Ornithobacterium rhinotracheale Infection in Poultry: an Updated Review

Ornithobacterium rhinotracheale is a Gram-negative bacterium of the rRNA superfamily V within the CytophagaFlavobacterium-Bacteroides phylum, which has become an emerging pathogen in the poultry industry. The purpose of this work is to review the current literature on O. rhinotracheale infection in domestic poultry. Despite the difficulties to identify the bacteria using biochemical tests, the polymerase chain reaction (PCR) technique has been successfully used to identify suspected isolates. Moreover, pulsed-field gel electrophoresis (PFGE) has allowed the genetic typing of O. rhinotracheale. Recent investigations carried out in the United States and Argentina described the unusual and extensive β-hemolytic activity of field strains isolated from affected chickens and turkeys. According to the latest serological studies, the causing agent has eighteen serotypes (A to R) and serotype A is the most prevalent among chicken and turkey strains. Gross findings typically found in broiler chickens are unilateral pneumonia, pleuritis, and abdominal airsacculitis with foamy, white yogurt-like exudate, which cause increased condemnation rates at slaughter, whereas those found in turkeys include tracheitis, pneumonia or bronchopneumonia, thoracic and/or abdominal airsacculitis, pericarditis, and peritonitis. Despite the worldwide efforts to control O. rhinotracheale infection, the present situation shows several problems not only because antibiotic treatments appear to be less effective due to the resistance acquired in the poultry industry but also because the vaccines developed so far show variable results in commercial poultry.


History
The first report related to the characterization of O. rhinotracheale was that of Charlton et al (1993). Then, Vandamme et al (1994) described the phylogenetic position and various genotypic, chemotaxonomic, and classical phenotypic characteristics of 21 strains were then described, and assigned the name O. rhinotracheale (Vandamme et al., 1994). However, this bacterium appears to have been isolated before 1993 . In 1991 a new respiratory disease in 28-day-old broiler chickens was observed in South Africa. The birds suffered sneezing, associated with increased mortality and poor performance parameters. At postmortem examination, pneumonia and foamy white and "yoghurt-like" exudate were observed in the abdominal air sacs and the bacteriological study revealed a slow growing, pleomorphic, Gram-negative rod unknown among the bacterial species previously reported (van Empel, 1998;van Empel and Hafez, 1999;Bisshop, 2003). A bacterium similar to Pasteurella spp. isolated from 10-week-old Pekin ducks with respiratory disease in Hungary in 1987(van Empel, 1998van Empel and Hafez, 1999;Pyzik, 2007;, as well as Riemerella anatipestifer-like strains isolated from turkeys suffering respiratory disease in Germany in 1991 and 1992 showed appearance and biochemical properties similar to those of the South African isolates (van Empel, 1998;van Empel and Hafez, 1999;Pyzik, 2007). This bacterium was named Pasteurella-like, Kingella-like, TAXON 28, or pleomorphic Gram-negative rod before the name Ornithobacterium rhinotracheale gen. nov. sp. nov. was suggested (van Empel and Hafez, 1999;. Investigations of German culture collections revealed that O. rhinotracheale had been already isolated from the respiratory tract of 5-week-old turkeys with nasal discharge, facial edema, and fibrinopurulent airsacculitis in 1981 and from rooks in 1983 (van Empel, 1998). In Belgium, France and Israel, O. rhinotracheale had also been isolated before 1990(van Empel, 1998. No isolates of O. rhinotracheale had been reported before 1981 (van Empel, 1998).

Etiology
2.1 Morphology, staining, growth requirements, antigenic structure, and colony morphology Ornithobacterium rhinotracheale is a Gram-negative, non-motile, highly pleomorphic, rod-shaped, and non-sporulating bacterium of the rRNA superfamily V within the Cytophaga-Flavobacterium-Bacteroides phylum. When cultured on solid media, the bacterium appears as short, and plump rods measuring 0.2~0.9 µm in width and 0.6~5 µm in length (van Empel and Hafez, 1999;, and less frequently as long filamentous rods or club-shaped rods (Chin and Charlton, 2008). No structures such as pili, fimbriae, and plasmids or properties such as specific toxic activities have been reported for the species (van Empel and Hafez, 1999;. The use of 5%~10% sheep blood agar plate is recommended for isolation and optimal growth of the causing agent. Ornithobacterium rhinotracheale does not grow on MacConkey agar, Endo agar, Gassner agar, Drigalski agar, or Simmons citrate medium . The bacterium grows aerobically, microaerobically, and anaerobically, but the best growth occurs in air enriched with 7.5~10 % CO 2 at 37℃ (Chin and Charlton, 2008). Under these conditions and 24 h post-incubation, O. rhinotracheale develops pin-point colonies smaller than 1 mm in diameter. After 48 h, the colonies are approximately 1~2 mm in diameter, gray to gray-white, circular, and convex with an entire edge, and some isolates from chickens have a reddish glow. Cultures of O. rhinotracheale have a distinct smell similar to that of butyric acid (van Empel and Hafez, 1999;Chin and Charlton, 2008). Because of the resistance to gentamicin and polymyxin B observed in 90% of O. rhinotracheale field isolates (Vandamme et al., 1994), 5 µL/mL of each antibiotic is recommended to be added to blood agar media for selective isolation of this bacterium (van Empel and Hafez, 1999;Chin and Charlton, 2008;. The use of 10 µg of gentamicin per mL of blood agar medium has also been suggested to isolate O. rhinotracheale from contaminated samples . van Empel and Hafez (1999) and Chin and Charlton (2008) also proposed the use of blood agar plates without antibiotic to prevent missing 10% of the antibiotic-susceptible isolates. Ornithobacterium rhinotracheale was first identified as a non-hemolytic microorganism (van Empel and Hafez, 1999;Hafez, 2002;Canal et al., 2005;Chin and Charlton, 2008;, such as the ATCC 51463 strain of O. rhinotracheale (Tabatabai et al., 2010;Gornatti Churria et al., 2011). However, the presence of extensive and unusual β-hemolytic activity has been recently reported among North American and Argentinean field isolates after the 48-h-period following incubation at room temperature (Walters et al., 2009;Tabatabai et al., 2010;Gornatti Churria et al., 2011). Tabatabai et al (2010) characterized and demonstrated the β-hemolytic activity of O. rhinotracheale isolates using in vitro kinetic hemolysis assays with sheep red blood cells, western blotting with leukotoxin-specific monoclonal antibodies, and isobaric tagging and quantitative analysis of O. rhinotracheale outer membrane protein digest preparation. Moreover, Walters et al (2011) have recently developed an embryo lethality assay to determine potential virulence differences between hemolytic and non-hemolytic O. rhinotracheale strains isolated from turkeys in Virginia´s Shenandoah Valley, USA. In search for the optimal dilution to differentiate the pathogenicity of the isolates, these authors compared mortality patterns for 8 days post-inoculation and found the most drastic differences in the 10 3 dilution groups. The results showed higher mortality numbers in an embryo model caused by hemolytic O. rhinotracheale isolates, which appear to be more virulent than non-hemolytic ones.
Ornithobacterium rhinotracheale is known for its hemagglutinating activity (Soriano et al., 2003;Tsai and Huang, 2006). Vega et al (2008) tested the hemagglutinating activity of serotypes A to I of O. rhinotracheale reference strains by using red blood cells from 15 different species, including avian, mammal, fish, and human erythrocytes, and concluded that rabbit erythrocytes were suitable to test O. rhinotracheale. In contrast, Chernyshev et al (2011) have recently reported the hemagglutinating activity of 19 Russian isolates with chicken and sheep erythrocytes.

Biochemical identification
The results of biochemical tests for the identification of O. rhinotracheale can be inconsistent (van Empel and Hafez, 1999;Chin and Charlton, 2008;. Therefore, Chin and Charlton (2008) proposed the following tests as those with more consistent reactions to identify O. rhinotracheale: oxidase (+), catalase (-), β-galactosidase (+), indole (-), and triple sugar iron agar (no change).  also described a cytochrome-oxidase negative strain isolated from turkeys in Germany.
On the other hand, Thachil et al (2007) studied the enterobacterial repetitive intergenic consensus (ERIC)-PCR and the random amplified polymorphic DNA (RAPD) assay with Universal M13 primer-based fingerprinting techniques to differentiate O. rhinotracheale isolates. To this end, they evaluated A total of 50 field strains and 8 reference strains for genetic differences using the primers ERIC 1R (5´-ATGTAAGCTCCTGGGGATTCAC-3´) and Universal M13 (5´-TTATGTAAAACGACGG CCAGT-3´). M13 fingerprinting revealed different patterns for six reference serotypes of O. rhinotracheale tested, namely, C, D, E, I, J, and K. Ornithobacterium rhinotracheale reference serotypes A and F showed indistinguishable fingerprints with M13 fingerprinting. The ERIC 1R technique discerned only five out of the eight reference serotypes. Distinct fingerprints were also found within the O. rhinotracheale serotypes with both techniques. From 58 isolates of O. rhinotracheale belonging to 8 O. rhinotracheale serotypes that were fingerprinted, 10 different fingerprints were obtained with M13 fingerprinting, whereas six different fingerprints were obtained with ERIC 1R fingerprinting.
The authors concluded that M13 fingerprinting technique was more discriminative in differentiating O. rhinotracheale isolates than the ERIC 1R fingerprinting technique. Moreno et al (2009) genotyped O. rhinotracheale strains obtained from Spanish red-legged partridges with neurological signs, otitis and cranial osteomyelitis, following adaptations of previously reported PFGE assays. The authors incubated isolates on Columbia agar in aerobic conditions at 37℃ for 24 h and prepared DNA plugs from fresh colonies of bacteria suspended into 1 mL of TE1 × buffer with a transmittance of 20%. The suspension was incubated at 37℃ for 10 min, and 60 mL of lysozyme was added. The suspension was mixed with an equal volume of 2% molten agarose, 10% sodium dodecyl sulphate, and recombinant proteinase K. The DNA plugs were incubated at 56℃ for 2 h in TE 1× buffer with 30 mL of recombinant proteinase K, and lysis buffer, and washed six times. The restriction enzymes ApaI and SmaI were used according to the manufacturer's recommendations. DNA fragments were resolved in a 1% gel on a PFGE equipment, CHEF-DR III, at 6 V/cm for 21 h, with switching times ramped from 0.1 to 25 sec at 14℃, with an angle of 120°. A Salmonella enterica serovar Braendorup H9812 restricted with XbaⅠ was included as the internal molecular ladder, and Mid Range PFG Marker Ⅱ was included as a molecular weight ladder for size determinations. The gel was stained with ethidium bromide, and destained in distilled water. The PFGE patterns were examined and the genetic relationship among isolates was evaluated. An epidemiologically unrelated French strain of O. rhinotracheale from a pheasant was added. The study revealed indistinguishable macrorestriction patterns of the O. rhinotracheale Spanish strains with the enzymes ApaI and SmaI and no relation was noted with the restriction pattern of the French isolate.

Serological typing
Eighteen serotypes (A through R) have been differentiated because of the results observed in enzyme-linked immunosorbent assays (ELISAs) and agar gel precipitation tests using boiled extract antigens and monovalent antisera (van Empel et al., 1997;van Empel and Hafez, 1999;Türkyilmaz, 2005;Chin and Charlton, 2008;. Some geographical differences exist in the serotypes found throughout the world. Most of the isolates in the United States and Europe belong to serotype A (Chin and Charlton, 2008), whereas serotype C has been found only found in chickens and turkeys of South Africa and the United States . Serotypes B, D, and E are the predominant ones after serotype A in Europe (Chin and Charlton, 2008). Also, serotype A has been found to be the most prevalent serotype among chicken strains (97%), and turkey strains (61%), (van Empel et al., 1997;. No relation has been found between host specificity and the O. rhinotracheale serotype . Nowadays, the Institute of Poultry Diseases (Free University Berlin, Germany) directed by Prof. Dr. Hafez Mohamed Hafez is the worldwide reference to identify the serotypes of O. rhinotracheale strains (Hafez Mohamed Hafez, 2011, private communications).

Transmission, incubation period, hosts, and clinical signs
Ornithobacterium rhinotracheale spreads horizontally by direct and indirect contact through aerosol or drinking water . Besides its isolation from ovaries, oviduct, hatching eggs, infertile eggs, dead embryos, and dead-in-shell chickens and turkeys (van Empel, 1998;, there are circumstantial evidences of vertical transmission in birds affected by ornithobacteriosis (van Empel, 1998). According to experimental reproductions of the disease, the clinical signs are seen 24~48 h post-inoculation . Ornithobacterium rhinotracheale infection has been reported mostly in broiler chickens and turkeys, and less frequently in other avian species such as pheasants, quails, gray partridges, chukar partridges, red-legged partridges, guinea fowls, ostriches, rooks, pigeons, ducks, geese, and gulls (van Empel and Hafez, 1999;Moreno et al., 2009).
Despite the successful experimental reproduction of O. rhinotracheale infection in broiler chickens without the presence of any concurrent infectious agents or environmental stressors (van Veen et al., 2000b;Pan et al., 2012), the role of O. rhinotracheale as the primary pathogen is still uncertain (Chin and Charlton, 2008). In many reported cases of affected broiler chickens and turkeys, O. rhinotracheale infection played an associated role with other respiratory pathogens such as Escherichia coli, Bordetella avium, Streptococcus zooepidemicus, Mycoplasma gallisepticum, Mycoplasma synoviae, Chlamydophila psittaci, Newcastle disease virus, avian metapneumovirus, infectious bronchitis virus, and Cryptosporidium spp. Sakai et al., 2000;Marien et al., 2005;2007;Thachil et al., 2009;Gornatti Churria et al., 2012;Pan et al., 2012). The severity of clinical signs, the duration of the disease and the mortality caused by O. rhinotracheale outbreaks are extremely variable, and can be influenced by a wide range of environmental factors .
Clinical signs of broiler chickens include depression, decreased food intake, reduced weight gains, nasal discharge, sneezing, and facial edema, and appear at 3~6 weeks of age. Also, sudden death with or without respiratory signs has been reported in chickens with nervous signs (Goovaerts et al., 1998;Chin and Charlton, 2008;.
Broiler breeders aged 20~50 weeks have been seen to be affected by O. rhinotracheale infection in the peak of the laying period or soon before entering the egg production period suffering mild respiratory signs, decreased food intake, variable to low mortality, decreased egg production, decreased egg size, and poor quality of eggshell (Chin and Charlton, 2008;. Commercial 20-to-50-week-old layers have been found to show increased mortality, decreased egg production, and increased numbers of misshapen eggs (Sprenger et al., 2000;Chin and Charlton, 2008).
In turkeys, O. rhinotracheale infections have been detected as early as at two-weeks of age, but, the severity of clinical signs and mortality has been found to be higher in older flocks (Chin and Charlton, 2008;. The normal mortality is approximately 1%~15%, but can be up to 50%. Coughing, sneezing, nasal discharge, followed by severe respiratory disease, sinusitis, dyspnea, reduction in food and water intake, and prostration are the signs involved . In turkey breeder flocks, ornithobacteriosis is associated with a decreased egg production, and an increased number of unsuitable hatching eggs (van Empel and Hafez, 1999). Ornithobacterium rhinotracheale infection has also been found to be the cause of neurological signs or paralysis due to arthritis, meningitis, encephalitis, otitis, and cranial osteitis and osteomyelitis in turkeys, chickens, and red-legged partridges (Goovaerts et al., 1998;van Empel and Hafez, 1999;Moreno et al., 2009).

Experimental reproductions
In order to investigate the role and pathogenesis of O. rhinotracheale in poultry, several works of experimental reproductions by using turkey embryo, chicken, turkey, layer hen, and quail models with or without the involvement of other bacterial and/or viral respiratory pathogens have been reported (Goovaerts et al., 1998;van Empel and Hafez, 1999;van Veen et al., 2000b;Marien et al., 2005;Eroksuz et al., 2006;Thachil et al., 2009;Pan et al., 2012). Intra-tracheal, intra-thoracic, intranasal, intravenous, and oculonasal administration routes were used to inoculate birds with the causing agent. Results show that neither the origin nor the serotypes of the strains appear to have an effect on the pathogenicity in experimental trials (van Empel and Hafez, 1999). Apart from the turkey embryo model described by Walters et al (2011), no experimental poultry trials have been used to study possible pathogenicity differences between hemolytic and non-hemolytic O. rhinotracheale field strains (Louisa Tabatabai, 2010, private communications). Previous viral infection with Newcastle disease virus, infectious bronchitis virus, or avian metapneumovirus was necessary for the presence of gross and/or microscopic lesions associated with O. rhinotracheale during experimental infections van Empel and Hafez, 1999;Marien et al., 2005;Eroksuz et al., 2006).
Although some bacteria such as E. coli and B. avium are considered less effective triggers than viruses for O. rhinotracheale infections (van Empel and Hafez, 1999;Thachil et al., 2009), Pan et al (2012 have recently reported that the association between O. rhinotracheale and S. zooepidemicus causes higher mortality and severe bronchial lesions than that observed in chickens experimentally infected with O. rhinotracheale or S. zooepidemicus alone. However, O. rhinotracheale has been found to be able to cause lesions in chickens and quails, without previous viral priming (Goovaerts et al., 1998;van Empel and Hafez, 1999;van Veen et al., 2000b;Eroksuz et al., 2006;Pan et al., 2012).

Gross pathology
The most common macroscopic findings in broiler chickens are unilateral pneumonia, pleuritis, and abdominal airsacculitis with foamy, white yogurt-like exudate . Other respiratory lesions, such as catarrhal tracheitis (Gavrilović et al., 2010) and bilateral exudative pneumonia (Gornatti Churria et al., 2011), has also been found in chickens affected by ornithobacteriosis.
Condemnation rates of 60% in broilers at slaughter due to airsacculitis in 84% of the birds examined, and due to pericarditis and pneumonia in a few birds have been reported to be associated with O. rhinotracheale infection (van Veen et al., 2000a). In addition, more than one third of the respiratory lesions in broiler chickens at slaughter age have been reported to be caused by O. rhinotracheale infection, indicating the wide distribution of this bacterium in the broiler industry of Europe (van Veen et al., 2005).
Uncommon lesions such as subcutaneous edema of the skull with severe osteitis and osteomyelitis together with encephalitis without the involvement of the respiratory tract have been described in 28-day-old broiler chickens (Goovaerts et al., 1998).
In turkeys, unilateral and bilateral consolidations of lungs due to pneumonic or bronchopneumonic lesions with fibrinous exudate of the pleura have been found Thachil et al., 2009;Tabatabai et al., 2010). Mild or severe tracheitis, fibrinosuppurative thoracic and/or abdominal airsacculitis, pericarditis, and peritonitis have also been described in turkeys Chin and Charlton, 2008;Thachil et al., 2009). Swelling of the liver and spleen, degeneration of the heart muscle, and infection of vertebrae and joints have been observed in some cases of O. rhinotracheale infection in turkeys .

Histopathology
Most histological lesions can be found in lungs, pleura, and air sacs . Lung lesions caused by O. rhinotracheale are similar to those produced by Pasteurella multocida (Fletcher et al., 2008). They are characterized by large and coalescing areas of necrosis centered in the lumen of parabronchi, filled with degenerated and necrotic heterophilic infiltrate or fibrinous exudate. Collections of fibrin with macrophages and heterophils occupying the interstitial tissues and air passages are also found Fletcher et al., 2008). According to Fletcher (2010), fibrinoheterophilic diffuse pneumonia in turkeys is suspected to be caused by O. rhinotracheale infection. Pleura and air sacs can be thickened with interstitial fibrin, diffuse heterophilic infiltrate, necrotic foci, and fibrosis .

Immunohistochemistry (IHC)
In spite of the low immunogenicity capacity of O. rhinotracheale (Paul van Empel, 2011, private communications), the immunohistochemical staining is a valuable and sensitive method to demonstrate the presence of this bacterium among the tissues of broiler chickens with respiratory signs in field trials (Hafez, 2000).
Immunohistochemistry has been reported with the use of a monovalent rabbit antiserum against the O. rhinotracheale strain named B3263/91 belonging to serotype A. The antiserum was prepared by injecting specific-pathogen-free New Zealand white rabbits twice in a 3-week interval, using an oil adjuvant bacterin. Serum was collected four weeks after the second injection .
A second option of antiserum elaboration recently suggested is to make a polyclonal antiserum. The use of three oil adjuvant bacterin doses of 0.5 mL each, administered subcutaneously in New Zealand white rabbits with 5~6 weeks of interval, has been recommended (Paul van Empel, 2011, private communications).
Detection of O. rhinotracheale by IHC has been carried out following the peroxidase anti-peroxidase staining method. Results of experimental trials have shown minimal, acute and short-lived microscopic air sac lesions associated with focal bacterial attachment to the epithelium of the respiratory tract of chickens exposed to O. rhinotracheale aerosol without prior viral infection .

Serology
Serological tests are useful for flock monitoring and/or the diagnosis of O. rhinotracheale infection . The advantage of serology over bacteriological examination is that antibodies persist for several weeks after infection and bacterial shedding. However, O. rhinotracheale excretion and its antibody response may also be affected by a number of factors such as antibiotic therapy and vaccination (Hafez, 2002). ELISAs, serum plate agglutination tests, and a dot-immunobinding assay have been reported for the detection of O. rhinotracheale antibodies (Back et al., 1998;van Empel and Hafez, 1999;Sakai et al., 2000;Erganiş et al., 2002;Hafez, 2002;Canal et al., 2003a;Allymehr, 2006;Eroksuz et al., 2006;Chansiripornchai et al., 2007;Ghanbarpour and Salehi, 2009;Uriarte et al., 2010;Walters et al., 2010).

ELISA
ELISA allows detecting the presence of antibodies against O. rhinotracheale in one-day-old chicks and egg-yolk as well as in birds with clinical signs (van Empel and Hafez, 1999). ELISAs have been developed using different serotypes and extracted antigens of O. rhinotracheale , and both self-made and commercially available ELISAs (Biochek, Inc., USA, and IDEXX Laboratories, Inc., USA) are used to detect antibodies against most O. rhinotracheale serotypes (Hafez, 2002).
In Thailand, 68% of broiler flocks were found to be positive. This value was higher than the 60% and 26% previously observed in Brazil and Germany, respectively (Chansiripornchai et al., 2007). The 32.5% prevalence of the individual broiler sera analyzed in Thailand was greater than the 6.52% found in Brazilian chickens and the 9.4% found in German chickens (Chansiripornchai et al., 2007). The serum analysis of the individual broiler breeder sera showed 87.8% of positivity in Thailand, which was lower than the 94.6% found in broiler breeders in Brazil and higher than the 13.9% found in broiler breeders in Japan (Chansiripornchai et al., 2007).
The differentiation between serological responses caused by hemolytic and non-hemolytic O. rhinotracheale infections was studied by Walters et al (2010). These authors tested a total of 1 200 serum samples from market hens with 12~14 weeks of age by the use of a specially designed ELISA test. Although 41% of the flocks were positive for both hemolytic and non-hemolytic O. rhinotracheale infections, 49.5% were positive for hemolytic O. rhinotracheale isolates and 64.6% were positive for non-hemolytic O. rhinotracheale isolates. Back et al (1998) developed and standardized the SPAT for rapid detection of antibodies against O. rhinotracheale infection. The antigen was prepared by growing O. rhinotracheale on 5% sheep blood agar at 37℃ for 24 h in 5% CO 2 , considering maximum growth of the bacterium at 24 h. The cells were harvested by scraping the bacterial growth, and resuspended in distilled water. The bacterium was inactivated with 0.8% formalin, and incubated at 37℃ for 3 h. The cells were harvested by centrifugation at 3 000 × g for 10 min, and the pellet was resuspended in distilled water. Serum samples from chickens and turkeys experimentally infected were tested, and antibodies against O. rhinotracheale were detected by SPAT in both avian species, whereas the serum samples from not exposed birds remained negative (Back et al., 1998). Erganiş et al (2002) developed a DIA and compared it with agglutination assays by testing serum samples from turkeys with respiratory signs. For DIA antigen, O. rhinotracheale was cultured on 5% sheep blood agar base at 37℃ for 48 h in 10% CO 2 , and harvested with phosphate buffer solution (PBS) containing 0.3% formaldehyde. Then, 1 mL of O. rhinotracheale suspension of 0.2 g wet weight/mL PBS was boiled for 60 min at 100℃. After centrifugation (2 500 × g for 45 min), the supernatant was passed through a 0.22 µm pore-size filter and used as antigen. Following standardization of the optimal dilutions of antigen, antiserum and conjugate, these were used in DIA tests. The authors concluded that the sensitivity of the DIA appeared to be lower than the agglutination assays studied (Erganiş et al., 2002).

Antimicrobial susceptibility
Antimicrobial resistance profiles of O. rhinotracheale isolates from broiler chickens, turkeys, and layer hens have been reported worldwide. However, the antibiotic treatment against this bacterium shows several difficulties, because of the variable susceptibility of the strains associated with the regime used by the poultry industry of the geographical area studied . Ornithobacterium rhinotracheale infections have become more common in the poultry industry and the treatment with antibiotics has become less effective due to an increased pathogenicity, an increased burden of infection, and/or an increased level of acquired antibiotic resistance (van Veen et al., 2001;Malik et al., 2003).
The susceptibility of O. rhinotracheale to antibiotics is very inconsistent and appears to depend on the source of the strains analyzed (van Empel and Hafez, 1999;Banani et al., 2004).
Although no standard methods are known for in vitro antibiotic susceptibility tests of O. rhinotracheale, several reports have followed the suggestions of the Clinical and Laboratory Standard Institute (CLSI) previously known as National Committee for Clinical Laboratory Standards (NCCLS) for fastidious Gramnegative microorganisms (Ak and Turan, 2001;Malik et al., 2003;Tsai and Huang, 2006;Murthy et al., 2008a;Gornatti Churria et al., 2011;. Works of antimicrobial susceptibility of O. rhinotracheale have been reported in the Middle-East (Ak and Turan, 2001;Banani et al., 2004) and Asian countries (Tsai and Huang, 2006;Murthy et al., 2008a). Ak and Turan (2001) tested eight antibiotics using the broth microdilution method against 11 strains of O. rhinotracheale isolated from broiler chickens in Turkey and found that oxytetracycline was the most effective antibiotic, with minimum inhibitory concentrations (MICs) ≤ 0.125~1 μg/mL, and that all isolates were resistant to gentamicin and neomycin. All field isolates were found to be sensitive to doxycycline. Except for one field isolate, the remaining ones were resistant to danofloxacine. Tilmicosin (MICs ranging from ≤ 0.5~4 μg/mL), erythromycin (MICs ranging from ≤ 0.5~8 μg/mL), and penicillin G (MICs ranging from ≤ 4~16 μg/mL) caused effective inhibition of Turkey field isolates but with higher antibiotic concentrations. Banani et al (2004) isolated 105 strains from chicken flocks and tested them against 19 antibiotics using the disc diffusion method. All the isolates were susceptible to tiamulin, and most of them were completely susceptible to chloramphenicol and the combination of lincomycin and spectinomycin. All the isolates were resistant to trimethoprim sulfa, colistin, and neomycin. Most of the field isolates were completely resistant to gentamicin, lincomycin, erythromycin, tetracycline, and enrofloxacin. This antibiotic was the antibiotic commonly used in the Iranian poultry industry during the period of time studied. Tsai and Huang (2006) obtained a total of 40 O. rhinotracheale Taiwanese field isolates from chickens (28 isolates), and pigeons (12 isolates), and used the disc diffusion method to study antimicrobial susceptibility patterns. Most of the chicken isolates (> 80%) were sensitive to amoxicillin, ampicillin, penicillin, and oxytetracycline, and resistant to clindamycin, erythromycin, and trimethoprim sulfa. There were significant differences in the resistance rates to clindamycin, erythromycin, gentamicin, and tetracycline between the pigeon and chicken isolates.
In India, Murthy et al (2008a) tested eighteen isolates from laying hens against twenty antibiotics and antibacterial agents by the disc diffusion method. These authors found that all the O. rhinotracheale field isolates were resistant to amikacin, cloxacillin, trimethoprim sulfa, gentamicin, metronidazole, and triple sulfa, and sensitive to amoxicillin, ampicillin, chloramphenicol, ciprofloxacin, doxycycline, enrofloxacin, erythromycin, oxytetracycline, and penicillin G. Susceptibility against cephalexin, norfloxacin, pefloxacin, streptomycin, and furazolidone was variable.
In Mexico, Soriano et al (2003) studied the MICs of ten Mexican field isolates together with reference strains from South Africa, Europe and the United States. Ten antibiotics were used following the microdilution broth method. Amoxicillin, enrofloxacin, and oxytetracycline inhibited all reference strains at 2~64 mg/mL, 4~64 mg/mL, and 4~32 mg/mL, respectively. For all the reference strains, MIC values of ≥ 128 mg/mL were recorded for gentamicin, fosfomycin, trimethoprim, sulfamethazine, sulfamerazine, sulfaquinoxaline, and sulfachloropyridazine. Amoxicillin, enrofloxacin, and oxytetracycline inhibited the Mexican isolates at 32 mg/mL to 128 mg/mL, 4 to > 128 mg/mL, and 8 to > 128 mg/mL, respectively. For all the Mexican isolates, MIC values of ≥ 128 mg/mL were recorded for gentamicin, fosfomycin, trimethoprim, sulfamethazine, sulfamerazine, sulfaquinoxaline, and sulfachloropyridiazine. Soriano et al (2003) observed no improvement in suspected O. rhinotracheale outbreaks in broiler chicken flocks when fosfomycin or gentamicin was used, in accordance with the marked (≥ 128 mg/mL) resistance found in vitro of the Mexican isolates tested against these two antibiotics. Malik et al (2003), on the other hand, obtained 125 field strains of O. rhinotracheale from lung tissues, tracheal swabs and sinus exudate swabs from Minnesota turkeys with respiratory disease during 1996-2002 in the United States. Field strains were tested by using the disc diffusion method following the CLSI guidelines. Most of the field isolates were sensitive to clindamycin, erythromycin, spectinomycin, and ampicillin. Resistance against sulfachloropyridiazine decreased from 1996 to 2002, whereas that gentamicin, ampicillin, trimethoprim sulfa, and tetracycline increased. The annual trend slopes for these antibiotics were 7.36%, 3.02%, 2.43%, and 1.95%, respectively. Resistance against penicillin remained constant, with a trend slope of only 0.54% per year. Therefore, based on their results, Malik et al (2003) suggested the need of continued antibiotic resistance monitoring of O. rhinotracheale and the establishment of baseline resistance patterns.
The European work conducted by Devriese et al (1995) was the earliest worldwide report describing the antimicrobial resistance of O. rhinotracheale field isolates. Fourteen isolates were obtained from gallinaceous birds (chickens, turkeys, a guinea fowl, and a partridge) from Belgian farms. The microdilution broth method was performed, and the results showed susceptibility to enrofloxacin, and acquired resistance to lincomycin, tylosin, doxycycline, and flumequine. Some years later, these authors described the MICs of 10 antibiotics for 45 strains of O. rhinotracheale isolated from Belgian broiler chickens from 1995 to 1998. All the strains were susceptible to tiamulin and resistant to lincomycin, ampicillin, and ceftiofur as well. Less than 10% of the strains were susceptible to tylosin and spiramycin, and a few strains were susceptible to enrofloxacin and doxycycline (Devriese et al., 2001).
In the Netherlands, van Veen et al (2001) obtained a total of 395 Dutch field isolates from affected broiler chickens from 1996 to 1999 and initially tested them against amoxicillin, tetracycline, enrofloxacin, and trimethoprim sulfa by the agar gel diffusion test. The susceptibility to amoxicillin and tetracycline decreased in successive years from 62% to 14%, and four strains were sensitive to enrofloxacin or trimethoprim sulfa. Seven alternative antibiotics (erythromycin, penicillin, gentamicin, tilmicosin, tylosin, ceftiofur, and clavulanic acid-potentiated amoxicillin) were later used to test 12 multiresistant isolates of O. rhinotracheale. Except for the susceptibility against clavulanic acid-potentiated amoxicillin observed in the isolates studied, all of them were resistant to the rest of antibiotics used.
In Russia, Chernyshev et al (2011) have recently identified and characterized 19 Russian isolates from chickens and turkeys and found resistance to amikacin, penicillin, gentamicin, clindamycin, norfloxacin, erythromycin, and trimethoprim sulfa.
In summary, European, American, and Asian studies from 1995 to 2012 have described the resistance of worldwide O. rhinotracheale field strains against the following antibiotics: gentamicin, neomycin, danofloxacine, trimethoprim sulfa, colistin, lincomycin, erythromycin, tetracycline, enrofloxacin, clindamycin, amikacin, cloxacillin, metronidazole, triple sulfa, fosfomycin, sulfamethazine, sulfamerazine, sulfaquinoxaline, sulfachloropyridazine, ampicillin, tylosin, doxycycline, flumequine, ceftiofur, penicillin, amoxicillin, tilmicosin, tylosin, and norfloxacin (Devriese et al., 1995;Ak and Turan, 2001;van Veen et al., 2001;Malik et al., 2003;Soriano et al., 2003;Banani et al., 2004;Tsai and Huang, 2006;Murthy et al., 2008a;Chernyshev et al., 2011;Gornatti Churria et al., 2011;. Only a few works have described some of the responsible mechanisms associated with antimicrobial resistance of O. rhinotracheale (Devriese et al., 1995;Marien et al., 2006). Devriese et al (1995) demonstrated the presence of the enzyme β-lactamase among Belgian field strains. In a later study, β-lactam antibiotics such as penicillin G, ampicillin, and ceftiofur were reported to be less effective when they were exposed against those selected strains (Devriese et al., 2001). Marien et al (2006) investigated the antimicrobial resistance of O. rhinotracheale isolates to enrofloxacin, particularly its association with the amino acid changes at position 87 of the GyrA subunit. To this end, they used O. rhinotracheale strains with increased MIC values to enrofloxacin, and isolated either from field cases or from turkeys treated with enrofloxacin under experimental conditions. They observed point mutations in amino acids at positions 83 or 87 of the GyrA subunit, in the field strains less susceptible or resistant to enrofloxacin. The isolates with reduced susceptibility following experimental enrofloxacin treatment showed a constant point mutation at nucleic acid position 646 of gyrA resulting in an amino acid change at position 87 of the GyrA subunit (Marien et al., 2006).

Vaccination
The best strategy for the control or prevention of O. rhinotracheale infection is probably vaccination, because most worldwide O. rhinotracheale isolates have acquired resistance against the antibiotics regularly used in poultry (Schuijffel et al., 2006b). However, in spite of the availability of autogenous vaccines, economic losses related to O. rhinotracheale infections in the poultry industry are estimated in hundreds of millions of dollars annually in the United States (Tabatabai et al., 2010). Bacterins, live vaccines, and subunit recombinant vaccines have been developed and reported, with variable results for the control of experimental and natural infections associated with O. rhinotracheale (van Empel and van den Bosch, 1998;Cauwerts et al., 2002;Lopes et al., 2002;Bisshop, 2003;Chansiripornchai, 2004;Schuijffel et al., 2006a;2006b;Murthy et al., 2007).
Injectable and inactivated vaccines were found to be impractical for commercial broiler flocks (van Empel and van den Bosch, 1998), whereas, autogenous bacterins were successfully used for the control of O. rhinotracheale outbreaks in turkeys in Israel . causing enough protection in the progeny against experimental challenges (van Empel and van den Bosch, 1998;Cauwerts et al., 2002;Bisshop, 2003). In addition, Cauwerts et al (2002) found lower condemnation rates, lower mean mortality and higher mean production index in the progeny from vaccinated broiler breeders. Indeed, De Herdt et al (2012) have recently reported the improved performance of broiler chickens derived from broiler breeders vaccinated with Nobilis ® OR inac (Merck Animal Health, USA and Canada). These authors demonstrated a significant 22.3% lower broiler loss and a significant 3.9% higher production index in the broiler progeny of the vaccinated broiler breeder flocks. Murthy et al (2007) then developed inactivated vaccines in order to find the bacterin candidate in pullets. To this end, they produced eighteen different vaccines with different inactivating substances (formalin and thiomersal) and with or without adjuvant (mineral oil, alum, and aluminum hydroxide gel). The experimental trials showed that bacterin in mineral oil adjuvant induced the highest serological response and a significant decrease in macroscopic lesions such as pneumonia and airsacculitis in the vaccinated group of birds (Murthy et al., 2007).
Other authors also developed subunit recombinant vaccines against O. rhinotracheale infection (Chansiripornchai, 2004;Schuijffel et al., 2006a;2006b). The genes encoding eight cross-reactive antigens were amplified, cloned in an expression vector, and expressed in E. coli and then used as subunit vaccine. The protection against an experimental challenge composed of homologous and heterologous serotypes of O. rhinotracheale was also evaluated. As a result, the production of reactive antibodies against the recombinant proteins on Western Blot assays provided homologous and heterologous protection in the challenged group of chickens . Although the eight previously reported antigens were not expressed by all O. rhinotracheale serotypes, a four component subunit recombinant vaccine was able to protect against O. rhinotracheale heterologous infection challenge (Schuijffel et al., 2006b).
Another candidate vaccine was the temperature-sensitive mutant strain of O. rhinotracheale considered by Lopes et al (2002). Turkeys aged one-day-old were vaccinated using drinking water and via oculonasal instillation. The trial showed successful results not only because the strain showed the ability to colonize the upper respiratory tract and was recovered for 13 days after administration, but also because the humoral response was observed only in the vaccinated group of turkeys compared with the unvaccinated control group (Lopes et al., 2002). * The first author would like to dedicate this article to the memory of his mother