Analysis of fluidphase mediators
- Leader of the Working Group: M.M. Kelly1,
- Members of the Working Group:,
- V. Keatings 2 ,
- R. Leigh 1 ,
- C. Peterson 3 ,
- J. Shute 4 ,
- P. Venge 5 and
- R. Djukanović 6
- 1Firestone Institute for Respiratory Health, Hamilton, Ontario, Canada. 2Dept of Medicine, St Vincent's Hospital, University College Dublin, Dublin, Ireland. 3McMaster University, Hamilton, Ontario, Canada. 4Pharmacia & Upjohn Diagnostics, Uppsala, Sweden. 5School of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, UK. 6Dept of Medical Sciences, Clinical Chemistry and Pharmacology, University Hospital of Uppsala, Sweden. 7Southampton University General Hospital, Southampton, UK
- M.M. Kell, Firestone Institute for Respiratory Health, St Joseph's Healthcare, 50 Charlton Avenue East, Hamilton, Ontario, L8N 4A6, Canada. Fax: 1 905 5216125. E‐mail: kellym@mcmaster.ca
Sputum cellular indices are valid, reliable and responsive to change 1–7. Increasingly, numerous inflammatory mediators are being measured in the fluid phase of sputum; these include cytokines, chemokines, granulocyte proteins, markers of vascular leakage, eicosanoids, proteases and others. Many of these mediators are included in table 1⇓, which gives details of methods used by investigators to process sputum and measure mediator levels. The table also provides the median/mean levels measured in studied subject groups to give an indication of the expected levels of these mediators in sputum. However, the reproducibility, precision and validity of many of these measurements in sputum have not been investigated and, therefore, their utility as a research and clinical tool remains uncertain and requires confirmation.
The following issues are important in the analysis of mediators: 1) choice of methods for measuring fluidphase mediators; 2) points to consider when planning an immunoassay; and 3) evaluation of the measurement of a soluble mediator in sputum, i.e. validation of the measurement method.
Methods for measuring fluidphase mediators
The three main types of method used for the measurement of sputum fluidphase mediators are bioassays, enzyme assays and immunoassays.
Bioassays
Bioassays rely on the retention of biological activity and the ability to exert a measurable effect, such as proliferation of cells, bone marrow colony formation or chemotaxis 3, 11, 58. Sputum processing with mucolytics such as dithiothreitol (DTT) or dithioerythritol, which are strong reducing agents, may decrease the biological activity of cytokines, many of which rely on disulphide bonds to provide a stable structure for bioactivity. Bioassays also have the disadvantage of being inconvenient, timeconsuming and lacking in specificity 59. In addition, the presence of commonly occurring endogenous cytokine inhibitors, although allowing an estimate of net activity, may result in significant underestimation of total cytokine levels 58.
Enzyme assays
Many of the mediators of the inflammatory response are enzymes, released from cellular sites of synthesis into an environment replete with enzyme inhibitors. Again, an estimate of net activity is important when evaluating their potential contribution to tissue responses. The proteases neutrophil elastase 20, 22, 60, 61, cathepsin G 20, 61 and cathepsin B 22 have been assayed in sputum using specific chromogenic substrates, from which proteases release a coloured product that can be quantified spectrophotometrically. Net protease activity is determined using purified enzyme standards. Active forms of matrix metalloproteinases in sputum are identified by means of substrate gel zymography, and net activity can be quantified using radiolabelled substrates 62. Other proteases, such as chymase and tryptase, are less robust. Their activity rapidly diminishes on freezing/thawing sputum samples, and immunoassays are the best means of detecting them (see below).
The activity of enzymes involved in oxidant/antioxidant balance, myeloperoxidase (MPO) 60, 61, 63 and catalase 63, similarly can be measured in sputum using spectrophotometry. The presence of sulphated glycoconjugates and deoxyribonucleic acid (DNA) in sputum in severe asthma 64 may inhibit MPO activity 63, and immunoassays carried out in parallel may also be required. However, studies in which sputum was spiked with MPO have yielded better recovery in the enzyme assay than the immunoassay 21.
Immunoassays
Immunoassays are the method of choice for measuring sputum fluidphase mediators because of their convenience, reproducibility and specificity. Their sensitivity is improving, and now approaches that of bioassays. A full description of the theoretical basis of immunoassay is beyond the scope of this report; however, the most widely used are competitive radioimmunoassays (RIAs), using radiolabelled antigen, and immunometric assays such as enzymelinked immunosorbent assays (ELISAs), which use enzymelabelled antibodies.
Immunoassays are indirect methods whereby the concentration of analyte is measured with respect to a parameter related to its concentration, using either counts per minute (radioactivity) or the optical density of the product of the antibodylinked enzyme at a certain wavelength 65. In competitive assays, the concentration of mediator present in the sample is inversely proportional to the radioactivity (RIA), whereas, in immunometric assays, the mediator concentration is proportional to the optical density of the product formed by the conjugated enzyme (ELISA). The concentration of the mediator is calculated with reference to a standard curve constructed by serial dilution of a standard mediator preparation. Since the standard is often a recombinant protein, its structure and degree of glycosylation and amidation depend on how it is produced 58. Therefore, the antibodies used often have different affinities for the standard as opposed to the endogenous mediator. Standardisation between different assays for valid comparisons is, therefore, problematic 66. In response to this, the World Health Organization Consultative Group on Cytokines is assigning International Standards or Reference Reagents 67–69 for each known cytokine, which are intended to be used to calibrate secondary and working standards. This will eventually allow valid interassay comparisons. Similar standards are also required for immunoassays of other soluble mediators 70.
Points to consider when planning an immunoassay
Immunoassay is the commonest means of measuring mediator levels. Therefore, its use is considered here in detail, although some of the following points could be applied to bioassays and enzyme assays. It is imperative that factors affecting the validity of the assay be recognised and controlled for, including those specific to: 1) the assay itself; 2) the soluble mediator being measured; and 3) the unique nature of sputum and the method used to process it.
Factors specific to the assay
The ease of production of monoclonal antibodies as well as recombinant antigens has permitted the development of numerous commercial and laboratoryspecific (“inhouse”) immunoassays, particularly ELISAs. The immunoassay operator should be trained in the application of immunoassays and be able to troubleshoot 71. Variability between production lots should be minimal, and the operator should be satisfied that the laboratory or company supplying the kits is reliable. Information on the validity, sensitivity, specificity, crossreactivity, predictive value and precision of each assay should be available 21, 71, but additional quality control and optimisation often need to be performed by the operator. The sensitivity of the assay refers to the lowest concentration that can be reliably detected above the background, and is, therefore, its detection limit. It is usually expressed as the amount of analyte detectable per millilitre. However, some manufacturers convert this figure to the amount of analyte per well, which is a much smaller figure, and it is necessary to be aware of this. The lowest standard suggested by the manufacturer is usually the detection limit of the assay, although operators can test this for themselves. The stability of reagents such as the standard used in the assay should be ensured by storage at the correct temperature. If necessary, the standard should be stored frozen in aliquots so that it is not continually subjected to freezing and thawing. Internal laboratory controls should be used routinely in each assay so as to determine intra and interassay variability in the hands of the user.
Factors specific to the soluble mediator
Cytokines are often bound to various molecules present in biological fluids (such as soluble receptors, α2‐macroglobulin and autoantibodies), and these are usually present in excess 72–78. A cytokine may not be recognised by the capture antibody of the assay if the relevant epitope is hidden when the cytokine is bound within a complex. Polyclonal capture antibodies recognise several epitopes and, therefore, may be more effective in detecting complexed cytokines. Competitive assays using monoclonal antibodies are likely to be more accurate in the presence of complexed cytokine but are not widely available. Chemokines, such as interleukin (IL)‐8, are small highly basic proteins, with an affinity for acidic negatively charged molecules, such as heparin, that are present in sputum samples 79. Heparin inhibits the chemotactic function of soluble IL‐8 in vitro 80, although binding does not necessarily imply an effect on function. Using gel filtration chromatographic analysis of IL‐8 in sputum samples, very highmolecularweight forms 81 that no longer bind to heparin are found, and these include IL‐8 bound to α2‐macroglobulin 80, 81, immunoglobulin A 82, DNA 83 and actin 83. These highmolecularweight molecules not only potentially mask detection of IL‐8 in immunoassays 83 but also affect the function of this chemokine. DNA is present in relatively high concentrations in sputum from patients with severe asthma 64, chronic obstructive pulmonary disease 85 and cystic fibrosis 64, 85; this may inhibit the function of IL‐8 in sputum 85. These factors should be recognised when interpreting results.
Factors specific to the unique nature of sputum and the method used to induce and process it
The unique nature of sputum
Many of the assays used to measure soluble mediator levels in sputum have been developed for serum or culture supernatants. However, in being a mixture of mucus, cellular degradation products, DNA, substances secreted by airway cells (including proteases, soluble receptors, autoantibodies, and binding and carrier proteins), and inflammatory and epithelial cells, sputum cannot be assumed to be equivalent to either of these. Sputum, like all biological fluids, may produce matrix effects in assays by alteration of pH or ionic strength. Spontaneously expectorated sputum from patients with infective bronchitis, bronchiectasis or cystic fibrosis is even more complex due to high levels of viscid DNA and actin released from necrotic inflammatory cells in addition to high concentrations of proteases 57. As mentioned in the Factors specific to the soluble mediator section, α2‐macroglobulin, which is present in sputum, binds certain cytokines (plateletderived growth factor, IL‐2, IL‐1β, IL‐6, tumour necrosis factor‐α (TNF‐α), interferon gamma and IL‐8), and may interfere with the recognition of cytokine epitopes by the assay 86.
The method used to obtain sputum
Induced versus spontaneous sputum
Fibrinogen levels have been found to be increased in spontaneous expectorated samples compared to induced sputum 87. There is a trend towards increased eosinophil cationic protein (ECP) levels in spontaneously expectorated sputum.
Factors that affect detection in induced sputum
β2‐Agonist pretreatment
Histamine levels are slightly reduced after salbutamol pretreatment (as might be expected given the mast cellstabilising properties of salbutamol), but ECP levels are unaffected 88.
Concentration of saline solution
One study looked at the effect of isotonic or hypertonic saline induction on ECP and histamine levels and found no effect 89.
Duration of inhalation
Since it has been shown that the composition of sputum changes as induction time increases (the proportion of granulocytes is reduced and that of macrophages and lymphocytes increased), it would be expected that concentrations of mediators might also change. During a 20min induction, 4‐min portions of sputum were examined and ECP and mucinlike glycoprotein levels found to be decreased, with surfactant protein A levels increasing during induction 90. This suggests that, if the study of proximal bronchial secretions is the objective, it is best to analyse samples collected early on in the induction. Similar findings were obtained in another study, in which the induction lasted 30 min; the concentration of ECP decreased in sputum collected sequentially every 10 min 91.
Method used for sputum processing
Delay before processing
This has not been systematically investigated, but recommendations are that sputum be processed as soon as possible and kept at 4°C in the meantime.
Selected sputum versus entire expectorate
As stated in the article entitled “Methods of sputum processing for cell counts, immunocytochemistry and in situ hybridisation” 95, two main methods are used for the processing of sputum: selection of viscid portions, in an attempt to minimise contamination by saliva, and processing of the entire expectorate, which contains variable amounts of saliva. Salivary contamination is often reduced by spitting out saliva separately during induction. Saliva generally contains much lower levels of ECP, tryptase, mucinlike glycoprotein, lactoferrin, DNA, elastase, α1‐antitrypsin, fibrinogen and albumin than unselected sputum 17, 26, 57, but histamine and endothelin‐1 levels have been found to be higher in saliva than in unselected sputum 26, 44. ECP levels have been found to be higher in selected portions of induced sputum than in the entire expectorate 96, 97. The repeatability (in repeated sputum samples from the same subject) of measurements of albumin, fibrinogen, IL‐8 and ECP has been found to be good in unselected/entire expectorates from asthmatics 3. The repeatability in selected sputum of measurement of ECP, major basic protein, albumin and fibrinogen has been also good (intraclass correlation coefficient (ICC) of >0.8) but not as good for IL‐5 (ICC 0.69) and tryptase (ICC 0.6) 2. Further studies, comparing the repeatability of mediator measurement in samples processed by both methods, need to be conducted.
Dispersal method
Because of its viscid nature, sputum requires dispersal in order to permit extraction of mediators into the fluid phase. Furthermore, solubilisation of mucus enables more complete removal of cells and debris which, if left in the supernatant, continue to release mediators. Inadequate dispersal results from suspending the sputum in saline and agitating and rocking the sputum 9, which results in lower cell counts and ECP 98, IL‐8 83 and MPO (J. Shute, unpublished data) levels. This suggests binding of these mediators to negatively charged mucins that are sensitive to sulphydryl reducing reagents such as DTT. Although DTT achieves good dispersal of most sputa, it may reduce disulphide bonds present in several mediators 99, such as IL‐1, ‐7, ‐10 and ‐12, chemokines (monocyte chemotactic proteins and eotaxin), granulocytemacrophage colonystimulating factor, transforming growth factor‐β (TGF‐β), the interferon family, TNF‐α, vascular cell adhesion molecule‐1, the selectin family and α2‐macroglobulin 100–104. Theoretically, DTT may also interfere with the disulphide bonds present in the capture antibodies of immunoassays, disrupting them and resulting in decreased assay sensitivity. Dithioerythritol is the optical isomer of DTT and exhibits similar actions.
In experiments using ultrasonication to disperse sputum followed by addition of DTT (final concentration 5 mM) to aliquots of sputum, DTT has been found to have no effect on ECP or eosinophil protein X (EPX) level but reduces recovery of eosinophil peroxidase and MPO 105. In spiking studies, DTT, when added for sputum processing, does not interfere with measurement of ECP, EPX, IL‐8, tryptase or immunoglobulin A 2, 9. However, fresh DTT added to the ECP standard interferes with the radioimmunoassay, suggesting that the activity of the DTT is reduced after it has dispersed the sputum and been frozen and thawed. In spiking experiments, DTT has not been shown to affect measurement of IL‐5 2, 106, 107. Good reproducibility (ICC 0.93) of TNF‐α concentrations over a 2‐week period has been found for samples dispersed with DTT. Conversely, the ICC is only 0.69 when sputum from the same subjects is ultracentrifuged only (V. Keatings, unpublished data). The levels detected in the ultracentrifuged samples were ∼10% of those processed with DTT.
Other types of chemical dispersal using enzymes 108 have not yet been fully evaluated, and the effect of enzymes on soluble mediator levels is unknown.
Physical methods of dispersal using ultracentrifugation 8, glass homogenisation 33 or ultrasonication 105 may be used. Ultracentrifugation, with separation into a sol and gel phase, may exclude mediators that remain bound to the gel phase 109, and mediators may “stick” to glass beads. These methods have the disadvantage of cell disruption with release of intracellular mediators; samples processed in this manner are not suitable for estimation of cellular indices.
Temperature during processing
The processing temperature varies from 4°C (on ice) to room temperature and 37°C. Its effects on mediators have not yet been fully investigated, but it does not appear to affect ECP, EPX, eosinophil peroxidase or MPO levels 9, 105.
Osmolarity
Increasing osmolarity of processing fluids has been shown to increase ECP levels in sputum supernatant 110.
Loss of mediators to nonspecific binding sites on filters and tubes
The dispersed cell suspension is commonly filtered to remove clumps of mucus and debris, and a proportion of the soluble mediators may be lost to nonspecific binding to the filter or the sides of tubes. This has been found to be in the order of 15% overall for IL‐5 when selected sputum is spiked with radiolabelled IL‐5 111. High dilution of the sample during processing may increase this loss. In addition, if there is poor dispersal, this may be aggravated by filtration, with the mediator remaining bound to nonfiltered mucus. Filters and tubes should be made from material that does not stick to protein, polypropylene tubes being superior in this respect to polycarbonate or glass ones. For highly cationic proteins such as ECP, it is advisable to dilute the sample before storage and assessment with a buffer containing the cationic detergent cetyltrimethylammonium bromide (CTAB; 0.2%) (see appendix).
Centrifugation
High centrifugal forces may result in activation of cells and release of mediators; a centrifugal force of 400×g is recommended.
Storage time and temperature
Supernatants should be stored in airtight tubes at −70°C. ECP has been shown to be stable for 72 h at 4°C or 6 h at 25°C 105, and is most stable if 0.4% CTAB is added (see appendix). In addition, repeated freeze/thaw cycles should be avoided. Therefore, if several mediators are to be measured, the supernatant should be frozen in aliquots.
Use of reagents to maximise levels of specific mediators
For some mediators, special treatment has been shown to result in better recovery. For example, TGF‐β is present in latent form and requires acid activation for quantification of total TGF‐β. The addition of CTAB to sputum samples during processing extracts intracellular proteins, e.g. ECP and MPO, and provides an estimate of the total content of the sample (see appendix) 112, 113. Storage of supernatant for later measurement of ECP and MPO is optimised by addition of 0.4% CTAB to the supernatant. (The addition of CTAB is not recommended if other mediators are to be measured.) Processing sputum after addition of protease inhibitors has been shown to improve measurement of IL‐5 (see appendix) 114.
Evaluation of the measurement of a soluble mediator in sputum
Bearing in mind the factors considered above, the discussion now focuses on the evaluation of the measurement of a specific soluble mediator in sputum and the questions that should be asked (fig. 1⇓).
Is the measurement in sputum valid?
The validity of a measurement is best established by comparison to a gold standard, but there is none available for most mediators. Therefore, validity should be examined in a series of “spiking” experiments 107. This involves adding a known amount of mediator to unprocessed sputum, processing the sputum as usual and then measuring recovery by immunoassay. It is important that the commercial source of the mediator used for spiking be the same as that used as the standard in the immunoassay. Unspiked sputum is simultaneously processed and assayed so that percentage recovery can be calculated. Any factors operating to reduce the amount of spike detected also reduce the recovery of endogenous mediator, masking to some extent the low recovery of the spike. However, the masking effect is usually slight, and poor recovery should be easily detected. The amount of mediator added to sputum should be such that the final concentration in supernatant, if most of it is recovered, lies on the straight part of the standard curve, in the middle/upper range of the limit of detection of the assay. A useful control experiment consists of spiking phosphatebuffered saline (PBS) containing 1% (weight/volume) bovine serum albumin (BSA) when the sputum is spiked. This is then frozen at the same time as the sputum supernatant, thawed and assayed simultaneously with it. This controls for errors due to pipetting technique and for any effect of freezing and thawing. The sputum should be spiked over a range of concentrations. It is not known whether recovery differs between individuals due to differences in factors in sputum which interfere with detection. Therefore, sputum from a range of subjects with different clinical conditions and degrees of severity should be used in spiking experiments. Dilution studies should be performed as a complement to or in combination with spiking. If the mediator is present in high concentration, the supernatant can be directly diluted and the mediator assayed. If not, exogenous mediator can be added to the supernatant after dilution and then assayed. The dilution buffer composition should be the same as for the standard. The concentration of mediator measured should not change significantly after correcting for dilution. If the measurement increases with dilution, this suggests the presence of interfering substances, such as soluble receptors of the mediator. In addition, parallelline studies may be performed, whereby mediator is added to supernatant which is then diluted. The mediator concentration can then be plotted against absorbance, and the gradient line should be parallel to that of the standard curve if no interfering factors are present.
If there is good recovery (>80%) of spike, it can be accepted that the assay is reasonably valid, and, if endogenous mediator is not detected, it can be assumed that there is insufficient mediator present to be detected. If recovery is poor, the next step will depend on whether the mediator is easily measured or not.
What to do if validity is poor
If the mediator level is readily measured, at least in some clinical settings, and is able to provide useful information, it may be appropriate to continue using the assay, accepting that a certain percentage of mediator is being lost. However, it cannot be assumed that the same proportion is lost in all subjects and at all mediator concentrations, unless this is established by spiking.
If the mediator is not readily detectable, it cannot be assumed that the level of mediator is below the detection limit of the assay. The following steps need to be taken in order to ascertain the cause of poor recovery and attempt to minimise loss.
What is the effect of dithiothreitol?
The effect of DTT on the mediator itself and on the immunoassay antibodies should be investigated. A solution of mediator of known concentration (in PBS containing 1% BSA) is “processed” with DTT (as for sputum) and stored as usual. It is later thawed and assayed, allowing the effect of DTT on the mediator to be examined (without sputum present). DTT may have an effect on the immunoassay itself; thus DTT is added to the assay standard (to the same final concentration as in sputum) and compared to the standard diluted in the usual way with assay buffer. In addition, if an ELISA is used, the effect on the capture antibody can be checked by adding DTT to wells coated with capture antibody and incubating for an additional 15–30 min before washing 107. If DTT does not affect measurement of mediator standard, other causes of poor recovery/loss should be considered.
Is the mediator physically lost during processing?
The mediator may be physically lost during processing by 1) binding to nonspecific sites on containers and filters, or 2) remaining bound to nonfiltered mucus particles. An effective means of investigating this is to spike the sputum with radiolabelled (e.g. with iodine125) mediator and measure recovery of radioactivity. Sodium dodecyl sulphate polyacrylamide gel electrophoresis of the radiolabelled mediator or analytical chromatography (alone, incubated with DTT and processed with sputum) may be performed to determine whether degradation of the mediator has occurred 111.
Has masking or alteration of epitopes occurred?
If the mediator is not physically lost during processing, masking or alteration of its epitopes or those of the antibodies used in the assay may have occurred. If recovery of radiolabelled mediator is better than that detected by immunoassay, it can be assumed that the low measured levels in the latter are due to poor immunological recognition by the antibodies. Parallelline analysis (dilution studies) can be used to examine whether or not the detecting antibody of the immunoassay is able to detect the endogenous mediator similarly to the standard 65. The straight part of the standard curve is compared to that produced by assaying serially diluted native supernatant or supernatant with added exogenous mediator. If the lines are parallel, it suggests that the assay is responding to active components in the preparation similarly. If the lines are not parallel, the presence of a substance that interferes with recognition of mediator epitopes by the immunoassay and which is dilutable is suggested. This may be due to masking of epitopes by autoantibodies 78, 115, soluble receptors 72, 73, 116, 117 or other binding proteins such as albumin and α2‐macroglobulin 78, 86. Specific assays are available for the soluble receptors of only certain cytokines, e.g. IL‐2 and IL‐4. Theoretically, use of polyclonal capture antibodies (more epitopes recognised) or competitive assays (which compete with the interfering substance for the epitope) would result in more sensitive measurement if interfering substances are suspected. The addition of a nonionic detergent such as 2% Tween 20 may decrease nonspecific protein interactions.
Assuming that DTT is not interfering, sputum proteases may also be responsible for degrading the mediator and altering its epitopes or interfering with the antibodies in the assay. Proteases may attain high levels in sputum, proportional to the clinical severity of asthma 64. There is evidence that they may be activated by DTT, which disrupts proteoglycan bonds and releases bound protease/antiprotease complexes 118. Stockley and Bayley 21 found that the addition of elastase to sputum resulted in decreased levels of secretory leucocyte protease inhibitor and MPO by immunoassay. The addition of protease inhibitors to sputum before processing has resulted in increased rates of detection and levels of IL‐5, suggesting that proteases are interfering substances in some cases 114.
Full reporting of results
When reporting results, the methods used for validation should be provided. The procedure used to calculate the sensitivity of the assay should be stated, and estimates below the lowest (nonzero) standard concentration used are not valid. Reproducibility should be assessed and described in terms of intra and interassay variation.
Before assays of soluble mediators can be used in a clinical situation, each laboratory needs to establish the normal range for that mediator as well as the levels expected in particular clinical settings.
Key points
1) The production, preparation and dispersal of sputum can all affect the levels of soluble mediators in sputum, and each mediator should be evaluated individually. 2) Different sources of antibodies used in immunological assays for soluble mediators can produce varying results. 3) If a soluble mediator is found to be below the level of detection of a specific assay, it cannot be assumed that the true level of that mediator is very low/zero unless spiking studies show good recovery of the mediator. If spiking studies show dramatic loss of added mediator, it should be assumed that it is unmeasurable using that assay. 4) The measurement of soluble mediators in sputum has been proven a useful investigative tool in airway inflammation as long as careful attention is paid to the use of appropriate laboratory methods and proper interpretation of results. 5) Where indicated, spiking experiments should be performed to validate results. 6) Sputum soluble mediator measurements reported in studies failing to follow a rigorous methodology should be interpreted with caution.
Outstanding questions
The following research questions still need to be addressed: 1) the differences in levels of mediators (other than ECP) between selected and unselected sputum; 2) the reproducibility of measurements of most of the mediators (apart from ECP, major basic protein, albumin, fibrinogen and IL‐8); 3) the validity of measurements of most of the mediators in both selected and unselected sputum (apart from ECP, IL‐5, histamine, total IL‐8 and tryptase in selected sputum and albumin, fibrinogen, IL‐8 and ECP in unselected sputum); 4) the effect of DTT on many mediators (apart from ECP, IL‐5, histamine, total IL‐8 and tryptase); 5) the effect of delay in processing sputum; 6) the effect of temperature (room temperature versus 37°C versus 4°C) during processing (apart from ECP, IL‐8, histamine and tryptase (room temperature versus 37°C)); 7) the effects of dilution, filtration and centrifugation of the sputum sample during processing; 8) the effect of soluble receptors, autoantibodies and other binding proteins in assays of most mediators (apart from IL‐8); 9) the effect of proteases on most mediators; 10) the effect of duration of sputum induction on mediators (other than ECP and mucinlike glycoprotein) and the optimal period of induction; 11) the effect of the time between inductions; and 12) the proportion of free versus bound mediator detected in each assay.
Appendix: Notes on optimal measurement of specific mediators
Eosinophil cationic protein
If measurement of total (intracellular and released) ECP is required, the sputum sample should be processed with the addition of 0.4% (weight/volume) CTAB in PBS (0.05 M sodium phosphate, 0.9% sodium chloride, 0.05% sodium azide, 0.01 M disodium ethylenediamine tetraacetic acid (EDTA), 0.1% BSA; pH7.4±0.05) after the dispersal step. If released ECP alone is required, an equal volume of 0.4% CTAB should be added to the sputum supernatant before storing at −70°C. Further dilutions, if required, should be made with 0.2% (weight/volume) CTAB in PBS. All PBS and PBScontaining solutions should be filtered through a 0.22µM mesh.
Myeloperoxidase
The same procedure as used for ECP should be followed. MPO is affected by freeze/thaw cycles and so the number of these should be minimised.
Interleukin‐8
Processing of sputum with DTT yields higher concentrations of a number of basic proteins such as IL‐8 in the soluble phase than in sputum treated with PBS; thus comparison can only be made between samples treated in the same way.
Interleukin‐5
The addition of a combination of protease inhibitors to sputum before processing has been shown to increase the levels of IL‐5 detected by enzyme immunoassay. The reagents used and their final concentrations are 4‐(2‐aminoethyl)benzensulphonyl fluoride (2 mM), pepstatin A (1.4 µM), leupeptin (1.0 µM) and EDTA; 1.3 mM). The effect of protease inhibitors on other mediators has not been examined.
- Received April 4, 2002.
- Accepted April 16, 2002.
- © ERS Journals Ltd