6- to 8-week-old female C57BL/6 mice were purchased from Charles River Laboratories (PQ, Canada). Female IL-1α-deficient (Il1a^-/- ) mice were obtained from MiceCenter for Experimental Medicine and Systems Biology, University of Tokyo, Japan and bred in-house. Female IL-1R1-deficient (Il1r1^-/- ) mice (C57BL/6 background) with respective wild-type controls were obtained from The Jackson Laboratories (Maine, USA). Mice were subjected to a 12-hour light–dark cycle and had ad libitum access to food and water and housed under specific pathogen-free conditions. All experiments were approved by the Animal Research Ethics Board at McMaster University (#19-08-23).

Mice were exposed to twelve 3R4F reference cigarettes with filters removed (University of Kentucky, Lexington, USA) or room air (RA), twice daily for five days per week. Mice were exposed for up to 24-weeks using a whole-body CS exposure system SIU-48 (Promech Lab AB, Vintrie Sweden). Upon exposure completion, mice were euthanized by exsanguination and cardiac puncture.

Single cell suspensions were produced as follows. Mouse right middle, inferior, and post-caval lobes were enzymatically digested (150U/mL collagenase type I) for 1-hour shaking at 37°C. Bronchoalveolar lavage (BAL) was collected following 2x 500µL phosphate-buffered saline (PBS) washes of the single left lung lobe and BAL samples were then spun at 300 g for 5-minutes to pellet cells. To assess bone marrow, a single femur was removed and cells were flushed out using 5mL RPMI. Digested lung, flushed bone marrow, and harvested spleens were all crushed through 40µm mesh. Lung cells were treated with ACK lysis buffer (1mL, 1-minute) to lyse red blood cells. Approximately 80μL of blood was drawn using retro-orbital bleeding, lysed using 1x Red Blood Cell lysis buffer (#00-4333-57, eBioscience) and spun at 300 g for 5-minutes. All single cell suspensions were stained for flow cytometric analysis using antibodies shown in Table S1 . Gating strategies are shown in Figures S1A–D . All samples were run on a BD LSRFortessa (BD Biosciences, ON, Canada).

CS- or RA-exposed mice were intraperitoneally injected with bromodeoxyuridine (BrdU) for three consecutive days prior to sacrifice. 200μg on day -3 then subsequently 100μg on day -2 and -1 to assess macrophage turnover. BrdU incorporation into cells was assessed by flow cytometry.

Tissue remodelling was induced with a single intratracheal instillation of bleomycin (0.05U/mouse in a volume of 50µl sterile saline) by oropharyngeal administration. Control animals received 50µl vehicle alone. Weights of animals were monitored regularly, and tissue harvest was conducted 7 (early inflammatory phase) or 21 (fibrotic phase) days following bleomycin intubation. Lung function assessments were conducted using a flexiVent^® mechanical respirator according to manufacturer’s protocol at day 21 (flexiVent^®, SCIREQ, Montreal, PQ, Canada) (20).

CD45⁺ cells from lung single-cell suspensions were positively selected using mouse CD45 microbeads and LS columns (#130-052-301 and #130-042-401, Miltenyi Biotech) and cultured on a flat 96-well plate for 90-minutes. Non-adherent cells were stringently washed with PBS before remaining adherent cells were treated for 24-hours with recombinant murine transforming growth factor beta 1 (TGF-β1) (30ng/mL; BioLegend, USA), IL-4 (20ng/mL; PeproTech, Canada) and IL-6 (5ng/mL; PeproTech, Canada). Macrophage “M2” polarisation was assessed by arginase activity in cell lysates, as described previously (21). Matrix metallopeptidase 9 (MMP9; #DY6718 R&D Systems, ON, Canada) ELISA, and CyQUANT™ lactose dehydrogenase (LDH; #C20301 ThermoFisher Scientific, ON, Canada) cytotoxicity assay, and soluble collagen assessed using sircol assay (Sircol™ Soluble Collagen Assay #CLRS1000, Biocolor, UK) were performed on cell culture supernatant.

Lung (right superior lobe) homogenate was processed using Bullet Blender 24 Gold (Next Advance, Troy NY USA) either in PBS or RLT lysis buffer (Qiagen, Valencia, CA, USA) for protein and RNA analysis respectively. Protein from lung homogenate supernatant was assessed using the Discovery Assay^® Mouse Cytokine Array/Chemokine Array 31-Plex (MD31; Eve Technologies Corp, Calgary, AB, Canada). Total RNA from RLT lysed lung homogenate was extracted using the RNeasy mini kit (Qiagen, Valencia, CA, USA, #74104). Following RNA integrity and quantity quality control the nCounter Elements system (NanoString Technologies, Seattle, WA, USA) was employed to quantify the expression levels of 25 mouse genes ( Table S2 ).

Mouse single left lungs were inflated and fixed in 10% formalin for 24-hours prior to embedment in paraffin wax. A tissue microarray (TMA) was generated containing 5µm lung sections cut and stained on a Bond RX fully automated research Stainer (Leica Biosystems) (21). Immunohistochemistry (IHC) staining on lung serial sections was performed for hematoxylin & eosin (H&E), α-smooth muscle actin (α-SMA), and Masson’s trichrome blue (MTri). IHC stained-microscope slides were digitalised using an Olympus VS120-L100-W slide scanner at a 20× magnification. HALO™ Image Analysis Software (Halo Plus 3.2, Indica Labs) was used to quantify α-SMA and MTri stain in the lung parenchyma following the exclusion of airways and blood vessels using the Multiclass IHC (v3.0) and Classifier modules as previously described (22).

Results expressed as mean ± standard error of the mean. Graphs and statistical tests were performed using GraphPad Prism 9.1 (GraphPad Software, Inc) and R (www.r-project.org). Two-way ANOVA followed by Tukey’s multiple comparisons test was used to determine significance when two concurrent variables were compared. Unpaired t test with Welch’s correction was used to assess significance between only two groups. A p < 0.05 was considered statistically significant. Data obtained form the NanoString profiling were preprocessed and normalised using nSolver 2.5 software (www.nanostring.com) using three housekeeping genes Actb, Pgk1, and Ywhaz plus negative and positive controls. 25 mouse genes ( Table S2 ) were used to perform principal component analysis (PCA); using prcomp function from rgl package (23), hierarchical clustering (Euclidean distance, complete agglomeration method, using heatmap.2 function from gplots package (24) in R) and differential expression analysis using limma package (25) in R. P-values were adjusted using BH correction for multiple testing (26) and adjusted p < 0.05 was considered statistically significant.

CS is known to cause the expansion of the total lung macrophages (4, 5) but the impact on macrophage subpopulation diversity is not known. Using the nomenclature proposed by Gibbings et al. (17) ( Table 1 ), we demonstrated an expansion of all CD11b⁺ populations including Mo-AM and IM1, -2 and – 3 at all CS timepoints ( Figure 1A ). Res-AM were unchanged at 2 and 12 weeks of CS but decreased at 24-weeks. SiglecF cell surface expression was highest in Res-AM with minimal expression in IM populations. To note, CS exposure decreased siglecF expression in Res-AM suggesting a downregulation of cell-cell interactions ( Figure S2A ). The alterations in macrophage subpopulation numbers were not driven by changes in total lung cellularity but represent changes to each subpopulation individually ( Figure S2B ). Independent of CS exposure length, each CD11b⁺ macrophage subpopulation was increased in CS-exposed mice.

The size and location of macrophages has been associated with phagocytic and inflammatory differences (7). In this study, CD11b⁺ populations, Mo-AM (16.0%), IM1 (8.2%), and IM3 (7.1%), decreased in size at 12 weeks of CS ( Figure S2C ), corresponding to clinical observations of expanded small macrophages in COPD lung sections (7). We further assessed the presence of macrophage subpopulations in the BAL as a measure of macrophage location within the airways. While Res-AM decreased, each CD11b⁺ subpopulation, including all IM, were increased in the BAL of CS-exposed mice at 8 weeks of CS ( Figure 1B ). Clinically, airway macrophages are reported to be lipid-laden, and thus more granular (27). In our model, Res-AM became 197.9% and CD11b⁺ subpopulations 121.5-150.1% more granular than respective RA populations in the BAL ( Figure S2D ). Overall, we observed increased number of small, more granular, CD11b⁺ macrophages subpopulations following CS exposure.

To elucidate mechanisms contributing to increased CD11b⁺ macrophage populations, we assessed immune mediators associated with macrophage recruitment and survival in the lung homogenate. At 12 weeks of CS exposure myeloid chemoattractants CCL2 and CCL3 increased, as well as IL-1α ( Figure 2A ). IL-1α has previously been associated with CS-mediated myeloid cell recruitment (6, 28). Next, we measured immune mediators associated with monocyte differentiation and macrophage survival, M-CSF, GM-CSF and IL-6, all of which increased in the lung tissue of CS-exposed mice ( Figure 2B ). Immune mediators associated with macrophage recruitment and survival were increased in CS-exposed mice.

To understand pulmonary lung dynamics, we assessed macrophage subpopulation turnover using thymidine analogue BrdU. BrdU uptake can be used as a surrogate measure of cell turnover and the balance between local proliferation and cell recruitment. While the total Res-AM trended to decrease following CS exposure the percentage of Res-AM positive for BrdU increased ( Figure 2C ). Of note, the decreasing trend in total Res-AM was associated with increased cell death in CS-exposed mice as measured by flow cytometry Live/Dead positive staining ( Figure 2D ). Taken together, this suggests Res-AM proliferated in response to CS exposure to refill the environmental niche opened by increased Res-AM cell death. In contrast, CD11b⁺ populations had proportionally equal or fewer percentage BrdU⁺ cells, but increased total BrdU⁺ cells, following CS exposure ( Figures 2E–H ). The fixed percentage of BrdU positive cells indicated a stable rate of BrdU incorporation. Consequently, the increase in total BrdU⁺ Mo-AM and IM3 likely reflected a predominant recruitment of cells to the lungs. Total lung cell numbers were equivalent in RA and CS mice suggesting that changes were reflective of compositional shifts within the lung. Overall, CS was associated with Res-AM proliferation and a recruitment of Mo-AM and IM3 cells to the lung.

Given the expansion of CD11b⁺ populations including Mo-AM and IM3 in the lung, we next assessed the impact of CS on lung, blood, spleen, and bone marrow monocytes. In the lung, we observed that Ly6C^Lo monocyte populations, primed for macrophage differentiation, were not altered by CS exposure at 12-weeks. In contrast, classical inflammatory Ly6C^Hi monocytes were reduced by CS ( Figure 3A ). There was no change in either monocyte population in the circulation ( Figure 3B ). In the spleen, a known monocyte reservoir (29), CS resulted in decreased total spleen cells ( Figure S3A ) and consequently both monocyte populations ( Figure 3C ). CS exposure decreased both monocyte populations in the bone marrow at 12-weeks ( Figure 3D ). No changes in either monocyte population were observed in any tissue at 2- and 24-weeks CS exposure, with the exception of Ly6C^Lo monocytes, which were decreased following 2 weeks of CS exposure in the bone marrow ( Figures S3B–E ).

Next, we assessed macrophage progenitor cells in the bone marrow. The linearity of commitment toward cell terminal differentiation in order is Lineage^NegSca1⁺c-Kit⁺(LSK), monocyte-macrophage dendritic cell progenitor (MDP), common monocyte progenitor (cMoP) and finally monocyte populations (30–32). We observed all LSK progenitor cell populations were decreased in CS-exposed mice at 12 weeks ( Figures 3E , S3F ). To note, more committed progenitor MDP and cMoP populations were unchanged by CS exposure ( Figure 3E ). CS exposure was associated with a transient reduction in lung, spleen and bone marrow monocyte and macrophage progenitor populations at 12 weeks of CS exposure.

The IL-1α axis is critical in the recruitment of myeloid cells to the lung following CS exposure (6, 33, 34). Consequently, we assessed the impact of IL-1α and IL1-R1 deficiency on macrophage subpopulations composition in CS-exposed mice. Res-AM decreased in both Il1a^-/- and Il1r1^-/- mice following CS exposure ( Figures 4A, B ). Mo-AM and IM3, which were the predominant subpopulations increased in total BrdU⁺ cells, had attenuated expansion in Il1a^-/- mice following CS exposure ( Figure 4A ). There was no change in Mo-AM or IM3 populations in Il1r1^-/- mice following CS exposure ( Figure 4B ). Moreover, IL-1α- and IL1-R1-deficiency was not associated with changes in IM1 or IM2 populations ( Figures 4A, B ). Cell number changes were not driven by differences in total cell number ( Figures S4A, B ). Thus, CS was associated with IL-1α dependent expansion of Mo-AM, IM3, and Res-AMs.

Monocyte-derived macrophages, which contribute to the expanded CD11b⁺ macrophages in our model, have been shown to be necessary for fibrogenesis (9, 10, 19). Using a model of concurrent mild bleomycin-induced lung injury, we assessed the impact of CS on macrophage function at two timepoints, day 7 (pre-tissue remodelling and early inflammation) and day 21 (peak fibrogenesis and late inflammation) ( Figure 5A ) (35). While bleomycin-treated mice had equivalent weight loss regardless of exposure ( Figure S5A ), 3/5 (60%) and 4/5 (80%) myeloid lineage-related genes were enriched in CS-exposed compared to RA-exposed bleomycin-treated mice at day 7 and 21 respectively ( Tables S3 and S4 ). These transcriptional changes, including the upregulation of the CD11b-encoding gene Itgam, were likely related to the observed expansion of all CD11b⁺ subpopulations at day 7 in CS-exposed bleomycin-treated mice ( Figure 5B ). This compositional phenotype was not maintained in CS-exposed bleomycin-treated mice at day 21, with only IM2 remaining expanded ( Figure 5C ). In contrast, at day 21, Res-AM and Mo-AM populations were decreased and IM1/IM3 were equivalent in CS-exposed compared to RA-exposed bleomycin-treated mice ( Figure 5C ).

These alterations in macrophage composition at day 7 and 21 were associated impaired fibrogenesis by day 21 as determined by biomechanical, transcriptional and histological measurements. CS-exposed compared to RA-exposed bleomycin-treated mice had reduced total lung elastance (Est) but increased inspiratory capacity (IC) and lung area (A) ( Figure 5D ) at day 21. These physiological changes were associated with the downregulation of 5/9 (55%) fibrotic/wound healing genes at day 7 and 3/9 (33%) at day 21 in CS-exposed compared to RA-exposed bleomycin-treated mice ( Tables S3 , S4 ). Notably, overall transcriptional changes were sufficient to define experimental groups by unsupervised hierarchical clustering analysis at day 7 but not at day 21, where clustering was split based on CS status ( Figures S5B, C ). These altered transcriptional and lung biomechanics were associated with attenuated expansion of α-SMA⁺ cells in CS-exposed compared to RA-exposed mice ( Figure 5E ). Moreover, while the percent of collagenous area increased in RA-exposed bleomycin-treated mice compared to non-treated RA controls, collagen deposition was variable in CS-exposed bleomycin-treated mice and did not increase compared to controls ( Figure 5F ). Thus, trichrome staining suggests impaired collagen production in CS-exposed mice. Representative H&E stains are shown in ( Figure S5D ). Overall, these data suggest CS alters the transcriptional environment and macrophage composition at day 7 post bleomycin in a manner conductive to impaired fibrogenesis, as evidenced by decreased α-SMA⁺ myofibroblasts, elastance and collagen deposition at day 21.

Given that we observed concurrent changes in macrophage composition and fibrotic outcome, we next sought to address whether CS specifically alters macrophage fibrogenic phenotype and function. M2-polarised, alternatively activated macrophages are proposed to contribute to tissue remodelling and fibrogenesis (21, 36). Consequently, using CD38 as a marker for M1-like and CD206 for M2-like macrophages, we assessed the polarisation state of each macrophage subpopulation at day 7 and 21 post bleomycin. While polarised macrophages comprised less than 30% of each subpopulation, of those cells polarised, CD38⁺ macrophages formed the greatest proportion at both day 7 and 21. CD38 expression was increased in all macrophage subpopulations of CS-exposed bleomycin-treated mice at day 7 ( Figures 6A , S6A ). By day 21, polarised macrophages remained predominately CD38⁺ but strikingly CD206⁺ Res-AM decreased in CS-exposed compared to RA-exposed bleomycin treated mice ( Figures 6B , S6B ). At both day 7 and day 21 CS-induced dual CD38⁺CD206⁺ expression, reflecting previous dual polarisation observations in CS only exposure models (37) ( Figures 6A, B , S6A, B ). Of note, no difference in total cell number was observed between bleomycin-treated groups at either day 7 or 21 ( Figures S6C, D ). To explore the impact of CS on macrophage function further, adherent lung CD45⁺ cells isolated following 12-weeks of CS or RA exposure were stimulated for 24-hours with a profibrotic cytokine mix (TGF-β1, IL-4, and IL-6). Comprising predominantly pulmonary macrophages, but not excluding monocytes or dendritic cells, we observed no difference in cell viability between experimental groups ( Figure S6E ). Arginase activity (as measured by urea production), a surrogate measure of alternatively-activated macrophage function, was equivalent in RA or CS-exposed cells following stimulation and demonstrating no loss in M2-like functionality ( Figure 6C ). In addition, while adherent CD45⁺ cells produced soluble collagen, CS exposure did not alter production and therefore contribute to changes in collagen deposition ( Figure 6D ). To note, the profibrotic cytokine mix induced MMP9, a peptidase implicated in CS-mediated epithelial-mesenchymal transition and myofibroblast development in fibrogenesis, in RA-derived cells. ( Figure 6E ) (38). This induction was not observed in CS-exposed cells treated with, or without, cytokine mix suggesting CS was sufficient to attenuate MMP9 release. In summary, CS exposure elicited a shift toward a M1-dominant phenotype and decreased MMP9 release.