We have previously shown that inducible overexpression of the human TGF-β1 gene results in the accumulation of CD45⁺Col-Iα1⁺ cells in the murine lung [³,²⁵]. While this combination of markers has traditionally been considered sufficient for the identification of fibrocytes [⁸], accumulating data from our group and from others indicate that this combination of markers may in fact identify a heterogeneous population of cells [⁶,⁷]. Thus, in order to better characterize the identity of TGF-β1-recruited intrapulmonary CD45⁺Col-Iα1⁺ cells, TGF-β1 transgenic positive (Tg⁺) and wild-type control (Tg^-) mice received doxycycline in their drinking water for up to 2 weeks after which they were killed and CD45⁺Col-Iα1⁺ cells quantified as we have previously described [³]. While we would have preferred to use an antibody specific for the immature form of collagen I, such an antibody is currently not available. Thus, detection of the mature form of collagen was employed. These cells were then further immunophenotyped based on their expression of CD14 and/or CD34.

Consistent with our prior findings, CD45⁺Col-Iα1⁺ cells were detected in all mice, with a robust increase seen in the TGF-β1 Tg⁺ animals (Figure 1a-d). Further assessment revealed that in all mice these cells displayed variable expression of CD14 and CD34 (Figure 1e-h). Interestingly, cells meeting classical definition of fibrocytes based upon the coexpression of CD34, CD45, and Col-Iα1 in the absence of CD14, were rare in both sets of animals and not significantly altered between groups (Figure 2a). In contrast, when compared to Tg^-animals, the lungs of TGF-β1 Tg⁺ mice contained 64.8% fewer CD45⁺Col-Iα1⁺ CD14⁺CD34+⁺ cells (P < 0.001, Figure 2b) but nearly tenfold more CD45⁺Col-Iα1⁺ CD14⁺CD34^-cells (P < 0.001, Figure 2c). The amount of CD45⁺Col-Iα1⁺ cells expressing neither CD14 nor CD34 (CD45⁺Col-Iα1⁺ CD14^-CD34^-) did not differ between groups (Figure 2d). These data indicate that CD45⁺Col-Iα1⁺ + cells appearing in the TGF-β1-exposed lung are primarily composed of cells that express CD14 and lack CD34.

In order to explore the role of intrapulmonary caspase activation and apoptotic responses in the accumulation and phenotype of CD45⁺Col-Iα1⁺cells, TGF-β1 Tg⁺ and Tg^-- mice were given doxycycline in their drinking water and randomized to receive intraperitoneal dosing of the caspase inhibitor carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone (Z-VAD/fmk) for between 2 to 14 days. Mice were killed at the height of cell death, which occurs at 48 hours in the model, and assessed for caspase-3 activation using immunohistochemistry and for cell death responses using terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining. Consistent with our prior reports [¹⁸], caspase-3 activation was abrogated in the presence of Z-VAD/fmk (Figure 3a) and assessment of TUNEL staining in the TGF-β1 Tg⁺ lung revealed a 79.9% reduction in cell death responses at this timepoint (P < 0.0003, Figure 3b).

Having confirmed that caspase inhibition does indeed reduce apoptosis in this model, we next explored its effects on the recruitment of CD45⁺Col-Iα1⁺ cells. Here we found that treatment of TGF-β1 Tg⁺ mice with Z-VAD/fmk reduced CD45⁺Col-Iα1⁺ cells by nearly tenfold (P < 0.0014, Figure 3c) and restored quantities of all CD45⁺Col-Iα1⁺ subtypes to wild type levels. Specifically, compared to sham-treated TGF-β1 Tg⁺ mice, the lungs of Z/VAD-fmk treated Tg+ mice showed no change in CD45⁺Col-Iα1⁺CD14^-CD34⁺ cells (P = 0.14, Figure 3d), an 85.8% reduction in CD45⁺Col-Iα1⁺CD14⁺CD34^-cells (P < 0.001, Figure 3e), a 71.8% increase in CD45⁺Col-Iα1⁺CD14⁺CD34⁺ cells (P < 0.001, Figure 3f), and essentially no change in CD45⁺Col-Iα1⁺CD14^-CD34^-cells (P = 0.35, Figure 3g). These data indicate that apoptotic cell death responses regulate the appearance and phenotype of CD45⁺Col-Iα1⁺ cells in the TGF-β1-exposed murine lung.

Our prior studies have revealed that alternatively activated (M2) macrophages regulate the development of fibrosis [³,²⁶]. However, the precise relationship between fibrocytes and macrophages in the TGF-β1-exposed lung has not been fully explored. Given the importance of the M2 macrophage in tissue repair and remodeling responses we thought it possible that M2 macrophages control the appearance of CD45⁺Col-Iα1⁺ cells in our model. In order to test this hypothesis, the effect of caspase inhibition on CD206/MRC⁺ alternatively activated macrophages was assessed via flow cytometry as we have previously described [³]. Results of these studies revealed only a trend toward reduced M2 macrophages in the Z/VAD-fmk treated mice that did not reach statistical significance (P = 0.09, Figure 3h). Analysis of M2-related genes such as CD206/MRC and MSR-1 using quantitative RT-PCR confirmed these results (P > 0.05 both comparisons, Figure 3i). Because caspase inhibition caused a profound reduction in CD45⁺Col-Iα1⁺ cells without a concomitant reduction in M2 macrophages, the expression of collagen by monocyte-derived cells is unlikely to be controlled solely by accumulation of M2 macrophages.

We next sought to determine the human relevance of these findings. In planning these studies we reasoned that if collagen-production in monocytes demonstrated a biological relationship with the development of fibrotic lung disease, then they would be detected in multiple forms of lung fibrosis. Thus, our murine studies were recapitulated in lung tissue from the discarded surgical margins of biopsy samples from patients with histopathological or clinical findings consistent with IPF or clinical diagnosis of connective tissue disease interstitial lung disease (CTD-ILD), or subjects with no known parenchymal lung disease. Immunohistochemistry performed on these samples revealed increased caspase 3 cleavage in the fibrotic samples (Figure 4a, b and data not shown) and a nearly twofold increase in TUNEL staining in both the IPF samples and CTD-ILD samples when compared to non-fibrotic control (P < 0.05 both comparisons, Figure 4c-e). Furthermore, while non-fibrotic lungs contained relatively low numbers of CD45⁺Pro-Col-Iα1⁺ cells, quantities of this population were increased nearly threefold in the samples with IPF and CTD-ILD (P < 0.05 for both comparisons, Figure 4e). Notably, accumulation of intrapulmonary CD45⁺Pro-Col-Iα1⁺ cells did not differ between IPF and CTD-ILD groups (Figure 4f). These data indicate that the lungs of patients with various forms of lung fibrosis demonstrate increased apoptosis and elevated numbers of CD45⁺Pro-Col-Iα1⁺ cells.

The pathogenesis of pulmonary fibrosis is thought to result in part from repeated bouts of injury leading to ongoing remodeling responses and dysregulated repair. The finding that fibrotic lung tissue is enriched for both apoptotic cell death responses and increased quantities of CD45⁺Pro-Col-Iα1⁺ cells supports an association between these processes. A mechanistic connection was explored in a human lab sample routinely accessible in clinical medicine: the peripheral blood. Here, circulating monocytes were obtained from the peripheral blood of patients with IPF and CTD-ILD, as well as of normal healthy controls, and cultured under serum-containing conditions that favor fibrocyte outgrowth [²⁷]. Characteristics of subjects are shown in Table 1. Assessment of spindle-shaped cells, which have traditionally been considered to be fibrocytes [²,⁷], revealed increased outgrowth of fibrocytes in the patients with ILD (Figure 5a, b and data not shown). Assessment of collagen expression via flow cytometry revealed that collagen expression was augmented in the subjects with IPF and CTD-ILD as well. Further analysis of phenotype revealed that total percentages of CD45⁺Pro-Col-Iα1⁺CD14^-CD34⁺ cells were quite low in cultures from all groups (P = 0.09, Figure 5c). In contrast, CD45⁺Pro-Col-Iα1⁺CD14⁺CD34^-cells were low in healthy subjects but increased by threefold to fourfold in the IPF and CTD-ILD samples (P < 0.02, Figure 5d). Percentages of CD45⁺Pro-Col-Iα1⁺CD14⁺CD34⁺ cells were low in controls but even further increased in the IPF and CTD-ILD subjects (P < 0.02, Figure 5e). Cells exhibiting expression of neither marker (CD45⁺Pro-Col-Iα1⁺CD14^-CD34^-) were rare in all subjects (P = 0.65, Figure 5f). Subgroup analysis of the CTD-ILD samples did not reveal a difference between disease subtypes (data not shown).

Finally, we determined whether caspase inhibition affected the phenotype of cultured monocytes from human subjects in the three groups. Cultured monocytes from each group were treated with 100 mM of Z-VAD/fmk or phosphate-buffered saline (PBS) control and assessed for changes in apoptosis and collagen production. Quantification of cellular apoptosis using annexin V labeling indicated a near complete eradication of apoptosis in the Z-VAD/fmk-treated cells (Figure 6a, b). These cells included cells in the early stages of apoptosis (as shown by single positive for annexin V in the right lower quadrant in Figure 6a) as well as apoptotic cells in the process of undergoing secondary necrosis (cells in the right upper quadrant in figure 6b). In addition, the accumulation of collagen-producing cells was also reduced to nearly zero in all samples (Figure 6c-e). Due to the extremely low frequency of Pro-Col-Iα1⁺ cells in these samples, further phenotyping could not be performed. These data indicate that apoptotic cell death responses promote collagen production in human monocytes and confirm the human relevance of our murine findings.

All mouse experiments were approved by the Yale School of Medicine Institutional Animal Care and Use Committee. The CC10-tTS-rtTA-TGF-β1 transgenic mice used in this study have been described [¹⁸]. These mice use the Clara cell 10-kDa protein (CC10) promoter to specifically express bioactive human TGF-β1 to the lung, and were backcrossed for > 10 generations onto a C57BL/6 background [¹⁸].

CC10-tTS-rtTA-TGF-β1 transgene positive (Tg⁺) or their wild-type littermate controls (transgene negative, Tg^-) aged 8-10 weeks old were given doxycycline 0.5 mg/ml in their drinking water for up to 2 weeks.

Mice were killed and bronchoalveolar lavage (BAL) performed as previously described [¹⁸]. Lung inflammation was assessed by assessing BAL samples as described previously [¹⁸].

Total right lung collagen was measured using the Sircol Assay following manufacturer's protocol (Biocolour, Carrickfergus, Ireland).

Lung samples were digested for flow cytometric identification of CD45⁺ Col-Iα1⁺ cells as previously described [³]. Total viable cells were quantified using Trypan blue staining. Collagen-producing leukocytes were detected using CD45 surface staining (BD Pharmingen, San Jose, CA) and intracellular staining for Col-Iα1 (Millipore, Billerica, MI). Flow cytometric analysis of CD45⁺Col-Iα1⁺ cells was performed by identifying live cells based on forward and side scatter characteristics, gating on the CD45⁺ cells, and then gating on the Col-Iα1⁺ cells within this population. Cells were then further subgated based on their expression of CD34 and/or CD14. Percentages of live cells coexpressing these markers were multiplied by total viable cell count of digested sample to determine the absolute number of collagen containing leucocytes.

TUNEL was performed as previously described [³].

Detection of caspase cleavage and activation using immunohistochemistry was performed as previously described [¹⁸].

Flow cytometric assessment of annexin V externalization was performed via flow cytometry as previously described [¹⁸].

Formalin-fixed and paraffin-embedded lung sections were stained with hematoxylin and eosin to assess gross morphology or Mallory's trichrome stains to visualize collagen deposition.

All studies were performed with HIC approval and written informed consent at Yale University School of Medicine. Individuals with no known medical conditions who self-identified as healthy were included as controls. Patients with SSc-ILD or amyopathic antisynthetase syndrome according to American College of Rheumatology (ACR) criteria or IPF according to current European Respiratory Society (ERS)/American Thoracic Society (ATS) criteria [⁴⁰] were recruited as the study group. Exclusion criteria included concurrent diagnosis of malignancy, pregnancy, the presence of known secondary lung disease such as pulmonary hypertension or chronic airway obstruction or inability (due to psychiatric or language limitations) to provide informed consent. A total of 30 ml of peripheral blood was drawn, peripheral blood mononuclear cells (PBMCs) isolated via density gradient centrifugation and CD14+ monocytes were enriched as previously described by our group [²⁶]. Cells were cultured in 96 well plates in the presence or absence of 100 μM Z-VAD/fmk (EMD Biochemical, Los Angeles, CA). After 10 days of culture cells were assessed qualitatively for fibrocytes based on spindle-shaped morphology. Cells were then harvested and assessed for CD45⁺Col-Iα1⁺ phenotype by fluorescence-activated cell sorting (FACS) as previously described [⁶].

Antibodies against human CD45, CD34, CD14, and appropriate isotype controls were obtained from BD Pharmingen. Flow cytometry and cell sorting was performed using a BD FACSCalibur. Data were analyzed using Flow Jo v 7.5 software (TreeStar, Inc, Ashland, OR). For all analyses, isotype control staining was subtracted from true antibody staining to determine the percentage of positive cells.

Gaussian distribution of data was determined using the D'Agostino and Pearson omnibus normality test. Normally distributed data are expressed as means ± SEM and assessed for significance by Student's t test or analysis of variance (ANOVA) as appropriate. Data that were not normally distributed were assessed for significance using the Mann-Whitney U test where appropriate.

AMAS = amyopathic antisynthetase syndrome; CT = computed tomography; CTD = connective tissue disease; DLCO = diffusing capacity of the lung for carbon monoxide; FVC = forced vital capacity; IPF = idiopathic pulmonary fibrosis; ILD = interstitial lung disease; N/A = not applicable; NS = not significant; NSIP = non-specific interstitial pneumonia; SSc = systemic sclerosis; UIP = usual interstitial pneumonia.