Zbl. Bakt. 277, 39-48 (1992) © Gustav Fischer Verlag, StuttgartlNew York
Monoclonal Antibody Based Capture ELISAIELIFA for the Detection of Chlamydia psittaci in Veterinary Clinical Specimens DETLEF THIELE, MARIJA KARO, and HARTMUT KRAUSS Institut fur Hygiene und Infektionskrankheiten der Tiere, Arbeitsgruppe Zoonosen, JustusLiebig-Universitat Ciefsen, 6300 GieBen, Germany
With 3 Figures· Received September 12, 1991 . Revision received November 11, 1991 . Accepted December 20, 1991
Summary A capture ELISAJELIFA system based on monoclonal capture and biotinylated monoclonal detection antibody with specificity for an epitope on chlamydial lipopolysaccharide (LPS) is described. The assay is fast, specific and detects a minimum dose of approximately 2000 chlamydiaI particles (c. psittaci). Unlike other commercially available test kits, it comprises a proteinase K treatment to exclude non-specific binding of antibodies to Fe receptors. In contrast to the sophisticated and cumbersome isolation procedures, even nonspecialized laboratories may use this assay system for the investigation of clinical samples of different origin for the presence of C. psittaci within a short period of time. Due to the genus specificity of the monoclonal antibody used C. trachomatis and C. pneumoniae can be detected as well.
Zusammenfassung Es wird ein "Capture-ELISAJELIFA-System" auf der Basis eines biotinylierten monoklonalen Antikorpers mit Spezifitat fur ein Epitop auf dem Lipopolysaccharid (LPS) von Chlamydien beschrieben. Der Test ist leicht durchfuhrbar, spezifisch und weist minimal bis zu 2000 Chlamydien (c. psittaci) nacho Anders als bei kommerziell erhaltlichen Test-Kits wird hier durch eine Proreinase-KBehandlung die unspezifische Bindung von Antikorpern iiber das Fe-Snick ausgeschlossen. Irn Gegensatz zur aufwendigen und schwierigen Zellkultur-Methode, die nur spezialisierten Laboratorien varbehalten bleibt, kann der Test von den meisten Labaratorien problemlos durchgefuhrt werden. Die Methode ist geeignet fur klinische Proben unterschiedlichen Ursprungs zur Untersuchung auf C. psittaci; sie kann innerhalb weniger Stunden durchgefuhrr werden. Auf Grund der Genus-Spezifitat des verwendeten monoklonalen Antikorpers konnen auch C. trachomatis und C. pneumoniae nachgewiesen werden.
40
D. Thiele, M. Karo, and H. Krauss
Introduction For a direct detection of chlamydiae, inoculation of cell cultures is still widely regarded as the best choice due to its high specificity; it is therefore often addressed as the "gold standard". However, culture as a diagnostic tool has also its disadvantages; it remains reserved for specialized laboratories with adequate equipment. Another important requirement for the use of cell cultures is that the organisms must have retained their infectivity. In order to conserve infectivity, special swabs and transport media are used and after collection, specimens should be processed as soon as possible, at best without freezing. In addition, cell cultures do not detect chlamydiaI infection in the lower genital tract in the presence of neutralizing cervical antibody (7). Several commercially available products have been introduced for direct detection of c. trachomatis antigens. These tests are based on polyclonal or monoclonal antibodies. Most of these test systems can be used to detect C. trachomatis only, e.g. Mikro Trak™ (Syva, USA), Chlamyset'" (Orion, Finland), Monofluo's (Diagnostic Pasteur, France) and PACE™ (Ccn-Probc, USA). Due to their genus specificity, some of them have been reported to detect both C. trachomatis and C. psittaci, e.g. Chlamydiazyme'P (Abbott, USA), IDEIA1M (Boots-Celltech, UK), Testl'ackv (Abbott, USA) and Clearview (Unipath, UK) but have only been released for the detection of C. trachomatis. Another important fact is that some of them may only be used for certain samples, i.e. conjunctival, cervical, or urethral swabs and urine. None of these tests have been described to be adequate for the analysis of samples like milk or feces. Additionally, a wide variety of cross reactivity has been demonstrated for the Chlamydiazyme test (1, 3, 13). Even with the monoclonal antibody based MicroTrak and IDEIA, cross reactivities have been found (4, 12, 14). Non-specific binding via the Fe portion of the antibody to Staphylococcus aureus protein A has been described for the MicroTrak and Imagen test systems (5, 6, 9). As a result, a "Chlamydia Blocking Reagent test" (Abbott, USA) has been introduced as a confirmatory test for the Chlamydiazyme and designed to strengthen the specificity and sensitivity of the former test (10). Our aim was to develop a test system mainly for the detection of C. psittaci in different samples like milk, feces, tissues and swabs taken from various sites. In addition, the test system should be based on monoclonal antibody with broad reactivity to chlamydiae, e.g. recognition of all three species including all isolates. Other aims were to set up a system which avoided non-specific binding to Fe receptors, with no cross reactivity to other organisms possibly found in the samples, as well as a low detection limit. Finally, the antibody should recognize an epitope which is found repeatedly on chlamydia for use of the antibody as capture as well as detection antibody in the antigen capture test, thus avoiding competitive effects.
Material and Methods Organisms. Chlamydiae were used as stocks harvested from organisms propagated in infected cell cultures using Buffalo Green Monkey (BGM) cells (Flow Lab., Meckenheim, Germany), or alternatively, after propagation in the allantoic cavity of embryonated chicken eggs. The following strains/isolates were used: Two standard isolates of C. psittaci (immunotype I: B-577-lamb, ATCC Vr-656; immunotype II: LW-613-calf, kindly supplied by]. Storz, Louisiana State University, Baton Rouge, USA)
ELISNELIFA for the Detection of Chlamyd ia psittaci
41
10 isolates of C. psittaci recovered fro m clinical material in our laboratory (1'1163/7 1cattle , Zll21 -cattle, 53-sheep, IS-goat, Z10 -cat, H4- rabbit, Z432-guineapig, 4043-duck, 1270-parrot, 1904-budgerigar ) 15 strains of C. trachomatis (type A: SA-I, B: TW-S, Ba: A-7, C: UW-1, D: IC-CAL-8, E: DK-20, F: MR C-301 , G: IOL-238, H: UW-4, I: UW 12, J: UW 36, K: UW 31 , L1: 440-L, L2: 434-B, U: 404-L ; all strains were kindly supplied by the Institute of Ophthalmology, University of London). One strain of C. pneumon iae (TWAR TW 183), supplied by the Washington Research Foundation, Washington, DC. For testing cross reactivity and non specific binding to Fc receptors, the following bacte rial isolates were used: A ch rom ob acter haem olysan s, A cinetoba cter calcoaceticus var. ani tratus, Bordetella bronchiseptica, Corynebacterium pyogen es, Cox iella burnetii, Escherichia coli, Listeria monocyt ogenes [ and IVb, Pasteurella multocida, Salm onella typhimurium group B, Staphylo coccus (S.) aureus, S. aureus Cowan I and Wood 46, S. hyicus, Stre ptococcus (St.) faecalis, St. zooepidem icus, and Yersinia entero colitica. All these bacteria were available from the collection of our institute. Additionally, Bart onella bacilliformis was obtained from J. Knobloch, Bernhard-Nocht-Institut, Hamburg, Germany and Legion ella pneumophila from H. G. Sch iefer, Institut fur Medizini sche Mikrobiologie, GieRen, Germany. Partial purification of C. psitt aci from cell culture. Supernatants from infected BGM cell cultures containing chlam ydiae were sonicated. Two low-speed centrifugations (300 x g, 10 min) were used to pellet cell debris. The superna tants were recentr ifuged at 20 ,000 x g, 30 min. The resulting pellet contai ning chlamydiae was washed twice with double-distilled water , dialysed exten sively against doubl e distilled water and lyophilized . Finally, smears from resuspended mate rial ( I mg/ml) were stained with Gimenez stain and examined for contaminating BGM mat erial. Immunization . Female Balble mice 6 to 10 weeks old were infected intraperitoneally with 6 x 10 6 inclusion-forming units of the C. psittaci 1904 (parrot) isolate resuspended in pho sphate-buffered saline (PBS) and har vested from infected embryonated chicken eggs. Mice were reinfecred at day 21 and fusion was performed 4 days later. Fusion and cloning. Briefly, spleen cell suspensions cont aining stimulated B lymphocytes from infected Balblc mice were fused with cells of the mouse myeloma cell line X63 Ag8.653 (8). Growing hybrids were tested for monoclonal anti body pr odu ction using an ELISA system. To assure monoclonal ity, hybrid s of interest were cloned 3 x using the limiting dilution techniqu e with rhyrnocytes as feeder layer (11). Results from limiting dilution were anal ysed using the Poisson statistical anal ysis for repetiti ve subcloning (2). EL ISA for detection and chara cterizatio n of antigen-specific m on oclonal antibodies . Antibody-secreting hybr idom a cells and ant ibody production were monitored by indirect ELISA. Briefly, wells on rnicrotiter Immulon M 129A plates (Greiner, Ntirtingen, Germany) were coated with C. psittaci 1904 organis ms originating from infected BGM cell cultures. Using chlamydiaI antigen propa gated in cell cultures for this purpose instead of antigen propagated in embryonated chicken eggs which was applied for immun ization was advantageous for not codetecting hybrid s producing antibodies against cont aminating egg proteins. As negative controls, wells were incubated either with RPMI 1640 medium or "conditioned medium" (supernatant from prop agated myeloma cells). Serum harvested from mice used for immun ization and as spleen cell donors served as positi ve control. The same ELISA format was used to determin e class, subclass and light chain type of monoclonal antibodies using the "Mouse Typer Kit" and " Mouse Typer Panel" (both BioRad, Munich, Germany) containing specific rabb it ant isera to mouse IgG, IgG1, IgG2a, IgG2b, IgG3, IgM, IgA, kappa- chain and lambda chain, and goat anti -rabbit IgG(H + L) peroxidase-labelled conjugate. H ybridoma maintenan ce and p rodu ction of antibodies. After establ ishing stable antibody-producing hybridomas in RPMl 1640 medium cont aining hypox anth ine and thymidine (omitt ing amin opte rin at this state) and 10% fetal calf serum, the cells were tested
42
D. Thiele, M. Karo, and H. Krauss
for the production of monoclonal antibodies under serum-free RMPI medium supplemented with 2% Ultroser HY (GIBCO, Eggenstein, Germany). Using this supplement culture, supernatants were free of foreign immunoglobulins such as bovine immunoglobulin. Characterization and purification of monoclonal antibodies. To further characterize established monoclonal antibodies, their binding characteristics were also determined in the direct and indirect immunofluorescence, dot test, indirect immunoperoxidase and immunoblot techniques. The monoclonal antibodies were purified on a protein A-Sepharose CL-4B column (Pharmacia, Freiburg, Germany), dialysed against double-distilled water and lyophilysed. Biotinylation of monoclonal antibodies. 12 mg of purified and lyophilyzed monoclonal antibody were dissolved in 500 !-II carbonate buffer (0.1 M NaHC0 3, 0.1 M Na2C03; pH 8.5).50 !-II ofdimethylsuiphoxide (DMSO) containing 1.0 mg of D-Biotinyl-E-aminocaproic acid-N-hydroxysuccinimide ester (Biotin-X-NHS) were added and incubated for 12 h while shaking at room temperature (RT). The reaction mix was extensively dialysed against 11 of carbonate buffer for 16 h at 4°C, with three changes of the buffer. The dialysate was then centrifuged (8000 g, 10 min, RT), the supernatant supplemented with 2% BSA for cryoprotection and shock-frozen in liquid nitrogen before storing at -20 0C. Antigen capture test. Several commercially available ELISA microtiter plates were tested and compared. Best results could be achieved with ELISA-F plates Immulon M129A made from polystyrene with a medium binding capacity for proteins (Greiner, Niirtingen, Germany). Plates were freshly coated with capture antibody 100 ul/well (1 mg/ml, diluted 1: 5000 in TRIS buffered saline [TBS]: 10 mM TRIS, 100 mM NaCl, pH 8.5) and incubated for 16 h at 4°C. Remaining binding sites were blocked with TBS containing 10 mM TRIS, 150 mM NaCl, 0.5% Tween 20, pH 8.0, supplemented with 1% fish gelatine (Sigma, Deisenhofen, Germany) for 15 min at RT. Plates were washed three times with 250 !-II washing buffer (10 mM Tris, 150 mM NaCl, 0.5% Tween 20, pH 8.0); each washing step requiring 5 min throughout the whole protocol, using always the same buffer. Samples were added in a volume of 100 !-II and incubated for 2 h at RT. After three more washings the monoclonal biotinylated detection antibody was added (100 !-II, 1: 10.000 in washing buffer) and incubated for 30 min at RT. Optimal concentration for each immunoglobulin compound was initially determined with a chequerboard ELISA titration. After three more washings, streptavidin, conjugated with alkaline phosphatase or peroxidase (Boehringer, Mannheim, Germany) was added (100 !-II, 1: 10.000 in washing buffer) and incubated for 30 min. Substrate was added, after three final washes. For streptavidin conjugated with peroxidase, 3.3'-5.5'-tetramethylbenzidine (TMB, Fluka, Neu-Ulm, Germany) was used as substrate (stock solution J: 1 M sodium acetate, adjusted to pH 5.8 with 0.1 citric acid, stored at 4°C; stock solution II: 10 mg/ml TMB in DMSO (Sigma, Deisenhofen, Germany), stored at 4°C. Working solution: 2 ml of stock solution J were mixed with 0.2 ml stock solution II and 20 ml double-distilled water. 20 !-II H 202 (6%) were added and the solution adjusted to 37°C). The enzyme reaction proceeded for 30 min and was stopped with 3 M H 2S0 4 (free of Fe ions). The optical density was measured at 450 nm in a Titertek Multiscan ELISA reader (Flow, Meckenheim, Germany). For the enzyme-linked immunofluorescent assay (ELIFA), streptavidin conjugated with alkaline phosphatase 4-methylumbelliferylphosphate (4-MUP; Sigma, Deisenhofen, Germany) was used as substrate (2 mg of 4-MUP were dissolved in JO ml diethanolamine buffer, always freshly prepared and adjusted to 37°C. Diethanolamine buffer: 97.0 ml diethanolamine - Merck, Darmstadt, Germany, 0.2 g NaN3, 0.01 g MgCb X 6 H 20, 800.0 ml double-distilled water; adjusted to pH 9.8 with HCl, double distilled water added to 1 1, stored at 4°C and protected from light). Fluorescence was measured after 60 min using a Titertek Fluoroscan II reader (Flow, Meckenheim, Germany) with excitation wavelength of 355 nm and emission wavelength of 460 nm. Samples were consequently tested in duplicates. A standard titration of C. psittaci isolate 1904 (1 mg/ml; 1: 1000-1 : 1.024.000) was always tested in parallel. Samples containing buffer instead of antigen served as negative controls.
ELISNELIFA for the Detection of Chlamydia psittaci
43
Samples. 129 routine clinical samples sent to our laboratory were tested in parallel with the Capture-ELISNELIFA system and cell culture isolation (BGM cells). These samples were obtained from a variety of animal species including cattle (47x), sheep (28x), goat (Zx), cat (29x), dog (14x), swine (3x), parrot (1X), chicken (2x), fruit bat (2x) and man (1 x). Samples consisted of swabs taken from different sites (eye, cervix, trachea, larynx, pharynx, nose, milking machine) and organs (liver, spleen, cornea, placenta, testis, lung and material from aborted fetus), or consisted of milk, semen, feces, wash fluid from lung, trachea, and joint, birth fluids and aspirate from prostate and a prostatic cyst. Preparation of antigen/samples. Analytical and clinical samples were treated with proteinase K prior to testing. Proteinase K (Sigma, Deisenhofen, Germany) non-self-digested was dissolved in 10 mM TRIS with 1 mM CaCl 2 (addition of Ca 2 + to proteinase K was carried out to degrade proteins that are commonly resistant to this enzyme) at a concentration of 400 ug/ml. The same volume of proteinase K was added to analytical and clinical samples (to give a final concentration of 200 ug/rnl) and incubated for 30 min at 56°C. For inactivation of the enzyme, the reaction mixture was heated to 100 °C for 15 min. Statistical analysis. ELIFA results with analytical samples containing 1 mg/ml of C. psittaci antigen were used to determine several statistical parameters of the test system. The following parameters were determined: detectability (cut off) and sensitivity. Detectability of the assay was defined as the lowest concentration of antigen exceeding the zero-dose precision and was estimated using the following equation: xsD . . t n_ 1; (1; one sided Detectabiliry = Absorbance.g., dose + yin Sensitivity was defined as the responsiveness of the assay to changes in concentration of the substance tested. Therefore, on a sigmoid dose-response curve, sensitivity decreases at either end of the curve. Sensitivity was estimated using the following equation: Sensitivity = dR/dC; where R = ELIFA units; C = reciprocal dilution (lOglO)'
Comparison of the Capture ELISA/ELIFA with commercially available Chlamydiaxymev and IDEIA™ test kits. Detecrabiliry and specificity of the new capture ELISNELIFA was directly compared with the Chlamydiazyme and !DEIA test systems using dilutions of a 1 mg/ml C. psittaci 1904 isolate stock solution and a panel of different bacteria (see material and methods) at a concentration of 10 6/ml as analytical samples. Tests were performed exactly as indicated by the manufacturers.
Results
Antibodies 16 hybridoma cell lines secreting antibodies directed against genus-specific chlamydial epitopes could be established. Nine of these hybridomas produced antibodies of the IgG2a and five those of IgG3 subclass, whereas two produced antibodies of the IgM type. All monoclonal antibodies produced carried kappa light chains. All clones of interest could be easily adapted to produce monoclonal antibodies in serum-free RPMI 1640 medium supplemented with "Ultroser HY". Cell culture supernatants contained antibodies in the 5 to 30 ug/rnl range. All IgG antibodies could be purified on protein A-Sepharose CL-4B with only insignificant loss of activity. Serum-free RPMI 1640 medium supplemented with "Ultroser HY" as control never yielded any detectable amount of immunoglobulins when applied to the protein A affinity chromatography column.
44
D. Thiele, M. Karo, and H. Krauss Binding characterization using different test systems
Preliminary binding studies using direct and indirect immunofluorescence, a dot test and the indirect immunoperoxidase test were carried out to evaluate the monoclonal antibodies. One of those recognizing all isolates and strains from the three Chlamydia species - Mab 22 (IgG3) - was finally chosen to set up the Capture ELISAJELIFA system. This monoclonal antibody was also tested with the immunoblotting procedure revealing reactivity with the genus-specific C. psittaci lipopolysaccharide (Fig. 1).
- - 92,5 * 66,2
•. •
LPS~
1
2
45,5 ..... 31,0
II • 3 4
21,5 14,4
5
Fig. 1. Western blots of C. psittaci, isolate 1904. Lane 1: SDS-PAGE gel stained with silver stain revealing protein and LPS bands. Lanes 2-5: Immobilon-P membranes. Lane 2: Protein bands stained with amido black. Lane 3: Immunoblot with Mab 22 purified from supernatants of cell cultures propagated in serum-free medium. Lane 4: Immunoblot with hyperimmune serum from BALB/c mouse. Lane 5: Marker proteins (MW in kDa).
Capture ELISA/ELIFA
Using this test system based on monoclonal antibody Mab 22 as capture and biotinylated detection antibody, all chlamydiaI strains and isolates were clearly recognized. All bacteria tested (20 different strains as negative control) did not react in this assay when using a concentration of 10 6 particles/m!. Using a stock solution of C. psittaci strain 1904 (1 mg/ml), the assay still gave a positive signal (cut-off) when applying a dose of 100 f.tl of a 1: 320,000 dilution (Figs.2, 3). This amount contained 312 pg of C. psittaci antigen, or approximately
ELISAIELIFA for the Detection of Chlamydia psittaci
45
2000 particles. The detection limit was the same when using streptavidin conjugated with peroxidase (ELISA) or streptavidin conjugated with alkaline phosphatase (ELIFA). Sensitivity defined as the responsiveness of the assay to changes in concentration of the suspension tested was maximal in the 1: 10,000 dilution range of the antigen stock suspension. Capture ELISA
2500
2000 Ul
~
c
1500
:J
Monoclonal Antibody Based Capture ELISAIELIFA for the Detection of Chlamydia psittaci in Veterinary Clinical Specimens DETLEF THIELE, MARIJA KARO, and HARTMUT KRAUSS Institut fur Hygiene und Infektionskrankheiten der Tiere, Arbeitsgruppe Zoonosen, JustusLiebig-Universitat Ciefsen, 6300 GieBen, Germany
With 3 Figures· Received September 12, 1991 . Revision received November 11, 1991 . Accepted December 20, 1991
Summary A capture ELISAJELIFA system based on monoclonal capture and biotinylated monoclonal detection antibody with specificity for an epitope on chlamydial lipopolysaccharide (LPS) is described. The assay is fast, specific and detects a minimum dose of approximately 2000 chlamydiaI particles (c. psittaci). Unlike other commercially available test kits, it comprises a proteinase K treatment to exclude non-specific binding of antibodies to Fe receptors. In contrast to the sophisticated and cumbersome isolation procedures, even nonspecialized laboratories may use this assay system for the investigation of clinical samples of different origin for the presence of C. psittaci within a short period of time. Due to the genus specificity of the monoclonal antibody used C. trachomatis and C. pneumoniae can be detected as well.
Zusammenfassung Es wird ein "Capture-ELISAJELIFA-System" auf der Basis eines biotinylierten monoklonalen Antikorpers mit Spezifitat fur ein Epitop auf dem Lipopolysaccharid (LPS) von Chlamydien beschrieben. Der Test ist leicht durchfuhrbar, spezifisch und weist minimal bis zu 2000 Chlamydien (c. psittaci) nacho Anders als bei kommerziell erhaltlichen Test-Kits wird hier durch eine Proreinase-KBehandlung die unspezifische Bindung von Antikorpern iiber das Fe-Snick ausgeschlossen. Irn Gegensatz zur aufwendigen und schwierigen Zellkultur-Methode, die nur spezialisierten Laboratorien varbehalten bleibt, kann der Test von den meisten Labaratorien problemlos durchgefuhrt werden. Die Methode ist geeignet fur klinische Proben unterschiedlichen Ursprungs zur Untersuchung auf C. psittaci; sie kann innerhalb weniger Stunden durchgefuhrr werden. Auf Grund der Genus-Spezifitat des verwendeten monoklonalen Antikorpers konnen auch C. trachomatis und C. pneumoniae nachgewiesen werden.
40
D. Thiele, M. Karo, and H. Krauss
Introduction For a direct detection of chlamydiae, inoculation of cell cultures is still widely regarded as the best choice due to its high specificity; it is therefore often addressed as the "gold standard". However, culture as a diagnostic tool has also its disadvantages; it remains reserved for specialized laboratories with adequate equipment. Another important requirement for the use of cell cultures is that the organisms must have retained their infectivity. In order to conserve infectivity, special swabs and transport media are used and after collection, specimens should be processed as soon as possible, at best without freezing. In addition, cell cultures do not detect chlamydiaI infection in the lower genital tract in the presence of neutralizing cervical antibody (7). Several commercially available products have been introduced for direct detection of c. trachomatis antigens. These tests are based on polyclonal or monoclonal antibodies. Most of these test systems can be used to detect C. trachomatis only, e.g. Mikro Trak™ (Syva, USA), Chlamyset'" (Orion, Finland), Monofluo's (Diagnostic Pasteur, France) and PACE™ (Ccn-Probc, USA). Due to their genus specificity, some of them have been reported to detect both C. trachomatis and C. psittaci, e.g. Chlamydiazyme'P (Abbott, USA), IDEIA1M (Boots-Celltech, UK), Testl'ackv (Abbott, USA) and Clearview (Unipath, UK) but have only been released for the detection of C. trachomatis. Another important fact is that some of them may only be used for certain samples, i.e. conjunctival, cervical, or urethral swabs and urine. None of these tests have been described to be adequate for the analysis of samples like milk or feces. Additionally, a wide variety of cross reactivity has been demonstrated for the Chlamydiazyme test (1, 3, 13). Even with the monoclonal antibody based MicroTrak and IDEIA, cross reactivities have been found (4, 12, 14). Non-specific binding via the Fe portion of the antibody to Staphylococcus aureus protein A has been described for the MicroTrak and Imagen test systems (5, 6, 9). As a result, a "Chlamydia Blocking Reagent test" (Abbott, USA) has been introduced as a confirmatory test for the Chlamydiazyme and designed to strengthen the specificity and sensitivity of the former test (10). Our aim was to develop a test system mainly for the detection of C. psittaci in different samples like milk, feces, tissues and swabs taken from various sites. In addition, the test system should be based on monoclonal antibody with broad reactivity to chlamydiae, e.g. recognition of all three species including all isolates. Other aims were to set up a system which avoided non-specific binding to Fe receptors, with no cross reactivity to other organisms possibly found in the samples, as well as a low detection limit. Finally, the antibody should recognize an epitope which is found repeatedly on chlamydia for use of the antibody as capture as well as detection antibody in the antigen capture test, thus avoiding competitive effects.
Material and Methods Organisms. Chlamydiae were used as stocks harvested from organisms propagated in infected cell cultures using Buffalo Green Monkey (BGM) cells (Flow Lab., Meckenheim, Germany), or alternatively, after propagation in the allantoic cavity of embryonated chicken eggs. The following strains/isolates were used: Two standard isolates of C. psittaci (immunotype I: B-577-lamb, ATCC Vr-656; immunotype II: LW-613-calf, kindly supplied by]. Storz, Louisiana State University, Baton Rouge, USA)
ELISNELIFA for the Detection of Chlamyd ia psittaci
41
10 isolates of C. psittaci recovered fro m clinical material in our laboratory (1'1163/7 1cattle , Zll21 -cattle, 53-sheep, IS-goat, Z10 -cat, H4- rabbit, Z432-guineapig, 4043-duck, 1270-parrot, 1904-budgerigar ) 15 strains of C. trachomatis (type A: SA-I, B: TW-S, Ba: A-7, C: UW-1, D: IC-CAL-8, E: DK-20, F: MR C-301 , G: IOL-238, H: UW-4, I: UW 12, J: UW 36, K: UW 31 , L1: 440-L, L2: 434-B, U: 404-L ; all strains were kindly supplied by the Institute of Ophthalmology, University of London). One strain of C. pneumon iae (TWAR TW 183), supplied by the Washington Research Foundation, Washington, DC. For testing cross reactivity and non specific binding to Fc receptors, the following bacte rial isolates were used: A ch rom ob acter haem olysan s, A cinetoba cter calcoaceticus var. ani tratus, Bordetella bronchiseptica, Corynebacterium pyogen es, Cox iella burnetii, Escherichia coli, Listeria monocyt ogenes [ and IVb, Pasteurella multocida, Salm onella typhimurium group B, Staphylo coccus (S.) aureus, S. aureus Cowan I and Wood 46, S. hyicus, Stre ptococcus (St.) faecalis, St. zooepidem icus, and Yersinia entero colitica. All these bacteria were available from the collection of our institute. Additionally, Bart onella bacilliformis was obtained from J. Knobloch, Bernhard-Nocht-Institut, Hamburg, Germany and Legion ella pneumophila from H. G. Sch iefer, Institut fur Medizini sche Mikrobiologie, GieRen, Germany. Partial purification of C. psitt aci from cell culture. Supernatants from infected BGM cell cultures containing chlam ydiae were sonicated. Two low-speed centrifugations (300 x g, 10 min) were used to pellet cell debris. The superna tants were recentr ifuged at 20 ,000 x g, 30 min. The resulting pellet contai ning chlamydiae was washed twice with double-distilled water , dialysed exten sively against doubl e distilled water and lyophilized . Finally, smears from resuspended mate rial ( I mg/ml) were stained with Gimenez stain and examined for contaminating BGM mat erial. Immunization . Female Balble mice 6 to 10 weeks old were infected intraperitoneally with 6 x 10 6 inclusion-forming units of the C. psittaci 1904 (parrot) isolate resuspended in pho sphate-buffered saline (PBS) and har vested from infected embryonated chicken eggs. Mice were reinfecred at day 21 and fusion was performed 4 days later. Fusion and cloning. Briefly, spleen cell suspensions cont aining stimulated B lymphocytes from infected Balblc mice were fused with cells of the mouse myeloma cell line X63 Ag8.653 (8). Growing hybrids were tested for monoclonal anti body pr odu ction using an ELISA system. To assure monoclonal ity, hybrid s of interest were cloned 3 x using the limiting dilution techniqu e with rhyrnocytes as feeder layer (11). Results from limiting dilution were anal ysed using the Poisson statistical anal ysis for repetiti ve subcloning (2). EL ISA for detection and chara cterizatio n of antigen-specific m on oclonal antibodies . Antibody-secreting hybr idom a cells and ant ibody production were monitored by indirect ELISA. Briefly, wells on rnicrotiter Immulon M 129A plates (Greiner, Ntirtingen, Germany) were coated with C. psittaci 1904 organis ms originating from infected BGM cell cultures. Using chlamydiaI antigen propa gated in cell cultures for this purpose instead of antigen propagated in embryonated chicken eggs which was applied for immun ization was advantageous for not codetecting hybrid s producing antibodies against cont aminating egg proteins. As negative controls, wells were incubated either with RPMI 1640 medium or "conditioned medium" (supernatant from prop agated myeloma cells). Serum harvested from mice used for immun ization and as spleen cell donors served as positi ve control. The same ELISA format was used to determin e class, subclass and light chain type of monoclonal antibodies using the "Mouse Typer Kit" and " Mouse Typer Panel" (both BioRad, Munich, Germany) containing specific rabb it ant isera to mouse IgG, IgG1, IgG2a, IgG2b, IgG3, IgM, IgA, kappa- chain and lambda chain, and goat anti -rabbit IgG(H + L) peroxidase-labelled conjugate. H ybridoma maintenan ce and p rodu ction of antibodies. After establ ishing stable antibody-producing hybridomas in RPMl 1640 medium cont aining hypox anth ine and thymidine (omitt ing amin opte rin at this state) and 10% fetal calf serum, the cells were tested
42
D. Thiele, M. Karo, and H. Krauss
for the production of monoclonal antibodies under serum-free RMPI medium supplemented with 2% Ultroser HY (GIBCO, Eggenstein, Germany). Using this supplement culture, supernatants were free of foreign immunoglobulins such as bovine immunoglobulin. Characterization and purification of monoclonal antibodies. To further characterize established monoclonal antibodies, their binding characteristics were also determined in the direct and indirect immunofluorescence, dot test, indirect immunoperoxidase and immunoblot techniques. The monoclonal antibodies were purified on a protein A-Sepharose CL-4B column (Pharmacia, Freiburg, Germany), dialysed against double-distilled water and lyophilysed. Biotinylation of monoclonal antibodies. 12 mg of purified and lyophilyzed monoclonal antibody were dissolved in 500 !-II carbonate buffer (0.1 M NaHC0 3, 0.1 M Na2C03; pH 8.5).50 !-II ofdimethylsuiphoxide (DMSO) containing 1.0 mg of D-Biotinyl-E-aminocaproic acid-N-hydroxysuccinimide ester (Biotin-X-NHS) were added and incubated for 12 h while shaking at room temperature (RT). The reaction mix was extensively dialysed against 11 of carbonate buffer for 16 h at 4°C, with three changes of the buffer. The dialysate was then centrifuged (8000 g, 10 min, RT), the supernatant supplemented with 2% BSA for cryoprotection and shock-frozen in liquid nitrogen before storing at -20 0C. Antigen capture test. Several commercially available ELISA microtiter plates were tested and compared. Best results could be achieved with ELISA-F plates Immulon M129A made from polystyrene with a medium binding capacity for proteins (Greiner, Niirtingen, Germany). Plates were freshly coated with capture antibody 100 ul/well (1 mg/ml, diluted 1: 5000 in TRIS buffered saline [TBS]: 10 mM TRIS, 100 mM NaCl, pH 8.5) and incubated for 16 h at 4°C. Remaining binding sites were blocked with TBS containing 10 mM TRIS, 150 mM NaCl, 0.5% Tween 20, pH 8.0, supplemented with 1% fish gelatine (Sigma, Deisenhofen, Germany) for 15 min at RT. Plates were washed three times with 250 !-II washing buffer (10 mM Tris, 150 mM NaCl, 0.5% Tween 20, pH 8.0); each washing step requiring 5 min throughout the whole protocol, using always the same buffer. Samples were added in a volume of 100 !-II and incubated for 2 h at RT. After three more washings the monoclonal biotinylated detection antibody was added (100 !-II, 1: 10.000 in washing buffer) and incubated for 30 min at RT. Optimal concentration for each immunoglobulin compound was initially determined with a chequerboard ELISA titration. After three more washings, streptavidin, conjugated with alkaline phosphatase or peroxidase (Boehringer, Mannheim, Germany) was added (100 !-II, 1: 10.000 in washing buffer) and incubated for 30 min. Substrate was added, after three final washes. For streptavidin conjugated with peroxidase, 3.3'-5.5'-tetramethylbenzidine (TMB, Fluka, Neu-Ulm, Germany) was used as substrate (stock solution J: 1 M sodium acetate, adjusted to pH 5.8 with 0.1 citric acid, stored at 4°C; stock solution II: 10 mg/ml TMB in DMSO (Sigma, Deisenhofen, Germany), stored at 4°C. Working solution: 2 ml of stock solution J were mixed with 0.2 ml stock solution II and 20 ml double-distilled water. 20 !-II H 202 (6%) were added and the solution adjusted to 37°C). The enzyme reaction proceeded for 30 min and was stopped with 3 M H 2S0 4 (free of Fe ions). The optical density was measured at 450 nm in a Titertek Multiscan ELISA reader (Flow, Meckenheim, Germany). For the enzyme-linked immunofluorescent assay (ELIFA), streptavidin conjugated with alkaline phosphatase 4-methylumbelliferylphosphate (4-MUP; Sigma, Deisenhofen, Germany) was used as substrate (2 mg of 4-MUP were dissolved in JO ml diethanolamine buffer, always freshly prepared and adjusted to 37°C. Diethanolamine buffer: 97.0 ml diethanolamine - Merck, Darmstadt, Germany, 0.2 g NaN3, 0.01 g MgCb X 6 H 20, 800.0 ml double-distilled water; adjusted to pH 9.8 with HCl, double distilled water added to 1 1, stored at 4°C and protected from light). Fluorescence was measured after 60 min using a Titertek Fluoroscan II reader (Flow, Meckenheim, Germany) with excitation wavelength of 355 nm and emission wavelength of 460 nm. Samples were consequently tested in duplicates. A standard titration of C. psittaci isolate 1904 (1 mg/ml; 1: 1000-1 : 1.024.000) was always tested in parallel. Samples containing buffer instead of antigen served as negative controls.
ELISNELIFA for the Detection of Chlamydia psittaci
43
Samples. 129 routine clinical samples sent to our laboratory were tested in parallel with the Capture-ELISNELIFA system and cell culture isolation (BGM cells). These samples were obtained from a variety of animal species including cattle (47x), sheep (28x), goat (Zx), cat (29x), dog (14x), swine (3x), parrot (1X), chicken (2x), fruit bat (2x) and man (1 x). Samples consisted of swabs taken from different sites (eye, cervix, trachea, larynx, pharynx, nose, milking machine) and organs (liver, spleen, cornea, placenta, testis, lung and material from aborted fetus), or consisted of milk, semen, feces, wash fluid from lung, trachea, and joint, birth fluids and aspirate from prostate and a prostatic cyst. Preparation of antigen/samples. Analytical and clinical samples were treated with proteinase K prior to testing. Proteinase K (Sigma, Deisenhofen, Germany) non-self-digested was dissolved in 10 mM TRIS with 1 mM CaCl 2 (addition of Ca 2 + to proteinase K was carried out to degrade proteins that are commonly resistant to this enzyme) at a concentration of 400 ug/ml. The same volume of proteinase K was added to analytical and clinical samples (to give a final concentration of 200 ug/rnl) and incubated for 30 min at 56°C. For inactivation of the enzyme, the reaction mixture was heated to 100 °C for 15 min. Statistical analysis. ELIFA results with analytical samples containing 1 mg/ml of C. psittaci antigen were used to determine several statistical parameters of the test system. The following parameters were determined: detectability (cut off) and sensitivity. Detectability of the assay was defined as the lowest concentration of antigen exceeding the zero-dose precision and was estimated using the following equation: xsD . . t n_ 1; (1; one sided Detectabiliry = Absorbance.g., dose + yin Sensitivity was defined as the responsiveness of the assay to changes in concentration of the substance tested. Therefore, on a sigmoid dose-response curve, sensitivity decreases at either end of the curve. Sensitivity was estimated using the following equation: Sensitivity = dR/dC; where R = ELIFA units; C = reciprocal dilution (lOglO)'
Comparison of the Capture ELISA/ELIFA with commercially available Chlamydiaxymev and IDEIA™ test kits. Detecrabiliry and specificity of the new capture ELISNELIFA was directly compared with the Chlamydiazyme and !DEIA test systems using dilutions of a 1 mg/ml C. psittaci 1904 isolate stock solution and a panel of different bacteria (see material and methods) at a concentration of 10 6/ml as analytical samples. Tests were performed exactly as indicated by the manufacturers.
Results
Antibodies 16 hybridoma cell lines secreting antibodies directed against genus-specific chlamydial epitopes could be established. Nine of these hybridomas produced antibodies of the IgG2a and five those of IgG3 subclass, whereas two produced antibodies of the IgM type. All monoclonal antibodies produced carried kappa light chains. All clones of interest could be easily adapted to produce monoclonal antibodies in serum-free RPMI 1640 medium supplemented with "Ultroser HY". Cell culture supernatants contained antibodies in the 5 to 30 ug/rnl range. All IgG antibodies could be purified on protein A-Sepharose CL-4B with only insignificant loss of activity. Serum-free RPMI 1640 medium supplemented with "Ultroser HY" as control never yielded any detectable amount of immunoglobulins when applied to the protein A affinity chromatography column.
44
D. Thiele, M. Karo, and H. Krauss Binding characterization using different test systems
Preliminary binding studies using direct and indirect immunofluorescence, a dot test and the indirect immunoperoxidase test were carried out to evaluate the monoclonal antibodies. One of those recognizing all isolates and strains from the three Chlamydia species - Mab 22 (IgG3) - was finally chosen to set up the Capture ELISAJELIFA system. This monoclonal antibody was also tested with the immunoblotting procedure revealing reactivity with the genus-specific C. psittaci lipopolysaccharide (Fig. 1).
- - 92,5 * 66,2
•. •
LPS~
1
2
45,5 ..... 31,0
II • 3 4
21,5 14,4
5
Fig. 1. Western blots of C. psittaci, isolate 1904. Lane 1: SDS-PAGE gel stained with silver stain revealing protein and LPS bands. Lanes 2-5: Immobilon-P membranes. Lane 2: Protein bands stained with amido black. Lane 3: Immunoblot with Mab 22 purified from supernatants of cell cultures propagated in serum-free medium. Lane 4: Immunoblot with hyperimmune serum from BALB/c mouse. Lane 5: Marker proteins (MW in kDa).
Capture ELISA/ELIFA
Using this test system based on monoclonal antibody Mab 22 as capture and biotinylated detection antibody, all chlamydiaI strains and isolates were clearly recognized. All bacteria tested (20 different strains as negative control) did not react in this assay when using a concentration of 10 6 particles/m!. Using a stock solution of C. psittaci strain 1904 (1 mg/ml), the assay still gave a positive signal (cut-off) when applying a dose of 100 f.tl of a 1: 320,000 dilution (Figs.2, 3). This amount contained 312 pg of C. psittaci antigen, or approximately
ELISAIELIFA for the Detection of Chlamydia psittaci
45
2000 particles. The detection limit was the same when using streptavidin conjugated with peroxidase (ELISA) or streptavidin conjugated with alkaline phosphatase (ELIFA). Sensitivity defined as the responsiveness of the assay to changes in concentration of the suspension tested was maximal in the 1: 10,000 dilution range of the antigen stock suspension. Capture ELISA
2500
2000 Ul
~
c
1500
:J
