T vaginalis genetic typing and diversity

Early approaches to typing T vaginalis isolates employed multiple techniques (table 1), including antigenic characterisation,1 isoenzyme analysis,2 monoclonal antibody binding,3 karyotype polymorphism by pulsed field gel electrophoresis,4 ,5 random amplified polymorphic DNA (RAPD),6–8 and restriction fragment length polymorphism (RFLP).9 ,10 However, each of these approaches has significant weaknesses. Isoenzyme analysis and monoclonal antibody serotyping are dependent on protein/antigen expression which can be variable, and are not sufficiently sensitive to distinguish individual isolates. Interpretation of band sizes for isoenzymes, chromosomal DNA, RAPD, or RFLP can be subjective, leading to problems with reproducibility. With the exception of RAPD, these procedures are time-consuming, requiring lengthy hybridisation steps or specific enzyme assays. The most frequently used approach, RAPD analysis, can also be compromised by variations in experimental PCR conditions and the presence of contaminating DNA, as has been shown for the presence of Mycoplasma DNA in T vaginalis clinical isolates.11 For example, three different RAPD schemes, using 18 isolates in common, produced similar phylogenetic trees, but failed to assign the same genetic relationships to all isolates.8 ,12 ,13 Two recently described DNA-based techniques for T vaginalis, multilocus sequence typing (MLST) that assays single nucleotide polymorphisms (SNPs)14 ,15 and microsatellite (MS) genotyping that determines the sizes of tandem repeats of DNA in the genome,14 ,16 are now considered more appropriate tools for investigating T vaginalis diversity due to their unambiguous nature, suitability for high-throughput analyses, and their ease of portability between different labs.15 ,17

Using RAPD, RFLP, MS and MLST typing, a significant degree of strain heterogeneity in T vaginalis isolates has been observed, with each method producing a unique genotype for nearly every isolate tested (table 1). T vaginalis isolates were also shown to maintain their phenotypic expression and genetic stability over time as monoclonal antibody profiles, PFGE, RFLP and MS patterns for isolates were reported to be unchanged after several years of continuous cultivation.1 ,5 ,10 ,17

T vaginalis population structure

There is strong evidence for the existence of a two-group or ‘type’ population structure for T vaginalis, as this organisation is seen in phylogenetic trees and the output of Bayesian clustering algorithms that analysed data based upon different typing techniques. For example, Snipes et al employed RAPD genotyping to 63 US strains, Meade et al used Eco-RI Hsp70 RFLP genotyping of 129 US isolates, Cornelius et al analysed 68 US clinical isolates using MLST, and Conrad et al used MS typing on 188 global isolates; each study identified a unique two-type structure.7 ,10 ,15 ,17 While the first three studies identified a two-type population structure in isolates collected in the USA, the work by Conrad et al demonstrated its global distribution. The number of isolates in the two types for each of these investigations appears to be approximately equal with the exception of a preponderance of Type 1 isolates in South Africa isolates and a bias toward Type 2 isolates in Mexico, which could be sampling artefacts.17 Correlations between the two types of T vaginalis isolates identified in each of these investigations will need to be determined, as on the whole, each has used unique sample sets with different genotyping techniques. The existence of two genetic types may offer an explanation for the conflicting results reported for associations of T vaginalis with its disease sequelae, such as risks for HIV acquisition or adverse pregnancy outcomes, as the magnitude of these associations may differ based on the local composition of T vaginalis population types. The potential for the manifestation of different clinical symptoms and outcomes from infection by the different types is also apparent as it is a well-established tenet of infectious disease that discrete strains or groups of isolates of a given microbial organism can exhibit differing, yet reproducible and stable, clinical properties.

T vaginalis recombination

T vaginalis reproduces by mitosis, and was originally proposed to be a clonal organism with sexual recombination either absent or insufficiently frequent to alter the consequences of a clonal population structure.18 The identification of clonal complexes in analyses of MLST data, including one composed of 17 MLST sequence types representing 23 individual isolates, with representatives of both MLST population groups present, and apparently stable within isolates collected >40 years apart, supports a role for clonality in T vaginalis.15 However, the first T vaginalis genome sequence19 revealed components of the meiotic recombination machinery, indicating a capacity for homologous recombination and genetic exchange.20 Additional evidence for the existence of sexual recombination in T vaginalis comes from population genetic analyses that looked for significant ‘linkage disequilibrium’ (LD) between different MS and SNP genetic markers.15 ,17 Results of these tests indicated that T vaginalis populations, and in particular Type 1 parasites, are in linkage equilibrium, indicative of genomes that have recently undergone recombination. However, recombination between Type 1 and Type 2 parasites appears to be a much less frequent occurrence, and suggests the existence of biological or spatial barriers to recombination between the types. It is apparent that although the capacity for clonal propagation exists, sexual recombination has significantly impacted the evolution of extant T vaginalis.17

Mixed genotype infections

The high prevalence of T vaginalis in some populations predicts that mixed infections of genetically different T vaginalis strains must occur. However, little evidence of mixed infections has been reported, principally due to the inability of many typing techniques to reliably and conclusively identify mixed genotypes. Demonstration of mixed infections also requires prompt analysis of the initial sample, as even small differences in in vitro growth rates can result in the elimination of one T vaginalis genotype during extended cultivation. MS typing of 211 T vaginalis cultures from across the globe identified 23 mixed infections (10.9%) isolated in five different continents.17 The interaction of isolates with differing phenotypic properties in mixed infections, particularly for those from different populations, has potentially important clinical implications in trichomoniaisis, and should be a priority topic in future investigations.

Genetic associations with clinical phenotypes

Numerous studies have demonstrated concordance between T vaginalis genotype and the clinical presentation of trichomoniasis, including metronidazole susceptibility, the presence of T Vaginalis virus (TVV), and concurrent Mycoplasma infection (table 2). The association between genotype and metronidazole susceptibility appears to be the strongest. For example, four studies based on RAPD genotyping demonstrate concordance between T vaginalis genotype and aerobic metronidazole susceptibility of isolates.7 ,8 ,13 ,21 In one of those studies, the association of genotype and metronidazole susceptibility was shown to correlate with the presence of a nucleotide polymorphism (C66T) in the 5.8S rRNA internal transcribed spacer region, although this is not likely to be a causative correlation.7 The mean minimum lethal concentration of metronidazole necessary to kill T vaginalis isolates was also significantly higher, a nearly threefold increase for Type 2 parasites as compared with Type 1 parasites.17

Evidence for concordance between genotype and clinical manifestations of the disease is less conclusive. Investigations using RAPD markers generated phylogenetic trees that grouped isolates based on the severity of clinical symptoms.6 ,8 ,22 ,23 However, other reports based on RAPD and RFLP patterns of a PCR-generated 28S-18S rRNA intergenic spacer failed to demonstrate concordance between genotype and symptoms.13 ,24 Additionally, Conrad et al were unable to demonstrate a correlation between T vaginalis genetic type and vaginal pH or a positive whiff test, although the numbers of samples analysed were small.17

T vaginalis can be infected with up to four distinct double-stranded RNA viruses of the family Totiviridae.25 The presence of TVV has been implicated in parasite pathogenesis via increased cysteine proteinase expression, which enhances cytoadherence, cytotoxicity and degradation of basement membrane, and the ability of TVV-infected T vaginalis to express immunogens on their surfaces and to undergo phenotypic variation.26–28 In one study, the presence of TVV in clinical isolates of T vaginalis resulted in higher levels of in vitro adherence to vaginal epithelial cells, and more severe clinical disease than uninfected isolates, and isolates infected with TVV-2 were more pathogenic than TVV-1 infected isolates.29 TVV presence in T vaginalis clinical isolates is widespread, with 50.5% of isolates collected in the USA infected with TVV.7 Snipes et al7 using RAPD markers, demonstrated concordance between TVV and genotype in phylograms of 29 metronidazole-resistant isolates from across the USA, and 33 isolates acquired in Atlanta, GA. A band-sharing index also demonstrated a high degree of relatedness between metronidazole-resistant isolates and TVV-infected isolates in that study, although a negative correlation between metronidazole resistance and presence of TVV was found in other studies.17 A significant association between the presence of TVV and phylogenetic position was also shown in a study based on RAPD data.23 Microsatellite typing of 154 isolates from across the globe that were infected with TVV also showed a significant difference with 72.9% identified as Type 1, and only 27.1% as Type 2 isolates, suggesting that the Type 1 genotype is more conducive to TVV infection.17 However, several other studies failed to show linkage between T vaginalis genotype and TVV presence, possibly due to the small sample sizes employed.8 ,13 Thus, unfortunately, due to the paucity of clinical information available, the relationship between the presence of TVV and clinical disease manifestations has not been definitively established. The discrepancies noted above highlight the need for larger, more comprehensive studies into the relationship between the genetic identity of T vaginalis isolates and clinical manifestations, drug resistance and the presence of TVV.

Future directions

The development of a common framework for genotyping T vaginalis is essential to provide the necessary tools for addressing potential correlations between T vaginalis the organism, and trichomoniasis the disease. The DNA-based techniques, MLST and MS genotyping, are most suitable for these investigations, but the relationship between the two population types identified in each of those approaches needs to be clarified. Analysis by both MLST and MS genotyping of a common set of clinically annotated T vaginalis isolates has recently been initiated by the authors. This will facilitate investigations into the relationships between mixed infections, T vaginalis genetic identity, the unique two-type population structure, and clinical manifestations of the disease. The interaction between the vaginal microbiome, patient immune response, and T vaginalis genetic diversity can then be investigated in order to obtain a comprehensive understanding of trichomoniasis and its associated disease sequelae. The extensive genetic diversity and two-clade population structure in T vaginalis will also need to be carefully considered in other investigations of trichomoniasis. The use of proteomic and genomic-based techniques to elucidate genetic factors controlling clinical manifestations and severity, association with other disease entities, and drug resistance will require T vaginalis isolates which, other than the relevant trait of interest, are as genetically similar as possible in order to maximise the discovery of the relevant genetic elements and avoid differences rooted in genetic diversity. By contrast, future vaccine design and development of antigen-based rapid diagnostic tests for trichomoniasis will need to focus on antigenic determinants which are broadly representative of the entire T vaginalis population. The existence of a two-type population structure, and the association of particular clinical and phenotypic traits with genetic identity also highlight the need for genome sequencing of additional T vaginalis strains. To date, the sequence of only a single T vaginalis strain, G3 (American Type Culture Collection reference number Pra98), has been generated, and it is imperative that the genome sequence of additional clinically relevant isolates representative of both T vaginalis population types be determined.