β2-microglobulin as a biomarker of pulmonary fibrosis development in COPD patients

COPD patients’ characteristics

This case-controlled, retrospective study included 450 COPD patients. Analysis of β2M levels and basic demographic information, systemic inflammatory levels, and pulmonary functions showed that β2M levels exhibited no correlation with gender, smoking index, BMI, RV/TLC, FEV1%pred, and PEF. However, the β2M levels showed a positive correlation with age, CRP levels, EOS%, FEV1/FVC, and MMEF, and a negative correlation with WBC, DLCO, and DLCO/V_A (Table 1).

β2M levels differ in COPD patients with better and worse DLCO values

The COPD patients were divided into four groups according to their DLCO values. The β2M levels gradually increased from DLCO I group to DLCO IV group (Table 2). Since there might be a β2M cut-off point, the COPD patients were divided into two groups – a “better DLCO group” (DLCO I and DLCO II) and a “worse DLCO group” (DLCO III and DLCO IV). Using Mann-Whitney U test, we found that the baseline characteristics were unmatched. The serum β2M values were 2.10±1.22 mg/L in the better DLCO group, and 2.24±0.89 mg/L in the worse DLCO group (Table 3). Using propensity score matching (PSM) to match the baseline, the β2M differences between the two groups were still significant (Table 3).

DLCO values differ in COPD patients with higher and lower β2M levels

In secondary analysis using the differences in β2M levels in the two DLCO groups, a cut-off point was set as 2.20 mg/L. Patients’ data were divided into a higher β2M group (serum β2M ≥2.20 mg/L) and a lower β2M group (serum β2M <2.20 mg/L). There were differences in age, CRP levels, and DLCO values between the two groups. COPD patients with emphysema represented less than 50% cases. Emphysema and RV/TLC % showed no differences between the two groups (Table 4).

Referring to the results of DLCO and β2M subgroups, the indicators (age, BMI, CRP, and β2M) with p-values < 0.05 were selected into Binary logical regression model (Table 5). Higher BMI was better for COPD patients’ diffusion capacity (OR 0.841), while higher serum β2M was harmful to COPD patients’ diffusion capacity (OR 4.050). The model indicated that an increased serum β2M level was an independent risk factor for diffusion impairment and pulmonary fibrosis development in COPD patients.

β2M serum and lung levels are associated with fibrosis in COPD patients

We measured β2M serum levels in 6 patients with COPD diagnosed by pulmonary function test, and analyzed their lung tissues collected during lung surgery. Patients with higher serum β2M levels had lower DLCO values compared to patients with low β2M levels. There was no difference in emphysema between the two groups; however, fibrosis was significantly different between the two groups, as reported by chest CT (Table 6).

Using hematoxylin-eosin (HE) staining, we found that the COPD patients with higher serum β2M levels had thicker alveolar wall/septum (Figure 1). Immunohistochemical (IHC) staining showed that the patients with increased serum β2M levels had an increased β2M expression in lung tissues, especially in the alveolar wall/septum. In addition, they had increased expression of TGF-β1, Smad4, and a-SMA compared to patients with low serum β2M levels (P<0.05; Figure 2).

β2M lung expression correlates with EMT and fibrosis progression in COPD rats

To rule out the effect of other diseases, such as lung cancer, aging, or obesity, we used a COPD rat model to validate our data indicating that increased β2M levels induce alveolar epithelial-mesenchymal transition (EMT) and pulmonary fibrosis in COPD. HE staining (Figure 3) showed that control rats (n=6) had normal alveolar construction (X_A=12.52μm), while COPD rats had moderately thick alveolar wall/septum (XB=21.22μm, n=7) and COPD rats had thickest alveolar wall/septum (XC=40.38μm, n=6). IHC staining demonstrated that the lung tissues of 6 COPD rats with the thickest alveolar wall/septum exhibited increased β2M expression compared to 7 COPD rats with moderately thick alveolar wall/septum and 6 rats with normal alveolar construction. In addition, the lung expression of TGF-β1, Smad4, a-SMA, col1, and col3 showed the same tendency as the β2M expression among the three rat groups, P<0.05 (Figures 4, 5). Masson staining showed that the 6 COPD rats with the thickest alveolar wall had a massively increased expression of collagen fibers in their lung tissues (Figure 6).

Study population

450 COPD patients were included in this case-control retrospective study. COPD was diagnosed by pulmonary function test according to the standards of Global Initiative for Chronic Obstructive Lung Disease (GOLD). Patients with chronic kidney disease (CKD) or other uncontrolled serious systemic disease were excluded. Patients’ data (containing basic demographic information, history of smoking, laboratory blood tests, and pulmonary function tests) were collected from Hospital electronic records between January 1, 2013 and December 31, 2017.

In addition, lung tissue specimens were obtained from 6 COPD patients (3 with higher serum β2M values and 3 with lower serum β2M), who were diagnosed with solitary lung nodule and needed resection. Then specimens were fixed by 10% formalin and embedded with paraffin for hematoxylin-eosin (HE) and immunohistochemical (IHC) staining. All procedures adhered to the Helsinki Declaration. This study was approved by Shandong Provincial Hospital Medical Ethics Committee (Ethical Review of Medical Research on Human Being No. 2016-23; LCYJ: NO. 2019-019).

Rat COPD model

Adult male Wistar rats weighing above 200 g were used for all experiments. Rats were fed in SPF animal facility (temperature 22±2° C, humidity 55±5%), with normal day and night cycle, and free access to common diet and water. Rats were randomly divided into 2 groups: control (n=6) and COPD/CSE (n=13). All experimental procedures followed the Guidelines of the Institutional Animal Care and Use Committee, Chinese Academy of Sciences.

Rat model of COPD was exposed to cigarette smoke (10/time, Huangshan Brand, China Tobacco Anhui Industrial Co., Ltd.) for 4 times each day in a special device, with more than 3-hour breaks in between. The exposure was performed for 5 days per week, and lasted for 12 weeks before final measurements of lung function. The non-COPD groups were put in the same case with free access to fresh air. After exposure, all rats were fed in the room and had access to fresh air. After lung function was measured and COPD model was established, rats were sacrificed and the lung tissues were obtained. The study was approved by Shandong Provincial Hospital Animal Experiment Ethics Committee (Ethical Review of Animal Experiment No. 2019-001).

Study design

We hypothesized that serum and tissue β2M values may serve as a biomarker of pulmonary fibrosis (determined by lung diffusion function - DLCO; chest CT; lung tissue HE staining - thickness of alveolar wall and thickness of alveolar septum; fibrosis indicators immunohistochemical staining of lung tissue; and Masson staining) in COPD. 450 COPD patients were divided into four groups according to their DLCO values (Normal group DLCO I≥80%, Mild Impairment group 80>DLCO II≥60%, Moderate Impairment group 60%>DLCO III≥40%, Severe Impairment group DLCO IV<40%). Then we divided these COPD patients into two groups: a better diffusing capacity group (combined Normal group and Mild Impairment group, DLCO≥60%) and a worse diffusing capacity group (combined Moderate Impairment group and Severe Impairment group DLCO<60%), and analyzed the differences in serum β2M values between these two groups. Secondary analysis was then performed to find a cut-off point of serum β2M to verify the difference in DLCO of COPD patients.

In addition, lung tissues and serum samples from 6 COPD patients (3 with high serum β2M and 3 with low serum β2M) and 19 rats (control n=6, COPD n=13) were analyzed by HE staining for changes in alveolar walls and alveolar septum, IHC staining of β2M and fibrosis indicators (TGF-β1, Smad4, α-SMA, col1, col3), and Masson staining for observing collagen fiber content in lung tissues.

Histology and morphometric analyses

The lung tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 4.5 μm-thick sections. The sections were stained with hematoxylin and eosin (HE). The measurement of distal airspace wall thickness was analyzed by using the Image Pro-Plus. In brief, the slides were observed at 200× magnification. The measurement was performed by drawing lines horizontally across the field of view. Bronchioles were intentionally ignored in the field. The thickness of each septum crossing the given horizontal line was measured perpendicular to its course at that crossing point. Three separate horizontal lines were drawn and analyzed for each field, and the average septal thickness was calculated for each grid. There were three lung specimens per group for analyses of distal airspace wall thickness.

IHC staining

The 4.5 μm-thick sections obtained from paraffin-embedded lung tissues were deparaffinized, and antigen retrieval was performed at a temperature over 95° C. Nonspecific reactions were blocked with goat serum for 15 minutes at 37° C, and the sections were incubated with primary antibodies against β2-Microglobulin (Proteintech 13511-1-AP, Rosemont, IL, USA, 1:50), TGF-β1 (Proteintech 21898-1-AP, Chicago, Il, USA, 1:100), Smad4 (Proteintech 10231-1-AP, Chicago, Il, USA, 1:50), a-SMA (Proteintech 14395-1-AP, Rosemont, IL, USA, 1:3000), col1 (YM3767, Immonoway, Beijing, China, 1:50), and col3 (Proteintech 22734-1-AP, Sanying, Wuhan, China, 1:500) for 1 hour at 37° C. The sections were incubated with a secondary antibody kit (SZGB-BIO, PV9003, Beijing, China) for 20 minutes at 37° C. The samples were viewed under a confocal FV 1000 SPD laser scanning microscope (Olympus, Japan), and labeling index analyses were performed with ImageJ.

Masson's trichrome staining

The sections were treated with Masson trichrome (MT) staining kit (Servicebio, G1006, Wuhan, China). Visual fields were randomly selected for each section, using the same conditions for light setting and contrast. Expression of collagen fibers in pulmonary alveoli was quantified by measuring the positive staining area using ImageJ.

Statistical analysis

All data were collected from Hospital electronic records and recorded by EpiData CRF (Version 3.1). Statistical analysis was performed by SPSS Version 23.0 (SPSS Inc.; Chicago, Illinois) and figures were made by GraphPad Prism7 (GraphPad, San Diego, CA) and ImageJ (National Institutes of Health, USA). Correlation between β2M and DLCO, DLCO/V_A was analyzed by Spearman bivariate analysis. In the primary analysis, parametric data were presented as mean ± standard deviation or mean (range). Non-parametric data were presented as median (interquartile range, IQR). Kruskal-Wallis test was used to compare differences among multi-groups and Mann-Whitney U test was used to compare data between pairs. Propensity score matching (PSM) test was performed between DLCO better and DLCO worse groups to exclude the effect of unmatched baseline. Then logistic regression was used to analyze multivariable parameters. In the secondary analysis, analyzing the cut-off point to divide COPD patients into high β2M groups and normal β2M groups, the differences between β2M groups were analyzed by Mann-Whitney U test. P<0.05 was considered significant for all statistical analyses.

Study approval

This study adhered to the Helsinki Declaration protocol, and was approved by Shandong Provincial Hospital Medical Ethics Committee (Ethical Review of Medical Research on Human Being No. 2016-23; LCYJ: NO. 2019-019). Animal experiments were approved by Shandong Provincial Hospital Animal Experiment Ethics Committee (Ethical Review of Animal Experiment No. 2019-001).