Immunological Methods for the Detection of Campylobacter spp . Current Applications and Potential Use in Biosensors

The genus Campylobacter belongs to the epsilon division of the class Proteobacteria. This genus comprises a group of bacteria that occupy diverse habitats and produce a wide range of diseases in different animal hosts (On 2001). Campylobacteriosis is a self-limited gastrointestinal illness that produces diarrhea, fever and abdominal cramps, and antimicrobial therapy is not generally indicated. However, treatment can reduce the duration and severity of illness if caught early, especially in those with the potential for severe illness, including infants, elderly, patients with underlying disease and immunocompromised individuals. In addition to acute enteritis, campylobacteriosis may result in reactive arthritis and Guillain-Barre syndrome (Hannu et al., 2002; Rees et al., 1995). The epidemiology of campylobacteriosis is unique in that most infections are sporadic. Meats, especially undercooked broiler chicken, are the main source of sporadic campylobacteriosis cases, while outbreaks are usually associated with the consumption of raw milk or unchlorinated water (Skirrow 1991). Campylobacter spp. are one of the leading agents of bacterial foodborne diseases worldwide (EFSA 2005, 2006). In the USA, there are approximately 12 reported cases per 100,000 persons per year, although the actual number has been estimated at 432 cases per 100,000 (Olson et al., 2008; Samuel et al., 2004). C. jejuni is responsible for approximately 95% of the human cases, while C. coli is responsible for the rest of the infections. C. concisus, C. fetus, C. upsaliensis and other Campylobacter species can cause sporadic cases.


Introduction
The genus Campylobacter belongs to the epsilon division of the class Proteobacteria.This genus comprises a group of bacteria that occupy diverse habitats and produce a wide range of diseases in different animal hosts (On 2001).Campylobacteriosis is a self-limited gastrointestinal illness that produces diarrhea, fever and abdominal cramps, and antimicrobial therapy is not generally indicated.However, treatment can reduce the duration and severity of illness if caught early, especially in those with the potential for severe illness, including infants, elderly, patients with underlying disease and immunocompromised individuals.In addition to acute enteritis, campylobacteriosis may result in reactive arthritis and Guillain-Barre syndrome (Hannu et al., 2002;Rees et al., 1995).The epidemiology of campylobacteriosis is unique in that most infections are sporadic.Meats, especially undercooked broiler chicken, are the main source of sporadic campylobacteriosis cases, while outbreaks are usually associated with the consumption of raw milk or unchlorinated water (Skirrow 1991).Campylobacter spp.are one of the leading agents of bacterial foodborne diseases worldwide (EFSA 2005(EFSA , 2006)).In the USA, there are approximately 12 reported cases per 100,000 persons per year, although the actual number has been estimated at 432 cases per 100,000 (Olson et al., 2008;Samuel et al., 2004).C. jejuni is responsible for approximately 95% of the human cases, while C. coli is responsible for the rest of the infections.C. concisus, C. fetus, C. upsaliensis and other Campylobacter species can cause sporadic cases.
Current methods of identification of campylobacters in stool and food samples rely on their growth on selective agar plates with and without prior broth enrichment.These methods are labor intensive and time consuming, taking four or more days for completion and require robust bacterial growth, which may limit the detection of stressed bacteria that do not grow well but are still infective.More rapid and accurate methods are necessary to identify campylobacters in stool samples in order to treat campylobacteriosis early, and to detect outbreaks as campylobacteriosis is a notifiable disease in the USA.Moreover, rapid detection of campylobacters in environmental samples is important for trace-back investigations to mitigate outbreaks and in food samples to catch contamination during commercial food processing.
During the 1960s and early 1970s, the search for alternative techniques to replace radioactive label reporters to identify biological molecules led to the development of the enzyme-linked immunosorbent assay (ELISA or EIA; Engvall and Perlmann 1971).In the mid-1970s the, the development of the hybridoma technique for monoclonal antibody synthesis (Köhler & Milstein 1975) helped expand the new field of immunodiagnosis and initiate development of immunoassays for the identification of bacterial pathogens.Since then, a variety of immunoassays have been developed for the detection of different foodborne pathogens in food and stool samples, including a few immunoassays for Campylobacter spp.detection.The simplest is latex agglutination, in which antibody-coated colored latex beads or colloidal gold particles are used for rapid confirmation of culture results or serotyping of culture isolates.The EIA, the most popular immunoassay used for pathogen detection in foods and stool, is typically designed as a "sandwich" assay, in which an antibody bound to a solid matrix is used to capture the bacterium and a second antibody conjugated to an enzyme then binds to the bacterium.Multi-well microplates are a commonly used solid support but other formats include dipsticks, paddles, membranes, or other solid matrices.
The lateral flow immunochromatographic method is a modified EIA, packaged in a simple device (dipstick or within a plastic casing) and used for rapid pathogen detection.Typically, ″sandwich" assays are used for large analytes such as bacteria.Samples migrate from the sample pad through a conjugate pad where the target analyte binds to the antibody conjugated to colored particles.The sample is drawn across the membrane to the capture zone where the target/conjugate complex binds to immobilized antibodies producing a visible line on the membrane.To ensure a working test, the sample migrates further until it reaches the control zone, where excess conjugate is bound to produce a second visible line on the membrane.Two clear lines on the membrane are a positive result.A single line in the control zone is a negative result.Lateral flow immunoassays have many advantages including their simplicity, production of a result within 15 minutes, stability with a long shelf life even in some cases without refrigeration and their low cost, but they may have a higher threshold of detection compared to EIA.
The present chapter reviews commercial and/or published immunological methods, mainly EIA and lateral flow immunoassays used to identify species of Campylobacter in food and stool samples.Methods that can be dovetailed with immunoassays to increase bacterial concentration and immunoassay detection, such as sample enrichment, sample filtration and immunomagnetic separation, are described.The strengths and limitations of immunoassays for identification of Campylobacter spp.are reviewed, along with suggestions to improve assay performance.The detection of Campylobacter spp.by antibody-based biosensors, primarily optical biosensors is also briefly discussed.

Identification of Campylobacter spp. using immunoassays
The presence of campylobacters in food and stool samples is based on the growth of the bacteria on a selective agar, incubated at 42°C under a reduced oxygen atmosphere (~5% O 2 ).The limitations of the culture method, especially the need for up to four days to identify campylobacters (Endtz 2000), have dictated the development of culture-independent methods, including immunoassays, for the detection of Campylobacter spp. in clinical stool and food samples.Although a variety of immunoassays have been developed for testing clinical and food samples for Campylobacter spp., these assays require approval by regulatory bodies, necessitating comparison of immunoassays to culture-based methods, considered the "reference methods."The validation of immunoassays for detection of Campylobacter spp.includes testing for inclusivity, to assure that different strains of C. jejuni and C. coli are detected, and exclusivity, to assure that closely related, non-target bacteria do not cross-react with the assay (Brunelle 2008).Other performance indicators used to validate assays include:


Sensitivity: The probability that an assay will yield a positive result when the culture is positive by the reference method. Specificity: The probability that an assay will yield a negative result when the sample is negative by the reference method.


Positive predictive value: The probability that a sample with a positive test contains the bacteria.


Negative predictive value: The probability that a sample with a negative test does not contain the bacteria.
In the following sections, commercial immunoassays available to detect Campylobacter spp. in stool and food samples will be described.

Clinical stool samples
Several commercial immunological assays are available for identification of C. jejuni and C. coli in stool samples, including some designed to confirm culture results and others that are culture-independent.1).These methods are reasonably rapid, from 20 minutes for the lateral flow assays to2-4 h for the microplate assays, and identify specific Campylobacter antigens common to C. jejuni and C. coli.
The ProSpecT TM Campylobacter assay (Remel Inc., Lenexa, KS), a microplate sandwich EIA that uses a polyclonal antibody conjugated to horseradish peroxidase for the detection of common antigen of C. jejuni and C. coli in fecal specimens and enriched fecal cultures, has received the most scrutiny.The results are read visually or spectrophotometrically, and the analytical sensitivity is approximately 10 5-6 colony forming units (CFU) per ml -1 .The test is accurate for samples stored at 4°C for several days (Dediste et al., 2003;Tolcin et al., 2000).
No cross-reactivity with other major fecal bacteria or with other Campylobacter spp., including C. lari and C. fetus, has been identified (Endtz et al., 2000;Hindiyeh et al., 2000).The manufacturer's studies, using three sites in the USA and Canada, demonstrated a pooled sensitivity of 97-100% and specificity of 98-100% using 1,049 stool samples (Table 1).
The PREMIER™ CAMPY microplate EIA (Meridian Bioscience, Inc, Cincinnati, OH) uses a monoclonal antibody that binds to an unspecified common antigen of C. jejuni and C. coli.
The limit of detection is 10 6-7 CFU per ml -1 , and according to the manufacturer's literature, this assay had a sensitivity of 97% and specificity of 96% in a study of 2,073 samples (Table 1).High specificity and sensitivity for this EIA were reported by Granato et al. (2010).
Another microplate EIA, Ridascreen Campylobacter (R-Biopharm AG, Germany), also uses monoclonal antibodies but does not appear to perform as well as the manufacturer's claims (Bessede et al., 2011;Tissari et al., 2007).The lateral flow assay, ImmunoCard STAT! CAMPY (Meridian Bioscience, Inc.), uses the same monoclonal antibody as the PREMIER CAMPY assay.The capturing of the target bacteria is by an antibody colloidal complex and the assay has similar specificity to the EIA assays, but with a sensitivity of 10 7 CFU per ml -1 (Bessede et al., 2011;Granato et al., 2010).The Xpect Campylobacter assay (Remel Inc.) is a rapid assay equivalent to the ImmunoCard STAT! CAMPY assay, but preliminary results suggest that it has poor sensitivity (Fitzgerald et al., 2011).
Currently, none of the described assays can be recommended for standalone identification of campylobacters in stool samples, in part because of the limited and conflicting findings regarding the performance of these immunoassays.The performance of these EIAs is variable and suggests a high probability of non-acceptable levels of false negatives in certain situations.The Centers for Disease Control and Prevention (CDC) has recently released preliminary data concluding that EIAs should not be used as standalone tests for direct detection of Campylobacter in stool samples.This study investigated 2,767 stool samples with a positive Campylobacter prevalence of approximately 3%.Although the specificity and negative predictive values in these tests were excellent, typically >95%, the sensitivities and positive predictive values (PPV) of the four EIAs were not acceptable.For example, the ProSpecT TM Campylobacter assay exhibited a sensitivity of 84% and a PPV of 56%, the PREMIER™ CAMPY assay exhibited a sensitivity of 83% and a PPV of 52%, the ImmunoCard STAT! CAMPY exhibited a sensitivity of 73% and a PPV of 39%, and the Xpect Campylobacter assay exhibited a sensitivity of 74% and a PPV of 80%.Therefore, a positive EIA test alone is not sufficient to consider a case "confirmed" and laboratories must confirm positive EIA results by culture methods (Fitzgerald et al., 2011).The basis for discordant results between culture, immunoassay and PCR-based methods needs to be determined.
It is not clear whether the poor performance of these EIAs is inherent in the assays themselves or is due to lack of optimization.The fact that variations appear to be dependent on the test format and manufacturer suggests that these assays could be improved (See Section 4).A limitation for the adoption of EIAs is the prohibitive cost of adopting these rapid tests in combination with routine culture methods (Abubakar et al., 2007).

Food samples
Current identification methods for Campylobacter in foods rely on bacterial growth on one of a variety of selective agar plates which contain antimicrobials to allow for the growth of the target organism, a method adapted from that used for stool samples (Corry et al., 1995).To achieve the required sensitivity of 1 cell per 25 g food, a 25 g food sample is typically enriched in a broth with selective agents to inhibit the growth of competing bacteria while allowing growth of the target organism before plating on selective agar (Oyarzabal 2005;Oyarzabal et al., 2007).Besides increasing bacterial concentration, enrichment is important because campylobacters are randomly distributed and can be present in clumps or aggregates in foods, limiting direct detection.Also, enrichment ensures that stressed or injured bacteria have a chance to recover prior to plating on selective agars.Culture-based methods are the accepted methods outlined in the Food and Drug Administration (FDA)'s Bacteriological Analytical Manual (BAM), the Microbiology Laboratory Guidebook (MLG) of the Food Safety and Inspection Services of the U.S. Department of Agriculture (FSIS USDA) and the International Organization for Standardization (ISO).
These culture-based methods for food screening suffer from the same constraints as those for stool samples: low specificity resulting in false positives, and long lag-time for results (Josefsen et al., 2011).Until recently, the lack of active surveillance and regulatory efforts to control Campylobacter spp. in poultry meat, a common vehicle for these pathogens (FSIS USDA 2009a), delayed the development of rapid methods.However, a new performance standard aimed at limiting the prevalence of Campylobacter-contaminated poultry meat products in the USA requires the screening of processed carcasses for the presence of this pathogen (FSIS USDA 2009b).Thus, there is a critical need to develop and validate rapid methods for Campylobacter identification, including those that are antibody-based.
Although antibody-based methods would speed assay time, only a few latex agglutination assays are included in reference manuals and only for confirmation of positive Campylobacter isolates.The FDA's BAM (Hunt et al., 2001) 2011).Commercial EIAs are available for culture-independent identification of campylobacters in food, but these assays have not been extensively validated.Typically applied to enriched cultures, these EIAs reduce assay time and may perform with equal specificity as culture methods, albeit with a sensitivity > 10 4 CFU per ml -1 , which is not an improvement over culture methods (Josefsen et al., 2011).Presently, a positive result by a rapid method is only regarded as a presum ptive result and must be confirmed by a standard method (Hunt et al., 2001;ISO 2006;MLG 2011).The next paragraphs describe the most common commercial EIAs used with food samples.
The VIDAS/MiniVIDAS CAM (bioMerieux, Hazelwood, MO), an automated EIA for detection of thermotolerant C. jejuni, C. coli and C. lari, has received the most attention over the last decade.In this assay, food or stool samples are incubated in an enrichment broth for 48 h.After enrichment, samples are boiled for 15 minutes and aliquots are analyzed by an automated enzyme immunoanalyzer.The few published studies indicate that this method performs similarly to culture methods for the identification of naturally occurring Campylobacter spp. in foods.Hoorfar et al. (1999) found that the MiniVIDAS CAM may provide a faster method for fecal samples from cattle and pigs with the capability of rapid screening of a large number of samples.The reported sensitivity of this assay ranged from 88-94%, which may not be adequate for screening samples with low numbers of Campylobacter spp.However, authors used a lower enrichment time (24 h) compared to the manufacturer's recommendation (48 h).Borck et al. (2002) compared the MiniVIDAS and the EIAFoss, an automated EIA which is no longer available, to culture methods for the identification of Campylobacter spp.The reported sensitivity for both EIA assays was higher than the sensitivity of selective agar plates for a variety of turkey meat and turkey fecal samples, but sensitivity varied as a function of sample and enrichment broth.The MiniVIDAS exhibited a sensitivity of approximately 94% for fecal samples with high bacterial loads and 65% for environmental samples with lower bacterial loads, and each methods had a specificity above 93%.Paulsen et el. (2005) has also indicated that the MiniVIDAS is as good as the standard method, with a sensitivity of 97.6% and a reduction in assay time by 24 h for chilled and frozen meat.In this study, Bolton enrichment broth in modified stomacher bags was better than Preston broth for enrichment prior to immunoassay.The MiniVIDAS CAM method has been successful for the detection of Campylobacter spp. in tissue samples (tonsils and lymph nodes) and fecal material from pigs during processing (Nesbakken et al., 2003), in a variety of chicken parts with positive results confirmed by culture methods (Reiter et al., 2005), and in artificiallycontaminated ground beef and fresh cut vegetables (Chon et al., 2011).Our experiences (OAO) with the MiniVIDAS for screening poultry meat are in accordance with the majority of reports mentioned above.However, the enrichment of the samples for 48 h is indispensable to reduce the high number of false negative samples that otherwise will be encountered (Liu et al., 2009).
Other commercial immunoassays have received less scrutiny than the MiniVIDAS CAM.
The TECRA Campylobacter Visual Immunoassay (3M, St. Paul, MN) for C. jejuni and C. coli employs a 40-h aerobic enrichment in a proprietary broth followed by a 2-h EIA.For 398 broiler carcass rinses from 19 processing plants, TECRA found 317 Campylobacter positives, four false positives and 22 false negatives, compared to 48 false negatives with the reference method (Bailey et al., 2008).All these false negatives came from rinses of carcasses collected toward the end of the production process, suggesting that severely injured campylobacters may not recover well in the TECRA enrichment broth.In another study, the TECRA VIA was less sensitive than a PCR method for detection of campylobacters in spiked raw and processed meat and poultry, but performed similarly or better than traditional culture media (Bohaychuk 2005).The low number of false negatives indicates that the TECRA VIA may be a viable screening tool as long as follow-up tests are performed on presumptive positive samples to rule out false positives.Unfortunately, no literature was found describing the antibody target and improvements might be possible with a better choice of antibody because, in the same study, an immunoassay outperformed PCR for detection of Salmonella.
Two other commercially available assays for rapid testing of food samples are the Singlepath TM Campylobacter microwell sandwich EIA (Merck KGaA, Germany) and the NH Immunochromato Campylobacter (Cosmo Bio Co., Tokyo, Japan).The Singlepath was comparable to standard culture methods in a preliminary study with artificially-inoculated www.intechopen.commeat (Hochel et al., 2004).The NH Immunochromato Campylobacter is a two-step enrichment assay using immunochromatographic identification with a reported sensitivity of 55 CFU per 25 g of spiked chicken meat for non-freezer stressed samples.Freezing samples decreased the sensitivity approximately 10-folds (Kawatsu et al., 2010).
Based upon the studies described above, commercial immunoassays show promise as one of the tools to speed food sample screening.However, more studies of these assays with naturally-contaminated samples from a variety of foods is necessary.The effect of bacterial stress that occurs during food processing and storage on assay performance should also be studied.Additionally, the antigen targets of the antibodies used in these assays need to be identified.A concerted effort to improve the performance of these assays should be undertaken now that a new standard for food screening poultry meat for Campylobacter spp.has been established in the USA.

Increasing concentration of campylobacters prior to immunoassay identification
Because EIA assays have a sensitivity of ≥ 10 4 CFU per ml -1 or g -1 , there is a need to increase the number of the target organisms in food samples prior to assay.The enrichment of the sample is the most common method to increase the number of bacterial cells.No enrichment protocol is 100% selective for the organism of concern; therefore other methods are used to separate or concentrate the target bacteria from the rest of the contaminants.This section will review enrichment broths, filtration, and immunomagnetic separation as techniques to increase the number of Campylobacter spp. in the food samples prior to identification by EIA.

Enrichment of the samples
The enrichment of food samples is imperative for the isolation of Campylobacter spp.from poultry or raw milk products.The most commonly used enrichment broth is Bolton broth, which has a basal component made up of meat peptone, lactalbumin hydrolysate, yeast extract, alphaketoglutaric acid, ssodium chloride, sodium pyruvate, metabisulphite and carbonate.The use of buffered peptone water performs similarly to Bolton broth for the isolation Campylobacter spp., suggesting that the basal medium does not need to be rich in nutrients to support the growth and multiplication of Campylobacter spp.(Oyarzabal et al., 2007).An important requirement is the incubation of food samples for at least 48 h in enrichment broth before identification of positive cultures by immunoassay because 24 h incubation results in a high number of false negative samples (Liu et al., 2009).
The antimicrobials added to the basal broth are cefoperazone, trimethoprim and vancomycin, at concentrations of 20 mg per L -1 each, and cycloheximide, at a concentration of 50 mg per L -1 .Originally, Bolton broth was supplemented with 5% lysed horse blood, but research has shown that the addition of blood is not necessary for isolation of Campylobacter spp.from retail broiler meat.Most importantly, blood in the enrichment broth is not necessary if an EIA-based method is employed to detect positive samples (Liu et al., 2009).
In the laboratory of one of the authors (OAO), a modification of Bolton broth was made by reducing the antimicrobials added to enrichment and plate media to only cefoperazone, at a concentration of 32 mg per L -1 , and amphotericin B at a concentration of 2 to 10 mg per L -1 (Liu et al., 2009;Zhou et al., 2011).Although this reduction of antimicrobials may allow for contaminants to grow, we control those contaminants by filtration prior to assay.

Use of filter membranes to separate contaminating bacteria
Filter membranes were used for the first isolation of C. jejuni from human stools (Dekeyser et al., 1972).In this report, the use of 0.65-µm filters was aimed at retaining most of the contaminating bacteria in fecal suspensions while allowing Campylobacter organisms to pass through the filter for isolation on agar plates.This differential filtration is different from most of the other filtration protocols in which the target organisms are concentrated by filter retention for subsequent identification.A search for different agar plates with more antimicrobials to control contaminants was initiated in the 1970s and by the early 1990s the filters were not used anymore, except for the isolation of less-known species of Campylobacter by Albert A filtration method coupled with a sandwich EIA that uses polymyxin B sulfate (PMB) instead of an antibody to capture C. jejuni and C. coli was developed by one of the authors (CAB).PMB is used to capture Gram-negative bacteria for EIA detection of Escherichia coli and Salmonella in a variety of foods (Blais 2005(Blais , 2006)).PMB binds to lipid A of the lipooligosaccharide (LOS) layer and improves bacterial capture, although treatment of bacteria with zwitterionic detergents is necessary (Blais 2005(Blais , 2006;;Brooks et al., 1998).In our experiments, various surface waters and tap water were filter sterilized prior to use, and samples of each were spiked with concentrations of 0 to 10 4 CFU ml -1 of Campylobacter cells.Aliquots (500 ml) were filtered using a conventional EPA filter system with Microfil V 0.2 µm filters and then filters were added to 1 ml phosphate-buffered saline containing CHAPS detergent (3-[(3-cholamidopropyl dimethylammonio]-1-propanesulfonate) with 1-2 minute vortexing.Aliquots of 100 µl were assayed in the PMB-capture ELISA.Sampling (100 µl) of the 500 ml prior to filtration did not yield values above background.Thus, filter concentration coupled with PMB-capture ELISA was able to detect 10 2 CFUml -1 , compared to the typical threshold of 10 4-5 CFU per ml -1 , although the net fluorescence intensity was much lower for ocean water (Fig. 1).

Concentration by immunomagnetic separation
Immunomagnetic separation utilizes magnetic spheres, such as Dynabeads TM (Dynal Biotech), coated with antibodies to capture target bacteria from a variety of matrices.Antibody-bound target cells are separated from the matrix and other bacteria by application of a magnetic field.However, pre-enrichment may still be necessary to obtain a high number of bacteria for detection.Immunocapture followed by plating on selective agar had a threshold of 10 4 CFU per g -1 for the detection of Campylobacter in ground poultry meat (Yu et al., 2001).The sensitivity did not improve when magnetic beads (Dynabeads TM ) were coated with a monoclonal antibody to the major 45 KDa porin and coupled with a DNA hybridization assay specific for the 23S rRNA gene of Campylobacter spp.(Lamoureux et al., 1997).A modification of immunomagnetic capture uses multiplexed magnetic microspheres, fluorescence-coded microspheres coated with antibodies, to detect bacteria by flow cytometry (MagPlex® Microspheres, Luminex, Austin, TX).Although this technique allows for rapid, multiplexed assays, the desired limit of detection is still a challenge for a variety of food matrices (Kim et al., 2010).

Improving antibody-based methods for Campylobacter spp. detection
One of the limitations of antibody-based methods for detection of pathogenic bacteria is the high threshold of detection (~10 4-6 CFUml -1 ), which results in unacceptable numbers of false negative samples especially for foods (Hochel et al., 2007).The detection limit is not as problematic for rapid detection in stool samples due to the high concentration of campylobacters in fecal samples.As discussed above, bacterial concentrations can be increased to EIA-detectable levels in food, such as retail broiler meats, by broth enrichment of up to 48 h (Liu et al., 2009), although components of enrichment broths may reduce immunoassay sensitivity (Chon et al., 2011).Improvements in EIAs such as reducing background noise or increasing the amplitude of the detection signal can improve performance.Fluorescence-based signaling improves the detection threshold by at least an order of magnitude over colorimetric assays, and the use of biosensors may allow for more efficient signaling to improve sensitivity.If sample matrices reduce EIA performance by increasing background signal, a variety of polymers, such as polyvinylpyrrolidione (Nyquist-Battie 2004) and Biolipidure tm (http://www.biolipidure.com/)can reduce background noise thereby increasing the signal to noise ratio.Another factor that may improve immunoassay performance is the molecular structure of the antibodies.F(ab') 2 fragments, or single chain Fv recombinant antibodies because they are smaller resulting in less steric hindrance between antibody molecules, although this premise needs to be investigated.
A possible reason for the high rate of false negatives in EIAs used to detect Campylobacter spp.may be the choice of antibody (Hochel et al., 2007;Rice et al., 1996).Some antibodies may not be able to detect all strains of C. jejuni and C. coli.Antigenic variation is common for externally-exposed molecule, such as outer membrane proteins and LOS, which limits the inclusivity of generic antibodies ( Logan & Trust, 1983;Taylor & Chang, 1987, Dubreuil et al., 1990;Hochel et al., 2004).Indeed, surface-exposed immunodominant molecules, such as flagellar proteins (Fernando 2007) and the O-antigen, undergo genetic drift to avoid immunological detection.An alternative target molecule is the core oligosaccharide of the LOS, which has been successfully targeted in EIA assays to detect C. fetus (Brooks 2002(Brooks , 2004;;Devenish 2005).High rates of continuous expression of the antigen is another important consideration, especially in Campylobacter cells that undergo stress during food production and processing.Detection of stressed bacteria by EIA has been shown to be reduced (Hahm & Bhunia 2006;Nyquist-Battie 2005).In this regard, the structure of the LOS, a target of many antibodies, may vary as a function of the temperature at which Campylobacter cells are grown (Semchenko et al., 2010).Thus, identifying stable surfaceexposed molecules for detection of C. jejuni and C. coli could be an avenue to improve antibody-based detection of these pathogens.
Improvements in proteomics over the last decade have made the identification of surface molecules easier (Cordwell et al., 2008;Prokhorova et al., 2006).Once ideal surface molecules are identified, newer methods of antibody production, such as the use of recombinant proteins, peptide fragments or DNA plasmids as immunogens, should be explored.For example, identifying epitopes that are stable but specific for C. jejuni in FlaA protein could be undertaken (Fernando 2007), and then peptide fragments with these epitopic sequences could be used to generate antibodies.If an ideal molecular target cannot be identified, it may be necessary to use antibody cocktails to increase the inclusivity of immunoassays (Hochel et al., 2007;Rice et al., 1996).

Polyclonal versus monoclonal antibodies for C. jejuni and C. coli detection
Several polyclonal antibodies have been developed for the detection of C. jejuni and C. coli, but few have been thoroughly tested for the ability to identify most if not all strains.In one study, nine strains were screened using two commercial anti-Campylobacter antibodies, B6601R (Biodesign, Saco ME) and 01-92-93 (Kirkegaard and Perry Laboratories; KPL, Gathesburg, MD ).Only the former antibody identified all nine Campylobacter strains (Wang et al., 2000).Recently, we screened polyclonal antibodies for detection of C. jejuni and C. coli using a non-sandwich indirect fluorescence ELISA.The KPL antibody 01-92-93, a rabbit anti-Campylobacter jejuni C1037-10 (US Biological B10), and C1037-14 (USB14) antibodies recognized all four strains of C. coli and all nine strains of C. jejuni with a limit of detection of approximately 10 6 CFU per ml -1 (unpublished results).The rabbit anti-Campylobacter spp.antibody 9-25B-PA (Cygnus Tech.) did not recognize all strains.Given the successful detection of a variety of C. jejuni and C. coli strains by the USB polyclonal antibodies and the need to determine useful antigenic targets, we performed Western blotting with two C. jejuni isolates (ADPH1608 and ADPH 1208, both human isolates).The USB antibodies did not bind to LOS epitopes, as Proteinase K digestion eliminated all bands.Next the surface proteins that are recognized by the USB antibodies were determined by biotinylating surface proteins of the bacteria using a non-membrane penetrant biotinylation agent, EZwww.intechopen.comLink Sulfo-NHS-LC-Biotin, at 5 mg per ml-1 for 1 h at room temperature, a modification of the method of Harding et al. (2007).Membrane proteins were extracted using two membrane protein isolation kits (G Bioscience Focus™ Membrane Protein Extraction kit and BioRad Ready Prep sequential extraction kit).Seven protein bands with the molecular weights of 73, 62, 43, 41, 35, 28, and 24 KDa were both biotinylated and recognized by the USB antibodies using both protein extraction kits.The molecular weight of these bands is similar to many proteins described previously as useful antibody targets for detection of campylobacters.A 62 KDa protein, possibly one of the flagellin proteins, has been mentioned as a useful by previous authors, target for campylobacter detection (Heo et al., 2009;Lu et al., 1997) and is one of the protein targets of the KPL polyclonal antibody (Rice et al., 1996).Lu et al. (1997) demonstrated that a monoclonal antibody to a62 KDa protein was able to detect campylobacters at concentrations of 10 5 CFU per ml-1 with great specificity.The 43 Kda protein is another protein that is recognized by Campylobacter antibodies such as the KPL antibody (Rice et al., 1996).This 43 KDa protein, bound by a monoclonal antibody developed after injecting mice with whole cell C. coli, was determined by MALDI-TOF mass spectrometry to be a major outer membrane protein (Qian et al., 2008a).This antibody was shown to bind to the cell surface using immunogold electron microscopy and detected 10 3 to 10 4 CFU ml -1 .A recombinant major outer membrane protein (43 KDa) was used to generate a monoclonal antibody that was specific for C. jejuni (Qian et al., 2008b).This antibody recognized an amino acid epitope which is 97% conserved but which is predicted to be exposed to the periplasm not the cell surface.The 28 KDa protein may correspond to PEB1, a surface exposed adhesion protein that is the basis of an ELISA kit (Pei et al., 1991).Antibodies to this protein were able to detect 35 C. jejuni and 15 C. coli isolates without cross-reactivity (Pei et al., 1991).Taken all together, the 62, 43, 28 KDa proteins may be useful target molecules for Campylobacter immunoassays and deserve further scrutiny.
Several monoclonal antibodies have been synthesized for Campylobacter detection.Monoclonal antibodies target a single epitope of an antigen, which may limit their use in detection of Campylobacter because the epitope may not be conserved across all Campylobacter species and strains (Rice et al., 1996 ), although they exhibit greater specificity than polyclonal antibodies (Velusamy et al., 2010). Indeed, Wang et al. (2000) reported a lack of success with the following monoclonal antibodies: 1744-9029 and 1744-9006 (Biogenesis Ltd, NH), MAB001 (Harlan Sera Lab, GB), and C65701M (Biodesign International, MD).Our experiences with different monoclonal antibodies (AbD Serotec monoclonal antibodies 1745-00, 1744-8508, 1744-9059, and 1744-9109; USB C1037-02A, C1037-04, and C1037-16) support the premise that monoclonal antibodies are not ideal for ELISA detection of Campylobacter spp.(data not published).In contrast, the monoclonal antibody to the 62 KDa protein, discussed above, was able to detect whole cells at concentrations of 10 5 CFU pe ml -1 with excellent specificity (Lu et al., 1997).Also the monoclonal antibody to the 43 KDa protein was specific for C. coli, bound to the cell surface, and detected 10 3 to 10 4 CFU per ml -1 (Qian et al., 2008a).However, it is not clear whether these antibodies bind to conserved epitopes expressed under different conditions.

Capture of C. jejuni and C. coli in EIAs
The retention of bacteria may be inefficient in EIAs without the use of a capture element such as an antibody as in sandwich EIAs (Hochel et al., 2007) and thus bacteria are typically captured by an antibody in commercial kits.The number of capture antibodies that attach to a microplate may be improved by using biotin-labelled antibodies that can attach to streptavidin coated plates.The use of microplates with bottom filters in conjunction with antibody-capture, such as multi-well filter plates (AcroWell, Pall Corp.), or hydrophobic grid membrane filters may improve capture of antibody-bacterial complexes (Tsai & Slavik 1994;Wang et al., 2000).Another approach is the colony-lift immunoassay.After bacterial colonies are grown on select agar plates, a membrane filter lifts bacteria from the plate for immunoassay.When coupled with an anti-goat polyclonal antibody, the assay was specific and detected all tests strains of C. jejuni, C. coli and C. lari in 18-28 h (Rice et al., 1996).However, testing of this method has not been published to the best of our knowledge.
Novel capture molecules such as polymyxin B sulfate (PMB), which binds to lipid A of the LOS may improve bacterial capture.The limitation of this method is that treatment of bacteria with zwitterionic detergents is necessary (Blais 2005(Blais , 2006;;Brooks et al., 1998).Fukuda et al. (2005 ) demonstrated that PMB capture was superior to antibody capture for detection of Salmonella in chicken.Our laboratory demonstrated that PMB-capture enhances ELISA detection of C. jejuni as indicated by a 10-fold improvement in the limits of detection with both the USB10 and USB14 antibodies (unpublished results).The PMB-capture ELISA with either of these antibodies detected 19 C. jejuni isolates obtained from a variety of sources, including human, pork, turkey and chicken.The minimum cell concentration yielding a positive assay signal was approximately 10 4 -10 5 CFU per ml -1 , and fluorescence intensity values at 10 6 CFU per ml _1 were similar for all isolates with no discernable difference between the two USB antibodies (Fig. 2).The mean signal to noise ratios for the 19 isolates were 16 ± 1 (USB10) and 17±2 (USB14) at 10 6 CFU per ml -1 , while the signal to noise ratios for the other Gram-negative bacteria screened at 10 7 CFU per ml _1 were below 2.5 except for H. pylori (Table 2).Therefore, the assay is specific for Campylobacter.

Potential use of antibodies in biosensors
Antibodies have the potential to be used in biosensors for the detection of specific bacterial pathogens.Biosensors are defined as analytical devices that combine a bioreceptor, which is a biological or biologically derived element, with a physicochemical transducer (Turner & Newman 2005).The term immunosensor describes a sensor that uses antibodies as the sensing element.When antibodies are used in bisosensors, they are the bioreceptors, capturing or sensing elements, and therefore the advantages and limitations in sensitivity and specificity of traditional antibody-based detection methods apply to these biosensors.
The transducers in biosensors are the components in charge of converting the biorecognition event, which happens when antibodies associate to the antigen, into signal that can be quantified.Several different transducers have been developed.In general, transducers can be categorized into four groups: electrochemical, calorimetric, acoustic and optical (Jönsson & Scott 2005).Optical transducers can measure variables such as changes in temperature (thermometric transducers), pressure, flow, etc., and sensors based on optical transducers are called optical sensors.Representative bacterial colonies were suspended in phosphate buffered saline prior to assay.Fluorescence intensity at 10 6 CFU per ml -1 was plotted for both USB10 and USB14 antibodies.Values are means ± SEM for n=5 colonies.The limit of detection was 2.5 greater than background.

Optical sensors
The optical sensors that use labeled antibodies, e.g.antibodies tagged with fluorescent labels, are similar in design to conventional EIAs.An optical sensor that has been used extensively to detect chemical compounds and bacterial pathogens is surface plasmon resonance (SPR).SPR is an optical phenomenon involving excitation of free oscillating metal electrons and it is based on the phenomenon called total internal reflection.When light traveling from a medium more optically dense (refractive index n 1 ) penetrated into a less optically dense medium (refractive index n 2 ), the light is bent away from the normal plane to the boundary (θ 2 > θ 1 ).(θ 2 ) approaches 90º when the incident angle (θ 1 ) increases to a critical angle (θ c ).When the incident angle (θ 1 ) is equal to or greater than the critical angle (θ c ), the light will be internally reflected (Fig 3 A).The electromagnetic field component of the incident light penetrates a short distance (tens of nanometers) into the less optically dense medium to create an exponentially decaying evanescent wave (Fig 3 B).If the incident light is monochromatic and plane-polarized, and a thin film of metal (most frequently gold) is coated at the interface between the two different optically dense media, the photon of the evanescent wave will resonance with free oscillating electrons (plasmons) in the metal film.The evanescent wave can be used to interrogate variations on the surface structure in a distance of up to 300 nm from the surface.The use of SPR as sensor emerged from the realization that SPR signal is sensitive to changes in the refractive index, which is the bending or refraction of a beam of light on entering a denser medium.The refractive index is influenced by the accumulation of mass on the metal surface.The optical excitation of plasmons only "when the energy of the photons of light exactly equals the quantum energy level of the plasmons", a circumstance that is called attenuated total reflection (Mol & Fisher 2010).
Some advantages of optical sensors include a wide dynamic range, the possibility of multiplexing to detect several analytes, and a compact, light built that allows for field applications.Another advantage that most commercial SPRs have is that the sensors are based on an open architecture and therefore the sensing element, the antibodies, can be defined by the final user and can be used without any labels to create a label-free biosensor (Ivnitski et al., 1999;Jönsson & Scott 2005).These types of sensors provide a close monitoring of the antibody/antigen reaction in a real-time manner.However, biosensors still have some limitations.For instance, there are constrains on the quantity of antibodies in an active state to capture the target bacteria that can be immobilized on the surface of a sensor.Increasing the density of antibodies is very important to increase sensitivity, but recent studies using three-dimensional aggregation of immunoglobulin G can increase the amount of antibodies by surface area and may provide a way to develop "high performance antibody biosensors" (Feng et al., 2011).

Conclusion
Current widely-accepted methods for identifying campylobacters in stool and food samples rely on their growth on selective agar plates with and without prior broth enrichment.These methods are labor-intensive and time-consuming.More rapid and accurate methods, including immunoassays, are necessary to identify campylobacters in stool samples, outbreaks investigations and more recently for screening processed poultry under a new performance standard established in the USA.Yet, the development of antibody-based methods for Campylobacter spp.has not received enough attention.At present, none of the commercially-available EIAs are recommended for standalone identification of campylobacters in stool samples, in part because of the limited and conflicting findings regarding their performances.Some of the commercial immunoassays show promise for identification of campylobacters in food although only after broth enrichment for up to 48 hours.More studies of these assays with naturally-contaminated samples from a variety of foods is necessary to validate their performance.Areas of research to improve antibodybased methods should include antibody/antigen characterization, biosensors platforms, and the dovetailing of concentration methods to improve sensitivity.

Fig. 1 .
Fig. 1.Capture of C. jejuni on Microfil V filters prior to polymyxin B EIA detection.Various water source samples (500 ml) were spiked with different concentrations of C. jejuni and then filtered.Net fluorescence intensity was obtained by subtraction of background.Means ± standard error of the means (n= 6 replicates) are plotted as a function of original bacterial concentration.

Fig. 3 .
Fig. 3.The optical bases for surface plasmon resonance.Changes in refracted light (A) and the production of evanescent light (B) to interrogate the thin gold surface.

Test Name Sample Size (% Positive) Sensitivity (95% CI) Specificity (95% CI) Positive Predictive Value Negative Predictive Value Source
).Although the number of false positives was low, (Diergaardt et al., 2003)a & Le Roux 2003) Roux& Lastovica 1998;Lastovica & Le Roux 2003).Although filters may not help isolate Campylobacter spp.from samples with low number of cells such as water samples(Diergaardt et al., 2003), filters help isolate Campylobacter spp.from highly contaminated samples.In enriched samples, the high motility of Campylobacter spp.allows for a relatively low number of cells to go through membrane filters and grow as pure colonies on plate media.Thus, filters are highly sensitive for the isolation of naturally occurring Campylobacter spp.present on retail broiler meat(Speegle et al., 2009).

Table 2 .
The exit angle Signal to noise ratios for the polymyxin B capture ELISA for various Gram-negative bacteria compared to C. jejuni and C. coli.Representative bacterial colonies,suspended in PBS at ~ 10 7 CFU per ml -1 were assayed using the optimised PMB-capture ELISA.Values are means ± SEM for n=5 experiments.The limit of detection was set at 2.5 or 2.5 greater than background.