Monoclonal antibody passive hemagglutination and capture enzyme ...

Vol. 25, No. 2

JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1987, p. 278-284 0095-1137/87/020278-07$02.00/0 Copyright © 1987, American Society for Microbiology

Monoclonal Antibody Passive Hemagglutination and Capture Enzyme-Linked Immunosorbent Assays for Direct Detection and Quantitation of F41 and K99 Fimbrial Antigens in Enterotoxigenic Escherichia colit T. J. G. RAYBOULD,t* C. F. CROUCH,§ AND S. D. ACRES Veterinary Infectious Disease Organization, Saskatoon, Saskatchewan S7N OWO, Canada Received 7 August 1986/Accepted

5

November 1986

Production of diarrhea in neonatal calves by enterotoxigenic Escherichia coli depends on its ability to attach to the epithelial cells of the intestine via surface adhesins called pili or fimbriae and to secrete enterotoxins. The most important of these fimbriae are designated K99 and F41. We produced and characterized a murine monoclonal antibody specific to F41. This monoclonal antibody and a K99-specific monoclonal antibody were used to develop sensitive and specific passive hemagglutination and capture enzyme-linked immunosorbent assays (ELISAs) for detection and quantitation of F41 and K99 antigens in E. coli cultures and culture supernatants. The capture ELISA systems exhibited excellent sensitivity and specificity, whereas the passive hemagglutination systems appeared to be oversensitive. The ability of the capture ELISAs to detect K99 and F41 fimbrial antigens in fecal specimens from calves was evaluated. Fimbrial antigens were detected in six of six specimens from scouring calves but not in four of four specimens from nonscouring calves.

ETEC isolates from calves; they used a K99-specific monoclonal antibody (obtained from Molecular Genetics Inc., Minnetonka, Minn.) as a capture and detector (indicator) antibody. This paper describes production and characterization of a monoclonal antibody specific to F41 fimbrial antigen. The F41-specific antibody and the anti-K99 antibody fr-m Molecular Genetics were used for development of capture ELISA and passive hemagglutination (PHA) systems. Both of these systems detected F41 and K99 antigens in ETEC cultures at concentrations as low as 50 ng m-1'. Preliminary data indicated that the capture ELISAs also enable direct detection of these fimbriae in fecal specimens from scouring calves without the necessity for culturing.

Enterotoxigenic Escherichia coli (ETEC) is a common of diarrhea (scours) in neonatal calves. Induction of diarrhea requires initial colonization of the mucosal surface of the small intestine (21, 36). The ability of an ETEC strain to adhere to the villus epithelium depends on the presence on that strain of surface attachment factors (9), of which the most important are pili or fimbriae designated K99 (4, 15, 22, 23, 27, 30, 37) and F41 (8, 24, 25). Neonatal immunity to ETEC probably depends on the presence of colostral antibody against these surface antigens (1, 26, 28, 38), and several commercial vaccines are available for administration to pregnant cows. For these vaccines to be effective, they must contain sufficient amounts of the protective antigens. Procedures previously used for detection of pili in ETEC include agglutination, fluorescentantibody (2, 3), and enzyme-linked immunosorbent assays (ELISAs) (7, 17, 20). The direct agglutination assay is rather insensitive and has been prone to nonspecific agglutination or false-negative results because of thick capsules on some ETEC isolates. The fluorescent-antibody assay is a good technique, but many small laboratories do not have facilities for fluorescence assays. The ELISA is extremely sensitive and can be done with very little equipment. The ELISA has been used to diagnose various bacterial, viral, and fungal diseases (29), and its application in diagnosing ETEC infections in calves and pigs by detecting K99 (11, 18) and K88 (19) fimbrial antigens has been previously described. Any immunoassay relies on the specificity of the antibody used. Mills and Tietze (18) obtained excellent specificity in an ELISA for identification of cultured K99-positive (K99+) cause

*

MATERIALS AND METHODS

E. coli strains. The following ETEC strains were used in this study: B41 (0101:K-:K99+;H-;F41+); Bi316-42 (09K9(L)H12); a field isolate from a scouring calf, designated wild type (0101:K99+;F41+); and Moon 1676 (0101:K30;H- ;K99- ;F41+). Culture media and growth conditions. Minca medium contained KH2PO4 (1.36 g. liter-'), Na2HPO4 (10.1 g. liter-1), glucose (1.0 g. liter-1), Casamino Acids (Difco Laboratories), (1.0 g liter-'), and trace salts solution (1 ml liter-') as described by Guinée et al. (14) but supplemented with 1% IsoVitaleX. E. coli strains were cultured in 10-liter batches in a Magnaferm 12-liter fermentor (New Brunswick Scientific Co., Inc.) while being stirred at 200 rpm with an air flow rate of 2 liters min` at 5 lb./in.' for 12 h at 37°C. Fecal specimens. Fecal specimens were collected from (i) six scouring neonatal calves affected during a field outbreak of ETEC bacillosis, and (ii) four calves born in the Veterinary Infectious Disease Organization isolation facility. A sample of each specimen was streaked onto a surface of Minca agar containing 1% IsoVitaleX and incubated for 24 h at 37°C. After this time, plates were examined and, when growth was present, 10 colonies per plate were tested by -

Corresponding author.

t Published with permission of the Director of the Veterinary Infectious Disease Organization as journal series no. 46. t Present address: Molecular Immunology Division, Allelix Inc.,

Mississauga, Ontario L4V lPi, Canada. § Present address: Amersham International plc, Amersham, Bucks HP7 9LL, England. 278

VOL. 25, 1987

MONOCLONAL ANTIBODY ASSAYS FOR ETEC FIMBRIAE

plate agglutination with K99-specific rabbit antiserum (5). Further samples of each specimen were heat inactivated at 60°C for 1 h and then tested in both capture ELISA systems. Purified fimbrial preparations and rabbit antisera. Purified F41 antigen was supplied by F. K. de Graaf, Department of Microbiology, Vrije Universiteit, Amsterdam, The Netherlands. This antigen preparation was designated F41-dG (dG is an abbreviation of de Graaf). F41-specific rabbit antiserum was also supplied by F. K. de Graaf. Purified K99 antigen was provided by R. E. Isaacson, Department of Epidemiology, University of Michigan, Ann Arbor. Antiserum against K99 was prepared by hyperimmunizing an adult New Zealand White rabbit with purified K99 antigen (5). Capsular K30 antigen, purified from E. coli serotype 09:K30:H-, was supplied by M. Perry, Division of Biological Sciences, National Research Council of Canada, Ottawa. Antiserum against K30 was prepared by hyperimmunizing a rabbit with E. coli cells of this same serotype as described by Edwards and Ewing (10). Preparation of mouse hybridoma cell Unes. Eight-week-old BALB/c mice were inoculated in the footpad with 5 x 10' E. coli Moon 1676 cells (O101:K30;H-;K99-;F41+) in Freund incomplete adjuvant. After 4 weeks, the mice were given three intravenous inoculations of 107 bacteria at 3-day intervals. Three days after the last inoculation, the mice were sacrificed, and their spleen cells were mixed with Sp2/0 Agl4 myeloma cells (kindly provided by W. E. Rawls, Department of Pathology, McMaster University, Hamilton, Ontario, Canada) and fused with 40% polyethylene glycol 1450 (lot no. 052332; J. T. Baker Chemical Co.) by a modification (T. J. G. Raybould, B. E. Duncan, and D. H. Watson, Abstr. IV Int. Conf. Comp. Virol. 1982, P-28, p. 243) ofthe method described by Kennett et al. (16). The cells were dispensed into 24-well tissue culture plates (Linbro) and incubated at 37°C in an atmosphere containing 7.5% carbon dioxide. Screening of hybridoma culture supernatant fluids was performed by ELISA with microtiter plates coated with ultrasonicated E. coli Moon 1676 cells. Positive hybridoma cultures were cloned by limiting dilution in 96-well microtiter plates. After 14 days, the supernatant fluids from growing clones were rescreened by ELISA, and subcultures from positive monoclones were stored in liquid nitrogen. Cloned hybridoma cell lines were subsequently cultured in RPMI 1640 medium (GIBCO Laboratories) containing 10% fetal bovine serum. Ascitic fluid antibody was produced by intraperitoneal injection of monoclonal hybridoma cells into pristanetreated BALB/c mice. In some cases, mice were also given an intraperitoneal injection of 4 mg of cyclophosphamide (Procytox; F. W. Horner, Montreal, Quebec, Canada) 24 h before injection of the hybridoma cells. Development of assays. Two assay systems were used in these studies, PHA and ELISA. ELISA procedure. ELISAs were performed by the microplate modification (40) of the method of Engvall and Perlmann (12). Microtiter plates (Immulon 2; Dynatech Laboratories, Inc.) were coated with 100 pul of an optimum concentration of antigen or antibody preparation (determined by titration) per well, diluted in 0.05 M carbonatebicarbonate buffer (pH 9.6), and incubated overnight at 4°C. A standard washing procedure for microplates of three washes in 0.15 M saline containing 0.05% Tween 20 was adopted. Test samples and conjugates were diluted in 0.01 M phosphate-buffered 0.15 M saline (pH 7.2) containing 0.05% Tween 20. All incubations were performed at 20°C for 60 min unless stated otherwise.

279

(i) ELISA for screening hybridoma culture supernatants. Microtiter plates were coated with 100 pil of an optimum concentration of E. coli Moon 1676 ultrasonicate per well. Before use, the coated plates were washed, and then 100 pul of culture supernatant was added to each well. Positive and negative controls were included on each plate. After incubation, the plates were washed and further incubated after addition of 100 pul of an optimum dilution of affinity purified, peroxidase-conjugated rabbit anti-mouse immunoglobulin G (IgG) (H and L chain specific; Zymed Laboratories) per well. Finally, the plates were again washed, and bound conjugate was detected by addition of chromogen and enzyme substrate (100 pl of recrystallized 5-aminosalicylic acid [Aldrich Chemical Co., Inc.] per well diluted to 1 mg/ml in 0.01 M phosphate buffer containing 0.01 M sodium EDTA [pH 5.95], to which 0.005% hydrogen peroxide was added immediately before use [13]). After 30 min of incubation, the A450 of each well was determined with a micro-ELISA reader (MR580; Dynatech).

(ài) ELISA for determination of monoclonal antibody isotype. Each well of a microtiter plate was coated with 100 ,ul of an optimal dilution of rabbit anti-mouse IgGl, IgG2a, IgG2b, IgG3, IgM, or IgA serum (Miles Laboratories, Inc.). Plates were washed, and 100 pul of hybridoma culture supernatant fluid containing monoclonal antibody was added to each well. Plates were incubated and washed, and then 100 pul of an optimal dilution of affinity-purified, peroxidaseconjugated rabbit anti-mouse IgG (H and L chain specific; Zymed) was added to each well. After incubation, plates were washed, and bound conjugate was detected as described above. (iii) Capture ELISA-F41. In the capture ELISA for detection of F41 fimbrial antigen (ELISA-F41), microtiter plates were coated with 100 pil of an optimal concentration of immunoglobulin per well, separated by rivanol and ammonium sulfate (33) from serum of a rabbit hyperimmunized with purified F41 fimbriae. After incubation overnight at 4°C, the coated plates were washed. Test samples were titrated twofold from an initial dilution of 1 in 2 (for E. coli cultures and culture supernatants) or 1 in 3 (for fecal specimens), and 100-pul samples of each dilution were added to duplicate wells. Control samples positive or negative for F41 were included in each series of tests. Plates were incubated, washed, and then further incubated after addition of an optimal dilution (previously determined by titration) of mouse ascitic fluid containing anti-F41 monoclonal antibody. Plates were again incubated and washed. A sample of 100 pul of an optimal dilution (previously determined by titration) of affinity-purified, peroxidase-conjugated rabbit anti-mouse IgG (H and L chain specific, Zymed) was then added to each well. After one more incubation, plates were washed, and bound conjugate was detected as described above. (iv) Capture ELISA-K99. The procedure for the capture ELISA for detection of K99 fimbrial antigen (ELISA-K99) was identical to that used for the capture ELISA-F41, except that rabbit antibody against K99 was substituted for rabbit anti-F41 antibody, and monoclonal antibody against K99 was used instead of anti-F41 monoclonal antibody. PHA procedure. (i) Sensitization of turkey erythrocytes. Turkey erythrocytes were fixed with formaldehyde by the method of Sequeira and Eldridge (34) and then treated with tannic acid, followed by sensitization with an optimal dilution (determined by titration) of appropriate purified fimbrial antigen or ascitic fluid containing monoclonal antibody. Assays for detection of specific antibody against F41 and K99 was designated PHA-F41Ab and PHA-K99Ab, respec-

280

J. CLIN. MICROBIOL.

RAYBOULD ET AL.

tively. Assays for detection of F41 and K99 antigens were designated PHA-F41 and PHA-K99, respectively. (ii) Passive hemagglutination tests. Test samples were titrated 4- or 10-fold from an initial dilution of 1 in 4 or 1 in 10, respectively, in V-well Microtiter plates (75 jxl per well) (Dynatech) with 0.1 M phosphate-buffered saline (pH 7.2) containing 1% heat-inactivated normal turkey serum. A 1% suspension (25 ,Jl) of appropriately sensitized turkey erythrocytes in 0.1 M phosphate-buffered saline (pH 7.2) containing 1% heat-inactivated normal rabbit serum was added to each well. After gentle agitation, the plates were covered and allowed to stand at room temperature for 30 min before endpoint agglutination titers were recorded. Inhibition (blocking) assays. The specificity of monoclonal antibodies for F41 antigen was confirmed by inhibition assays. In the ELISA system developed for screening hybridoma culture supernatants, this was performed by preincubation of either (i) the culture supernatant with purified F41-dG antigen before addition to the coated well or (ii) the coated well with rabbit antibody against purified F41-dG antigen before addition of culture supernatant. In the PHAF41Ab system, this was performed by preincubation of the culture supernatant with purified F41-dG antigen before addition of F41-dG-coated turkey erythrocytes. Controls were also run in each system by substituting purified K99 or K30 antigens and their specific rabbit antibodies for F41-dG antigen and its specific rabbit antibody. RESULTS Standardization of assays for detection of F41-specific antibodies in hybridoma culture supernatants. (i) ELISA. Microtiter plates coated with E. coli Moon 1676 ultrasonicate (diluted 1 in 2,000) were treated with titrations of rabbit antisera against purified F41-dG, K99, and K30 antigens followed by affinity-purified, peroxidase-conjugated goat anti-rabbit IgG (H and L specific) (Zymed), and then chromogen and enzyme substrate, by the ELISA procedure outlined in Materials and Methods. Strong reactivity was obtained in this system with anti-F41 antiserum down to a dilution of 1:20,480, whereas no reactivity was observed with rabbit antiserum against K99 or K30, even at dilutions of 1:20. (à) PHA. Preserved turkey erythrocytes treated with tannic acid were sensitized with a range of concentrations of purified F41-dG, K99, and K30 antigens and then tested for reactivity against titrations of homologous and heterologous hyperimmune rabbit antisera. Turkey erythrocytes sensitized with purified K99 and K30 antigens at concentrations of 2,500 and 1,250 ,ug of antigen per ml of packed cell volume, respectively, exhibited maximum sensitivity and specificity (Table 1). Turkey erythrocytes sensitized with F41-dG antigen agglutinated spontaneously when tested for PHA activity. It has previously been shown that the hemagglutinating activity of F41+ E. coli cells could be abolished by heating the bacteria to 65°C for 30 min or exposing them to 0.5% (wt/vol) formaldehyde at 370C for 4 h (24). Samples of purified F41-dG antigen was therefore heated to 650C for 30 min or exposed to 0.5 (wt/vol) formaldehyde at 37°C for 4 h before being used to sensitize tanned turkey erythrocytes. The latter treatment destroyed both the hemagglutinating activity and antigenicity of the F41-dG antigen, whereas heating to 65°C for 30 min abolished the hemagglutinating activity but did not destroy antigenicity in the PHA system. Turkey erythrocytes sensitized with heat-treated F41-dG

TABLE 1. Sensitivity and specificity of the PHA system with turkey erythrocytes sensitized with purified E. coli antigens against homologous or heterologous rabbit antiserum PHA titer (loglo) with rabbit antiserum Antigen concn on Turkey erythrocytes turkey erythrocytes against: sensitized with: (p.gIMl of pCVa)

Tensitizedrwithro:t F41-dGb K99 K30

a

b

25

2,500 1,250

F41

K99

K30

3 10.0

0.25 >1.0 >1,000.0

>100.0 0.05 >10.0

a In capture ELISA, a reading of -0.05 absorbance units; in PHA, complete agglutination of sensitized turkey erythrocytes.

Samples of each culture were (i) stored at 4°C (culture) or (ii) centrifuged to remove bacterial cells (culture supernatant). Each sample was then assayed for the presence of F41 and K99 antigens with the capture ELISA and PHA systems. Table 3 shows the endpoint titers obtained for each sample in each system. Higher titers were obtained with all samples in PHA-K99 than in capture ELISA-K99. In the assays for F41 antigen, ELISA titers were equal to or greater than PHA titers with F41+ samples but negative (.2) with F41+ samples. False-positive results were obtained with samples containing E. coli cells in the PHA-F41 and with both cultures and supernatants in the PHA-K99. These observations indicate that the ELISA systems exhibit excellent sensitivity and specificity, whereas the PHA systems appear to be oversensitive, resulting in false-positive reactivity. Concentrations of both fimbrial antigens, calculated from ELISA titers and the data in Table 2, are also shown in Table 3 for culture supernatants. Reproducibility of assay systems. The reproducibility of the capture ELISA and PHA systems for detection of F41 and K99 antigens was investigated by testing samples of F41- and K99-positive and -negative culture supernatants with different batches of reagents on different days. Assays were also performed by different people. In 10 replicate assays with each system, the endpoint titers of eight samples tested did not vary by more than 1 twofold dilution. Detection of fimbrial antigens in fecal specimens from scouring and nonscouring calves. In view of the efficiency of the capture ELISA systems at quantitating F41 and K99 antigens in cultures and culture supernatants of E. coli strains, it was decided to evaluate their use in detecting these fimbrial antigens in bovine fecal specimens. Use of the capture ELISAs in this way would provide a simple and rapid assay system to determine whether diarrhea in a calf was due to infection with F41+ or K99+ ETEC. Ten fecal

281

TABLE 4. Detection of K99 and F41 fimbrial antigens in fecal specimens from scouring and nonscouring calves by capture ELISA A4" in monoclonal antibody capture ELISA (1:3 dilution)

Source of fecal specimen

Scouring calf no.: 30 68 75 117 125 130

Capture ELISA-K99

0.169 0.386 0.076 0.193 0.049 0.078

Capture ELISA-F41

0.164 0.292 0.087 0.164 0.045 0.083

0.101 0.025 0.010 0.022 0.105 0.102 o

o

0.002 0.001 0

0.135

Nonscouring calf no.: 87 100 153 176

o

o

O

O

°

O

O

0.010

0.001 0.003 0.003

E. coli B41 cells

0.191

0.187

0.129

0.116 0.023 0.013 0.017 0.092 0.095

specimens were tested, of which six were from scouring calves and four were from nonscouring calves. K99+ E. coli isolates were identified by culture and serology in all six specimens from scouring calves. In each of these cases, 8 to 10 of the colonies tested were agglutinated by K99-specific rabbit antiserum. Fecal specimens from nonscouring calves were negative for K99+ E. coli when tested in the same way. For capture ELISAs, each specimen was diluted 1 in 3 with sterile saline, heat inactivated, and then centrifuged to sediment particulate matter. E. coli B41 cells at a concentration of 5 x 105 cells per ml were treated in an identical manner as a positive control. Samples of each supernatant were then tested in duplicate in both capture ELISAs. Fecal specimens from calves 68, 75, and 117 contained K99+ ETEC, the specimen from calf 125 contained ETEC expressing F41 and probably K99, and the specimens from calves 30 and 130 contained ETEC expressing both K99 and F41 fimbrial antigens (Table 4); fecal specimens from nonscouring calves 87, 100, 153, and 176 did not contain ETEC expressing either of these fimbrial antigens. DISCUSSION Capture ELISAs for detection of ETEC fimbrial antigens have previously been described. Milîs et al. (19) used a

TABLE 3. Detection of fimbrial antigens in different ETEC strains by the capture ELISA and PHA systems Capture ELISA titer (concn of fimbrial antigen [p.g/ml])

PHA titer

Fimbrial antigen present

Detection

F41+ K99+

F41 K99

4,096

8 (0.4) 1,024 (51.2)

16,384

16 16,384

Bi316-42

F41K99-

F41 K99