Validation of assays for inflammatory mediators in exhaled breath condensate

Study subjects

In total, 61 subjects were recruited for the study: 12 patients with bronchiectasis (confirmed on high-resolution computed tomography), 19 patients with COPD, and 30 normal subjects with normal spirometry and no history of lung disease. All subjects were assessed in a stable state (no antibiotics or oral corticosteroids for ≥2 months). Of the 19 COPD patients, 12 fulfilled standard criteria for the diagnosis of chronic bronchitis (CB; daily sputum production for ≥3 months of 2 consecutive yrs) 8. Patients with underlying immune deficiency, allergic bronchopulmonary aspergillosis or cystic fibrosis were excluded from the study.

Assessment and investigations

EBC was obtained using RTube^TM apparatus (Respiratory Research, Charlottesville, VA, USA) for a collection period of 20 min at -40°C, and was stored at -70°C until analysed. Collection was conducted with reference to the recommendations set out by Horváth et al. 9. Matched, spontaneous sputum sol phase was obtained from nine of the subjects with COPD and CB, and nine of the subjects with bronchiectasis. Sputum was collected in sterile containers over a 4-h period in the morning after rising. The sputum was ultracentrifuged at 50,000×g for 90 min at 4°C, and the sol phase was removed and stored at -70°C until analysed.

Following EBC collection, all subjects underwent spirometry with reversibility (20 min before and after inhalation of 400 μg of salbutamol) 10. Post-bronchodilator forced expiratory volume in one second (FEV₁) was expressed as percentage of the predicted normal reference values 11.

Assays on sputum sol phase and EBC

Mediator quantification was determined by interpolation of the signal from a standard curve of known concentrations. The intra- and interassay CV was determined for a pooled sample of EBC and samples prepared from the mediator standard measured on six occasions. In addition, recovery was determined in both EBC and standard buffer by comparing the assay result with that expected from a known mediator spike 12. Briefly, single samples or pure mediators were assayed on six occasions to obtain the intra-assay CV. At this point, a known quantity of the pure mediator was added to the sample, which was re-assayed. The result was compared with the standard curve and the new value was obtained by interpolation. This second value was subtracted from the predicted value to obtain the proportion “recovered”.

The LLD for an assay was defined as 2 sd above the mean signal for 20 sample blanks. The LLQ was defined as the point at which both intra- and interassay CV and spike recovery became acceptable for each assay (<12%). This concept is demonstrated for the SLPI ELISA in figure 1. ⇓⇓ All EBC samples were analysed without dilution.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

The performance of the secretory leukoprotease inhibitor (SLPI) assay is shown. The optical density of the ELISA (○) is related to the concentration of protein, a typical sigmoid curve is demonstrated (—). The coefficient of variation (•) of sample measurements at different SLPI concentrations is described. A major increase in the variability of the repeated measures is seen for the plateau of the curve (– – –).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

a) Leukotriene (LT)B₄, b) interleukin (IL)-8, c) secretory leukoprotease inhibitor (SLPI), d) α₁-antitrypsin (AAT) and e) myeloperoxidase (MPO) concentrations in individual exhaled breath condensate samples from normal subjects, patients with chronic obstructive pulmonary disease (COPD) and chronic bronchitis (CB), COPD patients without bronchitis, and bronchiectasis (BE) patients, shown relative to lower limit of detection (····) and lower limit of quantification (----). U: units. The logarithmic scales do not allow the representation of zero or negative values, hence these values are not shown.

The methods for LTB₄ (GE Bioscience), SLPI and MPO have been described in detail previously 12. The second commercial LTB₄ ELISA (Cayman Chemical Company) was used according to the manufacturer’s instructions.

IL-8 was measured by ELISA using a commercially available kit (R&D Systems Europe Ltd). The sol phase intra-assay CV was <8.5% throughout the working range of the assay. Sputum samples spiked with IL-8 resulted in >88% recovery.

AAT was measured by an in-house ELISA relative to a commercially available serum standard (The Binding Site Limited, Birmingham, UK). In brief, 200 μL of anti-human AAT (The Binding Site Limited) in 0.05 M sodium carbonate/bicarbonate pH 9.6 was added to a Nunc MAXISORP (Loughborough, UK) microtitre plate and incubated overnight at 4°C. The plate was then washed three times with PBS containing 1% (v/v) Tween 20 (Sigma-Aldrich Company Ltd, Poole, Dorset, UK) and 0.5% (w/v) dried skimmed milk (PBS-T), 200 μL of standard or sample was added to the plate, and it was then incubated for 1 h at 37°C. The plate was again washed three times with PBS-T, then 200 μL of anti-human AAT peroxidase conjugate (The Binding Site Limited) in PBS-T was added to each well, and the plate was incubated for 1 h at 37°C. Following three more washes with PBS-T, 200 μL of 3,3′,5,5′-tetramethylbenzidine ELISA substrate solution (Sigma-Aldrich Company Ltd) was then added to each well and incubated for 10 min at 25°C. The reaction was stopped with 50 µL of 0.1 M H₂SO₄ and the plate was read at 450 nm with a 570 nm correction, and the AAT concentration obtained by interpolation from the standard curve. The sol phase intra-assay CV was <10.2% throughout the working range of the assay and sputum samples spiked with AAT gave >85% recovery.

Statistics

Normally distributed data were expressed as mean±sd and the patient groups were compared using one-way ANOVA, with the Bonferroni correction being used if significant differences were detected. Correlations between matched sol phase sputum and EBC samples were assessed by Spearman rank correlation. Statistical significance was accepted at p<0.05.

Baseline characteristics

Baseline characteristics are shown in table 1⇓. The mean±sd age of COPD and bronchiectasis patients was similar (67.1±7.6 and 67.7±5.3 yrs, respectively; p = 1). Normal subjects were significantly younger (50±8 yrs) than both COPD (p<0.001) and bronchiectasis (p<0.001) groups. Patients with COPD had a significantly lower FEV₁ % predicted (48.9±14.7%) compared with both normal subjects (107±13.2%; p<0.001) and patients with bronchiectasis (70.3±25.8%; p = 0.007). Patients with bronchiectasis had a lower FEV₁ compared with normal subjects (p<0.001). The majority of COPD patients were either Global Initiative for Chronic Obstructive Lung Disease (GOLD) 13 stage 2 (42.1%) or stage 3 (47.4%). The remaining two (10.5%) patients were GOLD stage 4. All patients in the bronchiectasis group had an idiopathic disease (no evidence of causative/susceptibility factors such as AAT or immunoglobulin deficiencies, allergic bronchopulmonary aspergillosis, ciliary dyskinesia, etc.).

Assays on sputum sol phase and EBC

All assays had an intra- and interassay CV of <15% in the linear part of the standard curve and >85% spike recovery.

LTB₄

The LLD for the LTB₄ ELISA (GE Bioscience) was 7 pg·mL⁻¹, which was also consistent with the manufacturer’s value of ∼6 pg·mL⁻¹. Quantification was acceptable at 20 pg·mL⁻¹, when the intra-assay CV became <12% and spike recovery reached 77%. At values <20 pg·mL⁻¹, the intra-assay CV ranged from 21.4% at 12 pg·mL⁻¹ to 111.4% at 1.6 pg·mL⁻¹, with spike recoveries of between 81.8% at 12 pg·mL⁻¹ and 12.7% at 1.6 pg·mL⁻¹. Samples from only 19 subjects (12 controls, three COPD and CB, and four bronchiectatic patients) of the 61 study subjects had values that exceeded the LLD of the assay. However, no sample values exceeded the LLQ of the assay (fig. 2a).

The present authors also assessed a second LTB₄ assay (Cayman Chemical Company) and found that the LLD for this assay was 9.2 pg·mL⁻¹, which was consistent with the value reported by the manufacturer (∼13 pg·mL⁻¹). Quantification, however, only became acceptable at 30 pg·mL⁻¹, where the intra-assay CV was <12% and spike recovery was 86%. At values <20 pg·mL⁻¹, the intra-assay CV ranged from 21.4% at 16 pg·mL⁻¹ to 68.6% at 2.5 pg·mL⁻¹.

IL-8

The LLD for the IL-8 ELISA was consistent with the manufacturer’s value of ∼10 pg·mL⁻¹. Quantification became acceptable at 40 pg·mL⁻¹, where the intra-assay CV was <12% and spike recovery was >88%. Below 40 pg·mL⁻¹ the intra-assay CV ranged from 13.9% at 30 pg·mL⁻¹ to 90.7% at 8 pg·mL⁻¹, with spike recoveries of between 57.6% at 31 pg·mL⁻¹ and 6.1% at 3.9 pg·mL⁻¹. Samples from only four out of 61 subjects (two controls, one COPD, and one bronchiectatic individual) exceeded the LLD of the assay, but no sample reached the assay LLQ (fig. 2b).

SLPI

The LLD for the SLPI ELISA (25.5 pg·mL⁻¹) was consistent with the manufacturer’s value of ∼25 pg·mL⁻¹. Quantification became acceptable at 100 pg·mL⁻¹, where the intra-assay CV was <12% and spike recovery was 102.8%. At values <100 pg·mL⁻¹, the intra-assay CV ranged from 10.2% at 62 pg·mL⁻¹ to 56.2% at 8 pg·mL⁻¹, while spike return varied from 78% at 62 pg·mL⁻¹ to -60% at 8 pg·mL⁻¹. The relationship between the assay CV and the linear part of the ELISA sigmoid curve is shown in figure 1. ⇑No samples from the 46 subject samples for which this was assayed reached either the LLD or the LLQ (fig. 2c).

AAT

The in-house AAT ELISA had an LLD of 0.8 ng·mL⁻¹. Quantification was acceptable at 2 ng·mL⁻¹, when the intra-assay CV was <12% and spike recovery was >77%. At values <2 ng·mL⁻¹, the intra-assay CV ranged from 55.5% at 0.5 ng·mL⁻¹ to 119.2% at 0.03 ng·mL⁻¹, while spike recovery ranged from 75.5% at 1.75 ng·mL⁻¹ to 54.4% at 0.2 ng·mL⁻¹. In total, 46 samples were assayed. Samples from 28 of the subjects (13 controls, five COPD and CB, six COPD, and four bronchiectatic individuals) reached the LLD of the assay. Samples from nine subjects (three controls, five COPD, with CB and one COPD patient) exceeded the LLQ of the assay (fig. 2d).

MPO

The LLD for the MPO activity assay was 0.015 units·mL⁻¹. Quantification became acceptable at 0.05 units·mL⁻¹, at which point the intra-assay CV was <4.59% and spike recovery was 103.34%. Samples from 52 subjects were assayed. No samples reached either LLD or LLQ (fig. 2e).

Correlation with sputum

If the issues raised previously are not appreciated, interpolation of the signal from the standard dose–response curve would provide apparent “values” in the detectable range. For example, the LTB₄ levels, when above the LLD, would be on average 11.3 pg·mL⁻¹, by interpolation using the linear part of the standard curve (similar to levels reported in the literature). However, even when these derived values were related to the sputum value for the same patient sample, there was no significant direct correlation for any mediator, when corrected for the effects of multiple analyses (table 2⇓).