Immunology 1991 74 99-106
IgA-mediated inhibition of human leucocyte function by interference with Fcy and C3b receptors K. SAITO, C. KATO, H. KATSURAGI & A. KOMATSUZAKI Department of Oral Microbiology, School of Dentistry at Niigata, The Nippon Dental University, Niigata, Japan Acceptedfor publication 15 May 1991
SUMMARY The inhibitory effects of IgA from human colostrum, and IgA 1 and IgA2 from human serum on the chemiluminescence (CL) response and phagocytosis of polymorphonuclear leucocytes (PML) to Staphylococcus epidermidis and the CL response to formylmethionyl-leucyl-phenylalanine (FMLP) were studied. The dose-dependent inhibition of the luminol-mediated CL response of human PML to the bacteria was observed in the presence of more than 0-1 mg/ml IgA from both colostrum and serum. The preincubation of PML with a solution of IgA enhanced the suppressive effect of IgA on the cells. Removal of IgA from the reaction mixture after preincubation resulted in recovery, with time, of the response of PML to the bacteria. The bacteria treated with IgA did not give rise to any inhibition of the response. The CL response of PML to FMLP was not affected by the presence of IgA in the reaction mixture. The decrease of phagocytic activity of PML in the presence of IgA resulted in a decrease of NADPH oxidase activity of PML after stimulation with the bacteria as compared with the absence of IgA. The effect of IgA on the receptors of Fc and C3b (CR1 ) on the surface of PML was measured by monitoring erythrocyte-antibody (EA) or erythrocyte-antibody-complement (EAC) rosette formation and by direct and indirect immunofluorescence techniques using anti-CR1 antibody and Fc-specific antibodies. The presence of IgA in the reaction mixture led to a quantitative decrease in CR1 and the ability to bind IgG to the surface of PML.
specific IgG.4 9 Less IgG is present on the mucosal surface than in the serum, so the effect of IgA on the function of IgG may be increased more in the gingival crevice than in other parts of the body. Recently, several investigators have identified receptors specific for the Fc region of IgA on peripheral and alveolar macrophages and PML.2'0-'8 In our previous paper,'9 it was reported that saliva and an IgA-rich fraction from saliva inhibit both the CL response of murine PML to Staphylococcus epidermidis and the phagocytic activity of PML. In this report, the suppressive activity of IgA was confirmed using human PML and the cause of inhibition was investigated. It appears that blocking of receptors for C3b and Fc, both of which are closely associated with phagocytosis, may be the cause of the inhibition in the presence of IgA.
INTRODUCTION IgA is the predominant immunoglobulin in external secretions of humans and provides humoral protection against pathogens. The amount of IgA was found to be 600 mg/dl in colostrum, 160 mg/dl in serum and 19-4 mg/dl in mixed saliva. ' The obvious biological function of IgA is to protect the host from invading organisms which initiate their assault from outside the mucosal surfaces.2 In the oral cavity, the aggregation of micro-organisms by IgA retards their movement and the large aggregate of bacteria helps in the removal of the bacteria from the oral cavity to the alimentary tract. Polymorphonuclear leucocytes (PML) come into the oral cavity from blood vessels to the gingival crevice via the junctional epithelium of the gingiva. PML interact with micro-organisms in the crevices of subgingival plaque and are capable of killing the engulfed bacteria via oxygen-dependent and independent mechanisms.3 However, there have been several reports of IgA inhibiting the bactericidal activity and suppressing the leucocyte-associated function of
MATERIALS AND METHODS
Materials IgA (from normal human colostrum IgA), IgG (from normal human serum), human serum albumin (HSA; from fraction V, essentially fatty acid-free), prostaglandins El and E2 (PGE1, PGE2) and arachidonic acids (AA) were purchased from Sigma Chemical Co. (St Louis, MO). All globulins from Sigma had
Correspondence: K. Saito, Dept. of Oral Microbiology, School of Dentistry at Niigata, The Nippon Dental University, Hamauracho- 1-8, Niigata, Japan 951.
99
100
K. Saito et al.
been purified through ion-exchange or gel filtration chromatography. The purity was checked by immuno-electrophoresis and/or Ouchterlony double diffusion and 7-5% PAGE. No contamination was detected in IgA or IgG samples. IgAl and IgA2 from normal pooled human serum were obtained from Protagen AG imported by COSMO Ltd, Tokyo, Japan. FMLP (formylmethionyl-leucyl-phenyl-alanine) was purchased from LABOSCIENCE Co. Ltd, Tokyo, Japan and [3H]AA (93-6 Ci/ mmol) was obtained from New England Nuclear, Boston, MA. A solution of luminol, Lumilight, was purchased from Analytical Luminescence Laboratory Inc., San Diego, CA. IgAl and IgA2 were dialysed against 3 1 of distilled water and relyophilized before use to remove mannose. PAGE analysis revealed that the colostrum IgA was dimeric whilst IgAl and IgA2 were recovered from the dialysis as monomers.
Purification of IgA by Jacalin20 Ten mg of colostrum IgA in 2 ml of phosphate-buffered saline (PBS) were applied to the 15 ml of Sepharose 4B-Jacalin column equilibraed with PBS, pH 7-4. After washing with PBS, the bound IgA was eluated with 0-1 M melibiose and was continuously collected in 5-ml samples. IgA-containing fractions were pooled and dialysed against 2 1 of distilled water and lyophilized. Almost all of the IgA eluted from the column was IgA2 dimer form determined by PAGE and reduced SDSPAGE analysis. Human polymorphonuclear leucocytes (PML) Human PML were separated from the heparinized blood of healthy subjects by a Mono-Poly Resolving medium (Flow Laboratories, North Ryde, NSW, Australia) described previously. '9 The concentration ofcells was adjusted to I x 1 07 cells/ ml with 0-1% gelatin in Hanks' solution (GHS).
Stimulation of PML An aliquot of 0-1 ml of FMLP (1 x I0-I M) and a suspension of S. epidermidis (4 x I07 cells/ml) were used for stimulation of PML. The bacteria were harvested after growth in brain-heart infusion (BHI) broth (Difco Lab., Detroit, MI) at 370 for 18 hr. After two washes with PBS, the concentration of bacteria was adjusted to a turbidity of 0 3 (4 x I07 cells/ml) at 540 nm
(Shimazu Spectronic 20, Tokyo, Japan). Measurement of chemiluminescence (CL) of PML The assay of CL was performed according to the methods of our previous paper.'9 The total volume of 0-65 ml reaction mixture consisted of PML (2 x 106 cells), complement (0 05 ml of 1:10 guinea-pig serum), luminol (0-1 ml of 1: 5, 1 x 10-3 M Lumilight) and GHS. Intensity of CL response was measured by a 'Lumiphotometer TD 4000 (LABO Science Co. Ltd, Tokyo, Japan). After 5 min of incubation, the reaction mixture in the cuvette was stimulated with 0-1 ml of bacterial suspension or FMLP. Complement was omitted when FMLP was used as the stimulator of CL. The intensity of light emitted from the cuvette was recorded automatically. To observe the effect of IgA on the CL response, PML and various concentrations of IgA were mixed and the CL response was titrated with or without preincubation for 30 min. The CL values are shown as c.p.m. which were converted from the average relative intensity of CL during peak time + 5 min by comparison with the relative intensity of a 14C standard sample.
Estimation of phagocytic activities of PML A portion of the sample in the cuvette used for the CL reaction was placed on a glass microscope slide, air-dried, and Gramstained after fixation with ethanol. The phagocytic index and the total number of bacteria in 100 PML were counted in the same specimen, under a light microscope. Data are presented as the average of the results from more than 300 cells observed in more than three separate samples. Assay for NADPH oxidase activity in PML NADPH oxidase activity was assayed using a modified method of Oyanagi.2' The activity was assayed by the decrease in absorption of NADPH at 340 nm using a Hitachi 624 digital spectrophotometer (Nissei Sangyo Co. Ltd, Tokyo, Japan) at 370. A 3-ml cuvette contained 124 mM sodium phosphate buffer, pH 6-5, 0 05 mm EDTA, 1-2 U lactate dehydrogenase (LDH) and 320 ymol NADPH (Boehringer Mannheim-Yamanouchi Co., Tokyo, Japan), and 3 x 106 PML, either with or without 0-9 mg of IgA. The reaction was started by the addition of 0 3 ml of the suspension of S. epidermidis. A sample with IgA and the bacteria but without PML was used to obtain the basal value of NADPH oxidation. Measurement of release of AA from the PML A modified method of Maridenneau-Parini et al.22 was used. PML (I x 108/ml) were incubated for 15 min at 37° in Ca2+ and Mg2+-free PBS that contained 1 pCi of [3H]AA at pH 7-2. Cells were washed once with 0-1% fatty acid-free HSA in PBS and twice with PBS. The cells were suspended at 107 cells/ml in PBS which contained 0-9 mm CaCl2 and 0 1% HSA. The incorporation of [3H]AA by cells was 34-7 + 5-3% of the total radioactivity added to the cells. A plastic tube containing 1 ml of radiolabelled cells (1 x 107 cells/ml) was placed in a shaking water bath at 370 and incubated with either FMLP or S. epidermidis and complement in the presence or absence of IgA. The reaction was terminated by immersing the test-tube in ice. A duplicate sample of 100 pi of the cell supernatant was counted in a liquid scintillation counter (Packard model 3385, Packard Instrument Co. Inc., Downers Grove, IL).
Assay for prostaglandin E2 releasedfrom PML AA (Sigma) was dissolved initially in 0-1% NaCO3 and diluted with PBS, pH 7-2. A concentration of 0- 1 mg/ml of AA was used for incorporation into PML. The supernatant of the reaction mixture, which included PML-incorporated AA, S. epidermidis, and complement, with or without IgA, was prepared after incubation for 1 hr at 37°. The supernatant was transferred to a tube that contained EDTA, indomethacin, and 1.2.4 trazole and sent in frozen form to the Special Reference Lab. Co., Tokyo, Japan for an assay of PGE2.23 The titre was presented as pg/ml.
Assay for quantification of receptors for C3b and Fc on PML Presence of C3b receptor on PML was assayed by counting the rosette-forming cells with OxEAC (ox red blood cells sensitized with anti-ox-red blood cell-specific rabbit IgM antibody and reacted with human complement, and the reaction stopped at C3b; Otsuka Assay Lab., Tokushima, Japan). The extent of binding of fluorescein isothiocyanate (FITC)conjugated mouse IgG-specific rabbit antiserum (Becton-Dickinson Immunocytochemistry Systems, Mountain View, CA) directed against the murine monoclonal anti-human CR1 (CD,
101
IgA-mediated inhibition of human PML function 35: C3b receptor) specific mouse antibody (Becton-Dickinson, Monoclonal Center Inc., Mountain View, CA) which bound to the CRI on the surface of PML was also investigated. An aliquot of 0-5 ml of a suspension of 1 x 107/ml PML and the same volume of OxEAC (2 x 108 cells) were incubated together at 370 for 1 hr in a water bath. After a fivefold dilution of the incubated sample with GHS, one drop of a 0-5% solution of crystal violet was added to 0 5 ml of the sample. The number of rosettes was counted by a haemocytometer under a light
30
microscope.
For the assay of Fc receptor, OxEA (ox red blood cells sensitized with anti-ox red blood cell-specific rabbit IgG; Otsuka Assay Lab.) was used to make the EA rosettes. OxEA rosette formation was performed with the same procedure of OxEAC replacing OxEAC with OxEA. The relative intensity of fluorescence associated with CR1 was measured by a fluorescence microphotometer (Olympus System photometer TUR, Olympus Optical Co. Ltd, Tokyo, Japan). Aliquots of 107/ml cells were incubated in the presence or absence of 0-3 mg/ml IgA (final concentration) at 370 for 30 min. Each sample was gently dried on a glass slide and fixed with absolute acetone. A solution of CR1 -specific mouse antibody was placed on the specimen and the slide was incubated in a humidified box at 370 for 4 hr. Following two washes with PBS, FITC-conjugated goat antimouse IgG antibody (DAKO Immunoglobulins, Glostrup, Denmark)-was applied to the specimen. After 4 hr of incubation, the residual antibody on the specimen was removed by rinsing with cold PBS and the specimen was mounted with 30% glycerol in PBS. The intensity of fluorescence of 400 cells in three separate experiments was measured using a fluorometer. For the assay of Fc receptors, 0 6 ml of 1 x 107 cells/ml were incubated with human IgG at a final concentration of 0-3 mg/ ml. To observe the inhibitory effect of IgA on binding IgG to Fc receptor, PML were incubated in the mixture of 0-3 mg/ml IgG and 0-3 mg/ml of IgA, or 0 3 mg/ml IgA followed by the same concentration of IgG at 37°. The incubated cells were washed twice with 2 ml of PBS and smeared on a glass slide and stained with 10-fold diluted FITC-conjugated goat anti-human IgG specific antibody (Dako Immunoglobulins). After rinsing out the residual antibody, the specimens were mounted with mounting medium. The intensity of fluorescence from each of 400 cells of a specimen was measured. The experiment was repeated three times.
CL response of PML to S. epidermidis and its dependency complement (C)
on
The CL response of PML to the bacteria was measured with C, heat-inactivated C or without C. Figure 1 shows the time-course of the average intensity of the CL of PML in five separate samples. The mean intensity (c.p.m. and SD) at the peak time was as follows: 2 64+0A44 x 107 c.p.m. with C, 018+0 064x 107 c.p.m. withoutC, and 016+0 08 x 107 c.p.m. with heat-inactivated C. Effect of IgA on the CL response of PML to S. epidermidis The effect of IgA on the CL response was measured in the
40 30 20 Minutes after stimulation
I0
50
Figure 1. The CL response of PML to S. epidermidis and the dependency of complement (C). In the experiment, 1 x 106 cells were stimulated with 4 x 106 bacteria. The data show the mean and SD (vertical lines) of five A, separate experiments on five different donors. *0*, with C; without C; *U*, with heat-inactivated C.
Table 1. Effect of IgA
Concentrations of IgA (mg/ml) IgA 05
on
the CL
response
of PML to S. epidermidis
Inhibition Peak value of CL (%) Preincubation mean + SD x 106 C.p.m. + + + +
3-24 + 1-93 05+028 1-85 +0-36 0-98+0-18 19-29+1 00 2-60+0-25 25-27 + 1-60 28-5+5-36
0.1
+
1-9+0-08
93 4
IgA2 01
+
19+009
934
2-81+91-5
91 5
27-84+4-50
-11 0
0-25 01
0-0
RESULTS
CloL 20
E
87-2 98-2 92-8 96-6 23-7 90 09 0 0
IgAl
Jacalin-purified IgA (IgA2) 01 + HSA 10
-
Mean + SD was calculated from the results of five different experiments. +: PM L were incubated for 30 min with the given concentration of
IgA.
102
K. Saito et al. Table 2. Effect of IgA on phagocytic activities of PML
Concentrations of IgA (mg/ml)
Preincubation
Total no. of bacteria in 100 PML Mean + SD
Inhibition (0/)
Phagocytic index Mean + SD (%)
Inhibition (%)
+ + + +
143+47 133+29 271 +304 195+11 5 507+525 305+19-0 773+43 844+36
83 1 84 3 65-0 76-9 345 83-9 0 0
25+6-0 21-3+4-9 39-6+ 1-4 35-6+63 723+593 43-3+70 85 3+5-0 92-6+8-4
707 87-7 53-6 61-4 153 53 3 0 0
121 + 16-8
87-2
23-0+ 1-4
75-2
114+ 12-8
86 5
20-1 +2-8
78-3
234+24
73-3
38-4+6-2
58-6
811 +51
-4-9
83-6+8-2
2-0
IgA 05
025 0.1 0
IgA I 0.1 + IgA2 0.1 + Jacalin-purified IgA (IgA2) 0.1 + HSA 1-0 -
+: PML were incubated at 37° for 30 min with given concentration of IgA. The phagocytic ratio and total number of bacteria in 100 PML were estimated by microscopic observation of specimens which was made after the CL response was complete. Mean + SD was calculated from the results of three to five different experiments in each case.
presence of various concentrations of IgA, with or without preincubation. Nearly 90% inhibition of the CL response was observed at concentrations of IgA greater than 1-25 mg/ml without preincubation and greater than 0 1 mg/ml with preincubation at 370 for 30 min (Table 1). Jacalin-purified IgA, IgAl and IgA2 from serum inhibited the CL intensity of PML to less than 10% of the control after preincubation (Table 1). Human serum albumin 1 mg/ml did not inhibit the CL response of PML.
50
0
EC
,, I +IgA0
~ ~~
~
Effect of IgA on the phagocytic activity of PML Table 2 shows the phagocytic index and the total number of bacteria in 100 leucocytes at various concentrations of IgA. A dose-dependent decrease in the phagocytic function of PML is apparent. The decrease in phagocytic function at the same concentration of IgA was enhanced by preincubation for 30 min at 37°. The decrease of phagocytic function was also observed with Jacalin-purified IgA (IgA2), IgAl and IgA2.
~
C.)
CI ~0
25
IgAo
Figure 002. Effect of
Figue
2.Effet
ofIgA
n
NAPH oid
+ciiyofPL
h
reaction mixture of 3 ml contained 125 mm sodium phosphate buffer (pH 6 8), 12 U LDH, 320 mM NADPH, 3 x 106 PML and 0-15 ml of 10-fold diluted complement, with or without 0-9 mg IgA. The reaction was started by addition of S. epidermidis (4 x 106 bacteria).
Effect of IgA on the oxidation of NADPH by PML Figure 2 shows the NADPH oxidase activity of PML with 0-3 mg/ml of IgA or without IgA, when PML were stimulated with the bacteria in the presence of C. The rate of oxidation of NADPH was 4 29 + 052 nmol/106 PML/min without IgA and 1-06+0-2 nmol/106 PML/min with IgA. In the presence of IgA in the reaction mixture, the rate of oxidation of NADPH was greatly reduced.
Recovery from the inhibition of PML function by removal of IgA Figure 3 shows the recovery of the PML function after the
103
IgA-mediated inhibition of human PML function
Table 4. Effect of IgA on the receptor for either Fcy or C3b on PML
Concentration of IgA (mg/ml)
EA rosettes/500 cells
EAC rosettes/500 cells
423 + 38-6 374+523
256 + 32 2 126+16.5*
100 0 03 PI Data are expressed as the mean of five different experiments. * Significant P < 0 05.
ITT
8
..
Nearly complete recovery of phagocytic activity was observed after the removal of IgA whereas the recovery of the CL response still remained only 69% of the control.
50
Effect of IgA on PML stimulated with FMLP There was no inhibition of the CL response of PML to FMLP in the presence of IgA. In contrast to the stimulation by S. epidermidis, the peak intensity was increased in the presence of more than 05 mg/ml of IgA, and in the presence of 1 mg/ml IgA the response was 134-5% of the control (Table 3).
10 min 60 min Time after removal of IgA Figure 3. Recovery of function of PML from the effects of IgA after the removal of IgA. The PML were incubated with 0 3 mg/ml of IgA at 37° for 30 min and after washing twice in GHS with centrifugation at 300 g for 5 minm PML were kept in an ice-bath until the time indicated in the figure. The recovery rate of PML function is shown as a percentage of the control (without IgA). CL, chemiluminescence; PI, phagocytic index (%); No., number of bacteria in 100 cells.
Table 3. Effect of IgA
Concentration of IgA
(mg/mi) 1 05 03 0 (control)
the CL FMLP
on
response
Peak values of CL x
106 c.p.m. 3359 271 267 2-68
of PML to
% of
control 134-5 101 1 996 100
Data quoted are the mean of three different experi1 x 106 PML and 01 ml of I x 10-3 M FMLP were used. ments.
removal of IgA from suppression of PML that were preincubated for 30 min at 370 with 0-3 mg/ml of IgA. Ten minutes after the removal of IgA, the CL response of PML, phagocytic index, and the number of bacteria in 100 PML were 34%, 36%, and 39% of values in the absence of IgA (control), respectively.
Effect of IgA on EA or EAC rosette formation The number of EA rosette-forming cells in a group of 500 PML is shown in Table 4. The mean and SD of the number of EA rosettes was 423 + 38-6 without IgA while it was 374 + 52 3 with IgA, which is not significantly different. The number of EAC rosette-forming cells with IgA and without IgA was 256 + 32-2 and 126+ 16-5, respectively, and the difference was significant (P < 0 05). Effect of IgA on receptors for C3b and Fc on the surface of PML To determine the effect of IgA on the receptors of PML which participate in phagocytic functions, the intensity of immunofluorescence was counted using a fluorescence microphotometer, and the cells were divided into five groups (CR1) and four groups (Fc) according to the grade of intensity. Table 5 shows that in the presence of 0-3 mg/ml of IgA, 37-5% of cells which had a CR1 intensity of between 1000 and 3000 in the absence of IgA showed a decrease in CR1 intensity to less than 1000. The distribution of PML with fluorescence due to FITCconjugated Fc chain-specific antibody The distribution of four grades of fluorescence did not change in the presence of IgA from that without IgA. However, after treatment with IgG(0-3), PML were stained with y-chain specific antibody-conjugated FITC, and a large portion of cells developed fluorescence intensities of above 2000. In the mixture of both IgA and IgG, the distribution of relative intensities of fluorescence among the cells stained with FITC-conjugated Fcyspecific antibody was different from that in the presence of IgG alone. The cells with a lower level of fluorescence intensity were increased in number. Furthermore, when IgA had been added prior to the addition of IgG, the number of cells with a lower
K. Saito et al.
104
Table 5. Effect of IgA on the receptors for C3b or Fcy on the surface of PML PML were incubated with Ab (mg/ml)
% distribution of PML with fluorescence
500
1000
2000
3000
> 3000
Mean intensity of fluorescence
CRI
None IgA 0 3 HSA 1.0
44-1 81-6 46-2
68 2-8 4-8
164 0 172
265 8-2 256
65 7-4 6-2
1257-6+ 1168 (P
IgA-mediated inhibition of human leucocyte function by interference with Fcy and C3b receptors K. SAITO, C. KATO, H. KATSURAGI & A. KOMATSUZAKI Department of Oral Microbiology, School of Dentistry at Niigata, The Nippon Dental University, Niigata, Japan Acceptedfor publication 15 May 1991
SUMMARY The inhibitory effects of IgA from human colostrum, and IgA 1 and IgA2 from human serum on the chemiluminescence (CL) response and phagocytosis of polymorphonuclear leucocytes (PML) to Staphylococcus epidermidis and the CL response to formylmethionyl-leucyl-phenylalanine (FMLP) were studied. The dose-dependent inhibition of the luminol-mediated CL response of human PML to the bacteria was observed in the presence of more than 0-1 mg/ml IgA from both colostrum and serum. The preincubation of PML with a solution of IgA enhanced the suppressive effect of IgA on the cells. Removal of IgA from the reaction mixture after preincubation resulted in recovery, with time, of the response of PML to the bacteria. The bacteria treated with IgA did not give rise to any inhibition of the response. The CL response of PML to FMLP was not affected by the presence of IgA in the reaction mixture. The decrease of phagocytic activity of PML in the presence of IgA resulted in a decrease of NADPH oxidase activity of PML after stimulation with the bacteria as compared with the absence of IgA. The effect of IgA on the receptors of Fc and C3b (CR1 ) on the surface of PML was measured by monitoring erythrocyte-antibody (EA) or erythrocyte-antibody-complement (EAC) rosette formation and by direct and indirect immunofluorescence techniques using anti-CR1 antibody and Fc-specific antibodies. The presence of IgA in the reaction mixture led to a quantitative decrease in CR1 and the ability to bind IgG to the surface of PML.
specific IgG.4 9 Less IgG is present on the mucosal surface than in the serum, so the effect of IgA on the function of IgG may be increased more in the gingival crevice than in other parts of the body. Recently, several investigators have identified receptors specific for the Fc region of IgA on peripheral and alveolar macrophages and PML.2'0-'8 In our previous paper,'9 it was reported that saliva and an IgA-rich fraction from saliva inhibit both the CL response of murine PML to Staphylococcus epidermidis and the phagocytic activity of PML. In this report, the suppressive activity of IgA was confirmed using human PML and the cause of inhibition was investigated. It appears that blocking of receptors for C3b and Fc, both of which are closely associated with phagocytosis, may be the cause of the inhibition in the presence of IgA.
INTRODUCTION IgA is the predominant immunoglobulin in external secretions of humans and provides humoral protection against pathogens. The amount of IgA was found to be 600 mg/dl in colostrum, 160 mg/dl in serum and 19-4 mg/dl in mixed saliva. ' The obvious biological function of IgA is to protect the host from invading organisms which initiate their assault from outside the mucosal surfaces.2 In the oral cavity, the aggregation of micro-organisms by IgA retards their movement and the large aggregate of bacteria helps in the removal of the bacteria from the oral cavity to the alimentary tract. Polymorphonuclear leucocytes (PML) come into the oral cavity from blood vessels to the gingival crevice via the junctional epithelium of the gingiva. PML interact with micro-organisms in the crevices of subgingival plaque and are capable of killing the engulfed bacteria via oxygen-dependent and independent mechanisms.3 However, there have been several reports of IgA inhibiting the bactericidal activity and suppressing the leucocyte-associated function of
MATERIALS AND METHODS
Materials IgA (from normal human colostrum IgA), IgG (from normal human serum), human serum albumin (HSA; from fraction V, essentially fatty acid-free), prostaglandins El and E2 (PGE1, PGE2) and arachidonic acids (AA) were purchased from Sigma Chemical Co. (St Louis, MO). All globulins from Sigma had
Correspondence: K. Saito, Dept. of Oral Microbiology, School of Dentistry at Niigata, The Nippon Dental University, Hamauracho- 1-8, Niigata, Japan 951.
99
100
K. Saito et al.
been purified through ion-exchange or gel filtration chromatography. The purity was checked by immuno-electrophoresis and/or Ouchterlony double diffusion and 7-5% PAGE. No contamination was detected in IgA or IgG samples. IgAl and IgA2 from normal pooled human serum were obtained from Protagen AG imported by COSMO Ltd, Tokyo, Japan. FMLP (formylmethionyl-leucyl-phenyl-alanine) was purchased from LABOSCIENCE Co. Ltd, Tokyo, Japan and [3H]AA (93-6 Ci/ mmol) was obtained from New England Nuclear, Boston, MA. A solution of luminol, Lumilight, was purchased from Analytical Luminescence Laboratory Inc., San Diego, CA. IgAl and IgA2 were dialysed against 3 1 of distilled water and relyophilized before use to remove mannose. PAGE analysis revealed that the colostrum IgA was dimeric whilst IgAl and IgA2 were recovered from the dialysis as monomers.
Purification of IgA by Jacalin20 Ten mg of colostrum IgA in 2 ml of phosphate-buffered saline (PBS) were applied to the 15 ml of Sepharose 4B-Jacalin column equilibraed with PBS, pH 7-4. After washing with PBS, the bound IgA was eluated with 0-1 M melibiose and was continuously collected in 5-ml samples. IgA-containing fractions were pooled and dialysed against 2 1 of distilled water and lyophilized. Almost all of the IgA eluted from the column was IgA2 dimer form determined by PAGE and reduced SDSPAGE analysis. Human polymorphonuclear leucocytes (PML) Human PML were separated from the heparinized blood of healthy subjects by a Mono-Poly Resolving medium (Flow Laboratories, North Ryde, NSW, Australia) described previously. '9 The concentration ofcells was adjusted to I x 1 07 cells/ ml with 0-1% gelatin in Hanks' solution (GHS).
Stimulation of PML An aliquot of 0-1 ml of FMLP (1 x I0-I M) and a suspension of S. epidermidis (4 x I07 cells/ml) were used for stimulation of PML. The bacteria were harvested after growth in brain-heart infusion (BHI) broth (Difco Lab., Detroit, MI) at 370 for 18 hr. After two washes with PBS, the concentration of bacteria was adjusted to a turbidity of 0 3 (4 x I07 cells/ml) at 540 nm
(Shimazu Spectronic 20, Tokyo, Japan). Measurement of chemiluminescence (CL) of PML The assay of CL was performed according to the methods of our previous paper.'9 The total volume of 0-65 ml reaction mixture consisted of PML (2 x 106 cells), complement (0 05 ml of 1:10 guinea-pig serum), luminol (0-1 ml of 1: 5, 1 x 10-3 M Lumilight) and GHS. Intensity of CL response was measured by a 'Lumiphotometer TD 4000 (LABO Science Co. Ltd, Tokyo, Japan). After 5 min of incubation, the reaction mixture in the cuvette was stimulated with 0-1 ml of bacterial suspension or FMLP. Complement was omitted when FMLP was used as the stimulator of CL. The intensity of light emitted from the cuvette was recorded automatically. To observe the effect of IgA on the CL response, PML and various concentrations of IgA were mixed and the CL response was titrated with or without preincubation for 30 min. The CL values are shown as c.p.m. which were converted from the average relative intensity of CL during peak time + 5 min by comparison with the relative intensity of a 14C standard sample.
Estimation of phagocytic activities of PML A portion of the sample in the cuvette used for the CL reaction was placed on a glass microscope slide, air-dried, and Gramstained after fixation with ethanol. The phagocytic index and the total number of bacteria in 100 PML were counted in the same specimen, under a light microscope. Data are presented as the average of the results from more than 300 cells observed in more than three separate samples. Assay for NADPH oxidase activity in PML NADPH oxidase activity was assayed using a modified method of Oyanagi.2' The activity was assayed by the decrease in absorption of NADPH at 340 nm using a Hitachi 624 digital spectrophotometer (Nissei Sangyo Co. Ltd, Tokyo, Japan) at 370. A 3-ml cuvette contained 124 mM sodium phosphate buffer, pH 6-5, 0 05 mm EDTA, 1-2 U lactate dehydrogenase (LDH) and 320 ymol NADPH (Boehringer Mannheim-Yamanouchi Co., Tokyo, Japan), and 3 x 106 PML, either with or without 0-9 mg of IgA. The reaction was started by the addition of 0 3 ml of the suspension of S. epidermidis. A sample with IgA and the bacteria but without PML was used to obtain the basal value of NADPH oxidation. Measurement of release of AA from the PML A modified method of Maridenneau-Parini et al.22 was used. PML (I x 108/ml) were incubated for 15 min at 37° in Ca2+ and Mg2+-free PBS that contained 1 pCi of [3H]AA at pH 7-2. Cells were washed once with 0-1% fatty acid-free HSA in PBS and twice with PBS. The cells were suspended at 107 cells/ml in PBS which contained 0-9 mm CaCl2 and 0 1% HSA. The incorporation of [3H]AA by cells was 34-7 + 5-3% of the total radioactivity added to the cells. A plastic tube containing 1 ml of radiolabelled cells (1 x 107 cells/ml) was placed in a shaking water bath at 370 and incubated with either FMLP or S. epidermidis and complement in the presence or absence of IgA. The reaction was terminated by immersing the test-tube in ice. A duplicate sample of 100 pi of the cell supernatant was counted in a liquid scintillation counter (Packard model 3385, Packard Instrument Co. Inc., Downers Grove, IL).
Assay for prostaglandin E2 releasedfrom PML AA (Sigma) was dissolved initially in 0-1% NaCO3 and diluted with PBS, pH 7-2. A concentration of 0- 1 mg/ml of AA was used for incorporation into PML. The supernatant of the reaction mixture, which included PML-incorporated AA, S. epidermidis, and complement, with or without IgA, was prepared after incubation for 1 hr at 37°. The supernatant was transferred to a tube that contained EDTA, indomethacin, and 1.2.4 trazole and sent in frozen form to the Special Reference Lab. Co., Tokyo, Japan for an assay of PGE2.23 The titre was presented as pg/ml.
Assay for quantification of receptors for C3b and Fc on PML Presence of C3b receptor on PML was assayed by counting the rosette-forming cells with OxEAC (ox red blood cells sensitized with anti-ox-red blood cell-specific rabbit IgM antibody and reacted with human complement, and the reaction stopped at C3b; Otsuka Assay Lab., Tokushima, Japan). The extent of binding of fluorescein isothiocyanate (FITC)conjugated mouse IgG-specific rabbit antiserum (Becton-Dickinson Immunocytochemistry Systems, Mountain View, CA) directed against the murine monoclonal anti-human CR1 (CD,
101
IgA-mediated inhibition of human PML function 35: C3b receptor) specific mouse antibody (Becton-Dickinson, Monoclonal Center Inc., Mountain View, CA) which bound to the CRI on the surface of PML was also investigated. An aliquot of 0-5 ml of a suspension of 1 x 107/ml PML and the same volume of OxEAC (2 x 108 cells) were incubated together at 370 for 1 hr in a water bath. After a fivefold dilution of the incubated sample with GHS, one drop of a 0-5% solution of crystal violet was added to 0 5 ml of the sample. The number of rosettes was counted by a haemocytometer under a light
30
microscope.
For the assay of Fc receptor, OxEA (ox red blood cells sensitized with anti-ox red blood cell-specific rabbit IgG; Otsuka Assay Lab.) was used to make the EA rosettes. OxEA rosette formation was performed with the same procedure of OxEAC replacing OxEAC with OxEA. The relative intensity of fluorescence associated with CR1 was measured by a fluorescence microphotometer (Olympus System photometer TUR, Olympus Optical Co. Ltd, Tokyo, Japan). Aliquots of 107/ml cells were incubated in the presence or absence of 0-3 mg/ml IgA (final concentration) at 370 for 30 min. Each sample was gently dried on a glass slide and fixed with absolute acetone. A solution of CR1 -specific mouse antibody was placed on the specimen and the slide was incubated in a humidified box at 370 for 4 hr. Following two washes with PBS, FITC-conjugated goat antimouse IgG antibody (DAKO Immunoglobulins, Glostrup, Denmark)-was applied to the specimen. After 4 hr of incubation, the residual antibody on the specimen was removed by rinsing with cold PBS and the specimen was mounted with 30% glycerol in PBS. The intensity of fluorescence of 400 cells in three separate experiments was measured using a fluorometer. For the assay of Fc receptors, 0 6 ml of 1 x 107 cells/ml were incubated with human IgG at a final concentration of 0-3 mg/ ml. To observe the inhibitory effect of IgA on binding IgG to Fc receptor, PML were incubated in the mixture of 0-3 mg/ml IgG and 0-3 mg/ml of IgA, or 0 3 mg/ml IgA followed by the same concentration of IgG at 37°. The incubated cells were washed twice with 2 ml of PBS and smeared on a glass slide and stained with 10-fold diluted FITC-conjugated goat anti-human IgG specific antibody (Dako Immunoglobulins). After rinsing out the residual antibody, the specimens were mounted with mounting medium. The intensity of fluorescence from each of 400 cells of a specimen was measured. The experiment was repeated three times.
CL response of PML to S. epidermidis and its dependency complement (C)
on
The CL response of PML to the bacteria was measured with C, heat-inactivated C or without C. Figure 1 shows the time-course of the average intensity of the CL of PML in five separate samples. The mean intensity (c.p.m. and SD) at the peak time was as follows: 2 64+0A44 x 107 c.p.m. with C, 018+0 064x 107 c.p.m. withoutC, and 016+0 08 x 107 c.p.m. with heat-inactivated C. Effect of IgA on the CL response of PML to S. epidermidis The effect of IgA on the CL response was measured in the
40 30 20 Minutes after stimulation
I0
50
Figure 1. The CL response of PML to S. epidermidis and the dependency of complement (C). In the experiment, 1 x 106 cells were stimulated with 4 x 106 bacteria. The data show the mean and SD (vertical lines) of five A, separate experiments on five different donors. *0*, with C; without C; *U*, with heat-inactivated C.
Table 1. Effect of IgA
Concentrations of IgA (mg/ml) IgA 05
on
the CL
response
of PML to S. epidermidis
Inhibition Peak value of CL (%) Preincubation mean + SD x 106 C.p.m. + + + +
3-24 + 1-93 05+028 1-85 +0-36 0-98+0-18 19-29+1 00 2-60+0-25 25-27 + 1-60 28-5+5-36
0.1
+
1-9+0-08
93 4
IgA2 01
+
19+009
934
2-81+91-5
91 5
27-84+4-50
-11 0
0-25 01
0-0
RESULTS
CloL 20
E
87-2 98-2 92-8 96-6 23-7 90 09 0 0
IgAl
Jacalin-purified IgA (IgA2) 01 + HSA 10
-
Mean + SD was calculated from the results of five different experiments. +: PM L were incubated for 30 min with the given concentration of
IgA.
102
K. Saito et al. Table 2. Effect of IgA on phagocytic activities of PML
Concentrations of IgA (mg/ml)
Preincubation
Total no. of bacteria in 100 PML Mean + SD
Inhibition (0/)
Phagocytic index Mean + SD (%)
Inhibition (%)
+ + + +
143+47 133+29 271 +304 195+11 5 507+525 305+19-0 773+43 844+36
83 1 84 3 65-0 76-9 345 83-9 0 0
25+6-0 21-3+4-9 39-6+ 1-4 35-6+63 723+593 43-3+70 85 3+5-0 92-6+8-4
707 87-7 53-6 61-4 153 53 3 0 0
121 + 16-8
87-2
23-0+ 1-4
75-2
114+ 12-8
86 5
20-1 +2-8
78-3
234+24
73-3
38-4+6-2
58-6
811 +51
-4-9
83-6+8-2
2-0
IgA 05
025 0.1 0
IgA I 0.1 + IgA2 0.1 + Jacalin-purified IgA (IgA2) 0.1 + HSA 1-0 -
+: PML were incubated at 37° for 30 min with given concentration of IgA. The phagocytic ratio and total number of bacteria in 100 PML were estimated by microscopic observation of specimens which was made after the CL response was complete. Mean + SD was calculated from the results of three to five different experiments in each case.
presence of various concentrations of IgA, with or without preincubation. Nearly 90% inhibition of the CL response was observed at concentrations of IgA greater than 1-25 mg/ml without preincubation and greater than 0 1 mg/ml with preincubation at 370 for 30 min (Table 1). Jacalin-purified IgA, IgAl and IgA2 from serum inhibited the CL intensity of PML to less than 10% of the control after preincubation (Table 1). Human serum albumin 1 mg/ml did not inhibit the CL response of PML.
50
0
EC
,, I +IgA0
~ ~~
~
Effect of IgA on the phagocytic activity of PML Table 2 shows the phagocytic index and the total number of bacteria in 100 leucocytes at various concentrations of IgA. A dose-dependent decrease in the phagocytic function of PML is apparent. The decrease in phagocytic function at the same concentration of IgA was enhanced by preincubation for 30 min at 37°. The decrease of phagocytic function was also observed with Jacalin-purified IgA (IgA2), IgAl and IgA2.
~
C.)
CI ~0
25
IgAo
Figure 002. Effect of
Figue
2.Effet
ofIgA
n
NAPH oid
+ciiyofPL
h
reaction mixture of 3 ml contained 125 mm sodium phosphate buffer (pH 6 8), 12 U LDH, 320 mM NADPH, 3 x 106 PML and 0-15 ml of 10-fold diluted complement, with or without 0-9 mg IgA. The reaction was started by addition of S. epidermidis (4 x 106 bacteria).
Effect of IgA on the oxidation of NADPH by PML Figure 2 shows the NADPH oxidase activity of PML with 0-3 mg/ml of IgA or without IgA, when PML were stimulated with the bacteria in the presence of C. The rate of oxidation of NADPH was 4 29 + 052 nmol/106 PML/min without IgA and 1-06+0-2 nmol/106 PML/min with IgA. In the presence of IgA in the reaction mixture, the rate of oxidation of NADPH was greatly reduced.
Recovery from the inhibition of PML function by removal of IgA Figure 3 shows the recovery of the PML function after the
103
IgA-mediated inhibition of human PML function
Table 4. Effect of IgA on the receptor for either Fcy or C3b on PML
Concentration of IgA (mg/ml)
EA rosettes/500 cells
EAC rosettes/500 cells
423 + 38-6 374+523
256 + 32 2 126+16.5*
100 0 03 PI Data are expressed as the mean of five different experiments. * Significant P < 0 05.
ITT
8
..
Nearly complete recovery of phagocytic activity was observed after the removal of IgA whereas the recovery of the CL response still remained only 69% of the control.
50
Effect of IgA on PML stimulated with FMLP There was no inhibition of the CL response of PML to FMLP in the presence of IgA. In contrast to the stimulation by S. epidermidis, the peak intensity was increased in the presence of more than 05 mg/ml of IgA, and in the presence of 1 mg/ml IgA the response was 134-5% of the control (Table 3).
10 min 60 min Time after removal of IgA Figure 3. Recovery of function of PML from the effects of IgA after the removal of IgA. The PML were incubated with 0 3 mg/ml of IgA at 37° for 30 min and after washing twice in GHS with centrifugation at 300 g for 5 minm PML were kept in an ice-bath until the time indicated in the figure. The recovery rate of PML function is shown as a percentage of the control (without IgA). CL, chemiluminescence; PI, phagocytic index (%); No., number of bacteria in 100 cells.
Table 3. Effect of IgA
Concentration of IgA
(mg/mi) 1 05 03 0 (control)
the CL FMLP
on
response
Peak values of CL x
106 c.p.m. 3359 271 267 2-68
of PML to
% of
control 134-5 101 1 996 100
Data quoted are the mean of three different experi1 x 106 PML and 01 ml of I x 10-3 M FMLP were used. ments.
removal of IgA from suppression of PML that were preincubated for 30 min at 370 with 0-3 mg/ml of IgA. Ten minutes after the removal of IgA, the CL response of PML, phagocytic index, and the number of bacteria in 100 PML were 34%, 36%, and 39% of values in the absence of IgA (control), respectively.
Effect of IgA on EA or EAC rosette formation The number of EA rosette-forming cells in a group of 500 PML is shown in Table 4. The mean and SD of the number of EA rosettes was 423 + 38-6 without IgA while it was 374 + 52 3 with IgA, which is not significantly different. The number of EAC rosette-forming cells with IgA and without IgA was 256 + 32-2 and 126+ 16-5, respectively, and the difference was significant (P < 0 05). Effect of IgA on receptors for C3b and Fc on the surface of PML To determine the effect of IgA on the receptors of PML which participate in phagocytic functions, the intensity of immunofluorescence was counted using a fluorescence microphotometer, and the cells were divided into five groups (CR1) and four groups (Fc) according to the grade of intensity. Table 5 shows that in the presence of 0-3 mg/ml of IgA, 37-5% of cells which had a CR1 intensity of between 1000 and 3000 in the absence of IgA showed a decrease in CR1 intensity to less than 1000. The distribution of PML with fluorescence due to FITCconjugated Fc chain-specific antibody The distribution of four grades of fluorescence did not change in the presence of IgA from that without IgA. However, after treatment with IgG(0-3), PML were stained with y-chain specific antibody-conjugated FITC, and a large portion of cells developed fluorescence intensities of above 2000. In the mixture of both IgA and IgG, the distribution of relative intensities of fluorescence among the cells stained with FITC-conjugated Fcyspecific antibody was different from that in the presence of IgG alone. The cells with a lower level of fluorescence intensity were increased in number. Furthermore, when IgA had been added prior to the addition of IgG, the number of cells with a lower
K. Saito et al.
104
Table 5. Effect of IgA on the receptors for C3b or Fcy on the surface of PML PML were incubated with Ab (mg/ml)
% distribution of PML with fluorescence
500
1000
2000
3000
> 3000
Mean intensity of fluorescence
CRI
None IgA 0 3 HSA 1.0
44-1 81-6 46-2
68 2-8 4-8
164 0 172
265 8-2 256
65 7-4 6-2
1257-6+ 1168 (P
