Increased chemokines levels in patients with chronic obstructive pulmonary disease: correlation with quantitative computed tomography metrics

Abstract

Objective:

We sought to explore the relationships between multiple chemokines with spirometry, inflammatory mediators and CT findings of emphysema, small airways disease and bronchial wall thickness.

Methods:

All patients with COPD (n = 65) and healthy control subjects (n = 23) underwent high-resolution CT, with image analysis determining the low attenuation area (LAA), ratio of mean lung attenuation on expiratory and inspiratory scans (E/I MLD) and bronchial wall thickness of inner perimeter of a 10-mm diameter airway (Pi10). At enrollment, subjects underwent pulmonary function studies, chemokines and inflammatory mediators measurements.

Results:

Multiple chemokines (CCL2, CCL3, CCL5, CX3CL1, CXCL8, CXCL9, CXCL10, CXCL11 and CXCL12) and inflammatory mediators (MMP-9, MMP-12, IL-18 and neutrophil count) were markedly increased in the serum of COPD patients compared with healthy controls. There were associations between small airway disease (E/I MLD) and CCL11, CXCL8, CXCL10, CXCL11, CXCL12 and CX3CL1. Especially CXCL8 and CX3CL1 are strongly associated with E/I MLD (r = 0.74, p < 0.001; r = 0.76, p < 0.001, respectively). CXCL8, CXCL12 and CX3CL1 were moderately positively correlated with emphysema (%LAA) (r = 0.49, p < 0.05; r = 0.51, p < 0.05; r = 0.54, p < 0.01, respectively). Bronchial wall thickness (Pi10）showed no significant differences between the COPD and healthy controls,，but there was an association between Pi10 and FEV₁% in COPD patients (r=−0.420, p = 0.048). Our statistical results showed that there were not any associations between airway wall thickness (Pi10) and chemokines.

Conclusion:

Pulmonary chemokines levels are closely associated with the extent of gas trapping, small airways disease and emphysema identified on high-resolution chest CT scan.

Advances in knowledge:

This study combines quantitative CT analysis with multiplex chemokines and inflammatory mediators to identify a new role of pathological changes in COPD.

Introduction

Chronic obstructive pulmonary disease (COPD) is a highly heterogeneous disease characterized by persistent airflow limitation, which leads to functional impairment and associated symptoms. ¹ The mortality and disability rate of COPD is rising rapidly, which seriously endangers the public health. By 2020, COPD has ranked third in the number of disease deaths and fifth in the global economic burden of disease. ² The underlying mechanisms of disease are poorly understood, which has limited the development of new diagnostic approaches.

Airway inflammation, lung tissue degradation and airway remodeling are considered key pathogenesis of COPD. ³ Various immune cells, inflammatory mediators and multiple signaling pathways have been implicated in these processes. ⁴ Chemokines are small molecules (8 to 12 kDa) that belong to the large family of cytokines and consist of four major subfamilies: CXC, CC, CX3C and XC. ⁵ They mediate various cellular processes by interacting with cell surface G-protein coupled receptors (GPCRs). Since chemokines have various biological functions such as chemotaxis, ⁶ leukocyte degranulation, ⁷ hematopoiesis ⁸ and angiogenesis, ⁹ they play key roles in the pathophysiology of various diseases and represent therapeutic targets.

In vitro, CXCL8 is upregulated in the human bronchial epithelial cell line (HBE-16) when stimulated by LPS. ¹⁰ An animal experiment confirms that CX3CL1 expression is upregulated in emphysema, leading to the recruitment of CX3CR1⁺ cells to the lung parenchyma of mice exposed to tobacco smoke for a long time. ¹¹ Clinical studies illustrated increased expression of CXCL8, ¹² CXCL9/CXCL10/CXCL11, ¹³ CCL2 ¹⁴ CX3CL1 ^15,16 in the serum, induced sputum or situ lung tissue of subjects with COPD. However, in these previous studies, detailed analysis of chemokine profiles has not been performed through a combination of high-resolution CT and pulmonary function. CT imaging technology provides an important opportunity to study fundamental morphological changes such as emphysema and small airway remodeling in COPD. Quantitative analysis software can be quantitatively, objectively and non-invasively evaluate multiple parameters of COPD lung morphology. Study has reported that CX3CL1 ¹⁶ are associated with quantifiable measures of emphysematous change on CT in COPD patients. However, this study has not further explored the relationship between chemokines levels and bronchial wall morphology and small airway remodeling quantified by CT. Moreover, little is known about the relationships between multiple chemokines and CT parameters of emphysema, small airway disease and bronchial wall thickness in COPD patients.

Accordingly, we have, for the first time, combined CT analysis of lung pathology with multiple profiling of chemokines and inflammatory mediators, and identified a novel role for chemokines in COPD.

Methods and materials

Population study

Participants gave written informed consent and the study was approved by the Ethics Committee of the Yan'an University Hospital, Yan'an University, China. This was a cross-sectional study that recruited 65 stable COPD patients and 23 healthy controls from January 2016 to February 2019 in the Affiliated Hospital of Yan'an University. COPD was diagnosed on the basis of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria. ¹ Post-bronchodilator lung function was used to assess airflow obstruction with a ratio of FEV₁/FVC less than 0.7 required for enrolment. COPD was staged in accordance with the GOLD guidelines: GOLD 1 (mild): FEV₁ ≥80% predicted; GOLD 2 (moderate): FEV₁ <80% and ≥50% predicted; GOLD 3 (severe): FEV₁ <50% and ≥30% predicted; and GOLD 4 (very severe): FEV₁ <30% predicted. Exclusion criteria included diseases such as interstitial pneumonia, bronchiectasis, tuberculosis, lung cancer, diabetes, autoimmune disease, heart failure, cerebral infarction, liver damage and renal failure. Finally, a total of five COPD patients in our study met the exclusion criteria and were excluded from the original recruit, including two patients with interstitial pneumonia, two patients with diabetes and one patient with heart failure.

Quantitative CT analysis of airways and emphysema

All subjects underwent volumetric CT scans of the chest using Philips Brilliance 256 iCT scan machine (Philips Medical Systems, Eindhoven, The Netherlands) without contrast in the craniocaudal direction while they were placed in the supine position; scans were taken at full inspiration and at maximum expiration after coaching. The imaging protocol consisted of; slice thickness 0.75 mm, slice separation 0.5 mm, tube voltage 120KV, effective mAs 90mAs (using dose modulation), collimation 0.6 mm and a pitch of 1. Images reconstructed with the B35 kernel were used for image analysis using Chronic Obstructive Pulmonary Disease Analysis Software (Extended Brilliance Workplace, Philips Medical Systems, Eindhoven, The Netherlands). The extent of emphysema was assessed using the percentage of lung voxels with X-ray attenuation values less than (low attenuation areas; %LAA) −950 Hounsfield units (HU) (%LAA950) (Figure 1). A surrogate marker for small airways disease was measured using the ratio of mean lung attenuation on expiratory and inspiratory scans (E/I MLD). Bronchial wall thickness was quantified using the standardized parameter Pi10, which is the square root of the wall area of a hypothetical airway with a 10 mm internal perimeter.

Figure 1.

Quantitative measurements of pulmonary emphysema by HRCT. Red areas are emphysematous lesions identified by the Chronic Obstructive Pulmonary Disease Analysis Software (Extended Brilliance Workplace) (<−950 Hounsfield units). Utilizing the analysis software, the ratio of the emphysema volume of the whole lung or each lobe to the volume of the whole lung or each lobe was calculated (%LAA).

The airway tree was generated using an automated region-growing technique. Detailed airway analysis was completed to the segmental bronchi, as well as two generations distally, in six selected airway pathways (RB1, RB4, RB10, LB1, LB4 and LB10) (Figure 2). Airway wall thickening was evaluated on inspiratory CT using measures of a standardized wall thickness measure of an airway with an inner perimeter of 10 mm (Pi10). These were quantified in segmental bronchi, as well as two generations subsegmental. Calculate the regression line of the square root of the wall area of the bronchus cross-sections to the lumen circumference. Every scan underwent thorough visual assessment for accuracy of lung/lobar segmentations, as well as completeness of the airway tree and accuracy of airway labels. Manual editing of automated segmentations was performed when necessary.

Figure 2.

Quantification of standardized small-airway dimensions. (a) Volumetric acquisition permits segmentation of the airway tree, and curved reformatting. (b) Orthogonal cross-section of a subsegmental airway, derived from curved reformat, shows delineation of outer and inner bronchial wall, permitting calculation of airway dimensions.

Detection of chemokine and inflammatory cytokine levels

A total of 5 ml of peripheral blood was taken from all subjects on an empty stomach in the morning. The serum was separated by centrifugation at 3000 revolutions/min for 15 min and stored at −80°C until later use. Chemokine and inflammatory cytokines levels in serum were quantified using a microparticle-based multiplex immunoassay (R&D systems, Abingdon, UK) developed by Luminex. Samples were analyzed on the Luminex 200 platform (Biorad Bioplex 200, Hemel Hempstead, UK), as per manufacturer‘s instructions. The following chemokine were analyzed: CCL2, CCL3, CCL5, CCL11, CXCL8, CXCL9, CXCL10, CXCL11, CXCL12 and CX3CL1. Inflammatory cytokines analysis was performed for interleukin (IL) 1β, IL-18, MMP-9 and MMP-12.

Statistical analysis

GraphPad Prism 5 (GraphPad Software, Inc., San Diego, CA, USA) and SPSS 13.0 statistical software package (SPSS Inc.,Chicago,IL,USA) were used to assess statistical analyses. Data are expressed as mean ± SD. Statistical analysis was performed using Mann–Whitney U test and Fisher’s exact tests for differences between groups. In COPD patients, associations between chemokines, pulmonary function tests, CT parameters and inflammatory mediators were assessed using Spearman’s correlation. p values less than 0.05 were considered significant.

Results

Characteristics of participants

The characteristics of the 65 COPD patients and the 23 healthy controls are presented in Table 1. There was no difference between COPD patients and healthy controls regarding age, sex ratio and body mass index. Compared with the control group, FEV₁, FEV₁%, FEV₁/FVC, MMEF25%−75% of COPD patients were significantly lower, while %LAA and E/I MLD, were significantly elevated.

Table 1.

Baseline characteristics of COPD patients and control subjects

Variables	Control subjects (n = 23)	COPD subjects (n = 65)	p value
Age, years	57.6 ± 8.5	62.1 ± 10.7	0.12
Gender, Male/female	18/5	37/12	0.54
BMI, kg/m2	24.3 ± 4.7	23.7 ± 4.1	0.29
Current smoker	14	26	0.78
Gold stage (1/2/3/4)	–	9/27/23/6	–
Pulmonary function test
FEV1, L	2.7 ± 0.62	0.94 ± 0.25	<0.001 c
FEV1 %	95.7 ± 22.6	39.4 ± 19.8	<0.001 c
FEV1/FVC	77.8 ± 10.4	44.3 ± 14.9	<0.001 c
MMEF25%−75%	76.2 ± 15.6	20.9 ± 8.8	<0.001, c
CT parameters
Bronchial wall area (Pi10)	3.74 ± 0.23	3.81 ± 0.34	0.052
Small airway disease (E/I MLD)	0.75 ± 0.09	0.89 ± 0.11	0.002
Emphysema% (%LAA)	1.6 ± 0.51	10.8 ± 4.5	0.004

BMI, Body mass index = weight (kg)/height (m²); COPD, Chronic obstructive pulmonary disease; E/I MLD, Expiration/inspiration ratio of mean lung density; FEV₁%, FEV₁ percent predicted; FEV₁, Forced expiratory volume in 1 sec; FEV₁/FVC, Ratio of forced expiratory volume in 1 sec to FVC; FVC, Forced vital capacity; %LAA, The percentage of low attenuation area <950 hounsfield units; MMEF25%–75%, MMEF25–75 percent predicted.

Data are expressed as mean ± SD.

^a p < 0.05.

^b p < 0.01.

^c p < 0.001.

Chemokines and cytokines levels in COPD patients and healthy controls

We detected serum chemokines and cytokines levels in all subjects. The concentrations of CCL2, CCL3, CCL5, CX3CL1, CXCL8, CXCL9, CXCL10, CXCL11 and CXCL12 were markedly elevated in the serum of COPD patients compared with healthy controls (Figure 3). There was no significant differences in CCL11 between the groups. Moreover, the serum levels of MMP-9, MMP-12, IL-18 and neutrophil count were significantly increased in COPD patients compared with healthy participants (Figure 4). However, our findings showed no significant differences in IL-1β between the two groups.

Figure 3.

Expression levels of chemokines in COPD and healthy controls (HC). (a) CCL2, (b) CCL3, (c) CCL5, (d) CCL11, (e) CX3CL1, (f) CXCL8, (g) CXCL9, (h) CXCL10, (i) CXCL11 and (j) CXCL12. Data are expressed as mean ± SD. n = 65 for COPD and 23 for healthy controls. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001 using Mann–Whitney U test.

Figure 4.

Expression levels of cytokines in COPD and healthy controls (HC). (a) MMP-9, (b) MMP-12, (c) IL-1β, (d) IL-18, (e) serum neutrophil count (%). MMP-9, Matrix metalloproteinase-9; MMP-12, Matrix metalloproteinase-12; IL, interleukin. Data are expressed as mean ± SD. n = 65 for COPD and 23 for healthy controls. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001 using Mann–Whitney U test.

Relationships of chemokines with CT parameters, inflammatory mediators and respiratory function

There were associations between E/I MLD and CCL11, CXCL8, CXCL10, CXCL11, CXCL12 and CX3CL1. Especially CXCL8 and CX3CL1 are strongly associated with E/I MLD (r = 0.74, p < 0.001; r = 0.76, p < 0.001, respectively, Table 2). CXCL8, CXCL12 and CX3CL1 were moderately positively correlated with emphysema (%LAA) (r = 0.49, p < 0.05; r = 0.51, p < 0.05; r = 0.54, p < 0.01, respectively, Table 2).

Table 2.

Spearman’s correlation analysis between chemokines, pulmonary function test, inflammatory mediators and CT parameters in COPD subjects

	CCL2	CCL3	CCL5	CCL11	CXCL8	CXCL9	CXCL10	CXCL11	CXCL12	CX3CL1
CT parameters
Bronchial wall area (Pi10)	0.17	0.13	0.19	0.25	0.41	0.33	0.28	0.22	0.37	0.39
Small airway disease (E/I ratio MLD)	0.22	0.17	0.26	0.45 a	0.74 c	0.39	0.47 a	0.59 a	0.57 a	0.76 c
Emphysema% (%LAA)	−0.05	−0.07	0.12	0.24	0.49 a	0.29	0.34	0.37	0.51 a	0.54 b
Cytokines
MMP-9	0.52b	0.29	0.55 b	0.37	0.89 c	0.41	0.49	0.36	0.74 c	0.78 c
MMP-12	0.46 a	0.25	0.51 b	0.29	0.56 b	0.38	0.45	0.34	0.69 c	0.59 b
Neutrophils	0.29	0.37	0.18	0.14	0.67 b	0.19	0.24	0.22	0.39	0.41
IL-1β	0.27	0.21	0.11	0.31	0.38	0.21	0.15	0.28	0.35	0.23
IL-18	0.14	0.25	0.16	0.20	0.40	0.24	0.27	0.34	0.47 a	0.34
Spirometry
FEV1 %	−0.17	−0.34	−0.36	−0.19	−0.51 a	−0.23	−0.19	−0.26	−0.59 b	−0.64 b
MMEF25%−75%	−0.14	−0.25	−0.47 a	−0.12	−0.48 a	−0.25	−0.31	−0.27	−0.56 b	−0.62 b

E/I MLD, Expiration/inspiration ratio of mean lung density; FEV1%, FEV1 percent predicted; IL, Interleukin; %LAA, The percentage of low attenuation area <950 hounsfield units; MMEF25%−75%, MMEF25−75 percent predicted; MMP-9, Matrix metalloproteinase-9; MMP-12, Matrix metalloproteinase-12; Pi10, Square root of airway wall area of the theoretical airway with an internal perimeter of 10 mm.

Spearman’s r values given.

^a p < 0.05.

^b p < 0.01.

^c p < 0.001.

Bronchial wall thickness (Pi-10）showed no significant difference between the COPD subjects and healthy controls (Table 1), but there was an association between Pi10 and FEV₁% in COPD patients (r=−0.420, p = 0.048) (Table 3). Our statistical results also showed that there were no associations between airway wall thickness (Pi10) and chemokines (Table 2).

Table 3.

Correlation between CT parameters and lung function indices in COPD patients

Variables	FEV1%	MMEF25%−75%	DLCO%
Emphysema% (%LAA)	r = −0.192, p = 0.481	r = −0.258, p = 0.431	r = −0.541, p < 0.01 a
Small airway disease (E/I MLD)	r = −0.446, p = 0.035 a	r = −0.647, p < 0.001 b	r = −0.215, p = 0.517
Bronchial wall thickness (Pi10)	r = −0.420, p = 0.048 a	r = −0.316, p = 0.377	r = −0.186, p = 0.649

E/I MLD, Expiration/inspiration ratio of mean lung density; FEV₁%, FEV₁ percent predicted; %LAA, The percentage of low attenuation area <950 hounsfield units; MMEF25%−75%, MMEF25−75 percent predicted; Pi10, Square root of airway wall area of the theoretical airway with an internal perimeter of 10 mm.

Spearman’s r and p values given.

^a p < 0.05.

^b p < 0.001.

In addition, we also utilized the Spearman’s correlation to analyze the relationships between chemokines and inflammatory mediators. There is a strong association between MMP-9 and MMP-12 and chemokines CCL2, CCL5, CXCL8, CXCL12 and CX3CL1 (Table 2). IL-18 and neutrophil count, especially neutrophil count are closely correlated with CXCL8. However, our results showed no significant correlation between IL-1β and CXCL8. Finally, we analyzed and compared chemokines levels with spirometry indices of disease severity. FEV₁% and MMEF25%−75% were inversely correlated with chemokines CXCL8, CXCL12 and CX3CL1 in COPD patients. Moreover, our statistical results illustrated that CCL5 is significantly associated with MMEF 25%−75%.

Correlation between CT parameters and lung function indices in COPD patients

Small airway disease (E/I MLD) in COPD patients was negatively correlated with FEV₁% and MMEF25%−75% (r=−0.446, p = 0.035; r=−0.647, p < 0.001, respectively, Table 3). Bronchial wall thickness (Pi10) in COPD subjects was negatively correlated with FEV₁% (r=−0.420, p = 0.048). Moreover, our results showed a moderate correlation between emphysema% (%LAA) and DLCO% (r=−0.541, p < 0.01).

Discussion

The present study has, for the first time, illustrated an association between small airways disease measured by quantitative CT analysis and the expression levels of chemokines and inflammatory mediators in COPD. Chemokines and inflammatory mediators are also associated with other clinical indicators of disease severity, including FEV ₁% and emphysema severity. Our statistical results demonstrated that chemokines are closely related to small airway disease and emphysema rather than bronchial wall thickness.

Chemokines are small molecules with biological functions of chemotaxis and leukocyte degranulation that have been involved in the pathophysiology of airway inflammation and lung tissue degradation seen in COPD. ⁵ Our results demonstrate that CCL2, CCL3, CCL5, CXCL8, CXCL10, CXCL11, CXCL12 and CX3CL1 are elevated in the serum of COPD patients, which is in agreement with the results of previous studies. ^12–18 Moreover, we found no difference in serum CCL11 concentration between COPD patients and the healthy control subjects. However, there are several literature reports that are inconsistent with our findings. Bradford et al ¹⁹ found that CCL11 in the blood of COPD patients was significantly higher than that in the healthy control subjects and was associated with airflow obstruction (FEV₁%). An animal experiment illustrated that the expression of CXCL9 mRNA in bone marrow of CS-exposed guinea pigs was significantly decreased compared to the controls. ²⁰ Clinical study showed that CXCL12 mRNA expression levels in bone marrow mesenchymal stem cells (BM-MSCs) of COPD patients were reduced compared to control subjects. ²¹ Our findings are contrary to these previous studies due to the difference in sampling locations between induce sputum and bone marrow. Furthermore, our findings are inconsistent with author Bradford’s. The reason may be that COPD is a highly heterogeneous disease, and we have not compared and studied the different clinical phenotypes of COPD. In addition, our work also demonstrated increased levels of MMP-9, MMP-12 and neutrophil counts in COPD patients, which is consistent with previous findings. ^22,23 This provides further evidence for the role of MMPs and neutrophils in airway inflammation and lung tissue degradation in COPD.

Our study found that there was no significant difference in Pi10 between COPD and the healthy subjects, but it did show an association with FEV₁%, which is consistent with previous studies. ²⁴ There is no significant correlation between Pi10 and chemokines. The reasons are not clear. It may be that COPD is a highly heterogeneous disease. Furthermore, Pi10 may not be the best measure for evaluating changes in airway wall morphology. The bronchial wall thickening is not directly related to the intraluminal inflammation index. The remodeling of the proximal airway may be more associated with inflammatory infiltration of the submucosa. ²⁵ Further pathological biopsy or basic research is needed to clarify these mechanisms.

In histopathology, small airway remodeling is considered to be morphological abnormalities such as thickening of the airway wall, narrowing of the lumen and mucus obstruction, accompanied by the proliferation of mucus cells, goblet cells and fibroblasts. ²⁶ At present, there is not any uniform standard for investigating and assessing small airways disease in COPD. Studies have confirmed that E/I MLD is not only significantly related to lung function parameter of small airway diseases, but also the best CT parameter for capturing gas. ^27,28 Our results showed that E/I MLD is markedly elevated in COPD and is closely related to lung function, supporting its use as a CT parameter for small airway diseases.

Our study has, for the first time, explored the role of chemokines in small airway disease of COPD. Our findings suggested that levels of CCL5, CXCL8, CXCL12 and CX3CL1 are related to lung function MMEF 25% −75%. We also found that the quantitative CT parameter E/I MLD is closely associated with the concentrations of CCL11, CXCL8, CXCL10, CXCL11, CXCL12 and CX3CL1. However, we found that CXCL8 and CX3CL1 are most strongly associated with MMP-9 and MMP-12. This result suggests that related chemokines play a role in chemotaxis and adhesion to make neutrophils, monocytes, lymphocytes cells and macrophages migrate to the diseased tissue. It is widely accepted that MMPs play a critical role in the pathophysiology of emphysema and airway remodeling in COPD. ²⁹ Therefore, chemokines may directly or indirectly degrade components of the extracellular matrix and eventually lead to airway remodeling.

The present study demonstrated that CXCL8, CXCL12 and CX3CL1 are significantly associated with the extent of emphysema detected by CT scan (%LAA) in COPD patients. This is in keeping with previous studies which illustrated that the levels of CXCL8 and CX3CL1 are significantly related to %LAA in COPD subjects. ^16,19 It is generally believed that CXCL8 is produced by alveolar macrophages and bronchial epithelial cells. A clinical examine has confirmed that compared with smokers and non-smokers with normal lung function, CXCL8 in induced sputum of COPD patients is significantly elevated and is associated with increased neutrophil numbers in sputum. ³⁰ In addition, our study also found that the chemokine CXCL8 was significantly positively correlated with neutrophil count. These results support the key role of CXCL8 in neutrophil inflammation occurring in COPD airways. Although there are no clinical studies on the relationship between chemokine CXCL12 and emphysema, an animal experiment showed that the intermittent use of plrixafor in a mouse emphysema model exposed to cigarette smoke can reduce emphysema damage without affecting CXCL12 level and inflammation in bronchoalveolar lavage fluid (BALF). ³¹ The results of this basic experimental study indicated that targeting the CXCL12-CXCR4 axis may be promising for COPD treatment.

This work has quite a few potential limitations that deserve comment. Firstly, the study population was comparatively small. Secondly, due to limited statistical capacity, we were unable to analyze the effects of smoking on COPD patients and healthy subjects, and we did not further compare the differences between various COPD phenotypes in biomarkers, lung function, and quantitative CT parameters of small airway disease, emphysema and bronchial wall thickness. Thirdly, a cut-off of −950HU of %LAA was considered to be the best method that we currently have for emphysema quantification, ³² but it is not a perfect measurement. Meanwhile, it can be significantly altered by technical and patient factors such as breath holding, scan range, IV contrast, slice thickness, reconstruction kernel etc. Finally, we performed a multitude of correlation analysis and comparisons between CT parameters, pulmonary function indices and inflammatory mediators, and the results suggested that chemokines are closely related to small airway disease and emphysema rather than bronchial wall thickness. However, this is a single-center, observational study that requires further confirmation from basic research and animal experiments to give more evidence for the clinical and theoretical aspects of COPD.

In conclusion, our study showed that multiple pulmonary chemokines levels are closely associated with the extent of gas trapping, small airways disease and emphysema identified on high-resolution chest CT scan. To our knowledge, chronic inflammation of airway, lung tissue and pulmonary blood vessels was considered to be a pathological characteristic of COPD. A variety of inflammatory cells and cytokines are involved in the pathophysiological process of COPD, including multiple chemokines. Our previous studies have showed that the chemokine CX3CL1 as a biomarker is not only related to systemic inflammation, small airway obstruction and COPD assessment test scores, but also might be predict frequent exacerbation of COPD and assess the emphysema severity. ^15,16 Chemokines as biomarkers may be able to provide useful information about COPD inflammation and function. However, chemokines cannot provide related changes in lung anatomy and pathomorphology. HR CT can clearly show the key pathological changes of COPD, including emphysema, large airway remodeling and small airway diseases. ³³ Therefore, the combined application of chemokines and HRCT can not only provide useful information from the inflammatory and functional aspects, but more importantly, it can evaluate COPD from the anatomical and pathomorphological perspectives.