ISOLATION, CHARACTERISATION AND MOLECULAR TYPING OF FELINE MYCOPLASMA SPECIES

Transcription

1 ISOLATION, CHARACTERISATION AND MOLECULAR TYPING OF FELINE MYCOPLASMA SPECIES Sally Rae Robinson BVSc (Hons) Thesis submitted in fulfilment of the requirements of the Degree of Master of Veterinary Science (by research) Department of Veterinary Science, The University of Melbourne March 2009

2 2

3 ABSTRACT The exact role of mycoplasma in feline ocular and respiratory disease is not yet understood. The results of previous studies are contradictory in this regard. There is some evidence to suggest that M. felis has a pathogenic role in such diseases, but it is inconclusive. The aim of this study was to investigate the prevalence and anatomical distribution of mycoplasmas in a population of shelter cats, to determine which species were present, and establish the association of their presence with ocular or respiratory disease. The prevalence of mycoplasma in the 110 cats examined was 71.8%, as determined by in vitro culture. Mycoplasma was most commonly isolated from the pharynx, followed by the bronchus and conjunctiva. In infected cats, mycoplasmas were likely to be isolated from multiple anatomical sites. The polymerase chain reaction (PCR) was used to amplify part of the 16S rrna gene, and the mutation scanning technique non-isotopic single-strand conformation polymorphism (SSCP) was utilised to delineate mycoplasma isolates based on nucleotide sequence variation. PCR-SSCP proved to be a useful method to screen large numbers of samples for variation and to group them according to species. The species of mycoplasma identified by nucleotide sequencing were M. felis and M. gateae. It was not determined whether it was possible to differentiate between M. gateae and M. arginini based on SSCP profile results with the target DNA region used due to their almost identical nucleotide sequence. This group of M. gateae/m. arginini served as a useful non-pathogenic comparison group to M. felis. There was no statistically significant difference between M. felis and the M. gateae/m. arginini group with respect to prevalence or anatomic distribution. There was no evidence of any association of mycoplasma with disease linked to any of the anatomic locations studied. i

4 Mycoplasmas were isolated from the lower respiratory tract in 42.7% of cats. The isolation of mycoplasmas from the lower respiratory tract of healthy cats has been reported once, but this is the first report of M. felis being isolated from this location in healthy cats. This finding indicates that the isolation of mycoplasmas from the lower respiratory tract is not sufficient evidence to implicate a role in respiratory disease. Mycoplasmas were not significantly involved in ocular or respiratory disease in the population of cats studied. More likely, they are commensal organisms in the conjunctiva, pharynx and bronchus. Whether they are capable of playing an opportunistic role in disease, or what conditions may facilitate such a role remains to be determined. ii

5 DECLARATION This is to certify that the thesis comprises only my original work except where indicated in the preface; due acknowledgement has been made in the text to all other material used; the thesis is < 30,000 words in length, exclusive of tables, figures, appendices and bibliography. Sally Robinson March 2009 There are pages throughout this thesis intentionally left blank. iii

6 ACKNOWLEDGEMENTS I would like to acknowledge the assistance of my supervisors Steven Holloway, Kevin Whithear and Robin Gasser for their interest and ideas during the experimental phases, and especially Steven for his patience and understanding to the end of the thesis writing. Thank you to Tony Belfiore and Nathan Jeffery for their expertise and great company in the mycoplasma laboratory. Thanks also to Min Hu, Yousef El Osta, and the postgraduate students in the parasitology laboratory for their technical assistance and for accepting me as a non-parasitological impostor. I would like to acknowledge the Cat Protection Society, Victoria, for kindly allowing me access to the population of cats at the shelter for this study. Thank you to Garry Anderson for his time and enthusiasm in assisting me with the statistical analysis for this project. Mostly I give thanks for the tremendous support from all of my family; firstly to my husband Nick for his critical eye and unfailing belief in my abilities, and to my amazing daughters Sophie and Catherine who have made the completion of this project a little more challenging to achieve but ultimately more rewarding. I thank my parents Helen and Michael for their encouragement and belief in me to achieve anything, and for providing the educational opportunities for me to do so. I am also very grateful for the love and support from my parents-in-law Jen and Wayne. Especially, I am indebted to WFR for the guidance, wisdom and direction he provided at a critical time in the thesis writing process. iv

7 COMMUNICATIONS Work presented in this thesis has been communicated in the papers or presented at the meetings/conferences listed below: Sally Robinson, Robin Gasser, Steven Holloway Epidemiology and molecular typing of feline mycoplasmas from a population of shelter cats. (Manuscript in preparation) Sally Robinson, Robin Gasser, Steven Holloway (2004) Master of Veterinary Science confirmation seminar presented at The University of Melbourne Veterinary Clinical Centre, Melbourne, Australia. Sally Robinson, Robin Gasser, Steven Holloway (2004) Feline mycoplasmas. Oral presentation at the Australian College of Veterinary Scientists annual conference, young speakers section of the small animal medicine chapter. Gold Coast, Australia. Sally Robinson, Robin Gasser, Steven Holloway (2004) Feline mycoplasmas. Oral presentation at the University of Melbourne Academic Associates Meeting. Melbourne, Australia. v

8 TABLE OF CONTENTS CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW INTRODUCTION RESEARCH OBJECTIVES LITERATURE REVIEW Mycoplasmas general characteristics Species of mycoplasma isolated from cats Methods for the isolation and identification of feline mycoplasmas Collection techniques Culture media Culture conditions Morphology Biochemical profiles Serological methods for specific identification Molecular methods for specific identification Epidemiology of feline mycoplasmas Prevalence of mycoplasma in cats Prevalence with reference to anatomical location Prevalence - with reference to different mycoplasma species Possible roles of individual mycoplasma species in cats Age of cats Geographical and climatic factors Prevalence of mycoplasmas compared with other respiratory pathogens Relationship of mycoplasmas to other respiratory pathogens Association of mycoplasma with feline bronchial disease (FBD) Presence of feline mycoplasmas in other species Consideration of factors that may facilitate mycoplasmal involvement in disease CONCLUSION AIMS CHAPTER 2 EPIDEMIOLOGICAL SURVEY OF FELINE MYCOPLASMA INTRODUCTION MATERIALS AND METHODS Animal Collection Culture Statistical analysis RESULTS DISCUSSION CHAPTER 3 MOLECULAR CHARACTERISATION OF FELINE MYCOPLASMA INTRODUCTION MATERIALS AND METHOD DNA extraction by heat lysis Mycoplasma 16S rrna gene amplification Primer design for PCR-SSCP analysis PCR optimisation PCR of mycoplasma swab samples Non-isotopic Single-Strand Conformation Polymorphism (SSCP) Assessment of variability within samples SSCP of mycoplasma-positive PCR samples DNA Sequencing Sequence analysis Statistical analysis RESULTS PCR of controls and samples Screening of isolates by SSCP for genetic variability SSCP-coupled analysis of all samples, and identification vi

9 3.3.4 Numbers and proportion of each mycoplasma species found Comparison of colour change observed in liquid culture and identification based on PCR-SSCP and nucleotide sequencing Analysis of species distribution among anatomical sites Association between mycoplasma species and ocular/respiratory disease DISCUSSION CHAPTER 4 GENERAL DISCUSSION AND CONCLUSIONS CHAPTER 5 APPENDICES APPENDIX 1: DATA COLLECTION SHEET APPENDIX 2: LIQUID MYCOPLASMA MEDIA APPENDIX 3: MYCOPLASMA AGAR FORMULATION APPENDIX 4: CALCULATING THE BINOMIAL DISTRIBUTION APPENDIX 5: PROTOCOL FOR DNA EXTRACTION BY HEAT LYSIS METHOD APPENDIX 6: ALIGNMENT OF FELINE MYCOPLASMA SPECIES IN THE REGION AMPLIFIED BY PRIMERS FD1000 AND MYCOR APPENDIX 7: COMPARISON OF DNA EXTRACTION METHOD AND INFLUENCE OF CULTURE MEDIA IN SAMPLES BY RELATIVE INTENSITY OF PCR PRODUCTS VISUALISED ON AGAROSE GEL APPENDIX 8: DNA PURIFICATION PROTOCOL CHAPTER 6 REFERENCES vii

10 LIST OF TABLES Table 1.1: Comparative colony characteristics for feline mycoplasma species Table 1.2: Comparative biochemical characteristics of feline mycoplasma species Table 1.3: Comparative percentage of mycoplasma isolates from sick and healthy cats found to be infected with each of three species of mycoplasma Table 2.1: Study population characteristics Table 2.2: Comparison of prevalence of mycoplasma-positive samples and gender of cats Table 2.3: Comparison of prevalence of mycoplasma-positive samples in adult and juvenile cats Table 2.4: Comparison of prevalence of mycoplasma-positive samples in stray and owned cats Table 2.5: Comparison of prevalence of mycoplasma-positive samples and location of cats within shelter Table 2.6: Comparison of prevalence of mycoplasma-positive samples between cats with and without signs of ocular/respiratory disease Table 2.7 Comparison of paired proportions for mycoplasma status between the conjunctiva and pharynx using McNemar s test Table 2.8: Comparison of paired proportions for mycoplasma status between the conjunctiva and bronchus using McNemar s test Table 2.9: Comparison of paired proportions for mycoplasma status between the pharynx and bronchus using McNemar s test Table 2.10: Comparison of prevalence of mycoplasma-positive conjunctival swabs between cats with and without conjunctivitis Table 2.11: Comparison of prevalence of mycoplasma-positive pharyngeal swabs between cats with and without signs of upper respiratory tract disease Table 2.12: Comparison of prevalence of mycoplasma-positive bronchial swabs between cats with and without gross lung pathology Table 2.13: χ 2 test of goodness of fit of the binomial distribution, applied to the number of cats with a particular number of mycoplasma-positive anatomic sites Table 3.1: Mycoplasma 16S rrna gene PCR primer sequences viii

11 Table 3.2: Results of SSCP for the 129 mycoplasma-positive PCR samples Table 3.3: Number of variant SSCP profiles identified and comparison to health status of cats Table 3.4: Numbers of cats that had the same mycoplasma species or different mycoplasma species among sites when multiple sites within a cat contained mycoplasma Table 3.5: Comparison of the proportion of each mycoplasma species between different combinations of positive sites using Fisher s exact test (where the same species was present at each site) Table 3.6: Comparison of the proportion of each mycoplasma species between different combinations of positive sites using Fisher s exact test (where those with different species at each site were also considered in the overall proportion) Table 3.7: Number and proportion of positive samples for each species of mycoplasma at each anatomic site sampled Table 3.8: Proportion of observations and odds at each site of a positive result for M. felis (F) compared to M. gateae/m. arginini (R) Table 3.9: Comparison of likelihood of M. felis versus M. gateae/m. arginini at particular anatomic sites using the odds ratio Table 3.10: Comparison of paired proportions for mycoplasma status overall between the conjunctiva and pharynx using McNemar s test Table 3.11: Comparison of paired proportions for mycoplasma status overall between the conjunctiva and bronchus using McNemar s test Table 3.12: Comparison of paired proportions for mycoplasma status overall between the pharynx and bronchus using McNemar s test Table 3.13: Comparison of paired proportions for M. felis status between the conjunctiva and pharynx using McNemar s test Table 3.14: Comparison of paired proportions for M. felis status between the conjunctiva and bronchus using McNemar s test Table 3.15: Comparison of paired proportions for M. felis status between the pharynx and bronchus using McNemar s test Table 3.16: Comparison of paired proportions for M. gateae/m. arginini status between the conjunctiva and pharynx using McNemar s test ix

12 Table 3.17: Comparison of paired proportions for M. gateae/m. arginini status between the conjunctiva and bronchus using McNemar s test Table 3.18: Comparison of paired proportions for M. gateae/m. arginini status between the pharynx and bronchus using McNemar s test Table 3.19: Number of each mycoplasma species at any site for both healthy and diseased cats. Table 3.20: Comparison of proportion of each mycoplasma species present in conjunctival swabs between cats with and without conjunctivitis Table 3.21: Comparison of proportion of each mycoplasma species present in pharyngeal swabs between cats with and without upper respiratory tract signs Table 3.22: Comparison of proportion of each mycoplasma species present in bronchial swabs between cats with and without lower respiratory tract signs x

13 LIST OF FIGURES Figure 2.1: Location of signs of disease observed in the study population of cats Figure 2.2: Anatomic distribution of mycoplasma-positive swabs Figure 2.3: Anatomic distribution and combination of mycoplasma-positive swabs Figure 3.1: Mycoplasma 16S rrna gene, primers, and amplified products Figure 3.2: Gel showing PCR products amplified by primers MycoF and MycoR Figure 3.3: Gel showing PCR products amplified by primers FD1000 and MycoR Figure 3.4: SSCP gel of feline mycoplasma samples showing two distinct profiles; F and R Figure 3.5: SSCP gel of feline mycoplasma samples showing variant F profiles in addition to the two profiles F and R Figure 3.6: Nucleotide sequence alignment of sample 103C (profile F) with the sequence representing M. felis Figure 3.7: Nucleotide sequence alignment of F profile variant samples (110C and 111B) with M. felis and regular F profile (103C) Figure 3.8: Nucleotide sequence alignment of sample 14A (profile R) with M. gateae Figure 3.9: Comparison of mycoplasma species distribution across the different anatomic locations studied xi

14 ABBREVIATIONS AFLP Amplified fragment length polymorphism bp Base pairs BAL Bronchoalveolar lavage DGGE Denaturing gradient gel electrophoresis DNA Deoxyribonucleic acid EDTA Ethylene diamine tetraacetic acid FBD Feline bronchial disease FCV Feline calicivirus FeLV Feline leukaemia virus FHV Feline herpesvirus FIV Feline immunodeficiency virus Ig Immunoglobulin LAMP Lipid associated membrane protein LRT Lower respiratory tract PAMP Pathogen associated molecular pattern PBS Phosphate buffered saline PCR Polymerase chain reaction RNA Ribonucleic acid SPF Specific pathogen free SSCP Single-strand conformation polymorphism TLR Toll-like receptor URT Upper respiratory tract UV Ultraviolet 16S rrna gene The 16 small subunit ribosomal RNA gene 16-23S IGS Intergenic spacer region between the 16 and 23 small subunit ribosomal RNA genes xii

15 Chapter 1 Introduction and Literature Review 1.1 Introduction Mycoplasmas are one of the smallest known free-living organisms. Their small size, combined with plasticity due to the absence of a rigid cell wall leaves them well suited to close associations with cellular membranes. These intimate cellular environments not only provide osmotic protection, but the fastidious oxygen, temperature and nutritional requirements of mycoplasmas can also be met. Consequently, mycoplasmas are commonly found living on mucous membranes of their mammalian hosts (Razin and Freundt, 1984). The normal host defence mechanisms of the conjunctiva, upper respiratory and urogenital tracts tolerate a resident microflora population. As a consequence, mycoplasmas commonly colonise the upper respiratory tract (URT) compared with the relatively sterile environment of the lower respiratory tract (LRT). Within this niche, mycoplasmas must develop specific methods to avoid injuring the host cells and activating host defence mechanisms. The conjunctiva and URT primarily use secretions containing IgA and mucosal associated lymphoid tissue for their defence, while the lower respiratory tract also has a combination of ciliated epithelial cells and mucous covering, making it more difficult for organisms to attach. In addition, the lower respiratory tract has a wider array of lymphoid tissue, with location specific immune cells (alveolar macrophages and Clara cells) and antibodies (IgA, IgE, and IgG) (reviewed in (Cohn and Reinero, 2007; Lopez, 2007)). To survive, mycoplasmas need to attach to host epithelial cells so their physiologic requirements can be met, which also provides protection from mechanical flushing ((Razin, 1978), cited in (Cassell et al., 1985)). Irrespective of any specific pathogenic mechanisms, mycoplasmas must be able to evade the host defences to cause local damage or to enter the body, and thus must have substantial capacity for attachment as well as specific virulence factors (reviewed in (Cassell et al., 1985; Rottem, 2003)). Moreover, mycoplasmas may act synergistically with other organisms, such as bacteria or viruses, to cause disease. 1

16 Due to their attachment to mucous membranes, disease caused by mycoplasma is typically confined the conjunctiva, respiratory tract and/or urogenital tract (Rosenbusch, 1994). However, when mycoplasma infection becomes systemic, synovial infection is common (Rosenbusch, 1994). Mycoplasmal respiratory diseases of veterinary importance include Contagious Bovine Pleuropneumonia, Contagious Caprine Pleuropneumonia, Enzootic Pneumonia of pigs and Chronic Respiratory Disease of poultry (Cassell et al., 1985). Most mycoplasma diseases, although chronic, are often not fatal. Where mycoplasmas are known to cause disease, they may do so inconsistently and can be found in apparently healthy hosts. This proves challenging in determining their exact role in many diseases and despite what is known, the pathogenesis of mycoplasmal diseases are poorly understood. The conditions under which mycoplasmas are pathogenic are multifactorial, based not only on virulence but also host and environment factors. Respiratory disease in animals usually has a complex and multifactorial aetiology. The syndrome known as cat flu or feline upper respiratory disease complex (broadly including conjunctivitis) may involve a number of aetiologic agents, including feline herpesvirus (FHV), feline calicivirus (FCV), Chlamydophila felis, Bordetella bronchiseptica and Mycoplasma spp. In the past, it was thought that viral agents played the major role in feline upper respiratory disease, being exacerbated or complicated by superimposition of secondary bacterial infections. Although the pathogenic role of viruses has been clearly demonstrated, there is evidence to suggest that these bacteria may cause primary disease (Gaskell and Bennett, 1996; Dawson and Willoughby, 1999). The epidemiology of feline respiratory disease is likely to be complex. Host, pathogen and environmental factors are variable, and when considering the prevalence of mycoplasma in feline populations, it is important to differentiate between presence of the organism, and an association with disease. Furthermore, the boundaries between a commensal organism, opportunistic invader and pathogen can be difficult to determine. Microbiological disease of the respiratory tract has an inherent complexity, particularly when identifying multiple potentially causative organisms and determining their role. Additionally, mycoplasmas tend to be fastidious, with stringent growth conditions, making them either difficult to isolate or slow to grow. The 2

17 development of rapid and sensitive molecular techniques, such as PCR and nucleotide sequencing, have improved the characterisation of mycoplasmas and other organisms to the species level both for diagnostic purposes and research. 1.2 Research Objectives Syndromes involving ocular and respiratory diseases are widespread and significant problems in cats. Mycoplasmas have a demonstrated ability to act as primary pathogens in respiratory disease of other host species, and have been reported to be a primary pathogen in several feline case studies. The aim of this thesis is to investigate the epidemiology of feline mycoplasmas and any association with disease by surveying a population of shelter cats. The objective of this chapter is to concisely and critically review the current literature relating to feline mycoplasmas; in particular, the species of mycoplasma known to infect cats, their distribution and prevalence, and review the evidence of their role in disease. Current methods of isolation and characterisation of these organisms are also reviewed. This research, and hence review will be limited to surface mycoplasmas, not the haemotropic organisms recently reclassified as mycoplasmas. 1.3 Literature Review Mycoplasmas general characteristics Mycoplasmas are amongst the smallest known free-living prokaryotic organisms. They range in size from µm (Razin and Freundt, 1984), with a genome size ranging from 580 to 1300 kb (Morowitz and Wallace, 1973; Razin, 1992). They belong to the Class Mollicutes (meaning soft skin), Order Mycoplasmatales, Family Mycoplasmataceae and the Genus Mycoplasma (meaning fungus form). There are at least 112 defined species of mycoplasma (Tully and Bradbury, 2003). Mycoplasmas have evolved into their ecological niche by a process of genome minimisation (reviewed by (Bradbury, 2005)). The simplicity of the mycoplasma genome means they have the machinery for complex protein synthesis, but rely on hosts for their survival, and this also accounts for their slow growth. 3

18 Mycoplasmas have been found in an extensive range of avian and mammalian hosts, including humans, and are found as a common contaminant of cell cultures and veterinary vaccines (Thornton, 1986; Kojima et al., 1996). Mycoplasmas are usually host specific, particularly those species that are pathogenic (Whitford et al., 1994). Some mycoplasma species, such as M. arginini are found in multiple host species (Razin and Freundt, 1984). Mycoplasmas cannot synthesise peptidoglycans and thus have no rigid cell wall, but a flexible triple-layered outer plasma membrane, allowing pleomorphism. Owing to the absence of a cell wall, mycoplasmas are usually not resilient in the environment, and hence, are found living in association with their hosts, most commonly on mucous membranes. Thus, the types of diseases they are associated with are those of the respiratory tract, conjunctiva and urogenital tract. Less than half of known mycoplasma species cause a specific disease or have a defined pathogenicity (Razin and Freundt, 1984). Most species isolated in association with disease have a poorly defined role in the disease pathogenesis, while other mycoplasma species are considered to be part of the normal microbial flora of mucosal surfaces. Radostits et al. (1994) highlight this point: There is uncertainty about the real importance of mycoplasmas in many diseases from which they are consistently isolated (Radostits et al., 1994). Some mycoplasmas do, however, cause significant and widespread disease, particularly of the respiratory system. These have important health and economic consequences in pig, poultry and ruminant production industries; hence, there has been extensive research into these organisms, particularly focussing on pathogenesis, virulence and vaccine antigen targets (Whithear, 1996; Minion, 2002; Thiaucourt et al., 2003; Bradbury, 2005; Nicholas et al., 2009). Much less is known about the role and characteristics of mycoplasmas in companion animals, such as horses, dogs and cats (Whitford and Lingsweiler, 1994). Mycoplasmas have been associated with, implicated in, or isolated from cases of conjunctivitis (Cello, 1957; Cole et al., 1967; Tan and Markham, 1971b; Campbell et al., 1973b; Shewen et al., 1980; Haesebrouck et al., 1991; Low et al., 2007; Hartmann et al., 2008), upper respiratory disease (Schneck, 1972; Schneck, 1973; Bannasch and 4

19 Foley, 2005; Johnson et al., 2005; Huebner et al., 2007; Hartmann et al., 2008; Veir et al., 2008; Johnson and Kass, 2009), pneumonia (Tan, 1974; Tan and Miles, 1974b; Randolph et al., 1993; Foster et al., 1998; Bart et al., 2000; Chandler and Lappin, 2002; Foster et al., 2004b; Foster et al., 2004c; Trow et al., 2008), pyothorax (Malik et al., 1991; Gulbahar and Gurturk, 2002), bronchial disease (Foster et al., 2004a; Foster et al., 2004b; Foster et al., 2004c), abscesses (Keane, 1983; Crisp et al., 1987; Walker et al., 1995; Adamson, 2004), aural polyps (Cross, 2004; Klose et al., 2007), and arthritis (Moise et al., 1983; Hooper et al., 1985; Liehmann et al., 2006; Zeugswetter et al., 2007) in cats. However, evidence for the role of mycoplasmas as primary pathogens in these diseased cats is rarely conclusive. The limited investigation of feline mycoplasmas may be a reflection of this species lacking the same economic significance as production animals. Alternatively, feline mycoplasmas may not cause enough economic loss or morbidity. This may be because cats are seldom kept in conditions of close confinement similar to that of intensive piggeries or poultry establishments. In addition, routine cultures from cats do not usually include mycoplasma-specific media Species of mycoplasma isolated from cats The earliest report of mycoplasma isolation in a cat was from the lung of a 6 week old kitten with pneumonia in 1954 (Switzer, 1967). Mycoplasma was then found in cats with conjunctivitis (Cello, 1957) and in the saliva of healthy cats (Cole et al., 1967). Cole et al. (1967) characterised the isolates by their morphology, biochemical requirements and antigenic composition, and identified two distinct species, M. felis and M. gateae. Subsequently, two other species were discovered in domestic cats: M. feliminutum (Heyward et al., 1969) and M. arginini (Tan and Miles, 1974b). It is these four species of mycoplasma that are frequently isolated from cats, or are recognised as having cats as their primary host. Although M. felis is relatively host-specific, it has been isolated from other host species including horses (Ogilvie et al., 1983; Rosendal et al., 1986; Carman et al., 1997; Wood et al., 1997), and M. gateae has been frequently isolated from dogs (Rosendal, 1979). Although M. arginini is found in cats, it has a wide host distribution (Razin and Freundt, 1984). M. feliminutum differs from the other feline mycoplasmas, 5

20 with only one report of isolation in cats (Heyward et al., 1969), one in a horse and on more than one occasion in dogs (Razin and Freundt, 1984). Despite these four mycoplasma species having cats as a host, they are not phylogenetically close based on 16S rrna sequence data (Brown et al., 1995). Other mycoplasma species isolated from cats include M. pulmonis, M. arthritidis and M. gallisepticum (Tan et al., 1977a). These species were isolated from the oropharynx of a small number (representing less than 4% of positive isolates) of clinically normal cats (in addition to M. gateae, M. felis and M. arginini). In addition, this study reported a number of oropharyngeal mycoplasma isolates (2% of positive isolates) whose species status was unable to be determined by growth inhibition tests using reference antisera against 14 different mycoplasma species. A previous study by Tan and Miles (1974b) cultured 15 mycoplasma isolates from the oropharynx of diseased cats, and 19 isolates from the oropharynx of moribund cats which were not identified as either M. gateae, M. felis or M. arginini by growth inhibition. No attempt was made to investigate these isolates further. It is therefore possible that there are more mycoplasma species that inhabit the oropharynx of cats. Although it seems unlikely that the other species have a pathogenic role, it is important to characterise the normal flora of the oropharynx as there is potential for opportunistic secondary mycoplasmal infections. Other organisms related to mycoplasmas of the Genus Ureaplasma and Acholeplasma have also been recovered from domestic cats (Tan and Markham, 1971a; Tan and Miles, 1972; Tan and Miles, 1974a; Tan and Miles, 1974b; Tan et al., 1977a; Randolph et al., 1993; Senior and Brown, 1996). These are readily differentiated from mycoplasmas using biochemical characteristics; Ureaplasmas contain the enzyme urease and can therefore hydrolyse urea, whereas Acholeplasmas do not require sterols for growth (Razin and Freundt, 1984). Neither species have been significantly associated with disease in cats, apart from some experimental evidence of Ureaplasmas inducing abortions in pregnant queens (Tan and Miles, 1974a). Following early descriptions of mycoplasmas in cats, further research investigated the extent of the presence and role of mycoplasma in causing disease. Different research groups contributed significantly to this over a 10-year period from 1969, largely by 6

21 investigating the presence of mycoplasma in greater numbers of both healthy and diseased cats, by characterisation and identification of mycoplasma to the species level (utilising biochemical and serological methods), and the use of experimental transmission studies to determine their pathogenic potential (Heyward et al., 1969; Blackmore et al., 1971; Tan and Markham, 1971a; Tan and Markham, 1971b; Tan and Miles, 1972; Blackmore and Hill, 1973; Tan and Miles, 1973; Tan and Miles, 1974a; Tan, 1974; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b). Despite more recent advances in molecular and genetic technologies, very little is known about the genetic characteristics of feline mycoplasmas in relation to major genes, attachment structures, antigenic regions and virulence factors. There has, however, been characterisation of the 16S rrna gene nucleotide sequence, which has provided a molecular basis for species identification and phylogeny construction of feline mycoplasmas (Brown et al., 1995) Methods for the isolation and identification of feline mycoplasmas Collection techniques The conventional method for isolating mycoplasmas from mucosal surfaces is recovery via swab samples, which can be either streaked directly on to plates or placed into liquid culture media (Rosendal, 1979). Tissue mycoplasmacidal factors have been described by various authors (Tully and Rask-Nielsen, 1967; Kaklamanis et al., 1969; Rosendal, 1979) which may be liberated following the homogenisation of tissue samples. To prevent this problem, tissue samples should be pushed onto plates and then placed in broth (Rosendal, 1979). In contrast, Razin and Freundt (1984) suggested that tissues should be dissected coarsely and then added to liquid mycoplasma media followed by at least two ten-fold dilutions in broth to reduce concentrations of inhibitory components. Mycoplasmas have been detected directly by PCR from swabs of feline conjunctiva (Bannasch and Foley, 2005; Huebner et al., 2007; Low et al., 2007; Hartmann et al., 2008), oropharynx (Huebner et al., 2007; Hartmann et al., 2008; Veir et al., 2008), and nasal flush and biopsy samples (Johnson et al., 2004; Veir et al., 2008). The use 7

22 of direct swabs for subsequent PCR amplification (of a suitable DNA target region) might circumvent the effects of mycoplasmacidal components on culture, because viable organisms are not required for DNA amplification by the PCR. Other studies have utilised bronchoalveolar lavage (BAL) to isolate mycoplasmas from the lower respiratory tract of cats (Randolph et al., 1993), and retrospective studies of lower respiratory tract cytology, microbiology and disease are based on the use of BAL to culture the organisms (Padrid et al., 1991; Chandler and Lappin, 2002; Foster et al., 2004a; Foster et al., 2004b; Foster et al., 2004c). BAL is more applicable to, and widely used in, the clinical setting and may also avoid the release of mycoplasmacidal factors by avoiding tissue disruption. The relative efficacy of retrieving different mycoplasma species using lavage techniques may depend on their varying degrees of adherence or attachment. This aspect has not been studied in the LRT of cats but was considered in a study comparing the relative isolation of mycoplasmas using both nasal flush and biopsy for comparison (Johnson et al., 2004). Part of the rationale for the study by Johnson et al. (2004) was based on a study which investigated porcine mycoplasmas, and found pathogenic mycoplasmas adhered to cilia on epithelial cells whereas non-pathogenic species did not (Young et al., 2000). There was no significant difference between nasal flush and biopsy samples for the isolation of mycoplasmas in this, or a more recent analysis of cats with URT disease (Johnson et al., 2004; Johnson and Kass, 2009). As the respiratory health of the cats was not considered in the former study, and there was little discrepancy between the sampling methods, no further analysis was undertaken to determine any difference in mycoplasma species represented or to relate it to their pathogenicity. Handling of samples is important in the recovery of mycoplasma, both by culture and PCR. Mycoplasma-specific media or other transport medium (such as Amies without charcoal or Stuart s) is required to maintain viable organisms in samples to be cultured (Razin and Freundt, 1984; Whitford, 1994). There are three important rules for the transport of samples, which are to keep the sample moist, cool and to move fast (Whitford, 1994). Isolation is maximised if samples are placed in media within a 8

23 few hours of collection, but if delays of more than 24 hours are inevitable, freezing at -70 ºC is recommended (Whitford, 1994). Methods of collection and transport of feline swabs which have been utilised for direct PCR analysis of mycoplasmas have differed among studies. Generally, studies have been based on rolling dry or moistened swabs across the mucous membrane, placing the sample in sterile saline or phosphate-buffered saline and either processed within a few hours or frozen at -20 ºC, -70 ºC, or -80 ºC until processing (Johnson et al., 2004; Bannasch and Foley, 2005; Low et al., 2007; Sjödahl-Essén et al., 2008; Veir et al., 2008). Blackmore et al. (1971) demonstrated that the recovery of mycoplasmas at post mortem closely corresponded to results obtained from taking swabs from animals ante mortem Culture media Mycoplasmas are facultative anaerobes, requiring cholesterol or related sterols for growth, and use either sugars or arginine as their major energy source. Media requirements are complex due to the limited ability of mycoplasmas to produce many nutrients. Mycoplasma media and its components have previously been described (Edward, 1947; Hayflick, 1965; Frey et al., 1973), and with minor modifications, are widely used for the isolation of feline mycoplasmas (Cole et al., 1967; Heyward et al., 1969; Spradbrow et al., 1970b; Blackmore et al., 1971; Hill, 1971; Tan and Markham, 1971b; Tan and Miles, 1972; Blackmore and Hill, 1973; Tan and Miles, 1973, 1974b; Tan et al., 1977a; Tan et al., 1977b; Haesebrouck et al., 1991; Johnson et al., 2004; Johnson and Kass, 2009). An important consideration in the formulation of mycoplasma media is to provide osmolarity needed due to the absence of a cell wall (Rosenbusch, 1994). Animal serum is a source of sterols and fatty acids, while yeast extract, glucose or arginine can be used as a source of metabolisable energy (Rosenbusch, 1994). Different mycoplasma species preferentially utilise either one of these energy sources, and phenol red is added as an indicator to differentiate ph changes associated with different energy utilisation (Ruhnke and Rosendal, 1994). Antimicrobial agents 9

24 penicillin and thallium acetate are added to inhibit bacterial overgrowth or contamination (Edward, 1947). The comparative efficacy of recovering feline mycoplasmas using both solid and liquid media has been described (Hill, 1971). This study found that there were 6 (16%) additional isolates obtained from inoculating liquid media that were not apparent from solid media alone. Additionally, mycoplasmas could be isolated from up to 20% of the liquid media that had not shown any obvious change in ph after a 2- week culture period. This finding demonstrated the importance of inoculating liquid media in addition to culture plates, and also that ph change in liquid media may be an insensitive method for determining the presence of mycoplasmas Culture conditions As none of the known feline mycoplasmas are strict anaerobes, normal atmospheric conditions, or room air with 5% CO 2 is satisfactory for growth (Rosendal, 1979; Whitford and Lingsweiler, 1994). The optimal temperature for growth of mycoplasmas from animal origin is ºC (Razin and Freundt, 1984). Mycoplasma species recovered from cats have been grown aerobically at 37 ºC (Heyward et al., 1969) or in 5% CO 2 (Tan and Miles, 1972, 1974b) Morphology Compared with most other bacteria, mycoplasma colonies are very small (< 1 mm diameter) and are typified by a fried-egg appearance of a dense granular core resulting from their tendency to grow down into the solid media, with a flat translucent periphery of outward growth. Some species have distinct features of colony morphology, but this is not a sensitive criterion for identification, as variation can arise due to variations in media and growth conditions (Razin and Freundt, 1984). The morphological characteristics of each of the four feline mycoplasma species are listed in Table

25 Biochemical profiles A standard set of biochemical tests has been established for the characterisation of mycoplasma isolates by the International Committee on Systematic Bacteriology, Subcommittee on the Taxonomy of Mollicutes (Brown et al., 2007). These protocols are reviewed in (Goll, 1994), and are beyond the scope of this review. The biochemical profiles of most mycoplasma species of veterinary importance have been summarised (Razin and Freundt, 1984; Ruhnke and Rosendal, 1994). The distinguishing metabolic features of the four common feline mycoplasma species are given in Table Serological methods for specific identification Following an assessment of the morphological and biochemical characteristics of Mycoplamas, serological tools have, until recently, been used as the standard approach for the specific identification of mycoplasma (Brown et al., 2007). Specific identification is made by comparing the serological characteristics of an isolate with those of reference strains using various techniques such as growth inhibition, metabolic inhibition, immunofluorescence and immunobinding. These techniques have been reviewed elsewhere (Rosendal, 1979; Goll, 1994; Brown et al., 2007) and are beyond the scope of this review. However, it is important to understand the limitations of these techniques during critical review of the literature where mycoplasmas have been identified by these means. The use of serological methods for specific identification has certain disadvantages, which relate predominantly to cross-reactivity among mycoplasma species as a result of common antigens (Goll, 1994). Although each feline mycoplasma species can be demonstrated as serologically distinct, a degree of cross-reactivity can occur between them in particular tests (Cole et al., 1967; Heyward et al., 1969; Rosendal, 1979). Test interpretation must therefore be made with caution and consideration of test accuracy and limitations, which has been reviewed elsewhere (Goll, 1994) and are beyond the scope of this review. A major disadvantage of serological methods compared with more recent molecular methods is the requirement for maintenance of a wide array of reference type cultures and antisera. As antisera are produced in experimental 11

26 Table 1.1: Comparative colony characteristics for feline mycoplasma species (Heyward et al., 1969; Hill, 1971) Species Colony morphology Average colony Time for growth (h) size (µm) M. felis Large circular colonies, rust-like hue, prominent central core, dense periphery that becomes brown with age (brown colour can take 7-14 days to develop) M. gateae Clear, vacuolated colonies, small central core M. arginini As above with more prominent central core M. feliminutum Small colonies, irregular shape, no central core Table 1.2: Comparative biochemical characteristics of feline mycoplasma species (Razin and Freundt, 1984; Whitford et al., 1994) Species Glucose fermentation Arginine hydrolysis Phosphatase activity Expected colour change in liquid media * M. felis yellow M. gateae red M. arginini red M. feliminutum x - + x (* based on the use of phenol red indicator, whereby acidification of the media produces a yellow colour change, and alkalinisation a red colour change; x = unresolved)

27 animals, the use of molecular methods reduces the need for animal use, an important welfare consideration, and also averts the expense of maintaining animal colonies. Not withstanding these limitations, the serological characteristics of feline mycoplasma isolates have been described and are diagnostically useful (Cole et al., 1967; Heyward et al., 1969; Rosendal, 1979). M. felis has been demonstrated to be serologically distinct from M. gateae, M. arginini and M. feliminutum (Cole et al., 1967; Heyward et al., 1969). The single M. feliminutum isolated was serologically distinct from other feline species (Heyward et al., 1969). In addition to the very similar morphological and identical biochemical characteristics, the close relationship of M. gateae and M. arginini is also evident in their serological cross-reactivity (Heyward et al., 1969; Rosendal, 1979). In contrast, an immunobinding assay specifically designed to distinguish feline isolates of M. felis and M. gateae was tested against cultures of M. arginini and M. feliminutum; no cross-reactivity was evident using this method (Brown et al., 1990) Molecular methods for specific identification Molecular methods for the genetic characterisation of mycoplasmas have largely superseded the use of serological methods. The International Committee on Systematics of Prokaryotes Subcommittee on the Taxonomy of Mollicutes recognised this recently (Brown et al., 2007). The 16S rrna gene nucleotide sequence is now the primary method for specific identification and determination of phylogenetic relationships, provided that the serological and phenotypic data substantiates identification (Brown et al., 2007). Specific amplification of target DNA regions such as the 16S rrna gene by the PCR can be used to identify isolates to the species and/or subspecies level. This is usually performed by comparison of sequence data with those in databases such as GenBank (National Centre for Biotechnology Information, Bethesda, MD, USA which contain reference sequences. Differences in nucleotide sequence can be used to develop PCR based methods for the rapid identification of different species. 12

28 Some advantages of molecular techniques (compared with serological) are that they generally are more rapid, accurate (specific), and avoid problems with culture of nonviable, fastidious or contaminant organisms (as reagents are directed specifically at mycoplasma) (Brown et al., 1995; Chalker et al., 2004; Johnson et al., 2004). Using these techniques requires standardisation and validation, using appropriate controls, to maximise their accuracy and efficiency. Particular attention must be paid to meticulous technique due to the incredible sensitivity of PCR to contamination. Primers universal to the 16S rrna genes of mycoplasmas have been used for the purpose of differentiating among feline mycoplasma species (Brown et al., 1995). The PCR-amplified nucleotide sequences from feline mycoplasma isolates were determined. An approach was developed with the use of particular restriction endonucleases to cleave PCR products, yielding digestion patterns of characteristic fragment sizes for each mycoplasma species (Brown et al., 1995). Variations of these techniques have subsequently been applied to feline population studies to amplify DNA from clinical isolates for mycoplasma identification (Bannasch and Foley, 2005; Gray et al., 2005; Huebner et al., 2007; Klose et al., 2007; Low et al., 2007; Hartmann et al., 2008; Sjödahl-Essén et al., 2008; Veir et al., 2008). The genetic relationships of some mycoplasma species have been examined using the 16S-23S rrna intergenic spacer region (IGS) (Harasawa, 1999). This region is less conserved than the 16S rrna gene, but as it contains both conserved and variable regions in a bp sequence, it was considered to be a useful region for the genetic comparison of species within a genus (Harasawa, 1999). Results from the study demonstrated phylogenetic relationships of mycoplasma which were similar to those proposed previously using 16S rrna gene data (Harasawa, 1999). Sequence variability in the IGS within species must also be considered to determine its value for species differentiation. Investigation of the use of PCR employing the 16S-23S IGS region to identify M. felis in feline clinical samples demonstrated a high level of sequence conservation in this region among M. felis isolates (Chalker et al., 2004). The PCR was specific for M. felis when tested against three other feline mycoplasma species (Chalker et al., 2004). A variation of this PCR technique (Chalker et al., 2004) has recently been modified for use with real-time PCR for M felis, allowing 13

29 quantification of DNA and, by inference, organisms from swab samples (Nicholas et al., 2008b; Sjödahl-Essén et al., 2008). Other techniques such as restriction endonuclease analysis may be used in combination with PCR to differentiate among species of mycoplasma, particularly to reduce the requirement for sequencing every isolate (Brown et al., 1995). An alternative method integrating the use of restriction endonucleases is whole genome fingerprinting using the amplified-fragment polymorphism (AFLP) method (Vos et al., 1995). AFLP is based on the selective amplification of restriction fragments by PCR to create characteristic banding patterns for nucleotide sequences representing different species. AFLP has been used for identifying and characterising both interand intra-species variability in a range of mycoplasma species (Kokotovic et al., 1999). Other molecular typing methods have been reviewed in (Nicholas et al., 2008c), and are beyond the scope of this review. PCR-coupled mutation detection techniques have proven to be useful across a range of organisms for rapidly screening large numbers of samples for sequence variation, and in particular for low level detection of variation (reviewed by (Gasser, 1997; Gasser and Chilton, 2001; Gasser et al., 2006)). Denaturing gradient gel electrophoresis (DGGE) and single-strand conformation polymorphism (SSCP) are examples of such techniques. DGGE utilises the differential melting behaviour of DNA fragments resulting from sequence variation. Fragments are separated in a gel containing a gradient of chemical denaturants (such as urea and formamide) at a high temperature (Gasser, 2001; McAuliffe et al., 2003). The concentration of denaturant at which the DNA molecule separates is based on its nucleotide composition and sequence, and the electrophoretic mobility decreases as the DNA becomes partially denatured (Gasser, 2001). DGGE has been assessed as a means of rapid identification for a range of mycoplasma species, including M. felis (McAuliffe et al., 2003; Nicholas et al., 2008a). In this study (McAuliffe et al., 2003), intra- and inter-species variation was demonstrated in the V3 region of the 16S rrna gene by the use of DGGE. Mycoplasma species were readily differentiated with this technique if sufficient sequence variation existed in this region. Some closely related species contained identical nucleotide sequences in this region and were thus unable to be differentiated, which was a limitation of study design rather than of the technique. 14

30 The principle of SSCP is that single-stranded DNA assumes one or more conformations based on its nucleotide sequence, as pairing between nucleotides occurs to form structures. These conformations are highly dependent upon the nucleotide sequence, and variation in the sequence alters the structure of the molecule (Gasser et al., 2006). This variation (or polymorphism) can be detected based on differing migration through a non-denaturing electrophoretic gel, and is evident from unique banding profiles when visualised with a variety of staining methods or autoradiography. The sensitivity or mutation detection rate of the SSCP technique is dependent on a number of factors, most importantly being size of the DNA fragment used and electrophoresis conditions (Gasser, 2001). The DNA fragment sizes most suited to SSCP were sequences of 100 to 500 base pairs (Gasser et al., 2006). Although the mutation detection rate can decrease with sequences of > 200 bp (Gasser et al., 2006), point mutations have been detected in sequences of up to 530 bp (Zhu and Gasser, 1998). A detailed protocol for the use of SSCP has been described including the use of both radiolabelled and non-isotopic SSCP, and the different staining methods that can be used for detection (Gasser et al., 2006). In addition to detecting sequence variation or mutations between isolates or samples, sequence variation has been demonstrated in samples by SSCP (Gasser et al., 2003; Chalmers et al., 2005). This, and the ability of SSCP to detect point mutations over a relatively large sequence (compared with variability within restriction endonuclease cleavage sites) and to rapidly screen large sample numbers with relative technical ease (Gasser et al., 2006), demonstrate advantages over techniques such as restriction endonuclease analysis. PCR-SSCP has been applied to the study of mycoplasmas, first to study the molecular basis for resistance of M. hominis to fluoroquinolones (Gushchin et al., 1998), and more recently, to classify isolates of M. synoviae to the subspecies level (Jeffery et al., 2007). As both population and molecular epidemiological studies are important in investigating infectious disease agents, PCR-SSCP should find increased applications to mycoplasmas. 15

31 1.3.4 Epidemiology of feline mycoplasmas Although the exact role that mycoplasma plays in ocular and respiratory disease of cats, or specific host-organism relationships are not well known, current knowledge can be reviewed in an attempt to determine a possible epidemiologic model for disease and identify knowledge gaps. Epidemiological and ecological factors are considered in terms of where mycoplasmas are found in cats and what species are present. Additionally, geographical and climatic factors, the presence and interaction of mycoplasmas with other respiratory pathogens, and host factors such as age and immune response are reviewed Prevalence of mycoplasma in cats There are numerous challenges in comparing studies of mycoplasma prevalence in feline populations. Discrepancies across almost all studies hinder specific comparisons, however broad generalisations are possible. There are differences such as whether cats are owned or stray/shelter cats, gender, age, health status, geographical location, climate and time of year. In addition, the presence of other respiratory pathogens or systemic diseases, such as potentially immunosuppressive diseases, may influence the presence and isolation of mycoplasmas. Furthermore, variation in collection techniques, growth media and conditions, and anatomic sites sampled may affect results. Findings are often presented as numbers of positive samples out of total number of samples, which does not allow prevalence data to be compared (in terms of numbers of cats and/or anatomic site). Where mycoplasma species were defined, methods often differed from those (primarily serological) employed in earlier studies to molecular methods in more recent investigations. Differing sensitivity and specificity of different detection methods has consequences in relation to the accuracy of specific identification. It is therefore useful to consider prevalence data across groups or populations of cats, as they can then be compared and contrasted, drawing the most meaningful conclusions from the data. 16

32 Prevalence with reference to anatomical location Mycoplasmas have been isolated from a range of anatomical sites in cats including the conjunctiva, upper and lower respiratory tracts, urogenital tract, rectum, joint spaces and ear canals (Heyward et al., 1969; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b; Moise et al., 1983; Haesebrouck et al., 1991; Randolph et al., 1993; Senior and Brown, 1996; Cross, 2004; Klose et al., 2007). The significance of isolating mycoplasma and associations with disease can be inferred by establishing the relative frequency of isolation from both diseased and healthy cats. The upper respiratory tract of healthy cats is inhabited by a resident population of microorganisms which include Pasteurella multocida, non-haemolytic Streptococcus, and Escherichia coli (Lappin et al., 2007). Feline herpesvirus and feline calicivirus have also been isolated from the URT of healthy cats (Ellis, 1981). In addition, anaerobic bacteria from the oral cavity of healthy cats have been characterised, and include isolates from the genera Actinomyces, Bacteroides and Fusobacterium (Love et al., 1990). The association of microbial flora with the URT is dependent on both the nature of host defence mechanisms and the attributes of the organisms that allow them to remain there. Understanding the non-pathogenic association of organisms with their hosts in this way may provide a basis for comparison when disease results from these or other pathogenic organisms. Mycoplasmas live on the mucous membranes in the URT if they can attach or adhere firmly enough to withstand flushing by secretions, and if they can evade host defences. Mycoplasmas may avoid non-specific host defences such as phagocytosis by expressing antiphagocytic activity or resulting from close attachment of the organism to the host cells (reviewed in (Cassell et al., 1985)). Another way that mycoplasmas may be permitted to colonise the epithelium is a delay or prevention of antigenic recognition which may result from organisms being somewhat protected by their location on the epithelium. Biofilm formation has been demonstrated in some species of mycoplasma such as M. bovis, and some authors have speculated that this may have a role in resisting host defences (McAuliffe et al., 2006). Additionally, the ability to adsorb serum proteins or host cellular antigens (as demonstrated for M. hyorhinis) may mask mycoplasma-specific antigens or mimic host antigens (reviewed in (Cassell et al., 1985)). Alternatively, exposure to small amounts of mycoplasmal 17

33 antigen may induce low-dose tolerance ((Nossal, 1983), cited in (Cassell et al., 1985)). Mycoplasmas may evade effector mechanisms; for example mycoplasmas may be firmly established on the surface of host cells before an IgA response is activated (reviewed in (Cassell et al., 1985)). If mycoplasmas breach the epithelium or cause cell damage, the immune system will alert their presence and the resultant inflammatory response may destroy them. However even in the face of a specific immune response, the mycoplasmas may be protected by localisation on the epithelial surface where they are not readily accessible to phagocytic or lymphoid cells (Cassell et al., 1985). Thus they may still reside on mucous membranes of the URT after an immune response has cleared those organisms that gained systemic entry or invaded more distally into the respiratory tract where there is a greater armoury of immune mechanisms. The highest frequency of mycoplasma recovery has been from the URT. The prevalence of mycoplasma in the pharynx of healthy cats ranges from 35 to 100% (Jones and Sabine, 1970; Blackmore et al., 1971; Tan and Miles, 1972; Tan et al., 1977a; Haesebrouck et al., 1991; Randolph et al., 1993), which is comparable to diseased cats (Spradbrow et al., 1970b; Schneck, 1973; Tan and Miles, 1973; Randolph et al., 1993; Hartmann et al., 2008; Johnson and Kass, 2009), suggesting that mycoplasmas live as commensal organisms in the URT. Bannasch and Foley (2005) used an unmatched case control study, taking swabs from the conjunctiva and oropharynx of cats both with and without upper respiratory disease in animal shelters. Mycoplasma was found in 80 of 573 cats (14%); 65 isolates were from cats with upper respiratory disease and 15 without; the difference was statistically significant with an odds ratio of Cats with mycoplasma were found to have more severe ocular discharge, conjunctivitis and nasal discharge compared to cats without mycoplasma. This study provides evidence that mycoplasma can have a significant association with upper respiratory disease in cats. However, the study did not investigate mycoplasma-positive cats that had concurrent infection with other organisms, and only conjunctival samples were tested for mycoplasma. The presence of mycoplasma in the oropharynx was not included in the analysis, leaving the involvement of mycoplasma in disease in the oropharynx unclear. 18

34 The conjunctiva is a mucosal surface in contact with the external environment, and is accessible to microorganisms as a possible route of entry to the host. The host defences in this area consist of antibacterial factors and lysozymes in secretions of lacrimal and associated glands, mucosal immune responses from submucosal lymphocytic tissue and mechanical flushing by tears (Tizard, 1996). Organisms demonstrated in the conjunctiva of healthy cats are Staphylococcus epidermidis, Staphylococcus aureus, Streptococcus spp, Mycoplasma spp., Bacillus spp., and Corynebacterium spp. (Campbell et al., 1973a; Shewen et al., 1980; Low et al., 2007). Mycoplasmas have been isolated from the conjunctiva of up to 20% of healthy cats (Blackmore et al., 1971; Tan and Miles, 1972; Tan et al., 1977a; Haesebrouck et al., 1991; Low et al., 2007). The association of mycoplasma with conjunctivitis has been demonstrated in various studies (Tan and Miles, 1972, 1973, 1974b; Tan et al., 1977a; Haesebrouck et al., 1991; Low et al., 2007), with greater frequency of isolation from cats with conjunctivitis than without, but evidence for the pathogenicity or virulence of mycoplasmas in producing such disease has been conflicting. Two studies have demonstrated the development of conjunctivitis following experimental transmission of mycoplasmas (Tan, 1974; Haesebrouck et al., 1991), whereas multiple other studies failed consistently to do so (Cello, 1957; Cole et al., 1967; Blackmore and Hill, 1973; Tan et al., 1977b; Lappin et al., 2007). The conflicting results may reflect differences in pathogenicity of mycoplasma species or particular isolates, or in susceptibility to mycoplasma associated with age or other characteristics of the cats. The involvement of other infectious agents is difficult to determine, as not all of the above studies were performed in specific pathogen free (SPF) cats. As with upper respiratory tract disease, the involvement of mycoplasma with FHV or Chlamydophila felis in conjunctivitis or ocular disease must be considered. Low et al. (2007) demonstrated that cats with active conjunctivitis were significantly more likely to have mycoplasma than those without, and that multiple agent infections were uncommon. Thus, the evidence supports an involvement of mycoplasmas with conjunctivitis in cats, but the precise circumstances under which disease occurs and the specific characteristics of the organisms involved have not been determined. 19

35 The lower respiratory tract is substantially different from the URT as it is relatively free of microbes (Tizard, 1996; Lopez, 2007). The host defences in this region reflect this, being more sophisticated than in the upper respiratory tract. In addition to mucous secretions, lysozymes and IgA, the epithelium is ciliated. The latter may reduce attachment sites for microorganisms, provide a filter for particulate matter and a means to remove such matter via the mucociliary elevator. The lower respiratory tract also has a greater array of immunoglobulins (IgA, IgE and IgG), lymphoid nodules (in the form of bronchus-associated lymphoid tissue [BALT]) and resident immune cells such as the alveolar macrophages (Tizard, 1996; Cohn and Reinero, 2007; Lopez, 2007). Despite the host defences of the LRT, microorganisms have been isolated from this location in healthy cats. These microorganisms include Pasteurella, Pseudomonas, Staphylococcus, Streptococcus, Escherichia coli, Bordetella bronchiseptica and Micrococcus spp. (Padrid et al., 1991; Dye et al., 1996). Finding these organisms in the LRT raises questions as to whether they reside here as non-pathogenic commensals, or are simply an example of the transient introduction of oropharyngeal and nasal flora with inspired air that are continually removed by intact host defences (Lopez, 2007). Mycoplasmas are not considered normal inhabitants of the lower respiratory tract based on the comparative lack of isolation from this location in healthy cats (Tan et al., 1977a). This is highlighted when compared with the presence of mycoplasmas in cats with lower respiratory tract disease (Tan, 1974; Tan and Miles, 1974b; Randolph et al., 1993). In contrast, the isolation of M. gateae from the trachea of 29% of healthy cats is considered an important finding (Heyward et al., 1969). The lack of verification of the presence of mycoplasmas in the lower respiratory tract may be due to the infrequent sampling of this site in healthy cats or the sampling techniques (due to false negative results from liberation of tissue mycoplasmacidal factors). Alternatively, it may be that mycoplasmas are not normal inhabitants of this part of the respiratory tract. In individual studies where mycoplasmas have been isolated from clinical cases of lower respiratory tract disease, underlying causes or predisposing factors were 20

36 consistent features (Crisp et al., 1987; Malik et al., 1991; Foster et al., 1998; Gulbahar and Gurturk, 2002; Trow et al., 2008). Retrospective studies have demonstrated that the prevalence of mycoplasma in the lower respiratory tract of cats with respiratory disease is approximately 16 to 60% (Stein and Lappin, 2000; Foster et al., 2004b; Foster et al., 2004c), and 3.7% when associated with pyothorax in cats (Barrs et al., 2005). From these studies, it appears that mycoplasmas may be under-represented as significant organisms. This may be from lack of routine culture for mycoplasmas using specific media, and also from empirical antibiotic treatment putatively killing these organisms or stopping their growth. Despite these factors, Foster et al. (2004b) showed that mycoplasma were the most commonly isolated organisms in feline lower respiratory tract infections. These retrospective studies give some indication of the prevalence of mycoplasmas to be expected in lower respiratory tract disease in cats. Moreover, they describe findings which may support the involvement of mycoplasmas in this disease, such as positive culture and the resolution of disease following antimicrobial treatment (Chandler and Lappin, 2002; Foster et al., 2004b). However, there is still question regarding the role of mycoplasma in inducing disease. Other potential infectious pathogens, such as viruses were not excluded, and perhaps the frequent isolation of mycoplasma from these diseased animals was simply a reflection of the organism being present in this location as a commensal. They may, however, represent opportunistic pathogens under particular predisposing conditions. Hence the significance of the presence of mycoplasma species in the lower respiratory tract is far from resolved, and as such, further investigations are indicated Prevalence - with reference to different mycoplasma species Prevalence studies of different mycoplasma species in cats are limited. Many studies have either not identified isolates to the species level, or have made assumptions without definitively determining the species involved. Based on relatively comprehensive studies (Tan and Miles, 1974b; Tan et al., 1977a), the trend of species prevalence is: M. gateae > M. felis > M. arginini. This trend was reported for both diseased and healthy cats in these studies, and was also similar in an earlier study of healthy cats (Heyward et al., 1969). This pattern differed when considering 21

37 conjunctival samples of sick cats, whereby the prevalence of M. felis was greater than either M. gateae or M. arginini (Tan and Miles, 1974b). M. felis was found to be slightly more prevalent than M. gateae in both the conjunctiva and throat in another large study (Blackmore et al., 1971). These differences are not substantial, and may reflect differences in individual populations or geographical locations, and give an indication as to the most common feline mycoplasma species. Tan et al. (1977a) found that some swabs yielded more than one type of mycoplasma, however the combinations of species and from which anatomic sites they were recovered were not documented. An earlier study demonstrated that M. felis and M. gateae were frequently found concurrently in the pharynx of cats (Blackmore et al., 1971). In this study, 77 of 144 (53.5%) cats from one colony and 10 of 26 (38.5%) of cats from another colony (for which mycoplasma was demonstrated in the oropharynx) had both M. felis and M. gateae concurrently Possible roles of individual mycoplasma species in cats The overall trend for prevalence of different mycoplasma species was similar in a range of anatomic locations of healthy and diseased cats, while the relative prevalence of different species between diseased and healthy cats has been shown to differ (Table 1.3)(Tan and Miles, 1974b). This indicates that the different species may have varying roles or involvement in their hosts, particularly with respect to pathogenicity. Each species of mycoplasma in cats will be reviewed, considering what is currently understood about the host-mycoplasma relationship and pathogenesis. Frequent isolation of M. gateae from the URT tract of cats (Heyward et al., 1969; Blackmore et al., 1971; Tan et al., 1977a), combined with the lack of evidence of an association with disease (Tan and Miles, 1974b) or immunologic response to the organism (Blackmore and Hill, 1973; Tan and Miles, 1974b; Tan et al., 1977a), suggests it is a non-pathogenic, commensal organism of the feline upper respiratory tract. M. gateae has also been isolated from both the conjunctiva and lower respiratory tract of healthy cats (Heyward et al., 1969; Blackmore et al., 1971; Tan et al., 1977a). In the rare instances where M. gateae has been isolated from cases of feline arthritis, there has been evidence of concurrent immunosuppression suggesting 22

38 predisposition to infection (Moise et al., 1983; Crissman, 1986; Zeugswetter et al., 2007). The scant evidence available for M. gateae in cats suggests it does not independently initiate disease (Blackmore and Hill, 1973), apart from an isolate which displayed tropism to joints when inoculated intravenously (Moise et al., 1983). This latter finding is not remarkable given mycoplasmas of other species frequently localise in joints following dissemination via the blood-stream (Cole and Ward, 1979; Cole et al., 1985). Several studies suggest M. arginini is a non-pathogenic commensal organism of the conjunctiva and upper respiratory tract due to the frequent isolation from this location in healthy cats (Tan et al., 1977a; Tan et al., 1977b). In addition, the relative lack of a host immune response (Blackmore and Hill, 1973; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b) and an inability to induce disease following experimental infection (Blackmore and Hill, 1973; Tan et al., 1977b) supports the role as a commensal. M. arginini has a widespread distribution in a range of species (Razin and Freundt, 1984) and is a common contaminant of cell cultures (McGarrity and Kotani, 1985; McGarrity et al., 1992). However, the pathogenicity of M. arginini is unknown (Razin and Freundt, 1984). As is true for M. gateae, there have been rare isolations of M. arginini from cats with disease, but no evidence of them causing or playing a role in the disease (Tan and Miles, 1974b; Tan et al., 1977b), which is consistent with the ecological and biological similarities of these two species. In contrast to M. arginini and M. gateae, M. felis has been implicated to varying degrees in both feline conjunctivitis and respiratory disease (Cole et al., 1967; Campbell et al., 1973b; Tan, 1974; Haesebrouck et al., 1991; Randolph et al., 1993). Despite inconsistent findings regarding the expression of disease in experimental transmission studies (Cello, 1957; Cole et al., 1967; Blackmore and Hill, 1973; Tan, 1974; Tan et al., 1977b; Lappin et al., 2007), more frequent isolation of M. felis from cats with conjunctivitis than from unaffected cats has led to acceptance of this species as a causative agent (Tan and Miles, 1973; Tan et al., 1977a; Haesebrouck et al., 1991; Low et al., 2007). Serological studies have markedly differed for M. felis compared with M. gateae and M. arginini, with M. felis antibodies demonstrated in a relatively high proportion of 23

39 Table 1.3: Comparative percentage of mycoplasma isolates from sick and healthy cats found to be infected with each of three species of mycoplasma (adapted from Tan and Miles, 1974)* M. felis M. gateae M. arginini Sick cats 29.7% 46.2% 16.5% (total isolates 212) Dead/sacrificed 38.0% 43.1% 9.2% cats (total isolates 195) Healthy cats (total isolates 45) 4.4% 62.2% 6.7% *The majority of sick cats from this study were noted to have respiratory infections, but also included conjunctivitis, tonsillitis, glossitis, urolithiasis and other forms of urogenital tract infections, inappetence, enteritis, dysentery and various forms of food poisoning.

40 healthy and diseased cats (Tan and Miles, 1974b; Tan et al., 1977a). The demonstration of antibodies indicates exposure to and immunological recognition of an organism (Goldsby et al., 2003). The presence of antibodies does not, however, indicate the presence of disease as a result of this exposure. An example of this can be demonstrated by the findings of an experimental transmission study (Tan, 1974). Tan (1974) demonstrated the production of antibodies to M. felis in kittens that developed conjunctivitis following experimental inoculation with this species, and also in control kittens exposed to M. felis by close proximity to the inoculated kittens. Post exposure, the control kittens developed M. felis antibodies, but had no clinical signs of disease. The cats were not SPF and, in spite of the fact the kittens were negative for unspecified viral and bacterial agents, the involvement of other organisms cannot be excluded with certainty. This study provides evidence for the pathogenicity of M. felis, however it was not determined that this species of mycoplasma is sufficiently virulent under natural conditions of transmission to cause clinical disease (compared with the experimental introduction of a concentrated isolate). In this same study (Tan, 1974), M. felis was isolated from the lower respiratory tract of two kittens with post mortem evidence of interstitial pneumonia following experimental introduction of M. felis into the eyes and nostrils. However, the presence of M. felis in the lower respiratory tract was not tested for at the beginning of the study. It is possible that the pneumonia observed was induced by the introduction of M. felis into these kittens, but these results are equivocal. The importance of this finding is limited by the confounding factors of a concentrated inoculum and non-spf cats described previously. M. felis has been isolated from the LRT of a small number of diseased cats. A large study determined the incidence and significance of mycoplasmas in sick cats, isolating M. felis from the tracheas of 3 cats and lungs of 4 cats post mortem (Tan and Miles, 1974b). This study did not state from how many cats LRT samples were taken, limiting the significance of these findings. In addition, respiratory infections were noted to constitute the majority of disease observed in the population studied, however no correlation was made with individual animals from which mycoplasmas were isolated from the LRT. No direct evidence was presented for the proposed role 24

41 of mycoplasma causing pneumonia in cats, which raises questions about the validity of their conclusions. There have been no reports of respiratory disease being reproduced experimentally by inoculating cats with mycoplasmas isolated from feline cases of respiratory disease (reviewed in (Bemis, 1992)). Also, there has not been any description of mycoplasmas being contagious in cases of suspected natural infection. Similarly, there have been no reports of outbreaks of mycoplasma related respiratory disease in animal shelters where its presence has been demonstrated (Heyward et al., 1969; Blackmore et al., 1971; Bannasch and Foley, 2005; Veir et al., 2008). Extrapolating from mycoplasmal respiratory diseases in other animal species, an occurrence of Mycoplasmosis might be expected if M. felis were sufficiently virulent as a primary pathogen. However, whether M. felis plays a synergistic, secondary or opportunistic role in respiratory disease requires further investigation. Rare occurrences of arthritis linked to M. felis have been demonstrated (Hooper et al., 1985; Liehmann et al., 2006). These cases had underlying trauma or immunosuppression, similar to cases involving M. gateae (Moise et al., 1983; Crissman, 1986; Zeugswetter et al., 2007). This information reinforces the concept that in an immunocompetent feline host, mycoplasma species do not usually gain systemic entry. It has been suggested that this resistance is marked in cats compared with other species such as pigs and poultry, which may be a reflection of the consequences of intensive food animal production systems (Pedersen, 1998). It could however be related to feline mycoplasma species being less virulent. Despite the demonstrated pathogenic capacity of M. felis under certain conditions, there is sufficient doubt as to its role as a primary pathogen. The frequent isolation of M. felis from healthy cats and the inability to consistently initiate disease in many experimental transmission studies or following natural transmission would indicate that it is not sufficiently virulent to initiate disease alone. As the antigenic or virulent features of M. felis have not been defined, mechanisms for pathogenicity have not been determined. It is possible there are different variants of M. felis accounting for the apparent associations observed, or particular conditions are required for it to penetrate the mucosa. Factors influencing mucosal host defences, or presence of 25

42 inflammation or other microorganisms allowing mycoplasmas to gain host entry may be important. Further investigation is required to improve the understanding of M. felis characteristics to conclusively establish its role as a putative pathogen in cats. It would appear the isolation of M. feliminutum is a rare occurrence, being recovered in only one instance from a healthy cat (Heyward et al., 1969). One might question whether this is a mycoplasma species of cats, or whether the lack of other evidence relates to it having been overlooked or out-competed by other mycoplasmas in culture, due to its slower growth or differing metabolic requirements which have been previously suggested (Heyward et al., 1969). Reports of pneumonia, pyothorax, pulmonary and subcutaneous abscesses in cats from which mycoplasmas have been isolated have often not confirmed identification to the species level (Crisp et al., 1987; Malik et al., 1991; Foster et al., 1998; Gulbahar and Gurturk, 2002; Adamson, 2004; Cross, 2004; Trow et al., 2008). The significance of these isolations is questionable, given that mycoplasmas are frequently present in many healthy cats. Association with these diseases, even if via direct culture from the lesion, is not sufficient evidence for the organism being responsible for or involved in the disease Age of cats A large study of respiratory pathogen epidemiology from Californian shelter cats found age to be a significant risk factor for mycoplasma infection in the upper respiratory tract (Bannasch and Foley, 2005). There was an increased risk in age groups of 0 to 3 months and 7 to 11 months, and a decreased risk for cats over 12 months. These results were not divided into findings for the individual organisms identified, such that it is unknown whether this risk relates to mycoplasma alone. A different study showed upper respiratory disease was most common in young cats from 6 weeks to 8 months of age (Schneck, 1972). Mycoplasmas were isolated from 63% of ocular and nasal swabs of kittens with respiratory disease. The comparative prevalence of mycoplasma in healthy kittens was not included. The presence of virus 26

43 in the kittens was not described, even though the authors commented that viral respiratory disease was common leaving the significance of this study equivocal. The incidence of mycoplasma increased with age in another population of cats until approximately 6 months of age, with the highest incidence occurring in cats above 6-12 weeks of age (Blackmore et al., 1971). The prevalence of mycoplasma remained at consistently high levels beyond this age. This was suggested to be due to the microorganisms being transmitted primarily by contact with other cats at a young age. Evidence of transmission by contact has been demonstrated (Tan, 1974; Lappin et al., 2007). Another route of transmission may be vertically from the queen s vagina during birth; mycoplasmas have been isolated from this location (Heyward et al., 1969; Blackmore et al., 1971; Tan and Miles, 1972, 1974b; Tan et al., 1977a; Tan et al., 1977b). Under normal circumstances, kittens are in a germ-free environment in utero, and are subsequently exposed to a wide variety of organisms after parturition. Microbial flora is established in the first few weeks of life on the mucous membranes and in the gastrointestinal tract. Organisms that have colonised the mucosal surfaces then compete for life against other microorganisms and developing mucosal immunity (reviewed in (Baskerville, 1981; Kononen, 2000)). Commensal organisms that survive on the surface, such as mycoplasmas, usually avoid the host s immune response and form part of the normal microbial flora ((reviewed in (Heyward et al., 1969; Blackmore et al., 1971; Tan and Miles, 1972; Baskerville, 1981; Kononen, 2000)). Upper respiratory tract disease is more prevalent in young cats (Binns et al., 2000; Bannasch and Foley, 2005). This may be related to immaturity of the immune system in the first few months of life (reviewed in (Day, 2007; Morein et al., 2007)). In the period of waning maternal immunity, the development of the immune system is still occurring, which renders kittens more vulnerable to a range of diseases during this time compared with adult cats (reviewed in (Day, 2007; Morein et al., 2007)). Additionally, the resident microfloral populations of adult cats provide some protection by competing for attachment sites and nutrients with pathogenic organisms (reviewed in (Baskerville, 1981; Kononen, 2000)). However, the resident microflora may not be well established at this stage of the development of the immune system. 27

44 Geographical and climatic factors Geographical location and climate have not been considered specifically in any studies of mycoplasma in cats. Surveys and case reports in the literature cover areas in North America (Cello, 1957; Cole et al., 1967; Heyward et al., 1969; Campbell et al., 1973a; Campbell et al., 1973b; Shewen et al., 1980; Keane, 1983; Moise et al., 1983; Crisp et al., 1987; Brown et al., 1990; Randolph et al., 1993; Walker et al., 1995; Chandler and Lappin, 2002; Bannasch and Foley, 2005; Johnson et al., 2005; Klose et al., 2007; Low et al., 2007; Trow et al., 2008; Veir et al., 2008), England (Blackmore et al., 1971; Hill, 1971; Schneck, 1972; Schneck, 1973), Belgium (Haesebrouck et al., 1991), Switzerland (Bart et al., 2000), Germany (Huebner et al., 2007; Hartmann et al., 2008), Sweden (Sjödahl-Essén et al., 2008), Singapore (Tan et al., 1977a; Tan et al., 1977b), New Zealand (Tan and Markham, 1971b; Tan and Miles, 1972, 1973, 1974b) and Australia (Brisbane; (Spradbrow et al., 1970b), Sydney; (Malik et al., 1991; Foster et al., 1998; Foster et al., 2004a; Foster et al., 2004b; Foster et al., 2004c; Barrs et al., 2005) and Melbourne; (Jones and Sabine, 1970; Hooper et al., 1985)). Study comparisons are difficult because of different methods of investigation and presentation of results (as discussed previously under section ). Assessment of prevalence among these studies found no pattern to the overall geographical distribution, anatomical location or any differences in the organism characteristics. It can be concluded that feline mycoplasmas have widespread geographical distribution. One study considered the effect of season as part of a retrospective study of LRT infection in cats (Foster et al., 2004b). Of those considered to be mycoplasmal infections, clinical signs commenced in autumn in 4 of the 8 cats for which the time of onset was known. Additionally, in some cases there was recurrence of clinical signs when the weather became cold. Although mycoplasmal LRTI may be influenced by cold weather, greater numbers of cats need to be studied to determine whether this is a significant and widespread trend. 28

45 Prevalence of mycoplasmas compared with other respiratory pathogens In a survey of shelter cats, the prevalence of multiple respiratory pathogens was considered (Bannasch and Foley, 2005). Mycoplasma was found to be the third most prevalent group of organisms, after FCV and FHV (Bannasch and Foley, 2005). The prevalence of mycoplasma may have been underestimated, as only the conjunctiva was sampled for mycoplasma, whereas the viral samples were obtained from both the conjunctiva and oropharynx. From a group of 68 cats with clinical signs of cat flu, M. felis has been identified to be the most prevalent organism from the mouth and conjunctiva (Huebner et al., 2007). In this study, PCR assays were used to investigate the prevalence of both viral and bacterial agents typically associated with cat flu. A retrospective study examined records from 245 cats diagnosed with pneumonia or conjunctivitis/rhinitis. Bacterial infections constituted 33.5% of cases, viral infections constituted 18.4% of cases, and viral with secondary bacterial infections was suspected in 8.2% of cases (Bart et al., 2000). This study was not able to verify the specific aetiology of each of these cases. Differing diagnostic tests and results in some cases meant the specific agent was neither confirmed nor clear whether multiple microorganisms were isolated from individual cases. Cultures from lung samples taken post mortem of 40 cats with bacterial infection revealed that mycoplasma was the third most commonly isolated group of organisms (after Bordetella bronchiseptica and Pasteurella spp.) (Bart et al., 2000). A retrospective study of 21 cases of LRT infection in Australia indicated that mycoplasma was the most common cause of LRT infection (Foster et al., 2004b). This finding was based primarily on culture of BAL samples, but infectious agents cultured were only considered the aetiologic agents if historical, clinical, radiographic and cytologic findings were in agreement, combined with an unambiguous positive response to appropriate antimicrobial therapy. Cases included in the study were considered to be not to be viral, based on these criteria, but the presence of viral agents was not specifically excluded, and hence their involvement cannot be excluded with certainty. 29

46 In conclusion, mycoplasmas are one of the most common groups of organisms isolated from cats with upper and lower respiratory tract disease; however, their role in such diseases is uncertain, particularly given they are also commonly isolated from non-diseased cats. For many cases of respiratory disease in cats, no aetiological agent or combination of agents has been identified unequivocally. This aspect is in part due to the invasiveness in sampling the lower respiratory tract, but also as viral detection techniques are often not employed. The myriad of possible causes may require extensive and hence expensive study to investigate viral, bacterial, fungal, parasitic, inflammatory and neoplastic causes. These include imaging, sampling for cytology, microbiology and PCR. These tests are not always clinically indicated (or affordable to the client) in every case. Empirical antimicrobial therapy is therefore sometimes employed in such cases where bacterial infection is likely, as a tool to evaluate response to treatment. This may be appropriate in individual cases, but it limits accumulation of knowledge of individual organisms involved on a population basis Relationship of mycoplasmas to other respiratory pathogens Low et al. (2007) demonstrated that it was uncommon to find multiple infections in a population of 55 cats with conjunctivitis. Only 2 cats had M. felis and Chlamydophila felis concurrently, and one cat harboured M. felis and FHV concurrently. Similarly, another study demonstrated the concurrent presence of M. felis and FCV in 8 samples from the oral cavity and 3 from the conjunctiva from 68 cats with cat flu (Huebner et al., 2007). Earlier studies examined a population of cats with respiratory disease for the presence of both viruses and mycoplasmas (Spradbrow et al., 1970a; Spradbrow et al., 1970b). The findings for the viral and mycoplasmal isolations from this population were published separately, and the results were not correlated within individual animals and hence conclusions cannot be made as to the proportion that had concurrent infections. The incidence of mycoplasma was found to be higher in a population of healthy adult cats from a closed feline colony where viral respiratory infections were prevalent, than in healthy adult domestic cats (Schneck, 1972; Schneck, 1973). These findings 30

47 support, to some extent, a relationship between mycoplasma and virus infections known to be involved in the respiratory disease complex. However, they could also reflect other differences (e.g. environmental factors) between colony and household cats. These reports only considered 12 cats in each group, with 5 (42%) of the colony cats, and 1 (8.3%) of the household cats testing positive for mycoplasma by culture (Schneck, 1972; Schneck, 1973). It was not reported how many of the cats sampled had recent viral infections or were potential viral carriers. The relative prevalence of mycoplasma between diseased cats from the colony or from households was not considered which leaves the findings inconclusive. Although mycoplasma have not been proven to be associated with other respiratory pathogens it is possible that they are. Several authors speculated that mycoplasmas may act either as synergistic or opportunistic invaders following the damage of the respiratory epithelium by viruses (Heyward et al., 1969; Schneck, 1972; Campbell et al., 1973b; Pedersen, 1998; Dawson and Willoughby, 1999; Whitley, 2000). This speculation is partly based on the demonstration of similar associations in other animal species infected with mycoplasma. There are multiple examples where the severity of respiratory or ocular disease is worsened when other bacteria or viruses are involved, compared with the disease produced by the mycoplasma alone (Simecka et al., 1992). Even with scant evidence for mycoplasmas causing a defined disease condition as a result of exposure, the association of mycoplasmas with ocular and respiratory diseases in cats may be suggestive of a more complex aetiology. Relationships with other organisms may explain the discrepancies between experimental transmission studies conducted in SPF compared with those in non-spf cats, and would be a logical conclusion, given the multifactorial aetiologic nature of these diseases Association of mycoplasma with feline bronchial disease (FBD) Association of mycoplasma with feline bronchial disease (FBD), an inflammatory airway disease with no identifiable aetiology, has been suggested based on findings from retrospective studies of feline LRT infection and FBD (Foster et al., 2004a; Foster et al., 2004b; Foster et al., 2004c). In these studies, mycoplasmas were not cultured from any cases with pulmonary lesions other than FBD. Additionally, LRT 31

48 infections identified in a subset of the 25 cats diagnosed with FBD (purebred shorthair cats, excluding Siamese and Burmese) were all due to mycoplasmas. This subset of cats was more likely than domestic cats to have both FBD and LRT infections. These findings however, do not provide evidence for a definite link between FBD and mycoplasma. There is growing evidence from the human literature for links between mycoplasmal infection and asthma in humans ((Sabato et al., 1984; Seggev et al., 1986; Teo et al., 1986; Petrovsky, 1990; Gil et al., 1993; Kondo et al., 1994; Yano et al., 1994; Seggev et al., 1996; Kraft et al., 1998; Micillo et al., 2000) cited in (Foster et al., 2004b; Foster et al., 2004c)), which is outside the scope of this review. These authors suggested that based on their findings and similarly to the human literature cited in the article FBD may predispose to mycoplasmal LRT infections, although it is equally possible that mycoplasmal LRT infections may cause airway hyper-reactivity and induce FBD (Foster et al., 2004b; Foster et al., 2004c). Although this aspect certainly warrants further investigation, there is currently no evidence for an association in cats Presence of feline mycoplasmas in other species There are two reports of human infections with mycoplasma which were thought to have been transmitted from a cat (McCabe et al., 1987; Bonilla et al., 1997). The first occurred in a veterinarian scratched by a cat on the finger and subsequently developed a septic tenosynovitis from which mycoplasma was cultured (McCabe et al., 1987). The second relates to M. felis associated arthritis in an immunocompromised individual who had been bitten by a cat (Bonilla et al., 1997). The protracted nature of the arthritic disease in this individual, and the difficulty of antimicrobial treatment by bacteriostatic drugs in an immunocompromised state highlight the importance of the immune system in clearing the organisms. The feline zoonoses guidelines from the American Association of Feline Practitioners (Tuzio et al., 2005) include M. felis as a rare zoonosis, and cite the above 2 cases (McCabe et al., 1987; Bonilla et al., 1997). Methods to minimise the risks of cat bites or scratches are discussed, and in their event, the authors advise that medical advice be sought. The role of 32

49 immunosuppression in zoonoses is also raised, as it poses an increased risk for a range of zoonotic diseases. M. felis has also been isolated from horses in association with lower respiratory tract disease (Carman et al., 1997; Wood et al., 1997) and pleuritis (Ogilvie et al., 1983). This is the only mycoplasma species in horses that has been implicated as having a causal relationship with disease (Whitford and Lingsweiler, 1994). The significance of M. felis in these cases comes from demonstration of sero-conversion in paired sera (Ogilvie et al., 1983; Rosendal et al., 1986; Carman et al., 1997; Wood et al., 1997) and from the experimental induction of pleuritis when M. felis was introduced into healthy horses (Ogilvie et al., 1983; Rosendal et al., 1986). This was done by inoculation into the thoracic cavity which is not the natural route of transmission. These studies fail to conclusively determine a primary pathogenic role or virulence of M. felis in horses by not definitively excluding the involvement of other organisms. Mycoplasma infection in horses is unlikely to be a geographically widespread problem. Infrequent localised occurrences may occur due to unrecognised circumstances of horses coming in contact with the organism. Three of these four equine studies were from Ontario, Canada (Ogilvie et al., 1983; Rosendal et al., 1986; Carman et al., 1997), and the other from the United Kingdom (Wood et al., 1997). Another possibility is that mycoplasma has not been considered during investigation of pleural or lower respiratory tract disease in horses, with specific culture media not being employed to culture the organisms. How these horses came into contact with M. felis is unknown; whether this is an example of the organism acting in a more virulent way in an atypical or accidental species, or whether the horse is a true host of M. felis remains to be determined. The presence of M. felis in the healthy equine population has not been reported; hence, it is possible that it is present as a commensal organism in the wider population. The discovery of M. felis in horses has prompted some researchers (Brown et al., 1995) to question whether the organism isolated from these equine cases was the feline variant of M. felis or not. The question of identity arose due to the observation that pharyngeal swabs from cats identified as M. felis by an immunobinding assay were not confirmed by 16S rrna gene sequence data (Brown et al., 1990; Brown et 33

50 al., 1995). M. felis was identified by serological methods in each of the equine studies. Nucleotide sequence analysis of these isolates would provide insights into the genetic relationship of these organisms with the feline isolates. There have been a number of reports of mycoplasma isolation from captive felids, including lions (Pantherae leo), cheetahs (Acinonyx jubatus), pumas (Felis concolor) a leopard (Panthera pardus), a tiger (Panthera tigris), a lynx (Felis lynx) and a serval (Felis serval) (Hill, 1975; Hill, 1992; Brown et al., 1995; Johnsrude et al., 1996). Although mycoplasmas have been isolated from the same anatomical locations as domestic cats, some different mycoplasma species were present in these hosts in addition to M. felis, M. gateae and M. arginini (Hill, 1975; Hill, 1992). It would appear from these studies that mycoplasmas are present as normal flora in the upper respiratory tract, as they are in domestic cats. Exceptions related to the isolation of M. felis from a juvenile serval with severe pneumonia. This animal had been in contact with domestic cats as it was being hand-reared, and could have been predisposed to such an infection as a result of not receiving colostrum, and from a previous treatment with amoxicillin (Johnsrude et al., 1996). In another study, M. arginini was recovered at post mortem from the lungs and brain of a lion with encephalitis (Heyward et al., 1969; Tully et al., 1972) Consideration of factors that may facilitate mycoplasmal involvement in disease Even with extensive research into mycoplasmas, many specific factors of their involvement in disease remain obscure. This is true even for widely studied mycoplasma species such as M. hyopneumoniae (Pieters et al., 2009). The interaction between host, organism and environment is an integral part of any infectious disease, but the multifactorial nature of the role that mycoplasmas have in disease has proven to be challenging for investigators. Specific factors relating to the organisms which determine the relationship of mycoplasmas to their hosts and pathogenicity have not been characterised for feline mycoplasma species. Studies from other ycoplasmas have shown that attachment is an integral part of pathogenicity (reviewed in (Cassell et al., 1985; Simecka et al., 34

51 1992; Rottem, 2003)). Specialised attachment structures and specific binding to receptor sites on host cells have been demonstrated for some species of mycoplasma, such as M. gallisepticum and M. pneumoniae. These have been reviewed elsewhere ((Simecka et al., 1992; Rottem, 2003)) and are outside the scope of this thesis. In vitro studies of mycoplasmas associated with respiratory disease, such as M. hyopneumoniae have demonstrated their ability to adhere to ciliated epithelial cells, affect ciliary motility, and occasionally result in loss of cilia and epithelial necrosis (Simecka et al., 1992; DeBey and Ross, 1994). Mycoplasmas have also been shown to directly induce cell injury via production of toxic substances, and indirectly by inciting host inflammatory responses which may be responsible for the resultant tissue damage (Simecka et al., 1992). Mycoplasmas may modulate the host immune system in a variety of ways. Induction of host cell cytokine production is thought to be a major virulence mechanism of mycoplasmas, which can result in stimulation or down-regulation of different inflammatory pathways (reviewed in (Rottem, 2003)). Lipid associated membrane proteins (LAMPs) of mycoplasmas have been shown to induce signalling pathways via pathogen associated molecular pattern (PAMP) recognition receptors of the host innate immune system (reviewed in (Rottem, 2003; You et al., 2006)). Interaction of LAMPs with toll-like receptors TLR2 and TLR6 result in signalling cascades leading to stimulation of pro-inflammatory cytokine secretion and to apoptosis (reviewed in (You et al., 2006)). There has been great interest in determining and characterising surface components of mycoplasmas which may act as both antigenic and virulence factors. This will aid further characterisation of their pathogenic mechanisms and interaction with host cells. Some species of mycoplasma such as M. pulmonis and M. hyorhinis have the ability to change their phenotype, particularly with surface antigens (reviewed in (Simecka et al., 1992; Rottem, 2003)). This may influence the interaction of mycoplasmas with host cells, altering their ability to attach and to prevent immunological recognition by the host. Additionally, some surface mycoplasmas such as M. pneumoniae and M. gallisepticum have been shown to reside in non-phagocytic host cells, where they are partially protected from the host immune system (reviewed in (Rottem, 2003)). These factors may explain some variability of virulence observed 35

52 within particular species of mycoplasma (Simecka et al., 1992). The elucidation of immunogenic components has been utilised to develop vaccines aimed at reduction of production losses associated with mycoplasmas in some species, such as M. hyopneumoniae in pigs, M. mycoides subsp. mycoides in cattle, and M. gallisepticum and M. synoviae in poultry (reviewed in (Whithear, 1996; Thiaucourt et al., 2003; Maes et al., 2008; Nicholas et al., 2009)). Host factors are equally important in determining the result of exposure to mycoplasmas. The mechanisms of the immune response to mycoplasmas are not fully understood. Innate immunity is thought to provide the primary defence in the respiratory system, with the production of an acquired antibody response being particularly important for protection against systemic infections (Cartner et al., 1998). Both cell mediated and humoral immunity are required for maximal efficiency in eliminating mycoplasmas (Fernald, 1979; Cartner et al., 1998). The T lymphocyte response is important in the outcome of mycoplasmal respiratory infections, as the specific type of T cell response to mycoplasma infection influences the balance between host resistance and immunopathogenesis, particularly in chronic mycoplasmal disease (Jones and Simecka, 2003). The feline immune response to mycoplasmas has not been studied beyond the demonstration of antibody production in healthy and diseased cats (Tan, 1974; Tan and Miles, 1974b; Tan et al., 1977a). It is not known whether such responses are effective against current infections, capable of eliminating the organisms, or protective against subsequent challenge. However it would appear that if virulent, the continued presence of M. felis organisms in the host after resolution of disease could be defined as remaining in an asymptomatic carrier state. This has been demonstrated for M. hyopneumoniae pneumonia in pigs, where convalescent carriers were able to infect susceptible animals for up to 200 days post infection (Pieters et al., 2009). Irrespective of exact immune mechanisms involved, a functional immune system is required to eliminate mycoplasma infection. A competent immune response is important in both eliminating organisms and acting in the prevention of disease. This principle is highlighted by frequent reports of cats with immune compromise or other 36

53 underlying diseases and concurrent infection with mycoplasmas of little primary pathogenic significance (Cello, 1957; Moise et al., 1983; Hooper et al., 1985; Crissman, 1986; Malik et al., 1991; Pedersen, 1998; Zeugswetter et al., 2007). Environmental factors are known to be important in the outcome of exposure to mycoplasma in intensive production situations. The evidence and mechanisms of involvement in different animal species has been reviewed elsewhere ((Simecka et al., 1992)) and is outside the scope of this review. Environmental factors, such as air temperature and quality, essentially have their effect by altering host respiratory defence mechanisms (Jones and Simecka, 2003). Whether these factors are important in cats in shelter environments (which may be likened to intensive production systems) has not been determined, although risk factors for disease in such systems have been studied to some extent (Bannasch and Foley, 2005). Finally, other pathogens may play a crucial part in diseases involving mycoplasmas in cats. Although concurrent isolations have not often been demonstrated in diseased cats, or specific links or associations with particular pathogens have not been made, such interactions cannot be excluded. An additional factor that has not been considered and is outside the scope of this review is the vaccination of cats for respiratory pathogens. The prevention of viral respiratory disease may influence the prevalence and role of mycoplasmas in different populations. 1.4 Conclusion The current literature review has shown that mycoplasmas are common inhabitants of the upper respiratory tract of healthy domestic cats from widespread geographical locations. However, there is uncertainty regarding the prevalence of these organisms in the lower respiratory tract of cats. The infrequent isolation of mycoplasmas from this location may relate to previous sampling techniques, a relative lack of sampling of this location, or because mycoplasmas are not normal inhabitants of this region. Some non-pathogenic species of mycoplasma such as M. gateae and M. arginini appear to be commensal microorganisms of the conjunctiva and upper respiratory tract of cats. Their presence does not evoke a host immune response. These non- 37

54 pathogenic mycoplasmas may be involved with disease under certain conditions. Although these conditions are not yet defined, a compromised host defence is a factor, particularly if it allows mycoplasmas to penetrate the mucosa and become disseminated systemically, in which case, it seems likely that the mycoplasmas may preferentially localise to joints. M. felis, on the other hand, may be involved in disease of the conjunctiva and respiratory tract in cats. There is evidence to suggest that M. felis has some degree of virulence and is recognised by the immune system, which may be indicative of it being more invasive or antigenic in some way than the other mycoplasma species. M. felis is also a normal inhabitant of healthy cats. Even with its association with disease, M. felis does not consistently produce disease under experimental conditions, which may indicate that it is more of an opportunistic invader than a primary pathogen. No specific disease or syndrome has yet been associated with M. felis solely as the result of exposure of the host to the organism. It would appear, as for many other mycoplasmas, that any such involvement in disease is multifactorial. Very little is understood about the factors involved in mycoplasma-related feline diseases. The genetic characteristics of feline mycoplasmas are poorly defined, and the structural components for attachment, antigenicity and virulence have not yet been characterised. This limits the understanding of host-organism interactions. The understanding of the specific mechanisms by which the host immune system interacts with mycoplasma is limited, apart from a demonstration of seroconversion. Finally, interactions of mycoplasma with other pathogens or environmental factors have not been studied in detail. A limitation of published reports at present is the unresolved significance of the presence of mycoplasma in the lower respiratory tract of cats. This is due to the relatively infrequent identification of mycoplasma to the species level and the past usage of relatively unreliable methods for specific identification. There has been a distinct lack of statistical analysis of the association of mycoplasma with disease. With the development of rapid and sensitive molecular methods of identification, further investigation is warranted in order to define the extent, nature and significance of mycoplasmas in domestic cats. 38

55 1.5 Aims The specific aims of this thesis are to: survey a population of shelter cats for the presence of mycoplasma; examine the population of shelter cats for clinical or post mortem evidence of ocular or respiratory disease; culture mycoplasma from direct swabs obtained at necropsy from the conjunctiva, pharynx and bronchi; determine the overall prevalence of mycoplasma in a population of cats; determine the prevalence of mycoplasma at different anatomical sites; evaluate the association of the presence of mycoplasma with ocular, upper and lower respiratory tract disease; genetically characterise mycoplasma isolates to determine the species present in the population; and assess the association of particular mycoplasma species identified with their anatomical distribution and presence of disease. 39

56 40

57 Chapter 2 Epidemiological Survey of Feline Mycoplasma 2.1 Introduction A major challenge facing researchers who study infectious diseases relates to unequivocal demonstration that organisms isolated from tissue lesions are the aetiological agent. This is most evident in cases of respiratory disease in almost all domestic species, and is particularly the case with ubiquitous organisms such as mycoplasmas. While cause and effect has been shown for some species of mycoplasmas, it remains to be convincingly demonstrated for naturally occurring mycoplasmas in the cat. This aspect is complicated by the fact that experimental infection of susceptible cats with the most pathogenic feline mycoplasma, M. felis, has not always induced disease. Given that cause and effect can be difficult to prove, alternative approaches may rely upon association, but only where an association has been subjected to the rigour of statistical analysis. This chapter investigates the associations between mycoplasmas isolated from the conjunctiva, upper and lower respiratory tract and a number of characteristics of the host, particularly the presence or absence of clinically apparent ocular or respiratory disease. In addition, this chapter focuses on the prevalence of mycoplasma at these sites and any relationships of that prevalence among the sites. 41

58 2.2 Materials and methods Animal Collection Samples were collected over a three-week period in June Cats were obtained from a Melbourne animal shelter and were stray cats delivered to the shelter by a member of the public or picked up by a local council service, or domestic cats surrendered to the shelter by their owner. The cats used in this study were euthanised for being unhandlable, unsuitable to be re-homed due to temperament, age or health, or there not being enough space at the facility. The cats were held for up to one day in a holding room or for eight days in the shelter s housing facility. Prior to euthanasia, cats were clinically examined. Particularly, any signs of conjunctivitis or respiratory disease were noted. Data collection sheets (Appendix 1) were used to record age, breed, sex, weight and rectal temperature. As the actual age of most cats was unknown, age was recorded as either juvenile or adult based on dentition (Wiggs and Lobprise, 1997). Cats were then euthanised using Lethabarb (pentobarbitone 200 mg/ml, Virbac, NSW, Australia) diluted 1:2 in water and injected into the peritoneal cavity. Cats were labelled and transported individually in plastic bags to the Veterinary Clinical Centre of the University of Melbourne where necropsies were conducted. Full necropsies were performed and swabs taken from the conjunctiva, oropharynx and main stem bronchus, were placed immediately into 5 ml of liquid mycoplasma media (Appendix 2). Using sterile scissors the trachea was opened from the dorsal surface. This was performed distal to the pharynx to avoid contamination. The incision was extended as far distally as the carina to expose the mainstem bronchi. While being held open with the scissors, a sterile swab was inserted through the incision (without contacting the sides) into the left bronchus as deep as the diameter of the swab would allow. A representative tissue sample of ~ 1.5 cm 3 was taken from lungs in which macroscopic changes were evident and placed in 10% buffered formalin to preserve the tissue for histopathologic examination if required. 42

59 Table 2.1: Study population characteristics Cats Sex Age Origin Location / Time spent Adult Juvenile Stray Owned Holding room/1 day Pound/ 8 days Number %

60 One hundred and nineteen cats were examined. Samples from the initial 9 of them were disregarded from the study, as collection method was altered, leaving a study population of 110 cats (Table 2.1) Culture The inoculated broths were incubated in air at 37 C for 4 weeks. They were monitored daily for colour change. A change in colour from light pink-orange to red indicated an alkaline shift, as a result of arginine utilisation, whereas a change to yellow indicated an acid shift due to glucose utilisation by the mycoplasmas. Where colour change was noted, broths were frozen and kept at -20 C. Broths that became turbid were assumed to have bacterial contamination and were discarded. Broths with no change after 4 weeks were assumed to have no mycoplasma growth and were discarded. Each of the positive broth samples ( primary broth samples ) were thawed, agitated and plated on to mycoplasma agar plates (Appendix 3). Plates were divided into 5 vertical areas and streaked by allowing 20 µl of broth to run from a pipette down the plate in each section by gravity. The plates were incubated in air at 37 C and checked daily using a dissecting microscope for the presence of characteristic mycoplasma colonies (typified by a fried-egg appearance) (Razin and Freundt, 1984). Five individual colonies from each original sample were selected using a sterile glass pipette tip to lift a single colony from the agar. Each colony was suspended into 2 ml of mycoplasma liquid media. These broths were incubated at 37 C and checked twice daily for a colour change. In broths demonstrating growth ( secondary broth samples ), aliquots of 0.5 ml and 1.5 ml were transferred to 1.5 ml microcentrifuge tubes, and kept frozen at -70 C. The method described was repeated for plates with either bacterial or fungal contamination or no growth of mycoplasma colonies. Samples with no colony formation were considered to be test negative. 43

61 2.2.3 Statistical analysis Fisher s exact test was used to determine any statistically significant differences in the prevalence of positive samples between groups of cats, considering various factors, such as sex, age, origin of the cat, location of the cat, and presence of ocular or respiratory disease. P-values were determined using the computer software Stata (StataCorp., 2007), and a value of 0.05 was considered significant. Data were considered with respect to differences in prevalence between the three anatomic locations sampled. A Venn diagram was used to compare the pattern of distribution of mycoplasma-positive sites. McNemar s test was used to compare paired proportions of the presence of mycoplasma between the three anatomical locations. Stata software (StataCorp., 2007) was also used for McNemar s test, and a P-value of 0.05 was considered significant. Fisher s exact test was used to determine whether there was any significant difference in the prevalence of a positive swab in a particular anatomical location for which there was evidence of disease or absence of disease. The number of mycoplasma test-positive sites per cat was also considered (range 0 to 3 sites per cat). To determine whether the presence of mycoplasma in a particular anatomical location was independent of the presence of mycoplasma in other anatomical locations within a cat, the data were compared to a binomial distribution, as described by (Snedecor and Cochran, 1980) using a chi-squared test to determine the P-value (see Appendix 4 for calculation of expected numbers using the binomial distribution). Using a two-tailed test with 2 degrees of freedom (Snedecor and Cochran, 1978), a P-value of 0.05 was considered significant. The null hypothesis for this test was thus that the frequency of the number of positive sites per cat has a binomial distribution. For this hypothesis to be true, the presence of mycoplasma at a particular anatomical location must be independent of the presence of mycoplasma at any other site in each cat. 44

62 2.3 Results One hundred and ten cats were included in the study. Thirty-eight cats (34.5%) showed any signs of disease. The prevalence of cats with evidence of some form of ocular or respiratory disease (for the purpose of this study this includes conjunctival, upper and lower respiratory tract disease) was 31.8%. There were 24 cats (21.8%) with conjunctivitis or ocular discharge, 6 cats (5.5%) with upper respiratory tract signs (nasal discharge, sneezing, oral ulceration), and 14 cats (12.7%) had macroscopic lung changes which were considered to be post mortem artefact (Figure 2.1). The changes were limited to either reddish/pink mottling or small (< 2 mm) white nodules on the pleural surface of the lung. These were mostly observed on the caudal lung lobes, and were thought to be due to barbiturate precipitates resulting from the systemic absorption of barbiturate via intraperitoneal injection (personal communication, Prof. W. F. Robinson (BVSc MVSc, PhD, DipACVP, MACVSc), 2007). From the primary broth samples, 92/110 (83.6%) cats had one or more positive samples based on a colour change in mycoplasma liquid media. Of these 92 cats with samples plated on to agar, 79/110 cats (71.8%) had a sample with growth of typical mycoplasma colonies. Samples from which mycoplasma grew on solid medium were designated confirmed positives and used for further analyses. The prevalence of mycoplasma was compared for different groups of cats based on sex, age, origin and location within (and hence time spent in) the shelter. There was no significant difference observed between groups for any of these factors (Tables ). The prevalence of mycoplasma was compared for cats with any signs of ocular/respiratory disease and those with no evidence of ocular/respiratory disease (Table 2.6). There was no statistically significant difference in the frequency of mycoplasma isolation between these two groups (p = 1.0). Of a total of 330 swab samples (110 cats x 3 samples each), 134/330 (40.6%) swab samples were positive for mycoplasma. Of these 134 positive samples, considering 45

63 anatomical location, 15/110 (13.6%) of cats had a positive conjunctival sample, 72/110 (65.5%) a positive pharyngeal sample, and 47/110 (42.7%) a positive bronchial sample (Figure 2.2). The distribution and combination of mycoplasma-positive swabs in the cats, in terms of anatomical location, was compared using a Venn diagram (Figure 2.3). The most common occurrence was the presence of both a positive pharyngeal and bronchial swab (31.8% of cats), followed by no positive swabs (28.2%), a positive pharyngeal swab (21.2%), positive swabs in all three locations (6.4%), a positive result for both pharyngeal and conjunctival swabs (5.5%), and a positive bronchial swab (4.5%). There were no cats with the combination of a positive conjunctival and bronchial swab. mycoplasma was found at a significantly greater prevalence in both the pharynx and bronchus when compared with the conjunctiva, and also the pharynx compared to the bronchus. Each of these differences were significant with P-values of <0.001 (Tables ). The prevalence of cats with a positive mycoplasma swab in a particular anatomic site was then compared with respect to whether or not the cat was showing signs of disease in that anatomical location (Tables ). There were no significant differences observed between these groups; hence, the presence of mycoplasma was not found to be more common in a site for which there was evidence of disease. Significantly fewer cats had mycoplasma recovered from a single location than would be expected if the infection at that site was independent of infection at other sites. The observed number of mycoplasma-positive sites per cat, when compared with the expected value for a binomial distribution, is shown in Table The overall P- value of was statistically significant, showing that the data did not fit a binomial distribution. The results show that the presence of mycoplasma at a particular site is not independent of the presence of mycoplasma at other sites. 46

64 Figure 2.1: Location of signs of disease observed in the study population of cats (n = 110). Number of cats Conjunctiva Upper respiratory tract Lungs Location of disease a) Number of cats with disease observed in each of the three study sites Number of cats ocular URT lung ocular + URT ocular + lung URT + lung ocular + URT + lung other Location of disease b) Number of cats with any signs of disease including those with combinations of diseased sites

65 Table 2.2: Comparison of prevalence of mycoplasma-positive samples and gender of cats Swab + - Total % positive P-value Factor Male Female Table 2.3: Comparison of prevalence of mycoplasma-positive samples in adult and juvenile cats Swab + - Total % positive P-value Factor Adult Juvenile Table 2.4: Comparison of prevalence of mycoplasma-positive samples in stray and owned cats Swab + - Total % positive P-value Factor Stray Owned

66 Table 2.5: Comparison of prevalence of mycoplasma-positive samples and location of cats within shelter Swab + - Total % positive P-value Factor Holding room (1 day) Pound (8 days) Table 2.6: Comparison of prevalence of mycoplasma-positive samples between cats with and without signs of ocular/respiratory disease Swab + - Total % positive P-value Factor Ocular/respiratory disease present Ocular/respiratory disease not present

67 Figure 2.2: Anatomic distribution of mycoplasma-positive swabs Number of cats conjunctiva pharynx bronchus Anatomic site Mycoplasma-positive swabs

68 Figure 2.3: Anatomic distribution and combination of mycoplasma-positive swabs (n = 110 cats) pharynx conjunctiva bronchus negative = 31

69 Table 2.7: Comparison of paired proportions for mycoplasma status between the conjunctiva and pharynx using McNemar s test. pharynx Myco + Myco - Total % positive conjunctiva Myco /110 (13.6%) Myco Total % positive 72/110 (65.5%) P-value <0.001 Table 2.8: Comparison of paired proportions for mycoplasma status between the conjunctiva and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive conjunctiva Myco /110 (13.6%) Myco Total % positive 47/110 (42.7%) P-value <0.001 Table 2.9: Comparison of paired proportions for mycoplasma status between the pharynx and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive pharynx Myco /110 (65.5%) Myco Total % positive 47/110 (42.7%) P-value <0.001

70 Table 2.10: Comparison of prevalence of mycoplasma-positive conjunctival swabs between cats with and without conjunctivitis Conjunctival Swab Conjunctivitis Ocular discharge + - Total % P-value positive Yes No Table 2.11: Comparison of prevalence of mycoplasma-positive pharyngeal swabs between cats with and without signs of upper respiratory tract disease Pharyngeal Swab URT signs (nasal discharge, sneezing. oral ulceration) + - Total % P-value positive Yes No Table 2.12: Comparison of prevalence of mycoplasma-positive bronchial swabs between cats with and without gross lung pathology Bronchial Swab + - Total % P-value positive Lung Pathology Yes No

71 Table 2.13: 2 test of goodness of fit of the binomial distribution, applied to the number of cats with a particular number of mycoplasma-positive anatomic sites. The overall prevalence of mycoplasma-positive sites was 134/330 sites from 110 cats. Number of Number of cats 2 = (O-E) 2 /E P-value positive sites Observed (O) Expected (E) * Overall (2df) 0.005* *Statistically significant (p 0.05)

72 2.4 Discussion Of the population of cats presented to an animal shelter in Melbourne, Australia in June of 2003, approximately one third had clinically apparent ocular and upper respiratory tract disease. However, none of the cats were considered to have characteristics of acute or chronic pneumonia. A high percentage of cats (71.8%) were infected with mycoplasma and more often than not, mycoplasma infection was present at more than one anatomical site. Mycoplasma was most prevalent in the pharynx, followed by the bronchus then conjunctiva. While these findings are similar to other studies for the conjunctiva and upper respiratory tract of cats (Heyward et al., 1969; Blackmore et al., 1971; Tan and Miles, 1972, 1974b; Tan et al., 1977a; Tan et al., 1977b; Randolph et al., 1993; Bannasch and Foley, 2005; Low et al., 2007), there was a higher than expected number of cats (42.7%) with mycoplasma cultured from the lower respiratory tract. Isolation of mycoplasmas from the lower respiratory tract of healthy cats has been considered an abnormal finding due to lack of isolation from this location in previous studies (Tan et al., 1977a; Padrid et al., 1991; Randolph et al., 1993). Potentially, this may have been due to the liberation of mycoplasmacidal factors during sampling, or because the LRT is most often sampled when investigating the cause of disease and only rarely examined for the presence of mycoplasma in healthy cats. The present study shows that over 80% of the cats from which mycoplasma was isolated from the bronchus had no evidence of pulmonary lesions. This is directly contrasted by a study that found no evidence of mycoplasmas in tracheobronchial lavage samples from healthy cats compared to isolation from 21% of cats with pulmonary disease (Randolph et al., 1993). A difference between this and the current study is the sampling method of tracheobronchial lavage versus bronchial swab sampling, and may account for the different findings. However, the current findings are consistent with a different study where 29% of tracheal specimens from healthy cats contained M. gateae (Heyward et al., 1969). The current study suggests that mycoplasmas are a normal inhabitant of the LRT which is supported by other studies showing the LRT is not a sterile environment in cats (Padrid et al., 1991; Dye et al., 1996). This contradicts the assumption made by some authors that isolation of mycoplasmas from the LRT of cats with pulmonary disease is sufficient evidence to conclude it has a 47

73 pathogenic role in the disease (Malik et al., 1991; Foster et al., 1998; Chandler and Lappin, 2002). The current study clearly shows that mycoplasma isolation from an animal with pulmonary disease should be interpreted with caution before assuming an aetiological role. The absence of clinical observations of pneumonia in cats that either had clinically apparent URT disease or mycoplasmas in the LRT suggests that feline mycoplasmas are unlikely to cause primary pneumonia (nor was there any evidence that the URT disease seen in this study was caused by mycoplasmas). Moreover, the current study has shown that feline mycoplasmas may be considered commensal organisms. This is in contrast to studies that have suggested mycoplasmas cause disease in cats (Tan and Markham, 1971b; Campbell et al., 1973b; Tan, 1974; Tan and Miles, 1974b; Haesebrouck et al., 1991; Malik et al., 1991; Randolph et al., 1993; Foster et al., 1998; Chandler and Lappin, 2002; Foster et al., 2004b; Bannasch and Foley, 2005; Low et al., 2007). Mycoplasmas may have the capacity to cause disease under different conditions, such as in an immunocompromised host, or in the presence of existing respiratory disease. Existing mucosal damage, with the attendant inflammatory changes, might provide mycoplasmas with the opportunity to cross the mucosa and cause additional damage. If this were the case, they could be classed as commensal organisms with the capacity to act as opportunistic invaders. No significant association between gender, age, whether cats were stray or owned, the location in or time spent at the facility, and the presence of mycoplasma was demonstrated in the current study. Most importantly, there was no association between the presence or absence of mycoplasma and the presence or absence of clinical signs of ocular and respiratory disease. Specifically, there was no association between the presence of mycoplasma at a particular anatomical location with any evidence of disease in that location. Few other studies have considered such associations statistically. One study found no association between mycoplasmapositive cats with respect to age or gender (Randolph et al., 1993). However, a significant association was demonstrated between the presence of mycoplasma and upper respiratory tract disease in a population of 573 shelter cats (Bannasch and Foley, 2005). This is in contrast to the current study where the association with URT disease demonstrated by Bannasch and Foley (2005) may be indicative of either a 48

74 different or more virulent mycoplasma species or a high prevalence of other respiratory pathogens which influenced the mycoplasmal population. The absence of any statistical associations of ocular or respiratory disease with mycoplasma isolation suggests that these organisms are not pathogenic in the feline population in the current study. Although the precise pathogenic mechanisms and virulence factors of most pathogenic mycoplasma species have not yet been defined precisely, particular virulence characteristics have been determined for some. Mechanisms of adherence, cell injury and antigenic variation have been demonstrated in some mycoplasmas (reviewed in (Simecka et al., 1992)). It may be the inability of feline mycoplasmas to express such virulence factors that differentiates them from pathogenic mycoplasma species. A more thorough study of potential virulence factors in mycoplasmas isolated from cats with pneumonia or clinical disease needs to be undertaken. In addition, greater numbers of diseased cats, greater numbers of juvenile cats, or repeated studies at different times of the year and/or in different years, may show different associations. These aspects should be addressed in future investigations. The current study showed a statistical significance in isolation of mycoplasma between the anatomical sites sampled. In particular, there was a much higher prevalence of mycoplasma isolated from both the pharynx and bronchus than the conjunctiva. Other reports have not statistically analysed the difference in prevalence among these anatomical sites, although many studies have demonstrated this trend for the pharynx being the most frequent site for mycoplasma in the cat (Heyward et al., 1969; Blackmore et al., 1971; Tan and Markham, 1971b; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b; Randolph et al., 1993; Hartmann et al., 2008; Veir et al., 2008) To speculate on the results of this study, one could suggest that the pharynx is the favoured environment for feline mycoplasmas compared with both the conjunctival and bronchial environments. Although this site is generally more tolerant of a microfloral population, the prevalence of these organisms in the other sites also suggests that the organisms can evade the host defences sufficiently to reside there. It is apparent that when a cat is infected with mycoplasma, (most commonly in the 49

75 pharynx) it is much more likely to also have mycoplasma in one or more of the other sites (conjunctiva and bronchus). The presence of mycoplasma at a particular site is not independent of mycoplasma presence at other sites as indicated by frequency distribution data from the current study not fitting a binomial distribution. Previous studies of the anatomical location of mycoplasmas in feline populations have sampled multiple sites (such as (Heyward et al., 1969; Spradbrow et al., 1970b; Blackmore et al., 1971; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b)), but have not considered the number of positive sites or the relative frequency of the distribution and combinations of positive sites in the cats. The current study provides a more detailed description of the ecology of the mycoplasmal organisms within the population studied, which may assist in further determining the mechanism of host interactions. The combination of mycoplasma in the pharynx and bronchus was greater than any site alone. This appears to be a logical progression, rather than a post mortem occurrence. Evidence for this suggestion comes both from studies demonstrating that the LRT is not sterile in cats (Padrid et al., 1991; Dye et al., 1996), and also the findings of the present study that some cats had mycoplasma isolated from the bronchus only. Mycoplasmas in other species are able to access and attach to the LRT (reviewed in (Cassell et al., 1985)) and can colonise both ciliated and nonciliated epithelial surfaces in the respiratory tract (Bemis, 1992). It would not be surprising if feline mycoplasmas shared some of these characteristics. Therefore, once a cat becomes infected with mycoplasma, it colonises other sites that are either anatomically or functionally related (e.g., pharynx to bronchus), or into the conjunctiva by saliva when self grooming or vice versa via the nasolacrimal duct. It would be logical to assume that it is easier to transmit an organism between sites in an individual than between individuals. The ability of the mycoplasma to colonise any particular site is still reliant on avoiding the host defence mechanisms at that site. Perhaps this occurs because mycoplasma do not disturb the host defence mechanisms by remaining attached to epithelium, thus avoiding being presented as a foreign antigen to the cat s immune system. Based on the findings in this study and also the rate of isolation of mycoplasmas from other studies (Heyward et al., 1969; Blackmore et al., 1971; Tan and Miles, 1972, 50

76 1974b; Tan et al., 1977a; Tan et al., 1977b; Bannasch and Foley, 2005), one could suggest that mycoplasma could be isolated from almost any anatomical site from the conjunctiva, through the nasal, oral, pharyngeal, and tracheal regions to the bronchi. How far down the respiratory tract one might isolate mycoplasma is unknown; however, other evidence indicates that isolating mycoplasma from lung tissue might be particularly difficult due to mycoplasmacidal factors released during attempted isolation (Tully and Rask-Nielsen, 1967; Kaklamanis et al., 1969; Rosendal, 1979). This problem might be overcome by using PCR directly from tissue swabs, as this technique does not rely on the culture of viable mycoplasmas to detect their DNA. PCR also bypasses any potential issues with mycoplasma being slow or difficult to culture, or requiring specific growth conditions. Bronchoalveolar lavage (BAL) is a technique for sampling the alveoli (Tully et al., 1972; Padrid et al., 1991; Randolph et al., 1993; Foster et al., 2004c), and has the advantage of being a useful diagnostic tool in live animals. However, using this approach, it is not possible to specifically establish whether the organisms originate from the bronchi themselves or from the bronchioles or alveoli. Bronchial swab sampling was chosen as the method for mycoplasma recovery from the lower respiratory tract in the current study. Although techniques such as BAL sample more distal sites of the lower respiratory tract, bronchial swabs provided samples that were more consistent for comparisons to the conjunctival and pharyngeal samples collected concurrently in each cat. The current study demonstrates the overall prevalence of mycoplasma in a population of cats by culturing the organism from swabs and evaluates their relative prevalence in different subsets of this population, in order to determine if there were any associations between different groups of cats, and the presence or absence of disease. The effect of multiple species of mycoplasma may have complicated this study, particularly if pathogenic and non-pathogenic mycoplasma species were present, and will be considered in Chapter 3. 51

77 52

78 Chapter 3 Molecular Characterisation of Feline Mycoplasma 3.1 Introduction Identification of feline mycoplasma isolates to the species level is an important aspect of any population or clinical study, particularly for investigating association with disease. Previous reports have incompletely considered this aspect limiting the significance of the conclusions (Crisp et al., 1987; Malik et al., 1991; Randolph et al., 1993; Foster et al., 1998; Bannasch and Foley, 2005; Low et al., 2007; Trow et al., 2008; Veir et al., 2008). Even though it was demonstrated in Chapter 2 there was no significant association of mycoplasmas with ocular or respiratory disease, each mycoplasma species should be considered separately to avoid species related differences biasing results. Identifying feline mycoplasmas to the species level by PCR amplification and sequencing of the 16S rrna gene has been reported (Brown et al., 1995). Molecular methods have several advantages compared with traditional serological methods (see section ), and relate primarily to the wide accessibility of reagents and savings in time. A recent study has established the molecular method of SSCP to examine nucleotide variations of mycoplasma species (Jeffery et al., 2007). One of the advantages of this approach is the ability to rapidly screen a large number of samples, making it applicable to studies such as this one. This chapter describes analysis of part of the 16S rrna gene by PCR-based SSCP for identification of feline mycoplasma. This technique was used to classify mycoplasmapositive culture samples so different species could be identified by comparison to published 16S rrna sequences. 53

79 3.2 Materials and method DNA extraction by heat lysis Genomic DNA from mycoplasma-positive swab samples ( secondary broth samples described in section 2.2.2) was prepared for PCR. Mycoplasma media supernatant was removed by aspiration following centrifugation of broth at 14,000 x g for 3 min. The remaining pellet was twice washed and centrifuged, first in phosphate buffered saline (PBS), then in sterile deionised water (Appendix 5). The resuspended mycoplasmas were heated to 95 ºC for 5 min to release the DNA from the mycoplasma cells. This mycoplasmal lysate was stored at -70 ºC then thawed and used directly for all PCR studies without further purification Mycoplasma 16S rrna gene amplification The sequences of eight feline mycoplasma species for which the 16S rrna genes had been characterised (Brown et al., 1995) were obtained from GenBank (National Centre for Biotechnology Information, Bethesda, MD, USA and were aligned using the computer programs GeneWorks (Oxford Molecular Group, now part of Accelrys, San Diego, CA, USA) and ANGIS BioManager (Australian National Genomic Information Service, CLUSTALW(Accurate) application (Thompson et al., 1994) Primers designed to amplify the 16S rrna gene of feline mycoplasmas (Brown et al., 1995), named MycoF and MycoR (Table 3.1 and Figure 3.1), were commercially synthesised (GeneWorks Pty. Ltd. South Australia, Australia). This primer set binds to terminal 16S rrna sequences that are conserved in Mollicutes (Brown et al., 1990; Deng et al., 1991), and amplifies a 1.5 kb fragment. Primers were reconstituted to a 25 µm solution in sterile deionised water. PCR was performed in a 50 µl reaction containing 5 µl of 10 X magnesium free buffer, 1.5 mm MgCl 2, 0.2 mm nucleotide mix, µg each primer, 1.5 units of Taq DNA polymerase, and 10 µl of mycoplasma DNA. PCR was run in an Eppendorf Mastercycler 5330 plus (Hamburg, Germany). The PCR conditions were 50 cycles of the following: 94 C for 45 s, 59 C for 60 s and 72 C for 2 min. Positive control 54

80 samples used for the reaction were DNA from M. felis, M. arginini, M. synoviae, and M. gallisepticum (courtesy of The University of Melbourne Veterinary Science Department Microbiological Research Laboratories). A negative control for the reaction included 10 µl of RNAse free water. PCR products were subjected to electrophoresis through 1.5% agarose gels containing ethidium bromide in 1 X Tris-phosphate EDTA buffer; 10 µl of PCR product plus 3 µl loading buffer were loaded in each lane, and a molecular weight marker (Sigma P9577 DNA ladder marker (Sigma-Aldrich Co. NSW, Australia) used to estimate product size. Gels were electrophoresed at 100 V, and amplicons detected and photographed upon transillumination with UV light Primer design for PCR-SSCP analysis PCR primers were designed to produce a suitable sized DNA fragment for SSCP. The aim was to have the amplified fragment in the range of 250 to 350 bp and from a region of the 16S rrna gene with variability among the eight feline mycoplasma species previously described (Brown et al., 1995). Two sets of potentially suitable primers were identified, FB50 and RA50 to amplify a 233 bp fragment, and FD1000 and RC1000 to amplify a 274 bp fragment (Table 3.1 and Figure 3.1). Primers were commercially synthesised (GeneWorks Pty. Ltd. South Australia, Australia), and were reconstituted to a 100 mm solution with sterile deionised water. The PCR was run on a Perkin Elmer DNA Thermal Cycler 480 machine (Waltham, Massachusetts, USA) and was performed with 5 µl of 10 X magnesium free buffer, 3.0 mm MgCl 2, 0.2 mm nucleotide mix, µg each primer, 2 units of Taq DNA polymerase, and 1 µl of genomic DNA in a 50 µl reaction. PCR conditions were denaturation at 94 ºC for 5 min, 35 cycles of the following: 94 ºC for 30 s, 60 ºC for 30 s and 72 ºC for 30 s. This was followed by an elongation step at 72 ºC for 5 min before being cooled to 4 ºC. Positive controls used for the reactions were M. felis and negative controls of sterile deionised water in place of genomic DNA. 55

81 Table 3.1: Mycoplasma 16S rrna gene PCR primer sequences Primer Nucleotide sequence (5 to 3 ) Position MycoF AGA GTT TGA TCC TGG CTC AGG A MycoR GGT AGG GAT ACC TTG TTA CGA CT FB50 GCG AAC/T GGG TGA GTA ACA CG RA50 TCT CAG TCC CAG/A TGT GGC FD1000 CTC GTG TCG TGA GAT GT RC1000 TG/AC GAT TAC TAG CGA TTC C

82 Figure 3.1 : Mycoplasma 16S rrna gene, primers, and amplified products Myco F FB50 RA50 FD1000 RC1000 Myco R primers 16S rrna gene nt position ~1.5kb fragment 233 bp fragment 274 bp fragment 263 bp fragment 476 bp fragment

83 These smaller PCR products were electrophoresed through 2% agarose gels in 0.5 X Tris Boric EDTA at 100 V; 5 µl of PCR product plus 3 µl of 6 X loading dye (Promega, Madison, WI, USA) was added to each well, and ΦX174 DNA/HaeIII (Promega, Madison, WI, USA) was included as a molecular weight marker to estimate product size. Gels were stained with ethidium bromide to detect amplicons and were photographed upon transillumination with UV light. Primer combinations MycoF and RA50 (to amplify a 263 bp fragment), and FD1000 and MycoR (to amplify a 476 bp fragment) were then tested by PCR of M. felis controls to assess their suitability for PCR-SSCP. Primers FD1000 and MycoR were chosen as the most suitable primer pair for the study based on both size and quality of PCR product. These primers amplify a 457 bp fragment from position of the 16S rrna gene (Figure 3.1). The sequence alignment of the 16S rrna gene fragment these primers amplify for the eight feline species described is shown in Appendix PCR optimisation In addition to selecting the most suitable primers for both amplification of mycoplasma DNA from swab samples and to give an appropriate product (size and variability) to use in PCR-SSCP, the following steps were taken to optimise the PCR. M. felis was used as the positive sample. The PCR conditions and annealing temperature were selected based on the length and expected nucleotide composition of the primers and fragment to be amplified. These conditions provided suitable products with a clear band on an agarose gel. A magnesium titration was performed to determine the optimum MgCl 2 concentration for the reaction. There was no difference in intensity of bands observed across an MgCl 2 concentration range of mm. A concentration of 3.0 mm was used for the reactions. The amount of sample DNA used in the reactions was varied from 1-3 µl; 2 µl was chosen as giving bands of maximal intensity. 56

84 The method of mycoplasma DNA extraction was considered. The heat lysis method described in section was compared with a modified protocol which eliminated the heating step, and also to a column purification method with the DNeasy Tissue Kit (Qiagen Pty. Ltd. Victoria, Australia) according to manufacturer s instructions. There was no considerable difference in the intensity of PCR products using either extraction method (Appendix 7). The modified protocol without the heating step produced a product of reduced intensity using agarose gel separation. As there was no apparent advantage using the column purification method, the heat lysis method was selected for the study. The possibility of mycoplasma media having an inhibitory effect on the PCR was considered. The column purification method should have removed all components of the media, whereas the heat lysis method relies on removal of the majority of such components by heating and dilution. As described above (and shown in Appendix 7), there was no difference in the intensity of PCR products for the two extraction methods. Additionally, following DNA extraction by the heat lysis method, liquid mycoplasma media was added in increasing concentrations to the M. felis samples. Bands of lower intensity were evident in samples containing media, and this was more apparent at higher concentrations (Appendix 7). The heat lysis method of DNA extraction was considered suitable for the reactions. The final optimised PCR with primers FD1000 and MycoR in a Perkin Elmer DNA Thermal Cycler 480 machine (Waltham, Massachusetts, USA) was determined to be: 5 µl of 10 X magnesium free buffer, 3.0 mm MgCl 2, 0.2 mm nucleotide mix, µg each primer, 2 units of Taq DNA polymerase, and 2 µl of genomic DNA in a 50 µl reaction. PCR conditions were denaturation at 94 ºC for 5 min, 35 cycles of the following: 94 ºC for 30 s, 60 ºC for 30 s and 72 ºC for 30 s, then elongation at 72 ºC for 5 min and finally cooled to 4 ºC. Positive control samples used for the reactions were M. felis and negative control samples were sterile deionised water. PCR products were subjected to electrophoresis through 2% agarose gels in 0.5 X Tris Boric EDTA) at 100 V. 5 µl of PCR product with 3 µl of 6 X loading dye (Promega, Madison, WI, USA) was added to each well, and ΦX174 DNA/HaeIII (Promega, 57

85 Madison, WI, USA) was included as a molecular weight marker to estimate product size. Gels were stained with ethidium bromide to detect amplicons and were photographed after transillumination with UV light PCR of mycoplasma swab samples PCR was first performed on one of five colony isolates from each of the positive secondary broth samples (from section 2.2.2). Primers FD1000 and MycoR were used with PCR conditions and controls as per section Additionally for each positive sample, the five colony isolates were then pooled, and the PCR was performed on the isolate pool Non-isotopic Single-Strand Conformation Polymorphism (SSCP) Assessment of variability within samples Three cats with mycoplasma-positive samples were randomly chosen for each anatomical site studied (conjunctiva, pharynx and bronchus) to assess the likelihood of variability within the five colonies that had been selected for each sample in section DNA from all five colonies of each positive sample was amplified individually by PCR as described in section Feline DNA from an animal unrelated to this study was used as an additional negative (host) control. Precast SSCP gels (GMA for SSCP, Elchrom Scientific AG, Switzerland) were placed into the SEA 2000 Electrophoresis apparatus, (Elchrom Scientific AG, Switzerland) containing 1 X Tris-Acetic EDTA buffer, pre-cooled with a cooling system (MultiTemp III, Amersham Pharmacia Biotech) to 7.4 C and pre-run at 74 V for 15 min prior to loading PCR products. PCR products were prepared by adding 10 µl of PCR product to 5 µl of DNA sequence stop solution (Promega, Madison, WI, USA). The mix was briefly centrifuged, and denatured in a preheated PCR machine (Perkin Elmer DNA Thermal Cycler 480, Waltham, Massachusetts, USA) at 94 C for 15 min. The tubes were then snap-cooled in a freeze-block at -20 ºC and loaded into wells of the SSCP gel, and run for 17 h at 74 V and a constant 7.4 C. ΦX174 DNA/HaeIII (Promega, Madison, WI, USA) was used at as a reference marker. Gels 58

86 were stained in ethidium bromide for 15 min, de-stained in water for 10 min, and photographed upon transillumination with UV light SSCP of mycoplasma-positive PCR samples Each of the positive PCR products described in section (for both the single isolate and pooled isolate samples) had SSCP performed by steps described in section SSCP gels were analysed to determine the number of unique SSCP profiles. An SSCP profile is the pattern of bands visualised on a gel which correspond to the two single strands of DNA. Their migration through the gel is a function of the secondary and tertiary structure they form following denaturation which is based on their nucleotide sequence. Each sample may have two bands (for each strand), or greater than two if there are DNA fragments with nucleotide variations (which could arise due to the presence of mutations, subspecies differences, or more than one species being present). If more than one species were present, the SSCP profile would reflect this in the form of a combined profile, with two bands from each species present. Each time a unique profile was identified, the profile was assigned the number of the original cat the sample corresponded to, and all subsequent samples with the same profile were given this number. Once the different profiles from each gel were identified, representatives of these were run on summary gels in order to display profiles side by side. Samples with the same profile were disregarded in order to select only the unique SSCP profiles for sequencing DNA Sequencing A representative PCR product of each different SSCP profile was sequenced to determine which mycoplasma species the profile corresponded to. Prior to sequencing PCR products were purified using Wizard PCR Preps DNA Purification System (Promega, Madison, WI, USA) (Appendix 8). Prior to sequencing, purified DNA was then subjected to electrophoresis through 2% agarose gels using genomic DNA standards of known concentration (pgem control DNA, 200 pg/µl, Promega, Madison, WI, USA), to estimate sample concentration. 59

87 The sequencing reactions were performed according to the manufacturers directions in microcentrifuge tubes using ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kits (Versions 2.0 and 3.1) (Applied Biosystems, California, USA). Reactions of 20 µl were prepared by combining 6 µl of Terminator Ready Reaction mix (Applied Biosystems, California, USA), 3.5 µl of primer (diluted 1:100 from previously described 100 µm solution to make 1 µm) and 3-10 ng of DNA. For each sample, sequencing reactions were performed in both the forward and reverse direction (with primers FD1000 and MycoR respectively). Sequencing reactions were performed according to the manufacturer s directions in a Perkin Elmer DNA Thermal Cycler 480 (Waltham, Massachusetts, USA) with cycling conditions of 96 ºC for 5 min, 25 cycles of the following: 96 ºC for 30 s, 50 ºC for 15 s and 60 ºC for 4 min, then cooled to 4 ºC. Sequencing reactions were purified by isopropanol precipitation to remove excess mineral oil and reagents, according to the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (Versions 2.0 and 3.1) directions (Applied Biosystems, California, USA). This was achieved by addition of 75% isopropanol to precipitate the extension products, followed by removal of the supernatant leaving the purified products. Precipitated products were analysed by Griffith University DNA Sequencing Facility (GUDSF) (School of Biomolecular and Biomedical Science, Nathan, Queensland, Australia) Sequence analysis Sequences from GUDSF were analysed using the software program Chromas version 2.24 (Technelysium Pty. Ltd. Tewantin, QLD, Australia, In order to analyse multiple sequences or to compare the forward and reverse sequences, the CAP (contig assembly program) Sequence Assembly Machine software (Campus IFOM-IEO Milan, Italy, ((Huang, 1992))) was used to perform contigs of forward and reverse sequences, and sequence alignments. Once a contig sequence was established, BLAST (Basic Local Alignment Search Tool) searches of the sequences were conducted using the National Centre for Biotechnology Information 60

88 website ( BLAST searches the database GenBank (National Centre for Biotechnology Information, Bethesda, Maryland, USA) and compares the sample sequence to determine whether it was identical or similar to other mycoplasma sequences Statistical analysis SSCP results from section were used to determine the overall prevalence of the different mycoplasma species identified (as represented by the different SSCP profiles F and R). These data were then further analysed to determine if positive samples within an individual cat were of the same species, and to determine if there was any association between anatomical location and mycoplasma species, or between mycoplasma species and presence of disease. To determine if there were significant differences between species of mycoplasma and combination of anatomical sites they were present in (where there were multiple positive sites in an individual cat), Fisher s exact test was used. This compared proportions with respect to which combination of anatomic sites within each cat were positive for each of the two species found. P-values were determined using the computer software program Stata (StataCorp., 2007) with a value of 0.05 considered significant. To determine whether a particular mycoplasma species was associated more frequently with a particular anatomical location, comparisons were made for the three anatomic locations. This was determined by a random effects logistic regression to take account of the non-independence of observations, where multiple observations within a cat were correlated, as some samples came from the same cat (Dohoo et al., 2003). Stata software was used to calculate the random effects logistic regression using the xtlogit command (StataCorp., 2007). Data were organised into binary form with respect to positivity to profile F. For every cat, each site (conjunctiva (A), pharynx (B) and bronchus (C)) was given either the number 1 if positive for F, 0 if not positive for F (i.e. was profile R), or left blank if no sample, negative for mycoplasma or unknown. This data only considers those samples where a result for species based on SSCP was known, and therefore directly compares the proportion of the different 61

89 species found between each site. Samples that were negative for mycoplasma were not included in this test. The odds were defined as p/(1-p) where p was the proportion of observations that were positive to F. The odds ratio was the ratio of the odds when comparing sites. An odds ratio of 1 indicates the odds of having a positive sample at each site are the same. A P-value of 0.05 was considered significant. McNemar s test was used to compare paired proportions for inclusion of mycoplasma-negative cats in the analysis. This compared both the presence and absence of mycoplasma (both overall and then for each mycoplasma species found) between different anatomic locations. Stata software (StataCorp., 2007) was used for McNemar s test and a P-value of 0.05 was considered significant. Fisher s exact test was used to investigate any association between the presence of each mycoplasma species and the presence of disease. Although there was no statistically significant relationship found in Chapter 2 for mycoplasma-positive results, this test considered any association of each individual species with the presence of disease. Results were considered overall, and specifically where disease was present in the same anatomic location as the positive swabs. Fisher s exact test was performed with PEPI for Windows (Abramson, 2004) to determine a P-value. A P-value of 0.05 was considered significant. 62

90 3.3 Results PCR of controls and samples PCR amplification of the 16S rrna gene of reference for mycoplasma species M. felis, M. arginini, M. synoviae and M. gallisepticum using the primer pair MycoF and MycoR yielded the expected amplified product, of ~ 1.5 kb (Figure 3.2). The primer pair FD1000 and MycoR yielded a fragment of the expected size of ~ 450 bp (Figure 3.3). This occurred with both the reference strain M. felis, and the feline samples assumed to be mycoplasma-positive based on the colour change in broth and subsequent growth of typical mycoplasma colonies on agar. One hundred and twenty nine (96.3%) of the 134 feline samples that were mycoplasma culture-positive were also PCR positive. This represents an overall prevalence of 129/330 samples (39.1%), or 75/110 cats (68.2%) Screening of isolates by SSCP for genetic variability All five colony isolates were tested from three cats for each of the three anatomic sites investigated (i.e. 5 isolates each from 9 samples overall). Four different SSCP profiles were identified (data not shown). Seven of the nine samples had five identical isolate profiles whereas the other two samples had more than one profile identified amongst the five isolates. This demonstration of variability provided evidence for analysing both the single colony isolate, and a pooled sample of the five colony isolates by SSCP for all samples SSCP-coupled analysis of all samples, and identification There were two frequently occurring SSCP profiles, which were identified as F, as it was identical to the M. felis positive control, and R, as this profile was retarded (in migration) by comparison (Figure 3.4). An SSCP result attributable to either of these profiles was obtained for 111 samples, from 66 cats (Table 3.2). There were three additional profiles, which were all variants of F, as they contained the F profile as 63

91 Figure 3.2: Gel showing PCR products amplified by primers MycoF and MycoR 1,500 bp Key: Lane on gel Sample 1 Sigma P9577 marker 2 M. felis 3 M. felis(diluted*) 4 blank 5 M. felis 6 M. felis (diluted*) 7 blank 8 M. gallisepticum 9 M. synoviae 10 blank 11 M. arginini 12 M. arginini (diluted*) *(all dilutions 1 in 5 with water)

92 Figure 3.3: Gel showing PCR products amplified by primers FD1000 and MycoR 603bp 310bp Key: Lane on gel Sample + Primers 1 X174 DNA/Hae III Marker 2 M. felis with MycoR and RA50 3 M. felis with FD1000 and MycoR

93 Figure 3.4: SSCP gel of feline mycoplasma samples showing two distinct profiles; F and R R F R F ΦX

94 Table 3.2: Results of SSCP for the 129 mycoplasma-positive PCR samples M. felis M. gateae / M. arginini Both Negative (no SSCP Number of positive result) # Samples * % of positive # Cats * % of positive *This sample (106C) was excluded from further analysis Table 3.3: Number of variant SSCP profiles identified and comparison to health status of cats SSCP profile Number of Number of cats Diseased cats variant samples F(a) F(b) F(c) Total 8 6 1

95 well as an extra band (Figure 3.5). Variant F profiles were evident in 8 samples from 6 cats (Table 3.3). The SSCP result for both the single colony and the five pooled colonies for each sample was compared to determine whether they had the same or different profile. In every case except one (sample 106C), the result for the single and pooled samples were the same, or contained a variant of the same profile. Representative samples of the two frequently occurring SSCP profiles (F and R) were sequenced. Sequences were compared using the BLAST search tool of databases containing known/published sequences of the 16S rrna gene. Profile F showed 98% sequence similarity to M. felis (accession number U (Figure 3.6)). Using SSCP, this profile was identical to the M. felis reference strain used throughout the experiments. Representative samples from two of the three F profile variants were also sequenced, both demonstrating 98% sequence similarity to the reference strain of M. felis using BLAST (Figure 3.7). Overall, there was minimal variation between the different isolates, with 1-3 nucleotide differences present. Profile R was determined to be M. gateae using BLAST (accession number U ), with 99% sequence similarity to the reference strain (Figure 3.8). However, it was not possible to extrapolate this finding to differentiate between M. gateae and M. arginini for the SSCP profile R. There is only one nucleotide difference between M. gateae and M. arginini in this region of the 16S rrna gene, and although the sample that was sequenced was similar to M. gateae, the other samples in the study with the same SSCP profile could have been either of these species. This was due to a single nucleotide difference that may not alter the secondary structure of the DNA sufficiently to be detected by SSCP, particularly for an amplicon size as used here (~ 450 bp). For results analysis, samples that were considered to be profile F will be referred to as M. felis and those of profile R will be referred to as M. gateae/m. arginini. 64

96 3.3.4 Numbers and proportion of each mycoplasma species found M. felis was more prevalent than M. gateae/m. arginini for both the number of samples and number of cats (Table 3.2). There were 17/129 samples and 11/75 cats that were positive for mycoplasma by PCR, for which an SSCP result was not obtained. This was either because there was an unreadable result, no result, or for the purposes of data analysis, a variant profile. These will be considered further in the discussion. There was one sample, 106C, for which the SSCP profile was different for the single colony sample and the pooled sample of all five colonies, and although this will also be considered in the discussion, it was excluded from data analysis. For the remaining 7 cats with both mycoplasma species, the differing species were present in different samples and hence anatomical locations Comparison of colour change observed in liquid culture and identification based on PCR-SSCP and nucleotide sequencing Analysis of the correlation between colour changes observed in liquid culture of mycoplasmas and the species identified for each sample by PCR-SSCP and nucleotide sequencing was performed. Of the 111 samples for which an SSCP result was obtained, there were 32 where the colour change of some or all of the broths differed from that expected from the taxa identified. This is based on knowledge of the preferred energy utilisation of that species (Table 1.2). Where a primary broth changes colour to yellow, it is indicative of an acid shift and represents glucose fermentation, which would be expected of M. felis. Alternatively, a colour change to red is indicative of an alkaline shift, and represents arginine utilisation, which would be expected of M. gateae and M. arginini. The colour change in both the primary broth (a single sample from the direct swab) and then the secondary broths (five samples obtained from five mycoplasma colonies from the primary plates) was considered. Each of the five secondary broth samples displayed the same colour change as the primary broth in 80/111 samples. There were 13/111 samples where the colour change of the five secondary broth samples was different from that of the primary broth. Finally, there were 18/111 samples where both colour changes were represented in the five secondary broth samples. 65

97 Figure 3.5 SSCP gel of feline mycoplasma samples showing variant F profiles in addition to the two profiles F and R Key: M. felis positive control F profile F profile variant R profile F profile

98 Figure 3.6: Nucleotide sequence alignment of sample 103C (profile F) with the sequence representing M. felis consensus M.felis 103C CGAGCGCAACCCTTG CCTTAGTTAAATATTCTAGGGAGACTGCCCGAGTAA...GTC......T consensus TTGGGAGGAAGGTGGGGACGACGTCAAATCATCATGCCTCTTACGAGTGGGGCAACACAC M.felis C... consensus M.felis 103C GTGCTACAATGGATGGTACAAAGAGAAGCAATACGGCGACGT GAGCAAATCTCAAAAAA...N......T consensus CCATTCTCAGTTCGGATTGTAGTCTGCAACTCGACTACATGAAGTCGGAATCGCTAGTAA M.felis C consensus TCGTAGATCAGCTACGCTACGGTGAATACGTTCTCGGGTCTTGTACACACCGCCCGTCAC M.felis C consensus ACCATGGGAGCTGGTAATGCCCGAAGTCGGTTTTGTTAACTACGGAGACAACTGCCTAAG M.felis C consensus GCAGGGCCGGTGACTGGGG M.felis C...

99 Figure 3.7 Nucleotide sequence alignment of F profile variant samples (110C and 111B) with M. felis and regular F profile (103C) consensus TCGTGAGATGTTCGGTTAAGTCC GCA CGAGCGCAACCCTTGT--CCTTAG M.felis...T...A...GTC B...T...A C...N...T C consensus TTAAATATTCTAGGGAGACTGCCCGAGTAATTGGGAGGAAGGTGGGGACGACGTCAAATC M.felis B C C consensus ATCATGCCTCTTACGAGTGGGGCA-ACACACGTGCTACAATGGATGGTACAAAGAGAAGC M.felis B C...T C consensus AATACGGCGACGTTGAGCAAATCTCAAAAAACCATTCTCAGTTCGGATTGTAGTCTGCAA M.felis...N B C C consensus CTCGACTACATGAAGTCGGAATCGCTAGTAATCGTAGATCAGCTACGCTACGGTGAATAC M.felis B C C consensus GTTCTCGGGTCTTGTACACACCGCCCGTCACACCATGGGAGCTGGTAATGCCCGAAGTCG M.felis B...N C...C C consensus GTTTTGTTAACTACGGAGACAACTGCCTAAGGCAGGGCCGGTGACTGGGG M.felis B...N G C C...

100 Figure 3.8: Nucleotide sequence alignment of sample 14A (profile R) with M. gateae consensus GAGATGTTTGGTCAAGTCCTGCAACGAGCGCAACCCCTATCTTTAGTTACTAACGAGTCA M.gateae... 14A... consensus M.gateae 14A TGTCGAGGACTCTAGAGATACT CCTGGGTAACCGGGAGGAAGGTGGGGATGACGTCAA...G CA consensus ATCATCATGCCTCTTACGAGTGGGGCAACACACGTGCTACAATGGTCGGTACAAAGAGAA M.gateae... 14A consensus GCAATATGGCGACATGGAGCAAATCTCAAAAAGCCGATCTCAGTTCGGATTGGAGTCTGC M.gateae... 14A consensus AATTCGACTCCATGAAGTCGGAATCGCTAGTAATCGCAGATCAGCTACGCTGCGGTGAAT M.gateae... 14A consensus ACGTTCTCGGGTCTTGTACACACCGCCCGTCACACCATGGGAGCTGGTAATACCCAAAGT M.gateae... 14A consensus CGGTTAGCTAACCTCGGAGGCGACCGCCTAAGGTAGGACTGGTGACTGGGGTGA M.gateae... 14A...

101 There were 94/111 samples where the colour change in the primary broth was consistent with that expected for the taxa identified by SSCP. For those samples where each of the five secondary broth samples demonstrated the same colour change (93/111), it was in agreement with the identification based on the SSCP result in all but one sample Analysis of species distribution among anatomical sites Where there were multiple mycoplasma-positive samples in an individual cat, the distribution and proportion of the two groups of mycoplasma species were considered. The majority of cats 32/39 (82.1%) had a single species/group of mycoplasma at multiple anatomic sites (Table 3.4). There were, however, a small number of cats where different species were present in different locations. There were no cats for which multiple species were identified at a single site (except the case of 106C described in sections and 3.3.4, which is not included in this table). Although the majority of cats had the same species at multiple anatomic sites, neither species was associated more frequently with any particular combination of sites. Fisher s exact test was used to compare the proportions of species for different combinations of sites where data were available (Tables 3.5 and 3.6). There was no significant difference in the proportions of M. felis and M. gateae/m. arginini across these combinations of anatomical locations (p = 0.29 and 0.34). M. felis was found more often than M. gateae/m. arginini in the conjunctiva and pharynx, whereas these species had a slightly higher prevalence than M. felis in the bronchus (Table 3.7 and Figure 3.9). The presence of mycoplasma at a particular site was not necessarily independent of other sites (because they were from the same cat). The random effects logistic regression shows that M. felis and M. gateae/m. arginini were found in similar proportions at each site. The proportion of observations that were positive for M. felis (compared with M. gateae/m. arginini) were not statistically different for the overall effect of site (Wald p = 0.62, Likelihood Ratio Test p = 0.61).. The proportion of observations that were M. felis (including the odds of this occurring at a particular site) were compared with M. gateae/m. arginini (Table 3.8). The odds 66

102 of a mycoplasma-positive conjunctival sample being M. felis as opposed to M. gateae/m. arginini, was 4.5 to 1. Similarly, a mycoplasma-positive pharyngeal sample had odds of 1.26 to 1 of being M. felis. Bronchial samples, however, were slightly more likely to be M. gateae/m. arginini than M. felis, with odds of 0.95 to 1. While there was not a significant difference in the ratio of odds from the logistic regression model, and therefore proportion of M. felis and M. gateae/m. arginini between anatomic sites, there appeared to be a trend (Table 3.9). Although not significant (p = 0.34), the biggest difference noted was the proportion of M. felis (compared with M. gateae/m. arginini) was greater in the conjunctiva compared with the bronchus. M. felis was also proportionately greater in the conjunctiva than the pharynx (p = 0.47). When also considering those cats that were negative for mycoplasma in the statistical analysis, McNemar s test was used to compare paired proportions of mycoplasmapositive cats among sites. This enabled the comparison of the likelihood of either any mycoplasma-positive sample, or a particular mycoplasma species between sites. The proportion of mycoplasma-positive samples was significantly greater in both the pharynx (p < 0.001, Table 3.10) and bronchus (p < 0.001, Table 3.11) compared to the conjunctiva. Similarly, the proportion of mycoplasma-positive samples was significantly greater in the pharynx compared with the bronchus (p = 0.002, Table 3.12). When species of mycoplasma was also considered, M. felis was found at a significantly higher proportion in the pharynx (p < 0.001, Table 3.13) and bronchus (p = 0.031, Table 3.14) compared with the conjunctiva and also in the pharynx when compared with the bronchus (p = 0.007, Table 3.15). M. gateae/m. arginini was also found at a significantly higher proportion in the pharynx (p < 0.001, Table 3.16) and bronchus (p < 0.001, Table 3.17) compared with the conjunctiva, but in similar proportions between the pharynx and bronchus (p = 0.17, Table 3.18). Although there are significant differences in the prevalence and distribution of mycoplasma between anatomic sites studied, the pattern of the anatomical distribution is consistent between the different species of Mycoplasma found. Comparatively, neither mycoplasma species had a significantly greater prevalence at a particular anatomical site than the other species. 67

103 Table 3.4: Numbers of cats that had the same mycoplasma species or different mycoplasma species among sites when multiple sites within a cat contained mycoplasma, where A = conjunctiva, B = pharynx, C = bronchus Combination of anatomic sites with positive Number of cats with same mycoplasma species at all sites Number of cats with different mycoplasma Total samples M. felis M.gateae / M.arginini species at different sites 2 sites AB BC CA (total) (12) (15) 3 sites ABC (total) (5) (0) Total

104 Table 3.5: Comparison of the proportion of each species between different combinations of positive sites using Fisher s exact test (where the same mycoplasma species was present at each site), where A = conjunctiva, B = pharynx, C = bronchus Positive sites M. felis M. gateae / Total P-value M. arginini A + B 3 (75.0%) 1 (25.0%) 4 B + C 9 (39.1%) 14 (60.9%) Table 3.6: Comparison of the proportion of each species between different combinations of positive sites using Fisher s exact test (where those with different mycoplasma species at each site were also considered in the overall proportion), where A = conjunctiva, B = pharynx, C = bronchus Positive Same species Different Total P-value sites M. felis M. gateae / M. arginini species A + B 3 (75.0%) 1 (25.0%) 0 4 B + C 9 (32.1%) 14 (50.0%) 5 (17.9%)

105 Table 3.7: Number and proportion of positive samples for each species of mycoplasma at each anatomic site sampled Anatomic site Species Total number of M. felis M. gateae/ M. arginini mycoplasma-positive samples (111) Conjunctiva (A) 9 (81.8%) 2 (18.2%) 11 Pharynx (B) 34 (55.7%) 27 (44.3%) 61 Bronchus (C) 19 (48.7%) 20 (51.3%) 39 Figure 3.9: Comparison of mycoplasma species distribution across the different anatomic locations studied Number of cats Conjunctiva (A) Pharynx (B) Bronchus (C) Anatomic site M.felis M.gateae/arginini

106 Table 3.8: Proportion of observations and odds at each site of a positive result for M. felis (F) compared to M. gateae/m. arginini (R). Site F/(F + R) % Odds Conjunctiva 9/ A Pharynx 34/ B Bronchus C 19/ Table 3.9: Comparison of likelihood of M. felis versus M. gateae/m. arginini at particular anatomic sites using the odds ratio, where A = conjunctiva, B = pharynx, C = bronchus Sites Odds ratio* 95% CI P-value B vs A C vs A C vs B *Odds ratio of 1 means they are the same

107 Table 3.10: Comparison of paired proportions for mycoplasma status overall between the conjunctiva and pharynx using McNemar s test. pharynx Myco + Myco - Total % positive conjunctiva Myco /110 (10%) Myco Total % positive 61/110 (55%) P-value < Table 3.11: Comparison of paired proportions for mycoplasma status overall between the conjunctiva and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive conjunctiva Myco /109 (10.1%) Myco Total % positive 39/109 P-value (35.8%) < 0.001

108 Table 3.12: Comparison of paired proportions for mycoplasma status overall between the pharynx and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive pharynx Myco /109 (55%) Myco Total % positive 39/109 P-value (35.8%) Table 3.13: Comparison of paired proportions for M. felis status between the conjunctiva and pharynx using McNemar s test. pharynx Myco + Myco - Total % positive conjunctiva Myco /110 (8.2%) Myco Total % positive 34/110 P-value (30.9%) < 0.001

109 Table 3.14: Comparison of paired proportions for M. felis status between the conjunctiva and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive conjunctiva Myco /109 (8.3%) Myco Total % positive 19/109 P-value (17.4%) Table 3.15: Comparison of paired proportions for M. felis status between the pharynx and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive pharynx Myco /109 (30.3%) Myco Total % positive 19/109 P-value (17.4%) 0.007

110 Table 3.16: Comparison of paired proportions for M. gateae/m. arginini status between the conjunctiva and pharynx using McNemar s test. pharynx Myco + Myco - Total % positive conjunctiva Myco /110 (1.8%) Myco Total % positive 27/110 P-value (24.5%) < Table 3.17: Comparison of paired proportions for M. gateae/m. arginini status between the conjunctiva and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive conjunctiva Myco /109 (1.8%) Myco Total % positive 20/109 P-value (18.3%) < 0.001

111 Table 3.18: Comparison of paired proportions for M. gateae/m. arginini status between the pharynx and bronchus using McNemar s test. bronchus Myco + Myco - Total % positive pharynx Myco /109 (24.8%) Myco Total % positive 20/109 P-value (18.3%) 0.17 Table 3.19: Number of each mycoplasma species at any site for both healthy and diseased cats. P-value represents a comparison of proportions of each mycoplasma species at any site between healthy and diseased cats using Fisher s exact test. M. felis M. gateae/ M. arginini Both Total P-value Healthy Diseased* Total *all had disease that was associated with the conjunctiva, or upper/lower respiratory tract as scored in data sheets described in section 2.2.1

112 Table 3.20: Comparison of proportion of each mycoplasma species present in conjunctival swabs between cats with and without conjunctivitis Species Negative Total P-value M. felis M. gateae/ M. arginini Ocular Yes Signs No Table 3.21: Comparison of proportion of each mycoplasma species present in pharyngeal swabs between cats with and without upper respiratory tract signs Species Negative Total P-value M. felis M. gateae/ M. arginini URT Yes signs No Table 3.22: Comparison of proportion of each mycoplasma species present in bronchial swabs between cats with and without lower respiratory tract signs Species Negative Total P-value M. felis M. gateae/ M. arginini LRT Yes signs No

113 3.3.7 Association between mycoplasma species and ocular/respiratory disease Neither M. felis nor M. gateae/m. arginini were found to be associated more frequently with disease at either all anatomic sites, or specifically in the anatomical location in which signs of disease were observed. Fisher s exact test was used to compare proportions of each species in healthy and diseased cats (Table 3.19), and specifically for each of the three anatomic locations (Tables 3.20 to 3.22). There was no evidence of either M. felis or M. gateae/m. arginini being associated with disease in the population of cats studied. The samples with variant M. felis profiles were compared with the health status of the six cats they were isolated from (Table 3.3). All were healthy cats with the exception of one cat. The variant F(b) isolates from this cat were from the pharynx and bronchus, and the cat had bilateral conjunctivitis with mucopurulent ocular discharge and mottling of both caudal lung lobes. 68

114 3.4 Discussion The results of the current study confirmed the high prevalence of mycoplasma in the conjunctiva, pharynx and bronchi of cats. The combined use of PCR and SSCP enabled a large number of samples to be rapidly screened for genetic variation and identity. Two distinct profiles (F and R) were readily identified by SSCP. These were representative isolates of M. felis, and M. gateae respectively. The aim of this study was to determine the species these isolates belonged to, rather than to demonstrate any genetic variability between mycoplasmas isolated from different populations or geographical locations. It was not possible to distinguish between M. gateae and M. arginini by analysing the region of the 16S rrna gene by PCR-SSCP in the current study. This was due to there being only one nucleotide difference between the two species in the 450 bp target region. Although this difference may be identified by sequencing, it may not always provide sufficient variability to see a different SSCP profile. All samples represented by profile R were therefore grouped into a single cohort, even though they may contain both M. gateae and M. arginini. Currently, there is a lack of evidence suggesting either M. gateae or M. arginini is associated with disease in cats (Heyward et al., 1969; Blackmore et al., 1971; Blackmore and Hill, 1973; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b). M. felis however, has been implicated as having a pathogenic role in both conjunctival and respiratory disease in cats (Cole et al., 1967; Schneck, 1972; Campbell et al., 1973b; Tan and Miles, 1973; Tan, 1974; Tan and Miles, 1974b; Haesebrouck et al., 1991; Randolph et al., 1993). The main consideration was to separate M. felis from other non-pathogenic species to determine whether there was any evidence of it being associated with disease. M. gateae/m. arginini therefore provided a non-pathogenic group for comparison with M. felis. The current study showed M. felis had a slightly higher prevalence within total samples and number of cats than the non-pathogenic group M. gateae/m. arginini. This supports similar findings in feline populations from studies in New Zealand (Tan and Miles, 1974b; Tan et al., 1977a), England (Blackmore et al., 1971) and the USA (Heyward et al., 1969). 69

115 The criteria for determining the presence of multiple mycoplasma species in a sample (representing one anatomical site) is based on identification of multiple SSCP profiles in that sample. One sample in this study (106C) had more than one species of mycoplasma evident at a single anatomical site, with a different SSCP profile from the single colony and the pooled colony sample. This suggests one anatomical site contained two different mycoplasma species, which was atypical in this population of cats. This is in contrast to similar feline population studies where multiple mycoplasma species were commonly isolated from swabs (Heyward et al., 1969; Blackmore et al., 1971; Hill, 1971; Tan and Miles, 1974b; Tan et al., 1977a). Although this suggests there were multiple species isolated from that site, a mixed SSCP profile should also have been evident in the pooled colony sample, as the single colony isolate is represented in the pooled sample. However, neither this sample nor any others had a combination of two distinct SSCP profiles, which would indicate multiple species within the sample. Where there were multiple anatomic sites in a cat with mycoplasma, the same species/group of mycoplasma was isolated from each site in the vast majority of cases in this study. This occurrence was likely due to the anatomical or functional relationship between sites discussed in Chapter 2. There was no significant site predilection for either M. felis or M. gateae/m. arginini when compared at any one site (conjunctiva, pharynx or bronchi) or any combination of sites. Although not statistically analysed in previous studies, M. gateae tended to be slightly more prevalent than M. felis and M. arginini in the pharynx (Heyward et al., 1969; Tan and Miles, 1974b; Tan et al., 1977a), whereas M. felis was more prevalent in the conjunctiva (Blackmore et al., 1971; Tan and Miles, 1974b). However, in the present study, M. felis was slightly more prevalent than M. gateae/m. arginini in the overall number of samples and cats. This was also true for the conjunctiva and pharynx, although the differences were not significant when proportions of each species between sites were compared. This difference was most apparent in the conjunctiva, where the odds of a mycoplasma-positive conjunctival sample being M. felis, as opposed to M. gateae/m. arginini, was 4.5 to 1. This is supported by studies 70

116 where M. felis is the species of mycoplasma isolated from the conjunctiva most commonly (Blackmore et al., 1971; Tan and Miles, 1974b). Similar to the results reported in Chapter 2, there was a statistically significant difference in the prevalence of mycoplasma between sites, as confirmed by PCR. In these circumstances, the pharynx was by far the most common site from which mycoplasma was isolated, followed by the bronchus and then the conjunctiva. When the data were examined for each mycoplasma species separately, the same significant relationships existed among sites. That is, both M. felis and M. gateae/m. arginini were recovered significantly more frequently from the pharynx than the conjunctiva, and also from the bronchus than the conjunctiva. The exception was that M. gateae/m. arginini had no statistically significant difference in prevalence between the pharynx and bronchus as M. felis did. Although this demonstrates that both species have a predilection for particular anatomical sites, the two species were not significantly different to each other. Neither species was comparatively more prevalent at any one site. Therefore, there are no discernable differences in the ecology or relative preference for site by either species. The regular isolation of both M. felis, and M. gateae/m. arginini from the bronchus was interesting. Given the careful method of sampling, it is unlikely this was the result of contamination. As the samples were taken following euthanasia, it is possible that pharyngeal mucous moved (post mortem) into the trachea, and subsequently the bronchus from either agonal inhalation or during subsequent movement of the bodies. Evaluation of post mortem microbiological examination and the significance of such factors as bacterial invasion or translocation are inconclusive (reviewed in (du Moulin and Love, 1988; Tsokos and Puschel, 2001; Morris et al., 2006)), but do discuss important factors for obtaining representative samples such as time interval from death to post mortem, storage temperature, technique and also interpretation of culture results. In this study, routine bacterial culture may have provided some insight as to whether the mycoplasmas isolated from the bronchi originated from the oropharynx, based on the concurrent presence of a mixed flora of resident oropharyngeal organisms. 71

117 Further evidence for these being meaningful isolates comes from the fact that in five of the cats in the current study, mycoplasma was isolated from the bronchus but not from the pharynx. From four others, a different species was isolated from the bronchus compared with the pharynx, indicating they had not originated from this location. These findings are supported by a study where 35% of healthy cats had mycoplasma isolated from the trachea (Heyward et al., 1969). Each of these 10 tracheal isolates were found to be M. gateae (Heyward et al., 1969), whereas the current study demonstrated an equivalent prevalence of M. felis and M. gateae/m. arginini from the bronchial isolates. This is the first study to demonstrate the isolation of M. felis from the lower respiratory tract of a population of healthy cats. Differentiation of the mycoplasmas into species/groups did not demonstrate any statistical association with ocular or respiratory disease. This was true both overall (for any sign of disease and isolation of that species from any site), and specific to the isolation of each species from the location of the signs of disease. M. felis was isolated from eight cats with no sign of conjunctivitis compared with none of the cats with conjunctivitis. This contrasts a number of studies where M. felis had been isolated from the conjunctiva in greater numbers in cats with conjunctivitis or upper respiratory tract disease than without (Tan and Miles, 1973, 1974b; Tan et al., 1977a; Shewen et al., 1980; Haesebrouck et al., 1991; Bannasch and Foley, 2005; Low et al., 2007), implying that it may cause the disease. However, it was consistent with findings from other studies that have found M. felis in similar proportions in cats with and without conjunctivitis (Blackmore et al., 1971). M. felis isolated from the conjunctiva of cats in the current study was not associated with conjunctivitis. By studying larger numbers of cats or more cats with conjunctivitis, associations may be found to exist. The M. felis variant F(b) was isolated from the pharynx and bronchus of one cat with bilateral conjunctivitis and mucopurulent ocular discharge. Although no conclusions can be drawn from this one instance, in future it may be informative to further analyse isolates such as this one containing nucleotide variation. Further characterisation of the genotypic and phenotypic features of the organism, and experimental transmission 72

118 studies may assist to determine if the nucleotide variation contributes to any difference in virulence of these isolates. Compared to each other, M. felis and M. gateae/m. arginini both followed the same pattern of prevalence and distribution across anatomical sites. In addition, the lack of association between either group and the presence of disease in each of these sites (conjunctiva, pharynx and bronchus), suggests they are both commensal organisms in these sites. The results of the current study are strongly suggestive of neither M. felis nor M. gateae/m. arginini being significantly associated with ocular or respiratory disease in cats. It also suggests that M. felis and M. gateae/m. arginini could be considered as inhabitants of the ocular and respiratory mucous membrane environments. Technical considerations of PCR-SSCP From the initial provisional identification of mycoplasma colonies on solid media from 134/330 samples, 129 of these were PCR positive using the primers developed for SSCP analysis (FD1000 and Myco R). Of these, 111 were positive by SSCP analysis. Given what was considered to be optimal conditions for PCR, some mycoplasma appeared to be unable to be amplified, and furthermore of those that were amplified by PCR several were unable to be analysed by SSCP. This discrepancy may be due to bacteria identified as mycoplasmas by colony morphology which did not belong to the genus. It would be unusual, given the characteristic appearance of mycoplasma colonies, and the use of selective mycoplasma media, that these were mistaken for microorganisms other than mycoplasma. The second possibility is that they were mycoplasmas, but not mycoplasma species that were capable of being amplified by these primers. Given the current knowledge of mycoplasmas in the cat, this is unlikely. However, much of this knowledge was acquired prior to the advent of molecular technologies and may be incomplete. Thirdly, and most likely is that the negatives were false negatives due to technical error. There may be similar explanations for the loss of samples from PCR to SSCP, and it seems that these samples that were negative were possibly due to technical error. 73

119 It was noticed during the course of these experiments that the sample DNA appeared to degrade over time during storage at -70 ºC. This could possibly have been due to contamination with DNAses at the time of DNA isolation although no definitive study was performed to prove this. Therefore it may be prudent to perform PCR soon after sample preparation. The effect of temperature and specifically the degradation was not tested in this study, and would need to be established to determine the most appropriate method for sample handling. Alternatively, there may have been the carry over of inhibitors of PCR following genomic DNA isolation. Factors inhibitory to the PCR have been reported in previous studies, an example being those found in saliva (Ochert et al., 1994). In this study, inhibition of the PCR was demonstrated by addition of varying concentrations of mycoplasma media (see section 3.2.4). The heat lysis method of DNA extraction used for the experiments yielded equivalent results to a commercial kit column purification method during PCR optimisation (Appendix 7). Based on this it was assumed the heat lysis method was sufficient to overcome inhibitory factors from the sample itself or the mycoplasma media. However, this finding was based on subjective visual assessment of PCR product on an agarose gel. Quantification of levels of PCR product from a constant amount of genomic DNA by quantitative (real-time) PCR would be a more sensitive method to assess inhibition and would be beneficial in future studies. It is possible that the results of this study underestimate the true prevalence of mycoplasmas in the feline populations studied. In addition to the loss of samples (described above) from the original number of culture-positive samples, mycoplasmas have been isolated from up to 20% of liquid media which had not shown any obvious change in ph within 2 weeks (Hill, 1971). Colour change of liquid media was used as a method of screening large numbers of samples in the current study, being simple and cost effective, and only those eliciting such change were plated on to solid media to confirm the presence of mycoplasma colonies. However, the findings of Hill (1971) indicate that colour change may be a relatively insensitive method for determining the presence of mycoplasma, and implies the possibility of false negative results in the present study. This problem may be overcome by amplifying genomic DNA directly 74

120 from swab samples. This method has been demonstrated to be equivalent in sensitivity to culture (Johnson et al., 2004; Veir et al., 2008). Some samples in this study, whether single colony or pooled, contained one or two additional bands in their SSCP profile in addition to a typical F profile. The presence of multiple bands, representing unique conformers of the sense and antisense strands due to sequence variation in SSCP profiles, is a common finding in other organisms (Gasser et al., 2006; Jex et al., 2007) and has been demonstrated in the vlha gene of M. synoviae (Jeffery et al., 2007). The additional bands seen in the current study most likely represent minor nucleotide variation within species, rather than two separate species, either as a result of the presence of natural variation, strain differences, or infidelity of PCR. This was confirmed by nucleotide sequencing of representatives of these variant-type profiles, which were demonstrated to be M. felis. The M. felis variants sequenced had 98% similarity to the reference strain of M. felis, and contained 3 nucleotide differences in this region. Analysis of variation across the entire 16S rrna gene or genome was beyond the scope of this study. Although exact reasons for this variability remain uncertain, it confirms the usefulness of the SSCP technique to detect small nucleotide differences. To resolve the uncertainty of what the extra bands represented in this study, it would be necessary to excise them from the gels and sequence them (Gasser, 2001). Comparison with the sequencing reactions on the original broth culture samples corresponding to the SSCP sample could then be made to determine the origin of the sequence variability. In the current study no sample showed an SSCP profile consistent with the presence of more than one mycoplasma species. As SSCP has been demonstrated to detect multiple species or strains in a single sample based on their differing nucleotide sequence and hence different secondary structure (Gasser et al., 2003; Chalmers et al., 2005), it would be expected that a mixture of species should have been evident. It is possible that the five mycoplasma colonies selected for each sample (section 2.2.2) were all one mycoplasma species even though multiple species may have been present. The probability of selecting five colonies of one species from a plate containing two or more species cannot be estimated without already knowing the prevalence of each species in the population (personal communication, G. A. 75

121 Anderson (BAgrSc(Hons) GradDipApplStats MVSc), 2007). This is considered unlikely, as colonies with differing morphology were not evident, which would be expected if different species were present. Other possibilities include only one mycoplasma species present in the site sampled, more than one species but one present in greater numbers, or that one particular mycoplasma species was overgrowing the others in the primary broths due to preferential culture conditions. Although liquid media for the current study was suitable for the growth of M. felis, evidence does not suggest that other feline species have different requirements (Rosendal, 1979; Whitford et al., 1994). If this were the case, these studies should have shown a much greater prevalence of M. felis in comparison with the other species, which did not occur. However, this information relies on characterisation of mycoplasma isolates by culture, biochemical and serologic methods prior to the availability of molecular diagnostic techniques. This may not account for species that have been inadvertently selected against in due to a requirement for different culture requirements or growth conditions. The potential for such organisms has not been excluded in the population of cats in the current study. A comparison of mycoplasma PCR from direct swab samples in addition to cultured swab samples might resolve this issue. It is possible that the five mycoplasma colonies selected for each sample (section 2.2.2) were all one mycoplasma species even though multiple species may have been present. The probability of selecting five colonies of one species from a plate containing two or more species cannot be estimated without already knowing the prevalence of each species in the population (personal communication, G. A. Anderson (BAgrSc(Hons) GradDipApplStats MVSc), 2007). This is considered unlikely, as colonies with differing morphology were not evident, which would be expected if different species were present. Other possibilities include only one mycoplasma species present in the site sampled, more than one species but one present in greater numbers, or that one particular mycoplasma species was overgrowing the others in the primary broths due to preferential culture conditions. Although liquid media for the current study was suitable for the growth of M. felis, evidence does not suggest that other feline species have differing requirements (Rosendal, 1979; Whitford et al., 1994). If this were the case, there should have been 76

122 a much greater prevalence of this organism in comparison with the others, which did not occur. There were 18 samples for which the five secondary broth samples (each grown from a single mycoplasma colony taken from the primary plates) did not produce the same colour change as the primary broth. This may have occurred because there were multiple species present on the plates. It is possible that the differing colony morphology was not always evident when colonies were selected, as it can sometimes take 7-14 days for such morphological characteristics to become apparent (Hill, 1971). However, if this were the case, then multiple species should have been evident as mixed SSCP profiles, as discussed previously. Other possible reasons for this may be that colour change is an insensitive method for determining species, and a single species may induce either colour change, despite their preferred energy utilisation. This would need to be tested further to determine whether this does in fact occur. Without further evidence, it would seem that direct PCR on swabs with universal primers should circumvent any problems that may occur in relation to the mycoplasma population changing as a result of culture conditions or competition. The species of mycoplasma determined from SSCP profile results agreed with that expected from the colour change of the single colony secondary broth samples (there was only one sample that differed in this respect). However the colour change recorded from the primary broths was not consistent with that seen in the secondary broths in 13 samples. It appears that either colour change is insensitive as a means of presumptive species identification, or there is a shift in the mycoplasma population occurring in the transfer of primary liquid cultures to plates and then back to broths. The colour change is less important when molecular techniques are used as the means of mycoplasma identification. However, to avoid the possibility of population shifts occurring prior to PCR and SSCP, thereby influencing the study results, it would be prudent to use direct swabs for PCR, or to directly plate the swabs on solid media, in addition to using liquid cultures, to obtain larger amounts of DNA from single, representative colonies. 77

123 Differentiation of M. gateae and M. arginini M. gateae and M. arginini have very similar colony morphology, biochemical characteristics, some serological cross-reactivity, and 98% sequence similarity between them in the entire 16S rrna gene (Hill, 1971; Brown et al., 1995). Using molecular methods based upon the 16S rrna gene, it was not possible in this study to distinguish between these two species of mycoplasma. Although the mutation detection level for SSCP can be as sensitive as 100% for sequences of bp (Gasser, 2001), it has also been reported that the ability to differentiate to this level of variability may be reduced for sequences of greater than 200 bp (Gasser et al., 2006). The relative change in migration of a larger strand of DNA with only a small change in structure would presumably be less, and hence the ability to detect such a difference, or the sensitivity would consequently be reduced (reviewed in (Gasser et al., 2006)). Despite this, SSCP has been demonstrated to detect differences of 1 nucleotide in sequences of 400 bp (Jeffery et al., 2007) and 530 bp (Zhu and Gasser, 1998). However, the sensitivity of SSCP to determine such small nucleotide differences should be dependent on the exact location of the mutation and its effect on the resultant structure of the single-stranded DNA, and consequently its rate of migration through a non-denaturing gel (reviewed in (Gasser et al., 2006)). Further analysis of SSCP using reference samples of all feline mycoplasma species, and PCR amplification of different genes other than 16S rrna, may improve the specificity of SSCP to differentiate unknown mycoplasma species. The current study demonstrated that for a rapid test that clearly differentiates between mycoplasma species rather than study variation within one species, a set of primers that amplifies a region of the 16S rrna gene with greater variability than one nucleotide between species is required. The 16S-23S intergenic spacer (IGS) region of the rrna gene may prove to be a suitable region, due to it being less conserved that the adjacent 16S and 23S regions. The length of the 16S-23S IGS varies between bp in different mycoplasma species (Harasawa, 1999); this size should be suitable for SSCP (reviewed in (Gasser et al., 2006)). Intergenic spacer regions of ribosomal genes have been successfully utilised in other organisms for SSCP analysis (Gasser et al., 2003). A PCR for the detection of M. felis was developed using this region due to its species specificity, and 78

124 found it to be highly conserved between different isolates of M. felis (Chalker et al., 2004). In this same study, an alignment between the IGS of M. gateae and M. arginini contained only 3 nucleotide differences and a possible further 6 positions of variability due to deletions/insertions (Chalker et al., 2004). It remains to be determined whether universal primers in a conserved region of 16S-23S IGS region can be designed to PCR-amplify a small, sufficiently variable region from any mycoplasma species to readily differentiate between species by SSCP. The present study demonstrated that a large number of samples can be rapidly and easily screened by SSCP. This method allowed sequence variation of the samples to be displayed both within and between species. The variant M. felis profiles in this study showed that sequence variation within a species is represented by SSCP. Although not demonstrated in this study, SSCP may detect multiple species of mycoplasma in a single sample. This approach would also detect variants or previously uncharacterised species, eliminating the need to culture the organisms. Ultimately, using suitable target DNA regions, PCR-SSCP should provide a rapid and accurate tool for the identification of mycoplasmas to the species level compared with currently used methods. 79

125 80

126 Chapter 4 General Discussion and Conclusions The current thesis provides the first in depth investigation of the molecular identification and prevalence of mycoplasmas in the respiratory tract of a defined population of shelter cats. The prevalence of mycoplasma in more than 70% of the population of shelter cats investigated in this thesis was similar to studies from many parts of the world where cats are kept as companion animals (Heyward et al., 1969; Blackmore et al., 1971; Schneck, 1973; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b). Although mycoplasma has been isolated from and studied in cats from Australia previously (Jones and Sabine, 1970; Spradbrow et al., 1970b; Malik et al., 1991; Foster et al., 1998; Foster et al., 2004a; Foster et al., 2004b; Foster et al., 2004c), the prevalence of these organisms had not been determined, and only a single isolate from these studies had been identified to the species level (M. felis, (Hooper et al., 1985)). Isolates of Mycoplasma felis and Mycoplasma gateae were identified by nucleotide sequencing in the feline population examined in the current study, and these two species were the most frequently recovered from cats worldwide (Heyward et al., 1969; Blackmore et al., 1971; Tan and Miles, 1974b; Tan et al., 1977a). Differentiating M. gateae from M. arginini has, in the past, been problematic due to their morphological, biochemical, serological and sequence similarities (Heyward et al., 1969; Hill, 1971; Rosendal, 1979; Razin and Freundt, 1984; Whitford et al., 1994). Such difficulties were also encountered in the current thesis, as both were 99% similar in the 16S rrna gene sequence. It was not possible to unequivocally distinguish them using the current molecular approach. This may have been avoided during selection of primers to distinguish between the common feline mycoplasma species. However, with further characterisation of the mycoplasma genome, a shorter or more suitable region (with more sequence variation among species) could be used for PCR-SSCP analysis. The current thesis confirmed, through rigorous statistical analysis, the role of M. gateae/m. arginini in domestic cats as commensal organisms of the conjunctiva, oropharynx and respiratory tract. M. gateae/m. arginini were found in equivalent numbers from both healthy and diseased cats. There was no association between the 81

127 presence of these organisms and disease found in the current thesis. Previous studies have demonstrated that these organisms rarely induce a detectable immune response in cats (Blackmore and Hill, 1973; Tan and Miles, 1974b; Tan et al., 1977a; Tan et al., 1977b). Thus, it would seem they are not pathogenic, remain on the mucosa, rarely crossing it and thus do not stimulate an immune response. M. felis did not significantly differ in its frequency of isolation or anatomical distribution from the non-pathogenic group of M. gateae/m. arginini. This, combined with no evidence of disease association with M. felis, suggested it also is a commensal organism of the feline conjunctiva and respiratory tract. For the first time, the current thesis has shown that M. felis is a common inhabitant of the lower respiratory tract in healthy cats. Although M. felis mimicked the non-pathogenic group in this study with respect to the lack of any significant association with clinical disease, there is serological evidence from previous literature that M. felis is recognised by the immune system (Tan, 1974; Tan and Miles, 1974b; Tan et al., 1977a). The findings of the current thesis highlight the need to re-examine evidence that M. felis causes conjunctivitis and possibly also pneumonia. There are some studies that are indicative of the presence of M. felis being associated with disease (Cole et al., 1967; Campbell et al., 1973b; Shewen et al., 1980; Haesebrouck et al., 1991; Randolph et al., 1993), but these studies do not unequivocally demonstrate cause and effect of a pathogenic nature of these microorganisms. Furthermore, there are several supportive studies showing that M. felis has no association with disease, either as a result of its frequent isolation from healthy cats (Heyward et al., 1969; Blackmore et al., 1971) or failing to consistently induce disease experimentally (Cello, 1957; Blackmore and Hill, 1973; Lappin et al., 2007). M. felis has been associated with disease from cats that are immunosuppressed or have underlying diseases (Cello, 1957; Hooper et al., 1985), implying an opportunistic role of this microorganism in inducing disease. There has been a single study with substantial evidence for pathogenicity by M. felis (Tan, 1974). In this study, conjunctivitis was observed following the instillation of a strain of M. felis from a kitten with conjunctivitis into the eyes and nose of healthy kittens. Serum antibodies to M. felis were detected in these kittens following 82

128 inoculation. The absence in these kittens of other bacteria and viruses, the development of disease post inoculation, the organism being re-isolated from the kittens, and the serological response provided evidence for pathogenicity. However, the control kittens in the study also produced a low-level antibody response to M. felis (likely due to transmission from the inoculated cats in adjacent cages), but no clinical disease was evident in the control kittens; thus, natural transfer of M. felis did not produce clinical disease in healthy kittens. The results of the present thesis also suggest that M. felis, in the natural state, is not sufficiently virulent to delineate clinical disease caused by this organism from other causes of upper respiratory tract and ocular diseases. In contrast to the study by Tan (1974), Blackmore and Hill (1973) did not find a link between M. felis and overt clinical disease or seroconversion. This may indicate that there were other confounding factors in the study by Tan (1974) and may explain why there was a lack of disease induced in SPF cats, since in this one study where severe disease was produced, the cats were not SPF. Disease might be the result of coinfection with other known pathogens which breach the mucosa or produce mucosal damage permitting exposure of mycoplasma to the immune system, in spite of the authors comment that the cats were tested for non-specified viruses and bacteria. In conclusion, M. felis may be pathogenic in the broad sense, but is in all probability of very low virulence. Although no evidence of variation in virulence between different isolates or strains of M. felis has yet been demonstrated, the variation in disease and the seroconversion of cats experimentally infected with M. felis suggest that different strains may occur. It may be that there are different strains within each of the currently described mycoplasma species with differing levels of virulence, such as those of isolates of M. hyopneumoniae known to infect pigs (Meyns et al., 2007). This proposal may account for some differences in findings among studies in which no association of mycoplasma and disease was apparent. However, there are other possible reasons for these discrepancies that have not yet been addressed, the most important of which is the presence of other ocular or respiratory pathogens. Ideally, to definitively determine a pathogenic role of feline mycoplasmas, disease should be consistently produced by experimental transmission of suspected virulent 83

129 M. felis isolated from diseased cats into SPF cats. In addition, an unequivocal increase in antibody titres in paired samples would also provide evidence for infection. The multifactorial nature of mycoplasmal disease in species other than cats, and the multiagent nature of respiratory disease make such studies difficult. Even for species that are considered primary pathogens, such as M. mycoides subsp. mycoides, as the causative agent of Contagious Bovine Pleuropneumonia, Koch s postulates have not been fulfilled, and many species that are accepted as pathogens can also be found in apparently healthy hosts (reviewed in (Cassell et al., 1985)). With such limited knowledge of the factors involved, such as structural and functional components or genetic makeup that determine pathogenicity of mycoplasmas, the specific ways these organisms interact with the host immune system and other environmental influences, it is incredibly difficult to conclusively determine a defined role in disease. As the species of mycoplasma isolated in the current thesis were present in the conjunctiva, upper and lower respiratory tract, it would reasonable to expect to encounter disease associated with its presence in these locations if these organisms were sufficiently virulent. As there was no clinical or post mortem evidence of ocular or respiratory disease being significantly associated with the isolation of mycoplasmas at the respective anatomical sites, it appears that the species present in this population of cats were not pathogenic. However the possibility that some of these isolates may be pathogenic to particular animals in other situations cannot be excluded. Serological methods may have been useful to assist the determination of any past or present evidence of an immune response to these organisms. The majority of cats in the current thesis were adult, which may bias results given initial infection usually occurs in young cats during first contact with mycoplasma and may result in clinical signs of disease. These young individuals may recover and become carriers of the organism which are then commensal organisms with the potential to become opportunistic invaders under favourable conditions. It is possible that both groups of mycoplasma species isolated in the current thesis may be opportunistic pathogens, which under certain conditions, such as damage to the respiratory epithelium by other respiratory pathogens, stress or other disease leading to immune compromise, they may further contribute to disease or initiate damage or disease themselves as opportunists. Currently, there is no direct evidence 84

130 for such occurrences. There is some evidence from both case reports (Campbell et al., 1973b; Moise et al., 1983; Hooper et al., 1985; Malik et al., 1991; Walker et al., 1995; Foster et al., 1998; Chandler and Lappin, 2002; Barrs et al., 2005; Liehmann et al., 2006) and experimental transmission studies (Cello, 1957; Moise et al., 1983) where the presence of immune compromise or underlying disease is frequently associated with disease involving mycoplasma in cats. Therefore, it would appear to be more a lack of host defences than virulent features of the organisms allowing this opportunism. There is scant evidence showing mycoplasma is associated with disease in conjunction with other ocular or respiratory pathogens, as they have not often been concurrently recovered, and because as clearly demonstrated in the current thesis, the presence of mycoplasma does not indicate involvement in disease. The current thesis found no association of mycoplasma with ocular or respiratory disease in the study population. Greater study population numbers may enhance subtle differences; however this study of 110 cats showed no significant differences. The group numbers are considered sufficient to show any major or obvious associations, such as primary pathogenicity. However the small number of juvenile animals and diseased cats in the present study may be a contributing factor to some of the contrasting results of previous studies (e.g., (Cole et al., 1967; Tan and Markham, 1971b; Campbell et al., 1973b; Tan, 1974; Tan and Miles, 1974b; Haesebrouck et al., 1991; Randolph et al., 1993; Foster et al., 2004b; Bannasch and Foley, 2005; Low et al., 2007)). Different sampling methods and locations may account for some of the discrepencies between studies. Differences may also be apparent at other times of the year, between years, or where relationships with other respiratory pathogens or viruses affecting immune function exist. Further work may be indicated to define these variables. The use of PCR-SSCP as a means for rapidly screening large numbers of mycoplasma samples for genetic variation was demonstrated in the current thesis. With further validation and standardisation SSCP could provide clear advantages over currently used techniques of identification, being more rapid than culture to perform, and being able to detect organisms that may be slow growing or difficult to recover via culture. PCR and SSCP have the potential to overcome some of the limitations of serological methods of identification such as serological cross reactivity and antigenic variation 85

131 among mycoplasma species (Heyward et al., 1969; Brown et al., 1995), and eliminate the requirement for experimental animals to produce antisera. Additionally, they have the potential for applications beyond the identification of mycoplasmas. The technique could be used to detect multiple species present in samples which has been demonstrated previously for Cryptosporidium (Chalmers et al., 2005). It has also been utilised to genetically characterise regions associated with virulence, such as the vlha gene in M. synoviae (Jeffery et al., 2007). It may be more accurate and rapid than approaches such as PCR-RFLP as one gel can determine species present in a range of samples, compared with performing an algorithm of digests. In addition, sequence variation can be detected across a region of 100 to 500 bp, compared with only a few nucleotides of the endonuclease cleavage regions. Further sequencing and analysis of mycoplasma genomes is required to identify DNA targets better suited for use in PCR-SSCP analysis for species identification or identification of virulence factors in isolates. A major strength of the current thesis in clarifying the role of mycoplasmas in feline ocular and respiratory disease is the use of rigorous statistical analysis. These analyses comparing M. felis with the non-pathogenic species of mycoplasma demonstrated no significant difference in prevalence, distribution, or association with disease between groups, and provided evidence to support the proposal that they are not pathogenic under normal circumstances. Additionally, this study demonstrated important ecological characteristics of the organisms in the statistical comparison of mycoplasma species and anatomical sites they were isolated from. It clearly demonstrates that the isolation of mycoplasma from the conjunctiva or any part of the respiratory tract of a cat is not sufficient evidence to implicate a causative role in disease. In contrast, a number of studies have demonstrated a positive association between the isolation of mycoplasmas with ocular and respiratory disease (Randolph et al., 1993; Bannasch and Foley, 2005; Low et al., 2007), whereas previous reports of the prevalence of mycoplasma in feline population studies have been limited in their ability to make conclusions as to the significance of mycoplasma isolation by the lack of statistical analysis in determining any association with disease (Blackmore et al., 1971; Schneck, 1973; Tan and Miles, 1974b; Tan et al., 1977b; Haesebrouck et al., 1991). 86

132 The findings of the current thesis make an important contribution to both the overall epidemiological context of feline mycoplasmas, and also the means to identify them to the species level using PCR-based techniques. With continuing research into mycoplasmas of cats and other species, more detail will emerge as to the specific mechanisms by which mycoplasmas interact with their hosts and the conditions under which disease may occur. Future research should focus on establishing any direct evidence for pathogenicity in mycoplasmas isolated from cats in which they are thought to relate to disease. Isolates should then be investigated by performing experimental transmission studies in SPF cats. For any species or strains for which such pathogenic potential is evident, additional studies should be undertaken to characterise, compare and contrast the genetic and functional makeup of these organisms to better understand the genetic correlates of virulence. It has been suggested that the role of mycoplasmas in disease may well be influenced by a deficiency in the host s defence mechanism more so than the attributes of the organism (Whitford and Lingsweiler, 1994). Therefore, it may be equally important to determine the place of mycoplasma in the complex aetiology of respiratory disease, such that it can be determined when these organisms change from being commensals to opportunistic pathogens. This is not simple due to the multitude of factors that may influence such a process in any individual. The value of the current thesis is that it shows mycoplasmas are often present in the lower respiratory tract, which should initiate future case studies of feline respiratory disease to genetically characterise the mycoplasma isolated beyond the species level. Furthermore, it may be possible, using cats with resident mycoplasmal flora in the lower respiratory tract, to introduce respiratory viruses, and measure mycoplasma numbers over the course of disease, while monitoring serological responses. There may also be benefits in determining whether viruses, such as feline immunodeficiency virus (FIV) and feline leukaemia virus (FeLV), by reducing immune competence, influence mycoplasma numbers or behaviour in their hosts. Such studies may not only provide new information regarding the way mycoplasmas interact with their hosts, but aid the clinical management of such diseases. Perhaps a true symbiotic relationship exists between some species of mycoplasma such as M. gateae and M. arginini and their hosts. The benefits of such an association 87

133 for the mycoplasmas are obvious; warmth, protection and nutrition for organisms that, without cell walls are sensitive to changes in environmental conditions, and with a relatively limited capacity for synthesis of nutrients. However, for the hosts, mycoplasmas may provide a degree of protection to the host from other pathogenic organisms, which must compete for attachment surfaces and nutrients with the established resident population. In the least, it may be a commensal relationship, advantageous for the organism, which in turn does not cause undue harm to the host, and perhaps does not come into close contact with the immune system. Even for other resident mycoplasma species, such as M. felis, which may breach the mucosa and come into contact with the immune system, this low level pathogenicity may not be sufficient to cause clinical disease to the host under normal circumstances. This would reinforce the true opportunistic nature of these organisms as pathogens, as there must be other factors involved for disease to occur. Relatively few of the mycoplasma species characterised to date are primary pathogens. It may only be those mycoplasma species that have not evolved in association with that particular host to the same degree as commensal organisms that are the pathogens, with selective pressures reducing or eliminating virulence factors over time. It seems possible that periodically, more virulent mycoplasma that possess pathogenic characteristics emerge, being associated with disease. However, the vast majority of mycoplasmas appear to remain as minimally-virulent, commensal organisms. Future studies are needed to investigate pathogenic isolates to elucidate the true role of mycoplasma infection in cats. 88

134 Chapter 5 Appendices 5.1 Appendix 1: Data Collection Sheet 89

135 5.2 Appendix 2: Liquid mycoplasma media 90

136 5.3 Appendix 3: Mycoplasma agar formulation 91