COPD association and repeatability of blood biomarkers in the ECLIPSE cohort

Chronic obstructive pulmonary disease (COPD) is a major cause of global morbidity and mortality and is predicted to become the third leading cause of death by 2020 [1]. It is a multicomponent condition involving both local and systemic pathological processes that include airway obstruction, emphysema, mucus hypersecretion, loss of lean body mass and an increased risk of cardiovascular disease [2]. The most widely used marker of disease severity and progression is the forced expiratory volume in one second (FEV₁). However FEV₁ correlates poorly with both symptoms and other measures of disease progression [3]. Moreover it does not differentiate between the causes of airflow obstruction (i.e. small airways disease or emphysema) or identify the extra-pulmonary manifestations of COPD. There is clearly a need for biomarkers that can aid with the diagnosis, risk stratification and the assessment of therapeutic interventions. The search for biomarkers has centred around proteins and other molecules in exhaled breath condensate, sputum, urine, bronchoalveolar lavage and blood that have been implicated in the pathogenesis of COPD [4]. Profiling of blood biomarkers has identified a number of biomarkers that may distinguish individuals with COPD from control subjects [5]. However, additional research is needed to determine the value of biomarkers in characterizing COPD patients, following disease progression, and/or the ability to serve as intermediate or surrogate endpoints in clinical studies.

We have used the ECLIPSE (Evaluation of COPD Longitudinally to Identify Surrogate Endpoints) cohort to evaluate a number of previously reported biomarkers, and to identify new biomarkers for COPD [6]. This cohort has already been used to evaluate serum CC-16, surfactant protein D (SP-D) and CCL-18/PARC. SP-D and CCL-18/PARC were elevated and CC-16 was decreased in individuals with COPD when compared to control subjects [7‐9]. We report here the assessment of a panel of 34 biomarkers in a subgroup of individuals from the ECLIPSE cohort. The biomarkers were selected based on previous work suggesting their potential association with the pathogenesis of COPD [5, 10].

A sub-group of individuals was selected from the ECLIPSE study (SCO104960, Clinicaltrials.gov identifier NCT00292552) as the "biomarker cohort" in order to evaluate 34 potential blood biomarkers in COPD [7, 8]. The aims and operational aspects of ECLIPSE have been described elsewhere [6]. Briefly ECLIPSE is a 3-year multicentre, longitudinal, prospective study to identify novel endpoints in COPD. Individuals aged 40-75 years were recruited to the study if they had a smoking history of ≥ 10 pack-years, a post bronchodilator ratio between forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC) ≤ 0.7 and Global Initiative on Obstructive Lung Disease (GOLD) stage II (FEV₁ 50-80% predicted), III (FEV₁ 30-50% predicted) or IV (FEV₁ < 30% predicted) COPD. Smoking (≥ 10 pack-years) and non-smoking (< 1 pack-year) control subjects were enrolled if they were aged 40-75 years and had normal lung function (post bronchodilator FEV₁ > 85% predicted and FEV₁/FVC ratio > 0.7). The "biomarker cohort" consisted of 275 subjects (201 COPD subjects with GOLD stage II-IV disease, 37 ex-smoker controls and 37 non-smoker controls) matched to the full ECLIPSE cohort. Only former smokers were included to avoid the influence of acute cigarette smoke exposure that might increase variability in our relatively small sample size.

The methods used to evaluate lung function and the extent of emphysema have been described previously [8]. The number of exacerbations was assessed by telephone contact on a monthly basis enquiring about details of contact with doctors or hospital and the need to take medications for exacerbations (oral steroids and/or antibiotics) [11]. Exercise capacity was measured at enrolment with a standard supervised six-minute walk test. Assessments of depression, fatigue and level of dyspnoea were assessed with the CES-D (centre of epidemiological studies depression scale) [12], FACIT (functional assessment of chronic illness therapy) [13] and modified MRC (Medical Research Council) [14] dyspnoea scores, respectively.

A subgroup of the ECLIPSE cohort was used to assess thirty-four putative blood biomarkers for their associations with important characteristics of COPD. The analytes that were chosen have previously been associated with either local or systemic features of COPD [5, 10]. In subjects with stable COPD, fifteen analytes had significantly different concentrations in the COPD cohort as compared with smoker and non-smoker controls. Twelve of these were assessed for repeatability; this revealed wide variation over 3 months for many of the analytes, particularly CRP. Similar variability in a number of the biomarkers assessed in the ECLIPSE cohort has been shown by Man and colleagues in a cohort of 41 COPD subjects over a two week period [16] though less variable results were reported when measured at a yearly interval [17].

It is known that COPD is an inflammatory disease [18] and indeed CRP, fibrinogen and IL-6, all considered to be markers of systemic inflammation, are all significantly raised in COPD subjects compared to control subjects in this and in other studies [19‐22]. They have received much attention to determine their value as potential COPD biomarkers [23]. In addition, CRP, fibrinogen, and IL-6 have been reported to be elevated during exacerbations of COPD [24‐28]. Increased CRP and fibrinogen have also been associated with increased risk for COPD hospitalizations [29‐31]. In the ECLIPSE biomarker cohort, fibrinogen was elevated in COPD subjects, demonstrated excellent repeatability over time in stable disease and was associated with BODE index and MRC dyspnoea score. Data from this cohort have not confirmed a previously recorded correlation with GOLD stage [32], but baseline fibrinogen associates with exacerbation frequency in ECLIPSE and other cohorts [11, 33] and with exercise tolerance. Thus it may reflect a more global assessment of disease severity and therefore be a useful biomarker in COPD. IL-6 and CRP are both significantly raised in COPD and are yet further raised during exacerbations; however they display wide variability in stable subjects over three months and therefore their value as useful biomarkers of COPD would appear to be limited. Additionally, we found no significant association between TNF-α and COPD despite an association being described previously [34]. Importantly, circulating TNF-α was detectable in only approximately 10% of the subjects in the ECLIPSE biomarker cohort. The finding of low circulating TNF-α and the inability to detect it in many subjects in the current study is similar to that reported for COPD subjects included in a study of the effects of infliximab [35]. Further evaluation of serum TNF-α in the full ECLIPSE cohort confirmed that it is not measurable in a large percentage of subjects [36]. We also observed that serum concentrations of SP-D were both reproducible and elevated in COPD subjects with unresolved exacerbations compared with those with stable disease. This finding is consistent with our earlier report associating baseline serum concentrations of SP-D with exacerbations in the first year of ECLIPSE [8] and with the findings of others [37, 38].

Many of the analytes studied that were different between cases and controls may play important roles as chemoattractants for monocytes/macrophages (IL-8, IL-12p40, CCL2), lymphocytes (CCL18, CXCL10) and neutrophils (IL-8) in driving inflammation in individuals with COPD. Unfortunately, many of the analytes we assessed showed poor repeatability when measured at a 3-month interval and thus they may have limited utility as markers of disease progression or measures of the effects of drug intervention. It is however possible that they may be useful to help define subgroups of COPD subjects based on common underlying disease mechanisms [39]. Studies with larger numbers of subjects will be required to determine if this is the case.

Aberrant tissue destruction and repair is known to be central to the development of emphysema [10]. MMP-8 and MMP-9 are two proteases that are believed to play an important role in the tissue destruction that results in emphysema [40, 41] and both are raised in COPD subjects in the biomarker cohort. Consistent with other studies [42, 43], there was evidence that they correlated with GOLD stage and amount of emphysema and so may be useful in differentiating individuals with emphysema from those with predominantly airway disease. Hepatocyte growth factor plays a role in tissue regeneration [44] and was shown in a small study to be increased in bronchoalveolar lavage of COPD subjects and smokers with normal lung function [45]. The repeatability of MMP-8 remains untested; however the wide variability of MMP-9 and hepatocyte growth factor over 3 months suggests that they are unlikely to be useful in clinical applications.

β-defensin-2 and myeloperoxidase play a role in pulmonary immunity [46, 47]. We provide supportive evidence that circulating myeloperoxidase is raised in COPD [48] and show that β-defensin-2 is also higher in individuals with this disease. β-defensin-2 may also reflect disease severity as defined by GOLD stage. However additional profiling is required before it can be confidently deemed a useful biomarker in COPD.

Adiponectin and prolactin levels have been shown to be altered in subjects with COPD [5, 49], however the mechanisms underlying this remain uncertain. Both have wide ranging metabolic effects and possible roles in the inflammatory response [50]. For both analytes, significant differences were seen between COPD subjects and only one of the control groups and marked variability was seen in stable disease. Although highly elevated in COPD subjects, the temporal variability associated with serum myeloperoxidase may confound its use in clinical trials. However, adiponectin, prolactin and myeloperoxidase may have some utility in diagnostic tests that combine a panel of inflammatory markers to describe the underlying inflammatory process in some COPD patients.

Our study has several limitations. In assessing putative biomarkers for repeatability, measures were taken at baseline and at 3 months which is the approximate duration of many COPD Phase IIa intervention studies. Though this interval is relatively short, it is conceivable that during this time period there may have been significant clinical changes, changes in pharmacological treatment or as we observed in our cohort, intercurrent exacerbations. Any of these could create a "true" change in biomarker level and thus the assessment of repeatability may overestimate how variable markers are over time. For example, Hurst [24], reported that the concentrations of CCL18, adiponectin, CRP, and IL-6 are increased during exacerbations. In order to better estimate reproducibility during stable disease, results from subjects with exacerbations that were not resolved at least 30 days prior to the 3 month visit were excluded from the repeatability analysis as were subjects who had been taking oral steroids within 2 weeks of the 3 month visit. The control populations in the ECLIPSE cohort were chosen on the basis of smoking status and the absence of COPD. Individuals may have co-morbidities that are inflammatory in nature and therefore influence the levels of biomarkers. However this is unlikely as details of co-morbidities were obtained at recruitment to the study. Though many of the biomarkers examined showed statistically significant differences between groups, no allowances were made for multiple comparisons in interpreting the statistical significance of the results [51]. Additionally corrections had to be made for differences in baseline characteristics between cases and controls as this study was intended to be exploratory with the primary intent of identifying biomarkers for follow up in the full ECLIPSE cohort or other COPD cohorts.

We have identified several potentially clinically relevant blood biomarkers that are elevated in COPD subjects when compared to controls. In particular, fibrinogen appears to be the most repeatable over the three-month period tested. Surfactant protein D and CC-16 have also been shown to be repeatable [7, 8]. These now need to be more rigorously assessed to clarify their repeatability in stable disease and studied in a longitudinal fashion to ascertain whether they can be used as surrogates for disease phenotype, severity and/or progression. Interventional studies will also be needed to qualify them for use as intermediates in COPD drug development.