Ηeparan sulphate in infectious and non‐infectious exacerbations of COPD

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are associated with worsening health outcomes and effective treatment of each episode is essential. In this study, we aimed to investigate if plasma levels of heparan sulphate (HS) are associated with the aetiology of AECOPD.


INTRODUCTION
Chronic obstructive pulmonary disease (COPD) is a chronic inflammatory disease characterized by a progressive, largely irreversible, airflow limitation. Airway remodelling is a crucial process of COPD associated with modifications of extracellular matrix (ECM) molecules, resulting in serious alterations of airway wall thickness, resistance and elasticity. 1 Acute exacerbations of COPD (AECOPD) are episodes of accelerated deterioration of pulmonary function that affect quality of life and are associated with patients' morbidity and mortality. 2,3 Respiratory viral and bacterial infections have been implicated as causative factors for AECOPD. 4 However, it is not completely understood how these infections may alter airway inflammation and remodelling and how they are related to response to treatment. The heterogeneity associated with the aetiology and phenotypes of AECOPD remains an ongoing challenge for prompt and effective treatment in a clinical setting.
Glycosaminoglycans (GAGs) together with collagens are the main ECM molecules of the lung that provide structural and mechanical properties and moreover regulate a variety of cellular processes such as migration, proliferation and interaction with growth factors. 5 In COPD, under the catalytic action of GAG-degrading enzymes such as heparanase-1 (HSPE-1) and matrix metalloproteinases (MMPs), GAGs and collagens undergo a dynamic turnover that may determine disease severity and outcomes. [6][7][8][9][10][11] In this respect, we have shown that degradation products of collagens and elastin are associated with clinically relevant outcomes in COPD. 9,10 Furthermore, AECOPD are associated with increased degradation of hyaluronic acid (HA) in bronchoalveolar lavage (BAL) 6 and in serum 11 and this may contribute to airway inflammation and subsequent lung function decline during exacerbations. HS and chondroitin sulphate are also elevated in the BAL of COPD patients at exacerbation and correlated with MMPs in BAL, indicating that they may be associated with airway remodelling and lung function decline during AECOPD. 8 In the present study, we hypothesized that circulating levels of HS alter during AECOPD and that these alterations may predict the aetiology of exacerbations in a total of 1189 COPD patients from a discovery and a validation cohort.

Study design and participants
The present study comprises a discovery and a validation cohort. The discovery cohort included 638 patients from the PROMISE-COPD study (Predicting Outcome using Systemic Markers in Severe Exacerbations of Chronic Obstructive Pulmonary Disease) that was an investigator-initiated and driven, multicenter longitudinal trial, performed at 11 tertiary respiratory centers in eight European countries (ISRCTN99586989). 11 The PROMISE-COPD study was approved by the Institutional Review Board (EKBB295/07). Patients included in the PROMISE study had moderate to very severe COPD and were free of an exacerbation for at least 4 weeks before inclusion in the study. [12][13][14] After the baseline visit, patients had scheduled visits every 6 months and unscheduled visits at exacerbation and 4 weeks later. The median follow-up was 24 months. The diagnosis of stable COPD and of exacerbations was made according to the GOLD guidelines. AECOPD were diagnosed as an acute event characterized by a worsening of the patient's respiratory symptoms that is beyond normal day-to-day variations and leads to a change in medication. Plasma samples were obtained at stable state, at exacerbations and at 4 weeks after the exacerbations.
The validation cohort included a total of 551 patients, 101 patients from the BASCO study (Basel Study on COPD), a cross-sectional, observational, monocentric study and 450 patients from the PREVENT study, an investigator-initiated and driven, multicentre controlled trial (ISRCTN45572998). 3 Both studies were approved by the Institutional Review Board (EKBB05/06 and EKBB 306/10).
Plasma samples were obtained at stable state, at exacerbations and 3 weeks after the exacerbations. Spontaneous or induced sputum was collected and assessed for bacterial growth semi-quantitatively, as described by Podbielsky et al. 15 Oral and nasopharyngeal swabs were analysed using a commercial multiplex nucleic acid amplification testing (respiratory pathogen panel NxTAG-RPP, Luminex, MV's-Hertogenbosch, the Netherlands) 16 for the detection of 18 viruses (Appendix S1 in the Supporting Information).
Plasma samples from 30 age-and gender-matched healthy blood donors served as controls. The studies complied with the Helsinki Declaration and GCP Guidelines. All patients provided written consent for the study assessments.

Measurements of HS and HSPE-1
The concentrations of HS and heparanase-1 (HSPE-1) were measured in plasma aliquots obtained at stable state, at exacerbation and at 4 weeks follow-up visits, in duplicate, by enzyme-linked immunosorbent assay (ELISA). For HS the ELISA kit from Cusabio was used. Detection range, 20-8000 ng/mL; intraassay coefficient of variability (CV) < 6%; interassay CV < 11% (Appendix S1 in the Supporting Information).

Patients
A total of 1189 patients were enrolled in the study from a discovery and validation cohort. In the discovery cohort, from 638 COPD patients enrolled, 506 patients attended a scheduled visit at 6 months and were eligible for inclusion in subsequent analyses. Patients were followed during a median of 722 [395-762] days. During the follow-up, 317 patients (62.6%) suffered at least one exacerbation and 38 patients (7.5%) were deceased ( Figure 1).
In the validation cohort, 551 COPD patients were followed for 24 months. Out of a total of 292 AECOPD, 147 (50.3%) events were identified with no overt evidence of an infection in the previous stable visit, indicating the absence of previous colonization. These exacerbations without

SUMMARY AT A GLANCE
Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are associated with worsening health outcomes and effective treatment of each episode is essential. Circulating levels of heparan sulphate (HS) are increased at AECOPD and this increase may be associated with the aetiology of these events. Baseline demographic and clinical characteristics of the patients were generally similar in the discovery and validation cohorts (Table S1 in the Supporting Information). Mean age of patients was 67.0 ± 9.3 years, 33% of them were current smokers with a smoking history of 51.5 ± 28.9 pack years, had a duration of COPD symptoms of 101.7 ± 87.0 months, and a history of one exacerbation (IQR 0.5-1) in the previous year before inclusion in the study. Patients had COPD GOLD grade II (55.6%), grade III (33.1%) and GOLD grade IV (12.6%), post-bronchodilator FEV 1 of 50.2 ± 17.8% of the predicted value, residual volume of 152.4 ± 47.0% of the predicted value and diffusion capacity of 55.8 ± 20.7% of the predicted value (Table S1 in the Supporting Information).

Ηeparan sulphate and patients' characteristics
In the discovery cohort, HS levels at stable state were positively correlated with the age of patients (Spearman's rho: 0.342, p < 0.001) and with the age-adjusted (Spearman's rho: 0.326, p < 0.001) and unadjusted (Spearman's rho: 0.165, p = 0.024) Charlson score (Table S2 in the Supporting Information). There was no significant correlation of HS with lung function parameters or with emphysema, as indicated by the absence of a significant correlation with diffusion capacity or with quantitative CT scan parameters for emphysema.
In a linear regression model, HS was found to be significantly associated with MMRC (p = 0.011) and with Borg scale (p = 0.048), indicating that high levels of HS are positively associated with dyspnoea (Table S3 in the Supporting Information). However, HS was not associated with COPD outcome such as survival time, number of exacerbations or BODE index.
In the discovery cohort, HS was higher in patients who were not smokers [median (IQR): 1142 ng/mL (671)] as compared with patients who were current smokers [942 ng/mL (496)] (p = 0.001; Figure S1 in the Supporting Information).

HS is increased during AECOPD
Plasma levels of HS in COPD patients of the discovery cohort were significantly higher as compared to non-COPD controls F I G U R E 1 Study design. AECOPD, acute exacerbations of chronic obstructive pulmonary disease; COPD, chronic obstructive pulmonary disease.
(p < 0.001; Figure S2A in the Supporting Information). Furthermore, HS was significantly higher at exacerbations as compared to the stable state of the disease (p < 0.001). At follow-up, 4 weeks after exacerbation, HS remained at higher levels compared with stable state (p < 0.001).
When comparisons were stratified according to smoking status, HS was significantly increased at exacerbation in smokers ( p = 0.050) and non-smokers (p = 0.003; Figure S2B,C in the Supporting Information). At follow-up, compared to stable state values, HS remained significantly higher in non-smokers but not in smokers. F I G U R E 3 Plasma concentration of heparan sulphate in COPD patients of the validation cohort at stable state, at acute exacerbation of COPD (AECOPD) and at follow-up. Squares represent the central value of heparan sulphate and the whiskers the 95% confidence interval. Rank mixed linear regression models adjusted for smoking status were used to determine differences among groups.

HS and aetiology of AECOPD
group of no infection) (Figure 4). At follow-up, HS plasma levels dropped significantly as compared to AECOPD events in all groups (p < 0.001 for all) except in the group of no infections (p = 0.118).
The fold-increase of plasma HS from stable state to AECOPD was higher in the group with viral and bacterial coinfections (4.36-fold increase), than in the group of viral infections (3.26-fold increase) or bacterial infections (2.81-fold increase) or no infections (2.17-fold increase; Figure 5).
As the change of HS from the last stable visit to an AECOPD event increases, the probability of having an infection at AECOPD increases, therefore, the probability of having no infection decreases ( Figure 6). A change of HS of 5000 ng/mL from stable state to AECOPD was associated with a 15% chance of having no infection at AECOPD, 22% probability of having a mixed viral and bacterial infection, 44% of having a viral infection and 58% of having a bacterial infection. The probability of infections was similar in smokers and non-smokers (p = 0.144; Figure S3 in the Supporting Information).

HSPE-1 and aetiology of AECOPD
We did a longitudinal investigation of circulating HSPE-1 levels in the validation cohort. HSPE-1 was significantly increased at AECOPD [median (IQR): 3.81 ng/mL (5.49)] as compared to the stable state [3.33 (4.81)] (p = 0.007) and dropped to stable state levels at follow-up [3.65 (7.62)] ( Figure S4 in the Supporting Information). When AECOPD were grouped according to aetiology there was no significant F I G U R E 4 Plasma concentration of heparan sulphate in COPD patients of the validation cohort at stable state, at acute exacerbation of COPD (AECOPD) and at follow-up, in groups according to aetiology of COPD. Squares represent the central value of heparan sulphate and the whiskers its 95% confidence interval. Rank mixed linear regression models adjusted for smoking status were used to determine differences among groups.

HS to predict AECOPD
The area under the receiver operating characteristic curve for HS to diagnose any infection at AECOPD was 0.72 (95% CI, 0.62-0.91; Figure S6A in the Supporting Information). At max Youden-index (0.464) the sensitivity was 61.74% and the specificity 84.62%.
Using a logistic regression model adjusted for smoking we observed that an increase of HS by 1000 ng/mL from stable state to AECOPD would increase by 1.204 times the probability of having any infection at AECOPD with a sensitivity of 61.4% and a specificity of 76.9% (AUC: 0.698 95% CI, 0.596-0.800; Figure S6B in the Supporting Information).

DISCUSSION
In the present study, we investigated HS and its degrading enzyme HSPE-1 in AECOPD using longitudinal data from two independent, well-characterized cohorts of COPD patients. We demonstrate for the first time, that AECOPD are associated with increased circulating levels of HS and that longitudinal changes in plasma HS between stable state and AECOPD may be associated with the aetiology of AECOPD.
The current analysis provides new insights into the pathophysiology of AECOPD from the novel perspective of glycan polymers. HS is the most abundant sulphated GAG in the lung parenchyma. Its ability to sequester water provides HS the capacity to serve as a structural component of the lungs, whereas its ability to bind positively charged soluble ligands and cell surface receptors characterizes HS as a regulator of signalling pathways. [17][18][19] COPD is a complex and heterogeneous disease, therefore, it has been suggested that the use of a group of biomarkers reflecting various pathophysiological pathways and comorbidities would better serve to predict COPD severity and outcome. [20][21][22] In this respect, it has also been suggested that in patients with stable COPD, concurrently raised levels of adrenomedullin, arginine vasopressin and atrial natriuretic peptide are associated with increased risk of death. 22 In the present study, we show that HS is significantly correlated with inflammatory serum biomarkers such as copeptin, procalcitonin, atrial natriuretic peptide and adrenomedullin. It remains to be elucidated if concomitant measurements of plasma HS together with well-established and cheaper inflammatory markers such as CRP, would increase the accuracy of the prediction of disease severity and outcome, as well as of the aetiology of AECOPD.
Despite the evidence for HS in inflammation, in the current study HS was not associated with lung function parameters, exacerbation rate and death. In a previous study performed in the same cohort of patients, we showed that another GAG, HA, was strongly associated with COPD severity and predicted overall survival. 11 This indicates that even though HS and HA share structural similarities, they have distinct functions that underlie their discrete role in lung pathophysiology.
Paradoxically, HS was lower in smokers compared to non-smokers. In line with our findings, it has been shown that gene expression of glypican 3, a HS proteoglycan, was decreased by 60% in lungs of smokers as compared with non-smokers. 23 Pharmacologic treatment of lung cell lines indicated that glypican 3 expression was epigenetically silenced by promoter hypermethylation and thus glypican F I G U R E 6 Probability of infection type at acute exacerbation of COPD (AECOPD) in relation to the change of plasma HS levels from last stable visit to AECOPD event. Each line is the result of a ranked mixed logistic regression model.
3 may be a candidate lung tumour suppressor gene whose expression is regulated by exposure to cigarette smoke and functions to modulate cellular response to exogenous damage. Furthermore, it has been shown that cadmium, an important contaminant of tobacco, inhibits the synthesis of HS proteoglycans by 25%-45% and may play an important role in the pathogenesis of emphysema by inhibiting the production of connective tissue molecules. 24 Thus, the decreased plasma levels of HS in smokers that we report here may be attributed to decreased synthesis of this GAG.
HS was not correlated with diffusion capacity or with quantitative CT scan parameters representing the status of lung parenchyma. These results indicate that circulating HS is not associated with airway obstruction or emphysema and agree with the results obtained in BAL of COPD patients. 8 However, HS was associated with MMRC and Borg scale, suggesting that high levels of HS are related to dyspnoea and perceived exertion in COPD patients.
In the present study, we provide evidence that there is a differential increase of circulating HS from stable state to AECOPD which may be related to the aetiology of the events. Distinct pathophysiological changes in extracellular molecules are associated with the type of infection at AECOPD, implicating the potential predictive value of these molecules in AECOPD.
Respiratory viral and bacterial infections are the most common causes for AECOPD. Various types of pathogens such as bacteria, viruses and parasites bind GAGs as receptors to recognize and interact with host cells. 18,[25][26][27] HS is involved in the adherence of bacteria on the surface of lung epithelial cells and enzymatic digestion of HS reduced bacterial adhesion on lung cells. 27 The molecular kind of binding depends on the type of bacteria as well as on the type of pulmonary cells involved, as different bacterial adhesins and different sulfation pattern of HS chains seem to mediate this process. Viruses also utilize HS proteoglycans for attachment to the host cell, internalization, intracellular trafficking and spread. HS is an important attachment factor for SARS-CoV-2 on human lung epithelial cells. 28,29 Human metapneumovirus also binds to apical HS in the airways and this is essential for infection. 30 The highest increase of HS from stable state to AECOPD was observed in cases with coinfection with virus and bacteria. It remains unclear if increased HS is a causative factor for the coinfection by facilitating the binding of pathogens on the cell surface, or if it reflects the result of increased inflammation due to the coinfection. This could be clarified in AECOPD groups with limited changes in the inflammatory profile, as described by Bafadhel et al. and termed as 'pauciinflammatory'. 31 The results of our study cannot establish a causal relationship between increased levels of HS and infections of any aetiology.
Interestingly, even though HSPE-1 was numerically increased at AECOPD of different aetiology compared to previous stable state, this was not significant. Therefore, it is tempting to hypothesize that the increased levels of HS in AECOPD of different aetiology do not only reflect increased cleavage of HS from proteoglycan chains and subsequent release in the circulation, but also de novo synthesis of HS.
A further limitation of our study is that measurements of HS plasma levels by ELISA do not provide information about the sulfation pattern of the molecule that is an important determinant for its function. Major strengths of our study are the large sample size, the complete clinical characterization of the patients and the longitudinal sampling undertaken at stable state, at exacerbation and at follow-up. A further strength is the detailed characterization of distinct etiological AECOPD groups that enabled to demonstrate the differential association of HS in AECOPD according to aetiology.
In conclusion, the results of the current study indicate that circulating HS increased at AECOPD and that longitudinal changes in circulating HS between stable state and AECOPD may be associated with the aetiology of AECOPD.