Commercially Available Antibodies to Human Tumour Necrosis Factor-α Tested for Cross-Reactivity with Ovine and Bovine Tumour Necrosis Factor-α using Flow Cytometric Assays

A thorough understanding of the immune system, including the role of different cytokines, during inflammatory diseases in ruminants could lead to the development of new diagnostic methods and treatments. Tumour necrosis factor-α (TNF-α) is an important cytokine in the onset of the inflammatory responses. Unfortunately, the number of studies on cytokines, like TNF-α, in ruminants is limited due to a lack of species-specific reagents. As cytokines have remained rather conserved during evolution, cross-reactivity between animal species may occur. Therefore, the aim of the present study was to investigate 5 commercially available antibodies against human TNF-α for their ability to cross-react with ovine and/or bovine TNF-α, using a bead-based flow cytometric method. Two of the antibody clones (Mab 11 and 6401.1111) showed cross reactivity with ovine recombinant TNF-α in concentrations above 2.5 ng/ml. However, none of the antibodies detected TNF-α in bovine milk, or serum containing known concentrations of bovine TNF-α, as earlier determined with ELISA. The results could be due to inability of the antibodies to cross-react between species, but quenching of the signal by matrix proteins might also have lowered the response.


Introduction
The inflammatory response during bacterial infections in ruminants is mediated by pro-inflammatory cytokines, such as tumour necrosis factor-a (TNF-a) (Sordillo & Peel 1992). TNFα has an important role as a mediator of the inflammatory response as it induces expression of adhesion molecules on endothelial cells, activates neutrophils and macrophages, and induces production of nitric oxide and complement factors (Tracey 2002, Rainard et al. 2000. Quantitation of cytokines in biological fluids could be a useful tool in the diagnosis of inflammatory conditions since cytokine patterns give detailed information about the course of an inflammation. Infectious diseases, like mastitis, represent great health problems among dairy cows and sheep. A thorough understanding of how the immune system acts during such diseases could lead to the development of new diagnostic methods and/or treat-ments. Therefore, the development of new techniques for measuring ruminant cytokines is warranted. Several methods have been developed for cytokine detection, such as enzyme-linked immunosorbent assay (ELISA), enzyme-linked immuno spot assay (ELISPOT), bioassays, polymerase chain reaction (PCR) and intracellular staining (Bienvenu et al. 2000). Recently, multiplex cytokine assays have been developed, which measure several cytokines simultaneously in the same sample. In a study by Carson & Vignali (1999), 15 human cytokines were measured simultaneously in the same sample with a multiplex particle based flow cytometric method, the LabMAP™-assay (Luminex Corporation, Austin, TX, USA). Simultaneous quantification of several substances in a sample makes it possible to calculate the cytokine ratio. An altered cytokine ratio in an individual is considered a marker of disease state such as asthma and atopy (Maggi 1998). Studies on cytokines in ruminants are limited by a lack of species-specific reagents. However, as cytokines have remained conserved during evolution, cross-reactivity between animal species may occur. In support for this contention, Pedersen et al. (2002) found that an antibody against ovine TNF-α cross-reacted with bovine and human TNF-α. To our knowledge, no studies on cross-reactivity between antibodies to human TNF-α and ovine and bovine cytokines have been published. Therefore, the aim of this study was to investigate the ability of a number of commercially available antibodies to human TNF-α to cross-react with ovine and/or bovine TNF-α using particle based flow cytometry. The assays were performed using two different procedures. Procedure one is called the LabMAP™-assay, where both coupling of antibodies to microspheres and detection assays were performed. The other is called the Fluorokine ® MAP-assay (R&D Systems, Minneapolis, MN, USA), where microspheres already coupled with antibodies are used.

Antibodies, recombinant proteins and samples
The antibodies to human TNF-α used were all commercially available (Table 1). Two clones of recombinant human TNF-α (rhuTNF-α) were used. One was purchased from Serotec Ltd (Oxford, England) and the other was included in the Fluorokine ® MAP kit (R&D Systems). Recombinant ovine TNF-α (rovTNF-α) used was a kind gift from Dr. I. G. Colditz (CSIRO, Armidale, Australia). Recombinant bovine TNF-a is not commercially available. Therefore, bovine serum and milk samples with known concentrations of TNF-α were used as positive controls in this study. Three bovine serum samples containing 14000, 15400 and 15600 pg TNF-α/ml, respectively, were used as positive controls. The samples were a kind gift from Dr. C. Røntved, Danish Institute of Agricultural Sciences, Tjele, Denmark, who earlier had measured the concentration of bovine TNF-α in the serum sam-  (Grönlund et al. 2003). Milk samples were maintained at -70°C for one year before use. Two bovine serum samples and 2 bovine milk samples containing less than 100 pg/ml of TNF-α, originating from the same studies as the positive controls, were used as negative controls in this study. In experiments using rhuTNF-α and rovTNF-α, the buffer used was phosphate buffered saline (PBS) which also was used as a negative control.

Method optimization
Preliminary studies for the LabMAP™-assay were performed to determine the optimal detection system, which involved the assessement of different concentrations of capture and reporter antibodies, incubation times and temperatures. Two detection systems were tested, one direct method where the reporter antibody was PEconjugated, and one indirect method where a biotinylated reporter antibody was detected by PE-conjugated streptavidin. In assays on rovTNF-α, best results were achieved by indirect detection, whereas the best detection system for rhuTNF-α was the direct detection system. Seven different concentrations of the different capture antibodies (5mg/ml, 0.5mg/ ml, 50µg/ml, 5µg/ml, 100ng/ml, 50ng/ml and 10ng/ml) and reporter antibodies (3.2µg/ml, 1.6µg/ml, 800ng/ml, 400ng/ml, 320ng/ml, 200ng/ml, and 100ng/ml) were tested. Varying incubation times and temperatures were used for the incubation steps with antibodies (20 minutes at room temperature (RT) and 37°C, 45 min at RT, 1 h at RT, 1.5 h at RT, 2 h at RT and 37°C and overnight at RT). The best procedures for rhuTNF-α and rovTNF-α, with respect to signal intensity and background, are described below.

Coupling of monoclonal antibodies to microspheres
Coupling of antibodies to Luminex microspheres was performed in 1.7 ml eppendorf tubes (AB Göteborgs Termometerfabrik, Västra Frölunda, Sweden) according to the manu-facturer's instructions (Luminex 100 System  Training Material). Briefly, about 600 000 microspheres were dissolved and washed twice with 100µl activation buffer (0.1 M sodium phosphate buffer, pH 6.2) before resuspension in 80µl activation buffer. Ten µl each of freshly prepared 50 mg/ml sulfo-NHS and 50 mg/ml EDC solutions were added to the suspension. The microsphere solution was incubated for 20 min in the dark at RT. The activated microspheres were washed twice with 100µl coupling buffer (0.1M MES diluted in dH 2 O) before 500µl of TNF-α-antibody (0.5mg/ml) was added and the mixture was rotated for 2 h in the dark at RT (Fischer Hematology Mixer, Fisher Scientific International Inc., Hampton, NH, USA). After incubation the microspheres were washed twice with washing buffer (PBS with 0.05% Tween) and resuspended in PBS with 0.02% Tween (PBST) and 1mg/ml bovine serum albumin. The microspheres with attached antibodies were kept at 2-8 ºC, protected from light for at the longest for 14 days before use.

Cytokine assay using LabMAP™
RhuTNF-α (Serotec Ltd) or rovTNF-α was diluted in PBST to create appropriate standard curves. Bovine serum and milk samples were diluted 1:4 in PBST. Ninety ml cytokine solution, bovine serum or milk sample, was mixed with 3µl antibody-coupled microspheres, containing approximately 10 000 beads, and incubated for 20 min in the dark at RT. The mixture was subsequently washed twice with PBST. Forty ml biotinylated (5pg/ml) reporter antibody was added to rhuTNF-α samples whereas a PE-conjugated (0.3pg/ml) reporter antibody was added to the rovTNF-α samples. To milk and serum samples, either biotinylated reporter antibodies or PE-conjugated antibodies were added. All samples were incubated for 30 min in the dark at room temperature. After the incubation, rovTNF-α, milk and serum samples with PE-conjugated antibodies were washed twice with PBST and resuspended in 300ml PBS. Thereafter, they were transferred to 96well microtiter plates (Göteborgs Termometerfabrik) and analysed on the Luminex 100™ analyser (Luminex Corp.). The rhuTNF-α, milk and serum samples with biotinylated reporter antibodies were washed twice with 100µl PBST. Then 40ml streptavidin-PE, diluted 1:100 in PBST, was added and the samples were incubated for 20 min protected from light at RT. Samples were then washed twice with PBST, resuspended in 300µl PBS and analysed on the Luminex 100™ analyser. All experiments were performed in duplicates or triplicates and repeated at least 2 times.
Cytokine assay using Fluorokine® MAP The cytokine assay was performed according to manufacturer's instructions (R&D, Minneapolis, MN, USA). Briefly, RhuTNF-α (Serotec Ltd and R&D Systems) or rovTNF-α was diluted in PBST to appropriate standard curves. Bovine serum and milk samples were diluted 1:4 in PBST. A recovery experiment was conducted to verify if the assay could measure cytokines added to the milk or serum matrix. Duplicates of each sample were prepared and to one of the duplicates 50 µl of 2ng/ml rhuTNFα was added. Fifty ml microsphere mixture was added to 50µl standard, or sample, and was incubated in the dark at RT on rotation (Fischer Hematology Mixer, Fisher Scientific International Inc.) for 2 h. The microspheres were washed 3 times with 100µl washing buffer after which 100µl of conjugate cocktail was added and the solution was incubated for another 2 h. After incubation, the microspheres were washed 3 times with washing buffer, resuspended in 100µl washing buffer and analysed in the Luminex 100™ analyser. All experiments were performed in duplicates or triplicates and repeated at least 2 times.

Cytokine assay using LabMAP™
Results from the cytokine assay by Luminex showed that all antibodies investigated detected rhuTNF-α (data not shown). In addition, the antibody combination Mab11 and 6401.1111 showed cross-reactivity with rovTNF-α when the cytokine concentration was above 2.5ng/ml (Fig. 1). However, none of the 7 different antibody combinations tested detected TNF-α in the bovine serum or milk samples (data not shown).
Cytokine assay using Fluorokine ® MAP The limit of detection of the Fluorokine ® MAPkit for rhuTNF-a was approximately 4pg/ml (Fig. 2). The kit detected both rhuTNF-α from R&D Systems, and rhuTNF-α from Serotec Ltd. The kit did not detect rovTNF-α (data not shown) nor bovine TNF-α in the serum or milk samples (Figs. 3 and 4). The recovery test showed that the kit detected rhuTNF-α when 50µl of 2ng/ml rhuTNF-α was added to the milk or serum samples (Figs. 3 and 4). The magnitude of the signals was however weaker in milk and serum compared with values achieved when measuring the same amount of rhuTNF-α in buffer. The mean percent recoveries in serum and milk, based on results from duplicate samples from two separate experiments, were 36.4 ± 1.6% and 59.8 ± 2.0%, respectively. When calculating the results, the mean fluorescence intensity from the experiments with rhuTNF-a in buffer was set to 100%. Thereafter, the mean fluorescence intensity from the experiments with rhuTNF-α serum and milk were compared to the experiments in buffer and the results are shown as percent of rhuTNF-α in serum/milk ± standard deviation.

Discussion
All antibodies to human TNF-α investigated with the LabMAP™-assay in this study could detect rhuTNF-α. Those results indicate that the technique worked satisfactory. The antibodies against human TNF-α named Mab11 and 6401.1111 in combination also detected rovTNF-α in concentrations above 2.5ng/ml. However, this detection level is not satisfactory given that cytokines are very powerful signal molecules that give effects at low concentrations (Kelso 1998). None of the antibodies against human TNF-α tested with the LabMAP™-assay could detect TNF-α in bovine serum or milk samples with known TNF-α concentrations. These results could be explained by the inability of antibod-ies to cross-react between species, or that the fluorescent signal was somehow quenched. Results of Pedersen et al. (2002) suggest that the TNF-α protein has remained conserved during evolution since other studies have shown that at least 60% amino acid homology is required between proteins for cross-reactivity to occur (Scheerlinck 1999). An 84% homogeneity is found between human and bovine TNF-α while ovine and bovine TNF-α shows more than 90% homogeneity when performing a genetic sequence comparison (http://www.ncbi.nlm.nih. gov/BLAST/). This indicates that human, bovine and ovine TNF-a have a high degree of similarity. However, our results do not completely support that theory, as none of the human antibodies tested in this study detected TNF-α in bovine samples and only one pair of antibodies detected rovTNF-α using the En grundlig förståelse för immunsystemet inklusive betydelsen av olika cytokiner vid inflammationssjukdomar hos idisslare kan leda till utveckling av nya diagnostiska metoder och behandlingar. En viktig cyto-kin i den tidiga inflammationsreaktionen är tumour necrosis factor-α (TNF-α). Tyvärr begränsas ofta studier rörande cytokiner som TNF-α hos idisslare av brist på artspecifika reagenser. Eftersom cytokiner har visats vara konserverade genom evolutionen förekommer korsreaktivitet mellan arter. Målet med den här studien var därför att med hjälp av partikelbaserad flödescytometri undersöka om 5 kommersiellt tillgängliga antikroppar mot humant TNF-α kunde korsreagera med ovint och/eller bovint TNF-a. Två av antikroppsklonerna (Mab 11 och 6401.1111) korsreagerade med ovint rekombinant TNF-α vid koncentrationer över 2,5 ng/ml. Ingen av de undersökta antikropparna detekterade dock TNF-α i bovina serum-och mjölkprov med kända koncentrationer av TNF-α tidigare bestämda med ELISA-analys. Resultaten kan tyda på en oförmåga hos antikropparna att korsreagera mellan arter men matrixproteiner kan också ha minskat signalens magnitud.