|
|
||||||||
1 Depts of Otorhinolaryngology and 2 Clinical Chemistry, University of Ulm, Medical School, Prittwitzstr 43, 89075 Ulm, Germany
CORRESPONDENCE: H. Riechelmann, University of Ulm, Medical School, Prittwitzstr 43, 89075 Ulm, Germany. Fax: 49 7315026703. E-mail: herbert.riechelmann@medizin.uni-ulm.de
Keywords: biological factors, body fluids, immunological, nasal cavity, nasal lavage fluid, specimen handling
Received: August 8, 2002
Accepted November 14, 2002
This research was supported by a grant from the State of Baden-Württemberg, BWPLUS L98 002, Germany.
| Abstract |
|---|
|
|
|---|
Concentrations of various protein were repeatedly determined in nasal secretions of 16 healthy volunteers. The proteins were detected by using: 1)
2-macroglobulin as a marker for plasma contamination; 2) lactoferrin as a marker for glandular secretion; 3) lactate dehydrogenase as a marker for tissue injury; and 4) interleukin (IL)-1ß, IL-8, tumour necrosis factor-
, and eosinophil cationic protein and tryptase as indicators for tissue inflammation. A total of four different sampling methods, including nasal lavage (NL) and a new polyurethane foam sampler technique (PFST) were employed.
Analyte concentrations in NL were approximately10-times lower than in specimens obtained by PFST. Due to the unpredictable dilution during NL, various analytes were below the detection limit of the high sensitivity assays employed. With PFST, concentrations below the detection limit rarely occurred. The specimens did not significantly differ regarding plasma contamination, glandular secretion or tissue injury.
The considerable variability of reported analyte concentrations in nasal secretions mainly results from different sampling techniques. To collect nasal secretions, samplers are considered superior to nasal lavage techniques.
In recent years, investigations on upper airway mucosa inflammation, e.g. in response to inoculation with bacterial or viral pathogens 13, allergen challenge 46, or exposure to environmental pollutants 79, have focused on the detection of minute amounts of cytokines and inflammatory mediators. For this purpose, various methods to collect nasal secretions are employed (table 1
|
In dilution techniques, a liquid is instilled into the nose, recovered with an admixed and sample of epithelial lining fluid 13, 14. The dilution technique most commonly applied is nasal lavage 13. Since unknown fractions of the lavage fluid may be swallowed or absorbed, these techniques are associated with a substantial, often unpredictable, dilution of the nasal secretion 11. As a consequence, the concentrations of analytes may reveal high variability and frequently fall below the lower detection limits of the assays applied 12.
An alternative technique in the collection of nasal epithelial lining fluid involves the use of a sampler with absorptive properties which is placed within the nasal cavity. The absorption technique overcomes the problem encountered when only small quantities of spontaneous secretions are available, as it provides sufficient amounts of undiluted nasal secretions to enable various laboratory investigations to be undertaken 15. However, these techniques have been thought to traumatise the nasal mucosa and therefore alter the concentration of biomarkers under investigation.
In this study, the influence of four sampling techniques on several analytes in nasal secretions of healthy volunteers were explored, including nasal lavage (NL), a nasal spray-blow technique (NSB), a filter paper method (FPM), and a polyurethane foam sampler technique (PFST). Total protein (TP),
2-macroglobulin (A2M) as a marker of blood plasma contamination 16, lactate dehydrogenase (LDH) as a marker of tissue injury, lactoferrin (LTF) as a marker of glandular secretion 17, 18 and interleukin (IL)-1ß, IL-8, tumour necrosis factor (TNF)-
, eosinophil cationic protein (ECP), and tryptase (TRP) as commonly employed markers of airway inflammation were assessed. The following questions were addressed, in detail, in this study: 1) which method provides sufficient specimen volume for the assays intended; 2) how frequent are samples with analyte concentrations below the detection limit of the assay; 3) how is the inter-individual variability and intra-individual variability of repeated measurements within 1 day; and 4) are there indicators for mucosal trauma or specific alterations of the composition of obtained secretions depending on the collection technique.
| Methods |
|---|
|
|
|---|
Sampling techniques
A total of four sampling techniques were employed; NL, NSB, FPM and PFST. The volunteers were acclimatised to the indoor conditions and their noses were cleaned of any excess mucus. No nasal decongestants, anaesthetics or prewashes were applied.
For NL each nostril was washed with 5 mL isotonic sodium chloride solution. Volunteers tilted their head back at a 45° angle and closed the nasopharynx with the soft palate. After 10 s, the volunteer blew their nose forcefully onto a glass dish. For NSB technique isotonic saline solution was applied with a pump metered-dose dispenser (Allergopharma, Reinbek, Germany) to the sitting volunteer. A 400 µL aliquot was delivered into each nostril. The volunteer was asked to stop breathing during application. Following a 20-s period after the application, the volunteer blew their nose forcefully onto a glass dish. The sample was washed from the dish into a centrifuge tube with 1 mL isotonic saline solution. For FPM white filter paper stripes (40x10x0.37 mm, GB 002, 195 g·m2; Schleicher & Schüll, Dassel, Germany) were placed in the middle portion of the nasal septum, posterior to the mucocutaneous junction of each nasal cavity. The filters were removed after a 10-min period and placed on the piston of a syringe in a centrifuge tube (Labcon, San Raphael, CA, USA), and the fluid was separated from the strips by centrifugation (5 min, 500xg, 4°C). The pellet without the basal cell layer was further processed. For PFST, open cell flexible polyurethane foam with a specific weight of 30 kg·m3 (ISO 5999, 1982) was cut by the local distributor in squares of 28x18x6 mm. The fluid retention capacity of one sampler is
2.5 mL. Samplers were placed into each nasal cavity posterior to the mucocutaneous junction under direct visualisation and left in place for 10 min. Following removal, the fluid was extracted from the sampler by centrifugation as described earlier.
Handling and storage
Immediately following harvest the secretions were placed on ice. Cellular elements were instantly separated from the liquid phase by centrifugation at 500xg for 5 min at 4°C and further processed. The cell-free supernatants were then homogenised by ultrasonication at 160 Watts for 5 min (Sonorex RK 100 SH, Bandelin electronic Ltd, Berlin, Germany). The amount of fluid obtained was measured volumetrically. Then, aliquots of 100 µL of the homogenised fluids were stored at 20°C for further processing.
Sampling schedule
Each sampling technique was performed in each subject three times on each examination day (08:00, 08:30 and 16:00 h). Between each examination day, 2 days were left to minimise carry over effects between the different sampling techniques. Moreover, the sequence of sampling techniques followed a complete Latin square design.
Determination of biomarkers
The concentrations of all substances were determined in duplicate and calculated as the actual concentrations in the undiluted specimen, as it had been obtained, by each sampling technique. The specimens were diluted to adapt to the working range of the various assays according to previous checkerboard titrations (table 2
). Immunoassays were performed with the microplate reader MRX employing the software Biolinx (Dynatech Ltd, Denkendorf, Germany).
|
Quantikine or Quantikine HS Immunoassays (R&D Systems, Wiesbaden-Nordenstedt, Germany) were employed. ECP and TRP were measured with a matrix bound time-resolved fluorescence assay (UniCAP; Pharmacia, Freiburg, Germany). The amount of substrate needed for each assay, the lower detection limits and the employed dilution factors are given in table 2
Statistics
For each sampling method and biomarker, specimens with detectable concentrations were counted. The mean±sd and additional parameters were calculated from specimens with measurable analyte concentrations. Distributions of actual and logarithmically transformed analyte concentrations were explored graphically and with the Kolmogorov Smirnov test. For each sampling method, the inter-individual coefficient of variation (CV) was calculated in per cent and the intra-subject coefficients of the variation of the three samples obtained at 08:00, 08:30 and 16:00 h within 1 day were computed.
Partial correlations were performed after regression of each variable on the four sampling methods employed. Depending on the distribution and homogeneity of variance, differences between biomarker concentrations obtained with the four sampling techniques were explored with one-way analysis of variance or the Friedman test. Statistical calculations were performed using Systat® (Evanston, IL, USA).
| Results |
|---|
|
|
|---|
|
|
60%, 30 min after the initial sampling and reached the initial values again after 8 h (fig. 2
|
|
IL-1ß was log-normally distributed (p>0.8) and differed widely between the four sampling techniques employed (table 4
). The time/course of IL-1ß concentrations was similar to that of the total protein. The mean intra-subject CV was 54% with NL, 59% with NSB, 52% with FPM and 44% with PFST. IL-8 was detectable in almost all specimens (table 3
). The concentrations of IL-8 obtained from the four sampling methods were log-normally distributed. The inter-indivdual CV was 110% for NL and NSB, 85% for FPM and 70% for PFST, respectively. The time/course of IL-8 concentrations obtained at the three sampling times within 1 day was similar to that of total protein. The intra-subject CV was 65% with NL, 67% with NSB, 42% with FPM and 43% with PFST. Excluding the influence of the collection method by partial correlation analysis, logarithms of IL-8 concentrations correlated with logarithms of IL-1ß levels (r=0.66, p<0.01). This correlation became particularly remarkable with PFST alone (r=0.83, p<0.01, fig. 3
). TNF-
was frequently below the detection limit (table 3
). Data for the time/course of TNF-
concentrations are available for NL (30% of initial values at 08:30 h and 55% at 16:00 h) and NSB (55% of initial values at 8:30 h and 60% at 16:00 h). Intra-subject variation for NL was 49% and NSB 34%. ECP concentrations revealed an irregular distribution. It was detected in most specimens obtained (table 3
). The inter-individual CV was 127% with NL, 145% with NSB, 81% with FPM and 118% with PFST. The time/course of ECP concentrations was similar to that of LTF. The intra-subject coefficient of variation was 40% for NL, 58% for NSB, 63% for FPM, and 47% for PFST. Logarithms of ECP-concentrations correlated with the logarithms of IL-1ß (r=0.57, p<0.01) and IL-8 concentrations (r=0.53, p<0.01). In the specimens with detectable TRP concentrations, levels averaged at 5.6+4.9 ng·mL1 in NL, 5.3±3.7 mL1 in NSB, 37.6±31.4 mL1 in FPM and 40.0±48.0 mL1 in PFST. Data obtained did not provide information on inter-individual or intra-subject coefficient of variation. Logarithms of tryptase levels obtained with the various sampling techniques correlated with logarithms of ECP-levels (r=0.72, p<0.005).
|
| Discussion |
|---|
|
|
|---|
In this study, four sampling techniques were compared with regard to reliability of detection, mean values, and variability of IL-1, IL-8 and TNF-
, ECP and tryptase in healthy unchallenged volunteers. To assess the effects of repeated samplings, nasal secretions were collected in each subject at 08:00, 08:30 and 16:00 h with the same technique. In addition to the biomarkers of interest, LDH was assessed as an indicator of cell damage, A2M as an indicator of vascular leakage 6, and LTF was included as a marker of glandular secretion 17, 18. However, the study was carried out without any form of nasal provocation. Therefore, comparisons on how reproducible the different methods are with respect to differing amounts of nasal secretions e.g. a strong increase after an allergen challenge, cannot be evaluated.
Although sufficient amounts of sampling fluid were obtained with all NL, analyte concentrations were below the detection limit of 35% of the highly sensitive assays used. To overcome the problem of low analyte concentrations, nasal lavages can be concentrated using various techniques 5, 26. However, these techniques are associated with several disadvantages 27. For NL, participants had to be trained to acquire comparable recoveries and the intricate handling renders this method difficult for field studies. The FPM yielded dry specimens in 10 of 48 samples. These two techniques are considered less suitable for cytokine determinations when normal controls are included in the evaluation. NSB yielded sufficient specimen amounts in all participants, however, analyte concentrations were below the detection limit in 21% specimens. In this investigation, PFST was superior to the other sampling techniques. In one sample, an insufficient amount of sampling fluid was obtained. Despite the supposed mucosal trauma induced by the sampler, A2M and LDH were more frequently detected in NL than in PFST, suggesting that plasma transudation and tissue trauma does not disproportionately interfere with the results of PFST. Further advantages of PFST are easy handling, comparatively small variability and minor cooperation required by the examinee.
In nasal secretions, some proteins are found in concentrations considerably higher than in blood plasma. IL-8 was found in nasal secretions in concentrations 10100 times higher than in human sera 28. Unless the samples of FPM or PFST are diluted to adapt to the working range of the assay, falsely low IL-8 levels may be measured (fig. 4
). Possible causes for this matrix phenomenon include decreased accessibility to the immobilised antibody due to space occupying molecules and unspecific binding of competing molecules present in high concentrations 29. If the cytokine concentration in nasal secretions is not known in advance, it is advisable to perform checkerboard titrations to adapt to the working range of the assay.
|
The detection of cytokines and other inflammatory mediators in nasal secretions of healthy, nonallergic and nonsmoking volunteers is consistent with current evidence on airway epithelial cell cytokine expression 30, 31. Interleukin-1ß and tumour necrosis factor-
are constitutively produced at low levels, whereas interleukin-6 and particularly interleukin-8 are secreted in high concentrations. Eosinophil cationic protein is detectable in low concentrations also in normal nonallergic individuals, whereas tryptase was rarely detected. The pattern of cytokines assessed in this study suggests that the epithelial layer is the main source of protein in the four sampling techniques employed.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. Scagnolari and G. Antonelli Determinants of Reduced Interferon Gamma Expression in Severe Respiratory Syncytial Virus Disease: The Authors Respond Experimental Biology and Medicine, May 1, 2008; 233(5): 495 - 495. [Full Text] [PDF] |
||||
![]() |
J. R. Hurst, T. M. A. Wilkinson, W. R. Perera, G. C. Donaldson, and J. A. Wedzicha Relationships Among Bacteria, Upper Airway, Lower Airway, and Systemic Inflammation in COPD Chest, April 1, 2005; 127(4): 1219 - 1226. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Polzehl, M. Weschta, A. Podbielski, H. Riechelmann, and D. Rimek Fungus culture and PCR in nasal lavage samples of patients with chronic rhinosinusitis J. Med. Microbiol., January 1, 2005; 54(1): 31 - 37. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |