Get access to all handy features included in the IVIS website
- Get unlimited access to books, proceedings and journals.
- Get access to a global catalogue of meetings, on-site and online courses, webinars and educational videos.
- Bookmark your favorite articles in My Library for future reading.
- Save future meetings and courses in My Calendar and My e-Learning.
- Ask authors questions and read what others have to say.
Bacterial Infections Including Mycoplasmas
Get access to all handy features included in the IVIS website
- Get unlimited access to books, proceedings and journals.
- Get access to a global catalogue of meetings, on-site and online courses, webinars and educational videos.
- Bookmark your favorite articles in My Library for future reading.
- Save future meetings and courses in My Calendar and My e-Learning.
- Ask authors questions and read what others have to say.
Streptococcus equi and Rhodococcus equi are primary pathogens causing well-recognised diseases with significant impacts on the economics of horse keeping and on equine welfare. Streptococcus pneumoniae and Mycoplasma felis are less well recognised even though they have been particularly associated with respiratory tract diseases in field cases and in epidemiological studies. Furthermore, Koch’s postulates have been fulfilled for these latter two bacteria, including the ability to reproduce disease experimentally, establishing that they are significant equine pathogens. The types of epidemiological and case studies that associate S. pneumoniae and M. felis with in lower respiratory tract disease have even more strongly suggested significant and prevalent pathogenic roles for Streptococcus zooepidemicus and Actinobacillus/Pasteurella spp. However, due to the very high prevalence of naturally occurring disease involving these bacteria, it has not been possible to identify a suitable uninfected naïve population of horses that would enable all of Koch’s postulates to be tested. Their significance, therefore, remains to be finally determined or confirmed. Nevertheless, the combination of the weight of epidemiological evidence suggestive of their involvement, the very high prevalence and age related incidence and the absence of a better explanation for the aetiology of disease in animals from which they are isolated, suggests that they are also causes of highly important pathogenic infections. The clinical and histopathological signs associated with infection by all of these bacteria, disease diagnosis and epidemiology, pathogenesis and prospects for control are discussed.
Strangles results from nasopharyngeal infection by Streptococcus equi  that spreads rapidly to the lymph nodes of the head. Multiplication of S. equi and early lymphadenitis proceeds, unhindered by a massive infiltration by polymorphonuclear leukocytes into the lymph nodes, until an abscess capsule forms along with a sinus tract in order to drain pus to the nearest site of egress through either the skin or upper respiratory tract mucosa (Fig. 1 illustrates the final stages of this drainage process). This process including one or more nodes can last for 2 - 3 weeks and is usually associated with depression, loss of appetite, pyrexia, mucopurulent nasal discharge and inspiratory dyspnea. Abscesses can form in other body organs and their rupture is fatal in up to 10% of cases (a form known as "bastard strangles"). Purpura hemorrhagica (petechial haemorrhage associated with oedema of the limbs, eyelids and gums) may occur  in association with circulating antibody complexes with S. equi M-like protein . The peripheral oedema can be so extreme that circulatory failure and death ensue.
Figure 1. A Thoroughbred foals with pus draining from an S.equi infected abscess sinus through the skin.
S. equi (officially designated S. equi subspecies equi) is a member of the Lancefield Group C streptococci. On the basis of DNA hybridisation studies S. equi is regarded as the archetype for S. zooepidemicus (S. equi subspecies zooepidemicus; ), which causes a number of primarily suppurative mucosal diseases in several mammalian species. More recent studies employing multilocus enzyme electropherotyping  and sequencing of the intergenic spacer for the16S and 23S RNA genes suggest that S. equi is in fact a subtype of S. zooepidemicus . On blood agar both are haemolytic and classically S. equi is described as being highly capsulated giving it a honeydew appearance although approximately half of fresh clinical isolates appear indistinguishable from S. zooepidemicus. Identification of S. equi depends on testing for the ability to ferment lactose, sorbitol and trehalose. A negative result in all of these tests is taken as an identification ofS. equi. Typically, only S. zooepidemicus and S. equisimilis, another Lancefield Group C streptococcus, ferment one or more of these sugars, although atypical isolates of S. equi have been characterised that ferment one or both of lactose and trehalose .
The severity of strangles makes it one of the most feared diseases particularly since it is so prevalent; almost 30% of infectious disease episodes reported to the International Collating Centre is strangles making it one of the most frequently encountered equine clinical problems . Established outbreaks may last for months or even years particularly in large horse populations with frequent new arrivals that provide a supply of susceptible animals.
Treatment of strangles is usually ineffectual even though S. equi is susceptible to most antibiotics in the laboratory. During treatment, clinical signs have been reported to abate, and the incidence of new cases decline. However, when treatment is finished the clinical signs often return and a surge of new cases emerges, perhaps because the abscess or its surrounding capsule lacks sufficient vascularity to enable antibiotic penetration to therapeutic levels and because antibiotic treatment may not eliminate mucosal colonisation.
Epidemiological Control of Strangles
Control of strangles after outbreaks usually relies on a 4-week quarantine period to give time to detect intermittent shedding of S. equi that occurs with some convalescent cases; three negative weekly nasopharyngeal swabs are taken as minimal evidence of freedom from infection [9,10]. However, animals from over 50% of outbreaks carried S. equi for more than 6 weeks mainly in the guttural pouches [11,12]. Furthermore, carriers were apparently healthy and were negative for S. equi when serial nasopharyngeal swabs were collected and cultured over weeks or even months. However, these carriers were readily identified by culture of lavages collected endoscopically from the guttural pouch. Fig. 2 depicts a chondroid seen by endoscopy and Fig. 3 is a cross section through a similar chondroid revealing the presence of many Gram positive streptococci. Blood agar inoculated with material from the centre of the other half of this chondroid revealed a pure culture of S. equi. Failure of conventional techniques of culture from nasopharyngeal swabs is the probable explanation for carriers escaping detection and subsequently introducing strangles to herds of susceptible animals.
Figure 2. A chondroid in the guttural pouch of a carrier of S.equi seen via an endoscope.
Figure 3. A cross section through a guttural pouch chondroid from a carrier of S.equi. Note the many Gram positive streptococcal cells lying within fissures running through the denatured proteinaceous core of the chondroid.
Horses that are without prior contact with strangles do not carry S. equi. Consequently, effective carrier detection and elimination of infection, combined with disinfection and pasture management to counter transmission through fomites could prevent further outbreaks. Widespread guttural pouch endoscopy of convalescent cases is not practicable and a more sensitive method of detecting S. equi in nasopharyngeal swabs was needed to establish the extent of pathology and the presence of infection to better enable and direct treatment. Figure 4 depicts a collection of chondroids that were successfully removed, with the aid of a polyp basket and the patient skill of the veterinarian, from an infected guttural pouch through the Eustachian tube of a carrier that was subsequently effectively treated with antibiotics.
Figure 4. All of the chondroids that were collected from the guttural pouch of an S.equi carrier and the polyp basket that was used to manipulate them through the Eustachian tube with the aid of an endoscope.
Polymerase chain reactions (PCR) tests can specifically amplify a segment of DNA to detectable amounts so that highly sensitive S. equi detection is feasible even from dead organisms draining into the pharynx from the guttural pouches as an aid to presumptively identifying carriers. Large numbers of S. equi could be dislodged from inspissated pus in the guttural pouch by washing with saline  so it seemed likely that some would find their way to the nasopharynx where they could be detected by PCR. A modification of a PCR based on the M-like protein , supplemented by bacterial culture from nasopharyngeal swabs presumptively detected over 90% of established guttural pouch carriers whereas culture alone would have detected less than 60% . Unfortunately this rate of detection still depended on the results from a series of three nasopharyngeal swabs. Following treatment to eliminate infection some animals remained PCR-positive for many months emphasising the need to confirm carriage by culture from guttural pouch lavages collected by endoscopy. Endoscopy of potential carriers would, in any case, be needed to remove inspissated pus or chondroids and to instil antibiotics as the first step in treatment to eliminate the infection .
Almost one quarter of outwardly healthy carriers were colonised by variant S. equi expressing a truncated form of the surface M-protein from a gene containing a large in-frame deletion beginning just after the signal sequence, controlling surface transport and expression, and comprising approximately 20% of the complete gene . Conversely, less than 1% of isolates from active cases had the truncated gene. This raised the possibility that the variants may have arisen to escape immunity developing in the carrier and, since very few cases were caused by the variants, that they could be of reduced virulence but might still be transmitted and play an important role in elevating herd immunity. However, experimental infection with one of these variants was found to cause typical strangles (author; unpublished results) emphasising the need to regard variant S. equi as a potential source of new outbreaks. Interestingly, the M-protein gene deletion corresponded with the region of the protein identified as being responsible for binding to fibrinogen , an activity associated with increased resistance to phagocytosis , and to which a large proportion of opsonic antibodies in convalescent sera are directed . This could explain the basis of the selective pressure in carriers in which variants arise, with S. equi escaping the effects of opsonic antibody. However, the resistance to phagocytosis conferred by fibrinogen binding would presumably play only a minor role in strangles pathogenesis given the ability of the variant to cause typical strangles.
Early attempts to treat chronic guttural pouch carriers suggested that systemic antibiotics may be effective in animals without empyema but not so for those with empyema, concretions or chondroids .
Successful methods for the elimination of S. equi infection combined the removal of pus and chondroids with the local instillation of large doses of penicillin suspended in a gelatine solution, to increase viscosity and improve retention of the application in the guttural pouches, and a 7 days of systemic penicillin .
Alarmingly, during the course of the studies of S. equi carriers it was realised that over 75% of outbreaks result in the presence of one or more apparently silent carriers that are still difficult to detect and treat. Across all outbreaks investigated by guttural pouch endoscopy this was equivalent to approximately 10% of all cases and contacts becoming carriers. While the PCR test and treatment provide the potential to reduce problems in certain situations, the practical difficulties and expense associated with multiple nasal sampling and continued requirement for some endoscopy means that this approach does not provide a long-term solution for most horses. An effective and safe strangles vaccine is currently, therefore, the only foreseeable practical and affordable long-term solution for controlling this disease for the vast majority of horses.
Unfortunately, bacterin vaccines, based on whole bacterial cells produced from laboratory cultures killed by heat or chemical treatment have not been sufficiently effective [20,21], perhaps because production and inactivation methods reduced the immunogenicity of protective antigens. Furthermore, many protective immunogens of other bacterial pathogens are expressed poorly or not at all by laboratory grown cells and non-protective immunodominant antigens may induce a response diverting immunity away from the protective immunogens. Bacterin vaccines are used in some countries when they are faced with such a serious disease as strangles and the vaccines are perceived to have benefits based on anecdotal or circumstantial evidence, despite the lack of well designed and controlled studies formally demonstrating efficacy. Conversely, the lack of proven efficacy may explain why strangles bacterins are not licensed for use in many other countries.
Sub-unit vaccines are based on antigens extracted from laboratory cultures of the pathogen and separated from potentially reactogenic material, or produced as recombinant proteins from genetically cloned genes. For this, the antigens must first be identified and characterised as being both safe and sufficiently protective. Defined strangles vaccines have concentrated on a single cell-surface antigen, the S. equi M protein (SeM), because it elicits strong opsonic antibody responses . However, in vivo these antibodies are only poorly or not at all protective . Opsonic anti-SeM antibodies might make a useful contribution to immunity but alone they are unlikely to provide the basis for an effective strangles vaccine  in spite of speculation that immunity to strangles might depend on local mucosal responses to SeM . There is no direct evidence that mucosal SeM-specific antibodies are any more protective than SeM-specific opsonic antibodies. Nevertheless, SeM-based vaccines are used in some regions of the world but, like bacterins, none have worked sufficiently well for them to obtain a licence in many other countries.
Other subunits of S. equi are under investigation with the eventual objective of identifying the basis for effective vaccination. These include the hemolysin , fibronectin binding proteins [27-29], a metal binding and adhesin homologue of PsaA of S. pneumoniae , superantigens  and a hyaluronate associated protein . Currently, none of these studies is sufficiently well advanced to assess the potential for any of these approaches to underpin the development of an effective vaccine.
An attenuated live S. equi vaccine, was launched in the USA in 1998, with marketing literature claiming that it worked so well that it would no longer be necessary to control strangles by isolating animals with the disease. Live vaccination involves inoculation with a strain of the pathogen intended to be sufficiently attenuated so that it does not cause clinical disease whilst maintaining sufficient viability to induce a protective immune response. Most of the antigens encountered during natural infection will be produced, thus improving the prospects for protective immunity, provided the vaccine does not express products diverting immunity into unprotected or even harmful directions. Experience of field use of the live strangles vaccine has raised important questions about its safety. Reactions including nasal discharge, abscessation of lymph nodes and other sites, allergic reactions, systemic responses and purpura-like signs have been reported although at a lower frequency than during an outbreak of naturally occurring strangles. Additionally, S. equi abscesses have occurred at the site of intramuscular injection given immediately after intranasal administration of this vaccine, further demonstrating the potential of this product to produce strangles-like signs in horses.
Developing live vaccines requires a degree of attenuation to ensure safety without eliminating sufficient survival in the host to induce protective responses. These objectives can act contrary because attenuation is often based on removal of factors that play important roles in both virulence/persistence within the host and protective immunogenicity. Although, attenuation can be obtained by identifying naturally avirulent variant strains or by chemical mutagenesis more recently it is being achieved by introducing defined deletions into virulence factors. The greatest threat of the former strategies, if they were used for the live strangles vaccine, is that the basis of avirulence is not known and reversion to full virulence could easily occur. Furthermore, there may not be a good test to discriminate the vaccine strain circulating amongst the vaccinated population and their contacts from the naturally occurring pathogen, greatly confusing efforts to control the disease epidemiologically.
The precedent set by other streptococcal pathogens indicates that the pathogenesis of strangles is likely to be complex with the virulence of S. equi being dependent on a large number of virulence determinants. Good protective immunity, therefore, is also likely to be multifactorial, which explains the failures of approaches to vaccination that were wholly dependent on the SeM. It follows that the key to understanding the basis of immunity will be a systematic understanding of the molecular basis of pathogenesis. This in turn will require a full understanding of all of the potential virulence determinants before selection of those that are essential for developing multi-determinant immunity will be possible. Such understanding can only come after the full sequencing of the pathogen genome. Sequencing of the S. equi genome began in 2000 following a pivotal decision by the Home of Rest for Horses to fully support the project, which will be completed sometime in 2001. The test sequence assemblies representing over 99% of the genome are publicly available and can be searched at the Sanger Centre website (www.sanger.ac.uk/Projects/S_equi/). Already many scientists from around the world are analysing and using the sequence in studies of pathogenesis and as a result of these investigations, that will employ motif and homology searches, directed mutagenesis and multiple array analysis of differential gene expression, all of the S. equi virulence determinants will be revealed. A process of rational vaccine candidate selection will then be possible and provide the very best prospect for an effective vaccine against not only strangles but probably also against disease caused by the very closely related S. zooepidemicus (see below), the probable archetype of S. equi.
R. equi disease is one of the most commonly diagnosed bacterial respiratory tract infections of the lower airways of equines although epidemiological studies of Thoroughbred studs including "Rhodococcus farms" suggest that it is not quite as prevalent as S. zooepidemicus . Clinically apparent R. equi infection rarely occurs in equines older than 6 months of age with most cases less than 4 months old. Cases of disease most commonly result from chronic suppurative bronchopneumonia, extensive abscessation and suppurative lymphadenitis . Early short-lived mild fever is frequently overlooked so that steady spread of infection results in a progressive loss of lung tissue. Eventually this may be accompanied by re-occurrence of pronounced temperature, lethargy, decreased appetite, weight loss and marked abdominal effort in breathing. Around half of foals examined post mortem have multifocal ulcerative enterocolitis and typhlitis over the Peyer’s patches associated with granulomatous or suppurative inflammation of the mesenteric and colonic lymph nodes that is infrequently manifested as colic and diarrhoea . Immune mediated polysynovitis in approximately 30% of cases  and less frequently septic arthritis, osteomyelitis , ulcerative lymphangitis, cellulitis have been reported [34,38].
R. equi is predominantly a soil organism that is also found in the faeces of herbivores, particularly horses . The organism grows in the top layers of soil  and in faeces lying on the ground. Indeed, faeces left on the ground ultimately increases the weight of soil infection . The organism can also be found in air, particularly in the hot, dry and dusty conditions  that coincide with the peak incidence of the disease in foals. Whilst R. equi is more or less universally present, the disease in foals is not. The occurrence of equine disease is coincident with the distribution on certain farms of large numbers in the soil of strains carrying an 85kb or 90kb plasmid that encodes a number of proteins of between 15 and 17kDa . The expression of these "virulence associated proteins" or "Vap" is required for optimal virulence in experimentally infected mice  and plasmid-cured strains are practically avirulent in foals . Isolates from cases of natural disease are predominantly plasmid positive . R. equi is generally present in larger numbers in the faeces of foals than in adults but the weight of faecal infection is particularly high in cases of disease (106 or more bacteria /gram; ). The presence of diseased foals, therefore, increases the weight of soil infection with Vap+ bacteria to levels well above that required to infect further foals (around 104 bacteria; ). Typing of DNA restriction length polymorphism analysed by pulsed field gel electrophoresis has recently shown that there are several discernible types of Vap+ bacteria each detected from several locations but that a single type tends to predominate on individual endemic farms . This observation supports the hypothesis that endemicity is a cycle of environmental contamination from diseased foals that become infected from the environment.
Diagnosis of R. equi Disease
Diagnosis of R. equi respiratory disease is not straightforward. Several imaging techniques are useful aids but radiography in particular is helpful in assessing the severity of pneumonia and in gauging the progress and success of therapy but there are several other potential causes of pneumonia so it cannot be relied upon for specific diagnosis. High serum fibrinogen concentrations and neutrophilic leukocytosis are usually present in advancing cases but again other bacterial infections, not only of the airways, could cause the same effects. Bacteriological culture of tracheal aspirates taken transtracheally or via a catheter passed down the biopsy channel of an endoscope can yield R. equi in approximately 60% of cases that are subsequently shown to be affected at necropsy , whereas, recovery of the organism in healthy animals can result from inspiration of infected soil from the environment. Isolation of an abundantly growing organism from foals that is usually pigmented and contains large, pleomorphic Gram positive rods containing metachromatic granules is often sufficient to make an identification of R. equi. Isolation of R. equi from soil and faeces has only been practicable since the introduction of a selective culture medium . Cytological examination of the aspirates can be helpful since intracellular Gram positive rods can be seen, again in approximately 60% of cases . In any event, taking a tracheal sample from a foal in severe respiratory distress can be highly risky.
Various attempts at serological diagnosis have been made but tests have depended on crude antigens to which too many animals have antibody as a result of environmental exposure and detection of seroconversion requires a sample taken before the immune stimulus which is not often available. More recent examination of an ELISA test for antibodies reacting with a Vap A enriched antigen, and therefore detecting responses to potentially virulent R. equi, revealed that mares varied widely in their serum antibodies . Furthermore, although this variation may influence the emergence of active responses in the foals, seroconversion was evident around 8 - 10 weeks of age and well before 4 months of age. This test, therefore, is potentially useful but more work is needed to assess the relationship between seroconversion and subsequent emergence of evident disease.
R. equi Disease Treatment
Significant stages in the pathogenesis of R. equi infection seem to depend upon intracellular survival particularly in monocytes and the characteristic lesions are caseous and avascular. Treatment of R. equi infection, therefore, requires antibiotics that are both lipid soluble and so are likely to penetrate the lesions and that can enter the host cells and reach all susceptible bacteria. Combined oral treatment with bacteriostatic erythromycin and rifampin are the first choice [36,49] with the prognosis improved from 20% survival without treatment  to 80% or more with. Erythromycin may induce diarrhoea, and if this becomes severe the drug may have to be given intravenously. Rifampin given alone rarely induces diarrhoea but it can cause severe side effects if given intravenously so where it is necessary to give erythromycin IV, rifampin should be continued orally. In common with pneumonia in other mammals, the aetiology is sometimes complex involving several species of pathogenic bacteria so it may be necessary to give other antibiotics in refractory cases or where indicated by the results of bacteriological examination of tracheal samples. Little resistance to erythromycin or rifampin is seen in R. equi but where it occurs there is very little alternative. Success with enrofloxacin treatment has been reported in such cases  but the joint and cartilage side effects in growing animals mean that this drug is not licensed for use in the young.
Epidemiological and Immunological Control of R. equi Disease
Controlling R. equi disease in endemic farms has relied mainly on management to reduce the size of potential infectious challenges. In particular, pasture should be managed to avoid bare patches of soil developing in the summer that encourage air borne bacteria, and faeces in which R. equi readily multiplies should be frequently and completely cleared and disposed of by composting.
Active immunisation of mares and foals with crude bacterin based vaccines has had mixed and at best poor success. However, for reasons that are unclear, passive immunisation of foals with around one litre hyperimmune serum has been used prophylactically in foals on a number of endemic farms to significantly reduce the incidence of disease [54,55]. Administration of the serum after the onset of pathogenic infection does not seem to alter the course of disease. Unfortunately, early efforts at active immunisation with Vap enriched preparations have not been promising . Development of effective vaccination may await further analysis of the Vap encoding plasmids and of the potential chromosomal encoded virulence genes to identify a protective combination of immunogens or to aid the creation of a live avirulent vaccine. Recently, analysis of the Vap family of plasmid encoded genes has revealed Vap C, Vap D and Vap E expressed from genes arranged tandemly downstream of vapA . These proteins are secreted and hold some promise as potential protective immunogens in view of their cellular location and accessibility to host immunity and their lack of antigenic relatedness with the non-protective Vap A.
Many of the capsule types of Streptococcus pneumoniae (pneumococci) in man are an important world-wide cause of pneumonia, meningitis and febrile bacteraemia and the non-invasive diseases otitis media, sinusitis and bronchitis. A recent World Health Organisation publication estimated that over one million children die of pneumococcal disease every year, mainly in the developing world . Widespread development of antibiotic resistance in pneumococci, the great potential of the organism to acquire resistance genes from other bacteria, and the widespread prevalence of the organism, compromises the usefulness of treatment in the control of this disease [59,60]. The longer-term prospect for control lies in vaccination.
Pneumococci were first isolated from horses during a study of the virological causes of respiratory disease in Switzerland . Reports of sporadic isolation from various types and age of horse [62,63] followed. In the 1980's, the importance of pneumococci as a cause of inflammatory airway diseases (IAD, a moderate to gross excess of mucopus in the trachea sometimes with cough and pyrexia) was first noted in Thoroughbred horses in the United Kingdom [64-67]. The organism was generally isolated on horse blood agar on which growth was improved in 5% CO2, was inhibited by optochin and contained typical Gram positive diplococci. Subsequently, disease was reproduced in experimentally challenged ponies given equine pneumococci into the trachea through the biopsy channel of an endoscope . The challenged ponies had excess mucopus in the trachea, just like training Thoroughbreds, and post mortem examination revealed focal pneumonia especially in the cardiac area and accessory lobe, histologically very similar to human pneumococcal lung lesions, although less extensive. A later more detailed prospective study showed that IAD is extremely common in the training Thoroughbred between the ages of 2 and 4 . It had a prevalence of 13.8%, an incidence of 8.9 cases per 100 horses per month and a mean duration of 7.8 weeks. Streptococcus zooepidemicus and Actinobacillus/Pasteurella species were the most prevalent bacteria associated with IAD throughout the susceptible age group and there was very little involvement by viruses. However, the association of pneumococci with IAD in 2 year olds had a prevalence of 7% and an odds ratio of over 2000 . Furthermore, the incidence of pneumococcal infection before the age of 3 was over 95% making pneumococci the cause of a significant amount of disease in the horse. Although all of the isolates recovered from the horse to date have been capsule type 3 [70,71], early efforts to immunise with capsule and protect against experimental challenge were not promising (see below). Recently, allelic variation in restriction fragment patterns of six housekeeping genes revealed that equine isolates were a tight group very closely related to human type 3 pneumococci . However, they all lacked functional autolysin and pneumolysin genes due to recombination and deletion of DNA between the middles of each of these closely spaced genes. Furthermore, restriction fragment length polymorphism analysis of the genes of other putative virulence determinants showed that the equine type 3 pneumococci represented a tight clone distinct from the human type 3 pneumococci.
There are few antibiotics for which the relevant pharmacokinetics is established in the horse and, therefore, treatment of clinical cases cannot be performed with any assurance. However, there are empirically tested regimes for inflammatory airway disease and these are described under the section on Streptococcus zooepidemicus.
The potential for pneumococci to acquire and transfer antibiotic resistance genes and the close relationship between man and the domesticated horse means that antibiotic treatment of equine pneumococcal disease carries significant public health risks. Therefore, the best prospect for controlling pneumococcal lower airway disease of the horse lies with the possibility of vaccination. There is considerable evidence for the role of the pneumococcal capsule as a virulence determinant [72-74] and protective immunogen in experimental models of disease [75,76] and field trials of vaccines in adult humans that has led to a commercially available vaccine widely used in some countries . The vaccine, however, does not provide comprehensive protection. The two best understood reasons for vaccination failures are the non-cross protection between different capsule types and the T-cell independent nature of the human immune response to the capsule antigens [78,79]. The available vaccines attempt as wide as possible protection by incorporating the most frequently encountered capsule types. The vaccine contains 23 different capsule antigens so in some parts of the world the most prevalent types may not always be represented in the vaccine leaving vaccinates unprotected against a significant proportion of infections potentially caused by the other 60 - 70 capsule types.
The T-cell independent pneumococcal polysaccharides are large molecules with repeated epitopes of the same structure that can by-pass the need for T-cell help. The T-cell independent immune response in man is not mature until 2 years of age , thus leaving one of the most vulnerable sectors of the population unprotected by the existing vaccine. Furthermore, there is evidence that T-cell independent immunity may be difficult to boost and can in some circumstances induce a state of tolerance [80,81]. There is recent evidence that Ugandan HIV-1 infected women given the pneumococcal vaccine have a higher incidence of pneumococcal disease than placebo controls even though vaccination induced anti-capsule antibody . This is particularly interesting since the author found more severe disease, clinically and at post mortem examination, in ponies vaccinated with purified type 3 capsule and challenged with equine type 3 pneumococci than in challenged unvaccinated animals, even though the young horse produces antibodies to the capsule antigen (author; unpublished observation). The equine work was conducted in 1992 and was not published through fear that the results might be regarded as heretical. The observations in man and the horse suggests that immunity to pneumococcal capsule can augment disease by an unknown and potentially important mechanism in diverse circumstances.
The deficiencies of capsule vaccination have led to the examination of pneumococcal protein determinants as protective immunogens either singly, or as conjugates with capsule polysaccharide. These would have the advantages of T-helper-cell dependent immune responses, which are mature in young children, which can be boosted, and have the potential for much greater cross-type protective immunity. Foremost amongst these determinants has been the pneumolysin protein in which a large investment has been made due to its biological activities, such as hemolysis, induction of ciliostasis and complement activation, to name a few, that all suggest considerable importance for virulence . Although there is much evidence from experimental animal models that pneumolysin can confer a significant degree of protection  the equine pneumococci do not produce pneumolysin and so this avenue is not available for an equine vaccine. Nonetheless, immunisation with a critical combination of protein virulence determinants either surface expressed or excreted from the bacterium and hence accessible to neutralisation by antibody may hold out the best prospect for vaccination both in man and the horse. The recent availability of genome sequencing and sophisticated methods for the analysis of differential gene expression combined with mutagenic screening for loss of virulence in vivo should rapidly identify a list of promising vaccine candidates.
Mycoplasma spp. Infection
Mycoplasmas are bacteria without cell walls that require cholesterol for growth and are generally so small that they pass through filters which retain most other bacteria.They have been isolated from most animals and plants and their nutritional requirements are generally exacting, growing slowly to produce small colonies. Inhibition of growth with specific antisera is usually required for definitive identification.Mycoplasma infections are known to cause respiratory disease in many species including man, cattle and pigs. However, their role in equine diseases has not been clearly established.
A number of different Mycoplasma spp have been isolated from horses, including M. felis, M. equirhinis, M. subdolum, M. equigenitalis, a strain serologically cross reactive with M. pulmonis, strain N3 and Acholeplasma laidlawii [85-89]. The most commonly reported isolate is M. equirhinis and antibodies to M. equirhinis have been detected in 75 - 100% of racehorses in the UK, whereas antibodies to strain N3, another specific equine isolate, were detected in only 38% . All of these investigations in the UK, conducted 20 years ago, failed to find a relationship between nasopharyngeal isolation of mycoplasmas and equine respiratory disease. Usually, small numbers of organisms were detected in the nasopharynx and respiratory disease was not reproduced experimentally after intranasal challenge with either strain N3 or M. equirhinis [89,91]. More recent field investigations of respiratory disease in foals have revealed an association between disease and the presence of Mycoplasma spp in the nasopharynx. Clinical signs included fatal pneumonia with a mortality rate of up to 25% [92,93]. However, other investigations have not identified such an important role for mycoplasmas in causing respiratory disease in foals  and there is a clear need for further studies to clarify the role that mycoplasmas might play.
Despite the failure to clearly identify a link between nasal mycoplasma infection and equine respiratory disease using epidemiological and experimental methods, M. felis was isolated from cases of pleuritis in the horse more than 15 years ago . Furthermore, pleuritis was reproduced experimentally and naturally occurring cases seroconverted to M. felis [94-96]. In cases involving M. felis initial conventional bacterial investigations of inflammatory airway disease have often failed to identify a possible cause but a septic aetiology has been indicated by low glucose levels in both pleural and pericardial effusions. Although glucose measurement is rapidly performed, and can be used as an indicator of sepsis in the horse, differential diagnosis of the infectious cause is, nonetheless, needed to guide specific treatment and, consequently, rapid agent detection and identification is required.
More recently, an outbreak of respiratory disease in young Thoroughbred racehorses in training was found to be associated with M. felisin the absence of a significant presence of other detectable pathogens . A high proportion of the group was affected with increased tracheal mucopus visible on endoscopic examination. Coughing at exercise, pyrexia and limb filling were also observed. Morbidity was more than 85% and 19/22 animals tested seroconverted to M. felis. Large numbers of M. felis (>104 colony forming units) were detected in tracheal samples from 4 of the affected horses. A considerable proportion of cases of inflammatory lower respiratory tract disease that is not associated with infection by known equine viruses is observed annually in yearling and 2 year old racehorses in the UK, affecting as many as 80% such horses, are closely associated with large numbers of bacteria. Interestingly, Mycoplasma spp have been cultured from 54% of tracheal samples where the total bacterial count is greater than 104 cfu/ml but from only 6.5% of samples where the count is lower and from 42% of washes with evidence of pronounced airway inflammation, but from only 9% of samples with no or slight inflammation (J.L.N. Wood, unpublished data). As with earlier studies of the bacterial flora of the equine nasopharynx, M.equirhinispredominated. However, detailed multivariate analyses allowing for confounding and other factors has revealed no true association between mycoplasma isolation from the trachea and disease. M. felis was isolated too infrequently for the analyses to detect an association with inflammatory airway disease but the previous descriptions of relatively rare but nonetheless significant outbreaks together with the results of experimental challenges suggest that this infection should always be borne in mind as a possible cause of future outbreaks of disease.
There are few antibiotics for which the relevant pharmacokinetics is established in the horse and, therefore, treatment of clinical cases cannot be performed with any assurance. However, there are empirically tested regimes for inflammatory airway disease generally which are described under the section on Streptococcus zooepidemicus. However, wherever M. felis infection is suspected to be the main aetiology, treatments based upon the tetracycline group of antibiotics are recommended, but, as with S. zooepidemicus and S. pneumoniae infections, these treatment regimes are not fully tested and relapses are a common feature for inflammatory airway disease generally.
S. zooepidemicus is associated with several syndromes including endometritis in "susceptible mares"  and it is one of the most frequent causes of infectious abortion and placentitis [99-101], pneumonia [102,103] and shipping fever . It is now also associated with lower airway inflammation in foals and Thoroughbred horses in training [11,33,105-107] and there is growing evidence of close involvement with repeated respiratory tract infections from life as a foal to horses of 3 to 4 years of age.
A recent longitudinal study of lower airway disease in training Thoroughbreds  revealed the presence of moderate amounts of mucopus in the trachea, accompanied by a neutrophilia with a prevalence of 13.8% of horse months and an incidence of 106.8 cases/100 horses/year. Most horses up to 3 years old encountered at least one episode of disease (mean duration 9.4 weeks) annually, associated with infection by large numbers of S. zooepidemicus and one or more members of the Actinobacillus/Pasteurella spp (Fig. 5 depicts a moderately heavy deposit of inflammatory mucopus in the trachea of an S.zooepidemicus infected horse). But seroconversion to the known equine pneumotropic viruses occurred in only 7.5% of cases.
Figure 5. An endoscopic view of the inflammatory material in the trachea of an S.zooepidemicus infected young Thoroughbred rested from training whilst it recovered from inflammatory airway disease.
Our study of 12 Welsh Mountain ponies aged approximately 6 - 7 months, from a single isolated source after they were mixed and moved with 5 more ponies, each from different locations, showed that inflammatory airway disease associated with S. zooepidemicus was evident in many on arrival. The disease increased in severity and persisted throughout the following 10 weeks and there was a relationship between the mean visible mucopus score and numbers of S. zooepidemicus /ml of tracheal wash sample (unpublished observations). No relationship was found between tracheal pathology and seroconversion to the principal equine respiratory viruses. These results demonstrated that lower airway disease begins early in life in non-Thoroughbreds as it does in Thoroughbreds [33,106].
Treatment of S. zooepidemicus Disease
S. zooepidemicus respiratory tract disease in Thoroughbreds results in many lost training and racing days and there is evidence for important long term consequences. S. zooepidemicus infection of the lower airway in foals, but not infection by R. equi, adversely affected subsequent racing performance  by an unknown mechanism that may be related to residual lung damage or to autoimmune disease similar to rheumatic fever which can follow S. pyogenes infection in man. There are few antibiotics for which the relevant pharmacokinetics is established in the horse and, therefore, treatment of clinical cases cannot be performed with any assurance that therapeutic levels of drug are achieved in the respiratory tract although there are empirically derived regimes (James Wood, personal communication). These include penicillin, penicillin/gentamicin, ceftiofur IM daily, oxytetracycline IV daily, enrofloxacin injected daily and potentiated sulphonamides at 2 - 4 times the recommended dose. None of these regimes has a clearly effective record and all are often followed by relapses due to unsound pharmacokinetics, interference with the development of sound immunity or sequential infection by different non-cross-protective types of S. zooepidemicus (see below).
S. Zooepidemicus Epidemiology
The prolonged nature of inflammatory airway disease associated with S. zooepidemicus infection and its recurrent episodes in young animals suggest that immunity is not acquired effectively until later in life. In this respect, S. zooepidemicus in the horse appears to behave like S. pyogenes in man where a similar age-related incidence, prevalence and range of respiratory or other infections is seen. S. pyogenes has a surface antiphagocytic M-protein, which stimulates protective opsonogenic antibody . Unfortunately, the M-protein is present on different isolates in different non-cross-protective antigenic variations. The greater resistance of adults to S. pyogenes disease is widely regarded as being due to developing immunity to the most prevalent types of M protein. As with the M-protein of S. pyogenes, there is considerable evidence for antigenic, genetic and opsonogenic variation in the S. zooepidemicus M-like protein [111,112]. Homologous circulating opsonins to M-protein are protective against S. pyogenes infection in primates . It is not known whether antibody to M-like protein is protective in the horse against S. zooepidemicus but anti M-like protein is not protective against S. equi (see above), the most closely related bacterium to S. zooepidemicus. Nonetheless, the epidemiology of lower airway disease and S. zooepidemicus infection may still be related to the presence of different types of the organism with non-cross-protective pathogenic determinants which might yet include the M-like proteins.
The 16S - 23S RNA gene intergenic spacer of S. zooepidemicus is made up of 9 subregions most of which exhibit variations between different isolates . This variation allows the differentiation of isolates into 8 types. Some types have spacers that appear very different and suggest considerable genetic diversity that might also coincide with the production of antigenically variable and perhaps non-cross-protective pathogenic determinants.
PCR typing by the 16S - 23S RNA gene intergenic spacer has been extended by co-typing by the hypervariable region of the M-like protein (author, unpublished results). Fifteen serotypes were initially detected by immunodiffusion using cross-absorbed antisera . Sequences of the opsonogenic M-like proteins of some of these indicated that most have one of 5 central hypervariable regions [114,115]. Different intergenic spacer types can possess the same M-like protein hypervariable region (unpublished observations). Consequently, the intergenic spacer types could be subdivided to yield a potential 40 (5x8) sub-types. This degree of discrimination between types will provide a valuable tool to answer some fundamental questions about the epidemiology of S. zooepidemicus infection and lower airway disease.
PCR typing specific for the hypervariable region of the M-protein has indicated that pathogenic infection within individual horses may be clonal, with all isolates belonging to the same type, rather than infection by a mixture of types . Analysis of the incidence and prevalence of different types within individuals and defined populations may indicate if repeated and prolonged episodes of inflammatory airway disease are associated with persistent infection by one type, several types or a clonal succession of different types suggestive of a series of infections bearing non-cross-protective and antigenically variable pathogenic determinants. Consequently, infection by one type may significantly reduce the likelihood of a second infection by the same type indicating the existence of acquired type-specific immunity. Examination of the dynamics of transmission of infection and determination of the most common types of S. zooepidemicus encountered in lower airway disease may provide insight into the prospects for preventing spread at critical times in the young horse’s life.
S. zooepidemicus and Prospects for Vaccination
Currently, there are no vaccines available against S. zooepidemicus disease and there are no publications of attempts at vaccination. S. equi and S. zooepidemicus are very closely related and a similar complex pathogenesis can be expected of inflammatory airway disease as can be predicted for strangles (see above). Whilst the 2 diseases must clearly depend on virulence determinants unique to S. zooepidemicus and S. equi it is likely that these two bacteria will share the vast majority of their virulence determinants. Therefore, a detailed analysis of the S. equi genome may reveal vaccine candidates that will also be effective for S. zooepidemicus. A major difficulty with the development of S. zooepidemicus vaccines, however, is that there are no good models of disease in which to test immunisation. The young horse encounters too much intercurrent disease to enable the investigator to tell if the effects observed are due to experimentation or to immune responses to existing or previous infections. It may be possible to use experimental strangles as a model for vaccines based upon virulence determinants of S. zooepidemicus shared with S. equi. If such vaccines protected against strangles, a Group C streptococcal infection in the horse, then the same immunogens would stand a good chance of protecting against S. zooepidemicus and, furthermore, of cross protecting between different isolates of S. zooepidemicus. If such experimental evidence of the efficacy of S. zooepidemicus immunogens could be gained it would then be worth conducting field trials with an experimental vaccine perhaps in the first instance taking advantage of the natural disease seen in young Welsh Mountain ponies.
Actinobacillus equuli is a recognised cause of septicaemic disease and joint-ill in foals [117,118], but isolates from older horses were once regarded as commensal and unlikely to be pathogenic , with the exception of an occasional case of abortion . Although A. equuli has been renamed a number of times, its synonyms, and A. suis -like bacteria, have been isolated from horses on several occasions from as early as 1925 (reviewed; ). Horses are known to be oral carriers of A. equuli and A. suis -like bacteria that are distinct from porcine A. suis isolates.
A. equuli and A. suis -like bacteria were isolated from the upper respiratory tract of normal horses  but, as mentioned above, the Actinobacillus/Pasteurella group of bacteria recently has been found to be strongly associated with a significant proportion of cases of inflammatory airway disease in training Thoroughbred horses. During the studies of training Thoroughbreds, isolates from tracheal lavage samples were presumptively identified as Actinobacillus or Pasteurella species on the basis of colony morphology, cellular pleomorphism detectable by Gram staining and the result of API 20NE tests (BioMerieux). This was because definitive identification depended heavily on the detection of sugar fermentation in serum containing media that are difficult to produce on mass over a period of time to a standard quality. Therefore the association of this group of bacteria with lower airway disease in the Thoroughbred horse  might have been related to one or more species of this group of bacteria which may be acting directly as pathogens. Seventy-three isolates from 65 horses with and without evidence of lower airway disease were subsequently identified with the aid of a single batch of carefully prepared serum sugar broths, quality controlled with the use of type cultures as positive controls, to assess whether the association with disease was accounted for by a small or large number of species . Just over half (50.5%) were A. equuli, 17.8% were A. suis -like, 11% were Pasteurella pneumotropica, 8.2 % were A. lignieresii, 6.8% were Mannheimia haemolytica and 5.5% were P. mairii. Among the horses examined, lower airway inflammation was significantly associated with A. suis -like bacteria and A. lignieresii. However, since these two species constituted too small a proportion of the undifferentiated Actinobacillus/Pasteurella species to explain the overall association of the group of bacteria with lower airway disease it seems possible that A equuli constituting over half of the isolates, is likely to be playing a significant role in inflammatory airway disease.
Just as with S. zooepidemicus, the repeated and chronic infections in the trachea seen with the Actinobacillus/Pasteurella species suggests the possibility of a series of infections by poorly cross-protective types or species. Protective immunisation, therefore, may depend on the inclusion of a selection of isolates to achieve the greatest coverage of types likely to be encountered without jeopardising the quality of immune responses to the more common and perhaps more significant types. There is no published work to underpin vaccine development for this group of bacteria in the horse but effective approaches have been taken for investigating protective immunity for this group of bacteria for diseases they cause in the farm animal species  and these could be applied to the equine situation.
A very large number of other bacterial species have been isolated from the lung including practically all of those acting as opportunistic pathogens for many other mammals. However, there is no significant evidence to clearly identify these bacteria as anything other than opportunistic infections, in a favourable environment arising from pre-established disease. Indeed the exhaustive epidemiological studies of inflammatory airway disease described above in young training Thoroughbred horses or Welsh Mountain ponies found most of these bacteria in the trachea of these animals but no statistically significant association could be found with disease. Should any of these be suspected of causing disease in individual cases then there are the same problems with, and empirical solutions to, treatments, just as described above for bacterial infections that are shown to be significantly associated with tracheal inflammation.
How to reference this publication (Harvard system)?
Affiliation of the authors at the time of publication
Animal Health Trust, Lanwades Park, Kentford, Suffolk, Newmarket, UK.