Haemophilus influenzae Infections in Patients with Chronic
Obstructive Pulmonary Disease Despite Specific Antibodies in Serum and Sputum 1 •2
KEES GROENEVELD,3 PAUL P. EIJK,3 LOEK VAN ALPHEN, HENK M. JANSEN, and H. C. ZANEN Introduction Haemophilus influenzae is the most fre-
quently isolated pathogen from sputum of patients with chronic obstructive pulmonary disease (COPD) (1-3). It is not clear why these patients are persistently infected with this bacterial species. Since a lack of sufficient specific antibodies is a likely cause for this phenomenon, anti-H. influenzae antibodies in serum and sputum from COPD patients have been analyzed by several groups (4-10). In most of these studies (4-6, 8, 9) a single isolate of H. influenzae was used as antigen. This approach may lead to conflicting results since nontypable H. influenzae shows extensive antigenic variation in the composition of its surfacelocated major outer membrane proteins (MOMPs) (11-13) and lipopolysaccharide (14). In two studies the anti-H. influenzae antibody titers were determined for a small number of patients with the isolated strain as antigen: Reichek and coworkers (7) reported increasing as well as decreasing total antibody levels in patients with chronic bronchitis after an exacerbation caused by H. influenzae, and Musher and colleagues (10) found an increase in opsonizing antibody to the infecting H. influenzae in chronic bronchitis patients during acute febrile tracheobronchitis. In this study, wereport (1) the amounts of IgG, IgA, and IgM antibodies in serum and sputum from longitudinally followed infected COPD patients as measured by enzyme-linked immunosorbent assay (ELISA) with the cell envelope of the isolated strain as antigen, (2) a comparison of these antibodies with those of healthy controls measured to the cell envelope of five H. influenzae strains with various MOMP compositions, (3) the major outer membrane protein and lipopolysaccharide specificity of the serum and sputum antibodies as analyzed with immunoblotting, and (4) the influence of appearance of a "new" H. in1316
SUMMARY The titer and specificity of antibodies to the Infecting Haemophilus influenzae was determined in sera and sputa from 27 patients with chronic obstructive pulmonary disease (COPD) to analyze the specific immune response. COPD patients had significantly higher serum IgG and IgA antibody titers than 13 healthy controls (mean IgG titers 12,302 and 5,623, respectively; mean IgA titers, 2,398 and 912; P < 0.001). The mean IgM titers were comparable: 501 and 447, respectively. Specific IgA antibodies were also detectable in the sputum of the COPD patients (mean IgA antibody titer, 776). The local antibody production was determined by calculating the relative coefficient of excretion (RCE) to albumin. The mean RCE of 89.1 for IgA indicated statistically significant local production (p < 0.02), in contrast to a nonsignificant increase for IgG (mean RCE of 3.6). Specific IgM was below the detection level. Immunoblotting experiments showed that the antibodies in sera from COPD patients and controls were directed against most of the outer membrane proteins of H. influenzae, with individual differences between IgG, IgA, and IgM. The IgA and IgG antibodies in serum had a similar specificity as those in sputum. The appearance or persistence of H. influen· zae coincided with minor changes in antibody titer and specificity. From these results we conclude that COPD patients are infected with H. influenzae despite the presence of at least as many antibodies in sputum and serum as in controls and that these antibodies are directed against a variety of antigenic determinants of the infecting strain. AM REV RESPIR DIS 1990; 141:1316-1321
fluenzae strain or the persistence of H. influenzae on the antibody titer and specificity. Methods Study Populations A group of 27 COPD patients were selected according to the criteria of the American Thoracic Society (15). They all suffered from repetitiveinfections by nontypable H influenzae. Patients treated with corticosteroids and patients with malignancies wereexcluded. The mean age of the COPD patients was 57 yr (range 24 to 84 yr). Twenty-two patients were nonsmokers; the other five patients smoked fewer than five cigarettes per day. The results are presented for all 27 patients, since differences between smokers and nonsmokers were not detected. The COPD patients infected by lI. influenzae were compared with a group of healthy controls (n = 13, mean age 35 yr, range 1 to 59 yr). Sputum was collected in sterile containers and stored at 4 ° C. After taking a sample for culturing, the sputum was centrifuged for 90 min at 50,000 x g within 4 h after collecting. The sol phase was isolated and stored at - 70° C. Serum was stored at - 20° C.
Isolation and (Sub)Typing of H. influenzae Part of the sputum was washed twice in phos-
phate-buffered saline (PBS), pH 7.2 (16),and cultured on chocolate agar plates. Haemophi/us-like colonies wereisolated from the primary culture plate and examined for NAD and hemin requirement (17). The H influenzae strains were stored at - 70° C in nutrient broth supplemented with 15070 (vol/vol) glycerol. Serotyping of the capsular polysaccharides was performed by coagglutination with formalin-fixed cells of Staphylococcus aureus carrying protein A (strain Cowan I, NTCC 8530) coupled with antisera specific for the polysaccharide serotypes a through f as described previously (18). Subtyping was performed by analysis of the major outer membrane proteins with sodium dodecyl sul-
(Received in origina/form February /5, 1989 and in revised form August 21, 1989) I From the Department of Medical Microbiology, University of Amsterdam; and the Pulmonary Division, Academic Medical Centre, Amsterdam, The Netherlands. 2 Correspondence and requests for reprints should be addressed to Dr. Loek van Alphen, Department of Medical Microbiology, Room L162, University of Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands. 3 Supported by Grant 83.29 from the Netherlands Asthma Foundation.
1317
ANTIBODIES TO H. INFWENZAE IN COPD PATIENTS
fate-polyacrylamide gel electrophoresis (SDS-PAGE) as described by Lugtenberg and coworkers (19), modified as described previously (20).
Enzyme-linked Immunosorbent Assay The titers of antibodies to H. influenzae were determined by ELISA with the bacterial cell envelope as antigen. Cell envelopes were isolated as described previously (21), and the amount of protein in these preparations was determined by the method of Lowry and associates (22).Microtiter plates (Greiner, Frickenhausen B. Niirtingen, FRG) were coated overnight at room temperature with a suspension of bacterial cell envelope in PBS corresponding with 1 ug protein. The wells were washed three times with PBS + 0.05010 (vol/vol) Tween" 20 (PBS-Tw; Merck, Darmstadt, FRG), which was used as washingbuffer during the entire procedure. The plates were incubated with 150 ~I of 3% (wt/vol) bovine serum albumin (BSA) (A-7030; Sigma, St. Louis, MO) in PBS for 1 h at 37° C to saturate the wells. After washing, duplicate wells were incubated with 50 ~I of twofold dilutions of serum or sputum in PBS-Tw + 0.5% (wt/vol) BSA (PBS-Tw-BSA) for 2 hat 37° C. The starting dilutions were 11500 for serum IgG and 1150 for serum 19A and IgM and sputum antibodies. After removalof unbound antibodies by triple washings, the wells were incubated for 1 h at 37° C with 100 ~I horseradish peroxidase-conjugated rabbit antihuman IgG, IgA, or IgM antiserum (Tago, Burlingame, CAl diluted 113000 in PBS-TwBSA. After washing five times, the substrate solution was added, consisting of 0.04% (wt/vol) orthophenylenediamine dissolved in 24.3 mM citric acid and 51.4 mM NaH 2P04 (pH 5.0). H 20 2 (40 ~I, 30%) was added just before use, and the plates wereincubated with 100 ~I substrate solution per well for 20 min at 37° C. The reaction was stopped by adding 30 ~I of 50% (vol/vol) H 2S04 , and the optical density of the microtiter plates was read at 492 nm with an ELISA reader (Titertek Multiskan; Flow Laboratories, McLean, VA). A titration curve was obtained by plotting the optical density as a function of the logarithmic of the reciprocal serum dilution. The titer of a serum or sputum sample was defined as the dilution that corresponded with an optical density of 0.2 above the background as extrapolated from the slope of the curve. One batch of a pool of human serum was used as standard. Determination of the Relative Coefficient of Excretion (RCE) of Immunoglobulins To calculate the amount of local production of specific IgG and IgA antibodies, the RCE to albumin was calculated, since albumin is considered a marker for diffusion of serum proteins into the local secretions (23). The following formula was used: RCE
~
(titer specific Ig/albumin concentration) sputum (titer specific Ig/albumin concentration) serum
The concentration of albumin in sputum was measured by an immunoturbidimetric assay with a Cobas Bio Analyzer. Antialbumin was obtained from Dako (Glostrup, Denmark, code A-OOl), and a standard serum from Calbiochem (Behring CA, USA) was used for reference. The concentration of albumin in serum was determined with bromcresol green, a dye specific for albumin (24).
Immunoblotting Bacterial cell envelope components separated by SDS-PAGE weretransported to nitrocellulose filters (Schleicher & Schuell, Dassel, FRG) according to the method of Towbin and colleagues (25) in 10mM NaHC03 and 3 mM Na 2C03 , pH 9.9, containing 20% (vol/vol) methanol (26). The blots were washed for 15 min in washing buffer, consisting of 10 mM TRIS-HCI, 500mM NaCl, and 0.1% (vol/vol) Tween 20 (pH 7.4), rinsed for 1 min in running tap water and 1 min in washing buffer. The transportation of proteins to the nitrocellulose filter was checked by staining the blots with a solution of 0.1% (wt/vol) amidoblack in 45% (vol/vol) methanol and 10% (vol/vol) acetic acid. The blots werethen incubated with 500 ~I serum or sputum diluted in washing buffer for 1 h at room temperature. For IgG antibodies, the serum was diluted 1/500, and for IgA and IgM antibodies, 1150. Sputum wasdiluted 1150. Control blots wereincubated with washing buffer and subsequently treated as blots incubated with serum or sputum. After washing for 15 min in washing buffer followed by rinsing with running tap water and washing buffer, the blots were incubated with 500 ~I horseradish peroxidase-conjugated rabbit antihuman IgG, IgA, or IgM antiserum (Tago) diluted 1/3000 in washing buffer for 1 h at room temperature. Unbound conjugate was removed by 15 min washing in washing buffer and rinsing for 1 min in running tap water. Subsequently, the blots were incubated for 10 min at room temperature with 500ul of the substrate solution, consisting of 0.8% (wt/vol) dioctylsulfosuccinate and 0.24% (wt/vol) tetramethylbenzidine dissolved in ethanol added to 0.49 mM citric acid and 1.03mM NaH 2P04 (pH 5.0). Just before use, 20 ~I H 20 2 (30%) was added. The reaction was stopped by rinsing the blots in distilled water. Statistical Analysis The geometric mean of antibody titers was calculated from logarithms of the reciprocal values. The antibody titers in the sera from the infected COPD patients and healthy control subjects were compared using the Wilcoxon test for two samples. Antibody titers in several sputum or serum samples obtained from the same patient were compared using the Wilcoxon test for paired differences. To determine the significanceof the local production of IgA and IgG antibodies, Student's t test was used.
Results
Amount of Antibodies to H. influenzae We determined the antibody titer in serum and sputum of 27 COPO patients infected by H. influenzae against the cell envelope of the isolated strain. The sera of the healthy control subjects (n = 13) were tested to cell envelopes of five H. influenzae strains with strong differences in their MOMP composition. In the control sera the IgO antibody titers to the five different H. influenzae strains were highly comparable (mean variation 2.6%), indicating strong cross-reactivity between the strains. The variation in IgA and IgM antibody titers in each serum wasgreater (mean variation 9.5 and 6.6010, respectively), indicating more strain-tostrain variation in IgA and IgM antibody specificity (data not shown). To compare the data from the control group with data from infected COPO patients, the geometric mean of the five values for the IgO, IgA, and IgM antibody titers was used. The results are shown in figure 1. All COPO patients had specific antibodies with IgO titers above 1000 and IgA and IgM titers above 100. Statistical analysis of the comparison between infected patients and healthy controls (table 1)revealed that COPO patients infected by H. influenzae had significantly increased serum IgO and IgA antibody titers (p < 0.001). These differences remain significant even if for each control the highest reference value was used (p < 0.001). The IgM antibody titers in sera from infected COPO patients and healthy controls did not differ significantly (figure 1). Specific antibodies were also demonstrated in sputum of the COPO patients. The mean sputum IgA antibody titer (776) was higher than the sputum IgO antibody titer (191, figure 1).Specific sputum IgM antibody titers were below the detection level. To determine the proportion of IgO and IgA antibodies in sputum that was produced locally, the relative coefficient of excretion to albumin was determined for each patient (figure 1). The mean RCE for IgA was calculated as 89.1, indicating statistically significant local production of IgA antibodies (p < 0.02). The mean RCE for IgO was 3.6, indicating that the group of COPO patients did not show statistically significant local production of IgG. Only one patient had a RCE for IgO indicative of local production, since this was the only value above the statistical spread from 1.
GROENEVELD, EIJK, VAN ALPHEN, JANSEN, AND ZANEN
1318
IgA Serum IgA II I
Serum IgG I II
Serum IgM I II
10 3.
10'
104 oo:I t 1 '
104
I
"'I··:·· 1041 '~t I
I
3 10
1
I
x
•••
i
•
10'1
10
I
Tt ..••t. fie 103
r IIi
10
,',
10'1 ,
Sputum IgA
Sputum IgG
I
I
•,
•
••
..
••:v t • •••• • •
••
• • •
TITER
~~
,
2
,I,
00.
4
I
00
,',• I, ,It. It
103
10
•• : •
111£11
10'
IgG
•
tt I·'
•
• ••• ••• •••
-t • •
••• +
•••
Fig. 1. Serum and sputum antibody titers from COPD patients and healthy controls measured by ELISA. (A) Serum IgG, IgA, and IgM and sputum IgA and IgG antibody titers in 27 COPD patients infected by H. influenzae were determined against cell envelopes of their own isolate (I). Antibody titers in 13 healthy controls (II) were measured to the cell envelopes of five H. influenzae strains with different MOMP compositions . For each control the mean titer is indicated. The index of dispersity is indicated by the standard error of the mean. (B) RCE values for IgA and IgG antibodies in infected COPD patients.
Specificity oj Antibodies to H. influenzae The specificity of the antibodies to H. influenzae was analyzed by immunoblotting; five COPD patients and five healthy controls were selected for these analyses. The specificity of the antibodies in the patients was investigated with the isolated strain as antigen; the antibodies in the healthy controls were tested to the same five representative H. influenzae strains as used in the ELISA experiments. The sera from the five patients with H. influenzae infections almost all contained antibodies to the MOMPs a, b.c, d, and e, as wellas to high-molecular-weight proteins and lipopolysaccharide (LPS) of the infecting H. influenzae strain (table 2),
indicating that COPD patients are able to synthesize specific antibodies to a variety of antigens. The specificity of serum IgG, IgA, and IgM antibodies
showed quantitative differences, as illustrated for two patients in figure 2A and B. The antibodies in the sera of the healthy controls showed a high degree of cross-reactivity with the antigens of the five different strains, as illustrated in figure 2E. The specificity of the serum IgG, IgA, and IgM antibodies in healthy controls were comparable to the specificity of the serum antibodies in the COPD patients (table 2). The specificity of the serum antibodies to the cell envelope of one H . influenzae strain is shown for one healthy control in figure 2C. In sputum, IgA and IgG antibodies directed to the MOMPs of H . influenzae as well as to high-molecular-weight proteins and LPS were found in the majority of the COPD patients (figure 2 and table 2).
Injluence of Appearance oj a New H. influenzae Compared to Persistence of the Same Strain on the Antibody Titer and Specificity We analyzed whether the infection by a "new" H. influenzae strain, that is, reinfection by a H. influenzae strain with a MOMP composition different from the first isolate, resulted in a change in antibody titer or specificity. Antibodies in sputum from which aH. injluenzaestrain was first isolated and a corresponding serum sample werecompared with antibodies in a sputum and serum sample obtained before that strain was found. The results, summarized in figure 3 for seven COPD patients, show that the appearance of a new H. influenzae strain did not result in a statistically significant change in antibody titer in serum or in sputum (p > 0.05). In none of the infected patients was a difference in the specificity of antibodies observed after the appearance of a new strain. The persistence of H. influenzae in eight COPD patients coincided with maintenance of a high antibody titer (figure 4) with unchanged specificity in sputum and serum , indicat-
TABLE 1 COMPARISON BETWEEN SERUM ANTIBODY TITERS IN COPD PATIENTS INFECTED BY H. INFLUENZAE (n = 27) AND IN HEALTHY CONTROLS (n = 13)' IgG Mean COPD patients Healthy controls
12,302 5,623
IgA Range
(1000-252,000) (2950-14,125) P < 0.001
Mean
2398 912
IgM Range
(700-22,900) (218-2140) P < 0.001
Mean
501 447
Range
(100-1800) (218- 794) NS
• The sera of healthy controls were tested to five H. influenzae stra ins; for each control the mean antibody titer was determined. For both groups the antibody lilers were calculated as the mean after logarithmic conversion . Data were compared using the Wilcoxon signed-rank test.
1319
ANTIBODIES TO H. INFLUENZAE IN COPD PATIENTS
TABLE 2 COMPARISON BETWEEN SPECIFICITY OF SERUM AND SPUTUM ANTIBODIES IN COPD PATIENTS (P) AND HEALTHY CONTROLS (C)Serum IgGt Outer Membrane Protein High-Molecular-Weight Proteins (MW ± 110,000) a b,c d e LPS
Serum IgAt
Serum IgMt
Sputum IgAt
Sputum IgGt
P
C
P
C
P
C
P
P
5 5 5
5 5 5
5 5 5
5 5 5
3
5 5
5 5
5
4 4 4
4 4
5
3 5
5 5
4
5
4:1:
4
4
4
4
5 5
5 5
5 5
5 5
5 5
5
4 4
- Specificity of the antibodies in five infected COPD patients was analyzed with the cell envelope of the isolated strain as antigen. The specificity of the antibodies in five healthy controls was analyzed to the cell envelope of five H. fnfluenzae strains with differences in MOMP composition. t Number of patients with antibodies against the H. influenzae proteins. :t: In one H. inf/uenzae strain protein d was lacking.
az
c
B
A 1 23 45
az
c
1 2345
o
E
1 23
a
Ips Fig. 2. Specificity of serum and sputum antibodies from COPD patients and healthy controls as analyzed by immunoblolling. The specificity of the IgG, IgA, and IgM antibodies in two infected COPD patients is determined against cell envelopes of the isolated strain (A, B). The specificity of IgG, IgA, and IgM antibodies in one healthy control is determined to the cell envelope of one H. influenzae strain (C). The cross-reactivity of serum IgG antibodies in one healthy control (D) is determined against the cell envelopes of five H. influenzae strains with differences in MOMP composition as shown in amidoblack-stained immunoblot (E). (1) serum IgG, (2) serum IgA, (3) serum IgM, (4) sputum IgG, (5) sputum IgA, (az) amidoblack, (c) control.
Serum IgG
I
Fig. 3. Effect of appearance of H. influenzae on antibody titers in serum and sputum from COPD patients. Sputum from which a strain was firstly isolated and a corresponding serum sample (11) were compared with preceding sputum and serum (I).
II
Serum IgA
I
II
Serum IgM
I
II
Sputum IgA
I
II
Sputum IgG
I
II
a""", 'e e/
~7
a---.
:~ g~
ing the continuous presence of polyclonal antibodies. Discussion
In this study, the titer and specificity of antibodies in serum and sputum from COPD patients were analyzed with the
cell envelope of the patient's own H. influenzae strain as antigen to investigate if infection correlated with specific antibody patterns. The results can be summarized as follows. 1. COPD patients infected by H. influenzae have significantly higher serum
IgG and IgA antibody titers than healthy controls. The serum IgM antibody titers in infected COPD patients and healthy controls were comparable (table 1 and figure 1). 2. The RCE for IgA was indicative of strong local production of specific IgA antibodies in the lung of COPD patients. Local production of specific IgG antibodies was demonstrated for only one COPD patient (figure 1). 3. The specificity of the antibodies in infected COPD patients was comparable to the specificity in healthy controls (table 2 and figure 2). 4. Infection with a new H. influenzae strain or persistence of H. influenzae for several months did not result in a significant change in antibody titer or specificity (figures 3 and 4). We found serum IgG and IgA antibody titers in infected COPD patients to be increased compared to the titers in healthy controls (table 1 and figure 1). Interestingly, two COPD patients who had not been infected by H. influenzae for the last 6 months before the serum wasobtained had intermediate serum and sputum antibody titers compared to the infected patients and healthy controls (data not shown). Our results extend data from Burns (5) and Gump and associates (8), who found an increase in total serum antibody titers to H. influenzae between infected patients with chronic bronchitis and healthy controls, although they used one strain as antigen. We have in addition shown minor differences in the specificity of these antibodies between infected COPD patients and controls (figure 2). A problem in the analysis of antibodies in people without H. influenzae infections is that there is no homologous strain. Therefore, we used the cell envelopes of five H. influenzae strains with strong differences in MOMP composition as antigen. Although we cannot exclude that these five strains are not representative, the strong cross-reactivity observed between the different strains (figure 2) indicated that the selected strains were representative for nontypable H. influenzae. The high specificsputum IgA antibody titers found in the COPD patients (figure 1)are in agreement with data presented by Gump and associates (8), who used a reference strain as the source of antigen. To calculate the amount of local production of sputum IgA and IgG antibodies, we determined the RCE of the antibodies relative to albumin according to the method described by Delacroix and
1320
GROENEVELD, EIJK, VAN ALPHEN, JANSEN, AND ZANEN
Serum IgG TITER
104
•
a••- - -•
~::
o
.
•
1
3
5
....
•
7
9
11
TIME (MONTHS)
Serum IgA 4
10;~
•
Sputum IgA
TITER
105
(
TITER
Sputum IgG
4
10
Serum IgM 10
3
c.----
~
•
b'--------
a.
•
4 TIME (MONTHS)
liME (MONTHS)
Fig. 4. Effect of persistence of H. influenzae on antibody titers in sputum and serum from COPO patients. The interval between the subsequent sputa and sera is indicated in months.
colleagues (23), assuming that the diffusion of the immunoglobulins from the circulation into the local secretions was comparable to the diffusion of albumin (27). For IgA antibodies, the RCE was indicative for 89-fold local production over diffusion. This is in agreement with the high amount of IgA plasma cells found in the lungs of patients with chronic bronchitis (28). We could not demonstrate a significant change in antibody titer due to infection with a new H. influenzae strain (figure 3). In contrast, Reichek and coworkers (7) and Musher and associates (10)reported a significant increase in serum antibody titer in COPD patients with H. influenzae infections, but in these studies the patients had remarkably low preinfection antibody titers compared to the controls, suggesting that they were not infected by H. influenzae previously. The persistence of H. influenzae for several months did not cause a significant change in antibody titer to the infecting strain (figure 4). These results confirmed data obtained with a reference strain by Glynn (4) and Gump and col-
leagues (8). In addition, we showed that the specificity of serum and sputum antibodies was not influenced by the appearance or persistence of H. influenzae. Summarizing, we conclude that it is very likely that COPD patients are perfectly able to evoke an antibody response to the major antigens of the infecting H. influenzae. We can only speculate why these patients remain infected by H. influenzae for longer periods or become infected with a new H. influenzae strain despite high titers of specific antibodies. Wedo not know which epitopes on nontypable H. influenzae elicit antibodies necessary for protection. It might therefore be that in the background of the high antibody titers to the infecting H. influenzae strain deficiencies in antibodies to these epitopes are missed. Second, it may be that the antibody-dependent effector mechanisms, such as opsonophagocytosis or complement-dependent killing of the bacteria, are defective. Finally, it would be of interest to determine the proportion of IgA. and IgA z in the sputum antibodies, since H. influenzae is able to produce IgA. proteases, enzymes
that cleave IgA. and thereby destroy the functionality of these molecules (29, 30). Apart from the possible defectivedefense mechanisms of the host, we have recently shown that the MOMP composition of H. influenzae changes during persistence in COPD patients (31). If these changes prove to be of immunologic relevance, this might be a mechanism that enables the bacterium to persist. Acknowledgment The writers thank F. Garnier, M.D., J. Hemgreen, M.D., A. Leentvaar, M.D., and the physicians from the Pulmonary Department of the Academic Medical Centre for their kind cooperation in collecting the sputa and sera, and the Laboratory of Medical Microbiology of the Municipal Health Council for culturing bacterial strains. Dr. T. A. Out is gratefully acknowledged for critical comments. References May JR. In: Taverner D, Trounce J, eds. The chemotherapy ofchronic bronchitis and allied disorders. London: English Universities Press, 1972; 11-4. 2. Murphy TF, Apicella MA. Nontypable Haemophilus influenzae: A review of clinical aspects, I.
1321
ANTIBODIES TO H. INFLUENZAE IN COPD PATIENTS
surface antigens, and the human immune response to infection. Rev Infect Dis 1987; 9:1-15. 3. Groeneveld K, van Alphen L, Geelen-van den Broek NJ, Eijk PP, Zanen HC, van Ketel RJ. Detection of Haemophilus injluenzae with monoclonal antibody. Lancet 1987; 1:441-2. 4. Glynn AA. Antibodies to Haemophilus influenzaein chronic bronchitis. Br Med J (Clin Res) 1959; 2:911-4. 5. Burns MW, May JR. Haemophilus influenzae precipitins in the serum of patients with chronic bronchial disorders. Lancet 1967; 1:354-8. 6. Smith CB, Golden CA, Kamer RE, Renzetti AD. Haemophilus influenzae and Haemophilus parainfluenzae in chronic obstructive pulmonary disease. Lancet 1967; 1:1253-5. 7. Reichek N, LewinEB, Rhoden DL, WeaverRR, Crutcher JC. Antibody responses to bacterial antigens during exacerbations of chronic bronchitis. Am Rev Respir Dis 1970; 101:238-44. 8. Gump DW, Christmas WA, Forsyth BR, Phillips CA, Stouch WHo Serum and secretory antibodies in patients with chronic bronchitis. Arch Intern Med 1973; 132:847-51. 9. May JR, Peto R, Tinker CM, Fletcher CM. A study of Haemophilus influenzae precipitins in the serum of working men in relation to smoking habits, bronchial infection, and airway obstruction. Am Rev Respir Dis 1973; 108:460-8. 10. Musher DM, Kubitschek KR, Crennan J, Baughn RE. Pneumonia and acute febrile tracheobronchitis due to Haemophilus influenzae. Ann Intern Med 1983; 99:444-50. II. Musher DM, Hague-Park M, Baughn RE, Wallace RJ Jr, Cowley B. Opsonizing and bactericidal effects of normal human serum on nontypable Haemophilus influenzae. Infect Immun 1983; 39:297-304. 12. Hansen MV,Musher DM, Baughn RE. Outer membrane proteins of nontypable Haemophilus influenzae and reactivity of paired sera from infect-
ed patients with their homologous isolates. Infect Immun 1985; 47:843-6. 13. Murphy TF, Apicella MA. Antigenic heterogeneity of outer membrane proteins of nontypable Haemophilus influenzae is a basis for a serotyping system. Infect Immun 1985; 50:15-21. 14. Campagnari AA, Gupta MR, Dudas KC, Murphy TF, Apicella MA. Antigenic diversity of lipooligosaccharides of nontypable Haemophilus injluenzae. Infect Immun 1987; 55:882-7. 15. American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 136:225-44. 16. Mulder J, Goslings WRO, van der Plas MC, Lopes Cardozo P. Studies on the treatment with antibacterial drugs of acute and chronic mucopurulent bronchitis caused by H. influenzae. Acta Med Scand 1952; 143:32-49. 17. Kilian M. A taxonomic study on the genus Haemophilus, with the proposal of a new species. J Gen Microbiol 1976; 93:9-62. 18. Dirks-Go SIS, Zanen He. Latex agglutination, counterimmunoelectrophoresis and protein A coagglutination in diagnosis of bacterial meningitis. J Clin Pat hoi 1978; 31:1167-71. 19. Lugtenberg B, Meijers J, Peters R, van der Hoek P, van Alphen L. Electrophoretic resolution of the "major outer membrane protein" of Escherichia coli KI2 into four bands. FEBS Lett 1975; 58:254-8. 20. Van Alphen L, Riemens T, Poolman J, Zanen HC. Characteristics of major outer membrane proteins of Haemophilus influenzae. J Bacteriol1983; 155:878-85. 21. Van Alphen L, Romijn C, Brandt H, Geelen L, Zanen HC. Preparation of cellenvelopes of large numbers of individual bacterial strains with the use of an automatic cell disruptor. Anal Biochem 1987; 166:36-40. 22. Lowry OH, Rosebrough NJ, Farr AL, Ran-
dall RJ. Protein measurement with the folin phenol reagent, J Bioi Chern 1951; 193:265-75. 23. Delacroix DL, Marchandise FX, Francis C, Sibille y. Alpha-2-macroglobulin, monomeric and polymeric immunoglobulin A, and immunoglobulin M in bronchoalveolar lavage. Am Rev Respir Dis 1985; 132:829-35. 24. Rodkey FL. Direct spectophotometric determination of albumin in human serum. Clin Chern 1965; 11:478-87. 25. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Nat! Acad Sci USA 1979; 76:4350-4. 26. Dunn SD. Effects of the modification of transfer buffer composition and the renaturation of proteins in gels on the recognition of proteins on Western blots by monoclonal antibodies. Anal Biochem 1986; 157:144-53. 27. ReynoldsHY. Bronchoalveolar lavage.Am Rev Respir Dis 1987; 135:250-63. 28. Burnett D, Crocker J, Stockley RA. Cells containing IgA subclasses in bronchi of subjects with and without chronic obstructive lung disease.J Clin Pathol 1987; 40:1217-20. 29. Kilian M, Mestecky J, Schrohenloher RE. Pathogenic species of the genus Haemophilus and Streptococcus pneumoniae produce immunoglobulin A, protease. Infect Immun 1979; 26:143-9. 30. Mansa B, Kilian M. Retained antigen-binding activity of Fab fragments of human monoclonal immunoglobulin Al (IgA,) cleaved by IgA, protease. Infect Immun 1986; 52:171-4. 31. Groeneveld K, van Alphen L, Eijk PP, Jansen HM, Zanen He. Changes in outer membrane proteins of nontypable Haemophilus influenzae in patients with chronic obstructive pulmonary disease. J Infect Dis 1988; 158:360-5.
Obstructive Pulmonary Disease Despite Specific Antibodies in Serum and Sputum 1 •2
KEES GROENEVELD,3 PAUL P. EIJK,3 LOEK VAN ALPHEN, HENK M. JANSEN, and H. C. ZANEN Introduction Haemophilus influenzae is the most fre-
quently isolated pathogen from sputum of patients with chronic obstructive pulmonary disease (COPD) (1-3). It is not clear why these patients are persistently infected with this bacterial species. Since a lack of sufficient specific antibodies is a likely cause for this phenomenon, anti-H. influenzae antibodies in serum and sputum from COPD patients have been analyzed by several groups (4-10). In most of these studies (4-6, 8, 9) a single isolate of H. influenzae was used as antigen. This approach may lead to conflicting results since nontypable H. influenzae shows extensive antigenic variation in the composition of its surfacelocated major outer membrane proteins (MOMPs) (11-13) and lipopolysaccharide (14). In two studies the anti-H. influenzae antibody titers were determined for a small number of patients with the isolated strain as antigen: Reichek and coworkers (7) reported increasing as well as decreasing total antibody levels in patients with chronic bronchitis after an exacerbation caused by H. influenzae, and Musher and colleagues (10) found an increase in opsonizing antibody to the infecting H. influenzae in chronic bronchitis patients during acute febrile tracheobronchitis. In this study, wereport (1) the amounts of IgG, IgA, and IgM antibodies in serum and sputum from longitudinally followed infected COPD patients as measured by enzyme-linked immunosorbent assay (ELISA) with the cell envelope of the isolated strain as antigen, (2) a comparison of these antibodies with those of healthy controls measured to the cell envelope of five H. influenzae strains with various MOMP compositions, (3) the major outer membrane protein and lipopolysaccharide specificity of the serum and sputum antibodies as analyzed with immunoblotting, and (4) the influence of appearance of a "new" H. in1316
SUMMARY The titer and specificity of antibodies to the Infecting Haemophilus influenzae was determined in sera and sputa from 27 patients with chronic obstructive pulmonary disease (COPD) to analyze the specific immune response. COPD patients had significantly higher serum IgG and IgA antibody titers than 13 healthy controls (mean IgG titers 12,302 and 5,623, respectively; mean IgA titers, 2,398 and 912; P < 0.001). The mean IgM titers were comparable: 501 and 447, respectively. Specific IgA antibodies were also detectable in the sputum of the COPD patients (mean IgA antibody titer, 776). The local antibody production was determined by calculating the relative coefficient of excretion (RCE) to albumin. The mean RCE of 89.1 for IgA indicated statistically significant local production (p < 0.02), in contrast to a nonsignificant increase for IgG (mean RCE of 3.6). Specific IgM was below the detection level. Immunoblotting experiments showed that the antibodies in sera from COPD patients and controls were directed against most of the outer membrane proteins of H. influenzae, with individual differences between IgG, IgA, and IgM. The IgA and IgG antibodies in serum had a similar specificity as those in sputum. The appearance or persistence of H. influen· zae coincided with minor changes in antibody titer and specificity. From these results we conclude that COPD patients are infected with H. influenzae despite the presence of at least as many antibodies in sputum and serum as in controls and that these antibodies are directed against a variety of antigenic determinants of the infecting strain. AM REV RESPIR DIS 1990; 141:1316-1321
fluenzae strain or the persistence of H. influenzae on the antibody titer and specificity. Methods Study Populations A group of 27 COPD patients were selected according to the criteria of the American Thoracic Society (15). They all suffered from repetitiveinfections by nontypable H influenzae. Patients treated with corticosteroids and patients with malignancies wereexcluded. The mean age of the COPD patients was 57 yr (range 24 to 84 yr). Twenty-two patients were nonsmokers; the other five patients smoked fewer than five cigarettes per day. The results are presented for all 27 patients, since differences between smokers and nonsmokers were not detected. The COPD patients infected by lI. influenzae were compared with a group of healthy controls (n = 13, mean age 35 yr, range 1 to 59 yr). Sputum was collected in sterile containers and stored at 4 ° C. After taking a sample for culturing, the sputum was centrifuged for 90 min at 50,000 x g within 4 h after collecting. The sol phase was isolated and stored at - 70° C. Serum was stored at - 20° C.
Isolation and (Sub)Typing of H. influenzae Part of the sputum was washed twice in phos-
phate-buffered saline (PBS), pH 7.2 (16),and cultured on chocolate agar plates. Haemophi/us-like colonies wereisolated from the primary culture plate and examined for NAD and hemin requirement (17). The H influenzae strains were stored at - 70° C in nutrient broth supplemented with 15070 (vol/vol) glycerol. Serotyping of the capsular polysaccharides was performed by coagglutination with formalin-fixed cells of Staphylococcus aureus carrying protein A (strain Cowan I, NTCC 8530) coupled with antisera specific for the polysaccharide serotypes a through f as described previously (18). Subtyping was performed by analysis of the major outer membrane proteins with sodium dodecyl sul-
(Received in origina/form February /5, 1989 and in revised form August 21, 1989) I From the Department of Medical Microbiology, University of Amsterdam; and the Pulmonary Division, Academic Medical Centre, Amsterdam, The Netherlands. 2 Correspondence and requests for reprints should be addressed to Dr. Loek van Alphen, Department of Medical Microbiology, Room L162, University of Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands. 3 Supported by Grant 83.29 from the Netherlands Asthma Foundation.
1317
ANTIBODIES TO H. INFWENZAE IN COPD PATIENTS
fate-polyacrylamide gel electrophoresis (SDS-PAGE) as described by Lugtenberg and coworkers (19), modified as described previously (20).
Enzyme-linked Immunosorbent Assay The titers of antibodies to H. influenzae were determined by ELISA with the bacterial cell envelope as antigen. Cell envelopes were isolated as described previously (21), and the amount of protein in these preparations was determined by the method of Lowry and associates (22).Microtiter plates (Greiner, Frickenhausen B. Niirtingen, FRG) were coated overnight at room temperature with a suspension of bacterial cell envelope in PBS corresponding with 1 ug protein. The wells were washed three times with PBS + 0.05010 (vol/vol) Tween" 20 (PBS-Tw; Merck, Darmstadt, FRG), which was used as washingbuffer during the entire procedure. The plates were incubated with 150 ~I of 3% (wt/vol) bovine serum albumin (BSA) (A-7030; Sigma, St. Louis, MO) in PBS for 1 h at 37° C to saturate the wells. After washing, duplicate wells were incubated with 50 ~I of twofold dilutions of serum or sputum in PBS-Tw + 0.5% (wt/vol) BSA (PBS-Tw-BSA) for 2 hat 37° C. The starting dilutions were 11500 for serum IgG and 1150 for serum 19A and IgM and sputum antibodies. After removalof unbound antibodies by triple washings, the wells were incubated for 1 h at 37° C with 100 ~I horseradish peroxidase-conjugated rabbit antihuman IgG, IgA, or IgM antiserum (Tago, Burlingame, CAl diluted 113000 in PBS-TwBSA. After washing five times, the substrate solution was added, consisting of 0.04% (wt/vol) orthophenylenediamine dissolved in 24.3 mM citric acid and 51.4 mM NaH 2P04 (pH 5.0). H 20 2 (40 ~I, 30%) was added just before use, and the plates wereincubated with 100 ~I substrate solution per well for 20 min at 37° C. The reaction was stopped by adding 30 ~I of 50% (vol/vol) H 2S04 , and the optical density of the microtiter plates was read at 492 nm with an ELISA reader (Titertek Multiskan; Flow Laboratories, McLean, VA). A titration curve was obtained by plotting the optical density as a function of the logarithmic of the reciprocal serum dilution. The titer of a serum or sputum sample was defined as the dilution that corresponded with an optical density of 0.2 above the background as extrapolated from the slope of the curve. One batch of a pool of human serum was used as standard. Determination of the Relative Coefficient of Excretion (RCE) of Immunoglobulins To calculate the amount of local production of specific IgG and IgA antibodies, the RCE to albumin was calculated, since albumin is considered a marker for diffusion of serum proteins into the local secretions (23). The following formula was used: RCE
~
(titer specific Ig/albumin concentration) sputum (titer specific Ig/albumin concentration) serum
The concentration of albumin in sputum was measured by an immunoturbidimetric assay with a Cobas Bio Analyzer. Antialbumin was obtained from Dako (Glostrup, Denmark, code A-OOl), and a standard serum from Calbiochem (Behring CA, USA) was used for reference. The concentration of albumin in serum was determined with bromcresol green, a dye specific for albumin (24).
Immunoblotting Bacterial cell envelope components separated by SDS-PAGE weretransported to nitrocellulose filters (Schleicher & Schuell, Dassel, FRG) according to the method of Towbin and colleagues (25) in 10mM NaHC03 and 3 mM Na 2C03 , pH 9.9, containing 20% (vol/vol) methanol (26). The blots were washed for 15 min in washing buffer, consisting of 10 mM TRIS-HCI, 500mM NaCl, and 0.1% (vol/vol) Tween 20 (pH 7.4), rinsed for 1 min in running tap water and 1 min in washing buffer. The transportation of proteins to the nitrocellulose filter was checked by staining the blots with a solution of 0.1% (wt/vol) amidoblack in 45% (vol/vol) methanol and 10% (vol/vol) acetic acid. The blots werethen incubated with 500 ~I serum or sputum diluted in washing buffer for 1 h at room temperature. For IgG antibodies, the serum was diluted 1/500, and for IgA and IgM antibodies, 1150. Sputum wasdiluted 1150. Control blots wereincubated with washing buffer and subsequently treated as blots incubated with serum or sputum. After washing for 15 min in washing buffer followed by rinsing with running tap water and washing buffer, the blots were incubated with 500 ~I horseradish peroxidase-conjugated rabbit antihuman IgG, IgA, or IgM antiserum (Tago) diluted 1/3000 in washing buffer for 1 h at room temperature. Unbound conjugate was removed by 15 min washing in washing buffer and rinsing for 1 min in running tap water. Subsequently, the blots were incubated for 10 min at room temperature with 500ul of the substrate solution, consisting of 0.8% (wt/vol) dioctylsulfosuccinate and 0.24% (wt/vol) tetramethylbenzidine dissolved in ethanol added to 0.49 mM citric acid and 1.03mM NaH 2P04 (pH 5.0). Just before use, 20 ~I H 20 2 (30%) was added. The reaction was stopped by rinsing the blots in distilled water. Statistical Analysis The geometric mean of antibody titers was calculated from logarithms of the reciprocal values. The antibody titers in the sera from the infected COPD patients and healthy control subjects were compared using the Wilcoxon test for two samples. Antibody titers in several sputum or serum samples obtained from the same patient were compared using the Wilcoxon test for paired differences. To determine the significanceof the local production of IgA and IgG antibodies, Student's t test was used.
Results
Amount of Antibodies to H. influenzae We determined the antibody titer in serum and sputum of 27 COPO patients infected by H. influenzae against the cell envelope of the isolated strain. The sera of the healthy control subjects (n = 13) were tested to cell envelopes of five H. influenzae strains with strong differences in their MOMP composition. In the control sera the IgO antibody titers to the five different H. influenzae strains were highly comparable (mean variation 2.6%), indicating strong cross-reactivity between the strains. The variation in IgA and IgM antibody titers in each serum wasgreater (mean variation 9.5 and 6.6010, respectively), indicating more strain-tostrain variation in IgA and IgM antibody specificity (data not shown). To compare the data from the control group with data from infected COPO patients, the geometric mean of the five values for the IgO, IgA, and IgM antibody titers was used. The results are shown in figure 1. All COPO patients had specific antibodies with IgO titers above 1000 and IgA and IgM titers above 100. Statistical analysis of the comparison between infected patients and healthy controls (table 1)revealed that COPO patients infected by H. influenzae had significantly increased serum IgO and IgA antibody titers (p < 0.001). These differences remain significant even if for each control the highest reference value was used (p < 0.001). The IgM antibody titers in sera from infected COPO patients and healthy controls did not differ significantly (figure 1). Specific antibodies were also demonstrated in sputum of the COPO patients. The mean sputum IgA antibody titer (776) was higher than the sputum IgO antibody titer (191, figure 1).Specific sputum IgM antibody titers were below the detection level. To determine the proportion of IgO and IgA antibodies in sputum that was produced locally, the relative coefficient of excretion to albumin was determined for each patient (figure 1). The mean RCE for IgA was calculated as 89.1, indicating statistically significant local production of IgA antibodies (p < 0.02). The mean RCE for IgO was 3.6, indicating that the group of COPO patients did not show statistically significant local production of IgG. Only one patient had a RCE for IgO indicative of local production, since this was the only value above the statistical spread from 1.
GROENEVELD, EIJK, VAN ALPHEN, JANSEN, AND ZANEN
1318
IgA Serum IgA II I
Serum IgG I II
Serum IgM I II
10 3.
10'
104 oo:I t 1 '
104
I
"'I··:·· 1041 '~t I
I
3 10
1
I
x
•••
i
•
10'1
10
I
Tt ..••t. fie 103
r IIi
10
,',
10'1 ,
Sputum IgA
Sputum IgG
I
I
•,
•
••
..
••:v t • •••• • •
••
• • •
TITER
~~
,
2
,I,
00.
4
I
00
,',• I, ,It. It
103
10
•• : •
111£11
10'
IgG
•
tt I·'
•
• ••• ••• •••
-t • •
••• +
•••
Fig. 1. Serum and sputum antibody titers from COPD patients and healthy controls measured by ELISA. (A) Serum IgG, IgA, and IgM and sputum IgA and IgG antibody titers in 27 COPD patients infected by H. influenzae were determined against cell envelopes of their own isolate (I). Antibody titers in 13 healthy controls (II) were measured to the cell envelopes of five H. influenzae strains with different MOMP compositions . For each control the mean titer is indicated. The index of dispersity is indicated by the standard error of the mean. (B) RCE values for IgA and IgG antibodies in infected COPD patients.
Specificity oj Antibodies to H. influenzae The specificity of the antibodies to H. influenzae was analyzed by immunoblotting; five COPD patients and five healthy controls were selected for these analyses. The specificity of the antibodies in the patients was investigated with the isolated strain as antigen; the antibodies in the healthy controls were tested to the same five representative H. influenzae strains as used in the ELISA experiments. The sera from the five patients with H. influenzae infections almost all contained antibodies to the MOMPs a, b.c, d, and e, as wellas to high-molecular-weight proteins and lipopolysaccharide (LPS) of the infecting H. influenzae strain (table 2),
indicating that COPD patients are able to synthesize specific antibodies to a variety of antigens. The specificity of serum IgG, IgA, and IgM antibodies
showed quantitative differences, as illustrated for two patients in figure 2A and B. The antibodies in the sera of the healthy controls showed a high degree of cross-reactivity with the antigens of the five different strains, as illustrated in figure 2E. The specificity of the serum IgG, IgA, and IgM antibodies in healthy controls were comparable to the specificity of the serum antibodies in the COPD patients (table 2). The specificity of the serum antibodies to the cell envelope of one H . influenzae strain is shown for one healthy control in figure 2C. In sputum, IgA and IgG antibodies directed to the MOMPs of H . influenzae as well as to high-molecular-weight proteins and LPS were found in the majority of the COPD patients (figure 2 and table 2).
Injluence of Appearance oj a New H. influenzae Compared to Persistence of the Same Strain on the Antibody Titer and Specificity We analyzed whether the infection by a "new" H. influenzae strain, that is, reinfection by a H. influenzae strain with a MOMP composition different from the first isolate, resulted in a change in antibody titer or specificity. Antibodies in sputum from which aH. injluenzaestrain was first isolated and a corresponding serum sample werecompared with antibodies in a sputum and serum sample obtained before that strain was found. The results, summarized in figure 3 for seven COPD patients, show that the appearance of a new H. influenzae strain did not result in a statistically significant change in antibody titer in serum or in sputum (p > 0.05). In none of the infected patients was a difference in the specificity of antibodies observed after the appearance of a new strain. The persistence of H. influenzae in eight COPD patients coincided with maintenance of a high antibody titer (figure 4) with unchanged specificity in sputum and serum , indicat-
TABLE 1 COMPARISON BETWEEN SERUM ANTIBODY TITERS IN COPD PATIENTS INFECTED BY H. INFLUENZAE (n = 27) AND IN HEALTHY CONTROLS (n = 13)' IgG Mean COPD patients Healthy controls
12,302 5,623
IgA Range
(1000-252,000) (2950-14,125) P < 0.001
Mean
2398 912
IgM Range
(700-22,900) (218-2140) P < 0.001
Mean
501 447
Range
(100-1800) (218- 794) NS
• The sera of healthy controls were tested to five H. influenzae stra ins; for each control the mean antibody titer was determined. For both groups the antibody lilers were calculated as the mean after logarithmic conversion . Data were compared using the Wilcoxon signed-rank test.
1319
ANTIBODIES TO H. INFLUENZAE IN COPD PATIENTS
TABLE 2 COMPARISON BETWEEN SPECIFICITY OF SERUM AND SPUTUM ANTIBODIES IN COPD PATIENTS (P) AND HEALTHY CONTROLS (C)Serum IgGt Outer Membrane Protein High-Molecular-Weight Proteins (MW ± 110,000) a b,c d e LPS
Serum IgAt
Serum IgMt
Sputum IgAt
Sputum IgGt
P
C
P
C
P
C
P
P
5 5 5
5 5 5
5 5 5
5 5 5
3
5 5
5 5
5
4 4 4
4 4
5
3 5
5 5
4
5
4:1:
4
4
4
4
5 5
5 5
5 5
5 5
5 5
5
4 4
- Specificity of the antibodies in five infected COPD patients was analyzed with the cell envelope of the isolated strain as antigen. The specificity of the antibodies in five healthy controls was analyzed to the cell envelope of five H. fnfluenzae strains with differences in MOMP composition. t Number of patients with antibodies against the H. influenzae proteins. :t: In one H. inf/uenzae strain protein d was lacking.
az
c
B
A 1 23 45
az
c
1 2345
o
E
1 23
a
Ips Fig. 2. Specificity of serum and sputum antibodies from COPD patients and healthy controls as analyzed by immunoblolling. The specificity of the IgG, IgA, and IgM antibodies in two infected COPD patients is determined against cell envelopes of the isolated strain (A, B). The specificity of IgG, IgA, and IgM antibodies in one healthy control is determined to the cell envelope of one H. influenzae strain (C). The cross-reactivity of serum IgG antibodies in one healthy control (D) is determined against the cell envelopes of five H. influenzae strains with differences in MOMP composition as shown in amidoblack-stained immunoblot (E). (1) serum IgG, (2) serum IgA, (3) serum IgM, (4) sputum IgG, (5) sputum IgA, (az) amidoblack, (c) control.
Serum IgG
I
Fig. 3. Effect of appearance of H. influenzae on antibody titers in serum and sputum from COPD patients. Sputum from which a strain was firstly isolated and a corresponding serum sample (11) were compared with preceding sputum and serum (I).
II
Serum IgA
I
II
Serum IgM
I
II
Sputum IgA
I
II
Sputum IgG
I
II
a""", 'e e/
~7
a---.
:~ g~
ing the continuous presence of polyclonal antibodies. Discussion
In this study, the titer and specificity of antibodies in serum and sputum from COPD patients were analyzed with the
cell envelope of the patient's own H. influenzae strain as antigen to investigate if infection correlated with specific antibody patterns. The results can be summarized as follows. 1. COPD patients infected by H. influenzae have significantly higher serum
IgG and IgA antibody titers than healthy controls. The serum IgM antibody titers in infected COPD patients and healthy controls were comparable (table 1 and figure 1). 2. The RCE for IgA was indicative of strong local production of specific IgA antibodies in the lung of COPD patients. Local production of specific IgG antibodies was demonstrated for only one COPD patient (figure 1). 3. The specificity of the antibodies in infected COPD patients was comparable to the specificity in healthy controls (table 2 and figure 2). 4. Infection with a new H. influenzae strain or persistence of H. influenzae for several months did not result in a significant change in antibody titer or specificity (figures 3 and 4). We found serum IgG and IgA antibody titers in infected COPD patients to be increased compared to the titers in healthy controls (table 1 and figure 1). Interestingly, two COPD patients who had not been infected by H. influenzae for the last 6 months before the serum wasobtained had intermediate serum and sputum antibody titers compared to the infected patients and healthy controls (data not shown). Our results extend data from Burns (5) and Gump and associates (8), who found an increase in total serum antibody titers to H. influenzae between infected patients with chronic bronchitis and healthy controls, although they used one strain as antigen. We have in addition shown minor differences in the specificity of these antibodies between infected COPD patients and controls (figure 2). A problem in the analysis of antibodies in people without H. influenzae infections is that there is no homologous strain. Therefore, we used the cell envelopes of five H. influenzae strains with strong differences in MOMP composition as antigen. Although we cannot exclude that these five strains are not representative, the strong cross-reactivity observed between the different strains (figure 2) indicated that the selected strains were representative for nontypable H. influenzae. The high specificsputum IgA antibody titers found in the COPD patients (figure 1)are in agreement with data presented by Gump and associates (8), who used a reference strain as the source of antigen. To calculate the amount of local production of sputum IgA and IgG antibodies, we determined the RCE of the antibodies relative to albumin according to the method described by Delacroix and
1320
GROENEVELD, EIJK, VAN ALPHEN, JANSEN, AND ZANEN
Serum IgG TITER
104
•
a••- - -•
~::
o
.
•
1
3
5
....
•
7
9
11
TIME (MONTHS)
Serum IgA 4
10;~
•
Sputum IgA
TITER
105
(
TITER
Sputum IgG
4
10
Serum IgM 10
3
c.----
~
•
b'--------
a.
•
4 TIME (MONTHS)
liME (MONTHS)
Fig. 4. Effect of persistence of H. influenzae on antibody titers in sputum and serum from COPO patients. The interval between the subsequent sputa and sera is indicated in months.
colleagues (23), assuming that the diffusion of the immunoglobulins from the circulation into the local secretions was comparable to the diffusion of albumin (27). For IgA antibodies, the RCE was indicative for 89-fold local production over diffusion. This is in agreement with the high amount of IgA plasma cells found in the lungs of patients with chronic bronchitis (28). We could not demonstrate a significant change in antibody titer due to infection with a new H. influenzae strain (figure 3). In contrast, Reichek and coworkers (7) and Musher and associates (10)reported a significant increase in serum antibody titer in COPD patients with H. influenzae infections, but in these studies the patients had remarkably low preinfection antibody titers compared to the controls, suggesting that they were not infected by H. influenzae previously. The persistence of H. influenzae for several months did not cause a significant change in antibody titer to the infecting strain (figure 4). These results confirmed data obtained with a reference strain by Glynn (4) and Gump and col-
leagues (8). In addition, we showed that the specificity of serum and sputum antibodies was not influenced by the appearance or persistence of H. influenzae. Summarizing, we conclude that it is very likely that COPD patients are perfectly able to evoke an antibody response to the major antigens of the infecting H. influenzae. We can only speculate why these patients remain infected by H. influenzae for longer periods or become infected with a new H. influenzae strain despite high titers of specific antibodies. Wedo not know which epitopes on nontypable H. influenzae elicit antibodies necessary for protection. It might therefore be that in the background of the high antibody titers to the infecting H. influenzae strain deficiencies in antibodies to these epitopes are missed. Second, it may be that the antibody-dependent effector mechanisms, such as opsonophagocytosis or complement-dependent killing of the bacteria, are defective. Finally, it would be of interest to determine the proportion of IgA. and IgA z in the sputum antibodies, since H. influenzae is able to produce IgA. proteases, enzymes
that cleave IgA. and thereby destroy the functionality of these molecules (29, 30). Apart from the possible defectivedefense mechanisms of the host, we have recently shown that the MOMP composition of H. influenzae changes during persistence in COPD patients (31). If these changes prove to be of immunologic relevance, this might be a mechanism that enables the bacterium to persist. Acknowledgment The writers thank F. Garnier, M.D., J. Hemgreen, M.D., A. Leentvaar, M.D., and the physicians from the Pulmonary Department of the Academic Medical Centre for their kind cooperation in collecting the sputa and sera, and the Laboratory of Medical Microbiology of the Municipal Health Council for culturing bacterial strains. Dr. T. A. Out is gratefully acknowledged for critical comments. References May JR. In: Taverner D, Trounce J, eds. The chemotherapy ofchronic bronchitis and allied disorders. London: English Universities Press, 1972; 11-4. 2. Murphy TF, Apicella MA. Nontypable Haemophilus influenzae: A review of clinical aspects, I.
1321
ANTIBODIES TO H. INFLUENZAE IN COPD PATIENTS
surface antigens, and the human immune response to infection. Rev Infect Dis 1987; 9:1-15. 3. Groeneveld K, van Alphen L, Geelen-van den Broek NJ, Eijk PP, Zanen HC, van Ketel RJ. Detection of Haemophilus injluenzae with monoclonal antibody. Lancet 1987; 1:441-2. 4. Glynn AA. Antibodies to Haemophilus influenzaein chronic bronchitis. Br Med J (Clin Res) 1959; 2:911-4. 5. Burns MW, May JR. Haemophilus influenzae precipitins in the serum of patients with chronic bronchial disorders. Lancet 1967; 1:354-8. 6. Smith CB, Golden CA, Kamer RE, Renzetti AD. Haemophilus influenzae and Haemophilus parainfluenzae in chronic obstructive pulmonary disease. Lancet 1967; 1:1253-5. 7. Reichek N, LewinEB, Rhoden DL, WeaverRR, Crutcher JC. Antibody responses to bacterial antigens during exacerbations of chronic bronchitis. Am Rev Respir Dis 1970; 101:238-44. 8. Gump DW, Christmas WA, Forsyth BR, Phillips CA, Stouch WHo Serum and secretory antibodies in patients with chronic bronchitis. Arch Intern Med 1973; 132:847-51. 9. May JR, Peto R, Tinker CM, Fletcher CM. A study of Haemophilus influenzae precipitins in the serum of working men in relation to smoking habits, bronchial infection, and airway obstruction. Am Rev Respir Dis 1973; 108:460-8. 10. Musher DM, Kubitschek KR, Crennan J, Baughn RE. Pneumonia and acute febrile tracheobronchitis due to Haemophilus influenzae. Ann Intern Med 1983; 99:444-50. II. Musher DM, Hague-Park M, Baughn RE, Wallace RJ Jr, Cowley B. Opsonizing and bactericidal effects of normal human serum on nontypable Haemophilus influenzae. Infect Immun 1983; 39:297-304. 12. Hansen MV,Musher DM, Baughn RE. Outer membrane proteins of nontypable Haemophilus influenzae and reactivity of paired sera from infect-
ed patients with their homologous isolates. Infect Immun 1985; 47:843-6. 13. Murphy TF, Apicella MA. Antigenic heterogeneity of outer membrane proteins of nontypable Haemophilus influenzae is a basis for a serotyping system. Infect Immun 1985; 50:15-21. 14. Campagnari AA, Gupta MR, Dudas KC, Murphy TF, Apicella MA. Antigenic diversity of lipooligosaccharides of nontypable Haemophilus injluenzae. Infect Immun 1987; 55:882-7. 15. American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 136:225-44. 16. Mulder J, Goslings WRO, van der Plas MC, Lopes Cardozo P. Studies on the treatment with antibacterial drugs of acute and chronic mucopurulent bronchitis caused by H. influenzae. Acta Med Scand 1952; 143:32-49. 17. Kilian M. A taxonomic study on the genus Haemophilus, with the proposal of a new species. J Gen Microbiol 1976; 93:9-62. 18. Dirks-Go SIS, Zanen He. Latex agglutination, counterimmunoelectrophoresis and protein A coagglutination in diagnosis of bacterial meningitis. J Clin Pat hoi 1978; 31:1167-71. 19. Lugtenberg B, Meijers J, Peters R, van der Hoek P, van Alphen L. Electrophoretic resolution of the "major outer membrane protein" of Escherichia coli KI2 into four bands. FEBS Lett 1975; 58:254-8. 20. Van Alphen L, Riemens T, Poolman J, Zanen HC. Characteristics of major outer membrane proteins of Haemophilus influenzae. J Bacteriol1983; 155:878-85. 21. Van Alphen L, Romijn C, Brandt H, Geelen L, Zanen HC. Preparation of cellenvelopes of large numbers of individual bacterial strains with the use of an automatic cell disruptor. Anal Biochem 1987; 166:36-40. 22. Lowry OH, Rosebrough NJ, Farr AL, Ran-
dall RJ. Protein measurement with the folin phenol reagent, J Bioi Chern 1951; 193:265-75. 23. Delacroix DL, Marchandise FX, Francis C, Sibille y. Alpha-2-macroglobulin, monomeric and polymeric immunoglobulin A, and immunoglobulin M in bronchoalveolar lavage. Am Rev Respir Dis 1985; 132:829-35. 24. Rodkey FL. Direct spectophotometric determination of albumin in human serum. Clin Chern 1965; 11:478-87. 25. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Nat! Acad Sci USA 1979; 76:4350-4. 26. Dunn SD. Effects of the modification of transfer buffer composition and the renaturation of proteins in gels on the recognition of proteins on Western blots by monoclonal antibodies. Anal Biochem 1986; 157:144-53. 27. ReynoldsHY. Bronchoalveolar lavage.Am Rev Respir Dis 1987; 135:250-63. 28. Burnett D, Crocker J, Stockley RA. Cells containing IgA subclasses in bronchi of subjects with and without chronic obstructive lung disease.J Clin Pathol 1987; 40:1217-20. 29. Kilian M, Mestecky J, Schrohenloher RE. Pathogenic species of the genus Haemophilus and Streptococcus pneumoniae produce immunoglobulin A, protease. Infect Immun 1979; 26:143-9. 30. Mansa B, Kilian M. Retained antigen-binding activity of Fab fragments of human monoclonal immunoglobulin Al (IgA,) cleaved by IgA, protease. Infect Immun 1986; 52:171-4. 31. Groeneveld K, van Alphen L, Eijk PP, Jansen HM, Zanen He. Changes in outer membrane proteins of nontypable Haemophilus influenzae in patients with chronic obstructive pulmonary disease. J Infect Dis 1988; 158:360-5.
