Alveolar macrophages drive lung fibroblast function in cocultures of IPF and normal patient samples

Abstract

Idiopathic pulmonary fibrosis (IPF) is characterized by increased collagen accumulation that is progressive and nonresolving. Although fibrosis progression may be regulated by fibroblasts and alveolar macrophage (AM) interactions, this cellular interplay has not been fully elucidated. To study AM-fibroblast interactions, cells were isolated from IPF and normal human lung tissue and cultured independently or together in direct 2-D coculture, direct 3-D coculture, indirect transwell, and in 3-D hydrogels. AM influence on fibroblast function was assessed by gene expression, cytokine/chemokine secretion, and hydrogel contractility. Normal AMs cultured in direct contact with fibroblasts downregulated extracellular matrix (ECM) gene expression whereas IPF AMs had little to no effect. Fibroblast contractility was assessed by encapsulating cocultures in 3-D collagen hydrogels and monitoring gel diameter over time. Both normal and IPF AMs reduced baseline contractility of normal fibroblasts but had little to no effect on IPF fibroblasts. When stimulated with Toll-like receptor (TLR) agonists, IPF AMs increased production of pro-inflammatory cytokines TNFα and IL-1β, compared with normal AMs. TLR ligand stimulation did not alter fibroblast contraction, but stimulation with exogenous TNFα and TGFβ did alter contraction. To determine if the observed changes required cell-to-cell contact, AM-conditioned media and transwell systems were utilized. Transwell culture showed decreased ECM gene expression changes compared with direct coculture and conditioned media from AMs did not alter fibroblast contraction regardless of disease state. Taken together, these data indicate that normal fibroblasts are more responsive to AM crosstalk, and that AM influence on fibroblast behavior depends on cell proximity.

INTRODUCTION

Pulmonary fibrosis occurs when there is failure to control the wound healing response resulting in the replacement of normal, functional tissue with fibrous scar tissue. Fibroblasts, the main effector cells in pulmonary fibrosis, respond to the insult by activating and initiating a wound repair response by laying down additional extracellular matrix (ECM) components and exerting force on the surrounding ECM to pull the wound together. This response is necessary to preserve lung function and maintain gas exchange; however, it can cause scar-like tissue formation characterized by excessive collagen accumulation, reduced lung elasticity, and decreased pulmonary function. A severe form of pulmonary fibrosis is idiopathic pulmonary fibrosis (IPF), which is a chronic, progressive disorder characterized by excessive ECM accumulation that is progressive and nonresolving. Although the cause of IPF is unknown, previous studies have determined men who smoke and are over the age of 65 are at the highest risk of developing the disease (1). Patients with IPF have a median survival of only 2–5 yr following diagnosis (2) and treatment options are limited to antifibrotic therapies, such as pirfenidone and nintedanib. These drug treatments do not reverse damage, but instead work to slow the progression of the disease. The only cure for patients with IPF is lung transplantation (3).

Although fibroblasts are the primary effector cell in IPF, they are not the only cell dysregulated in the setting of pulmonary fibrosis. The lung has a complex microenvironment comprised of a variety of structural and immune cells that are subjected to a variety of different mechanical forces. Macrophages, a phagocytic immune cell, are found in the alveolar space and in the interstitium of the lung. Circulating monocytes can also be recruited and differentiated into lung macrophages following injury or insult such as during pulmonary fibrosis. The primary function of macrophages is to monitor the lung microenvironment and act as the first line of defense against inhaled particles and pathogens. The ability of macrophages to respond to signals in the environment is mediated by pattern recognition receptors, such as Toll-like receptors (TLRs) that are expressed both extracellularly and intracellularly and work to identify conserved motifs present in a variety of pathogens, including viruses, fungus, bacteria, and parasites (4). Macrophages release and respond to cytokines, which drives their activation state and affects the functional behavior of both themselves and surrounding cells. Macrophages are capable of coordinating the inflammatory response through the release of both proinflammatory (IL-1β, TNFα, IL-6, IL-8) and anti-inflammatory cytokines (IL-4, IL-10, IL-13, IFN-α, TGFβ) (5). In a typical wound healing response, macrophages are critical to fibrosis resolution where their uptake of dead cells and excessive ECM helps to degrade the scar tissue and aid in resolving the injury (6). Although several studies have investigated the role of alveolar macrophages (AMs) in IPF (7, 8), few have investigated the combinatory role of AMs on fibroblast functions, such as contraction and gene expression of ECM components.

The impact of macrophage-fibroblast coculture has been extensively studied in the realm of cancer (9), foreign body response (10), and wound healing (11). Both direct and indirect coculture methods have demonstrated crosstalk and influence on gene expression and cytokine secretion that mediates both macrophage and fibroblast functions. Macrophages influence fibroblast gene expression based on their activation state, by augmenting expression of collagens, αSMA, and TGFβ (12) as well as ECM synthesis and fibroblast proliferation (13). Fibroblasts regulate the ability of macrophages to produce proinflammatory cytokines, such as IL-6 (12), and chemokines, such as macrophage inflammatory protein 1α (MIP-1α) (14). Recent work has demonstrated that alternatively activated AMs secrete S100a4, which promotes activation of lung fibroblasts via αSMA and collagen expression as well as proliferation (15). Although this previous work has been useful in determining baseline communication methods between fibroblasts and macrophages, previous studies focused on transcriptomic changes and did not evaluate how AMs alter the functional ability of fibroblasts. In addition, most studies utilized either cell lines or murine cells to generate these coculture studies.

To expand on these studies, we established a model to investigate macrophage-fibroblast interactions using primary human AMs and lung fibroblasts isolated from patients with IPF or normal, donor controls. These cells were cultured either directly or indirectly using coculture systems and assessed for changes in gene expression, functional contractility, and protein production to elucidate possible mechanism of cell-to-cell interactions. These data generated demonstrate how intracellular interactions can contribute to changes in ECM expression, fibroblast contractile function, and fibrosis progression within an in vitro model system.

METHODS

Cell Culture

The following supplies were utilized for tissue culture maintenance and cell isolation as specified in the following methods: RPMI tissue culture medium (Gibco, 11875119), fetal bovine serum (Avantor, 89510-186), antibiotic-antimycotic (Gibco, 15240096), human serum (Millipore Sigma, S1-100ML), EDTA (ThermoFisher, BM150), Trypan blue (Invitrogen, T10282), trypsin (Gibco, 15400054), bovine collagen (PureCol, 103700-630), NaOH, PBS, phenol red (abcam, ab146336), lipopolysaccharide (LPS;Sigma, L2880), Pam3CysSerLys4 (Pam3Cys, Invivogen, tlrl-pms), polyinosinic:polycytidylic acid (poly I:C; Invivogen, tlrl-pic), TGFβ (PeproTech, 100-21), IL-13 (PeproTech, 200-13), and TNFα (PeproTech, 300-01 A).

Normal patient lungs deemed not viable for transplantation and explant lungs of patients with IPF following transplantation were procured via a collaboration with the Comprehensive Transplant Center (CTC) Human Tissue Biorepository at The Ohio State University, (IRB protocol 2014H0367). Lungs were processed within 24 h of procurement. All samples collected from patient lungs as well as clinical data were deidentified and classified as either IPF or normal for experimental use. Patient sample demographics are provided in Supplemental Table S1. Alveolar macrophages (AMs) were isolated via ex vivo lavage by instilling ∼500 mL of PBS with EDTA into the bronchi and rinsing each lobe to recover the fluid and cells. AMs from IPF and normal, donor control lungs were centrifuged down and frozen at −80°C in CryoStor cryopreservation media (CS10, Stemcell technologies, 100–1061) at a concentration of 10e6–20e6 cells/mL then thawed for coculture experiments. Fibroblasts were harvested from patient samples by mechanical mincing of the lung tissue and plating on tissue culture plates for 21–28 days in RPMI cell culture medium containing, 1% antibiotic-antimycotic, and 15% FBS until confluent. Fibroblasts were then passaged every 3–4 days as necessary and maintained in either 10% or 15% FBS RPMI cell culture medium depending on rate of growth. Lung fibroblasts were cryopreserved in CryoStor cryopreservation media at a concentration of 1–5e6 cells/mL and stored in liquid nitrogen. All cells were used for these studies were below passage 8. Duration of cryopreservation varied depending on date of receiving the donor sample and date of experimental setup. Viability of thawed fibroblasts remained high (90%–100%) and recovery rate of thawed AMs was consistent between normal and IPF samples at ∼50%–60% (Supplemental Fig. S1D).

Cytokine and TLR Ligand Treatment

Macrophages were treated with TLR ligands at a concentration of 100 ng/mL (TLR1/2 = Pam3Cys, TLR3 = poly I:C, TLR 4 = LPS) on day 0. Cells and conditioned cell culture medium were collected after 1 day for analysis.

For fibroblast contraction studies the following doses were administered at days 0 and 4 after cell culture medium changes: 100 ng/mL of LPS, Pam3Cys, or poly I:C; 20 ng/mL of IL-13 or TNFα; 2 ng/mL TGFβ. Gene expression studies of fibroblasts subjected to cytokines or TLR ligands were performed on 2-D, 6-well, tissue culture plates at a concentration of 1e6 cells/well. Exogenous stimuli were placed in 15% FBS RPMI cell culture medium on day 0 at concentrations used for the contraction assay described above. After 24 h of stimulation, media was harvested and cells in culture wells were collected in TRIzol reagent for downstream RNA analysis.

In Vitro Macrophage-Fibroblasts Coculture Systems

AMs derived from human donor BAL were cocultured with primary human fibroblasts in direct contact on tissue culture plates, indirectly in transwell systems, and jointly in 3-D collagen hydrogels. Fibroblasts were first grown to 80% confluency and maintained in RPMI cell culture medium containing 15% FBS and 1% antibiotic-antimycotic. Alveolar macrophages (when cultured alone) were maintained in RPMI cell culture medium containing 10% human serum and 1% antibiotic-antimycotic.

Direct and Indirect Coculture of Fibroblasts and Alveolar Macrophages

Fibroblasts were cultured at 1e4 to 1e5 cells per well within 6-well plate for 24 h. AMs were thawed, washed, and measured for viability. They were then resuspended in cell culture media (RPMI with 10% FBS and 1% antibiotic-antimycotic) and plated in direct or indirect contact with fibroblast cultures at 1e6 cells per well. For indirect contact assays, macrophages were plated on 0.4 µm transwell culture inserts (Fisher). Cocultures were maintained for 1–3 days before cells were collected with TRIzol reagent for RNA analysis and conditioned media was collected for ELISA.

Colometric Assays

ELISA assays were obtained from R&D systems (Human DuoSet ELISA): TGFβ1 (DY240), TNFα (DY210-05), and CCL2 (DY27905) and utilized according to manufacturer’s protocols. TGFβ ELISAs required additional acidification steps required for protein activation which were carried out according to provided protocols. Cellular viability was determined using MTS assay from Promega (PAG5421) after 120–130 min of incubation with indicator solution according to the manufacturer’s protocols. Both assays were read using a Synergy H1 microplate reader.

Collagen Hydrogel Contraction Assay

Fibroblasts were collected from tissue culture plates, spun down, and resuspended in 15% FBS RPMI medium at a live cell concentration of 1.32e6 cells/mL. AMs were thawed and suspended in RPMI medium at a live cell concentration of 4e6 cells/mL. The collagen hydrogel components were prepared by combining 10× PBS with phenol red as pH indicator, sterile water, and NaOH which were well mixed before adding 333uL of 6% bovine collagen. After careful mixing to avoid the introduction of air bubbles, the cell suspensions in cell culture medium were added in single gel aliquots and mixed well. This provided a final collagen concentration of 2 mg/mL containing either fibroblasts alone at 0.33e6 cells/gel, AMs alone at 1e6 cells/gel, or fibroblasts and alveolar macrophages cocultured at a concentration of 0.33e6 and 1e6 respectively per 1 mL of gel. Optimum ratios of AM:fibroblast cultures were determined via concentration investigations, Supplemental Fig. S1, for significant influence in contraction and conservation of cell counts. Cell containing gel mixtures were then transferred to a 12-well plate (1 mL/well) and allowed to solidify in the cell culture incubator for a minimum of 45 min. Gels were then transferred to 6-well plates containing 3 mL of 15% FBS RPMI cell culture medium using a plastic spatula to detach the gel from the well. Cell encapsulated collagen hydrogels were then maintained for 7–10 days with daily monitoring of hydrogel diameter in mm. Cell culture medium was changed every 4 days and collected on days 4 and 7 for ELISA assays. Contraction assays were performed in technical replicates of 3 for each patient number and biological replicates of 4 or more.

Conditioned Media Contraction Assay

Cell culture RPMI medium containing either 10% human serum or 15% FBS was conditioned by alveolar macrophages cultured on 15 cm tissue culture plates at a concentration of 25e6 cells/20 mL of medium and incubated overnight. Cell culture medium was then collected, spun down, and used as culture medium on collagen encapsulated fibroblast hydrogels (0.25e6 cells/gel) for 7 days. Hydrogel diameter was measured daily, and cell culture medium was changed on day 4 and collected at days 0, 4, and 7 for ELISA assays. All conditions were repeated in technical replicates of 3 and biological replicates of 3 or more. Supplemental Fig. S3 shows the contraction profiles of fibroblast hydrogels cultured in AM-conditioned human serum media, which was not as significant as results obtained with FBS. Therefore, FBS was used in all contraction assays presented in the article figures.

Quantitative Real-Time PCR (qPCR)

RNA was extracted from cells by using a Direct-zol RNA miniPrep Plus Kit (Zymo Research, R2072) according to the manufacturer’s instruction. RNA was quantified using a NanoDrop 1000 Spectrophotometer (ThermoFisher). Synthesis of cDNA was performed using the RevertAid First Strand cDNA Synthesis Kit (ThermoFisher, K1691) according to product protocol, and gene expression was measured by qPCR using a QuantStudio 3 Real-Time PCR System with 40 cycles using PowerUP SYBR Green Master Mix (Applied Biosystems, A25742). Data were analyzed by the 2^−ΔΔCt method with housekeeping gene expression as the endogenous control. Primers used for qPCR analysis are included in Supplemental Table S2.

Flow Cytometry

Frozen ex vivo lavage cell samples from normal patients and patients with IPF were thawed for cell type quantification using flow cytometry. Cells were counted, stained for live/dead for 30 min using BV510 Aqua (Thermo Fisher L34966). They were then washed and incubated with FUN 2 CD32 blocking solution (Biolegend, 303202) for 20 min followed by the antibody panel listed in Supplemental Table S3. Cells were then fixed in 2% paraformaldehyde and analyzed on a Cytek Aurora spectral cytometer. Gating was based on previous literature (16, 17) using unstained cell samples as controls and performed in FlowJo with specific gating schema provided in Supplemental Fig. S2A.

Statistical Analysis

All data are expressed as means ± SEM and were analyzed using GraphPad Prism 9 statistical program. All data were assessed for outliers, analyzed for normality, and then assessed for statistical significance accordingly. If data were found to be normally distributed, then a one-way ANOVA was used. Alternatively, a two-way ANOVA was used depending on the number of independent variables. If data were not normally distributed, then a nonparametric one-way ANOVA was used. For lognormal distributions, data were log transformed before ordinary one- or two-way ANOVA was used.

Contraction assays were analyzed using two separate approaches. The first utilized the appropriate ANOVA to assess significant differences between conditions at each timepoint as well as between days. Second, a regression curve for one phase decay was fit to each sample, grouped in technical gel replicates (n = 3). The model equation used for best fit exponential decay is shown in Eq. 1.

$$Y=\left(Y_0-\mathrm{P}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{a}\mathrm{u}\right)\times \exp \left(-K\times X\right)+\mathrm{P}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{a}\mathrm{u}$$ (1)

where Y₀ is Y value at time 0 and was fixed at initial gel diameter 22 mm; plateau is y value at infinite time; K is decay constant; X is time in days. Using this curve fit function in GraphPad Prism 9, the decay constant and plateau was determined for each sample and compared between conditions using t tests or one-way ANOVAs as appropriate and specified in each figure.

RESULTS

IPF Fibroblasts Have Elevated Gene Expression and Contraction Rate Compared with Normal Fibroblasts

To compare the innate differences between IPF and normal fibroblasts, changes in gene expression and contraction were compared. IPF fibroblasts significantly upregulated expression of collagen 1, collagen 3, and αSMA compared with normal fibroblast (Fig. 1A). When fibroblasts were encapsulated in 3-D collagen hydrogels and monitored over 7 days for changes in gel diameter, there was initially significant contraction observed in both IPF and normal samples. There was a significant difference in the hydrogel diameter of normal versus IPF fibroblasts at days 5–7 (Fig. 1B). Best fit decay curves were utilized to assess the contraction profiles. There was a significant increase in the decay constants of IPF fibroblasts which indicates a faster rate of contraction compared with normal fibroblasts (Fig. 1C). There was enhanced spread in the decay constants of IPF fibroblasts compared with normal, which correlated with the high variability between patient samples. The plateau values of the regression curves were also calculated and showed no significant difference between disease states (Fig. 1D), which demonstrates that there is no difference in the final hydrogel size despite differences in contraction rates. Overall, these data demonstrate that IPF fibroblasts have increased expression of collagen and fibroblast activation makers as well as enhanced contractility rate compared with normal fibroblast controls.

Figure 1.

IPF fibroblasts display heightened gene expression and contractility compared with normal pulmonary fibroblasts. Gene expression changes were compared between normal and IPF human fibroblasts for collagen 1, collagen 3, fibronectin, and fibroblast activation marker αSMA (A). Hydrogel contraction was monitored over the course of 7 days for both IPF and normal fibroblasts (B; * significance between IPF and normal on the same day, % significance between normal sample and the previous day, and & significance between IPF sample and the previous day). Regression curves were fit to each sample’s 7-day contraction and decay constants for each curve were plotted for comparison of contraction rate between IPF and normal samples (C). The regression curve plateau values were plotted for both IPF and normal contraction (D). Statistical significance was determined through t test comparisons for gene expression changes, two-way ANOVA for contraction studies with Tukey’s multiple comparisons test, and t tests for comparisons of decay constants and plateau values. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. IPF, idiopathic pulmonary fibrosis.

Fibroblast-Macrophage Coculture Influences Gene Expression and Fibroblast Contractility

To study the influence of AM disease state on fibroblast phenotypes as well as their role in fibrosis progression, direct fibroblast-AM cocultures were assessed for gene expression changes. Normal AMs cocultured with either normal lung fibroblasts (Fig. 2A) or IPF lung fibroblasts (Fig. 2B) in a 6-well plate were found to significantly decrease expression of ECM genes collagen 1 and collagen 3. Normal AMs reduced expression of fibroblast activation marker αSMA only when incubated with normal, but not IPF, lung fibroblasts. Interestingly, there was a significant increase in the expression of fibronectin when normal AMs were cocultured with either normal or IPF fibroblasts (Fig. 2, A and B). Incubation of IPF AMs with normal fibroblasts resulted in a significant increase in collagen 3 expression (Fig. 2A). Although incubation of IPF AMs with IPF fibroblasts only resulted in a significant increase in expression of fibronectin, cocultured IPF AMs with IPF fibroblasts exhibited more collagen 1 and collagen 3 expression as compared with normal AMs with IPF fibroblasts (Fig. 2B). It should be highlighted that all of the coculture conditions were collected together, and inclusion of multiple cell types may influence the gene expression changes observed. Taken together these data suggest that AM disease state can regulate fibroblast gene expression.

Figure 2.

Fibroblast-macrophage coculture influences gene expression and contractility of fibroblasts. Normal fibroblasts were cultured alone or directly in contact with normal or IPF AMs and assessed for changes in gene expression for ECM components collagen 1 and 3 and fibronectin and fibroblast activation marker αSMA (A). IPF fibroblasts were cultured independently or in direct contact with normal (blue) or IPF (red) AMs and assessed for changes in gene expression (B). Normal fibroblasts were encapsulated independently (gray) or in conjunction with normal (blue) or IPF (red) AMs and monitored for contractile changes over the course of 7 days (C). Contractility of IPF fibroblasts was then measured independently (gray) or in hydrogel coculture with normal AMs (blue) or IPF AMs (red; D). Best fit curves for exponential decay were fit to all contraction studies and resulting decay constants were plotted for normal (E) and IPF (F) coculture contraction studies. Plateau values for resulting decay curves were also plotted for normal (G) and IPF (H) 7-day contractions. For gene expression changes, statistical significance was determined using one-way ANOVA with Tukey’s multiple comparisons when normally distributed (A) and using a nonparametric one-way ANOVA with Kruskal-Wallis test for multiple comparisons when nonnormally distributed (B). Contraction assays were assessed using two-way ANOVAs. Decay constants and plateau values were compared using ordinary one-way ANOVA and uncorrected Fisher’s LSD test for multiple comparisons. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). AMs, alveolar macrophages; ECM, extracellular membrane; IPF, idiopathic pulmonary fibrosis; LSD, least significant difference.

To extend these studies, we investigated the role of AMs in regulating fibroblast contraction within 3-D collagen hydrogels. The concentration of AM to fibroblast ratio was chosen based on a concentration response study shown in Supplemental Fig. S1. There were no changes in the diameter of the collagen gels when AMs from either normal patients or patients with IPF were encapsulated alone (gray lines in Supplemental Fig. S1). When normal AMs were cocultured together with normal fibroblasts in the collagen hydrogel, they significantly reduced fibroblast contraction compared with normal fibroblasts alone (Fig. 2C). Coculture of IPF AMs with normal lung fibroblasts resulted in a significant decrease in gel contraction early (days 1 and 2), but this trend did not continue and there was no significant difference in the diameter of the collagen hydrogel at day 7 (Fig. 2C). Incubation of normal AMs with IPF fibroblasts reduced contractility initially, at day 1, but increased contraction at day 7 (Fig. 2D). IPF AMs significantly influenced IPF fibroblast contractility only on days 6 and 7, showing enhanced contractility compared with controls (Fig. 2D).

To analyze this contraction data further, regression curves were fit to the contraction profiles for each unique patient sample. The decay constants and plateau values of these curves were compared between IPF and normal samples cultured alone and in direct coculture. The rate of contraction in the normal fibroblasts was significantly reduced when cocultured with AMs from both normal patients or patients with IPF, as evident by the smaller decay constants (Fig. 2E). Alternatively, AMs, regardless of their disease phenotype did not alter the decay constant of the IPF fibroblasts (Fig. 2F). The plateau value of the decay curves indicate the predicted end point diameter of the gels and when compared, AMs had no significant influence on this value for the normal (Fig. 2G) nor IPF (Fig. 2H) fibroblast cocultures. Taken together, these data demonstrate that fibroblast functional phenotypes can be significantly influenced by AM coculture. Normal fibroblasts are more responsive to coculture with AMs altering both gene expression and contraction profiles, whereas AMs regulated gene expression changes, but not contractility in IPF fibroblasts.

Stimulation of Alveolar Macrophages with TLR Ligands Alters Gene and Protein Expression of Cytokines and Chemokines

As stated in the material and methods section, all AMs were obtained by ex vivo bronchoalveolar lavage (BAL) of explant lungs. To verify there were similar cell populations present in the normal and IPF samples, a representative group of samples were analyzed for immune cell populations using flow cytometry. The various innate and adaptive immune cell populations and a representative flow cytometry panel is provided in Supplemental Fig. S2A. Based on our flow cytometry analysis, there was no significant difference between relative cell populations between IPF and normal samples (Fig. 3A, Supplemental Fig. S2C).

Figure 3.

Alveolar macrophages alter gene expression and protein production in response to TLR stimulation. Bronchoalveolar lavage samples from normal and samples of patients with IPF were stained and quantified for relative proportion of cell populations (A). Normal (B) and IPF (C) human AMs were treated with various TLR ligands including LPS, PAM3Cys, or poly I:C and assessed for gene expression changes compared with nonstimulated controls. Protein production of TNFα (D), MCP1 (E), IL-10 (F), and IL-1β (G) were assessed using ELISA. Significant changes in gene expression were assessed using nonparametric one-way ANOVA with Kruskal-Wallis test for multiple comparisons. Protein production was assessed using a nonparametric one-way ANOVA for either normal AMs under treatment or IPF AM treatments respectively and nonparametric t tests for treatment specific comparison of IPF to normal AMs. The * symbols indicate significance when compared with nonstimulated control from same disease or nondiseased condition (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). The & symbols indicate significance when comparing IPF vs. normal nondiseased samples under same treatment condition (&P < 0.05, &&P < 0.01, &&&P < 0.001, &&&&P < 0.0001). AMs, alveolar macrophages; IPF, idiopathic pulmonary fibrosis; TLR, Toll-like receptor.

When macrophages are activated, they produce a variety of cytokine and chemokines that can play roles in fibrogenesis and myofibroblast activation. Expression of cytokines and chemokines are regulated through TLR-dependent signaling pathways (18). Since lung macrophages from patients with IPF have an altered TLR expression profile compared with normal donor controls (Supplemental Fig. S4A), we investigated how normal and IPF AMs stimulated with TLR ligands altered pro- and anti-inflammatory cytokine and chemokine expression. Stimulation of both normal (Fig. 3B) and IPF (Fig. 3C) AMs with the TLR4 ligand, LPS, resulted in a significant increase in gene expression of IL-1β, IL-8, IL-6, and IL-10. Both IPF and normal AMs upregulated IL-8 in response to PAM3Cys, a TLR1/2 ligand (Fig. 3, B and C); however, only IPF AMs had increased expression of IL-6, IL-10, and CCL2/MCP-1 after stimulation PAM3Cys (Fig. 3C). Expression of TNFα was only upregulated in IPF AMs when stimulated with either LPS or PAM3Cys (Fig. 3C). To confirm differences in gene expression correlated with protein production, the supernatants from stimulated AMs cultures were assessed for a select group of cytokines and chemokines by ELISA. Overall, the IPF AMs were more responsive to TLR agonists compared with the normal AMs (Fig. 3, D–G). LPS induced a significant increase in TNFα (Fig. 3D), IL-10 (Fig. 3F), and IL-1β (Fig. 3G) from IPF AMs. In addition, LPS and TLR3 ligand, poly I:C, induced increased CCL2/MCP-1 production (Fig. 3E). Of all the cytokine and chemokines measured, only LPS increased production of TNFα in normal AMs (Fig. 3D). In fact, IPF AMs produced significantly more TNFα at baseline as well as under LPS and PAM3Cys stimulus compared with respective normal AMs under the same stimulus (Fig. 3D). These data indicate that IPF AMs have heightened sensitivity to TLR ligands, displaying exaggerated gene expression and protein production of various cytokines.

Fibroblast Contractility and CCL2/MCP-1 Secretion Is Not Influenced by TLR Stimulation

Fibroblasts isolated from patients with IPF also have an altered expression of TLR receptors compared with fibroblasts from normal donor controls (Supplemental Fig. S4B). To determine the role of TLR stimulation in regulating fibroblast contractile behavior and CCL2/MCP-1 production, normal and IPF lung fibroblasts were encapsulated in collagen hydrogels and treated with TLR 1/2, TLR 3, and TLR 4 ligands (PAM3Cys, poly I:C, LPS, respectively). Contraction of the hydrogels was monitored over 7 days, and no significant change in contractility was found under any treatment for either diseased or normal fibroblasts (Fig. 4, A and B). Cell culture medium was collected from these cultures on days 4 and 7 and were assessed for CCL2/MCP-1 production via ELISA. After 7 days in culture (media change at day 4), IPF fibroblast made less CCL2 compared with normal fibroblasts (Supplemental Fig. S5C). No significant change in CCL2/MCP-1 production was found for normal fibroblasts under TLR stimulus (Fig. 4C); however, there was an increase in the amount of CCL2 produced at day 7 compared with day 4. IPF fibroblasts significantly increased CCL2 production under LPS stimulation at day 7 compared with both nonstimulated controls (day 7) and LPS-treated cultures at day 4 (Fig. 4D). TGFβ production in normal and IPF samples were also compared at days 4 and 7 and no significant difference in baseline production levels between disease states was found (Supplemental Fig. S5D). Fibroblasts grown on 2-D plates and treated with TLR ligands were assessed for changes in gene expression (Fig. 4, E and F). Normal lung fibroblasts increased collagen 1 and αSMA when stimulated with PolyI:C. IPF fibroblasts significantly decreased collagen 1 and αSMA in response to LPS. Fibroblast Taken together, these data demonstrate that fibroblast contractility is not directly influenced by TLR ligand stimulation, though CCL2 secretion may be increased in IPF fibroblasts under LPS treatment.

Figure 4.

TLR ligands do not affect fibroblast contractility, but TLR4 activation may contribute to CCL2/MCP-1 production from IPF fibroblasts. The functional contractility of normal (A) or IPF (B) fibroblasts was tracked over the course of 1 wk under various TLR ligand treatments (LPS = purple, PAM3Cys = fuchsia, poly I:C = pink). CCL2 protein production due to these treatments was assess using ELISA at days 4 and 7 for both normal (C) and IPF (D) fibroblasts. Two-way ANOVAs were utilized for significance assessment with Tukey’s test for multiple comparisons. Changes in gene expression for normal (E) and IPF (F) fibroblasts were assessed after 24 h of TLR ligand exposure on 2-D plates. Significance was determined using two-way ANOVAs (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). IPF, idiopathic pulmonary fibrosis; TLR, Toll-like receptor.

TNFα Reduces Normal Fibroblast Contractility, Increases IPF Fibroblast CCL2/MCP-1 Secretion, and Alters ECM Gene Expression

To determine the role of AM cytokine secretion in driving the fibroblast contraction and fibrotic genes within the coculture system, normal and IPF fibroblasts were treated with various cytokines and contraction and ECM gene expression was assessed. Treatment of normal and IPF fibroblasts with exogenous cytokines decreased cellular viability, but both IPF and normal fibroblasts responded similarly (Supplemental Fig. S5, A and B). The addition of TNFα to normal fibroblasts significantly reduced contractile behavior early (Fig. 5A); however, IPF fibroblast contractility was not significantly altered by TNFα treatment (Fig. 5B). Interestingly, treatment of TGFβ, but not IL-13, increased hydrogel contractility in both normal and IPF fibroblasts but it did not reach significance until day 7 (Fig. 4, A and B). TGFβ treatment also significantly reduced the amount of CCL2/MCP-1 produced by fibroblast cultures when compared with control (Fig. 5, C and D). In both the IPF and normal fibroblasts treated with TGFβ, significantly less CCL2/MCP-1 was produced at day 7 compared with day 4 (Fig. 5, C and D). TNFα treatment resulted in a significant increase in CCL2/MCP-1 production only in IPF fibroblasts compared with nonstimulated controls (Fig. 5D). Treatment of normal and IPF fibroblasts with exogenous TNFα resulted in a significant decrease in expression of αSMA for both conditions. It also caused an increase in both collagen 1 and 3 in IPF fibroblasts but a decrease in collagen 1 and an increase in collagen 3 for normal fibroblasts (Fig. 5, E and F). These data demonstrate that TNFα can alter the contraction function of normal fibroblasts, the expression of CCL2/MCP-1 production of IPF fibroblasts and gene expression of both normal and IPF fibroblasts.

Figure 5.

Fibroblast respond to exogenous cytokine treatment by altering contraction phenotype and gene expression. Normal (A) or IPF (B) human fibroblast cultures encapsulated within collagen hydrogels were subjected to various cytokine treatments and tracked for changes in contractility over the course of 1 wk. Protein production of CCL2 was measured using ELISA on days 4 and 7 for normal (C) and IPF (D) treatment conditions. Gene expression changes after 24 h of exogenous TNFα treatment on 2-D cultures was investigated for normal (E) and IPF (F) fibroblasts. Statistical significance was determined using a two-way ANOVA with Sidak’s multiple comparisons test (A and B), Tukey’s multiple comparison test (C and D), or t tests (E and F; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). IPF, idiopathic pulmonary fibrosis.

Indirect Coculture of Fibroblasts and Macrophages Increased CCL2/MCP-1 Secretion by Fibroblasts

To assess whether changes in fibroblast phenotype were due to secreted cytokines or chemokines from alveolar macrophages, AM transwell systems and AM-conditioned media were utilized with fibroblast cultures to measure changes in gene expression and fibroblast contraction. In contrast to what was observed in a direct culture system (Fig. 2A), normal AMs cultured in a transwell insert with normal or IPF fibroblasts showed no decrease in expression of collagen 1 or collagen 3 (Fig. 6, A and B). Indirect culturing of normal AMs with normal fibroblasts resulted in a significant decrease in αSMA expression where IPF fibroblasts did not. Normal fibroblasts cultured in a transwell system with IPF AMs resulted in a significant increase in both collagen 3 and αSMA, whereas IPF fibroblasts were not affected by IPF AMs (Fig. 6, A and B). There were significant differences in fibronectin expression in IPF fibroblasts cultured in transwell with normal AMs compared with IPF AMs, but not compared with IPF fibroblasts alone (Fig. 6B). To determine if conditioned media from AMs altered fibroblast contractility, conditioned media was collected from IPF and normal AMs and fibroblast hydrogels were cultured in conditioned media for 7 days. Unlike what was observed with direct coculture, there was menial impact on contractility, with normal fibroblasts increasing contraction with IPF AM-conditioned media on days 3–5, and day 7. Normal AM-conditioned media showed an increase in contractility on days 5 and 7 for normal fibroblasts (Fig. 6C). IPF fibroblasts showed reduced contraction with normal AM-conditioned media on days 6 and 7 and there was a significant difference between normal and IPF conditioned media contraction on days 5–7. When these contraction profiles were assessed using curve fit regression equations no significant differences were observed for decay constants or plateau values (Supplemental Fig. S6, D–G). Both normal (Fig. 6E) and IPF (Fig. 6F) fibroblasts incubated in conditioned media from normal AMs resulted in increased CCL2/MCP-1; however, IPF AM-conditioned medium only increased CCL2/MCP-1 production on IPF fibroblasts (Fig. 6F). Days in culture increased production of TGFβ from both normal and IPF fibroblasts (Fig. 6, G and H), and normal AM-conditioned media increased TGFβ in normal fibroblast cultures compared with controls. Baseline levels of TGFβ, TNFα, and CCL2/MCP-1 in the normal or IPF AM-conditioned media was not significantly different (Supplemental Fig. S6). Overall, these data demonstrates that indirect coculture influences gene expression and cytokine secretion of fibroblasts but is not sufficient to induce contractile differences from diseased or normal lung fibroblasts.

Figure 6.

Effects of AMs on fibroblast phenotypes in nondirect coculture. Normal fibroblasts (A) or IPF fibroblasts (B) were cultured in a transwell system with either normal (blue) or IPF (red) AMs and assessed for gene expression changes. Changes in normal fibroblast (C) and IPF fibroblast (D) contraction was assessed when cultured with conditioned media from normal AMs (blue) or IPF AMs (red). Protein production for cytokines CCL2/MCP-1 (G and H) and TGFβ (I and J) was assessed at days 4 and 7 from the contraction assays using ELISA. Dotted lines indicate baseline levels of CCL2/MCP-1 protein (E and F) and TGFβ protein (G and H) in AM-conditioned media. Significant changes in gene expression were determined using one-way ANOVA with Tukey’s multiple comparison tests when data tested with a normal distribution (A), and with nonparametric one-way ANOVA with Kruskal–Wallis test for multiple comparisons when the data was nonnormal (F). All significance in contractility was determined using two-way ANOVA with Sidak’s multiple comparisons test. Statistical significance of protein production was done using two-way ANOVAs Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. AMs, alveolar macrophages; IPF, idiopathic pulmonary fibrosis. % indicates significance between fibroblasts alone and fibroblast + IPF AM conditioned media; # indicates significance between fibroblasts + normal AM conditioned media and fibroblasts + IPF AM conditioned media.

DISCUSSION

This study examined the impact of direct versus indirect macrophage-fibroblast communication as well as the role of disease state in regulating ECM gene expression and fibroblast contraction. Coculturing primary human AMs isolated from normal donors with either IPF or normal lung fibroblasts resulted in reduced collagen gene expression. Interestingly, cocultures of IPF AMs with IPF fibroblasts did not change ECM gene expression; however, there was a significant increase in collagen 3 expression when IPF AMs were incubated with normal lung fibroblasts. Contraction of normal fibroblasts was significantly reduced when cocultured with either normal or IPF AMs, whereas IPF fibroblasts exhibited a significant decrease in contraction at day 7 when cocultured with IPF AMs. Both normal and IPF AMs responded more to the TLR4 ligand LPS, but only IPF AMs showed a heightened response to TLR1/2 and TLR3 ligand by increasing production of proinflammatory cytokines, such as TNFα and IL-1β. There was no difference in the contractility phenotype or production of CCL2/MCP-1 of normal or IPF lung fibroblasts when stimulated with a variety of different TLR agonists. The addition of exogenous TNFα significantly reduced normal fibroblast contractility, at early time points, whereas exogenous TGFβ significantly increased the contraction of both normal and IPF 3-D hydrogel cultures at later time points. To assess the effect of cell-to-cell communication between AMs and fibroblasts in regulating gene expression, we utilized a transwell coculture system and found that AMs from patients with IPF elevated collagen 3 and αSMA expression from normal fibroblasts; however, subjecting fibroblasts to AM-conditioned media had only a marginal influence on fibroblast behavior, increasing CCL2 production but not altering contractility. Taken together, these data demonstrate that AMs have a significant impact on fibroblast contractility, gene expression, and cytokine production and this response is dependent on the disease state of the cells. AM-mediated modulation of fibroblast contractility requires direct cell-cell communication and is only partially driven by soluble factors. A summary of these findings are provided in Fig. 7.

Figure 7.

Summary of fibroblast and alveolar macrophage coculture responses. Pulmonary fibroblasts and alveolar macrophages derived from normal or IPF human patient samples were cocultured in either direct or indirect contact. Changes in gene expression, contraction and CCL2 are summarized for the various coculture settings. IPF, idiopathic pulmonary fibrosis. [Image created with BioRender.com and published with permission.]

The study of macrophage-fibroblast crosstalk in the context of pulmonary fibrosis has been largely unexplored. Pakshir et al. (19) demonstrated that macrophages are drawn to fibroblast remodeling areas through force generation within the ECM and that this form of communication and macrophage trafficking ranges over greater distances than chemotactic gradients. The influence of AM-conditioned media was investigated by Hult et al. (20) who found that the fibrotic response of fibroblasts, as measured through their migration, proliferation, and mRNA expression, in AM coculture was recapitulated by AM polarization cytokines IL-4 and IL-13 rather than AM-conditioned media. Our data are consistent with these findings as our AM-conditioned media did not contain macrophage-skewing cytokines and did not significantly alter contractility profiles of either normal or IPF fibroblasts. Additional research investigating transwell coculture of AMs and fibroblasts showed influences on fibroblast migration (21), CCL18 production (22), and osteopontin crosstalk (23) between the two cell types. Though we did not investigate fibroblast migration, our gene expression findings are consistent with the work of Ding et al. (24) who demonstrated that αSMA and collagen 1 expression levels are significantly altered by macrophage-conditioned media depending on the state of macrophage polarization and can predict a positive feedback loop in pulmonary fibrosis. Overall, our data and the previous work on AM-fibroblast coculture systems indicate that AMs have a direct impact on fibroblast gene expression, but that soluble factors have a modest effect on fibroblast contractility. Macrophage-fibroblast contact may be an important mechanism that requires further study to understand the coculture crosstalk critical to the perpetuation of IPF.

There are transcriptional differences in macrophages from patients with IPF compared with normal donor controls (25), and this study focused on expanding those results to understand differences in the function of AMs in the setting of IPF. The relative cell composition of normal and IPF BALs were identified using flow cytometry, and no significant difference in proportion of cell types was found (Supplemental Figs. S2A and S3A). In the setting of murine models of pulmonary fibrosis, reduced embryonically derived tissue resident macrophages and increased recruited monocyte-derived macrophages resulted in a more fibrotic phenotype (26). To determine if this also occurred in human lungs, we attempted to distinguish tissue resident versus monocyte-derived alveolar macrophages in our BAL samples. Based on published literature, we decided to use CD163 to distinguish recruited versus. resident macrophages (16). Using flow cytometry, we did not observe any difference in expression of CD206⁺ CD169⁺CD163⁺ AMs between IPF and normal samples (Supplemental Fig. S2B). however, this marker is not well established and may not be an accurate indicator.

Our work demonstrates a role for TNFα in directing macrophage-fibroblast interactions. This proinflammatory cytokine is instrumental in the inflammatory response, immune cell recruitment, and can be produced by a variety of cells, including both macrophages and fibroblasts (27). TNFα has a critical role in fibrotic disease, but use as a potential therapy for IPF is controversial because it has been attributed to both progression and initiation of inflammatory-related diseases, such as liver fibrosis (28) and cancer (29). Previous work has proposed TNFα as a promising treatment for IPF as it has shown prevention of pulmonary fibrosis (30) and resolution of ECM remodeling (31) in murine models. A randomized, prospective, double-blind, placebo-controlled study showed anti-TNFα therapy, etanercept, showed no difference in pulmonary function tests, but did decrease the rate of disease progression (32). Studies using human fetal lung fibroblasts demonstrated that exogenous TNFα reduced fibroblast contractility (33), which aligns with our current findings. Collagen production by cultured fibroblasts decreased when treated with TNFα in a dose- and time-dependent manner (34), whereas we observed increased collagen gene expression under TNFα treatment (Fig. 4, E and F). The recent work by Xu et al. (35) showed that pulmonary macrophages stimulated with LPS induced release of TNFα via the JNK pathway. When the conditioned media from LPS-stimulated macrophages were placed on a fibroblast cell line, enhanced aerobic glycolysis and lactate production were observed (35) demonstrating that activated macrophages influence fibroblasts via TNFα signaling. In addition, our work showed that conditioned media from AMs was not sufficient to induce changes in contraction from normal or IPF fibroblasts (Fig. 5), but exogenous TNFα did influence normal fibroblast contraction (Fig. 5A) and gene expression from both normal and IPF lung fibroblasts (Fig. 5, E and F). These data suggest either a possible dose-dependent response or difference in receptor expression that could alter the ability of the fibrotic fibroblasts to respond. Future studies will investigate the dose-dependent response between IPF and normal patient samples to determine if doses of TLR ligands and cytokines affect functional behaviors of fibroblasts.

TNFα is also a critical regulator of CCL2/MCP-1 production (36–38), a known monocyte chemoattractant. Our data demonstrate that the disease state regulates the response profile of fibroblasts to TNFα treatment. Normal fibroblasts under TNFα treatment show an antifibrotic response, reduced contraction, and no increase in CCL2/MCP-1 production, whereas IPF fibroblasts did not alter their contraction phenotype and significantly upregulated CCL2/MCP-1 production. CCL2 is a profibrotic mediator in IPF fibroblasts and has been shown to induce an upregulation of receptors for TGFβ and IL-13 but not in normal fibroblasts (39). Though we did not investigate the role of exogenous CCL2 treatment, previous studies have shown it to increase αSMA gene expression as well as ECM components in IPF fibroblasts which were hyperresponsive compared with normal fibroblasts (39, 40). This both aligns and contrasts with our study investigating TNFα treatment which showed that IPF fibroblasts were more responsive, producing a significant increase in CCL2 but also lack a functional change in IPF fibroblast contractility. Further exploration on the influence of exogenous CCL2 in our samples would be interesting to compare the TGFβ-CCL2-IL-13 axis under coculture conditions and the extent of IPF versus normal fibroblast hyperresponsiveness.

One interesting theme observed throughout our experiments was the fact that the normal fibroblasts were more responsive to AM signals by altering contraction and gene expression compared with the IPF fibroblasts. Addition of either AMs or exogenous cytokines resulted in significant differences in behavior as well as gene expression of normal fibroblasts, whereas IPF fibroblasts appear to be less responsive overall. One reason for this could be due to the large degree of heterogeneity within the IPF fibroblast samples. Fibroblasts from patients with IPF had a large degree of variability in their contractile properties as well as an elevated decay constant indicating more rapid contraction. Although all these patients were diagnosed with IPF, there were differences in age, sex, and previous smoking history. Previous IPF research has shown that age affects fibroblast gene expression and functional ability (41), and that sex has a significant influence on IPF susceptibility and progression (42). Due to our sample size these variables could not be explored in depth but are areas of ongoing research. Another possible reason for the differences in fibroblast function and gene expression could be epigenetic differences. Multiple previous genome-wide studies have demonstrated that there are significant changes in methylation patterns and histone modification sites in patients with IPF compared with normal donor controls (43). Since there is a higher incidence of smoking among patients with IPF (1), this could lead to increased epigenetic changes resulting in the abnormal wound healing and repair mechanisms observed in the setting of pulmonary fibrosis. Taken together with these previous studies, our work suggests that patient variability plays a large role in IPF progression and fibroblast response to external mediators as well as in cell-cell communication with macrophages.

When working with human samples in vitro there are natural limitations present within the findings. We recognize that serum is a prominent source of TGFβ and may be occluding an observable difference in baseline levels between IPF and normal AMs. However, as all our in vitro studies were performed in supplemented cell culture media, TGFβ levels reported in Supplemental Fig. S6B are representative of what the cells in these studies were consistently exposed to. The TLR expression in Supplemental Fig. S4 show IPF AMs to have a decreased expression of various TLRs despite their heightened responsiveness to TLR ligands (Fig. 3, D–G). Further studies targeting additional TLRs that are differentially expressed (TLR 5, 7, 8, 9) would be an important way to investigate this trend and may suggest the importance of other non-TLR-mediated mechanisms of inflammatory gene expression, such as RIG-I.

In the literature, there is a lack of functional data exploring the impact of observed changes in gene expression between IPF and normal macrophage-fibroblast coculture experiments. Many studies assay transcriptomics without tying results to functional outcomes in either patients or cellular phenotypes. To address this gap, our model assesses not only the changes in gene expression due to various coculture conditions but also how coculture-induced changes influence fibroblast functional contractile behavior. Our study demonstrates that AMs isolated from normal subjects can significantly alter both gene expression and fibroblast-mediated contractility. Although TLR ligands and pro- and anti-inflammatory cytokines can alter fibrotic gene expression, these soluble factors do not significantly alter fibroblast contractility. Our data suggest that AM modulation of fibroblast contractility requires direct coculture and future studies should investigate the potential contact-mediated mechanisms of AM-fibroblast interactions.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Tables S1–S3 and Supplemental Figs. S1–S6:https://doi.org/10.6084/m9.figshare.20503482.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants R01HL141217 and 1F32HL164020 (to M. N. Ballinger); Mistletoe Research Fellowship from the Momental Foundation (to C. M. Novak); and the Presidents Accelerator Grant from Ohio State University Foundation (to S. N. Ghadiali and M. N. Ballinger).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.M.N., S.S., K.L.L., B.F.R., B.W., and M.N.B. conceived and designed research; C.M.N., S.S., K.L.L., B.F.R., J.S.W., and M.N.B. performed experiments; C.M.N., K.L.L., B.F.R., S.N.G., and M.N.B. analyzed data; C.M.N., B.F.R., S.N.G., and M.N.B. interpreted results of experiments; C.M.N., S.N.G., and M.N.B. prepared figures; C.M.N., K.L.L., J.S.W., and M.N.B. drafted manuscript; C.M.N., S.S., B.F.R., J.S.W., B.W., S.N.G., and M.N.B. edited and revised manuscript; C.M.N., S.S., K.L.L., B.F.R., S.N.G., and M.N.B. approved final version of manuscript.