Alveolar fibroblasts in acute lung injury: biological behaviour and clinical relevance

Study population

The protocol was approved by the ethical committee of the Paris-Hotel-Dieu Hospital (Paris, France). Ventilated patients from three ICUs were prospectively enrolled if a BAL procedure was performed to confirm clinically suspected ventilator-associated pneumonia 11. The patients were classified into three groups after inclusion (ALI, ARDS and a ventilated-control group without ALI/ARDS) according to the criteria of the American–European Consensus Conference on ALI/ARDS 12. All the patients were ventilated and weaned from the ventilator according to current guidelines 13. Patients with pre-existing fibrotic lung disease, corticosteroid medication use, HIV infection, end stage cancer, or who were aged <18 yrs or currently pregnant were excluded. On the day of inclusion, the following clinical data were recorded: age; sex; reason for ICU admission; presence or absence of sepsis and pulmonary infection; arterial oxygen tension (P_a,O2)/inspiratory oxygen fraction (F_i,O2) ratio; arterial carbon dioxide tension; Simplified Acute Physiologic Score II on admission to ICU 14; Sequential Organ Failure Assessment score 15; and lung injury score 16. In addition, the length of time between the onset of the mechanical ventilatory support and the BAL, as well as the length of ventilatory support, the time spent in the ICU and mortality 28 days after the BAL were recorded for all patients.

BAL protocol and BAL fluid sample processing

The BAL was performed and processed as previously described 17. A differential BAL cell count was performed on a cytocentrifuge smear with a Diff-Quik stain kit (Dade International, Miami, FL, USA). BAL fluid and plasma protein concentrations were measured with the analyst Hitachi-911 (Roche, Meylan, France), and the protein ratio was determined to evaluate the alveolar permeability induced by lung injury.

Isolation and culture of fibroblasts

The BAL cell viability was determined before initial plating by Trypan blue vital staining (viability always >92% in the different groups tested). The BAL cells (n = 68) were cultured at a density of 2.10⁵ cells·cm⁻² in a 25-cm² flask (Corning Costar, Lowell, MA, USA) in 5 mL of RPMI-1640 (Gibco, Grand Island, NY, USA) supplemented with 10% heat decomplemented fetal calf serum (FCS), 2 mM l-glutamine and antibiotics (Invitrogen, Cergy Pontoise, France) for at least 28 days. Cells grown from BAL were used for immunocytochemical characterisation 24 h and 21 days after initial plating (n = 5), and then at passage 1 and 3 (n = 12). Nine BAL cell lines were randomly chosen and their proliferation, migration and collagen 1 production capacities were compared to those of fibroblasts derived from normal lung explants obtained from five patients (two females and three males, mean age 60 yrs) undergoing lung surgery for removal of a primary lung tumour. Three patients had never smoked and two were ex-smokers. As previously described 18, normal lung from a healthy segment distant from the solidary lesion was obtained. Fibroblasts were cultured with complete medium and 10% FCS supplemented with antibiotics and were used at passage 3.

Measurement of C-terminal propeptide of type I pro-collagen, monocyte chemoattractant protein-1, interleukin-8 and transforming growth factor-β1

C-terminal propeptide of type I pro-collagen (PICP) was measured by enzyme immunoassay (OSTEOmedical, Paris, France). The detection threshold of the assay was 0.2 ng·mL⁻¹. Transforming growth factor (TGF)-β1, monocyte chemoattractant protein (MCP)-1 and interleukin (IL)-8 concentrations were measured by ELISA (R&D systems, Minneapolis, MN, USA). The detection thresholds of TGF-β1, IL-8 and MCP-1 assays were 5 pg·mL⁻¹, 3.5 pg·mL⁻¹ and 5 pg·mL⁻¹, respectively.

Immunocytochemistry

Alveolar fibroblasts cultured on Lab-Tek slides (Nunc, Naperville, IL, USA) were fixed with acetone. Anti-human antibodies (Ab) directed against collagen 1 (Sigma Aldrich, St Louis, MO, USA), vimentin (Immunotech, Marseille, France), prolyl-4-hydroxylase (α-4H), desmin, pancytokeratin, CD31, α-SMA, CD45 (Dako SA, Trappes, France), smooth muscle myosin 1 (Abcam, Paris, France), CD14, CD34 (BD Pharmingen, Le Pont-De-Claix, France) or isotype-matched control Ab were used. Positive cells were revealed using the Vectastain ABC-alkaline phosphatase kit mouse immunoglobulin G (Vector Abcys, Paris, France) and the fast red substrate (Dako SA).

Confocal microscopy analysis

Cells cultured on Lab-Tek slides were fixed after 21 days of culture with acetone and saturated with 50 mM NH₄Cl and permeabilised with 0.1% Triton X100. After saturation with normal horse serum, slides were incubated with appropriate dilution of anti-collagen 1 monoclonal Abs (1:290) (Sigma Aldrich) prior to the addition of fluorescent-labelled Abs (biotinilated horse antimouse (1:300) plus Alexa Fluor 568-conjugated streptavidin (1:100); Invitrogen, Carlsbad, CA, USA). After washing, the cells were incubated with anti-CD45-Alexa Fluor 488 Abs (1:25) (Santa Cruz Biotechnology, San Diego, CA, USA). Cell nuclei were localised by DNA staining with thiazole orange protein 3 (0.4 μg·mL⁻¹; Molecular Probes, Eugene, OR, USA). Slides were mounted with Prolong Gold anti-fade reagent (Invitrogen) and examined by confocal laser scanning microscopy (LSM-510-META microscope; Zeiss, Oberkochen, Germany). To test for the specificity of immunostaining, Abs were replaced by an isotype-matched control Ab.

Cell proliferation assay

After 72 h of incubation with 0.1% FCS alone (basal condition) or with prostaglandin (PG) E₂ (10⁻⁶ M), recombinant human (rh)IL-1β (10 ng·mL⁻¹), rh platelet-derived growth factor (PDGF)-BB (10 ng·mL⁻¹) or rhTGF-β1 (10 ng·mL⁻¹; R&D systems), proliferation was measured as bromodeoxyuridine (BrdU) incorporation according to the manufacturer's protocol (Roche).

Cell migration assay

Cell motility was examined using a modified Boyden chamber apparatus (Transwell Costar, Cambridge, MA, USA) fitted with an 8-μm pore membrane coated with fibronectin (10 μg·mL⁻¹). Serum deprived fibroblasts (2.10⁵ per well) were placed into the upper well and allowed to migrate for 24 h in the presence or absence (basal condition) of PGE₂ (10⁻⁶ M), Oleoyl-l-α-lysophosphatidic acid sodium salt (LPA, 10⁻⁶ M; Sigma Aldrich), or rhPDGF-BB (10 ng·mL⁻¹) into the lower well. Non-migrated cells were removed and the membrane was fixed and stained with the Diff-Quick staining kit (Dade Behring, Paris, France). The number of cells that migrated to the lower surface of the membrane was counted (400× magnification). Five high-power fields (HPF) were counted per sample and run in duplicate. The chemotactic index was determined as the ratio of cells moving in response to a chemoattractant relative to cells moving in response to media control.

Assessment of collagen I and TGF-β1 protein secretion

Cells (5×10⁴ cells·well⁻¹) were cultured in serum-free medium (basal condition) or with mediators: PGE₂ (10⁻⁶ M), rhIL-1β (10 ng·mL⁻¹) or rhTGF-β1 (10 ng·mL⁻¹) for 48 h. The concentration of PICP or TGF-β1 in fibroblast supernatants was divided by the total protein amount of the cell monolayer (Bio-Rad Laboratories, Hercules, CA, USA).

Assessment of collagen I or TGF-β1 mRNA expression

After 18 h of culture, the relative content of mRNA was analysed after reverse transcription by RT-PCR as previously described 17. The following primer sequences were used. Collagen 1-α1 (COL1A1): CGAGGTAGTCTTTCAGCAACACAGT and AGCCACCAGCCCCTCACT; collagen 1-α2 (COL1A2): CAGTTCTTGGCTGGGATGTTT and TTGAGACTCAGCCACCCAGAGT; TGF-β1: TGCTGTCACAGGAGCAGTG and AGGGGAAATTGAGGCTTT; Ubiquitin C: TTTTTTGGGAATGCAACAACTTT and CACTTGGTCCTGCGCTTGA.

Statistical analysis

Demographic data were expressed as mean±sd and biological data as median (range). All proportional values were compared with the Fischer's exact test or the Chi-squared test for multiple comparisons. The continuous variables were compared by the Mann–Whitney test or Wilcoxon paired test when appropriate. Multiple comparisons were tested by the Kruskal–Wallis or Friedman test followed by Dunn's multiple-comparison post hoc analysis. Durations of ventilation after the BAL procedure were analysed by Kaplan–Meier survival curves and compared by log-rank analysis. Correlations were assessed with the Spearman rank-order test. Statistical significance was accepted as p≤0.05.

Characteristics of the study population

31 patients with ARDS, 17 with ALI and 20 ventilated patients without ALI/ARDS criteria were enrolled. The patient characteristics are shown in table 1⇓. ALI/ARDS patients were admitted for septic shock (n = 23), multiple organ failure (n = 18), cerebral stroke or convulsive state (n = 4), severe trauma (n = 2) or cardiac arrest (n = 1). The ventilated patients without ALI/ARDS were admitted for acute heart failure and post-operative complications following cardiopulmonary bypass (n = 14), post-operative localised pneumonia (n = 3), acute renal failure (n = 1), haemorrhagic shock (n = 1) or acute intoxication (n = 1).

Cells cultured from BAL expressed fibroblast markers

All BAL cell pellets (n = 68) were maintained in culture for 28 days without any bacterial or fungal contamination. At the time of initial plating, BAL cells were mainly neutrophils and mononuclear cells (macrophage/monocyte). In positive cell cultures (n = 12), isolated spindle-shaped cells were usually detected after the first week of culture (median time 10 (4–23) days) (fig. 1a⇓). After 3 weeks of culture, spindle-shaped cells formed foci (fig. 1b⇓), while round cells (likely macrophages) were quiescent. After 4 weeks of culture, the cell monolayers were harvested to produce a homogenous cell population in the subsequent cultures (fig. 1c⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

a–c) Morphology of adherent bronchoalveolar lavage (BAL) cells at amplification when a positive culture occurred. a) After 8 days, b) at 21 days and c) at passage 1. The arrows indicate typical adherent spindle-shaped cells. d–g) Immunocytochemical characteristics of adherent BAL cells after initial plating (H24) with d) control antibody, e) CD34, f) CD45, g) prolyl-4-hydroxylase (α-4H). The arrows indicate typical adherent spindle-shaped cells. h–k) Characterisation of fibrocytes (spindle-shaped cells co-expressing CD45 and collagen 1) by confocal microscopy analysis of BAL cell culture at day 21. h) CD45, i) collagen 1, j) thiazole orange protein 3 (TO-PRO-3) and k) merge of the three fluorescence channels. l–p) Representative immunocytochemical stain prepared from alveolar fibroblasts at passage 1. l) CD45, m) vimentin, n) α-4H, o) collagen 1, p) α-smooth muscle actin (α-SMA). The arrows indicate CD45+ spindle-shaped cells. q–u) Representative immunocytochemical stain prepared from alveolar fibroblasts at passage 3. q) CD45, r) vimentin, s) α-4H t) collagen 1 and u) α-SMA. a–c) Scale bars = 250 μm. d–g) Scale bars = 25 μm. h–k) Scale bars = 15 μm. l–u) Scale bars = 100 μm.

We characterised the phenotype of adherent BAL cells from five different patients (three ARDS and two ALI) 24 h after initial plating. At that time point, 95% of adherent cells were mononuclear round cells expressing the common leukocyte marker (CD45+) (fig. 1f⇑). The remaining cells were CD45-, including some endothelial cells (CD31+) and epithelial cells (pancytokeratin+). No mature form of fibroblast (CD45-) with large cytoplasm expressing collagen 1 was detectable 24 h after plating. In contrast, we observed a side population (<1% of the total cell population) of very typical elongated spindle-shaped cells with a ratio of cell length/cell width equal to 16. All these cells were CD45+, α-4H + and weakly CD34+ (fig. 1e–g⇑). This pattern of expression is consistent with that of fibrocytes 19 and was confirmed by confocal microscopy showing the co-expression of collagen 1 and CD45 on day 21, before the first passage (fig. 1h–k⇑). At passage 1, most of the spindle-shaped cells were positive for α-4H, collagen 1 and vimentin, but only a few population were still CD45+ (fig. 1l–p⇑). A limited population of cells with large cytoplasm expressing αSMA was also observed at this passage. The cells characterised at passage 3 demonstrated a stable mesenchymal phenotype with positive staining for vimentin, collagen 1 and α-4H (fig. 1q–u⇑). The α-SMA was expressed in eight out of 12 cell lines in which 20% of cells were α-SMA positive. Markers for vascular smooth muscle cells (desmin, smooth muscle myosin 1), epithelial cells (pancytokeratin), endothelial cells (CD31), monocytes (CD14), leukocytes (CD45) and haematopoieteic progenitor cells (CD34) were negative at this passage. Together, this pattern of markers is consistent with a differentiated fibroblast phenotype. According to these characteristics, the cells will be hereby referred to as “alveolar fibroblasts”.

Proliferative rate and migration of alveolar fibroblasts

After 72 h of culture, no significant difference in the proliferative rate was observed between unstimulated alveolar and control fibroblasts studied at passage 3 (BrdU median absorbance 0.58 versus 0.60, p>0.05). IL-1β, TGF-β1 or PDGF-BB induced a similar weak mitogenic effect in both types of fibroblasts (125% of basal condition, p>0.05). PGE₂ induced a weak and similar inhibitory effect on proliferation in both cell types (75% of basal condition, p>0.05). We examined the migration of alveolar fibroblasts in the presence or absence of mediators known to modulate mesenchymal cell migration. In the basal condition, the migration of alveolar fibroblasts was three-fold higher than that demonstrated by control fibroblasts (26 versus 8 cells·HPF¹, p = 0.04) (fig. 2a⇓). PDGF-BB (10 ng·mL⁻¹) and LPA (10⁻⁶ M) stimulated control and alveolar fibroblast migration to the same extent (fig. 2b–e⇓). In contrast, alveolar fibroblasts were hyporesponsive to the inhibitory effect of PGE₂ as compared with control fibroblasts (73 versus 23% of basal condition, p = 0.01).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

a) Alveolar fibroblast migration capacities after 24 h. Basal migratory capacity of alveolar fibroblasts (n = 9) was assessed in comparison with control fibroblasts (n = 5) by cell count in five random high-power fields (HPF) run in duplicate. The horizontal bars indicate the median of individual values. b) Cell migration in the presence of prostaglandin (PG) E₂ (10⁻⁶ M), lysophosphatidic acid (LPA; 10⁻⁶ M) and platelet-derived growth factor (PDGF)-BB (10 ng·mL⁻¹). Results are expressed on a logarithmic scale as chemotactic index (ratio of the basal condition performed in the same assay). c–e) Representative observations of alveolar fibroblasts traversing the chamber membrane. c) Basal condition, d) lysophosphatidic acid (10⁻⁶ M) and e) PDGF-BB (10 ng·mL⁻¹). Scale bars = 30 μm. ^#: p = 0.04; ^¶: p = 0.01.

Collagen I production was elevated in alveolar fibroblasts and correlated with TGF-β1 production

Collagen 1 production was assessed by measuring PICP in cell culture supernatants. PICP secretion by alveolar fibroblasts was four-fold higher than that of control fibroblasts (7.69 versus 1.63 ng·μg⁻¹·well⁻¹, p<0.01) (fig. 3a⇓). In addition, COL1A1 and COL1A2 mRNA expression in alveolar fibroblasts was higher than in control fibroblasts (p<0.01, fig. 3b⇓) and was positively correlated with collagen 1 protein secretion (ρ = 0.75, p = 0.03). Recombinant TGF-β1, and to a lesser extent rhIL-1β, stimulated collagen 1 production by control and alveolar fibroblasts (p<0.05). After stimulation, collagen 1 production remained higher in alveolar fibroblasts than in control fibroblasts (p≤0.04) (fig. 3a⇓). Interestingly, as compared with fibroblasts obtained from normal lung explants, alveolar fibroblasts were hyporesponsive to the inhibitory effect of PGE₂ (p<0.05). As TGF-β1 is a key factor in collagen 1 production by fibroblasts, its role was specifically studied. The basal levels of TGF-β1 secretion by alveolar and control fibroblasts were similar (5.3 versus 4.8 pg·μg⁻¹·well⁻¹, p = 0.24) (fig. 4a⇓). A strong correlation was found between TGF-β1 and collagen 1 production by alveolar fibroblasts at both protein and transcriptional levels (fig. 4b⇓ and c). Such a correlation was not found in control fibroblasts (fig. 4d⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Alveolar fibroblast type I collagen production capacities. a) Type I collagen synthesis of alveolar fibroblasts (n = 9) in comparison with control fibroblasts (n = 5). Fibroblasts were cultured in duplicate for 48 h in serum-free medium alone (basal) or with prostaglandin (PG) E₂ (10⁻⁶ M), interleukin (IL)-1β (10 ng·mL⁻¹) or transforming growth factor (TGF)-β1 (10 ng·mL⁻¹). b) Unstimulated alveolar and control fibroblast type I collagen α-1 (COL1A1) and α-2 (COL1A2) mRNA relative content analysed at 18H by quantitative real time RT-PCR and expressed as a ratio to the Ubiquitin C (UBC) mRNA. The horizontal bars indicate the median of individual values. ^#: p = 0.04; ^¶: p = 0.03. **: p<0.01.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4—

Correlation between transforming growth factor (TGF)-β1 and type I collagen production in alveolar fibroblasts. a) TGF-β1 synthesis of alveolar fibroblasts (n = 9) in comparison with control fibroblasts (n = 5). Fibroblasts were cultured in duplicate for 48 h in serum-free medium alone (basal condition). The horizontal bars indicate the median of individual values. b) Correlation between TGF-β1 and type I collagen concentration in unstimulated alveolar fibroblast (n = 9) supernatants. PICP: propeptide of type I pro-collagen. ρ = 0.82; p = 0.01. c) Correlation between both type I collagen alpha-1 (COL1A1) and alpha-2 (COL1A2) mRNA and TGF-β1 mRNA expression in unstimulated alveolar fibroblasts (n = 9). ρ = 0.87 and 0.72, respectively; p<0.04. d) Lack of correlation between COL1A1 and COL1A2 mRNA and TGF-β1 mRNA expression in unstimulated control fibroblasts (n = 5, p>0.05). Results are represented on a logarithmic scale. UBC: Ubiquitin C. COL1A1; □; COL1A2; ▪.

Clinical determinants of alveolar fibroblast culture from BAL and influence on outcome

Alveolar fibroblasts were grown from BAL of patients with ALI (n = 6; 35%) or ARDS (n = 6; 19%) but never from ventilated controls (n = 0) (p≤0.04, control versus ALI or ARDS). The proportion of positive fibroblast cultures was higher in ALI than in ARDS without reaching significance (35% versus 19%, p = 0.30). There was a greater proportion of positive cultures (eight (35%) out of 25 BAL) in the early phase (<7 days) of ALI/ARDS when compared to the late phase (≥7 days) (four (17%) out of 23 BAL), although this difference was not statistically significant (p = 0.32) (fig. 5a⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5—

a) Kinetics of fibroblast isolation in the acute lung injury and acute respiratory distress syndrome groups. Positive (▪) and negative (▓) fibroblast cultures were plotted as a function of mechanical ventilation duration before bronchoalveolar lavage (BAL). b) Duration of mechanical ventilation after BAL procedure according to the growth of alveolar fibroblasts. Kaplan–Meier survival curves comparing duration of mechanical ventilation after BAL procedure for patients with positive (n = 12: –––––) and negative fibroblast culture (n = 36: - - - -). •: censored data (death during ventilation or ventilation continuation > 28 days after BAL procedure). p<0.001, Log-rank test.

To clarify the determinants of positive fibroblast culture, the ALI/ARDS patients were divided into two groups according to the cell culture results (positive culture (n = 12) versus negative culture (n = 36)). As summarised in table 2⇓, the clinical parameters were similar in the two groups. There was no difference in P_a,O2/F_i,O2 ratio, severity scores, aetiology of ALI, length of ICU stay and 28-days mortality. However, the total duration of mechanical ventilation was notably reduced in patients with positive culture (14 versus 28 days, p = 0.03). This difference was explained by a reduction of mechanical ventilation duration after BAL procedure (8 versus 21 days, p<0.001) (fig. 5b⇑) and was observed in patients with early (8 versus 22 days, p = 0.0024) and late (9 versus 21 days, p = 0.0068) ALI/ARDS.

Patients with positive fibroblast culture exhibited specific BAL fluid characteristics (table 3⇓). In this group, a higher collagen 1 concentration associated with a reduction of inflammatory response assessed by a decreased neutrophil percentage and a lower IL-8 concentration in BAL was noted. In parallel, an increased percentage of monocytes/macrophages was found, which was positively correlated with BAL fluid MCP-1 concentration (ρ = 0.67, p = 0.02).