytoplasmic granules revealing neutrophil presence, was found to be significantly increased in BALF on day 7 after bleomycin exposure in HIF1A^f/fLysM^Cre mice (Fig. 4C) (30). These findings reveal that myeloid-specific HIF1A deletion followed by acute lung injury through BLM treatment will not only increase neutrophil count but also contribute to worsening pulmonary edema, greater neutrophil recruitment, and ultimately, increased inflammation, thereby suggesting a protective role for HIF1A in acute lung injury.

Figure 4.

Deletion of hypoxia-inducible factor 1A (HIF1A) on myeloid cells exacerbates acute lung injury markers in bleomycin (BLM)-induced lung fibrosis. A–C: albumin concentration in bronchoalveolar lavage fluid (A), C-C motif chemokine ligand 2 (CCL2) mRNA expression in whole lung lysate (B), and myeloperoxidase (MPO) activity (C) in bronchoalveolar lavage fluid after bleomycin exposure at day 7 in HIF1A^f/fLysM^Cre and LysM^Cre mice. Individual values are presented with mean ± SD; n = 4–8. All data were checked for normality using the Shapiro-Wilks test. Significance levels ***P < 0.001, **0.001 < P < 0.01, and *P < 0.05 refer to one-way ANOVA comparisons with the Šídák’s multiple comparisons test. CCL2 mRNA expression assessed using real-time PCR. Data are presented as ratio to 18S RNA. For A: LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 3; LysM^Cre + BLM: n = 3; snf HIF1A^f/fLysM^Cre + BLM: n = 4. For B: LysM^Cre + PBS: n = 5; HIF1A^f/fLysM^Cre + PBS: n = 6; LysM^Cre + BLM: n = 6; and HIF1A^f/fLysM^Cre + BLM: n = 8. For C: LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 3; LysM^Cre + BLM: n = 3; HIF1A^f/fLysM^Cre + BLM: n = 4; ns, not significant.

HIF1A Deletion on Myeloid Cells Increases Pulmonary Fibrosis

Thus far, our data reveal an anti-inflammatory role for myeloid HIF1A in the acute stages of bleomycin exposure as evidenced by increases in acute inflammatory mediators, neutrophil recruitment, count, MPO, and loss of barrier function in bleomycin-treated myeloid-specific HIF1A knockout mice. This worsening of acute lung injury due to loss of HIF1A on myeloid cells led us to investigate whether pulmonary fibrosis would also be exacerbated, as has been shown previously (31). LysM^Cre control mice and HIF1A^f/fLysM^Cre mice were treated with BLM intraperitoneally biweekly for 4 wk. Histological assessment of fibrosis was performed using Masson’s Trichrome staining and quantified using Ashcroft scores. This analysis revealed an increased fibrotic injury in LysM^Cre mice after BLM treatment on day 14 and continued to increase by days 21 and 33 (Fig. 5, A and B). Mice lacking myeloid HIF1A presented with increased fibrotic injury assessed by Ashcroft scores on days 14 and 33 post-BLM, compared to LysMCre controls (Fig. 5B). We note that earlier time points (days 0 and 7) primarily reflect the inflammatory/acute injury phase. Therefore, our fibrosis analysis was focused on day 14 onwards, when histological, lung function, and genetic evidence of fibrosis were present (23). Further assessment of fibrosis, using immunofluorescence for α-smooth muscle actin (SMA), a marker of myofibroblasts in lung sections, revealed increased signals in BLM-treated control mice as compared to PBS, which was further increased in HIF1A^f/fLysM^Cre mice receiving BLM in association with reductions in arterial oxygen saturation (Fig. 6, A and B). Assessing fibrotic markers in whole lung lysate on day 33 revealed increased collagen 1a1 transcript levels in mice lacking HIF1A on myeloid cells as compared to bleomycin treatment of control mice alone (Fig. 6C). This was consistent with increased fibronectin protein expression in BLM-treated HIF1A^f/fLysM^Cre compared to PBS-treated LysM^Cre mice (Fig. 6, D and E). Since AAM have been implicated in promoting lung fibrosis (20), we next evaluated ADORA2B expression and markers for AAM in myeloid-HIF1A deficient mice. Herein, we report that CD206 protein expression was increased with day 33 BLM-treated LysM^Cre mice as compared to PBS treated but reduced in those bleomycin-treated mice lacking HIF1A on myeloid cells (Fig. 6, E and F). In line with this, transcript levels in lung lysates for arginase-1 were increased with BLM-treated LysM^Cre mice but reduced in HIF1A^f/fLysM^Cre mice (Fig. 6G). Our previous experiments have shown that activation of ADORA2B on myeloid cells upregulates profibrotic AAMs and profibrotic mediators and contributes to the development of pulmonary fibrosis (32). We have also shown that HIF1A modulated ADORA2B expression on BMDMs and reduced experimental lung fibrosis (20). In agreement with these findings, ADORA2B transcript levels were increased after bleomycin exposure in LysM^Cre mice yet reduced in BLM-treated mice lacking HIF1A on myeloid cells (Fig. 6H). Together, these data reveal that myeloid-specific HIF1A deletion leads to worsened pulmonary fibrosis, preceded by elevated markers of acute lung injury, and is associated with reduced expression of alternative activation markers and ADORA2B following bleomycin exposure.

Figure 5.

Myeloid hypoxia-inducible factor 1A (HIF1A) deletion exacerbates the development of pulmonary fibrosis after bleomycin (BLM) treatment. A and B: Masson’s trichrome staining in lung sections (A) with quantitative, fibrotic histologic scores as given by Ashcroft scores (B) in HIF1A^f/fLysM^Cre and LysM^Cre mice after intraperitoneal PBS or bleomycin exposure on days (D) 14, 20, and 33. Results are presented as mean ± SE. ***P < 0.001 LysM^Cre + PBS vs. LysM^Cre + bleomycin and ###P < 0.001, ANOVA comparison with the Bonferroni correction between LysM^Cre + bleomycin and HIF1A^f/fLysM^Cre + bleomycin. Scale bar = 100 μm. For B: LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 5; LysM^Cre + BLM: n = 4); HIF1A^f/fLysM^Cre + BLM: n = 6.

Figure 6.

Fibrotic markers are increased on day 33 after bleomycin (BLM) exposure in mice lacking hypoxia-inducible factor 1A (HIF1A) on myeloid cells. A and B: immunofluorescence for alpha-smooth muscle actin staining (magenta signals) and counterstained for DAPI (blue signals) from formalin-fixed paraffin-embedded (FFPE) lung sections (A) and arterial oxygen saturation (B) in HIF1A^f/fLysM^Cre and LysM^Cre mice after 33 days of bleomycin (BLM) or control (PBS) treatment. Images are representative of n ≥ 4 animals from each group. Scale bar: 100 μm (×10). C–E: collagen 1a1 (Col1a1) mRNA (C), densitometry analysis for gibronectin (FN; D) and corresponding Immunoblot for FN, CD206, and βActin (E). Lanes 1–2: LysM^Cre + PBS; lanes 3–4: HIF1A^f/fLysM^Cre + PBS; lanes 5–8: LysM^Cre + BLM; lanes 9–12: HIF1A^f/fLysM^Cre + BLM. F–H: corresponding densitometry analysis for CD206 (F) and expression levels for arginanase 1 (Arg1; G), and adenosine A_2B receptor (ADORA2B; H) Adora2b from HIF1A^f/fLysM^Cre and LysM^Cre mice after intraperitonael PBS or bleomycin exposure on day 33. Data are presented as mean ± SE in A and as individual values are presented with mean ± SD in C, D, F, G, and H; n = 4–8. All data were checked for normality using the Shapiro-Wilks test. In A, significance levels ***P < 0.001 and **0.001 < P < 0.01 refer to ANOVA followed by a Bonferroni post hoc test. Comparison between LysM^Cre + PBS and LysM^Cre + bleomycin: ###P < 0.001, #P < 0.05 by ANOVA followed by a Bonferroni post hoc test comparison between LysM^Cre + bleomycin and HIF1A^f/fLysM^Cre + bleomycin. Significance levels *P < 0.05 refer to Kruskal-Wallis test for C and D and one-way ANOVA comparisons with the Tukey multiple comparisons test for F, G, and H. For B: LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 5; LysM^Cre + BLM: n = 4; HIF1A^f/fLysM^Cre + BLM: n = 6. For C: LysM^Cre + PBS: n = 3; HIF1A^f/fLysM^Cre + PBS: n = 3; LysM^Cre + BLM: n = 3; HIF1A^f/fLysM^Cre + BLM: n = 3. For D and F: LysM^Cre + PBS: n = 2; HIF1A^f/fLysM^Cre + PBS: n = 2; LysM^Cre + BLM: n = 4; HIF1A^f/fLysM^Cre + BLM: n = 4. For G and H: LysM^Cre + PBS: n = 4; HIF1A^f/fLysM^Cre + PBS: n = 4; LysM^Cre + BLM: n = 4; HIF1A^f/fLysM^Cre + BLM: n = 5; A.U., arbitrary units; ns, not significant.

Myeloid-Specific HIF1A Deletion Attenuates ADORA2B Expression and AAM Polarization after BLM Exposure

HIF1A deletion in bone marrow-derived macrophages lowers arginase-1 and the production of profibrotic mediators.

In vivo, BLM treatment in myeloid-specific HIF1A conditional knockout mice exacerbated pulmonary inflammation and subsequent increases in fibrosis despite reductions in ADORA2B and AAM expression. To confirm the role of hypoxia through HIF1A in regulating macrophage differentiation (20), BMDMs from HIF1A^f/fLysM^Cre and LysM^Cre mice were cultured in the presence of IL-4 and IL-13 and exposed to hypoxia for 12 h. BMDMs were selected for these experiments as they provide higher cell yields and reproducibility compared to AMs, enabling parallel treatments under defined cytokine and hypoxic conditions. While AMs represent the in vivo resident macrophage population, the BMDM system is an established model for dissecting HIF1A- and adenosine-dependent pathway. Treatment of hypoxia and TH2-cytokine-treated LysM^Cre BMDMs with NECA, a panadenosine receptor agonist, resulted in increases in arginase-1, interleukin-6, and CXCL1 transcript levels, which were significantly reduced in BMDMs lacking HIF1A (Fig. 7, A–C). This suggests HIF1A on myeloid cells is needed for ADORA2B-mediated stimulation of macrophage differentiation into AAMs and profibrotic mediator production.

Figure 7.

Alternatively activated macrophages (AAM) and profibrotic mediator expression is reduced after adenosine receptor activation in bone marrow-derived macrophages (BMDMs) lacking hypoxia-inducible factor 1A (HIF1A). HIF1A^f/fLysM^Cre and LysM^Cre BMDMs were treated with IL-4 and IL-13 for 72 h and then incubated in hypoxia (2% O₂) for 12 h in the presence of 10 μM NECA. A–C: arginase-1 (A), interleukin-6 (B), and CXCL1 (C) mRNA expression was assessed using real-time PCR. Data presented as ratio to β-actin. Significance levels ****P < 0.0001, ***P < 0.001, and **0.001 < P < 0.01 refer to ANOVA followed by the Šidák’s multiple comparisons test; n = 4 per group and conditions; ns, not significant.

Reduced inflammatory markers in NR-COVID-19.

In line with BLM-treated HIF1A^f/fLysM^Cre mice, we report reduced expression of IL-6 and in NR-COVID-19 compared to controls (Fig. 8A). Assessment of Th1 markers revealed a reduced expression of IL1B and an upregulation of interferon-gamma (IFNG) (Fig. 8, B and C). No differences in the TH2 markers IL-4 and IL-3 were observed between control and NR-COVID-19 lungs (Fig. 8, D and E). Similarly, expression levels for the inflammatory markers CXCL12 and TNFA were not significantly different between control and NR-COVID-19 samples (Fig. 8, F and G). Assessment of genes associated with the metabolism of adenosine revealed reduced expression of adenosine deaminase (ADA), albeit no changes in adenosine kinase expression (ADK) in NR-COVID-19 versus control lungs (Fig. 8, H and I). These results point to a hypoinflammatory response in NR-COVID-19 lungs that mirrors our in vivo results in myeloid-HIF1A-deficient mice.

Figure 8.

Inflammatory markers are reduced in nonresolvable (NR)-COVID-19. Expression levels for IL-6 (A), IL-1B (B), interferon-gamma (IFNG; C), IL-4 (D), IL-13 (E), CXCL12 (F), TNFA (G), adenosine deaminase (ADA; H), and adenosine kinase expression (ADK; I) were quantified by quantitative (q)PCR from lung lysates obtained from control donors, (white squares, n = 13) and patients with nonresolving COVID-19 (NR-COVID-19; black triangles, n = 23). Gene expression was normalized to TATA box-binding protein (TBP) and is shown as relative expression levels. Statistical comparisons between groups were performed using unpaired two-tailed Student’s t test. Significance levels were **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant.

DISCUSSION

HIF1A has received recent attention as a transcription factor that can be protective in various types of acute lung injury, including normoxic ventilator-induced lung injury (VILI) and ovalbumin challenge models of airway hyperresponsiveness and asthma (33, 34). Most relevant to our study, activation of HIF1A has been demonstrated to reduce coronavirus infection and resulting injury (14). These and other findings have led to developing strategies to target HIF pharmacologically in ARDS (15). In line with these studies, we demonstrate that deletion of myeloid HIF1A exacerbated acute lung injury. Previously published studies by our group have demonstrated a pathogenic role for macrophage HIF1A and ADORA2B expression in lung fibrosis (20, 32). Contrary to these studies, our data show that genetic abrogation of myeloid HIF1A worsened lung fibrosis. An important factor that explains this discrepancy is timing; in our previous publication (20), pharmacological blockade of HIF1A was initiated once fibrotic lung injury was established, specifically on day 15 of BLM (20), as opposed to before injury, as in our current study. This is important given the well-known protective effects of HIF in acute lung disease (8, 14, 15). However, prolonged HIF1A activation and the resulting activation of the adenosinergic system are detrimental to chronic lung diseases (9, 20, 32, 35–39). Thus, deletion of HIF1A, once fibrosis is established, can attenuate the adenosinergic response that drives fibrotic processes (9, 40). This is supported by studies where deletion of ADORA2B, using the same LysM^Cre promoter as in the current study, reduced fibrotic injury in the IP BLM model (32). In the current study, the deletion of HIF1A before lung injury resulted in an upregulation of both an inflammatory at early stage and a fibrotic response at late stage. These findings highlight the importance of HIF stabilization as an essential system that can prevent “run-away” injury responses to the lungs. Furthermore, our results also shed light on the potential mechanism that promotes the development of fibrotic lung injury necessitating lung transplantation following viral infection, particularly SARS-COV2 (4, 41–44).

Our study integrates findings across human IPF, NR-COVID-19, and a chronic low-dose bleomycin model. While IPF lungs show increased HIF1A expression including in macrophages, NR-COVID-19 lungs uniquely exhibit loss of HIF1A in macrophages despite elevated tissue-wide HIF1A/ADORA2B. The bleomycin model provided a mechanistic bridge, demonstrating that conditional deletion of myeloid HIF1A recapitulates this loss, leading to early inflammation/barrier dysfunction and subsequent exacerbated fibrosis. Thus, IPF represents a chronic reference state, NR-COVID-19 highlights a fulminant fibrotic trajectory, and the bleomycin model mechanistically connects macrophage HIF1A deficiency to this accelerated injury-fibrosis sequence. These findings are further supported by our dual IHC analyses, where HIF1A and CD14 coexpression was absent in macrophages from NR-COVID-19 lungs with fulminant fibrosis, but was detectable in macrophages from ARDS-COVID-19 lungs. This observation underscores that the lack of HIF1A in macrophages is not a general feature of COVID-19 but is specifically associated with the fibrotic, nonresolving form of the disease.

Although the development of lung fibrosis following SARS-COV2 infection requiring lung transplantation represents a very small minority of cases, it is important to uncover the mechanisms that may lead to this fulminant presentation of lung fibrosis. This is important for two key factors: (1) we do not know how a history of COVID-19 may affect the future development of interstitial lung disease (1, 45–47), and (2) the development of fulminant lung fibrosis was extremely rapid and affected a much younger population (4) than other interstitial lung diseases that typically affect older individuals (5). Collectively, our data demonstrate that HIF1A stabilization is abolished from macrophages in the lungs of patients with fatal COVID-19 ARDS and NR-COVID-19, dominated by fibrotic lung injury (4). This aligns with our experimental results in vivo and demonstrates that myeloid deletion of HIF1A potentiated both the acute inflammatory response and fibrotic lung injury induced by intraperitoneal BLM.

Numerous pulmonary diseases are known to result in hypoxia, and the interplay between hypoxia and inflammation is well known (48). HIF1A is known to be stabilized in patients with idiopathic pulmonary fibrosis, in mouse models of BLM-induced lung injury, and in lung macrophages in fibrotic lung disease (37, 49, 50). In line with our previous studies linking HIF1A with ADORA2B expression (20, 32), we demonstrate that HIF1A deletion results in decreased markers of AAM, namely reduced CD206 and arginase-1, and ADORA2B transcript levels in whole lung lysates. This appears paradoxical given our prior work showing that myeloid deletion of ADORA2B reduced fibrosis (32). We interpret this difference as a reflection of timing and scope: HIF1A deletion occurs before injury and abolishes protective acute-phase functions, tipping the balance toward greater early injury and ultimately worse fibrosis. In contrast, direct deletion of ADORA2B primarily attenuates its profibrotic signaling once fibrosis is underway. Thus, while HIF1A loss reduces ADORA2B, it also removes additional macrophage programs that normally limit injury, explaining the divergent outcomes. The reduction in ADORA2B is consistent with prior studies identifying HIF1A as a positive regulator of ADORA2B expression in macrophages (20, 32). While AM-specific RNA profiling would provide more cell-restricted evidence, our results indicate that HIF1A deficiency broadly attenuates ADORA2B expression across lung myeloid compartments. These findings align with studies indicating that HIF1A inhibition and myeloid-specific HIF1A-deficient mice presented with reduced airway hyperresponsiveness and eosinophil recruitment in a model of allergen-induced airway inflammation (34). Supporting a similar anti-inflammatory role, Eckle et al. demonstrated that HIF1A stabilization in normoxia under mechanical stretch and VILI leads to decreased pulmonary edema and inflammation associated with maximal carbohydrate metabolism in vivo. These observations were exacerbated in alveolar-epithelial cell-specific HIF1A-deficient mice after VILI (33). These results agree with previous work from revealing an anti-inflammatory, protective role for ADORA2B in acute lung injury (51, 52). Intriguingly, our study also demonstrated that myeloid deletion of HIF1A led to the worsening of lung fibrosis; these results are surprising given that prior studies have demonstrated that deletion of macrophage HIF1A attenuated lung fibrosis (20, 53, 54). An important distinction between these studies and ours is that in all these cases, HIF1A deletion was initiated once lung injury was established and utilizing therapeutic strategies that may be able to target HIF1A in cells other than macrophages. In addition, our studies are also in line with previous observations where macrophage deletion exacerbated lung fibrosis (55, 56). AAMs typically function as key regulators of tissue repair and wound healing. However, in lung fibrosis, their activity shifts towards promoting excessive collagen deposition, ultimately driving the progression of fibrosis. Our finding aligns with previous studies showing that depletion of macrophages during the late resolution stage exaggerates lung fibrosis (57). Our study adds to the complexity of macrophages in the pathogenesis of lung fibrosis, where their role as a friend or foe has been postulated (58, 59). A possible molecular explanation for these findings may reside in previous studies where CD73−/− mice treated with BLM presented with reduced adenosine levels and increasing cellular infiltration, proinflammatory mediators including interleukin-1β, TNFα, and elevated lung fibrosis (31). These findings are significant since CD73 is known to have an HRE on its promoter (24). As such, it may be that myeloid-specific HIF1A deletion leads to decreased levels of CD73, lowering local adenosine levels, which can exacerbate injury in the acute phase of a bleomycin-induced lung injury model (60).

It is important to acknowledge that LysMCre, while widely used for targeting myeloid cells, can also produce recombination in a subset of epithelial cells. However, studies using epithelial-specific Hif1a deletion (e.g., Sftpc-CreER, Scgb1a1-Cre) have already demonstrated that epithelial HIF1A supports barrier integrity and protects against acute lung injury (33, 34). These findings are consistent with our interpretation that the exacerbated injury and fibrosis in our model arise primarily from myeloid HIF1A deficiency, as supported by the BAL cellular profiles, reductions in macrophage-derived AAM markers, and mechanistic deficits in BMDMs. A limitation of our study is that only male mice were used. This choice was based on the greater reproducibility of the IP bleomycin model in males and the fact that our NR-COVID-19 cohort was predominantly male, allowing for translational alignment. Nonetheless, sex differences are recognized in acute lung injury (61), and future studies are needed to determine whether the protective role of myeloid HIF1A extends to females.

Collectively, our data point to reduced myeloid HIF1A as a potential mechanism that potentiates BLM-induced acute lung injury and worsens lung fibrosis. These observations are in line with reduced macrophage HIF1A signals in NR-COVID-19 and point to a potential molecular mechanism responsible for the fulminant presentation of lung fibrosis following infection with SARS-COV2 (1, 2, 4, 42). These findings advocate for using HIF1A stabilizers; however, it is essential to recognize that prolonged HIF1A signaling can also promote the development of chronic lung injury.

Supplementary Material

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.30287317.

DATA AVAILABILITY

Data will be made available upon reasonable request.