A Sensitive Double-sandwich ELISA for Neutrophil Elastase Assay Results in Unconcentrated Bronchoalveolar Lavage Fluid 1 - 4
JAN OCHNIO, RAJA T. ABBOUD, ELiE M. SMYRNIS, and SHIVRAJ S. JOHAL
Introduction
T he current hypothesis of the pathogenesis of emphysema is that emphysema develops as a result of elastolytic lung injury induced by an imbalance between proteases and antiproteases in the lung (l). The two likely sources of elastase release in the lung are neutrophils and pulmonary alveolar macrophages, both of which are increased in numbers in the lungs of smokers. Current evidence suggests that the main source of elastase release in the lung is the neutrophil (2). Previous studies on bronchoalveolar lavage (BAL) indicated that there is increased elastase or elastase-like activity in smokers compared with nonsmokers (3-6). The enzyme activity was likely of both macrophage and neutrophil origin. There have been only two studies on immunologic neutrophil elastase (NE) levels in BAL in healthy subjects, however. In our previous study on the acute effect of smoking on BAL elastase (5) we demonstrated an increase in immunologic NE levels in BAL immediately after smoking, indicating that there was a release of neutrophil elastase in the bronchoalveolar lining fluid. In that study immunologic NE levelsweredetermined by a competitive inhibition method using an enzyme-linked immunosorbant assay (ELISA) (7). Another study compared NE levelsin the BAL of nonsmokers with those of smokers who had abstained from smoking (8), using an ELISA for determining NE complexed to a t protease inhibitor (9). These investigators reported similar levels of NE in the BAL of nonsmokers and smokers. In both these studies (5, 8) the range of the ELISA was such that the assays were not sensitive enough to detect NE levels in unconcentrated BAL. We have developed a double-sandwich ELISA for NE, capable of detecting NE at levels of 0.2 ng/ml, which we have applied to unconcentrated BAL. We compared NE levels in young healthy smok-
SUMMARY We developed a sensitive double-sandwich ELISA assay for neutrophil elastase (HE) uslnSil affinity-purified HE antibody. The assay was capable of detecting HE levels of 0.2 ng/ml and was used to determine HE in bronchoalveolar lavages (BAL) of 12 healthy subjects (6 nonsmokers and 6 smokers) with a mean age of about 27 yr. HE levels In the unconcentrated cell-free supernatant of BAL, subjected to high-speed centrifugation (17,000 x 9 for 30 min) to sediment subcellular debris, were similar In the smokers who abstained overnight from smoking and In the nonsmokers (24.4 ± 13.9versus 23.7 ± 12.3 ng/mg [SD] albumin). HE levels were significantly higher In lavage fluid not subjected to high-speed centrifugation, reflecting the presence of HE bound to subcellular debris that was sedlmented by high-speed centrifugation. Concentration by ultrafiltration through a MlIIlpore CX-10 filter was accompanied by loss of protein with a relatively greater loss of HE than albumin, resulting In lower HE/albumin ratios In concentrated than In unconcentrated lavage. It Is therefore recommended that HElevels be determined on unconcentrated SAL after high-speed centrifugation to sediment subcellular debris. AM REV RESPIR DIS 1991; 143:61-65
ers who abstained overnight from smoking with NE levels in nonsmokers. We also evaluated the effects of high-speed centrifugation of BAL and of concentration of BAL by ultrafiltration on NE levels. Methods ELISA Assay The assay is based on the double-sandwich ELISA method (10)in which an ELISA plate is coated with antibody, the sample applied, and the bound NE detected by alkaline phosphatase-linked antibody. Human NE was purified in our laboratory according to the technique of Baugh and Travis (11). No impurities weredetected by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. Antibody to NE was obtained by immunizing rabbits with the purified NE using a technique similar to that described previously by Feinstein and Janoff (12). The pure NE dissolved in sodium acetate buffer, pH 5.5, was neutralized with NaOH to pH 7.0 and mixed with Freund's complete adjuvant for the initial immunization (500 ug NE) and with incomplete Freund's adjuvant for booster immunization (300 J.1g NE) at intervals of 3 wk. The immune antiserum was obtained after Booster Dose 2 or 3. When tested by a double-Ouchterlony agarose gel immunodiffusion, the crude antiserum gave a single precipitin live against pure NE and against crude neutrophil lysosomal extract, indicating that it recognized only NE in the crude
neutrophil extract. The crude antiserum as wellas the affinity-purified NE antibody (prepared as described subsequently) gaveequivalent precipitin lines against pure NE, NE preincubated with nonimmune rabbit serum, and NE preincubated with normal human serum. This indicated that both crude and purified NE antibody (purified by affinity binding to inactivated NE) recognized NE whether it was free or complexed with inhibitors in rabbit or human serum. We purified NE antibody from the crude rabbit antiserum using affinity chromatography (13) with Sepharose-4B linked to NE that had been inactivated by pretreatment with phenylmethionylsulfonylfluoride (PMSF). The NE antibody was then eluted from the Sepharose-NE column using 0.1 M glycine Hel buffer, pH 2.5. The fractions eluted with glycine containing the purified NE antibody were neutralized with Tris, pooled, and dialyzed against phosphate-buffered saline
(Received in original form September 11, 1989 and in revised form July 30, 1990) I From the Respiratory Division, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia. 2 Supported by a grant from the B.C. Health Care Research Foundation. 3 Presented in part at the American Thoracic Society meeting in Las Vegas, May 1988. 4 Correspondence should be adcressed to Dr. R. Abboud, Respiratory Division, 2775 Heather Street, Vancouver, B.C., Canada, V5Z 3J5.
61
OCHNIO, ABBOUD, SMYRNIS, AND JOHAL
62
(PBS). The NE antibody was then concentrated in dialysis tubing against an osmotic gradient obtained with sodium carboxymethylcellulose (Aquacidef II; CalbiochemHoechst, La Jolla, CAl. Alkaline phosphatase was then linked to the NE antibody by incubating 1 mg of affinity-purified NE antibody and 2 mg of highly purified alkaline phosphatase with 0.2070 glutaraldehyde (14) in a total volume of 1 ml PBS for a period of 2 h. The mixture was then dialyzed against 0.05 M 'Iris, pH 8.0, and the conjugate was then diluted to a volume of 2 ml with 1% bovine serum albumin (BSA) in the same buffer containing 0.02% sodium azide. Thus the final concentration of the NE antibodyalkaline phosphatase conjugate was equivalent to 0.5 mg NE antibody/ml. The use of the same antibody from the same species for the bottom and top of the sandwich ELISA has been documented and validated in the literature (10, 15-18) and also used specifically in studies of bronchial secretions (17) and bronchoalveolar lavage (18). Initial experiments to perform the ELISA assay using the IgO fractions of the NE antiserum purified by a protein A-Sepharose column (19) were not successful because of a high background reading in the absence of NE leading to poor assay sensitivity. Thus the use of a highly purified NE antibody separated by specific affinity chromatography was necessary. For the ELISA assay 96-well microtiter plates (Immulon'" I plates; Dynatech Lab., Inc., Chantilly, VA)were coated with affinitypurified antibody in carbonate buffer, pH 9.6, using 0.1 J.1g NE antibody in 100 J.11 buffer for each well. The antibody was allowed to coat overnight at 4°C, and the plates were then washed three times with phosphatebuffered saline, pH 7.0, containing 0.1% 1\veen~-20. Nonspecific binding sites were blocked by loading the wells with 100 III of PBS-Threen 20 containing 1% BSA and incubating for 0.5 h at 37° C. The plates were then washed with PBS-Threen. In our initial experiments we prepared NE standards in I M NaCI and NE standards in PBS with 1% BSA and obtained reliable and superimposable standards curves with both. In later experiments, however, we found that when two ELISA plates were done in sequence with the same set of standards the second plate gave lower readings, suggesting that there was loss of NE from the solution when the standards were not immediately applied to the plate. To guard against nonspecific loss and binding of NE during the preparation of the set of standards weincorporated the addition of 1% nonimmune rabbit serum to the PBS-IOJo BSA used to prepare the standards; the use of serum to prevent nonspecific binding is documented in the literature (20). We determined that NE standard curves were identical when prepared with or without the addition of 1% nonimmune rabbit serum or 0.1070 normal human plasma; the NE present in human plasma (21)diluted 1,OOD-fold would have no detectable effect on the assay. We also demonstrated that the standard curves did not
differ if the NE was previously inactivated by PMSF. irrespective of the type of diluent used. We subsequently incorporated the routine addition of 1% BSA and 1% nonimmune rabbit plasma to the standards to avoid nonspecific binding and loss of NE during the preparation of the standards. A set of NE standards prepared in PBSTween containing 1% BSA and 1% nonimmune rabbit serum was applied to each plate. Zero NE concentrations were provided by the same diluting buffer (containing 1% BSA and 1% nonimmune rabbit serum) applied to wells already coated with affinity-purified NE antibody. Blanks were provided by the same buffer applied to wells not coated with NE antibody. The zero NE readings were always low, less than 10% of the reading at an NE concentration of 2 ng/ml. To achieve similar conditions during the assay and to avoid nonspecific binding and loss of NE when samples were diluted for assay, the samples to be tested were also diluted in PBS-Threen with 1% BSA and 1% nonimmune rabbit serum. For samples with low levels of NE to be tested without dilution, we added PBS containing IOtVo rabbit serum, 100/0 BSA, and 1070 1\veen-20 in a volume of 1/10 of the sample to achieve final concentrations of 1% for both BSA and nonimmune rabbit serum. Each sample of standard or pure NE was applied to three or four wells of the plate using 100 Ill/well, and the average spectrophotometric reading was used for calculating the results. The plate was then incubated for 2 h at 37° C and then washed with PBS-Threen. We then added to each wel1l00/l1PBS-1\veencontaining NE antibody linked to alkaline phosphatase at a concentration of 0.25 Ilg antibody/ml. The plate was then incubated for 2 h at 37° C and washed three times, and 100 III alkaline phosphatase substrate (diluted in diethanolamine buffer, pH 9.8) was added to each well. After incubation at 37° C for 0.5 h for the high-range assay and 2 h for the lowrange assay, the optical density (OD) at 405 nm was read with an automated microplate reader (Flow Laboratories Inc., McLean, VA). Sepharose-4B was obtained from Pharmacia Chemicals (Uppsala, Sweden). High-purity alkaline phosphatase (for conjugation), alkaline phosphatase substrate, BSA, diethanolamine, Freund's adjuvants, Tween-zu, and glutaraldehyde were all obtained from Sigma Chemical Co. (St. Louis. MO). Our initial experiments to standardize the assay indicated that the assay yielded identical standard curves for NE diluted in PBSTween with and without the addition of 1% nonimmune rabbit serum or 0.1% normal human plasma, as well as for active NE or for NE inactivated by treatment with PMSF.. Thus the assay was able to detect and quantitate free NE. NE bound to rabbit serum, and NE complexed with human a-protease inhibitor in 0.1% normal plasma. However, since we used a final concentration of 1% nonimmune rabbit serum for both our standards and samples, any free NE would be bound to the at-protease inhibitor in the rabbit serum.
Wevalidated the ELISA assay by preincubating varying amounts of pure NE with BAL and showing that the corresponding NE levels were correctly determined. The assay can be used to determine NE levels in two ranges. a high range of 2 to 8 ng/ml and a low range of 0.2 to 2 ng/ml, with more prolonged incubation times with the alkaline phosphatase substrate. The variability of the assay was evaluated by running aliquots ofthe same lavage sample on nine different days; the coefficient of variation was 7.8% of a mean value of 1.8 ng/ml.
Bronchoalveolar Lavage Studies We applied this assay to determine neutrophil elastase levels in unconcentrated bronchoalveolar lavage from six young healthy nonsmokers (mean age 26 ± 4 yr) and six healthy young smokers (mean age 28 ± 4 yr). All smokers had smoked at least IS cigarettes/day for at least 5 yr; they abstained from smoking overnight before the lavage that was done in the morning. Each subject had two bronchoalveolar lavages done in sequence using 5 X 50 ml of warm saline. one from the right lung and the other from the left lung. The lavage was done under local lidocaine anesthesia using a fiberoptic bronchoscope wedged into a segment of subsegment of the middle lobe or lingula. as described previously (5). The interval between the two lavages was about 15 min. We used plastic syringes for the lavage and plastic tubes throughout in the subsequent handling of the BAL. The five aliquots from each lavage werepooled, cellcounts were performed, and smears for differential were obtained using cytocentrifuged preparations. The cells were sedimented at 500 x g for 10 min to obtain a cell-free supernatant. In each of the first five subjects (three nonsmokers and two smokers) the supernatants from the two lavages were combined to obtain more volume of fluid, necessary for comparison of assay results with the different methods of BAL preparation (i.e., with and without high-speed centrifugation or with and without concentration). In the remaining seven subjects (three nonsmokers and four smokers) the fluid from the two lavages was handled separately to allow comparison between sequential lavages from the same individuals. The cell-free supernatant from each lavage (or from the pooled lavage from each of the first five subjects) was then split into two aliquots, with only one undergoing highspeed centrifugation (17,000 x g for 30 min at 4°C). Aliquots from each type of fluid were then placed in conical 50-ml centrifuge tubes in crushed ice and concentrated approximately lOO-fold by negative-pressure ultrafiltration through a CX-IO filter (Millipore Corp., Bedford. MA). Samples of lavage fluid, both unconcentrated and concentrated. were stored in aliquots at - 70° C and assayed for both albumin and NE within 3 wk. It was shown that storage for this period did not affect the assay results. Albumin was quantitated by radial immu-
ELISA FOR NEUTROPHIL ELASTASE IN BRONCHOALVEOLAR LAVAGE
63 TABLE 2
0.0. 1.2
/
1.0
~ ........•
EFFECT OF HIGH-SPEED CENTRIFUGATION AND CONCENTRATION BY ULTRAFILTRATION ON NEUTROPHIL ELASTASE LEVELS IN BRONCHOALVEOLAR LAVAGE"
.~
/
0.8 0.6 0.4
Lavage Unconcentrated Centrifuged, nglml nglmg albumin Uncentrifuged, ng/ml nglmg albumin Concentrated Centrifuged, nglmg albumin Uncentrifuged, nglmg albumin
0.2
2
4
6
8
10
Neutrophil Elastase (ng/mJ) Fig. 1. Neutrophil elastase ELISA, high range, obtained after incubation with the alkaline phosphatase substrate for 0.5h at ~ C. The standard curve was prepared from a stock solution of NE diluted in saline (closed circles). The usable range ofthis assay is from 2to 8 ng/ml. Serial dilutions of SAL (closed squares) with neutrophil elastase added to a final concentration of 20 nglml tit the standard curve obtained from stock NE diluted in saline.
0.0. 1.2
• Mean ± SO. Significantly less than nonsmokers, p :I: Significantly less than nonsmokers, p
t
three sets of standards on each plate, with the lowest 25 ug/ml, The four samples that had albumin concentrations lower than 25 ug/ml were assayed by double-loading the samples in the wells of the plate and by using lower standards down to 12.5 ug/ml, which were also applied by double-loading. Under these conditions the limit of assay was reduced to 12.5 ug/ml,
Results
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Neutrophil Elastase (nq/rnl) Fig. 2. Duplicate standardcurvesfor neutrophil elastase ELISA in the low range, 0 to 2 nglml, obtained after incubation with alkaline phosphatase substrate tor 2 h at ~ C. The lower limit of detection is about 0.2 nglml. Standard 1; closed squares StanClosed circles dard 2.
=
=
nodiffusion using low-level plates (LC Partigen~; Behring Diagnostics, Marburg, Germany). According to the manufacturer, these plates have a range of 25 to 440 ug/ml, and this was confirmed in our assays. We used
Smokers
1.15 ± 0.48 23.7 ± 12.3
0.69 ± 0.40t 24.4 ± 13.9
2.46 ± 1.11 43.0 ± 23.3
1.43 ± 0.75* 55.5 ± 35.8
14.0 ± 4.1 35.8 ± 28.2
15.4 ± 9.7 32.8 ± 21.1
< 0.01. < 0.05.
1.0
0.2
Nonsmokers
Standard curves for the ELISA assays are shown in figures 1 and 2. A standard curve of zero to 10 ng/ml NE obtained after 0.5 h of incubation with alkaline phosphatase substrate is shown in figure 1. The usable assay range is from 2 to 8 ng/ml; at higher concentrations of NE the curve plateaus. The figure also indicates that a curve prepared by assaying different dilutions of BAL previously supplemented with NE (final concentration of 20 ng/ml) superimposes closely on the standard curve of NE. Figure 2 shows two duplicate ELISA standard curves of zero to 2 ng/ml obtained after incubation for 2 h at 37° C with the alkaline phosphatase substrate; the limit of detection of NE is about 0.2 ng/ml. The results of BAL findings in the initiallavage of both groups of subjects are
shown in table 1. As expected, smokers had a higher number of macrophages and neutrophils in their BAL, although the percentage of neutrophils was similar in the two groups. The recovery of instilled fluid was slightly lower in smokers than in nonsmokers, and the albumin levels were significantly lower. NE levels in unconcentrated lavage subjected to high-speed centrifugation are shown in table 1; NE was detected in all subjects, both smokers and nonsmokers. Unexpectedly, NE levelsin the nonsmokers when expressedas nanograms per milliliter fluid were higher than in the smokers, but NE levelsexpressedas nanograms per milligrams albumin were similar in the two groups. The lowerconcentrations of NE (in nanograms per milliliter)in the lavage fluid of smokers may be due to a somewhat poorer recovery or mixing of the instilled saline with epithelial lining fluid in the smokers, as suggested by the somewhat lower albumin levelsin the smokers. In both smokers and nonsmokers the levels of neutrophil elastase in those lavage fluid samples that did not undergo high-speed centrifugation (table 2) were about double the levels in fluid centrifuged at high speed. High-speed cen-
TABLE 1 COMPARISON OF FLUID RECOVERY, CELL COUNTS, AND NEUTROPHIL ELASTASE LEVELS IN BRONCHOALVEOLAR LAVAGE OF NONSMOKERS AND SMOKERS"
Volume (ml)
Macrophages (10 3/m /)
Nonsmokers, n = 6
129 ± 16.9
92.6 ± 15.9
Smokers, n = 6
111 ± 26.9
440 ± 240*
Total No. of Macrophages (1CJB)
Neutrophils (%)
Neutrophils (10 3/ml)
13.0 ± 2.2
1.3 ± 0.5
1.43 ± 0.3
49.1 ± 28.7*
1.1 ± 0.6
5.0 ± 3.6*
• Mean ± SO. on unconcentrated lavage subjected to high-speed centrifugation (17,000 x g for 30 min). :I: Significantly different from nonsmokers by the unpaired t test. p < 0.01.
t Determined
Total No. of Neutrophils (10 3 )
Albumint (Jlglml)
Neutrophil Elastaset (ng/ml)
(nglmg Albumin)
189 ± 33
49.1 ± 21.0
1.15 ± 0.48
23.7 ± 12.3
531 ± 405*
30.4 ± 9.2*
0.69 ± 0.40*
24.4 ± 13.9
64
trifugation did not affect albumin levels, which were similar in both centrifuged and uncentrifuged BAL. Levels of NE expressed in nanograms per milliliter fluid and in nanograms NE per milligram albumin showed a similar proportional difference between BAL with and without high-speed centrifugation. It was shown that the sediment from the highspeed centrifugation contained neutrophil elastase, which would account for the decrease in NE levels in BAL after high-speed sedimentation. Furthermore, electron microscopy demonstrated that the sediment consisted of cellular and/or subcellular membranes, and it is likely that the sedimented neutrophil elastase represented elastase adherent to cell or subcellular membranes and was not initially in solution in the lavage. In three subjects the sediment from lavage fluid subjected to high-speed centrifugation was suspended in 1 ml PBS and frozen at - 70° C. Before assay the suspension was thawed and sonicated (3 x 15 s pulses at a 30070 setting with a microprobe sonicator) in the presence of 0.1070 Triton X-loo and was then centrifuged at 13,000 x g for 15 min. The supernatant was then assayed for NE by ELISA. NE levels in the suspension of the sediment corresponded to NE levelsof 0.49 ± 0.28 (SEM) ng/ml of the original fluid volume from which the sediment was obtained. The NE recoveredin the sediment accounted for a mean of 66 ± 10070 of the difference between uncentrifuged and centrifuged lavage. The incomplete recovery is consistent with previous experience (22), indicating difficulty in releasing NE bound to lysosomal membranes. Comparison of the sequential lavages obtained in the sevensubjects (three nonsmokers and four smokers) did not show a significant difference in the volume recovered (124 ± 29 versus 129 ± 7 ml [SD)) or in the total number of macrophages in the first and second lavages. There was a slightly increased percentage of neutrophils in the second lavage in smokers compared with the first (1.4 ± 0.5 versus 2.4 ± 0.9070; p < 0.05). NE levelsin centrifuged unconcentrated BAL were similar in the first and second lavages (0.58 ± 0.32 versus 0.44 ± 0.33 ng/ml), and there was no significant difference in albumin levels (39.6 ± 25.8 versus 34.4 ± 21.3 ug/rnl) or NE/albumin ratios (15.9 ± 8.5 versus 12.9 ± 5.0 ng/mg). NE levels expressed as nanograms per milligram albumin in concentrated BAL both for uncentrifuged and high-speed
OCHNIO, ABBOUD, SMYRNIS, AND JOHAL
centrifuged samples are also shown in table 2 and confirm the higher values found in uncentrifuged BAL. However, the concentration procedure resulted in a loss of protein similar to the previously reported protein losses with the use of membranes for concentration (23). Following the concentration procedure, the mean loss of NE was about 44070 and the mean loss of albumin was about 12070, as estimated from the total content of the two proteins (volume x concentration) in the aliquots of BAL before and after concentration. The greater loss of NE relative to albumin following the concentration procedure resulted in lower values of NE per milligram albumin in concentrated than in unconcentrated BAL, as shown in table 2. Mean levels of NE per milligram albumin in both nonsmokers and smokers combined were 24.0 ± 12.5 (SD) for centrifuged unconcentrated BAL and 14.7 ± 7.1(SD) for centrifuged concentrated BAL. For uncentrifuged BAL, NE levelsin unconcentrated lavage were 49.2 ± 28.7, and in concentrated BAL 34.3 ± 23.7. The levels of NE in concentrated lavage expressed per milligram albumin were about 60 to 70070 of those in unconcentrated BAL. Discussion
We have developed a sensitive ELISA for determining neutrophil elastase based on a double-sandwich antibody technique using affinity-purified NE antibody. We used the same affinity-purified antibody to coat the plate and bind NE and after having linked it with alkaline phosphatase to detect the bound NE. The use of the same antibody for the bottom and top of the ELISA sandwich is documented and validated in the literature and has been used in ELISA assays to measure bronchial leukocyte protease inhibitor or antileukoprotease (18, 24), as well as ceruloplasmin and az-macroglobulin in BAL (19). In our ELISA the binding of the second NE antibody linked to alkaline phosphatase was specific to NE bound to the plate. This specificity was indicated by the low readings obtained in buffer without added NE (zero level NE in figures 1 and 2), and by the absence of a reaction to lavage samples containing NE when the second NE antibody was substituted by alkaline phosphataselinked antibody to bronchial leukocyte protease inhibitor (24). The ELISA can be used in two ranges, a high range of 2 to 8 ng/ml obtained by incubating the ELISA plate with the alkaline phosphatase substrate for 0.5 h
at 37° C, and a low range of 0.2 to 2 ng/ml obtained by incubation with the substrate for 2 h at 37° C. In the initial standardization of the ELISA assay we demonstrated identical standard curves whether the NE standards were inactivated by PMSF or the NE was active and diluted in PBS-1\veen-l07o BSA with or without the addition of 1070 nonimmune rabbit serum or 0.1070 normal human plasma. Thus the assay was capable of detecting free NE or NE bound to a l protease inhibitor in rabbit serum or human plasma. In practice we incorporated the routine use of 1070 rabbit serum added to the PBS-Tween-l07o BSA used to prepare standards and dilute samples. Thus in the assay, since standards and samples were incubated with 1070 nonimmune rabbit serum, any free NE binds to the excess rabbit ai-protease inhibitor present. This assay is well suited for determining NE levelsin unconcentrated BAL. We found NE in BAL of healthy young nonsmokers in concentrations similar to those in BAL of healthy young smokers. However, heavier smokers may have higher levels of NE in BAL, since we had previously shown that NE levels in BAL can increase after heavy smoking (5). We demonstrated that concentration of BAL fluid results in a loss of NE, which is greater than the loss of albumin, resulting in lower NE/albumin ratios in concentrated BAL compared with unconcentrated BAL. Afford and coworkers (23) reported that the concentration of BAL by ultrafiltration resulted in a variable loss of different proteins and suggested that analysis of BAL should be performed on unconcentrated samples. We have also shown NE levels to be lower in BAL subjected to high-speed centrifugation, apparently due to the presence ofNE bound to cellular or subcellular membrane fragments. This cellular debris is likely to develop during the handling of the lavage fluid, and therefore NE levelsin BAL without highspeed centrifugation may not represent in vivo elastase levels in the bronchoalveolar lining fluid. Thus we recommend that studies of NE levels in BAL be performed on unconcentrated lavage fluid subjected to high-speed sedimentation of cellular debris after the initial low-speed centrifugation of cells. The values of NE in concentrated centrifuged lavage (expressed relative to albumin) are similar to those we obtained previously (5) using a competitive inhibition ELISA (7); NE levels in BAL af-
ELISA FOR NEUTROPHIL ELASTASE IN BRONCHOALVEOLAR LAVAGE
ter overnight abstinence from smoking ing NE levelsin unconcentrated BAL. We in the six smokers in this study (table 2: found similar NE levelsin healthy young 15.4 ± 4.0 ng/mg albumin [SEM]) were nonsmokers and healthy young smokers. not different from results in similarly We recommend that NE levels in BAL processed BAL from 14smokers studied be performed on unconcentrated BAL previously (16.2 ± 2 ng/mg albumin) (5). subjected to high-speed centrifugation to However, our presently determined NE sediment cellular debris. values for uncentrifuged concentrated laAcknowledgment vage, expressed relative to albumin, are about one-fifth of those obtained by The writers thank Dr. David C. Walker, U.RC. Jochum and coworkers (8) in BAL con- Pulmonary Research Laboratory, St. Paul's centrated approximately fivefold without Hospital, Vancouver, RC., for providing elechigh-speed centrifugation. These differ- tron microscopic studies of the sediment from ences may be due to subject selection or the high-speed centrifugation of BAL. handling of BAL or to technical differReferences ences in the assays. Jochum and cowork1. Mittman C. Summary of symposium on pulers (8) obtained BAL by infusing 5 x monary emphysema and proteolysis. Am Rev Respir 60-ml aliquots of saline and recovering Dis 1972; 105:430-48. the fluid by suction in a vacuum trap. 2. Janoff A. State of the art: elastases and emIt is conceivable that the suction proce- physema. Current assessment of the proteasedure could traumatize cells and result in anti protease hypothesis. Am Rev Respir Dis 1985; 132:417-33. the release of elastase from neutrophils. 3. Janoff A, Raju L, Dearing R. Levelsof elastase To compare our technique of lavageusing activity in broncho-alveolar lavage fluid of healthy a syringe to instill and aspirate the saline smokers and nonsmokers. Am Rev Respir Pis 1983; with the technique used by Jochum and 127:540-4. coworkers (8) we studied two additional 4. Niederman MS, Fritts LL, Merrill WW, et 0/. Demonstration of a free elastolytic metalloenzyme subjects, each of whom underwent two in human lung lavage fluid and its relationship to lavages, one with each technique, We alpha-antiprotease. Am Rev Respir Dis 1984; found that the NE levels in BAL obtained 129:943-7. using the lavage technique of Jochum 5. Fera T, Abboud RT,Richter A, Johal SS. Acute effect of smoking on elastaselike esterase activity and coworkers (8) were not significantly and immunologic neutrophil elastase levelsin brondifferent from the NE levels in BAL ob- choalveolar lavage fluid. Am Rev Respir Dis 1986; tained with our technique. Thus the 133:568-73. differences in the levels of NE in BAL 6. Smith SF, Guz A, Cooke NT, Burton GH, Tetbetween the two studies are unlikely to ley TD. Extracellular elastolytic activity in human lung lavage: a comparative study between smokers be due to differences in the technique of and non-smokers. Clin Sci 1985; 69:17-27. lavage. It is of interest that Jochum and 7. Senior RM, Campbell EJ, Landis JA, Cox FR, coworkers (8) also found no significant Kuhn C, Koren HS. Elastase of U-937 monocyte differences in NE levels in BAL between cells: comparison with elastases derived from human monoeytes and neutrophils and murine smokers and nonsmokers; their smokers macrophage-like cells. J Clin Invest 1982;69:384-93. had an age and smoking history similar 8. Jochum M, Pelletier A, Boudier C, Pauli G, to ours and had not smoked for 24 h be- Bieth JG. The concentration of leukocyte elastasefore smoking. However, one would ex- aI-proteinase inhibitor complex in bronchoalveopect NE levels in BAL to be higher in lar lavage fluids from healthy human subjects. Am Rev Respir Dis 1985; 132:913-4. older and heavier smokers and in smok- 9. Neumann S, Gunzer G, Hennrich N, Lang H. ers lavaged without prior abstinence from "PMN-elastase assay": enzyme immunoassay for human polymorphonuclear elastase complexed with smoking. In conclusion, we have developed a acproteinase inhibitor. J Clin Chern Clin Biochem 22:693-7. sensitiveELISA using a double-sandwich 1984; 10. VollerA, Bidwell D, Bartlett A. Enzyme-linked technique and affinity-purified NE an- immunosorbent assay. In: Rose NR, Friedman H, tibody. This assay is capable of detect- ed. Manual of clinical immunology. Washington,
65 DC: American Society for Microbiology, 1980; 339-71. . 11. BaughRT, Travis J. Human leucocyte granule elastase: rapid isolation and characterization. Biochemistry 1976; 15:836-41. 12. FeinsteinG. Janoff A. Rapid method of purification of human granulocyte cationic neutral proteases. Purification and further characterization of human granulocyte elastase. Biochim Biophys Acta 1975; 403:493-505. 13. Wilchek M, Miron T, Kohn J. Affinity chromatography. Methods Enzymol 1984; I04C:3-55. 14. Avrameas S. Coupling of enzymes to proteins with glutaraldehyde. Use of the conjugates for the detection of antigens and antibodies. Immunochemistry 1969; 6:43-52. 15. Ferrua B, Vincent C, Revillard JP, et al. A sandwich method of enzyme immunoassay. III. Assay for human I3rmicroglobulin compared with radioimmunoassay. J Immunol Methods 1980; 36: 149-58. 16. Gemmell RS. Determination of plasma total acantitrypsin by enzyme-linked immunosorbent assay. Med Lab Sci 1986; 43:135-9. 17. Kramps JA, Franken C, Dijkman JH. ELISA for quantitative measurement of low molecular weight bronchial protease inhibitor in human sputum. Am Rev Respir Dis 1984; 129:959-63. 18. Out TA, Jansen HM, van Steenwijk KP, de Nooijer MJ, van de Graaf EA, Zuijderhoudt FMJ. ELISA of ceruloplasmin and a2-macroglobulin in paired bronchoalveolar lavage fluid and serum samples. Clin Chim Acta 1987; 165:277-88. 19. Hjelm H, Hjelm K, Sjoquist J. Protein A from Staphylococcus aureus. Its isolation by affinity chromatography and its use as an immunosorbent for isolation of immunoglobulins. FEBS Lett 1972; 28:73-6. 20. Meurman O. Detection of antiviral IgM antibodies and its problems. Curr Top Microbiol Immunol 1983; 104:101-31. 21. Plow EF. Leucocyte elastase release during blood coagulation. A potential mechanism for activation of the alternative fibrinolytic pathway. J Clin Invest 1982; 69:564-72. 22. Abboud RT, Rushton J-M, Grzybowski S. Interrelationships between neutrophil elastase, serum ai-antitrypsin, lung function and chest radiography in patients with chronic airflow obstruction. Am Rev Respir Dis 1979; 120:31-40. 23. Afford SC. Stockley RA, Kramps JA, Dijkman JH, Burnett D. Concentration of bronchoalveolar lavage fluid by ultrafiltration: evidence of differential protein loss and functional inactivation of proteinase inhibitors. Anal Biochem 1985; 151:125-30. 24. Ochnio J, Abboud RT, Lam S, Johal SS, Smith CE, Johnson DA. Bronchial leucocyte proteinase inhibitor levels in bronchial washings in asthma. Chest 1988; 93:1008-13.
JAN OCHNIO, RAJA T. ABBOUD, ELiE M. SMYRNIS, and SHIVRAJ S. JOHAL
Introduction
T he current hypothesis of the pathogenesis of emphysema is that emphysema develops as a result of elastolytic lung injury induced by an imbalance between proteases and antiproteases in the lung (l). The two likely sources of elastase release in the lung are neutrophils and pulmonary alveolar macrophages, both of which are increased in numbers in the lungs of smokers. Current evidence suggests that the main source of elastase release in the lung is the neutrophil (2). Previous studies on bronchoalveolar lavage (BAL) indicated that there is increased elastase or elastase-like activity in smokers compared with nonsmokers (3-6). The enzyme activity was likely of both macrophage and neutrophil origin. There have been only two studies on immunologic neutrophil elastase (NE) levels in BAL in healthy subjects, however. In our previous study on the acute effect of smoking on BAL elastase (5) we demonstrated an increase in immunologic NE levels in BAL immediately after smoking, indicating that there was a release of neutrophil elastase in the bronchoalveolar lining fluid. In that study immunologic NE levelsweredetermined by a competitive inhibition method using an enzyme-linked immunosorbant assay (ELISA) (7). Another study compared NE levelsin the BAL of nonsmokers with those of smokers who had abstained from smoking (8), using an ELISA for determining NE complexed to a t protease inhibitor (9). These investigators reported similar levels of NE in the BAL of nonsmokers and smokers. In both these studies (5, 8) the range of the ELISA was such that the assays were not sensitive enough to detect NE levels in unconcentrated BAL. We have developed a double-sandwich ELISA for NE, capable of detecting NE at levels of 0.2 ng/ml, which we have applied to unconcentrated BAL. We compared NE levels in young healthy smok-
SUMMARY We developed a sensitive double-sandwich ELISA assay for neutrophil elastase (HE) uslnSil affinity-purified HE antibody. The assay was capable of detecting HE levels of 0.2 ng/ml and was used to determine HE in bronchoalveolar lavages (BAL) of 12 healthy subjects (6 nonsmokers and 6 smokers) with a mean age of about 27 yr. HE levels In the unconcentrated cell-free supernatant of BAL, subjected to high-speed centrifugation (17,000 x 9 for 30 min) to sediment subcellular debris, were similar In the smokers who abstained overnight from smoking and In the nonsmokers (24.4 ± 13.9versus 23.7 ± 12.3 ng/mg [SD] albumin). HE levels were significantly higher In lavage fluid not subjected to high-speed centrifugation, reflecting the presence of HE bound to subcellular debris that was sedlmented by high-speed centrifugation. Concentration by ultrafiltration through a MlIIlpore CX-10 filter was accompanied by loss of protein with a relatively greater loss of HE than albumin, resulting In lower HE/albumin ratios In concentrated than In unconcentrated lavage. It Is therefore recommended that HElevels be determined on unconcentrated SAL after high-speed centrifugation to sediment subcellular debris. AM REV RESPIR DIS 1991; 143:61-65
ers who abstained overnight from smoking with NE levels in nonsmokers. We also evaluated the effects of high-speed centrifugation of BAL and of concentration of BAL by ultrafiltration on NE levels. Methods ELISA Assay The assay is based on the double-sandwich ELISA method (10)in which an ELISA plate is coated with antibody, the sample applied, and the bound NE detected by alkaline phosphatase-linked antibody. Human NE was purified in our laboratory according to the technique of Baugh and Travis (11). No impurities weredetected by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. Antibody to NE was obtained by immunizing rabbits with the purified NE using a technique similar to that described previously by Feinstein and Janoff (12). The pure NE dissolved in sodium acetate buffer, pH 5.5, was neutralized with NaOH to pH 7.0 and mixed with Freund's complete adjuvant for the initial immunization (500 ug NE) and with incomplete Freund's adjuvant for booster immunization (300 J.1g NE) at intervals of 3 wk. The immune antiserum was obtained after Booster Dose 2 or 3. When tested by a double-Ouchterlony agarose gel immunodiffusion, the crude antiserum gave a single precipitin live against pure NE and against crude neutrophil lysosomal extract, indicating that it recognized only NE in the crude
neutrophil extract. The crude antiserum as wellas the affinity-purified NE antibody (prepared as described subsequently) gaveequivalent precipitin lines against pure NE, NE preincubated with nonimmune rabbit serum, and NE preincubated with normal human serum. This indicated that both crude and purified NE antibody (purified by affinity binding to inactivated NE) recognized NE whether it was free or complexed with inhibitors in rabbit or human serum. We purified NE antibody from the crude rabbit antiserum using affinity chromatography (13) with Sepharose-4B linked to NE that had been inactivated by pretreatment with phenylmethionylsulfonylfluoride (PMSF). The NE antibody was then eluted from the Sepharose-NE column using 0.1 M glycine Hel buffer, pH 2.5. The fractions eluted with glycine containing the purified NE antibody were neutralized with Tris, pooled, and dialyzed against phosphate-buffered saline
(Received in original form September 11, 1989 and in revised form July 30, 1990) I From the Respiratory Division, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia. 2 Supported by a grant from the B.C. Health Care Research Foundation. 3 Presented in part at the American Thoracic Society meeting in Las Vegas, May 1988. 4 Correspondence should be adcressed to Dr. R. Abboud, Respiratory Division, 2775 Heather Street, Vancouver, B.C., Canada, V5Z 3J5.
61
OCHNIO, ABBOUD, SMYRNIS, AND JOHAL
62
(PBS). The NE antibody was then concentrated in dialysis tubing against an osmotic gradient obtained with sodium carboxymethylcellulose (Aquacidef II; CalbiochemHoechst, La Jolla, CAl. Alkaline phosphatase was then linked to the NE antibody by incubating 1 mg of affinity-purified NE antibody and 2 mg of highly purified alkaline phosphatase with 0.2070 glutaraldehyde (14) in a total volume of 1 ml PBS for a period of 2 h. The mixture was then dialyzed against 0.05 M 'Iris, pH 8.0, and the conjugate was then diluted to a volume of 2 ml with 1% bovine serum albumin (BSA) in the same buffer containing 0.02% sodium azide. Thus the final concentration of the NE antibodyalkaline phosphatase conjugate was equivalent to 0.5 mg NE antibody/ml. The use of the same antibody from the same species for the bottom and top of the sandwich ELISA has been documented and validated in the literature (10, 15-18) and also used specifically in studies of bronchial secretions (17) and bronchoalveolar lavage (18). Initial experiments to perform the ELISA assay using the IgO fractions of the NE antiserum purified by a protein A-Sepharose column (19) were not successful because of a high background reading in the absence of NE leading to poor assay sensitivity. Thus the use of a highly purified NE antibody separated by specific affinity chromatography was necessary. For the ELISA assay 96-well microtiter plates (Immulon'" I plates; Dynatech Lab., Inc., Chantilly, VA)were coated with affinitypurified antibody in carbonate buffer, pH 9.6, using 0.1 J.1g NE antibody in 100 J.11 buffer for each well. The antibody was allowed to coat overnight at 4°C, and the plates were then washed three times with phosphatebuffered saline, pH 7.0, containing 0.1% 1\veen~-20. Nonspecific binding sites were blocked by loading the wells with 100 III of PBS-Threen 20 containing 1% BSA and incubating for 0.5 h at 37° C. The plates were then washed with PBS-Threen. In our initial experiments we prepared NE standards in I M NaCI and NE standards in PBS with 1% BSA and obtained reliable and superimposable standards curves with both. In later experiments, however, we found that when two ELISA plates were done in sequence with the same set of standards the second plate gave lower readings, suggesting that there was loss of NE from the solution when the standards were not immediately applied to the plate. To guard against nonspecific loss and binding of NE during the preparation of the set of standards weincorporated the addition of 1% nonimmune rabbit serum to the PBS-IOJo BSA used to prepare the standards; the use of serum to prevent nonspecific binding is documented in the literature (20). We determined that NE standard curves were identical when prepared with or without the addition of 1% nonimmune rabbit serum or 0.1070 normal human plasma; the NE present in human plasma (21)diluted 1,OOD-fold would have no detectable effect on the assay. We also demonstrated that the standard curves did not
differ if the NE was previously inactivated by PMSF. irrespective of the type of diluent used. We subsequently incorporated the routine addition of 1% BSA and 1% nonimmune rabbit plasma to the standards to avoid nonspecific binding and loss of NE during the preparation of the standards. A set of NE standards prepared in PBSTween containing 1% BSA and 1% nonimmune rabbit serum was applied to each plate. Zero NE concentrations were provided by the same diluting buffer (containing 1% BSA and 1% nonimmune rabbit serum) applied to wells already coated with affinity-purified NE antibody. Blanks were provided by the same buffer applied to wells not coated with NE antibody. The zero NE readings were always low, less than 10% of the reading at an NE concentration of 2 ng/ml. To achieve similar conditions during the assay and to avoid nonspecific binding and loss of NE when samples were diluted for assay, the samples to be tested were also diluted in PBS-Threen with 1% BSA and 1% nonimmune rabbit serum. For samples with low levels of NE to be tested without dilution, we added PBS containing IOtVo rabbit serum, 100/0 BSA, and 1070 1\veen-20 in a volume of 1/10 of the sample to achieve final concentrations of 1% for both BSA and nonimmune rabbit serum. Each sample of standard or pure NE was applied to three or four wells of the plate using 100 Ill/well, and the average spectrophotometric reading was used for calculating the results. The plate was then incubated for 2 h at 37° C and then washed with PBS-Threen. We then added to each wel1l00/l1PBS-1\veencontaining NE antibody linked to alkaline phosphatase at a concentration of 0.25 Ilg antibody/ml. The plate was then incubated for 2 h at 37° C and washed three times, and 100 III alkaline phosphatase substrate (diluted in diethanolamine buffer, pH 9.8) was added to each well. After incubation at 37° C for 0.5 h for the high-range assay and 2 h for the lowrange assay, the optical density (OD) at 405 nm was read with an automated microplate reader (Flow Laboratories Inc., McLean, VA). Sepharose-4B was obtained from Pharmacia Chemicals (Uppsala, Sweden). High-purity alkaline phosphatase (for conjugation), alkaline phosphatase substrate, BSA, diethanolamine, Freund's adjuvants, Tween-zu, and glutaraldehyde were all obtained from Sigma Chemical Co. (St. Louis. MO). Our initial experiments to standardize the assay indicated that the assay yielded identical standard curves for NE diluted in PBSTween with and without the addition of 1% nonimmune rabbit serum or 0.1% normal human plasma, as well as for active NE or for NE inactivated by treatment with PMSF.. Thus the assay was able to detect and quantitate free NE. NE bound to rabbit serum, and NE complexed with human a-protease inhibitor in 0.1% normal plasma. However, since we used a final concentration of 1% nonimmune rabbit serum for both our standards and samples, any free NE would be bound to the at-protease inhibitor in the rabbit serum.
Wevalidated the ELISA assay by preincubating varying amounts of pure NE with BAL and showing that the corresponding NE levels were correctly determined. The assay can be used to determine NE levels in two ranges. a high range of 2 to 8 ng/ml and a low range of 0.2 to 2 ng/ml, with more prolonged incubation times with the alkaline phosphatase substrate. The variability of the assay was evaluated by running aliquots ofthe same lavage sample on nine different days; the coefficient of variation was 7.8% of a mean value of 1.8 ng/ml.
Bronchoalveolar Lavage Studies We applied this assay to determine neutrophil elastase levels in unconcentrated bronchoalveolar lavage from six young healthy nonsmokers (mean age 26 ± 4 yr) and six healthy young smokers (mean age 28 ± 4 yr). All smokers had smoked at least IS cigarettes/day for at least 5 yr; they abstained from smoking overnight before the lavage that was done in the morning. Each subject had two bronchoalveolar lavages done in sequence using 5 X 50 ml of warm saline. one from the right lung and the other from the left lung. The lavage was done under local lidocaine anesthesia using a fiberoptic bronchoscope wedged into a segment of subsegment of the middle lobe or lingula. as described previously (5). The interval between the two lavages was about 15 min. We used plastic syringes for the lavage and plastic tubes throughout in the subsequent handling of the BAL. The five aliquots from each lavage werepooled, cellcounts were performed, and smears for differential were obtained using cytocentrifuged preparations. The cells were sedimented at 500 x g for 10 min to obtain a cell-free supernatant. In each of the first five subjects (three nonsmokers and two smokers) the supernatants from the two lavages were combined to obtain more volume of fluid, necessary for comparison of assay results with the different methods of BAL preparation (i.e., with and without high-speed centrifugation or with and without concentration). In the remaining seven subjects (three nonsmokers and four smokers) the fluid from the two lavages was handled separately to allow comparison between sequential lavages from the same individuals. The cell-free supernatant from each lavage (or from the pooled lavage from each of the first five subjects) was then split into two aliquots, with only one undergoing highspeed centrifugation (17,000 x g for 30 min at 4°C). Aliquots from each type of fluid were then placed in conical 50-ml centrifuge tubes in crushed ice and concentrated approximately lOO-fold by negative-pressure ultrafiltration through a CX-IO filter (Millipore Corp., Bedford. MA). Samples of lavage fluid, both unconcentrated and concentrated. were stored in aliquots at - 70° C and assayed for both albumin and NE within 3 wk. It was shown that storage for this period did not affect the assay results. Albumin was quantitated by radial immu-
ELISA FOR NEUTROPHIL ELASTASE IN BRONCHOALVEOLAR LAVAGE
63 TABLE 2
0.0. 1.2
/
1.0
~ ........•
EFFECT OF HIGH-SPEED CENTRIFUGATION AND CONCENTRATION BY ULTRAFILTRATION ON NEUTROPHIL ELASTASE LEVELS IN BRONCHOALVEOLAR LAVAGE"
.~
/
0.8 0.6 0.4
Lavage Unconcentrated Centrifuged, nglml nglmg albumin Uncentrifuged, ng/ml nglmg albumin Concentrated Centrifuged, nglmg albumin Uncentrifuged, nglmg albumin
0.2
2
4
6
8
10
Neutrophil Elastase (ng/mJ) Fig. 1. Neutrophil elastase ELISA, high range, obtained after incubation with the alkaline phosphatase substrate for 0.5h at ~ C. The standard curve was prepared from a stock solution of NE diluted in saline (closed circles). The usable range ofthis assay is from 2to 8 ng/ml. Serial dilutions of SAL (closed squares) with neutrophil elastase added to a final concentration of 20 nglml tit the standard curve obtained from stock NE diluted in saline.
0.0. 1.2
• Mean ± SO. Significantly less than nonsmokers, p :I: Significantly less than nonsmokers, p
t
three sets of standards on each plate, with the lowest 25 ug/ml, The four samples that had albumin concentrations lower than 25 ug/ml were assayed by double-loading the samples in the wells of the plate and by using lower standards down to 12.5 ug/ml, which were also applied by double-loading. Under these conditions the limit of assay was reduced to 12.5 ug/ml,
Results
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
Neutrophil Elastase (nq/rnl) Fig. 2. Duplicate standardcurvesfor neutrophil elastase ELISA in the low range, 0 to 2 nglml, obtained after incubation with alkaline phosphatase substrate tor 2 h at ~ C. The lower limit of detection is about 0.2 nglml. Standard 1; closed squares StanClosed circles dard 2.
=
=
nodiffusion using low-level plates (LC Partigen~; Behring Diagnostics, Marburg, Germany). According to the manufacturer, these plates have a range of 25 to 440 ug/ml, and this was confirmed in our assays. We used
Smokers
1.15 ± 0.48 23.7 ± 12.3
0.69 ± 0.40t 24.4 ± 13.9
2.46 ± 1.11 43.0 ± 23.3
1.43 ± 0.75* 55.5 ± 35.8
14.0 ± 4.1 35.8 ± 28.2
15.4 ± 9.7 32.8 ± 21.1
< 0.01. < 0.05.
1.0
0.2
Nonsmokers
Standard curves for the ELISA assays are shown in figures 1 and 2. A standard curve of zero to 10 ng/ml NE obtained after 0.5 h of incubation with alkaline phosphatase substrate is shown in figure 1. The usable assay range is from 2 to 8 ng/ml; at higher concentrations of NE the curve plateaus. The figure also indicates that a curve prepared by assaying different dilutions of BAL previously supplemented with NE (final concentration of 20 ng/ml) superimposes closely on the standard curve of NE. Figure 2 shows two duplicate ELISA standard curves of zero to 2 ng/ml obtained after incubation for 2 h at 37° C with the alkaline phosphatase substrate; the limit of detection of NE is about 0.2 ng/ml. The results of BAL findings in the initiallavage of both groups of subjects are
shown in table 1. As expected, smokers had a higher number of macrophages and neutrophils in their BAL, although the percentage of neutrophils was similar in the two groups. The recovery of instilled fluid was slightly lower in smokers than in nonsmokers, and the albumin levels were significantly lower. NE levels in unconcentrated lavage subjected to high-speed centrifugation are shown in table 1; NE was detected in all subjects, both smokers and nonsmokers. Unexpectedly, NE levelsin the nonsmokers when expressedas nanograms per milliliter fluid were higher than in the smokers, but NE levelsexpressedas nanograms per milligrams albumin were similar in the two groups. The lowerconcentrations of NE (in nanograms per milliliter)in the lavage fluid of smokers may be due to a somewhat poorer recovery or mixing of the instilled saline with epithelial lining fluid in the smokers, as suggested by the somewhat lower albumin levelsin the smokers. In both smokers and nonsmokers the levels of neutrophil elastase in those lavage fluid samples that did not undergo high-speed centrifugation (table 2) were about double the levels in fluid centrifuged at high speed. High-speed cen-
TABLE 1 COMPARISON OF FLUID RECOVERY, CELL COUNTS, AND NEUTROPHIL ELASTASE LEVELS IN BRONCHOALVEOLAR LAVAGE OF NONSMOKERS AND SMOKERS"
Volume (ml)
Macrophages (10 3/m /)
Nonsmokers, n = 6
129 ± 16.9
92.6 ± 15.9
Smokers, n = 6
111 ± 26.9
440 ± 240*
Total No. of Macrophages (1CJB)
Neutrophils (%)
Neutrophils (10 3/ml)
13.0 ± 2.2
1.3 ± 0.5
1.43 ± 0.3
49.1 ± 28.7*
1.1 ± 0.6
5.0 ± 3.6*
• Mean ± SO. on unconcentrated lavage subjected to high-speed centrifugation (17,000 x g for 30 min). :I: Significantly different from nonsmokers by the unpaired t test. p < 0.01.
t Determined
Total No. of Neutrophils (10 3 )
Albumint (Jlglml)
Neutrophil Elastaset (ng/ml)
(nglmg Albumin)
189 ± 33
49.1 ± 21.0
1.15 ± 0.48
23.7 ± 12.3
531 ± 405*
30.4 ± 9.2*
0.69 ± 0.40*
24.4 ± 13.9
64
trifugation did not affect albumin levels, which were similar in both centrifuged and uncentrifuged BAL. Levels of NE expressed in nanograms per milliliter fluid and in nanograms NE per milligram albumin showed a similar proportional difference between BAL with and without high-speed centrifugation. It was shown that the sediment from the highspeed centrifugation contained neutrophil elastase, which would account for the decrease in NE levels in BAL after high-speed sedimentation. Furthermore, electron microscopy demonstrated that the sediment consisted of cellular and/or subcellular membranes, and it is likely that the sedimented neutrophil elastase represented elastase adherent to cell or subcellular membranes and was not initially in solution in the lavage. In three subjects the sediment from lavage fluid subjected to high-speed centrifugation was suspended in 1 ml PBS and frozen at - 70° C. Before assay the suspension was thawed and sonicated (3 x 15 s pulses at a 30070 setting with a microprobe sonicator) in the presence of 0.1070 Triton X-loo and was then centrifuged at 13,000 x g for 15 min. The supernatant was then assayed for NE by ELISA. NE levels in the suspension of the sediment corresponded to NE levelsof 0.49 ± 0.28 (SEM) ng/ml of the original fluid volume from which the sediment was obtained. The NE recoveredin the sediment accounted for a mean of 66 ± 10070 of the difference between uncentrifuged and centrifuged lavage. The incomplete recovery is consistent with previous experience (22), indicating difficulty in releasing NE bound to lysosomal membranes. Comparison of the sequential lavages obtained in the sevensubjects (three nonsmokers and four smokers) did not show a significant difference in the volume recovered (124 ± 29 versus 129 ± 7 ml [SD)) or in the total number of macrophages in the first and second lavages. There was a slightly increased percentage of neutrophils in the second lavage in smokers compared with the first (1.4 ± 0.5 versus 2.4 ± 0.9070; p < 0.05). NE levelsin centrifuged unconcentrated BAL were similar in the first and second lavages (0.58 ± 0.32 versus 0.44 ± 0.33 ng/ml), and there was no significant difference in albumin levels (39.6 ± 25.8 versus 34.4 ± 21.3 ug/rnl) or NE/albumin ratios (15.9 ± 8.5 versus 12.9 ± 5.0 ng/mg). NE levels expressed as nanograms per milligram albumin in concentrated BAL both for uncentrifuged and high-speed
OCHNIO, ABBOUD, SMYRNIS, AND JOHAL
centrifuged samples are also shown in table 2 and confirm the higher values found in uncentrifuged BAL. However, the concentration procedure resulted in a loss of protein similar to the previously reported protein losses with the use of membranes for concentration (23). Following the concentration procedure, the mean loss of NE was about 44070 and the mean loss of albumin was about 12070, as estimated from the total content of the two proteins (volume x concentration) in the aliquots of BAL before and after concentration. The greater loss of NE relative to albumin following the concentration procedure resulted in lower values of NE per milligram albumin in concentrated than in unconcentrated BAL, as shown in table 2. Mean levels of NE per milligram albumin in both nonsmokers and smokers combined were 24.0 ± 12.5 (SD) for centrifuged unconcentrated BAL and 14.7 ± 7.1(SD) for centrifuged concentrated BAL. For uncentrifuged BAL, NE levelsin unconcentrated lavage were 49.2 ± 28.7, and in concentrated BAL 34.3 ± 23.7. The levels of NE in concentrated lavage expressed per milligram albumin were about 60 to 70070 of those in unconcentrated BAL. Discussion
We have developed a sensitive ELISA for determining neutrophil elastase based on a double-sandwich antibody technique using affinity-purified NE antibody. We used the same affinity-purified antibody to coat the plate and bind NE and after having linked it with alkaline phosphatase to detect the bound NE. The use of the same antibody for the bottom and top of the ELISA sandwich is documented and validated in the literature and has been used in ELISA assays to measure bronchial leukocyte protease inhibitor or antileukoprotease (18, 24), as well as ceruloplasmin and az-macroglobulin in BAL (19). In our ELISA the binding of the second NE antibody linked to alkaline phosphatase was specific to NE bound to the plate. This specificity was indicated by the low readings obtained in buffer without added NE (zero level NE in figures 1 and 2), and by the absence of a reaction to lavage samples containing NE when the second NE antibody was substituted by alkaline phosphataselinked antibody to bronchial leukocyte protease inhibitor (24). The ELISA can be used in two ranges, a high range of 2 to 8 ng/ml obtained by incubating the ELISA plate with the alkaline phosphatase substrate for 0.5 h
at 37° C, and a low range of 0.2 to 2 ng/ml obtained by incubation with the substrate for 2 h at 37° C. In the initial standardization of the ELISA assay we demonstrated identical standard curves whether the NE standards were inactivated by PMSF or the NE was active and diluted in PBS-1\veen-l07o BSA with or without the addition of 1070 nonimmune rabbit serum or 0.1070 normal human plasma. Thus the assay was capable of detecting free NE or NE bound to a l protease inhibitor in rabbit serum or human plasma. In practice we incorporated the routine use of 1070 rabbit serum added to the PBS-Tween-l07o BSA used to prepare standards and dilute samples. Thus in the assay, since standards and samples were incubated with 1070 nonimmune rabbit serum, any free NE binds to the excess rabbit ai-protease inhibitor present. This assay is well suited for determining NE levelsin unconcentrated BAL. We found NE in BAL of healthy young nonsmokers in concentrations similar to those in BAL of healthy young smokers. However, heavier smokers may have higher levels of NE in BAL, since we had previously shown that NE levels in BAL can increase after heavy smoking (5). We demonstrated that concentration of BAL fluid results in a loss of NE, which is greater than the loss of albumin, resulting in lower NE/albumin ratios in concentrated BAL compared with unconcentrated BAL. Afford and coworkers (23) reported that the concentration of BAL by ultrafiltration resulted in a variable loss of different proteins and suggested that analysis of BAL should be performed on unconcentrated samples. We have also shown NE levels to be lower in BAL subjected to high-speed centrifugation, apparently due to the presence ofNE bound to cellular or subcellular membrane fragments. This cellular debris is likely to develop during the handling of the lavage fluid, and therefore NE levelsin BAL without highspeed centrifugation may not represent in vivo elastase levels in the bronchoalveolar lining fluid. Thus we recommend that studies of NE levels in BAL be performed on unconcentrated lavage fluid subjected to high-speed sedimentation of cellular debris after the initial low-speed centrifugation of cells. The values of NE in concentrated centrifuged lavage (expressed relative to albumin) are similar to those we obtained previously (5) using a competitive inhibition ELISA (7); NE levels in BAL af-
ELISA FOR NEUTROPHIL ELASTASE IN BRONCHOALVEOLAR LAVAGE
ter overnight abstinence from smoking ing NE levelsin unconcentrated BAL. We in the six smokers in this study (table 2: found similar NE levelsin healthy young 15.4 ± 4.0 ng/mg albumin [SEM]) were nonsmokers and healthy young smokers. not different from results in similarly We recommend that NE levels in BAL processed BAL from 14smokers studied be performed on unconcentrated BAL previously (16.2 ± 2 ng/mg albumin) (5). subjected to high-speed centrifugation to However, our presently determined NE sediment cellular debris. values for uncentrifuged concentrated laAcknowledgment vage, expressed relative to albumin, are about one-fifth of those obtained by The writers thank Dr. David C. Walker, U.RC. Jochum and coworkers (8) in BAL con- Pulmonary Research Laboratory, St. Paul's centrated approximately fivefold without Hospital, Vancouver, RC., for providing elechigh-speed centrifugation. These differ- tron microscopic studies of the sediment from ences may be due to subject selection or the high-speed centrifugation of BAL. handling of BAL or to technical differReferences ences in the assays. Jochum and cowork1. Mittman C. Summary of symposium on pulers (8) obtained BAL by infusing 5 x monary emphysema and proteolysis. Am Rev Respir 60-ml aliquots of saline and recovering Dis 1972; 105:430-48. the fluid by suction in a vacuum trap. 2. Janoff A. State of the art: elastases and emIt is conceivable that the suction proce- physema. Current assessment of the proteasedure could traumatize cells and result in anti protease hypothesis. Am Rev Respir Dis 1985; 132:417-33. the release of elastase from neutrophils. 3. Janoff A, Raju L, Dearing R. Levelsof elastase To compare our technique of lavageusing activity in broncho-alveolar lavage fluid of healthy a syringe to instill and aspirate the saline smokers and nonsmokers. Am Rev Respir Pis 1983; with the technique used by Jochum and 127:540-4. coworkers (8) we studied two additional 4. Niederman MS, Fritts LL, Merrill WW, et 0/. Demonstration of a free elastolytic metalloenzyme subjects, each of whom underwent two in human lung lavage fluid and its relationship to lavages, one with each technique, We alpha-antiprotease. Am Rev Respir Dis 1984; found that the NE levels in BAL obtained 129:943-7. using the lavage technique of Jochum 5. Fera T, Abboud RT,Richter A, Johal SS. Acute effect of smoking on elastaselike esterase activity and coworkers (8) were not significantly and immunologic neutrophil elastase levelsin brondifferent from the NE levels in BAL ob- choalveolar lavage fluid. Am Rev Respir Dis 1986; tained with our technique. Thus the 133:568-73. differences in the levels of NE in BAL 6. Smith SF, Guz A, Cooke NT, Burton GH, Tetbetween the two studies are unlikely to ley TD. Extracellular elastolytic activity in human lung lavage: a comparative study between smokers be due to differences in the technique of and non-smokers. Clin Sci 1985; 69:17-27. lavage. It is of interest that Jochum and 7. Senior RM, Campbell EJ, Landis JA, Cox FR, coworkers (8) also found no significant Kuhn C, Koren HS. Elastase of U-937 monocyte differences in NE levels in BAL between cells: comparison with elastases derived from human monoeytes and neutrophils and murine smokers and nonsmokers; their smokers macrophage-like cells. J Clin Invest 1982;69:384-93. had an age and smoking history similar 8. Jochum M, Pelletier A, Boudier C, Pauli G, to ours and had not smoked for 24 h be- Bieth JG. The concentration of leukocyte elastasefore smoking. However, one would ex- aI-proteinase inhibitor complex in bronchoalveopect NE levels in BAL to be higher in lar lavage fluids from healthy human subjects. Am Rev Respir Dis 1985; 132:913-4. older and heavier smokers and in smok- 9. Neumann S, Gunzer G, Hennrich N, Lang H. ers lavaged without prior abstinence from "PMN-elastase assay": enzyme immunoassay for human polymorphonuclear elastase complexed with smoking. In conclusion, we have developed a acproteinase inhibitor. J Clin Chern Clin Biochem 22:693-7. sensitiveELISA using a double-sandwich 1984; 10. VollerA, Bidwell D, Bartlett A. Enzyme-linked technique and affinity-purified NE an- immunosorbent assay. In: Rose NR, Friedman H, tibody. This assay is capable of detect- ed. Manual of clinical immunology. Washington,
65 DC: American Society for Microbiology, 1980; 339-71. . 11. BaughRT, Travis J. Human leucocyte granule elastase: rapid isolation and characterization. Biochemistry 1976; 15:836-41. 12. FeinsteinG. Janoff A. Rapid method of purification of human granulocyte cationic neutral proteases. Purification and further characterization of human granulocyte elastase. Biochim Biophys Acta 1975; 403:493-505. 13. Wilchek M, Miron T, Kohn J. Affinity chromatography. Methods Enzymol 1984; I04C:3-55. 14. Avrameas S. Coupling of enzymes to proteins with glutaraldehyde. Use of the conjugates for the detection of antigens and antibodies. Immunochemistry 1969; 6:43-52. 15. Ferrua B, Vincent C, Revillard JP, et al. A sandwich method of enzyme immunoassay. III. Assay for human I3rmicroglobulin compared with radioimmunoassay. J Immunol Methods 1980; 36: 149-58. 16. Gemmell RS. Determination of plasma total acantitrypsin by enzyme-linked immunosorbent assay. Med Lab Sci 1986; 43:135-9. 17. Kramps JA, Franken C, Dijkman JH. ELISA for quantitative measurement of low molecular weight bronchial protease inhibitor in human sputum. Am Rev Respir Dis 1984; 129:959-63. 18. Out TA, Jansen HM, van Steenwijk KP, de Nooijer MJ, van de Graaf EA, Zuijderhoudt FMJ. ELISA of ceruloplasmin and a2-macroglobulin in paired bronchoalveolar lavage fluid and serum samples. Clin Chim Acta 1987; 165:277-88. 19. Hjelm H, Hjelm K, Sjoquist J. Protein A from Staphylococcus aureus. Its isolation by affinity chromatography and its use as an immunosorbent for isolation of immunoglobulins. FEBS Lett 1972; 28:73-6. 20. Meurman O. Detection of antiviral IgM antibodies and its problems. Curr Top Microbiol Immunol 1983; 104:101-31. 21. Plow EF. Leucocyte elastase release during blood coagulation. A potential mechanism for activation of the alternative fibrinolytic pathway. J Clin Invest 1982; 69:564-72. 22. Abboud RT, Rushton J-M, Grzybowski S. Interrelationships between neutrophil elastase, serum ai-antitrypsin, lung function and chest radiography in patients with chronic airflow obstruction. Am Rev Respir Dis 1979; 120:31-40. 23. Afford SC. Stockley RA, Kramps JA, Dijkman JH, Burnett D. Concentration of bronchoalveolar lavage fluid by ultrafiltration: evidence of differential protein loss and functional inactivation of proteinase inhibitors. Anal Biochem 1985; 151:125-30. 24. Ochnio J, Abboud RT, Lam S, Johal SS, Smith CE, Johnson DA. Bronchial leucocyte proteinase inhibitor levels in bronchial washings in asthma. Chest 1988; 93:1008-13.
