Fibrinogen mediates cadmium-induced macrophage activation and serves as a predictor of cadmium exposure in chronic obstructive pulmonary disease

Abstract

The etiologies of chronic obstructive pulmonary disease (COPD) remain unclear. Cadmium (Cd) causes both pulmonary fibrosis and emphysema; however, the predictors for Cd exposure and the mechanisms by which Cd causes COPD remain unknown. We demonstrated that Cd burden was increased in lung tissue from subjects with COPD and this was associated with cigarette smoking. Fibrinogen levels increased markedly in lung tissue of patients with smoked COPD compared with never-smokers and control subjects. Fibrinogen concentration also correlated positively with lung Cd load, but inversely with the predicted % of FEV1 and FEV1/FVC. Cd enhanced the secretion of fibrinogen in a cdc2-dependent manner, whereas fibrinogen further mediated Cd-induced peptidylarginine deiminase 2 (PAD2)-dependent macrophage activation. Using lung fibroblasts from CdCl₂-treated Toll-like receptor 4 (TLR4) wild-type and mutant mice, we demonstrated that fibrinogen enhanced Cd-induced TLR4-dependent collagen synthesis and cytokine/chemokine production. We further showed that fibrinogen complexed with connective tissue growth factor (CTGF), which in turn promoted the synthesis of plasminogen activator inhibitor-2 (PAI-2) and fibrinogen and inhibited fibrinolysis in Cd-treated mice. The amounts of fibrinogen were increased in the bronchoalveolar lavage fluid (BALF) of Cd-exposed mice. Positive correlations were observed between fibrinogen with hydroxyproline. Our data suggest that fibrinogen is involved in Cd-induced macrophage activation and increases in fibrinogen in patients with COPD may be used as a marker of Cd exposure and predict disease progression.

INTRODUCTION

Environmental pollution, such as particulate matter (PM), and cigarette smoke are two main sources for cadmium (Cd) exposure (1–3). Cd is enriched in PM2.5 (2) and exposure to PM2.5 exacerbates fibrosis and chronic obstructive pulmonary disease (COPD) (4). Specially, clinical manifestations of coal workers’ pneumoconiosis include diffuse fibrosis and chronic airway diseases, such as emphysema, a primary form of COPD (5). Cigarette smoke is considered an independent risk factor for COPD (6) which is characterized by macrophage-driven chronic inflammation and tissue destruction (7). Cigarette smoke has also been associated with several pathological features of COPD, such as pulmonary emphysema (6) and peribronchiolar fibrosis (8, 9). Peribronchiolar fibrosis contributes importantly to loss of airways and later airway remodeling (10–12), which often presents with emphysema (13).

Cd deposits mainly in lung tissues (3) and is significantly higher in stage 4 COPD (14). We have demonstrated that a low dose of Cd causes peribronchiolar fibrosis and phosphorylated vimentin plays important roles in Cd-induced peribronchiolar fibrosis and airway remodeling (12). On the other hand, Cd also induces airspace enlargement in pigs, hamsters, and mice (15, 16).

Several biomarkers have been reported which have diagnostic or prognostic values in COPD, such as plasma fibrinogen (17–19), C-reactive protein (17, 19), interleukin 8 (IL-8) (17, 19), and matrix metalloproteinase-9 (19). Within these biomarkers, plasma fibrinogen is a promising biomarker which is currently being evaluated by the US Food and Drug Administration (20). However, the mechanisms underlying the increased fibrinogen/fibrin accumulation in COPD are not clear.

Accumulation and persistence of fibrinogen/fibrin have been linked to the development of many lung diseases including adult respiratory distress syndrome (21), idiopathic pulmonary fibrosis (IPF) (22), pneumocystis carinii pneumonia (23), and COPD (17, 18). Fibrinogen/fibrin is present in the alveolar space of bleomycin-exposed mice (21) and can induce fibroblast proliferation (24, 25) and collagen deposition (25). Plasminogen activator inhibitor-1 (PAI-1) inhibits the activity of urokinase plasminogen activator (uPA), a major PA in alveolar space, and thereby the activation of plasminogen, leading to impairment of fibrinolysis (21). PAI-1 level is also significantly higher in COPD serum (26) and fibrotic lungs (27). In contrast with PAI-1 that is mainly produced from endothelial (28) and epithelial cells (29, 30), PAI-2 originates from monocytes and macrophages (31). PAI-2 can be upregulated by lipopolysaccharide and transforming growth factor-β1 (TGF-β1) and is found to be highly expressed in lung tissue from subjects with IPF (22), suggesting an important role of PAI-2 in fibrotic diseases.

The implications of fibrinogen in metal-induced lung diseases remain largely unclear. We therefore investigated whether fibrinogen was highly expressed in subjects with COPD and could be induced by Cd exposure in an experimental mouse model of COPD. Furthermore, we examined the hypothesis that Cd exposure disrupts the normal balance among coagulation and fibrinolysis, leading to fibrin deposition/formation and that overproduction of fibrinogen promotes Cd-induced macrophage activation and Toll-like receptor 4 (TLR4)-dependent collagen accumulation and profibrotic cytokine production.

MATERIALS AND METHODS

Human Subjects

Lung tissues (n = 37, including 4 COPD never smoker, 24 ex-smokers, and 9 active smokers) were obtained from diseased parts removed from lung transplantation for advanced COPD or by video-assisted thoracic surgery for suspected lung lesions in the University of Alabama at Birmingham (UAB) Clinic with approval from the institutional review board (IRB). Age-matched control specimens (n = 16, including 11 never-smoker and 5 control smoker without airflow limitation) were carefully excised from the resections or lobectomy during the surgery, or the uninvolved lobes of control subjects during thoracic surgery who did not have any lung parenchymal abnormality. Effective informed consent form (IRB-300002249) was obtained for human studies. Subjects with COPD (without α1-antitrypsin deficiency) were defined with baseline postbronchodilator forced expiratory volume in 1 s (FEV1) and FEV1/forced vital capacity (FVC) < 0.7 and COPD severity was graded according to the Global Initiative for COPD (GOLD) staging system (6). Lung function parameters including FEV1, FEV1/FVC, and diffusing capacity for carbon monoxide (DLCO) were performed. Cd concentrations were evaluated by inductively coupled plasma mass spectrometry (ICP-MS) (12) and primary human lung macrophages were purified and cultured as described previously (32).

Cd-Associated Mouse Models

Animal experiments were approved by the Institutional Committee of Animal Care of the UAB. Experimental groups were randomly assigned and equal number of female and male mice (6- to 8-wk-old) was selected in this study. To evaluate the role of peptidylarginine deiminase 2 (PAD2) in Cd or fibrinogen-induced macrophage activation, PAD2 wild type (WT) or PAD2^−/− mice (kindly provided by Scott A. Coonrod, Cornell University, Ithaca, NY) were treated by intratracheal instillation (IT) once with cadmium chloride (CdCl_2, Sigma-Aldrich, CAS 10108-64-2, 0.458 mg/kg). Macrophages from BALF were isolated from mice at day 3 and treated with CdCl₂, fibrinogen (Sigma-Aldrich, F4883) or combination. C3H/HeOuJ TLR4 WT and C3H/HeJ TLR4 mutant (MUT, a spontaneous mutation in the TLR4 gene, Tlr4^lps-d) mice purchased from the Jackson Laboratory were treated as above and mouse lung fibroblasts were isolated and cultured as previously reported (12) to assess the function of fibrinogen in Cd-induced collagen synthesis and cytokine/chemokine production. For the blockade of connective tissue growth factor(CTGF) signal, mice were administered with anti-CTGF antibody (Abcam, ab209780, 0.1 mg/mouse) in 50 µL of PBS via IT 2 h after treated with CdCl₂. To make an experimental mouse model of COPD, mice were administrated with saline or CdCl₂ (0.458 mg/kg) by IT. Mouse lungs, BALF, and plasma were collected at days 3, 7, 14, 21, and 28. For BALF, three separate aliquots of 1 mL of PBS were performed for each mouse. Lung tissues from mice were dried, acid hydrolyzed and hydroxyproline concentration was examined as manufacture’s protocols (Chondrex, 6017). To determine the mean alveolar airspace (MAA), morphometric analysis of H&E sections was performed as described (33).

Immunoblot Analysis and Immunoprecipitation

Cell lysates from primary human, mouse lung macrophages or U937 monocytes/macrophages (ATCC, CRL-1593.2), and concentrated supernatants using an Amicon Ultra-15 centrifugal filter units (Millipore Sigma, UFC901008) were resolved on 4%–12% gels (Bio-Rad) and transferred to nitrocellulose membrane (Bio-Rad, 1620112). After blocking, membranes were incubated with primary antibodies. Secondary antibodies were HRP-conjugated anti-rabbit, goat or anti-mouse antibody and images were developed using a Bio-Rad system. The following inhibitors, proteins, or primary antibodies were used: cell division cycle protein 2 (cdc2) kinase inhibitor roscovitine (Santa Cruz, CAS 186692-46-6); PAD2 inhibitor AFM-30a (a gift from Paul R. Thompson, University of Massachusetts Medical School, Worcester, MA); recombinant human CTGF protein (active) (Abcam, ab269222); anti-PAD2 (Abcam, ab16478); anti-β-actin (Sigma-Aldrich, A5316); anti-human fibrinogen (Invitrogen, PA1-26809); anti mouse fibrinogen (Abcam, ab27913); anti-PAI-1 antibody (Abcam, ab222754); anti-PAI-2 antibody (Abcam, ab137588); anti-GAPDH (Sigma-Aldrich, G8795); and anti-CTGF antibody (Abcam, ab209780). For immunoprecipitation, cell lysates were precleared with protein A/G agarose (Santa Cruz, sc-2003) for 2 h at 4°C and supernatants were incubated with 2 µg of anti-fibrinogen or anti CTGF overnight in RIPA buffer (Sigma-Aldrich, R0278) containing 1% protease inhibitor cocktail (Sigma-Aldrich, 539134) and 1% phosphatase inhibitor cocktail (Sigma-Aldrich, 524625) before performing immunoblot analysis.

Measurements of Fibrinogen, PAI-2, and Cytokines/Chemokines

ELISA was used to determine the concentrations of fibrinogen, PAI-2, monocyte chemoattractant protein 1 (MCP-1), keratinocyte-derived chemokine (KC), CTGF, and active TGF-β1 according to the manufacturer’s instructions. Human fibrinogen was detected using the ELISA kit from Abcam (ab108841) and mouse fibrinogen kit was from ICL Lab (E-90FIB). Mice PAI-2 was quantified using a kit from Amsbio (AMS.E03P0057). ELISA kit for mouse MCP-1 was provided by the BD Cytometric Bead array (BD Biosciences, 552364), for mouse KC by R&D Systems (DY453), for mouse CTGF by Lifespan Biosciences (LS-F21342-1), and for active TGF-β1 by BioLegend (437707).

Collagen Expression in Lung Biopsies

To evaluate the expression of collagen, lung sections from human subjects or mice were fixed with formalin, embedded with paraffin and immunostaining was performed using Diaminobenzidine (DAB) staining kit (Dako, K5361). Immunofluorescence was quantified using Bioquant Imaging Analysis software and the image density in each subject was calculated from 10 random areas.

Knockdown of Fibrinogen α, β, and γ by siRNA

All siRNAs were purchased from Thermo Fisher Scientific including fibrinogen α (s5115), fibrinogen β (s5119), and fibrinogen γ (s5179) as well as scrambled control siRNA (4390843). Before siRNA transfection, U937 cells (4 × 10⁵/mL) were resuspended in six-well plates with 1.7 mL of RPMI 1640 complemented with 10% fetal bovine serum without antibiotics. Sock solutions of each siRNAs (25 µM) were solubilized in nuclease-free water provided by the manufacturer. About 50 pmol of each siRNA duplex (a triple combination, total 150 pmol) was diluted in 150 µL Opti-MEM I Reduced Serum Medium (Thermo Fisher Scientific, 31985062) before mixing with an equal volume of Opti-MEM containing Lipofectamine RNAiMAX (6 µL, Thermo Fisher Scientific, 13778030). After 10 min of incubation at room temperature, 300 µL of the complex of siRNA/RNAiMAX was added to cells in a final volume of 2 mL medium. The cells transfected with a scrambled siRNA were used as negative control. Forty eight hours after transfection, cells were treated or not with CdCl₂ (1.8 µg/mL) for 2 h and processed for immunoblot analysis.

Detection of Phagocytosis

For phagocytosis analysis, macrophages isolated from CdCl₂-exposed WT or PAD2^−/− mice were cultured with fibrinogen (200 µg/mL) for 24 h and stimulated with FITC-coupled zymosan particles (25 µg/mL, Invitrogen, Z2841) for 18 h in the presence or absence of cytochalasin (10 µM, Sigma, C8273). The cells were then stained with macrophages marker, F4/80-APC (BioLegend, 123116) before analysis by flow cytometry. Data were acquired with a LSR II (BD Bioscience) and analyzed using Flowjo software (v. 10.6.1).

Quantitative Real-Time PCR

Total RNA was extracted from mouse lung macrophages or fibroblasts using TRIzol RNA Isolation Reagent (Thermo Fisher Scientific, 15596018) and cDNA was synthesized from RNA using high-capacity cDNA Reverse Transcription kit with RNase inhibitor according to the manufacturer’s protocol (Thermo Fisher Scientific, 4374966). RT-PCR were employed to measure the transcript levels of the mice Col1a1, Col3a1, fibrinogen α, β, and γ genes and performed as described previously (34) using an SYBR Green PCR Master Mix (Thermo Fisher Scientific, 4309155). The following primer pairs were used: mouse Col1a1, 5′-GTC, TGG, TTT, GGA, GAG, AGC, AT-3′ and 5′-CTT, CTT, GAG, GTT, GCC, AGT, CT-3′; mouse Col3a1, 5′-TGA, TGT, CAA, GTC, TGG, AGT, GG-3′ and 5′-TCC, TGA, CTC, TCC, ATC, CTT, TC-3′; mouse fibrinogen α, 5′-TGA, TTC, CGA, CAT, CCT, CAC, AA-3′ and 5′-GGG, CAC, CTG, AAG, TTT, GTG, TT-3′; mouse fibrinogen β, 5′-TCT, TCA, GCA, CGT, ACG, ACA, GG-3′ and 5′-GAG, TAC, CAC, GAT, CCC, TTC, CA-3′; mouse fibrinogen γ, 5′-CAC, CCA, GAC, ACC, ATG, AGT, TG -3′ and 5′- TCA, GCC, CGG, AAT, AAG, ATG, TC-3′; and mouse β-actin, 5′-AGC, CAT, GTA, CGT, AGC, CAT, CC-3′ and 5′-TCT, CAG, CTG, TGG, TGG, TGA, AG-3′. β-Actin was used as a reference gene to normalize the data and the arbitrary unites were calculated using the cycle threshold method.

Measurement of Fibrinolytic Activity

BALF was obtained from mice treated with saline or CdCl₂ along with or without anti-CTGF antibody. A fibrin clot analysis assay was used to examine the fibrinolytic activity as previously described (35). Briefly, after centrifugation, 50 µL of BALF was added to 96-well plates and mixed with an equal volume of plasminogen-depleted human plasma fibrinogen (2 µM, Sigma-Aldrich, F4883) and human glu-plasminogen (0.5 µM, Thermo Fisher Scientific, PIRP43078). After adding 50 µL of human thrombin (0.2 U/mL, Abcam, ab269042), the plates were incubated at 37°C and O.D values at 405 nm were monitored. The standard curve was generated by serial dilution of the human uPA (Abcam, ab167714) and the rate of fibrinolytic activity was evaluated by using t1/2 which indicated the time required for 50% of the clot to be degraded.

Respiratory Mechanics Analysis

Mice were treated by IT with saline or CdCl₂ (0.458 mg/kg). At days 3, 7, 14, 21, and 28, mice were anesthetized and respiratory mechanics analysis was performed as previously described (34). Static compliance (Cst) was determined using the PV-Loop Salazar Knowles equation and airway resistance (Rrs) was evaluated using the single-compartment model.

Statistics

Data were analyzed by two-tailed t test for two group comparisons and ANOVA followed by Tukey’s post hoc test for three or more group’s comparisons. To investigate the associations between variables, spearman rank correlations were used for the analysis. Significance was accepted when a P value was < 0.05. Values were shown as means ± SE unless specified.

RESULTS

The Amount of Fibrinogen Is Increased in Lung Tissue from Subjects with COPD and Correlates with Cd Load and Lung Function

Increasing reports have demonstrated that urinary and blood Cd concentrations do not reflect the Cd levels in the target organs (3, 14). Thus, we first assessed the concentration of Cd in the lung, in subjects with COPD. ICP-MS analysis revealed that the content of Cd was significantly increased in the lungs of patients with COPD compared with the control lungs (Fig. 1A). Furthermore, we showed that Cd content in the lung tissue of active smokers (AS) was markedly higher than that in ex-smokers (Ex-S) and patients with COPD who never smoked (Fig. 1B). Cd content in patients with COPD who never smoked was also notably higher compared with the control never smokers (NS), suggesting that cigarette smoke was not the only source for Cd exposure in COPD (Fig. 1B). The clinical characteristics and smoking status analysis of subject with COPD are summarized in Tables 1 and 2.

Figure 1.

Cd accumulates highly in lung tissue from subjects with COPD, especially those who smoked. Cd contents by ICP-MS from lung tissues for A (control, n = 16; COPD, n = 37) and B [control never smokers (NS), n = 11; control smokers (S), n = 5; COPD never smokers, n= 4, ex-smokers (ex-S), n = 24, and active smokers (AS), n = 9]. *P < 0.05, **P < 0.01; ***P < 0.001 using two-tailed t test for (A) and ANOVA followed by Tukey’s post hoc analysis for B. Cd, cadmium; COPD, chronic obstructive pulmonary disease; n, represents subject number; ICP-MS, inductively coupled plasma mass spectrometry.

Table 1.

Clinical characteristics of the subjects with COPD

		Lung Tissue
	Control	COPD	P Value
Number	16	37
Age, yr	62 ± 10	67 ± 8	0.450
Male, %	25	73	0.001
White, %	77	87	0.110
Smoker, %a	31	89	<0.001
FEV1, %	82 ± 24	71 ± 23	0.021
FEV1/FVC, %	81 ± 13	62 ± 14	<0.001
DLCO, %	80 ± 24	65 ± 21	0.001
GOLD
I	2	8
II	1	10
III	1	8
IV	0	1

Values are means ± SD. ^aSmoker: active smokers (AS) or ex-smokers (ex-S, have not smoked in the last 28 days) with at least a 5 pack-year smoking history; FEV1 (%): % of predicted; FEV1/FVC (%): % of predicted; DLCO (%): % of predicted; GOLD, Global Initiative for COPD; COPD, chronic obstructive pulmonary disease; DLCO, diffusing capacity for carbon monoxide; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity.

Table 2.

Smoking status analysis of subjects with COPD

	Control (Lung Tissue)				COPD (Lung Tissue)
	NS	S	P Value	NS	Ex-S	AS	Ex-S+AS	P Value
Number	11	5		4	24	9	33
Age, yr	61 ± 12	64 ± 6	0.560	73 ± 6	67 ± 8	66 ± 6	67 ± 7	0.210a
Male, %	9	60	<0.001	50	79	67	76	0.001a
White, %	73	83	0.070	100	83	89	85	0.061a
FEV1, %	89 ± 22	66 ± 23	<0.001	85 ± 24	65 ± 23	77 ± 20	68 ± 23	<0.001a
FEV1/ FVC, %	84 ± 11	70 ± 17	0.006	67 ± 12	61 ± 16	63 ± 11	61 ± 15	0.039a
DLCO, %	82 ± 22	71 ± 30	0.023	86 ± 12	60 ± 16	64 ± 25	61 ± 19	0.009a
GOLD
I	1	1		1	4	2	6
II	1	0		1	8	1	9
III	0	1		0	5	2	7
IV	0	0		0	1	0	1

Values are means ± SD. Smoker: active smokers (AS) or ex-smokers (Ex-S, have not smoked in the last 28 days) with at least a 5 pack-year smoking history; No significant differences between Ex-S and AS. ^aEx-S+AS vs. NS; FEV1 (%): % of predicted; FEV1/FVC (%): % of predicted; DLCO (%): % of predicted. GOLD, Global Initiative for COPD; COPD, chronic obstructive pulmonary disease; DLCO, diffusing capacity for carbon monoxide; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity.

Relatively higher amounts of plasma fibrinogen in the subjects with COPD (17, 36) prompted us to investigate its source. For the first time, we found that lung fibrinogen, including intact fibrinogen and individual fibrinogen α, β, and γ, was highly expressed in macrophages from COPD (Fig. 2, A and B). An ELISA analysis showed higher amounts of fibrinogen in lung tissue lysates from subjects with COPD smokers, especially active smokers (Fig. 2, C and D). Importantly, fibrinogen content positively correlated with Cd load, both Cd and fibrinogen correlated inversely with the predicted % of FEV1 (Fig. 2, F and H) and FEV1/FVC (Fig. 2, G and I), suggesting the predictive role of fibrinogen in the progression of COPD.

Figure 2.

Fibrinogen positively correlates with Cd load, whereas both Cd as well as fibrinogen negatively correlate with lung function in COPD. A: immunoblot analysis of fibrinogen in lung macrophages isolated from control and subjects with COPD. B: quantification of fibrinogen expression from (A). C: fibrinogen amounts by ELISA in lung tissues and analyzed by smoking status (D). E: correlation analysis of Cd and fibrinogen in lung tissue from subjects with COPD (n = 37). Correlation analysis of lung tissue Cd load with predicted % of FEV1 (F) and FEV1/FVC (G) as well as lung tissue fibrinogen with predicted % of FEV1 (H) and FEV1/FVC (I). P and r values were determined using Spearman rank correlations. ** P < 0.01; ***P < 0.001 using two-tailed t test for B and C and ANOVA followed by Tukey’s post hoc analysis for D. Cd, cadmium; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity.

We further confirmed the appearance of fibrosis and emphysema in the same patients with COPD (Supplemental Fig. S1; all Supplemental Material is available at https://doi.org/10.6084/m9.figshare.19342841). Collagen deposition was more notable in the right lower lobe (RLL) (Supplemental Fig. S1, B and C), whereas airspace enlargement was more substantial in the right upper lobe (RUL) (Supplemental Fig. S1, B and D), indicated by the increased MAA, a measure of pulmonary airspace enlargement. Note that signs of peribronchiolar fibrosis were more significant in the subjects with COPD compared with the control (Supplemental Fig. S1E).

Fibrinogen Mediates Cd-Induced Macrophage Activation

Alveolar macrophages are the first line of defense following inhaled toxin exposure, and Cd exposure increases macrophage number in mouse bronchoalveolar lavage fluid (BALF) (34, 37, 38). We next explored the mechanisms of fibrinogen induction using lung macrophages isolated from subjects with COPD. Our previous studies have shown that Cd induced oxidative stress, demonstrating the important role of redox regulation in Cd-induced fibrosis (12, 34). Here, we demonstrated that Cd could cause a noticeable increase of fibrinogen and PAI-2 (but not PAI-1) (Fig. 3A). The induction of fibrinogen and PAI-2 by Cd was blocked by cdc2 inhibitor, roscovitine (RO), suggesting that cdc2 signaling was required for the induction of fibrinogen and PAI-2. We further demonstrated that Cd induced the secretion of fibrinogen and pretreatment with a cdc2 inhibitor could effectively reverse this action (Fig. 3, B and C).

Figure 3.

Fibrinogen mediates Cd-induced PAD2-dependent macrophage activation. A: immunoblot analysis of macrophages isolated from subjects with COPD treated with or without CdCl₂ (1.8 µg/mL) for 2 h. Cells were pretreated with vehicle (Veh), AFM-30a (5 µM) overnight, or roscovitine (RO, 10 µM) for 2 h. Supernatants were concentrated and analyzed by immunoblot analysis at different time points by CdCl₂ (B) or pretreatment with AFM-30a overnight (C) or RO for 2 h, followed by CdCl₂ for 48 h. D: immunoblot analysis of U937 cells treated with negative control (NC)-siRNA or fibrinogen-α, β, and γ siRNA for 2 days and challenged with or without CdCl₂ for 2 h. Mouse BALF macrophages were isolated from WT or PAD2^−/− mice treated intratracheally with CdCl₂ (0.458 mg/kg body weight) for 3 days. The isolated macrophages (5 × 10⁵/mL) were then treated with increased amounts of fibrinogen (FBG) in the presence of CdCl₂ (1.8 µg/mL) (FBG+Cd) for 24 h. For control conditions, macrophages were cultured only with increased amounts of CdCl₂ at 0, 0.45, 0.90, 1.8, 3.6, 7.2 µg/mL (Cd), or only with increased amounts of fibrinogen at 0, 100, 200, 400, 600, and 800 µg/mL (FBG), respectively. The concentrations of MCP-1 (E and F), active TGF-β1 (G and H), and KC (I and J) in the supernatants were measured by ELISA. After culture with fibrinogen (200 µg/mL) for 24 h, macrophages were stimulated with FITC-labeled zymosan (25 µg/mL), and phagocytosis was measured 18 h later in the presence (dashed line) or absence of phagocytosis inhibitor, cytochalasin D (Cyto. D, 10 µM) (solid line). K: representative histograms on flow cytometry with the mean fluorescence intensity (MFI) for each condition. Data shown are representative of three experiments. L: quantification of zymosan-FITC using the MFI ratio (MFIR) calculated by dividing the MFI in the absence of cytochalasin D (solid line) by the MFI of isotype control (gray-filled histogram), respectively. *P < 0.05; **P < 0.01; ***P < 0.001 vs. dose 0 and ##P < 0.01; ###P < 0.001 vs. CdCl₂ only using ANOVA followed by Tukey’s post hoc analysis. BALF, bronchoalveolar lavage fluid; Cd, cadmium; CdCl₂, cadmium chloride; COPD, chronic obstructive pulmonary disease; MCP-1, monocyte chemoattractant protein 1; PAD2, peptidylarginine deiminase 2; TGF-β1, transforming growth factor-β1; KC, keratinocyte-derived chemokine.

PAD2, a citrullinating enzyme, is highly expressed in COPD (39). We next examined whether Cd could induce the expression of PAD2 in macrophages isolated from subjects with COPD. As shown in Fig. 3, A and C, PAD2 could be induced by Cd and pretreatment with a PAD2 inhibitor, AFM-30a, showed no effects on the expression and secretion of fibrinogen. We further investigated whether Cd-induced PAD2 could be blocked when the fibrinogen gene was silenced. Three isoforms of fibrinogen were simultaneously knocked down using siRNA and then cells were treated with CdCl₂. Immunoblot analysis showed that knockdown of fibrinogen markedly decreased the expression of PAD2 compared with scrambled negative control (NC) siRNA-transfected cells (Fig. 3D). To analyze the role of PAD2 in fibrinogen-induced macrophages activation, macrophages were isolated from BALF of PAD2 WT or PAD2^−/− mice and treated with CdCl₂, fibrinogen, or a combination. Our data showed that fibrinogen or CdCl₂ treatment alone caused noticeable production of MCP-1, active TGF-β1, and KC only from WT macrophages, whereas the combination of fibrinogen and CdCl₂ produced significantly more compared with CdCl₂ alone (Fig. 3, E–J). We further evaluated the phagocytic ability of macrophages by adding FITC-labeled zymosan after fibrinogen exposure for 24 h. Compared with the control, fibrinogen-exposed macrophages phagocytosed more zymosan particles, whereas the increased phagocytic capacity by fibrinogen exposure was markedly reduced in the macrophages isolated from PAD2^−/− mice (Fig. 3, K and L). Our data suggest that fibrinogen may be involved in Cd-induced macrophage activation in a PAD2-dependent manner.

Fibrinogen Enhances Cd-Induced TLR4-Dependent Collagen Synthesis and Cytokine/Chemokine Production

We next examined whether fibrinogen mediated Cd-induced collagen synthesis and cytokine/chemokine production. The effects of reduced fibrinolysis were imitated by the addition of α2-antiplasmin (α2-AP), a specific plasmin inhibitor, to mouse lung fibroblasts isolated from TLR4 WT or TLR4 MUT mice. Treatment with α2-AP, fibrinogen, or CdCl₂ alone induced a significant increase in Col1a1 and Col3a1 mRNA expression and combination of both or three of these agents produced more Col1a1 and Col3a1 (Fig. 4, A and B). Furthermore, in the presence of CdCl_2, collagen mRNA was found highly expressed with increased doses of fibrinogen in fibroblasts isolated from TLR4 WT, but not from TLR4 MUT mice (Fig. 4, C–F). Analysis of cytokines showed that in response to fibrinogen stimulation, production of CTGF and TGF-β1 was also increased notably in fibroblasts isolated from TLR4 WT mice (Fig. 4, G–J). Compared with fibrinogen alone, combined treatment of fibrinogen, α2-AP, and CdCl₂ produced more collagen and cytokines (Fig. 4, C, E, G, and I). These data suggest that by reducing fibrinolysis, over accumulation of fibrinogen/fibrin could enhance Cd-induced TLR4-dependent collagen synthesis and cytokine/chemokine production.

Figure 4.

Fibrinogen enhances Cd-induced TLR4-dependent collagen synthesis and regulates cytokine/chemokine production in Cd-treated mouse lung fibroblasts. Mouse lung fibroblast were isolated from TLR4 WT or TLR4 MUT mice treated intratracheally with CdCl₂ (0.458 mg/kg body weight) for 3 days. Col1a1 (A) and Col3a1 (B) mRNA levels in WT mouse fibroblasts stimulated with α2-AP (1 µM), fibrinogen (FBG, 200 µg/mL), CdCl₂ (1.8 µg/mL), or combination of them for 48 h. The isolated fibroblast (5 × 10⁵/mL) were treated with increased amounts of fibrinogen only (FBG), or in the presence of CdCl₂ (1.8 µg/mL) (FBG+Cd) or α2-AP plus CdCl₂ (FBG+α2-AP+Cd). mRNA expression of Col1a1 in WT (C), Col1a1 in TLR4 MUT (D), Col3a1 in WT (E), and Col3a1 in TLR4 MUT (F) mouse fibroblasts treated with indicated stimuli for 48 h. The concentrations of CTGF in WT (G), CTGF in TLR4 MUT (H), active TGF-β1 in WT (I), and active TGF-β1 in TLR4 MUT (J) mouse fibroblasts were measured by ELISA after different treatments for 24 h. *P < 0.05; **P < 0.01; ***P < 0.001 vs. control or dose 0 and ##P < 0.01; ###P < 0.001 vs. FBG only using ANOVA followed by Tukey’s post hoc analysis. Cd, cadmium; CdCl₂, cadmium chloride; CTGF, connective tissue growth factor; MUT, mutant; TGF-β1, transforming growth factor-β1; TLR4, Toll-like receptor 4; WT, wild type.

Cd-Induced CTGF Is Responsible for Fibrinogen Synthesis in Lung Tissues

The induction of CTGF by fibrinogen prompted us to investigate the pathologic roles of CTGF in Cd-induced fibrinogen synthesis and reduced fibrinolysis. We found that CTGF enhanced Cd-induced fibrinogen and PAI-2 expression, whereas anti-CTGF antibodies blocked this action (Fig. 5A). To support the hypothesis that CTGF binds with fibrinogen and cooperatively regulates fibrinogen synthesis and function, immunoprecipitation was performed either with CTGF or fibrinogen antibodies and immunoblot analysis demonstrated that CTGF formed a complex with fibrinogen in response to Cd stimulation (Fig. 5B). We further investigated the effects of blockade of CTGF on the production of fibrinogen in vivo. Treatment with anti-CTGF antibodies significantly attenuated Cd-induced fibrinogen α, β, and γ mRNA expression (Fig. 5, C–E). The amounts of PAI-2 were highly expressed in lung macrophages and peaked at day 3 after CdCl₂ treatment (Supplemental Fig. S2) although no PAI-2 was detectable in BALF (data not shown). Likewise, the increased amounts of fibrinogen in BALF and PAI-2 in lung macrophages were significantly inhibited in the presence of anti-CTGF antibodies (Fig. 5, F and G). In addition, the fibrinolytic activity was evaluated using a clot lysis assay. Cd-induced lung injury caused marked inhibition of clot lysis indicated by the longer time required for 50% clot lysis (t1/2), whereas the BALF from anti-CTGF-treated mice demonstrated rapid clot lysis with the average of mice having t1/2 values of < 500 min (Fig. 5H). Furthermore, anti-CTGF antibody treatment also mitigated CdCl₂-induced lung fibrosis indicated by the decreased Col1a1 and Col3a1 mRNA expression as well as hydroxyproline concentrations (Supplemental Fig. S3). Overall, these data suggest that Cd-induced fibrinogen synthesis and inhibition of fibrinolytic activity in lungs are mediated by CTGF.

Figure 5.

Blockade of CTGF attenuates Cd-induced inhibition of fibrinolytic activity and synthesis of PAI-2 and fibrinogen. A: immunoblot analysis of U937 cells treated with or without CdCl₂ (1.8 µg/mL) for 2 h in the presence or absence of active recombinant CTGF (2 µg/mL) or anti-CTGF (10 µg/mL). B: immunoblot analysis of macrophages treated with or without CdCl₂ for 2 h. Before immunoblot analysis, cell lysates were immunoprecipitated with or without anti-fibrinogen or CTGF antibodies. Mice (n = 5/group) were treated intratracheally with anti-CTGF antibody (0.1 mg/mouse) in 50 µL of PBS 2 h after treated with saline or CdCl₂ (0.458 mg/kg). Mouse lungs and BALF were collected at days 3, 7, 14, 21, and 28. mRNA expression of Fibrinogen α (C), Fibrinogen β (D), and Fibrinogen γ (E) in lung macrophages at day 3. F: fibrinogen amounts by ELISA in BALF at day 7. G: PAI-2 amounts by ELISA in lung macrophages at day 3. H: fibrinolytic activity using a clot assay in BALF. Results shown as scatter plot of t1/2. *P < 0.05; **P < 0.01; ***P < 0.001 using ANOVA followed by Tukey’s post hoc analysis. BALF, bronchoalveolar lavage fluid; Cd, cadmium; CdCl₂, cadmium chloride; CTGF, connective tissue growth factor; PAI-2, plasminogen activator inhibitor-2.

Fibrinogen Increases in a Mouse Model of Cd-Induced Peribronchiolar Fibrosis and Late Stage Alveolar Enlargement

Combined pulmonary fibrosis and emphysema (CPFE) possesses distinct clinical and pathological characteristics compared with IPF and emphysema (40). Next, we examined whether airway delivery of Cd was able to induce fibrosis and alveolar enlargement in the same mice as in COPD and/or CPFE in human. The dose of CdCl₂ (0.458 mg/kg) was selected and the Cd concentration of CdCl₂-treated mice at week 4 (1.48 ng/mg protein) was similar to that in lung tissue from subjects with COPD (Fig. 1A), indicating that we were able to achieve a similar concentration of Cd in our mouse model as that found in the lung parenchyma of subjects with COPD.

We found that a single dose of CdCl₂ could induce signs of peribronchiolar fibrosis at day 3 and parenchymal fibrosis at day 7 through 28 shown by increased collagen deposition in the lungs as revealed by immunohistochemistry staining (Fig. 6, A and B) and biochemical confirmation of hydroxyproline concentrations (Fig. 6C). Collagen accumulation was increased as time of exposure to Cd increased, peaking at day 7 (Fig. 6, B and C). Quantitative analyses of MAA revealed that the alveolar space in CdCl₂-exposed mice did not change before day 7, gradually increased at day 14, and peaked at day 28 (Fig. 6D), suggesting that Cd-induced airspace enlargement was a late-stage development. Respiratory mechanics analysis demonstrated the increases in whole lung airway resistance (Fig. 6E), whereas noticeable decreases of lung compliance occurred at day 7 (Fig. 6F). The decreased lung compliance gradually recovered at day 21 and was significantly higher compared with the control at day 28.

Figure 6.

Concentrations of fibrinogen are increased in mice with Cd-induced peribronchiolar fibrosis plus later stage airspace enlargement. Mice (n = 5/group) were treated intratracheally with Sal or CdCl₂ (0.458 mg/kg). Mouse lungs, BALF, and plasma were collected at days 3, 7, 14, 21, and 28. A: representative lung histology with collagen-1 staining at different time points. Scale bars, 100 µM. B: DAB staining density from A was quantified. C: hydroxyproline content at different time points. D: analysis of MAA at different time points. Respiratory mechanics analysis was performed in paralyzed and mechanically ventilated mice following Sal or CdCl₂ administration. At different time points, the parameters of Rrs (E) and Cst (F) were assessed using the baseline values. G: immunoblot analysis of fibrinogen and PAI-2 in mouse lung macrophages. Fibrinogen amounts by ELISA in BALF (H) and plasma (I) at different time points. J: correlation analysis of BALF fibrinogen with hydroxyproline. P and r values were determined using Spearman rank correlations. Each dot represents individual mouse. *P < 0.05, **P < 0.01, ***P < 0.001 vs. 0 day using ANOVA followed by Tukey’s post hoc analysis. BALF, bronchoalveolar lavage fluid; Cd, cadmium; CdCl₂, cadmium chloride; MAA, mean alveolar airspace; Cst, static compliance; PAI-2, plasminogen activator inhibitor-2; Rrs, airway resistance.

Importantly, Cd exposure induced the protein expression of fibrinogen and PAI-2 in mouse lung macrophages (Fig. 6G) and fibrinogen amounts were also significantly increased in the BALF of “COPD mice” (Fig. 6H) whose pattern was similar to that of lung tissue collagen deposition (Fig. 6, B and C). Although fibrinogen amounts increased in peripheral blood (Fig. 6I), no correlation was found between plasma fibrinogen with lung hydroxyproline (Supplemental Fig. S4). Correlation analysis showed clearly that BALF fibrinogen was positively correlated with lung hydroxyproline (Fig. 6J), indicating the prognostic role of BALF fibrinogen in Cd-induced fibrosis and also in late-stage airspace enlargement.

DISCUSSION

Plasma fibrinogen has been considered a promising biomarker for COPD (17–20) and Cd exposure is linked to COPD pathogenesis (3, 14). We demonstrate here that the amounts of fibrinogen in lung tissue are increased in COPD subjects, especially those who smoked. Importantly, the amounts of fibrinogen correlate positively with Cd load but correlate negatively with lung function, suggesting that fibrinogen may be a marker of Cd exposure and a prognostic factor in COPD, at least in Cd-associated COPD. Here we demonstrate that lung macrophages synthesize fibrinogen polypeptides in response to Cd stimulation. Although the abundant levels of fibrinogen in mouse plasma suggest the likelihood of synthesis and secretion of fibrinogen in other organs remote from the lung, our current studies indicate that fibrinogen can be synthesized and secreted in the lung and BALF fibrinogen is more reliable to predict Cd-induced fibrosis. The mechanisms controlling the synthesis and secretion of fibrinogen are not understood. Our data suggest the upstream signal of cdc2 may be responsible for the regulation of fibrinogen expression. On the other hand, Cd-induced CTGF may also regulate the synthesis of fibrinogen/fibrin, likely by upregulating the PAI-2 expression.

Airway remodeling due to the fibrosis of small airways may occur before the presence of parenchymal fibrosis in IPF (41). Small airway disease is associated with the deterioration of lung function such as in COPD (42). COPD is a heterogeneous disease whose clinical manifestations differ in each individual (43, 44). We found that fibrosis and emphysema occurred in the same human organs of patients with COPD. In this regard, fibrosis was more likely localized in the lower lobes, whereas airspace enlargement appeared usually in the upper lobes which are similar to the findings from CPFE (40). This finding could be due to, at least in part, the difference of pulmonary airflow regional ventilation and heterogeneity among the upper and lower lobes (45). In fact, similar phenomena have also been reported in another small airway disease, respiratory bronchiolitis-associated interstitial lung disease, in which the areas of hypoattenuation (feature of small airway disease) are predominantly seen in the lower lobes, whereas airspace enlargement is frequently present in the upper lobes (46).

By using CdCl₂ aerosols, peribronchiolar fibrosis has been reported in animal models such as in rats (47). Based on our observations, we propose three potential types of involvement that could explain the development of Cd-associated COPD. First, peribronchiolar fibrosis occurs during an initial stage following instillation of CdCl₂ in mouse lungs. This may be attributed to Cd-induced PAI-2 activation and subsequent inhibition of fibrinolytic activity and consequently the accumulation of fibrinogen/fibrin. This action further creates a prothrombotic environment in the alveoli which may foster Cd-induced peribronchiolar fibrosis. In the second stage, CdCl₂ could also induce parenchymal fibrosis, featured as a diffuse involvement of the lung parenchyma, similar to that found in asbestosis and other pneumoconioses (48). Finally, our data allow for a hypothesis that Cd-associated emphysema represents a late stage of abnormalities characterized by enlarged airspaces and increased lung compliance. Static lung compliance recovered at day 21 and was markedly increased at day 28, whereas a noticeable increase in airway resistance and a significant amount of total lung hydroxyproline were still observed 7 days after Cd exposure. These data demonstrated the appearance of emphysema and fibrosis in the same Cd-exposed mice and suggested that impaired lung function were not always in good agreement with histological observations. It seems that collagen deposition extensively observed in peribronchiolar fibrosis and parenchymal fibrosis is involved in Cd-induced airspace enlargement and this hypothesis can be demonstrated by the fact that blocking of collagen deposition by treatment with a proline analogue inhibits the airspace enlargement (49). We were struck by the frequent finding of prominent fibrosis in association with emphysema in Cd-exposed mice. These data suggest that Cd-induced peribronchiolar fibrosis, parenchymal fibrosis, and late stage of emphysema mimic many of the characteristics of the fibrotic response and enlargement of airspace seen in patients with COPD and in humans after accidental Cd exposure (50).

Macrophage activation is characterized by the increased phagocytic capacity and increased production of the proinflammatory cytokines such as MCP-1, TGF-β1, and KC (mouse IL-8) (51–53). The production of both MCP-1 and KC is inhibited markedly in the PAI-1 knockout (KO) mice (53). In the current study, the induction of MCP-1, KC, and TGF-β1 only in the macrophages isolated from PAD2 WT, but not from PAD2^−/− mice, indicated that fibrinogen mediated Cd-induced macrophages activation which is PAD2-dependent. It is reasonable to speculate that PAD2, an enzyme catalyzing the posttranslational deamination of proteins, mediates Cd-induced balance disorder of coagulation and fibrinolysis.

Increased production of collagen mRNA in the presence of plasmin inhibitor or fibrinogen suggests that inhibition of plasmin-mediated fibrin breakdown contributes to Cd-induced increased level of fibrinogen/fibrin and subsequent collagen synthesis. Lung dysfunction can be induced due to the deposition of extravascular fibrin (53) in which PAI-1 plays predominant role in these injuries, especially acute lung injuries (30, 54, 55). TGF-β enhanced the expression of PAI-1 and blockage of PAI-1 by using PAI-1 RNAi or PAI-1 inhibitor attenuated bleomycin or TGF-β-induced lung fibrosis (27, 56, 57). In PAI-1 KO mice, less fibrin and cytokines were seen which was caused by increased lysis of endogenous fibrin (53). Cd can induce the expression of PAI-1 in human vascular endothelial cells (58). Our data showed that Cd enhanced the expression of PAI-2, but not PAI-1, in human macrophages and Cd-exposed mouse macrophages, indicating that PAI-2 may be a regulator of the overexpression and secretion of fibrinogen/fibrin from macrophages except for the vascular leak due to ongoing lung injury. Increase in PAI-2’s anti-protease activity may lead to the deposition of fibrinogen/fibrin and collagen’s accumulation due to decreased plasmin-mediated activation of matrix metalloproteinase. Similar to PAI-1 (59), PAI-2 may also bind the matrix protein, vitronectin, and this binding could strengthen the anti-protease activity of PAI-2. Considering the higher expression of PAI-2 in BALF cell lysate of patients with IPF (22) as well as Cd-treated macrophages and mice, these data also suggest that blockage of PAI-2 may be useful for suppression of Cd-induced fibrinogen/fibrin deposition and subsequent lung fibrosis.

We also demonstrated that fibrinogen mediated Cd-induced TLR4-dependent collagen synthesis and profibrotic cytokine production. Fibrinogen is the most abundant extracellular matrix and secreted fibrinogen is a source of fibrinogen deposition in the extracellular matrix (60). The significance of fibrinogen secretion and its modulation of collagen accumulation have been extensively explored. First, fibrinogen/fibrin matrix may regulate the tissue repair response by enhancing migration, adhesion, and proliferation of fibroblasts (23, 24, 61). In this study, we speculated that fibrinogen/fibrin may act as a provisional reservoir or scaffold for Cd-induced inflammation factors including MCP-1, TGF-β1, and IL-8, further enhancing Cd-induced pulmonary fibrosis. Fibrinogen/fibrin may also bind to fibroblasts by the cell surface integrins to activate fibroblasts and induce collagen synthesis via a mitogen-activated protein kinase pathways (62). Another explanation for an increase of collagen accumulation is that Cd may alter fibrinous matrix remodeling by inducing PAI-2 expression and therefore inhibiting fibrinolysis.

In the current study, the association of CTGF with fibrinogen suggests that binding of CTGF with fibrinogen may augment fibrinogen synthesis. The interaction of CTGF with fibrinogen could lead to the reduced fibrinolytic activity at sites of extracellular matrix remodeling, therefore allowing the persistence of CTGF/fibrinogen scaffolds. Fibrinogen contains heparin binding domains (63) which bind integrins in macrophages (51), and integrins can be induced in response to CTGF stimulation (64). Therefore, it is likely that Cd-induced CTGF secretion from fibroblasts may in turn promote the adhesion of fibrinogen with integrin following fibrinogen secretion. Synthesis of local fibrinogen in lung may serve as matrix scaffold to aid in extracellular matrix remodeling and development of lung fibrosis. Furthermore, secretion of fibrinogen and deposition of fibrinogen/fibrin in lung tissues and sites of injury may further activate TLR4, induce profibrotic cytokine production, and thereby result in pathologic fibrotic events.

Our data demonstrated that macrophages in mouse lung tissues expressed fibrinogen and secrete fibrinogen in response to local inflammation caused by Cd exposure. Considering the significant association between BALF fibrinogen with lung hydroxyproline, these data indicated that increased fibrinogen due to Cd exposure may be the cause of macrophage activation and a potential indicator of COPD airway remodeling caused by Cd-induced fibrosis.

In conclusion, we have successfully used a human descriptive study and a COPD mouse model to demonstrate that fibrinogen is a predictive indicator of Cd exposure and mediates Cd-induced macrophage activation in COPD. Cd-induced increased lung fibrinogen is likely responsible for the reported increase of plasma fibrinogen in patients with COPD. Our results demonstrated that Cd-induced fibrinogen mediates PAD2-dependent macrophages activation and TLR4-mediated collagen synthesis and profibrotic cytokine production.

DATA AVAILABILITY

The data that support the findings of this study are available on request from the corresponding author.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S4: https://doi.org/10.6084/m9.figshare.19342841.

GRANTS

This work was supported by National Institute of Environmental Health Sciences Grants P42 ES027723 (V. B. Antony), R01 ES029981 (V. B. Antony), and U54 ES030246 (M. Athar and V. B. Antony).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

F.J.L. conceived and designed research; F.J.L., R.S., P.S., K.G.D., and Z.W. performed experiments; F.J.L., C.T.S., R.-M.L., S.B., Y.-I.K., and V.B.A. analyzed data; F.J.L. interpreted results of experiments; F.J.L. prepared figures; F.J.L. drafted manuscript; F.J.L., R.-M.L., M.A., M.T.D., and V.B.A. edited and revised manuscript; V.B.A. approved final version of manuscript.