Role of CD248 as a potential severity marker in idiopathic pulmonary fibrosis

Lung tissue samples for histology were obtained by diagnostic video assisted thoracoscopic (VATS) lung biopsy and transplant explants. These procedures were performed by surgical teams at Heartlands Hospital and the Queen Elizabeth Hospital, Birmingham between the time-frame January 2005 to December 2012. Lung resections were fixed in formalin, embedded into paraffin and sectioned according to standard procedures. A total of 22 patients’ samples were included in the study. Demographic and clinical data of patients were obtained from electronic and paper records. Patient data is summarized in Table 1. No patients received disease modifying treatment (pirfenidone or nintedanib) prior to surgery.

Primary human lung fibroblasts were isolated from explanted lung tissue of patients with IPF, acquired at the time of lung transplantation (n = 6) and lung tissue from healthy donors rejected for transplantation as controls (n = 6) at the Institute of Transplantation, Newcastle Upon Tyne Hospitals NHS Trust. All procedures in this study were performed in accordance with approval from the local research ethics committees at the University of Birmingham and Newcastle University, respectively. All patients gave written informed consent for the use of their tissue and clinical data for research purposes. Ethics committee approval number is 07/MRE08/42.

Lung tissue samples for immunohistochemistry (IHC) were obtained by VATS from patients with mild-to-moderate fibrosis (n = 11) or were taken from lungs of patients with end-stage fibrosis undergoing lung transplantation (n = 11). Sections were processed for CD248 immunohistochemical staining using standard methods for diaminobenzidine (DAB) chromogen and haematoxylin for nuclear staining. The B1/35.1 anti-human CD248 antibody [14] was used for IHC in an automated IHC staining system (Roche Benchmark Ultra). The area staining positive for CD248 was compared between the two patient groups using the ImageJ software for digital image analysis as follows: we took 3–7 high resolution digital images per slide at random locations. Patient identification were blinded during the whole analysis. CD248 signal and haematoxylin signal on the images were digitally separated using the Colour Deconvolution plugin for ImageJ [22]. Threshold value was set manually at the same levels in both of the channels on all of the images. The area above the threshold was summarised for each image, representing the area stained positive for CD248 or haematoxylin on each image. CD248 positive area was compared to the haematoxylin staining area representing relative CD248 expression in the lung samples of IPF patients. A graphical explanation of this method is shown on Fig. 2a. The relative CD248 staining area were compared between transplant and VATS biopsy patients. We used both Student’s t-test and Mann–Whitney U-test for statistical analysis of the data derived from images, where p < 0.05 denoted statistical significance.

Normal human lung fibroblasts (NHLF) were purchased from Promocell (Heidelberg, Germany). At least 3 batches from different patents were used for every experiment. NHLFs were initially cultured and expanded in Fibroblast Growth Medium (Promocell) according to the supplier’s instructions. A549 human lung adenocarcinoma cell line was maintained in DMEM supplemented with 2 mM L-glutamine, HEPES, non-essential amino acids, 100 U/ml penicillin and 100 mg/ml streptomycin and 10 % FCS.

Additional primary fibroblasts were isolated at Newcastle University from the lungs of patients with advanced IPF and the lung tissue from healthy donors rejected for transplantation as controls at the Institute of Transplantation, Newcastle Upon Tyne Hospitals NHS Trust. The isolation method based on a cell outgrowth technique. Briefly, lung tissue pieces (<1 mm³) were cultured in DMEM/F12 (Sigma) supplemented with 10 % FCS, 1 % L-glutamine, 100U/ml penicillin and 100 μg/ml streptomycin to allow cells to migrate out of the tissue. After 7 days the tissue is removed and the cells grown to confluence. Mesenchymal phenotype was confirmed by positive expression of fibronectin, vimentin and α-smooth muscle actin and little/no expression of E-cadherin, ZO-1 and CD45. Fibroblasts were used at the third passage in all experiments.

Human Primary Lung epithelial cells were isolated from tissue samples from lobectomy patients with normal lung function. Cells were isolated and cultured as described elsewhere [23]. TGF-β1 was obtained from R&D Systems (Abingdon, UK). Final concentration of TGF-β1 was 10 ng/ml.

CD248 knock-down (KD) in NHLFs was carried out using specific siRNA (Life Technologies). We used Lipofectamine 2000 reagent and Opti-MEM (both from Life Technologies) according to the manufacturer’s instructions. Cells were cultured after transfection for 48 h. Efficiency of KD was determined using PCR and flow cytometry (Additional file 1: Figure S4A and B, respectively). Cell viability was >95 % after 48 h of siRNA transfection as determined by the trypan-blue exclusion test.

A commercial ELISA-based bromodeoxyuridine (BrdU) proliferation assay kit was used from Calbiochem (Watford, UK) according to the manufacturer’s instructions. Briefly, previously serum-starved cells were seeded into a 96-well tissue culture plate in DMEM + 0.1 % FCS; BrdU and stimulants were added afterwards. Cells were incubated for 24 h to incorporate BrdU, then the proliferation was stopped using the fixative agent supplemented with the kit. The development and colorimetric measurements were performed according to the manufacturer’s instructions. Proliferation experiments were repeated for at least 3 times with cells from 3 different donors.

Cultured cells were detached using a non-enzymatic cell dissociation solution (Sigma Aldrich) to preserve the trypsin-sensitive CD248 epitope. Labelling was carried out using an anti-CD248-FITC monoclonal antibody [24]. Samples were analysed on a CyanADP Analyzer flow cytometer and Summit v4.3 software. CD248 protein expression is presented as median fluorescent intensity (MFI).

Total RNA was isolated from cultured cells using the NucleoSpin RNA isolation kit with on-column DNase digestion (Machery-Nagel), cDNA synthesis was performed using a High Capacity RNA-to-cDNA kit (Applied Biosystems) following manufacturer’s protocols. For real-time qPCR experiments, master mixes with or without SYBR Green were used (Roche). The list of primers is available in the online supplement. (Additional file 1: Table S1) PCR experiments were performed on a Light Cycler 480 Instrument (Roche). In the plots reverse ΔCt values versus GAPDH expression are presented; the calculation formula was reverse ΔCt = Ct(GAPDH) - Ct(Target) [25]. The mean Ct values were calculated for 3 or more independent experiments.

Data were analysed on using SPSS for Windows 16.0 (SPSS, Inc., Chicago, IL). Data were tested for normality using Spearman’s chi squared test and analysed by non-equal variance t test or Mann–Whitney U test. Data are expressed as mean ± SE unless otherwise indicated. Correlations were calculated using least squares linear regression test.

CD248 expression was compared in lung samples from patients with IPF who underwent either VATS biopsy (mild or moderate fibrotic changes, n = 11) or lung transplantation (severe or end-stage fibrotic changes, n = 11). CD248 expression was observed specifically on fibroblast-like stromal cells but not on epithelium, endothelium, smooth muscle, alveolar macrophages or inflammatory cells (Fig. 1, Additional file 1: Figure S1). Areas of fibrosis stained positive for CD248. There was a range of weak to strong staining in the fibrotic regions with the more established fibrosis showing greater CD248 expression levels. There was a visible difference between CD248 expression patterns in lung samples with mild or moderate fibrosis and severe fibrosis (Fig. 1a and b, respectively.) Fibroblastic foci generally stained less intensive or not at all for CD248 in comparison to the surrounding fibrosis (see Fig. 1c).

In contrast to fibroblasts, epithelial cells in the lung stained negative for CD248 (Fig. 1d) and some of the physiological structures in the lung proved to be consequently positive for CD248, like the adventitia of larger muscular arteries, interlobular septa and pleural regions (Additional file 1: Figure S1). Staining in these latter structures was universally greater than that observed in fibrotic areas. In addition, staining around the bronchioles was seen but areas of inflamed interstitium exhibited very little CD248 expression. (Additional file 1: Figure S1).

Since staining intensity with DAB is not proportional to expression level [22] we compared the relative area stained positive for CD248 by comparing DAB staining (e.g. CD248 positive areas) to the area of nuclear staining with haematoxylin (Fig. 2a). We found, that the area stained positively for CD248 was significantly higher in the transplant explant samples than in patients undergoing VATS biopsy (Fig. 2b) For further statistical information on please see Additional file 1: Figure S2).

Importantly in univariate analysis there was a significant negative correlation with the relative staining area of CD248 with lung function. Our data show that FEV1, FVC, TLCO and TLC are significantly correlated to CD248 staining in fibrotic lungs (r² value between 0.35 and 0.5; p < 0.05 using non-parametric ANOVA, detailed data are presented in Table 2 and graphical presentation of the data are in Additional file 1: Figure S3.

To confirm that CD248 expression is higher in IPF fibroblasts compared to normal fibroblasts, expression was compared by qPCR and flow cytometry.

There was no significant difference in mRNA expression of CD248 between normal and IPF derived fibroblasts (Fig. 3c). In contrast CD248 protein expression was significantly increased when measured using flow cytometry (Fig. 3a and b). Interestingly, CD248 expression seemed to be sensitive to TGF-β1 treatment in IPF fibroblasts but not in normal fibroblasts (Fig. 3b) with TGF-β1 reducing expression.

In order to establish a functional role for CD248 in fibroblast function we transfected NHLFs with negative control or CD248-specific siRNA for 48 h, then assayed the bromodeoxyuridine (BrdU) uptake of the cells with ELISA. CD248 expression levels after knockdown were assessed with qPCR and FACS (Additional file 1: Figure S4). We found, that CD248 knockdown significantly reduced NHLF BrdU uptake (Fig. 4a). We tested, whether CD248 knockdown had any effect on in vitro myofibroblast transdifferentiation by assessing alpha smooth muscle actin (alpha-SMA) and collagen-1 mRNA and protein levels. We found, that CD248 knockdown did not induce myofibroblast differentiation as there were no significant differences in collagen and alpha-SMA levels between control and CD248 KD samples.

CD248 is a mesenchymal marker expressed on fibroblasts but not on epithelial cells in the lung. We tested whether CD248 is expressed by pulmonary epithelial cells undergoing epithelial-mesenchymal transition in vitro using both reverse-transcription qPCR and flow cytometry. We found that CD248 mRNA is expressed in epithelial cells at a very low level and protein expression is negative with flow cytometry. Addition of TGF-β1 did not induce CD248 expression but induced changes characteristic for EMT (Fig. 5).

In conclusion, CD248 appears to be a specific marker of mesenchymal cells that is elevated in the lungs of patients with IPF. It is noteworthy, that epithelial cells do not express CD248 when undergoing EMT, in contrast to other commonly used EMT biomarkers. CD248 expression correlated with markers of disease severity. CD248 was functionally important in terms of proliferation of primary pulmonary fibroblasts but does not appear to alter myofibroblast differentiation. CD248 may represent a novel therapeutic target in IPF to reduce fibro-proliferation.

All procedures in this study were performed in accordance with approval from the local research ethics committees at the University of Birmingham and Newcastle University. All patients included in this study gave written informed consent for the use of their tissue and clinical data for research purposes. Ethics committee approval number 07/MRE08/42.

Not applicable for this research.

All relevant data and materials are published in the manuscript and supplementary materials.