Development of multi-parametric tests for the diagnosis of feline tuberculosis

Abstract

Mycobacterial infections cause a substantial burden of disease in both human and animal populations and continue to provide a challenge in terms of both their diagnosis and treatment. Feline mycobacterial infections are increasingly recognised by veterinary staff but obtaining a rapid and reliable diagnosis in these cases remains problematic. Tuberculosis (TB) is the most frequently diagnosed mycobacterial disease in cats in Great Britain, where it is caused by infection with either Mycobacterium (M.) bovis or M. microti. Non-tuberculous mycobacteria also cause disease in cats, but a species-level diagnosis of mycobacterial infection cannot be accomplished on clinical signs alone. Accurately diagnosing these infections is of importance so that appropriate treatment can be initiated, and to ascertain the risk posed to humans due to the zoonotic potential of some species of mycobacteria, namely M. bovis. Specialist mycobacterial culture is limited by the fact it can take several months to attain a positive result, and that some species of mycobacteria are near-impossible to grow in the laboratory. Commercially available polymerase chain reaction (PCR) assays can be financially restrictive, especially for obtaining a species-level diagnosis in cases of TB, and extracting sufficient high-quality deoxyribonucleic acid (DNA) from formalin-fixed paraffin-embedded (FFPE) tissue samples for PCR testing can be a limiting factor. Finally, while the feline-specific interferon-gamma (IFNγ) release assay (IGRA) has been in use for over six years, there is a lack of data assessing the performance of this test. Therefore, there is plentiful scope for the advancement of diagnostics for cases of feline mycobacteriosis, especially feline TB, which would be of benefit for both cats and humans. Ocular TB (OTB) is an unusual presentation of disease in humans and remains controversial in terms of its pathogenesis and treatment; OTB has also been reported in cats. Given this, it was hypothesised that the cat may provide insights into human OTB, but further work was needed to characterise feline ocular mycobacterial lesions. This study characterised the histopathological changes present in 24 cases of feline ocular mycobacteriosis, and described the immune cells present using immunohistochemistry. Granulomatous to pyogranulomatous choroidal lesions with concurrent retinitis and retinal detachment was the most common finding, but many choroidal lesions also showed infiltration with a substantial number of B-lymphocytes. Mycobacteria were identified in 83% of globes, sometimes in large numbers, in contrast with data from humans. Histopathology is often the first diagnostic step performed that raises the suspicion of mycobacterial infection in cats. However, descriptions of the histopathological features of these lesions and characterisation of the cell populations present are lacking, particularly for cases of TB. One aim of this study was to identify whether differences in the appearance and composition of feline (muco)cutaneous and lymph node TB lesions could discern between infection with M. bovis or M. microti. A novel, descriptive scoring system was devised which suggested that the type of granuloma present differed between cases of M. bovis and M. microti, and this may provide a means of differentiating between these two aetiological agents at an earlier stage of the diagnostic investigation. It was also shown that cases of TB can present with abundant acid-fast bacilli (AFB) on Ziehl-Neelsen staining; therefore, identification of large numbers of AFB does not rule out infection with tuberculous pathogens. The next aim of this study was to assess the performance of tests of cell-mediated immunity, namely the IGRA, for the diagnosis of feline mycobacteriosis as well as the utility of the IGRA for monitoring treatment responses. Receiver-operating characteristic curve analysis suggested reducing test cut-off values that would improve sensitivity of the IGRA, in particular for cases of M. bovis infection, while maintaining test specificity. Adopting these adjusted thresholds also improved test agreement between the IGRA and culture and/or PCR-confirmed cases of mycobacteriosis. Assessment of IGRA responses in cats treated for mycobacterial infection showed that 78% of cats remain positive at the point of clinical resolution. Additionally, remaining positive at the end-of-treatment IGRA was not associated with recurrence of disease. Therefore, the IGRA is both a sensitive and specific test for the diagnosis of feline mycobacterial infections, especially for cases of TB, but based on these data it should not be recommended as a tool for monitoring responses to treatment. The IGRA is an example of a test that has been successfully adapted for use in multiple species; the feline-specific IGRA has also been used in other felids, namely the lion. However, data to confirm that anti-cat IFNγ antibodies can bind to IFNγ from other species is lacking. A combined bioinformatics and bench-work approach was undertaken to identify, compare and characterise IFNγ from Felidae species, and then produce recombinant IFNγ (rIFNγ) from these species to determine cross-reactivity of anti-cat IFNγ antibodies. Cross-reactivity of Felidae rIFNγ with anti-dog and anti-bovine IFNγ antibodies was also performed. Anti-cat IFNγ antibodies successfully cross-reacted with five rIFNγ Felidae proteins. By extension, the feline-specific IGRA may be applicable for use in other Felidae, overcoming the limitation of a lack of specific reagents for use of diagnostic tests for TB in wildlife species. Tests of humoral immunity for the diagnosis of TB are controversial in human medicine. While serodiagnostic tests have been developed for diagnosing TB in a range of animal species, the use of these tests in cats has been somewhat overlooked. To that end, an enzyme-linked immunosorbent assay for the comparative response between purified protein derivative (PPD) from M. bovis and PPD from M. avium for the diagnosis of feline TB was developed as part of this study. This assay had a sensitivity of 41% and specificity of 100%. Notably, when applied to a larger cohort of samples, a positive comparative PPD response was documented in a small number of cats that had histopathological changes suggestive of mycobacterial infection but were negative on culture, PCR and/or IGRA. This would suggest there is some scope of antibody-based diagnostics for identifying cases of feline TB. Molecular-based diagnostics such as PCR allow for the direct identification of mycobacteria, rather than assessing for an indirect response of host immunity. Given the difficulties in extracting DNA from FFPE tissue samples for conventional PCR testing, a quantitative PCR (qPCR) for use on FFPE tissues was developed. Mycobacterial DNA was successfully amplified in 79% of samples. The TB-complex specific gene IS6110 was detected in 64% of FFPE samples from cases of culture- or PCR-diagnosed TB. In cases of M. microti infection, DNA for the region of difference (RD)12 portion of the genome was identified in six out of nine cats. Detection of RD1 DNA was only achieved in three out of ten cases of infection with M. bovis. For the identification of NTM species of importance in cats, novel primer/probe sets were designed against the genes IS1311 and MLM_3300; these detect the presence of M. avium and M. lepraemurium, respectively. Four out of six samples from cats with M. avium infection were positive for IS1311; the two IS1311 negative samples were positive for MLM_3300, suggestive of M. lepraemurium infection. Three of the four M. lepraemurium samples were positive for the primer/probe set targeting this species-specific gene. These data show that further work is required for the development of a qPCR assay to diagnose cases of TB from FFPE tissues, but in its current format it may be more successful for the identification of infection with some NTM species. Overall, the work presented in this thesis advances our knowledge of the underlying immunopathology of feline mycobacterial infections, particularly for cases of TB. These results allow for the improvement of current diagnostic tests available for cases of feline mycobacteriosis, and the potential application of the IGRA for use in other felid species. The data also suggests a potential role for the use of antibody-based tests for diagnosing TB and provides further advancement of molecular approaches using qPCR.