1 Part 1: Introduction

With a global prevalence approaching 400 million (Fiori et al. 2013; WHO 2018) and annual new infections globally reported to be 156 million (WHO 2018), trichomoniasis is the third-most common sexually transmitted infection, highlighting the potent ability of T. vaginalis to colonize its human host. Trichomoniasis prevalence is only exceeded by HPV (Human Papilloma Virus) and HSV-2 (Herpes Simplex Virus Type 2, genital herpes). Discovered in 1836, T. vaginalis was historically considered to be a member of the commensal vaginal microbiome, however its pathogenic properties began to be appreciated about a century later. Still, the complexity of trichomoniasis is not well understood, with reports of symptoms ranging from none, to moderate, to severe, putatively due to variation in (1) human immune responses, (2) indigenous vaginal microbes, and (3) the inherent virulence of the strain, including whether the strain harbors any microbial symbionts (Mercer and Johnson 2018). From a cell biology perspective, T. vaginalis flagella and cytoskeletal elements are some of the parasite’s most visually strinking features and poised to reveal unique T. vaginalis biology. Various biochemical processes in the parasite, in addition to promoting survival within a unique compartment in the human host, also likely intersect with pathogenesis. Studies have zeroed in on particular molecular players and cellular pathways that contribute to T. vaginalis interactions with host cells. Adhesins, proteases, secreted vesicles, regulatory gene expression, and signaling events help mediate T. vaginalis adhesive properties and its cytotoxic effects. It is now important to investigate how T. vaginalis employs these factors and pathways during an in vivo infection. The latter is also needed to fully understand the parasite’s impact on the surrounding host microenvironment including effects on promoting inflammation and modulation of the urogenital microbiome. In this chapter we will delve deeper into these multifaceted aspects of the T. vaginalis lifecycle.

2 Part 2: Trichomoniasis; Disease and Outcomes

T. vaginalis can affect both sexes, and most infections are mild or asymptomtic. Below, we outline current knowledge of T. vaginalis pathology in women, men, and neonates.

2.1 Effects on Women

The morbidity burden of trichomoniasis is disproportionately carried by women. With an overall prevalence of 1.8%, the United States carries a higher burden than other high-income countries, pointing to the lack of public health attention paid to the parasite in the US, where it is currently classified as a neglected parasitic infection (Secor et al. 2014; Van Gerwen and Muzny 2019). African Americans are disproportionately affected by 17-fold, compared to other ethnic backgrounds (Van Gerwen and Muzny 2019). T. vaginalis is also reported to have a prevalence of up to 43% in Human Immunodeficiency Virus (HIV)+ women (Fastring et al. 2014; Van Gerwen and Muzny 2019). Common symptoms include vaginitis and vaginal discharge. However, serious adverse consequences include pelvic inflammatory disease (Wiringa et al. 2019), increased risk of both HIV transmission and acquisition, and increased risk of malignant cervical cancers (RN Fichorova 2009; Mielczarek and Blaszkowska 2016). One study reported a 2.62 odds ratio of high-grade cervical squamous intraepithelial lesions with T. vaginalis (Tao et al. 2014), and another showed that among HPV+ women, those also infected with T. vaginalis were more likely to have cervical cancer lesions (Ghosh et al. 2017). A recent meta-analysis also upholds this link in all ethnic backgrounds studied, which included European, African, and Asian women (Yang et al. 2018). How T. vaginalis contributes to cervical cancer progression remains unknown, but it could be due to inciting more tissue damage and inflammation in HPV+ neoplasms, which in turn promotes their metastasis.

2.2 Effects on Reproductive Outcome and Neonates

T. vaginalis is associated with female and male infertility, and with pre-term/low-weight infant births (RN Fichorova 2009; Mielczarek and Blaszkowska 2016; Wiringa et al. 2019). T. vaginalis prevalence in pregnant women was found to range from 3.9% to 24.6% in low- to middle income countries (Van Gerwen and Muzny 2019). The strongest adverse consequences are thought to be reductions in carrying a fetus to term. However, aside from sequelae of premature birth, T. vaginalis is not thought to directly cause disease in neonates. A recent meta-analysis of all reported studies on T. vaginalis and pregnancy, upholds a link between T. vaginalis and perinatal morbidity (Silver et al. 2014), and a new study in Uganda upholds a specific link between T. vaginalis infection and premature uterine membrane rupture among STIs; Chlamydia and HSV-2 had no link (Nakubulwa et al. 2015). Recently, a report also cited a single case of a neonate born with T. vaginalis, found in a brain abscess (Hamilton et al. 2018), however this is not a common occurrence.

2.3 Effects on Men

Although most men are asymptomatic and can have high spontaneous clearance rates, T. vaginalis can cause inflammation in the prostate (prostatitis) (Van Gerwen and Muzny 2019). Furthermore, T. vaginalis sero-positivity was found to be associated with prostate cancer (PC) (Sutcliffe et al. 2006), particularly in cases that are metastatic (Stark et al. 2009). However, more recent studies observing larger cohorts of African American men, who are at increased risk for both T. vaginalis and PC, failed to support this correlation (Marous et al. 2017). Another recent study also failed to find a correlation between T. vaginalis sero-positivity and metastatic or fatal prostate cancer (Shui et al. 2016), and a recent metanalysis of all correlation studies including T. vaginalis and PC performed across many ethnic groups also identified only a modest increase that was not statistically significant (Najafi et al. 2019). Therefore, T. vaginalis is unlikely to be a major risk factor for prostate cancer. Intriguingly however, T. vaginalis produces a protein called TvMIF, a protein mimic of human Macrophage Migration Inhibitory Factor (MIF), which is associated with PC (Meyer-Siegler et al. 2006, 2005). Researchers also demonstrated that TvMIF could promote prostate cell growth and metastatic behavior in vitro (Twu et al. 2014). Therefore, it is possible that TvMIF could be involved in promotion of some cases of metastatic prostate cancers despite the fact that recent studies fail to uphold signitifcant increases in PC among T. vaginalis+ men, since PC is highly multi-factorial in etiology. Furthermore, TvMIF could promote cell growth and invasion during cervical cancer metastasis. TvMIF also contributes to inflammation (Twu et al. 2014), which is associated with myriad T. vaginalis-associated pathologies and also contributes to metastatic processes.

2.4 Transmission

As a sexually-transmitted parasite, T. vaginalis can be transmitted through genital contact from male-to-female, from female-to-male, and from female-to-female (Muzny et al. 2012, 2011; Van Gerwen and Muzny 2019). It is less commonly transmitted from male-to-male, but reports of rectal transmission in men who have sex with men are now emerging (Hoffman et al. 2018). While oral sex is not thought to be a major mode of transmission, one case of vaginal-oral-pharyngeal transmission was recently reported (Kitty Carter-Wicker and Omole 2016). Safer-sex practices can reduce transmission, as condoms confer good protection (Crosby et al. 2012). Furthermore, a recent study confirmed old reports that male partner circumcision reduces male-to-female transmission by approximately 50% (Gray et al. 2009; Pintye et al. 2017). However, studies have had variable results in determining whether circumcision decreased the acquisition of T. vaginalis by males from females (Mehta et al. 2009; Sobngwi-Tambekou et al. 2009).

2.5 Diversity of Strains

As discussed above, the clinical presentations of T. vaginalis can vary greatly. One reason for the variability in disease is thought to be because many different strains of the parasite are circulating in the population, which likely vary in virulence (Lustig et al. 2013; Mercer et al. 2016). Genetic analysis of T. vaginalis strains shows that two main types exist (MD Conrad et al. 2012). Types 1 and 2 can be distinguished using a panel of micro-satellites, and exist in equal proportions worldwide (Conrad et al. 2011, 2012, 2013). Type 1 is more likely to harbor a viral symbiont, Trichomonasvirus (discussed in Sect. 5.1), and is correlated with increased HIV viral load in T. vaginalis-HIV co-infected women (Conrad et al. 2013). T. vaginalis is known to divide asexually. However, modes of sexual reproduction are suspected, since comparison of different strains showed evidence of genetic exchange (Conrad et al. 2012). However, sexual reproduction has never been demonstrated in the laboratory (Bradic and Carlton 2018). It is currently unclear whether the parasite undergoes any sexual reproduction during infection, but it is clear that a great amount of diversity has evolved in this parasite throughout its history within the human host.

Having outlined important clinical features of trichomoniasis, we will now discuss some basic biology of this trichomonad, including its morphology and organelles.

3 Part 3: The Cell Structure and Biology of T. vaginalis

T. vaginalis exists as an ovoid trophozoite when grown axenically. However, the parasite transforms to an ameboid form upon binding to surfaces of host epithelial cells (Arroyo et al. 1993; Benchimol et al. 2001; Midlej and Benchimol 2010). Transmission electron microscopy (TEM), scanning electron microscopy (SEM), and recent electron tomography with 3D-reconstruction of T. vaginalis have allowed detailed observation of T. vaginalis morphology and cellular features, which include interesting structures and organelles outlined below, and shown in Fig. 1. While the specific proteins comprising each of these cellular structures and a full understanding of the biogenesis of these organelles are unknown, recent powerful omics work presented below has generated datasets bound to illuminate new aspects of T. vaginalis cell biology.

Fig. 1
figure 1

T. vaginalis morphology. (a) Schematic diagram of T. vaginalis. Anterior flagella (AF), recurrent flagellum (RF), undulating membrane (UM), costa (C), parabasal filament (Pf), pelta (P), nucleus (N), Golgi complex (G), axostyle (Ax), hydrogenosome (H), vacuole (V). (b, c) 3D structure of T. vaginalis. Using electron tomography, 100-nm sections (b) or 250-nm sections (c) of T. vaginalis were utilized to generate the 3D reconstructions shown. Anterior flagella (AF), axostyle (Ax), costa (C), Golgi complex (G), hydrogenosomes (H), recurrent flagellum (RF), parabasal filaments (Pf 1–3). This figure was reprinted from Copyright (2009) Oxford University Press. Lee et al. Three-dimensional structure of the cytoskeleton in Trichomonas vaginalis revealed new features, J Electron Microsc (Tokyo) and publisher

3.1 Structural Features and Dimension

3.1.1 Flagella and Cytoskeleton-Related Structures

T. vaginalis is a motile parasite propelled by flagellar beating (Lee et al. 2009). Five basal bodies (kinetosomes) at the most anterior region of the cell anchor the four anterior flagella and one recurrent flagellum (Lee et al. 2009). The flagellar axonemes have a typical eukaryotic 9 + 2 arrangement of microtubules (Benchimol 2004; Lee et al. 2009). The anterior flagella exits the cell through a flagellar canal that is supported by the crescent-shaped pelta (Benchimol 2004). The pelta also supports the anterior portion of the cell and overlaps with the axostyle in the top part of the cell; this area of overlap is referred to as the pelta-axostyle complex (Marlene Benchimol 2010). The recurrent flagellum exits through another region on the anterior part of the cell and folds downwards towards the posterior of the parasite, remaining attached to the cell body and helping to form the undulating membrane (Benchimol 2010; Benchimol and De Souza 1990; Honigberg et al. 1984). The undulating membrane extends about halfway to three fourths down the T. vaginalis cell body (Cheon et al. 2013; Smith and Stewart 1966).

The costa is a unique fiber structure present in trichomonads that contain undulating membranes (Viscogliosi and Brugerolle 1994). The costa emerges from the second basal body, passes underneath the basal body that gives rise to the recurrent flagellum, and continues in the cytoplasm, following the path of the undulating membrane (Lee et al. 2009; Viscogliosi and Brugerolle 1994) (Fig. 1). Due to its adjacent localization to the undulating membrane, the costa is hypothesized to provide structural support to the undulating membrane (Viscogliosi and Brugerolle 1994). Proteins with molecular weights between 100 and 135 kDa are present in the costa-enriched fractions from T. vaginalis and other trichomonads (Viscogliosi and Brugerolle 1994). However, the identity of these proteins remains to be determined.

Researchers have developed a methodology to isolate T. vaginalis flagella, by physically shearing the parasite and performing sucrose density-gradient centrifugation. Analysis of these isolated flagella also shows the presence of proteins with a wide variety of molecular weights (Jemilohun 1998). However, the complete protein makeup of the T. vaginalis flagella is still unknown. Upon exogenous expression of tagged proteins it has been shown that many proteins that localize to the T. vaginalis flagella also localize to the plasma membrane. These proteins include two tetraspanin family members TSP1 (TVAG_019180) and TSP6 (TVAG_460770), TVAG_267320 a nicastrin precursor, TVAG_020780 a serine/threonine protein phosphatase, and TVAG_258230, a hypothetical protein (de Miguel et al. 2010, 2012). Some of the identified flagellar proteins also have likely roles in interaction with host cells. For instance, upon co-incubation of ectocervical cells with T. vaginalis overexpressing an HA-tagged TSP6 fusion protein (TSP6-HA), TSP6-HA underwent a striking redistribution to the flagella. The C-terminal tail of TSP6 was necessary for flagellar targeting (de Miguel et al. 2012), and when the TSP6 C-terminal tail was fused to two different tetraspanin proteins, it conferred stronger flagellar targeting of these proteins as well, indicating that T. vaginalis contains flagellar-targeting motifs. The concentration of plasma membrane proteins to the flagella upon host cell encounter, together with the identification of other signaling proteins in the flagella, suggest a sensory role for T. vaginalis flagella (de Miguel et al. 2012). Interestingly, using freeze-facture electron microscopy, researchers observed the presence of 9-12 intramembranous particles arranged in rosettes on the anterior flagella of T. vaginalis (Benchimol and De Souza 1990; Honigberg et al. 1984). In the area where flagella emerge from the parasite, rosettes are also found positioned around the flagella in the form of a “flagellar necklace” (Benchimol 2010). These intriguing flagellar rosettes have also been visualized in other ciliates, but their function similarly remains to be elucidated. It also remains to be investigated whether the five T. vaginalis flagella have distinct or complementary roles, which may be revealed as we learn more about their locomotive properties and how these flagella are assembled in the parasite.

3.1.2 Axostyle

Similar to some other protists, T. vaginalis also contains an axostyle. The axostyle is a ribbon of longitudinally-oriented microtubules that span from the top to the bottom of the parasite and extends past the end of the cell, forming a small narrow tube that protrudes from the cell posterior (Lee et al. 2009; Ribeiro et al. 2000) (Fig. 1). The axostyle in T. vaginalis does not contract and thus is not thought to contribute to motility. Instead, the axostyle helps to maintain the cell axis and has an important role in T. vaginalis cell division (Marlene Benchimol 2010), as discussed in Sect. 3.3.

3.1.3 Parabasal Apparatus

As a member of the group Parabasalia, T. vaginalis also contains a parabasal apparatus, composed of the Golgi complex (discussed in Sect. 3.1.4) in proximity to parabasal filaments (Benchimol et al. 2001) (Fig. 1a, b). Benchimol et al. observed fibrils connecting the cis-most region of the Golgi to the parabasal filaments and thus the parabasal filaments have been hypothesized to support the Golgi structurally (Benchimol et al. 2001). Recent 3D reconstruction of T. vaginalis via electron tomography has now revealed the presence of three parabasal filaments (Lee et al. 2009) instead of two, as previously observed. These filaments arise from the basal bodies and extend downward inside the cell (Benchimol et al. 2001). Parabasal filament 1 splits into two strands and one of these strands then points into the interior of the cell (Fig. 1c). It has also been hypothesized that the twisting structure of parabasal filament 1 may help provide tension and force to aid flagellar beating (Lee et al. 2009), however this model also remains to be tested.

3.1.4 Dimensions

Using scanning electron microscopy, researchers found that the T. vaginalis cell body measures 9.5 (7.4–11.4) μm in body length and 6.8 (5.3–7.7) μm in width, and that the total size including the length of the flagella, the body, and the axostyle is 26 (21–32) μm (Cheon et al. 2013). Interestingly, Cheon et al. reported that the axostyle of freshly isolated T. vaginalis is longer than that of long-term passaged parasites, measuring 8.2 and 4.0μm, respectively. However, the biologic implication of this finding remains to be investigated.

3.2 Organelles and Metabolism

3.2.1 Golgi Apparatus

As discussed in Sect. 3.1.3, The T. vaginalis Golgi has a peculiar association with the parabasal apparatus, a defininig feature of the group Parabasalia. However, many other features of the T. vaginalis Golgi are similar to other eukaryotes. It contains 8–12 cisternae (segments), and reaches dimensions of 6μm in length and 1μm in width (Benchimol et al. 2001). Cytochemistry analysis of the T. vaginalis Golgi shows the presence of carbohydrates, indicating a role in post-translational glycosylation of proteins, as it does in other eukaryotes (Benchimol et al. 2001). Proteins within the T. vaginalis Golgi have been defined, such as AP51 and AP65 (Benchimol et al. 2001), TvROM2 (TVAG_359500), TvROM3 (TVAG_476950) (Riestra et al. 2015), and a BspA protein (TVAG_240680) (Handrich et al. 2019). Some of these proteins have served as Golgi markers (Riestra et al. 2015), indicating specific roles in the organelle, however the majority of the Golgi content and the biochemical reactions that take place there remain to be uncovered.

3.2.2 Endoplasmic Reticulum

T. vaginalis also contains a rough endoplasmic reticulum (ER) system that is located around the nucleus (Smith and Stewart 1966). Along with the Golgi, the ER is predicted to be a site of calcium mobilization, also displaying positive Ca++-ATPase activity (Benchimol 2004), indicating that the Golgi and ER are key regulatory organelles for calcium-dependent cellular signaling. T. vaginalis proteins that appear to localize to the ER include the adhesin TVAG_212570 (Handrich et al. 2019) and the tetraspanin protein TvTSP4 (TVAG_458280 (Coceres et al. 2015).

3.2.3 Nucleus and the Mastigont System

The nucleus of T. vaginalis is ovoid and is surrounded by a double membrane (Smith and Stewart 1966), and freeze fracture studies have revealed the presence of nuclear pores on the T. vaginalis nucleus (Honigberg et al. 1984). Researchers have also observed that the nucleus is closely associated with the mastigont system of T. vaginalis, which arises at the T. vaginalis cell anterior (Viscogliosi and Brugerolle 1994). The mastigont system refers to kinetosomes and flagellar axonemes, appendages and rootlet filaments that surround the basal bodies, the costa, parabasal filaments, the pelta, and the axostyle (Benchimol 2010). Its close association with the nucleus is important in cell division, as we will discuss in Sect. 3.3.

Other organelles such as hydrogenosomes, lysosomes, and glycogen granules are also present in T. vaginalis (Benchimol 2004; Honigberg et al. 1984; Smith and Stewart 1966). Interestingly, hydrogenosomes and glycogen granules are often observed in association with the axostyle (Benchimol 2004). Hydrogenosomes and glycogen granules participate in the parasite’s anaerobic metabolism, which we will discuss below.

3.2.4 Hydrogenosomes

T. vaginalis lack mitochondria and instead contain hydrogenosomes. Phylogenetic and biochemical analyses indicate that the two metabolic organelles share a common ancestor (Schneider et al. 2011; Shiflett and Johnson 2010). Similar to mitochondria, hydrogenosomes are double-membrane bound organelles, but unlike mitochondria, they lack a genome. Fermentative metabolism of pyruvate takes place inside hydrogenosomes, leading to the production of ATP. During this process, molecular hydrogen is generated as a by-product, hence the organelle’s name (Shiflett and Johnson 2010). Schneider et al. published a proteome of T. vaginalis hydrogenosomes, consisting of 569 proteins and identifying putative hydrogenosome biochemical pathways such as amino acid and energy metabolism, iron-sulfur cluster assembly, flavin mediated catalysis, oxygen stress responses and ATP hydrolysis (Schneider et al. 2011). A recently published metabolomic analysis of T. vaginalis hydrogenosomes, has provided functional evidence that hydrogenosomes are indeed sites of amino acid metabolism (Huang et al. 2019). The metabolic adaptability of T. vaginalis to varying conditions that the parasite might face during infection, particularly low glucose availability and oxidative stress, together with the production of ATP, highlight the vital role of hydrogenosomes in T. vaginalis biology.

3.3 Cell Division

In vitro, T. vaginalis doubles approximately every 4–6 h. The parasite undergoes cell division through closed mitosis, a form of mitosis in which the nuclear envelope does not break down. Here, we discuss structures that are important for understanding closed mitosis and describe how the cell morphology changes during cell division.

Prior to mitosis, DNA is replicated and the cell also duplicates the components of the mastigont system (discussed in Sect. 3.2.4), which travel together with the cell bodies during T. vaginalis cell division (Benchimol 2004; Ribeiro et al. 2000). A microtubule organizing center (MTOC), also known as an attractophore, appears underneath the second kinetosome in each daughter cell and near the parabasal filament (Benchimol 2010; Bricheux et al. 2007). The attractaphore is a unique morphological feature of parabasalids (Bricheux et al. 2007). Scanning electron microscopy and indirect immunofluorescence assays utilizing anti-tubulin antibodies that recognize the spindle pole, flagella, basal bodies, and the pelta-axostyle together with nuclear staining, have allowed visualization of the movement of these structures during T. vaginalis cell division (Bricheux et al. 2007; Ribeiro et al. 2000; Zuo et al. 1999). A summary of these overall findings is described below and shown in Fig. 2.

Fig. 2
figure 2

Immunofluorescence of T. vaginalis during cell division. (a) Visualization of the attractophore and growing mitotic spindle during cell division. (A–C) Double immunofluorescence microscopy of extracted cytoskeletons. (A) anti-attractophore (At), green, (B) Flagella (F) and axostyle (Ax), anti-tubulin, red. (C) Attractophore co-localized near the the base of the flagella and axostyle, yellow. (E–G) A growing mitotic spindle/paradesmose (S) observed in dividing T. vaginalis cells. Scale bar=10 μm. (b) Duplication and movement of cytoskeletal structures during cell division. (A–I) anti-tubulin staining of T. vaginalis cytoskeletal components: axostyle (Ax), flagella (F), recurrent flagella (rF), pelta (P) and the mitotic spindle/paradesmose (S) during Interphase (A) and in dividing cells (B–I). Scale bar=3 μm. Image 3A was adapted from Copyright (2007) Elsevier. Used with permission from Bricheux et al. Identification of a new protein in the centrosome-like “atractophore” of Trichomonas vaginalis, Mol Biochem Parasitol. and publisher. Image 3B was adapted from Copyright (2000) Wiley. Used with permission from Ribeiro et al. Contributions of the axostyle and flagella to closed mitosis in the protists Tritrichomonas foetus and Trichomonas vaginalis, J Eukaryot Microbiol. and publisher

During interphase and prophase, the attractophore, is initially observed as a round “dot-like” structure next to one pole of the nucleus (Bricheux et al. 2007; Gomez-Conde et al. 2000; Juliano et al. 1986) (Fig. 2a, A). The dividing nucleus also grows in size and it is initially situated between the duplicated axostyles arranged side-by-side (Fig. 2b, B–C). During metaphase, the attractaphore appears to duplicate (Marlene Benchimol 2010; Bricheux et al. 2007). A set of microtubules emerges from the attractophores and connects both attractophores to each other, forming an extranuclear spindle called the paradesmose (Petrin et al. 1998). As the paradesmose elongates it drives the division of the nucleus and the daughter cells (Juliano et al. 1986; Petrin et al. 1998) (Fig. 2a, E–G). Another set of microtubules extends from each attractophore over the outside of the nucleus and forms a mitotic spindle like-structure (Gomez-Conde et al. 2000). These microtubules connect to the kinetochores of the duplicated chromosomes residing inside the nucleus via attachment of each microtubule to the nuclear membrane (Benchimol 2004; Gomez-Conde et al. 2000; Juliano et al. 1986; Petrin et al. 1998). Thus the spindle remains extranuclear at all times. The attractophores and the anchored mitotic spindle like-structures migrate towards the opposite poles of the nucleus. Thus, at the end of metaphase, two attractophores are observed flanking the nucleus, forming the poles for nuclear division (Petrin et al. 1998).

In parallel to these processes, the axostyles start to move in opposing directions causing the dividing cell to initially display a triangle or heart-shaped morphology (Fig. 2b, C–D). The basal bodies and mastigont system continue to move in opposite directions and at metaphase end up at about 180° from each other, with the two daughter cells pointing in opposite directions but remaining connected at the cell posterior, giving rise to an elongated cell (Benchimol 2004; Ribeiro et al. 2000) (Fig. 2b, H–I). In telophase, the nucleus is stretched and elongates, and finally separates into two daughter nuclei during the process of karyokinesis (Gomez-Conde et al. 2000). Video microscopy has revealed that flagellar beating in the opposing cells may help to complete cytokinesis, driving the physical separation of the two daughter cells which then regain their normal cell motility (Ribeiro et al. 2000). After T. vaginalis divides, two daughter cells are formed, each with 6 haploid chromosomes (Gomez-Conde et al. 2000).

Prior to mitosis, the Golgi in T. vaginalis divides through the process of Golgikinesis. The Golgi ribbon divides through medial fission cisternae by cisternae forming two smaller Golgi ribbons (Benchimol et al. 2001). When the cell divides, the Golgi ribbon in each daughter cell also travels with their associated parabasal filaments. Therefore, the two small Golgi ribbons are initially adjacent to each other side by side and then move apart and grow as the cell bodies migrate towards opposing ends. After the cell undergoes cytokinesis, a fully formed Golgi is observed in each daughter cell. Similarly, fully formed hydrogenosomes are also present in both daughter cells, in part generated through hydogenosomal fission (M. Benchimol et al. 1996; Wexler-Cohen et al. 2014).

Ultrastructural imaging has therefore uncovered a very intriguing T. vaginalis cytoskeletal and organellar arrangement that may have some unique and important contributions to T. vaginalis biology. However the specific proteins comprising these structures and organelles are less characterized. As described ahead (Table 1), large datasets from recent omics work are aiding in the depiction of a more comprehensive molecular characterization of these structures and organelles in T. vaginalis.

Table 1 Available omics resources useful for the study of Trichomonas vaginalis

3.4 T. vaginalis omics: Genome, Epigenetics, and Gene Expression

3.4.1 The T. vaginalis Genome

The genome of the G3 T. vaginalis strain was published in 2007 as a highly fragmented assembly of sequences, thus claimed as a draft of the entire genome (Carlton et al. 2007). This ∼160 Mb T. vaginalis draft genome is deposited in GenBank, and is also accessible at TrichDB (http://TrichDB.org), along with a variety of functional genomics data (Aurrecoechea et al. 2009). It is the largest genome among sequenced protists, containing ~60,000 predicted protein-coding genes residing in six chromosomes with many gene families highly expanded and containing many repetitive sequences (up to 65% of the genome) (Carlton et al. 2007). Since its publication, the genome has been used to reveal interesting functional properties of T. vaginalis and facilitated significant progress in the development of molecular biology and omic approaches to study the infection process. Despite the hurdles, the first T. vaginalis genome sequence was undoubtedly a success, improving our understanding of the biology of this organism. However, postgenomic applications frequently require comparative genomics at the single nucleotide level and performing these analyses with the currently available highly fragmented genome is problematic. Hence, it is critical to continue efforts to improve T. vaginalis genome assembly and gene annotations of the reference strain. Additionally, sequencing of new T. vaginalis strains will provide invaluable information for future studies that will reveal knowledge about T. vaginalis genome architecture and structure.

In the last few years, high-through put methods for large-scale genetic comparison have been useful to confirm the existence of two genetically distinct T. vaginalis populations (Bradic et al. 2017). Interestingly, genomic characterization of different isolates showed that the extensive clinical variability in trichomoniasis is correlated with significant genetic diversity in the organism itself (Meade and Carlton 2013), as discussed in Sect. 2.5.

3.4.2 Epigenomics

Epigenomics analyzes epigenetic changes across the whole genome, that could lead to changes in gene activity and expression without altering DNA sequences. DNA methylation and histone modifications are examples of tightly regulated mechanisms that produce changes in gene expression, potentially leading to phenotypic changes. There is growing interest in T. vaginalis epigenetics due to its recently discovered role in the regulation of gene transcription and pathogenesis (Lizarraga et al. 2020; Pachano et al. 2017; Song et al. 2017). DNA methylation sequencing (MeDIP-Seq) and chromatin immunoprecipitation followed by sequencing (ChIP-Seq), enable precise genome-wide localization of epigenetic markers to decipher gene activity and chromatin states. In 2017, the first epigenome mapping was performed in T. vaginalis (Song et al. 2017). The authors described the genome-wide localization of two histone H3 modifications (H3K4me3 and H3K27Ac) and found that these marks are positively associated with active gene expression in both steady and dynamic transcriptional states (Song et al. 2017), providing the first direct evidence that epigenetic mechanisms play an essential role in transcriptional regulation of T. vaginalis. Pachano et al. also demonstrated that the binding of the initiator binding protein IBP39 to the metazoan-like initiator (Inr) element is strictly dependent on the acetylation state of the histones and the chromatin structure, exemplifying for the first time the mutual dependence and complex crosstalk among epigenetic players, chromatin structure, and basal gene expression machinery in T. vaginalis (Pachano et al. 2017).

Following these studies, DNA methylation of the parasite genome was first revealed by MeDIP-seq analysis. Genome-wide distribution of N6-methyladenine (6 mA), the main DNA methylation mark in T. vaginalis, reveals enrichment at intergenic regions (Lizarraga et al. 2020). Interestingly, bioinformatics analysis revealed the presence of transcriptionally active or repressive intervals flanked by 6 mA-enriched regions; with chromatin conformation capture experiments suggesting that these 6 mA flanked regions are in close spatial proximity (Berndorff et al. 1994). Furthermore, with a combination of omics and molecular biology, a new role for 6 mA in modulating 3D genome architecture and gene expression has been proposed for T. vaginalis (Lizarraga et al. 2020).

3.4.3 Transcriptomics

The system-wide investigation of parasite gene expression is vital in understanding the coordinated set of events underlying particular stages of the parasite lifecycle, and can provide clues about which genes may be important during T. vaginalis: host interactions. These studies might possibly help us elect targets for new T. vaginalis-specific drugs. The transcriptomic data currently available for T. vaginalis reflects the complexity of the genome. Only about half of the encoded genes appear to be expressed (Gould et al. 2013). Interestingly, 93% of the encoded gene families are expressed together with many hundred pseudogenes and long non-coding RNAs (lncRNAs) whose expression is driven independently from neighboring genes (Woehle et al. 2014). Transcriptome comparisons have further demonstrated that individual members of gene families are differentially regulated upon environmental changes (Gould et al. 2013; Horvathova et al. 2012; Huang et al. 2014). In particular, the exposure to host cells results in differential expression of hundreds of genes within minutes (Gould et al. 2013). Additionally, changes in gene expression upon glucose restriction (Huang et al. 2014), cold-stress response (Fang et al. 2014), oxygen stress (Gould et al. 2013) or iron deficiency (Cheng et al. 2015) were observed; reflecting the rapid adjustment of the parasite to changing environments and the existence of a sophisticated mechanism that regulates adaptation. Finally, changes in gene expression associated with resistance to metronidazole (Bradic et al. 2017), the drug utilized to treat T. vaginalis infections (discussed in Sect. 6.1), or upon tetracycline treatment (Huang et al. 2015) have also been observed.

Analysis of the T. vaginalis genome and the chromatin landscape has given insight into various aspects of T. vaginalis biology. Employment of various omics approaches has further revealed how the parasite varies the expression of its large coding arsenal and how quickly it can do so. There is a great interest to further explore how the parasite globally coordinates expression of the molecular factors that facilitate host colonization, described in the next section.

4 Part 4: Host-Parasite Interactions

T. vaginalis is an extracellular pathogen of the human mucosa. As such, it must attach to host cells to establish infection. This process, termed cytoadherence, is thus crucial for its lifestyle. Often, cytoadherence is followed by killing of the host cell. However, this is not always the case. We have seen that this parasite also displays contact-independent modes of eliciting host cell death. In this section, we will describe the process of parasite cytoadherence and the involvement of specific molecular mediators. We will depict how this parasite promotes the killing of host cells by cytolytic factors or by modulating known pathways that induce host cell death. The underlying mechanisms of virulence that might be contact-dependent or -independent, including recent descriptions of extracellular vesicles produced by this parasite, will also be discussed here. Finally, looking at the other side of the equation, we will examine the involvement of host immune responses—including effector cells and molecules of the innate and adaptive immune response—in helping clear the infection and playing decisive roles on immunopathogenesis of trichomoniasis. As we move closer toward understading host-parasite interactions, this knowledge will be of therapeutic interest to block specific players to stymie parasite colonization, prevent ensuent tissue damage, and/or help boost a protective immune response against this parasitic infection.

4.1 T. vaginalis Cytoadherence and Host Cell Killing

4.1.1 Morphology and Dynamics of T. vaginalis in Contact with Host Cells

T. vaginalis binds to a myriad of epithelial cells of the female and male reproductive tract as well as spermatozoa, red blood cells, and leukocytes (Benchimol et al. 2008; Fiori et al. 1997; Mercer and Johnson 2018; Rendon-Maldonado et al. 1998; Rosset et al. 2002; Vazquez-Carrillo et al. 2011). Microscopy studies have characterized the morphology and dynamics of T. vaginalis adhesion to host cells. Immediately or within 5 min after co-incubation, T. vaginalis can be observed tightly bound to human epithelial cells (Arroyo et al. 1993; Furtado and Benchimol 1998; Rasmussen et al. 1986) changing its tear drop/ovoid shape to a flat, ameboid form as shown in Fig. 3 (Arroyo et al. 1993; Furtado and Benchimol 1998; Midlej and Benchimol 2010; Rasmussen et al. 1986). Arroyo et al. described this morphological change upon contact with the host cell, wherein the parasite body elongates, pseudopodia are formed, and the parasite continues to flatten out on top of the host cell gaining a “fried-egg” appearance (Arroyo et al. 1993). In scanning electron micrographs, tight apposition between the parasite and the host cell plasma membranes can be observed throughout the cell body of the parasite, potentially to maximize adherence with the host cell (Arroyo et al. 1993; Furtado and Benchimol 1998; Rasmussen et al. 1986). In addition, cytoplasmic projections protrude from the parasite and interdigitate with microvilli of vaginal epithelial cells (Arroyo et al. 1993; Furtado and Benchimol 1998). Intriguingly, transmission electron microscopy has revealed areas of continuity between the parasite and the host cell plasma membranes (Furtado and Benchimol 1998). Midlej and Benchimol also imaged T. vaginalis in the process of “pinching-off” or pulling up microvilli of bovine oviduct epithelial cells and endocytosing microvilli (Midlej and Benchimol 2010). These events resemble trogocytosis, a process by which cellular nibbling by the gut extracellular protozoan Entamoeba histolytica leads to membrane exchange and contributes to host-cell lysis (Ralston 2015b). However, whether T. vaginalis uses trogoycytosis to kill host epithelial cells remains to be formally tested.

Fig. 3
figure 3

Scanning electron microscopy of T. vaginalis trophozoites under ovoid and amoeboid forms. (left) T. vaginalis (green), with flagella (blue), recurrent flagella (orange) and axostyle (green protruding from the cell posterior) in its ovoid form. (right) Amoeboid T. vaginalis (green) adhered to a vaginal epithelial cell (pink). Images show significant morphological changes that occurs during host-cell adhesion. Image was provided as a gift by Dr. Antonio Pereira-Neves, Oswaldo Cruz Foundation-Fiocruz, Aggeu Magalhães Institute (Pernambuco, Recife, Brazil)

T. vaginalis motility also contributes to the nature of host-cell interactions. Videomicroscopy of T. vaginalis reveals that parasites can remain attached long-term to host cells or display dynamic attachment: attaching and then detaching and moving on to other host cells (Furtado and Benchimol 1998). Thus, it has been proposed that T. vaginalis can display a “hit-and-run” effect (Furtado and Benchimol 1998). However, the dynamics of T. vaginalis interactions with host cells remains to be quantified and more fully characterized with different host cell types.

T. vaginalis can also adhere to host cells in aggregates, also referred to as clumps or microcolonies. Aggregate adherence to host cells occurs rapidly (Arroyo et al. 1993; Rasmussen et al. 1986), and clumps attached to host cells remain visible after several hours (Midlej and Benchimol 2010; Rasmussen et al. 1986). Parasite aggregate adherence appears to be a more common phenotype of fresh clinical isolates and more virulent strains of T. vaginalis (Coceres et al. 2015; Midlej and Benchimol 2010), indicating that clumping could be an important mechanism of enhanced pathogenesis. Interestingly, parasite-parasite membrane interdigitations amongst aggregates attached to host cells are also visible by scanning electron microscopy, indicating a very strong association between individual parasites within clumps (Arroyo et al. 1993). Recent studies have started to identify the T. vaginalis membrane proteins that help mediate these parasite-parasite interactions. Overexpression of the tetraspanin protein 8 (TvTSP8, TVAG_008950) and a cadherin-like protein (TvCLP, TVAG_393390) led to increased parasite clump formation (defined as an aggregate of ∼ 10 or more parasites in both studies) (YP Chen et al. 2019; Coceres et al. 2015; Nievas et al. 2018b) indicating that these proteins could be virulence factors aiding in clumping. We are beginning to learn how specifc features of these surface proteins help to coordinate their involvement in the clumping process. For example, calcium ions are required for maximal aggregate formation mediated by CLP, possibly due to conformational stabilization of CLP similarly to calcium’s effects on mammalian cadherin proteins (YP Chen et al. 2019). The C-terminal tail of TvTSP8 has been found to be essential to target TvTSP8 to the parasite cell surface and the C-terminal tail may also be important to engage with intracellular cytoskeletal and signaling proteins (Coceres et al. 2015). Additionally, palmitoylation of TvTSP8 may also be necessary to promote parasite clumping, as inhibition of S-acylation led to reduced clump formation by TSP8 (Nievas et al. 2018b). Interestingly, TvTSP8 and TvCLP are both upregulated upon binding to ectocervical cells, indicating that clumping may be increasingly induced upon contact with host cells, although this remains to be formally investigated. Collectively, this indicates that T. vaginalis clumping behavior can be regulated.

4.1.2 Mechanisms of Cytoadherence-Dependent and Independent Host Cell Killing

Generally, after T. vaginalis attaches to host cells, lysis of host cells ensues (Lustig et al. 2013). Host cell damage inflicted by T. vaginalis has been observed via various experimental findings. For example, co-incubation of host cells with parasites can lead to visible gaps in cell monolayers as shown in Fig. 5b–c (Furtado and Benchimol 1998; Lin et al. 2015; Lustig et al. 2013; Ma et al. 2011; Puente-Rivera et al. 2017). Measurements of decreased transepithelial electrical resistance, indicative of reduced monolayer integrity following T. vaginalis infection has also been observed (Guenthner et al. 2005). Lastly, co-incubation of fetal membranes with T. vaginalis also causes tissue weakening (Draper et al. 1995). Therefore in vivo, T. vaginalis cytoadherence may compromize the integrity of epithelial barrier function in infected tissues. The dynamics of host cell destruction also vary by T. vaginalis strain, host cell type, and the ratio of parasites to host cells (Mercer and Johnson 2018), findings that together may account for the varied nature of symptomology in humans after T. vaginalis infection.

As described above, after parasite attachment to host cells, the parasite can ingest fragments of host epithelial cells. One study found that a large majority of epithelial cells had signs of necrosis, including plasma membrane rupture and release of cytosolic content, following adherence of T. vaginalis. Researchers also found that host cell organelles were ingested by the parasite (Midlej and Benchimol 2010). However, membrane blebbing and condensed chromatin, indicative of apoptosis, have also been observed following T. vaginalis interaction with host cells (Midlej and Benchimol 2010). A recent study by Lin and collegues reported that after incubation of a cervical epithelial cell line with T. vaginalis, 70% of cells displayed “signs of disruption”, while other cells displayed a morphology indicative of necrosis or apoptosis (Lin et al. 2015), indicating that different forms of cell death can be infilicted in a population of host cells after contact with T. vaginalis. It is also possible that T. vaginalis “feeds off” ingested host cell material, as host cell debris has been found inside T. vaginalis vacuoles, that are subsequently trafficked to the lysosomes (Midlej and Benchimol 2010). The specific molecular interactions and events that occur during parasite adherence to host cells and how they trigger host cell death remain largely unknown and may vary by host cell type. Therefore, we will present knowledge about cytoadherence-dependent cytolysis in general and then examine findings about host cell death by cell types.

Parasite attachment to host cells is detected quickly, and highly virulent strains can lyse host cells soon after (Alderete and Garza 1985; Lustig et al. 2013; Okumura et al. 2008). In the most comprehensive comparison of T. vaginalis strains, Lustig et al. compared the attachment and lytic properties of 26 T. vaginalis strains co-incubated with the ectocervical cell line (Ect-1) or the benign prostate hyperplasia cell line (BPH-1). A significant correlation between the parasite’s adhesive and cytolytic properties was found among the weakly cytolytic T. vaginalis stains, whereas this correlation was not significant in the strongly cytolytic strains, indicating that a threshold of cytoadherence may need to be reached, after which host-cell killing proceeds. In addition, separating parasite and host cells using transwell inserts that prevented physical contact blocked lysis of all 26 T. vaginalis strains tested (Lustig et al. 2013), showing that killing of epithelial cells is contact-dependent. Increasing the T. vaginalis multiplicity of infection in cytolysis assays also leads to increasing amounts of host cell lysis (Alderete and Pearlman 1984; Gilbert et al. 2000). Thus, increasing the amount of parasites that bind to and physically damage host cells potentially augments the lysis of host cells. However, future single-cell tracking analysis is necessary to discern how parasite contact and adherence to host cells is mechanistically linked to the induction of host cell death.

Inhibition of several T. vaginalis cellular processes has shed light on factors that contribute to T. vaginalis cytoadherence and host-cell killing. Initial studies involving T. vaginalis treatment with cyclohexamide and proteases revealed an inhibition on host cell binding, indicating that active protein synthesis by T. vaginalis and surface-localized proteins mediate parasite attachment to host cells (Alderete and Garza 1985). As mentioned in Sect. 3.4.3, recent transcriptomic analysis of T. vaginalis uncovered that exposure to host cells induces differential expression of hundreds of genes within minutes, identifying genes that may contribute to T. vaginalis interaction with host cells (Gould et al. 2013). Upregulation of several proteins with potential roles in pathogenesis have also been found following host cell encounter (YP Chen et al. 2019; Coceres et al. 2015; de Miguel et al. 2010; de Miguel et al. 2012; Gould et al. 2013). Additionally, post-translational modifications may also regulate the activity of these T. vaginalis adhesins, as dephosphorylation by the phosphatase TvPP1γ and protein palmitoylation also affect parasite attachment (Munoz et al. 2012; Nievas et al. 2018b). Therefore, T. vaginalis likely employs a speedy and multi-faceted transcriptionally and post-translationally-regulated response to maximize colonization.

To date, the only T. vaginalis ligand and host receptor pair fully characterized to participate in attachment is T. vaginalis surface lipoglycan (TvLG) binding to the host cell galectin-1 protein (Bastida-Corcuera et al. 2005; Fichorova et al. 2016; Okumura et al. 2008; Ryan et al. 2011b; Singh et al. 1994). T. vaginalis contains ~3 × 106 TvLG molecules that form a dense glycocalyx on the parasite surface (Singh et al. 1994). Chemically-mutagenized parasites that have altered TvLG composition display reduced adherence and lysis of ectocervical cells (Bastida-Corcuera et al. 2005). Similarly, knockdown of galectin-1 on host cells caused a 20% decrease in parasite attachment (Okumura et al. 2008). Galectin-1 is a homodimer with two carbohydrate recognition domains, and is hypothesized to act as a bridge binding T. vaginalis TvLG on one end, and galactose on the host cell on the other end (Ryan et al. 2011a). Specifically, the TvLG glycoconjugate’s outermost N-acetyllactosamine (LacNAc) and lacto-N-biose (LNB) side chains help mediate the parasite binding to galectin-1 on host ectocervical cells (Ryan et al. 2011a).

Two additional potential adherence ligands are TVAG_140850, and TVAG_240680, which are upregulated upon host-cell encounter, and whose overexpression contributed to enhanced attachment to host cells (Gould et al. 2013; Handrich et al. 2019). Interestingly, these proteins are members of the Pmp and BspA families, respectively. These gene famlilies are highly expanded in the T. vaginalis genome, predicting a functional importance (Carlton et al. 2007; Hirt et al. 2011). Furthermore, exogenous expression of TVAG_140850 and TVAG_240680 in a bird-infecting Tetratrichomonas gallinarum, also increased ability of this trichomonad to bind vaginal epithelial cells (Handrich et al. 2019). However, the host cell receptor for these adhesins remains unknown.

T. vaginalis proteases also clearly play a role in cytoadherence and killing. It is estimated that the T. vaginalis genome encodes hundreds of peptidates,  with one study identifying 310, and another reporting 447 (Carlton et al. 2007; Hirt et al. 2011). Thus T. vaginalis has one of the largest predicted degradomes amongst eukaryotic protists. The use of protease inhibitors that target cysteine, serine, and metalloproteases has helped to reveal the involvement of proteases in T. vaginalis: host interactions. Protease inhibitors reduce the ability of T. vaginalis to attach to and/or lyse host cells (Alvarez-Sanchez et al. 2000; Arroyo and Alderete 1989; De Jesus et al. 2009; Hernandez-Gutierrez et al. 2004; Ma et al. 2011; Mendoza-Lopez et al. 2000; Puente-Rivera et al. 2017; Quan et al. 2014; Rendon-Gandarilla et al. 2013; Riestra et al. 2015; Sommer et al. 2005). Additionally, inhibiting protease function with blocking antibodies also decreases parasite attachment and/or host cell lysis (Alvarez-Sanchez et al. 2000; Cardenas-Guerra et al. 2013; Hernandez et al. 2004; Mendoza-Lopez et al. 2000; Miranda-Ozuna et al. 2019; Puente-Rivera et al. 2017; Rendon-Gandarilla et al. 2013; Rivera-Rivas et al. 2020). Furthermore, T. vaginalis protease expression and activity levels differ under varying levels of Fe2+, Zn2+, polyamines, oxygen, glucose, and between T. vaginalis strains with different virulence properties (Alvarez-Sanchez et al. 2008, 2007; Arroyo et al. 2015; Cuervo et al. 2008; De Jesus et al. 2009; Garcia et al. 2005; Gould et al. 2013; Hernandez-Romano et al. 2010; Kummer et al. 2008; Miranda-Ozuna et al. 2019; Puente-Rivera et al. 2017; Quintas-Granados et al. 2013; Rivera-Rivas et al. 2020). Therefore, proteases likely mediate and modulate T. vaginalis cytoadherence and host-cell killing, and certain proteases may mediate particular roles in different microenvironments within the host.

Specifically, cysteine proteases constitute the largest mechanistic and most characterized class of proteases in T. vaginalis (Carlton et al. 2007; Hirt et al. 2011). T. vaginalis cysteine proteases localize to lysosomes, the Golgi complex, cytosolic vacuoles, vesicles, the parasite cell surface, and can also be secreted (Hernandez-Gutierrez et al. 2004; Mendoza-Lopez et al. 2000; Rendon-Gandarilla et al. 2013; Riestra et al. 2015). Cysteine proteases are known to be active during natural infection, because vaginal lavages of infected women contain the active proteases, as well as antibodies raised against them. While cleavage of extracellular matrix proteins by some of these cysteine proteases has been reported (Alvarez-Sanchez et al. 2000; Hernandez-Gutierrez et al. 2004; Rendon-Gandarilla et al. 2013), how inhibition of T. vaginalis proteases leads to the observed effects on modulating T. vaginalis adherence to host cells and host cell lysis remains largely unknown. The identification of the parasite and/or host substrates during these processes will shed light on the mechanisms of action at play.

Gain of function studies utilizing recent advances in T. vaginalis molecular biology coupled with bioinformatics or proteomics have also allowed the discovery of additional T. vaginalis surface proteins and parasite factors involved in cytoadherence and killing. De Miguel and colleagues conducted a quantitative proteomics study that identified and compared the T. vaginalis surface proteins in highly adherent strains and weakly adherent strains (de Miguel et al. 2010). The study revealed 11 proteins that were more abundant in the highly adherent T. vaginalis strains. Overexpression of two of these proteins, TvBAP1 (TVAG_166850) and TvBAP2 (TVAG_244130), led to a >2 fold increase in attachment to the ectocervical cells. The putative intramembrane protease TvROM1 was also identified in the T. vaginalis cell surface proteome, and membrane localization was confirmed in a later study (Riestra et al. 2015). Exogenous expression of TvROM1 led to a 1.6 fold increase in parasite attachment and a 4.2 fold increase in lysis of ectocervical cells compared to empty vector control transfectants (Riestra et al. 2015). Riestra and colleagues also identified two substrates, TvBAP1 and TVAG_280090, cell surface proteins functionally annotated as hypothetical proteins. Mutation of the TvROM1 cleavage site in TvBAP1 led to increased TvBAP1 surface amounts. These findings suggest that T. vaginalis may modulate the presence of certain cell surface adhesins during parasite attachment and host cell lysis. It remains to be determined whether the observed effects on host cell attachment and host cell lysis are mediated by cleavage of the same or different parasite and/or cell surface proteins.

Structure predictions have helped to reveal additional T. vaginalis proteins involved at the parasite: host interface. Bioinformatic analysis of two surface proteins implicated in adherence, TvBAP1 and TVAG_280090, predicted cadherin-like tertiary structures (Riestra et al. 2015). In multicellular organisms, cadherin-like proteins mediate cell-cell interactions and cell-cell adhesion (Abedin and King 2010). Further searches for cadherin-like proteins in T. vaginalis, led to the functional characterization of TVAG_393390 (TvCLP). Exogenous expression of TvCLP led to a 3.5 fold increase in T. vaginalis adherence to the ectocervical cells and a 3.3 fold increase in T. vaginalis lysis of these cells, compared to empty vector transfectants (Chen et al. 2019). Calcium binding was also found to be necessary for TvCLP’s role in promoting host cell attachment and cytolysis. Investigating the evolutionary history of the cadherin-like proteins and possible homologs in other parasitic protozoans is of interest, as these may be evolutionary relics of mammalian cadherin-like proteins and have conserved functional roles in pathogenesis.

T. vaginalis may also utilize its putative large arsenal of virulence factors to respond differently in the female vs. male host. For example, differences in iron availability occur during female menses and several studies have revealed that iron can modulate the expression of various cysteine proteases (Arroyo et al. 2015; Figueroa-Angulo et al. 2012). In the male, prostate zinc levels are the highest compared to other organs (Cui et al. 2015). Quintas-Granado and colleagues identified and characterized a T. vaginalis metalloprotease, TvMP50 (TVAG_403460), that was recognized by sera collected from a T. vaginalis-infected male but was not detected by sera from a T. vaginalis-infected female or uninfected male and female controls, indicating that this metalloprotease is only expressed during infection in men. TvMP50 expression and proteolytic activity was also upregulated in the presence of increased Zn2+ concentrations. TvMP50 also caused killing of a prostate epithelial cell line (Puente-Rivera et al. 2017). Interestingly, an antibody raised against TvMP50, reduced the cytolytic activity of TvMP50 and the inhibitory effects were more pronounced under higher Zn2+ concentrations. Together, these studies highlight that Zn2+ concentrations may regulate the cytolytic activity of TvMP50 and that this metalloprotease could play an important and differential role in male trichomoniasis.

Lastly, the secretion of extracellular vesicles by T. vaginalis (described in Sect. 4.2) also helps to bolster parasite attachment to host cells. Exosomes isolated from highly adherent strains can increase the adherence of the weakly adherent G3 strain to ectocervical cells (Twu et al. 2013). Furthermore, T. vaginalis exosomes from two strains that display higher adherence to a prostate cell line compared to an ectocervical cell line (Lustig et al. 2013) similarly confer this preferential binding ability to the G3 strain. Therefore, exosomes may prime host cells for maximal T. vaginalis attachement and for successful colonization of the two different environments encountered within the male and female hosts.

4.1.3 A Closer Examination of Host Cell Death Induced by T. vaginalis

As described in Sect. 4.1.2, signatures of both host-cell necrosis and apoptosis have been detected by different groups of researchers following exposure to T. vaginalis. However, recent advances in the field of cell death have identified several additional cell death pathways, and revealed extensive crosstalk between different cellular death pathways (Galluzzi et al. 2018; Rogers and Alnemri 2019; Rogers et al. 2019; Yu et al. 2014), adding some nuance to the necrosis vs. apoptosis dichotomy. It is therefore likely that host cells may die in context and cell-type specific ways following T. vaginalis encounter, as we outline below, and summarize in Fig. 4. Furthermore, it may be necessary to build on the knowledge generated from the existing studies and perform future analysis of host cells en route to dying after contact with T. vaginalis, to better define how T. vaginalis induces cell death. Below, we discuss the evidence supporting each cell-death pathway in contexts where it was observed.

Fig. 4
figure 4

T. vaginalis cytoadherence and lysis of various host cell types. Schematic of the various cell types to which T. vaginalis attaches and kills via the various cell death pathways shown. T. vaginalis can also kill B-cells, T-cells, and red blood cells via contact-independent cytolysis. Figure created with BioRender.com

4.1.3.1 Host Cell Apoptosis

Researchers have found that DNA fragmentation, a key signature of apoptosis, occured in two cervicovaginal cell types following culture with T. vaginalis (Quan et al. 2014; Sommer et al. 2005). In fact, in the absence of whole-live parasites, secreted products from the parasites showed a similar effect (Sommer et al. 2005), indicating that cytoadherence-independent cell death could proceed via apoptosis during infection. Quan and colleagues also detected many other signs of apoptosis such as cleavage of the initiator caspase-9, the executioner caspase-3, the caspase-3 substrate PARP, and increased cytolosolic cytochrome c , all signatures of intrinsic cellular apoptosis (Rogers and Alnemri 2019). However, since detection of these and other markers of apoptosis largely relies on assessing cleavage products, T. vaginalis’ large degradome activity may complicate discernment of cell death pathways. Single cell analysis of dying host cells after encounter with T. vaginalis and secreted products is therefore necessary in order to fully understand the order of signaling events that induce host cell killing.

How T. vaginalis encounter with host cells could physically lead to the activation of intrinsic apoptosis also remains largely unknown. In the studies described above, Quan et al. identified that T. vaginalis metalloproteases contribute to activate apoptosis (Quan et al. 2014). A later study by Quan and colleagues identified that upon co-incubation of a cervical cell line with T. vaginalis, host cells increased production of intracellular and mitochondrial reactive oxygen species (ROS) production, and that inhibition of ROS production during infection with T. vaginalis caused reduced levels of apoptosis markers (Quan et al. 2017). Therefore, upon T. vaginalis infection of cervical epithelial cells, a T. vaginalis-generated signal as well as increased oxidative stress in the host cells could contribute to apoptosis activation.

4.1.3.2 Host Cell Pyroptosis

Pyroptosis, or “a fiery death,” is a newly discovered form of cell death that is commonly activated as part of the innate immune response to pathogens and host cell injury, and results in strong inflammatory signals being sent from the dying cell (Galluzzi et al. 2018). Pyroptosis starts when Pattern-Recognition Receptors (PRRs) in the cytosol form multi-protein complexes called inflammasomes. Inflammasome assembly in turn leads to the activation of inflammatory caspases which can cleave the gasdermin D protein (Kovacs and Miao 2017; Shi et al. 2015). Importantly, the gasdermin D N-terminal cleavage fragment can then insert itself in the plasma membrane, oligomerize, and form pores (Ding et al. 2016), thus causing cell death (Kovacs and Miao 2017; Liu et al. 2016; Shi et al. 2015). The overall cellular features of pyroptosis are chromatin condensation, cellular swelling, and fast cell lysis (Galluzzi et al. 2018). Additionally, pyroptosis is almost always accompanied with secretion of the pro-inflammatory cytokines IL-1β and IL-18 (Galluzzi et al. 2018; He et al. 2015; Shi et al. 2015).

As discussed in Sect. 4.3.2, two separate research groups found that T. vaginalis induces inflammasome activation in a prostate epithelial, and macrophage cell line, respectively (Gu et al. 2016; Riestra et al. 2019). Both groups detected an increase in processing of the pyroptosis-specific caspase-1. During pyroptosis, the gasdermin D pores that assemble in the host cell membrane are 10–14 nm in diameter and thus proteins that are smaller than the pore size, such as LDH (8-10 nm), can exit the cell (Ding et al. 2016; He et al. 2015). Riestra and colleagues found that inflammasome and caspase-1 inhibition caused a reduction in LDH release after co-incubation of THP-1 macrophages with T. vaginalis. Furthermore, T. vaginalis also had a reduced ability to lyse gasdermin D knockout cells compared to control cells, further supporting pyroptosis as the mechanism of macrophage host-cell killing. The authors of these studies thus hypothesized that T. vaginalis induces pyroptosis in host cells. Further investigation at a morphological level is warranted to more fully characterize the induction of pyroptosis after contact with T. vaginalis and to investigate the effects of inflammasome activation on the host.

Recently, it has also been identified that inflammasome activation can lead to apoptosis, because the N-terminal gasdermin D cleavage fragment can also form pores in the mitochondrial membrane, contributing to mitochondrial outer membrane permeabilization and the classic downstream apoptotic events such as cytochrome c release (Rogers and Alnemri 2019; Rogers et al. 2019; Yu et al. 2014). Chang et al. detected increased Annexin V staining, cytochrome c release into the cytosol, and apoptotic caspase activity after co-incubation of a macrophage cell line with T. vaginalis (Chang et al. 2004), all indicators of apoptosis. However, nucleosomal degradation and Annexin V+ staining are also detected during pyroptosis, as gasdermin D pores allow Annexin V access to the inner leaflet of the plasma membrane (Abe and Morrell 2016; Man and Kanneganti 2016). Therefore, it remains to be determined, whether induction of apoptosis downstream of pyroptosis may reconcile signs of pyroptosis versus apoptosis observed in macrophage cell lines exposed to T. vaginalis, or whether each type of cell death is induced depending on differential conditions or parasite strains.

4.1.3.3 Haemolysis

T. vaginalis also lyses red blood cells (RBCs) and its hemolytic activity has been predicted to provide the parasite with a source of iron and lipids (Fiori et al. 1996), however the precise molecular interactions that drive RBC lysis also remain to be identified. Using scanning electron microscopy, Fiori and colleagues as well as Rosset and colleagues reported that tight contact is observed when T. vaginalis interacts with RBCs (Fiori et al. 1997; Rosset et al. 2002). Studies have also highlighted the potential involvement of both contact-dependent and contact-independent haemolysis (Fiori et al. 1993, 1996, 1997; Pindak et al. 1993; Rosset et al. 2002). However, the experimental conditions of how T. vaginalis secretions were collected was critical for detection of contact-independent haemolysis, as this experimental variable affected the final pH of the T. vaginalis supernatants, and a more acidic pH was found to maximize haemolytic activity (Fiori et al. 1996; Pindak et al. 1993). Researchers have also found that osmoprotectants can reduce or completely inhibit the lysis of RBCs by T. vaginalis (Fiori et al. 1993, 1996). Thus it has been hypothesized that haemolysis may be driven by T. vaginalis pore-forming proteins that insert into the RBC membrane, and that a low pH that is locally induced by T. vaginalis adherence is needed in order to trigger this haemolytic activity (Fiori et al. 1993, 1996). Analysis of the T. vaginalis genome by Hirt and colleagues has revealed the presence of 12 genes coding for putative proteins belonging to the saposin-like protein family (TvSAPLIP), which display homology to pore-forming proteins of the pathogenic protists Entamoeba histolytica and Naegleria fowlerii. Some TvSAPLIPs have been detected in T. vaginalis secretions (Riestra et al. 2015; Štáfková et al. 2018), thus TvSAPLIP family members may participate in cell lysis including haemolysis (Hirt et al. 2011).

The weakening of the membrane-associated cytoskeleton of RBCs appears to be a key event in T. vaginalis-induced heamolysis. Fiori and colleagues showed that a non-secreted protease of T. vaginalis specifically targets and degrades the RBCs spectrin, the main component of membrane cytoskeleton: this step takes place before haemolysis occurs (Fiori et al. 1997). Recently, Cardenas-Guerra and colleagues identified that the T. vaginalis cysteine protease TvCP4 may also help to mediate haemolysis, as pre-incubation of the parasite with an anti-CP4 antibody or with a cysteine protease inhibitor caused a reduction in RBC lysis by T. vaginalis (Cardenas-Guerra et al. 2013). Future studies will help to reveal whether shared features exist between how T. vaginalis lyses RBCs compared to other host cell types.

4.1.3.4 Lymphocyte Lysis

Similar to the lysis of RBCs, researchers have also detected contact-dependent and contact-independent lysis of B-cells and T-cells (Mercer et al. 2016), as discussed in Sect. 4.3.5. In these studies, a highly adherent clinical isolate could lyse about 30% of T cells and about 70% of B-cells during a 4 h co-incubation, whereas the G3 lab-adapted T. vaginalis strain caused less B-cell lysis and no T-cell lysis. Therefore, T. vaginalis appears to have a preference for lysing B-cells. Furthermore, when physical contact between the parasite and lymphocytes was prevented via insertion of a membrane filter, cell lysis of both cell lines was reduced by two-fold, indicating that contact-independent factors also contribute to the killing of B- and T-cells by T. vaginalis. To date, no specific T. vaginalis factors that mediate the initial contact or downstream lysis of lymphocytes have been identified.

A plethora of molecular determinants that facilitate contact with host cells and killing of host cells have been identified. Future investigations now rely on integrating the various processes and players that have been highlighted to contribute to pathogenesis, and dissect how and when they are specifically used to mediate host cell killing and how exactly that cell death ensues. Extracellular vesicles, host commensals, T. vaginalis’ symbionts, and the host immune system are additional layers to be considered at the host-pathogen interface and likely to contribute to parasite pathogenesis as well, as we discuss below.

4.2 T. vaginalis Extracellular Vesicles

Exciting work in the past decade has revealed that T. vaginalis secretes Extracellular Vesicles (EVs), that the parasite putatively uses for communication among different trichomonads and with host cells, during infection (Nievas et al. 2018a; Twu et al. 2013). For a discussion of how EVs affect T. vaginalis cytoadherence please see Sect. 4.1, and for a discussion of how they have immunomodulatory roles see Sect. 4.3. Here, we will discuss their identification, biogenesis, contents, and mechanism of delivery to host cells.

EVs are a membrane-bound secretome product that contribute to the adaptive capabilities of microbial cells through delivery of proteins, lipids and nucleic acids to both adjacent and distant cells. The international society of extracellular vesicles (ISEV) endorses “extracellular vesicles” as the generic term for particles naturally released from the cell that are surrounded by a lipid bilayer and cannot replicate (Thery et al. 2018). Depending on their size and mechanisms of biogenesis, EVs are classified into three major categories: exosomes, microvesicles, and apoptotic bodies. Exosomes range in size from 30 to 150 nm in diameter, and are formed by the fusion of intracellular multivesicular bodies with the plasma membrane. Ectosomes, or microvesicles (MVs), are 100–1000 nm in diameter, and are shed directly from the plasma membrane (van Niel et al. 2018). Apoptotic bodies are larger than 1000 nm in size and released during apoptotic cell death (Caruso and Poon 2018). Despite apparent differences in the mechanism of formation, it is experimentally difficult to discriminate among these vesicle populations after they are secreted from cells, due to the overlap in their sizes. However, as researchers identified more vesicular proteins, it became apparent that EVs contain a specific subpopulation of proteins rather than a random pool of molecules from their cells of origin (Choi et al. 2015), indicating that it should be feasible to define a panel of markers for different kinds of EVs. Despite the enrichment of EV contents, there is currently no clearly descriptive physical property or molecular marker that can unambiguously distinguish exosomes from MVs (Choi et al. 2015). Hence, assigning specific functions to the different EVs populations remains a challenge.

Nonetheless, the analysis of T. vaginalis EVs has become an active endeavor in the study of pathogenesis for this parasite, and so far, the release of both exosomes and MVs has been documented (Nievas et al. 2018a; Olmos-Ortiz et al. 2017; Twu et al. 2013). Twu and colleagues isolated vesicles smaller than 200 nm (with a mean diameter of 95 nm) mainly enriched in exosomes, and Nievas and colleagues designed a protocol to isolate vesicles from the same highly adherent strain (B7RC2) ranging between 200 and 800 nm (with a mean diameter of 380 nm) enriched mainly in MVs (Nievas et al. 2018a), indicating that the same strain of T. vaginalis can secrete both exosomes and MVs. Interestingly, the secretion of both types of vesicles increases upon exposure of the parasite to host cells, suggesting that EVs play an important role in pathogenesis (Nievas et al. 2018a; Twu et al. 2013). To begin to make predictions about their function, researchers performed proteomic analysis of these vesicles, and identified 215 and 592 proteins in exosomes and MVs proteomes, respectively. Interestingly, there was a good deal of overlap, as 167 out of 215 exosome proteins were also found in the MVs proteomic survey, similar overlap has also been observed in other studies (Abels and Breakefield 2016). However, it is important to consider that a minimum level of contamination is inevitably present in the samples due to vesicle aggregation. Nevertheless, as 425 proteins were unique to MV’s, this could indicate that these vesicles may be selectively loaded with specific groups of proteins that exert specific functions. Many of the molecules identified in both proteomes are commonly found in EVs proteomes from other organisms. These include members of the vesicle trafficking-related complex such as ESCRT factors (VPS32 and CHMP1) (Iriarte et al. 2018; Schmidt and Teis 2012)⁠; Rab GTP-ases and SNARE-complex proteins (VAMP, syntaxin, Vti1) (Fader et al. 2009; Ostrowski et al. 2010)⁠; cytoskeletal proteins (actin, tubulin, villin, plastin) (Mathieu et al. 2019)⁠; metabolic enzymes (enolases, malic enzyme, glyceraldehyde 3-phosphate dehydrogenase); ribosomal proteins and tetraspanins (TvTSP1 in exosomes (Twu et al. 2013) and TvTSP8 in MVs (Nievas et al. 2018a).

Tetraspanin proteins (CD9, CD63, and CD81) are often used as markers of EVs in mammalian cells (Kowal et al. 2016; Mathivanan et al. 2010). Tetraspanins are transmembrane proteins that organize specialized membrane domains termed tetraspanin-enriched microdomains (TEMs) which may play a role in EVs biogenesis, cargo selection and/or binding and uptake by target cells (Andreu and Yanez-Mo 2014). Considering the presence of the two tetraspanins, TvTSP1 and TvTSP8 in the exosome and MVs proteome, respectively (Nievas et al. 2018a; Twu et al. 2013), these proteins could potentially be used as markers of EVs in T. vaginalis. However, it is unlikely that TSP proteins will be useful as general markers of all protozoan EVs, as they are present in some, but not all protozoan genomes. Specifically, TSP-like sequences have been found in E. histolytica, Leishmania spp., Trypanosoma cruzi, and Trypanosoma brucei, but no TSP-like sequences were identified in Giardia spp., Plasmodium spp., Toxoplasma gondii, or Cryptosporidium parvum (HMMer search using Pfam PF00335). Considering the role of TSPs in EV biogenesis in mammalian cells, it would be interesting to evaluate if the mechanism of EV formation differs among parasites that contain or lack TSPs within their genomes. For example, it has been demonstrated that TvTSP1 accumulates in Multivesicular Body (MVB)-like structures after T. vaginalis is exposed to vaginal epithelial cells (Twu et al. 2013). This accumulation of TvTSP1 in MVBs upon host cell exposure could suggest a role of the protein in cargo selection or biogenesis of exosomes. It is interesting to note that the expression of both TvTSP1 and TvTSP8 are up-regulated upon exposure of the parasites to the host cell (Coceres et al. 2015); indicating that they might be regulating the process of T. vaginalis: host interaction⁠s. On the other hand, TvTSP8 is highly expressed in adherent compared to less adherent T. vaginalis strains and overexpression of the surface-localized full-length TvTSP8 promoted increased parasite aggregation (Coceres et al. 2015) (discussed in Sect. 4.1.1⁠). This clumping phenotype is exhibited by some parasite strains, and correlates with their capacity to adhere to host cells (Lustig et al. 2013; Nievas et al. 2018a).

Besides tetraspanins, T. vaginalis EV proteomics analysis of EVs uncovered an enrichment for other proteins that have been implicated in host–parasite interactions and immunomodulation (Nievas et al. 2018a, 2020; Twu et al. 2013). Specifically, these include metalloproteases that are hypothesized to be involved in the infection process (Ma et al. 2011), BspA proteins that participate in host cell adherence (Handrich et al. 2019; Noel et al. 2010)⁠, the surface immunogen protein p270 (Alderete 1999), TvMIF that induces inflammatory responses (Twu et al. 2014), and proteins that are highly expressed in adherent strains (TVAG_020780 and TVAG_239650), albeit with unknown function (de Miguel et al. 2010; Gould et al. 2013). In order to fully understand the function of these proteins in EVs during infection, it would be interesting to analyze the protein content of the different vesicle populations in the presence and absence of host cells.

How host cells internalize EVs secreted by T. vaginalis is another important question. Recently, Rai and colleagues demonstrated that T. vaginalis exosomes fuse with and are internalized by host cells via caveolae and lipid raft-dependent endocytosis (Rai and Johnson 2019). Additionally exosomes were found to interact with glycosaminoglycans on the surface of host cells and to specifically bind heparan sulfate proteoglycans (HSPGs), indicating that HSPGs on host cells act as receptors that mediate the first step to internalize EVs (Rai and Johnson 2019). A ligand on the surface of EVs, named 4-α-glucanotransferase (Tv4AGT), was also identified to be critical for EV uptake (Rai and Johnson 2019). Interestingly, this protein is also present in the MV proteome (Nievas et al. 2018a), suggesting that mechanisms used to drive T. vaginalis: host interactions could be conserved among both types of vesicles. Future comparative analyses of the internalization pathways for different types of EVs will be necessary to fully understand the mechanisms involved in EV-mediated cell: cell communication.

4.3 Host Immune Responses to T. vaginalis and Immune-Evasion by the Parasite

Our understanding of how the host immune response to T. vaginalis proceeds, whether it effectively clears the parasite, and whether effective immunological memory forms against the parasite is certainly undercharacterized. However, researchers have gathered some information from in vitro studies and clinical correlations. While T. vaginalis is exclusively a human-infective pathogen, some preliminary murine models have also given insight into mammalian immune responses in vivo. Here we will discuss evidence of the innate and adaptive immune responeses to T. vaginalis, and findings that indicate immune-evasive behaviors of the parasite.

4.3.1 Cytokine Responses to T. vaginalis

Cytokines are important signaling proteins secreted from immune cells that are used to orchestrate immune responses. It has long been known that the pleitropic cytokine Interleukin-8 (IL-8) is released from human cells in response to T. vaginalis infection (Ryu et al. 2004; Shaio et al. 1994, 1995). Mice do not encode for the IL-8 gene, therefore it is not possible to test the effects of IL-8 in an experimental murine model. However, a cross-sectional study involving 65 women recently showed an increase in cervical IL-8 levels in T. vaginalis-exposed women compared to controls (Jarrett et al. 2015), giving in vivo validation to the in vitro observations of IL-8 secretion. Recent reports have also shown the cytokines IL-1β, CCL2, IL-17, IL-22, IFN-β, IL-6, RANTES, MIP-3α, TNF, and IL-23 to be associated with T. vaginalis infection either in vitro or in vivo (RN Fichorova et al. 2016; Fiori et al. 2013; Gu et al. 2016; Makinde et al. 2013; Mercer et al. 2016; Seo et al. 2014). Parasite-produced exosomes were also shown to induce IL-8 and IL-6 from host cells (Twu et al. 2013).

The cytokines released from cells that encounter T. vaginalis can give us clues about immune responses that form against the parasite. IL-1β, TNF, and IL-6 are all important cytokines that can get released systemically and act as fever-inducers (Garlanda et al. 2013). IL-1β, also indirectly promotes the recruitment of neutrophils, the first line of defense against most microbes, to the infection site (Slaats et al. 2016), and IL-8 does this directly. IL-1β, IL-6, and IL-23 are polarizing cytokines promoting the differentiation of T-helper 17 (Th17) cells (Z Chen and O’Shea 2008), pointing to a possible role of Th17 cells in T. vaginalis immunity, similar to Giardia lamblia, another mucosal-dwelling protozoan (Singer et al. 2019). RANTES, MIP-3α, and CCL2 are all involved in T-cell recruitment, and could thus be important for execution of T-cell responses against T. vaginalis. TNF and IL-1β are broadly inflammatory and consistent with the vaginitis and prostatitis associated with T. vaginalis infection.

As mentioned in Part 1, T. vaginalis is associated with increased spread of HIV. While the mechanism by which the parasite could promote enhanced HIV infection has not been well-characterized, recent data indicate that cytokines could play a role. T. vaginalis induces the inflammatory cytokines RANTES and MIP-3α (Fichorova et al. 2012), which are chemotactic ligands for the receptors CCR5 and CCR6 respectively. Therefore, T. vaginalis infection could promote the recruitment of CCR5+ and CCR6+ cells to the vaginal mucosa. As T-cells have high levels of expression of these two receptors, and CCR5 is a co-receptor for HIV entry, the parasite could increase the recruitment of HIV target cells to the vaginal mucosa, thus increasing susceptibility to HIV infection. Interestingly, RANTES and MIP-3α were found to be released only in response to T. vaginalis strains that harbor the symbiont Trichomonasvirus (discussed in Sect. 5.1) (Fichorova et al. 2012), indicating that Trichomonasvirus could faciliate increasing T. vaginalis-associated HIV contraction.

In fact, a majority of the cytokine induction by T. vaginalis appears not to be due to the parasite itself, but to T. vaginalis symbionts Trichomonasvirus or Mycoplasma hominis (described in Sect. 5.1). When Fichorova and colleagues compared a panel of strains harboring Trichomonasvirus to strains without Trichomonasvirus, they observed that Trichomonasvirus+ strains were greatly enhanced in their ability to induce IFN-β, IL-8, IL-6, RANTES, MIP-3α and IL-1β in cervicovaginal epithelial cells (Fichorova et al. 2012), with strains lacking the virus showing almost no cytokine induction. Interestingly, this effect was enhanced in the presence of metronidazole (discussed in Sect. 7.1), a drug that kills T. vaginalis, presumably due to release of virions from dead parasites that could then more easily access host cell viral recognition receptors like TLR3. Similarly, Fiori and colleagues found that that M. hominis triggered the release of IL-8, IL-23, and TNF from a macrophage cell line, while T. vaginalis alone only stimulated small amounts of IL-1β. However, the M. hominisT. vaginalis consortium was able to induce greater amounts of cytokine than M. hominis or T. vaginalis alone, suggesting a synergism between the two microbes in inducing cytokines (Fiori et al. 2013). Other researchers also observed an M. hominis-dependent induction of IL-1β, IL-8, and IL-6 in primary human monocytes (Mercer et al. 2016). Each of these studies indicate that none, or very little cytokine is induced from the parasite alone. However, these studies may not be sensitive enough to detect cytokine induced from the parasite alone. Furthermore, during infection, the parasite likely elicits tissue damage, releasing dead-cell products that also likely stimulate cytokine induction. In addition, studies have not yet considered both symbionts at the same time. It would therfore be interesting to test a panel of strains containing both symbionts, or cleared of either, to better define the true contribution of each symbiont to cytokine stimulation. Data from Fiori and collagues, which used a T. vaginalis isolate naturally Trichomonasvirus- and M. hominis-free to stimulate a macrophage cell line, saw only weak activation of NF-κB and small amounts of IL-1β indicating weak immunostimulatory activity of the parasite in the absence of its symbionts. This might indicate that vaccines containing T. vaginalis immunogens need strong adjuvants since the parasite alone may not elicit strong responses.

4.3.2 T. vaginalis Recognition by the Innate Immune System

To determine how host immunity first perceives a T. vaginalis infection, an important question is whether the parasite stimulates Pattern-Recognition-Receptors (PRRs) on innate immune cells, and if so, which one(s). PRRs are essential for initial recognition of pathogens by the innate immune system and influence antimicrobial cellular responses and cytokine secretion by innate immune cells. The PRR Toll-Like Receptor −4 (TLR4) was posited to be triggered by T. vaginalis, because T. vaginalis was found to be less immunogenic in mutagenized mouse cells lacking TLR4 (Zariffard et al. 2004). However, it is not clear if these cells had other deficiencies besides lacking functional TLR4, or whether the parasites used in these experiments harbored T. vaginalis symbionts, which could confound the results. Furthermore, Fichorova and colleagues reported induction of cytokine secretion from purified T. vaginalis dominant surface lipoglycan (TvLG), in cervicovaginal epithelial cells that do not express detectable levels of TLR4, and instead proposed a TLR4-independent mode of T. vaginalis recognition (RN Fichorova et al. 2006). The host lectin receptor, Galectin 3, has recently been shown to modulate IL-8 secretion through engagement of TvLG (Fichorova et al. 2016). Since host cell galectins are becoming appreciated as PRRs (Vasta 2012), the TvLG-Galectin interaction is an exciting area of future research into how T. vaginalis initially activates the innate immune system.

Furthermore, the observation that IL-1β is induced from host cells by T. vaginalis sparked interest in a potential role for the inflammasome in anti-T. vaginalis immune responses. IL-1β processing and secretion is largely driven by assembly of cytosolic PPRs called Nod-Like-Receptors (NLRs), that form multiprotein complexes called inflammasomes. Inflammasomes can activate caspase-1, which leads to cleavage of the pro-form of IL-1β to its bioactive form. IL-1β is then released from host cells as they die through pyroptosis (discussed in Sect. 4.1.3). Recently, researchers identified that T. vaginalis can activate the NLRP3 inflammasome, since knockdown or inhibition of caspsase-1 and NLRP3, decreased the amount of IL-1β observed in response to T. vaginalis (Gu et al. 2016; Riestra et al. 2019), and specifically the bioactive form (Riestra et al. 2019). Furthermore, Riestra and colleagues also detected the production of IL-1β in a murine model of T. vaginalis infection (Riestra et al. 2019), indicating that inflammasomes are activated by T. vaginalis in vivo. Additionally, a murine gut trichomonad, Tritrichomonas muris was observed to induce an inflammasome- dependent program of immunity in vivo (Chudnovskiy et al. 2016). However, it remains to be investigated, whether inflammasome activation contributes to control of T. vaginalis infection or whether it leads to chronic inflammation. It will also be interesting to determine whether T. vaginalis strains vary in their potency to activate inflammasomes.

4.3.3 Innate Immune Responses to T. vaginalis

While the innate immune response to T. vaginalis is not fully characterized, in vitro work has shown that phagocytes (mainly neutrophils) and the complement system are likely to be important for immune clearance of the parasite. As mentioned above, IL-8, which has functions in neutrophil recruitment, is the most abundant cytokine induced by T. vaginalis in in vitro assays, and has been found at the site of infection (Cauci and Culhane 2007; Lazenby et al. 2013; Mercer et al. 2016; Shaio et al. 1994, 1995). In 1980, researchers observed that T. vaginalis is killed by neutrophils in vitro after they swarm the parasite in the presence of serum, indicating that neutrophils work in aggregate, cooperating to kill T. vaginalis (Rein et al. 1980). Researchers also found that supernatants from T. vaginalis cultured with epithelial cells induce neutrophil migration over a filter (Seo et al. 2014), a proof-of-principle experiment that neutrophils chemotax towards T. vaginalis infection products. Collectively, these results support a crucial role for neutrophils in elimination of T. vaginalis.

More recently, researchers have sought to determine the specific cellular mechanism(s) that neutrophils use to kill T. vaginalis. Neutrophils are known to have 3 classical modes of killing pathogens, all of which involve the neutrophil granules, which are vesicles present in neutrophils that contain toxic substances like pore forming toxins, reactive oxygen species, and enzymes (Kolaczkowska and Kubes 2013). The first mode is phagocytosis, in which neutrophils engulf pathogens whole. The vacuole containing the pathogen then fuses with lysosomes or neutrophil toxic granules to kill the pathogen. The second mechanism is extracellular degranulation, in which the toxic granules are exocytosed to intoxicate pathogens extracellularly. The third mechanism is by using Neutrophil Extracellular Traps (NETs), a process termed NETosis, in which the neutrophil ejects its DNA, producing an unraveled NET studded with the contents of toxic granules so that ensnared pathogens are intoxicated by granule contents. Generally, phagocytosis and extracellular degranulation are very rapidly employed (within minutes), while NETosis is a late-stage killing mode, possibly employed by neutrophils as a last-ditch effort if the other two mechanisms have failed.

When researchers tested which mode of killing neutrophils employ against T. vaginalis, they found that the parasite was killed very rapidly and in a contact and engulfment-dependent process, but that as opposed to phagocytosis, neutrophils mainly kill T. vaginalis by taking piece-meal bites of the parasite surface until it dies (Mercer et al. 2018). Such cellular “nibbling” is termed trogocytosis (trogo = to nibble), contrasted with traditional phagocytosis (phago = to eat), in that the target does not get engulfed whole (Ralston 2015a). It is not yet known whether the neutrophil toxic granules are involved in generating the nibbles upon cell-cell contact, or whether the neutrophil has other mechanisms to degrade the parasite surface. However, since neutrophil trogocytic killing of T. vaginalis was almost completely inhibited when neutrophils were pre-incubated with a serine protease inhibitor, it is hypothesized that serine proteases present in neutrophils may contribute to trogocytic killing (Mercer et al. 2018). In addition, since human serum is required for neutrophil trogocytic killing of T. vaginalis (Mercer et al. 2018), it is likely that complement proteins, reactive proteins found in serum that can mediate pathogen killing, mediate neutrophil binding to T. vaginalis to initiate trogocytosis. Complement proteins have also been shown to directly kill some strains of the parasite (Gillin and Sher 1981). While complement proteins and trogocytosis by neutrophils are likely to be rapid killing mechanisms used against T. vaginalis, it is not yet known whether neutrophils can also kill the parasite using NETosis, at later time points.

4.3.4 Adaptive Immune Responses to T. vaginalis

It is not known if, or how often, effective immunological memory to T. vaginalis develops in hosts. Antibodies that recognize the parasite or its components, have been detected in patients, indicating that an adaptive immune response is formed. For example antibodies reactive against TvMIF have been detected in patients (Kaur et al. 2008; Paintlia et al. 2002; Twu et al. 2014) and antibodies against the cysteine proteases TvCP2, TvCP4, TvCPT, and TvLEGU-1 have also been identified in vaginal lavages and in serum collected from T. vaginalis-infected women (Alvarez-Sanchez et al. 2000; Hernandez-Gutierrez et al. 2004; Mendoza-Lopez et al. 2000; Ramon-Luing et al. 2010). Additionally, both IgG and IgA antibodies specific to the dominant surface lipoglycan TvLG (Bastida-Corcuera et al. 2013), have been found in both the blood and vaginal secretions of infected patients. Furthermore, a serological study conducted on a cohort of patients in Vietnam described the kinetics of circulating antibodies in response to T. vaginalis. Successful treatment of the infection resulted in a relatively fast decline of ciruculating antibody titers, while antibody persistence was related to possible drug resistance, or alternatively to “ping-pong” reinfection by untreated sexual partners, thus suggesting that the humoral response to T. vaginalis might be transient and not entirely protective (Ton Nu et al. 2015). In support of this, several studies have noted higher rates of rebound trichomoniasis when partners were not treated, further indicating that immunological memory may not effectively form against T. vaginalis (Forna and Gulmezoglu 2003). Interestingly, adverse pregnancy outcomes are associated with lower vaginal IgG, indicating that IgG could be protective, without inciting immunopathology (Bastida-Corcuera et al. 2013). Therefore, future prophylactic vaccines may require boosters or strong adjuvants to achieve protection.

Furthermore, neutrophil trogocytic killing of T. vaginalis is dependent on human serum, and roughly 50% of neutrophil trogocytic activity was inhibited in the presence of Fc proteins (Mercer et al. 2018), which compete for antibody receptors. It is therefore likely that antibodies play a role in neutrophil killing of T. vaginalis, highlighting that neutrophil trogocytosis is part of the adaptive immune process against T. vaginalis as well.

CD4+ T-cells are also thought to be recruited to the site of infection, as T-cell recruitment cytokines are induced from other cells by the parasite (Fichorova et al. 2012) and CD4+ T-cell levels increase in the vaginal mucosa after parasite inoculation in a mouse model (Smith and Garber 2015). How T-cells respond to T. vaginalis is not known, and which T-helper cell subsets differentiate and help shape immune responses to T. vaginalis has not been formally tested, however IL-17 and IL-22 responses have been reported as elevated in T. vaginalis + women compared to non-infected controls (Makinde et al. 2013), indicating that Th17 responses could play a role. Since major functions of a Th17 adaptive response is promotion of neutrophil recruitment and combating extracellular pathogens, this is a logical hypothesis. Of note, the presence of T. vaginalis’ symbionts (discussed in Sect. 5.1) will likely also influence T-cell polarization.

4.3.5 T. vaginalis Subversion of Host Immune Responses

The relatively high percentage of partner re-infection observed in some studies (Forna and Gulmezoglu 2000), and the long-term persistence of some infections, indicates that many strains of T. vaginalis are able to evade the immune response. Studies of neutrophil trogocytic killing show that on average, five neutrophils were found to surround one trichomonad during killing in vitro (Mercer et al. 2018). Therefore, it is highly likely that the routinely observed behavior of T. vaginalis forming microcolonies (i.e. clumping) (Coceres et al. 2015; Nievas et al. 2018b) discussed in Sect. 4.1, could serve as a neutrophil-evasive behavior, making it challanging for neutrophils to surround individual trichomonads. Additionally, T. vaginalis has been reported to induce apoptosis of neutrophils in vitro, indicating that killing of neutrophils could be a strategy that some strains use in some circumstances to evade killing (Song et al. 2008). Furthermore, T. vaginalis produces cysteine proteases that show the ability to degrade antibody proteins in vitro, suggesting that the parasite could evade trogocytosis and other modes of antibody-dependent killing (Hernandez et al. 2014; Provenzano and Alderete 1995).

As mentioned above, while neutrophils employ trogocytosis as an early-stage killing mechanism against T. vaginalis, it is not yet known whether neutrophils can also kill T. vaginalis using NETosis, at later time points. However it is interesting that a DNAse was recently discovered to be part of the T. vaginalis constitutive secretome (Stafkova et al. 2018), indicating that the parasite could harbor a strategy to evade killing by NETosis. Furthermore, the T. vaginalis symbiont M. hominis (discussed in Sect. 5.1) was recently discovered to express a surface nuclease, MH_0730, which was also found to be capable of degrading DNA (Cacciotto et al. 2019). In the same paper, researchers showed that M. hominis could also degrade NETs in vitro (Cacciotto et al. 2019). Therefore, one interesting possibility is that while M. hominis may make T. vaginalis more immunogenic because it induces more cytokine secretion, it may also help the parasite evade killing by neutrophil NETosis.

Another potential mechanism of immune-evasion could be the cytolytic properties of the parasite. T. vaginalis was shown to be able to kill lymphocytes in vitro (Mercer et al. 2016). Interestingly, strains that were more lytic to epithelial cells (Lustig et al. 2013), were also found to be lytic to lymphocytes, indicating that these strains thought to be more pathogenic could also evade adaptive immunity by killing its cellular effectors. Furthermore, this lytic activity preferentially targeted B-cells, the producers of antibody, further supporting the hypothesis that T. vaginalis employs evasive behaviors against antibodies.

Interestingly, T. vaginalis has also been found to make a cytokine mimic, and a mimic of leukotriene, a chemical mediator of inflammation. TvMIF, a homolog of human MIF (Macrophage Migration Inhibitory Factor), is secreted from the parasite, and can bind the human MIF receptor, increasing host inflammatory cytokine production (Twu et al. 2014). It is not known whether TvMIF-induced inflammation promotes T. vaginalis clearance, or contributes to immune subversion through distraction, as previously observed for a MIF ortholog encoded by the unicellular parasite Plasmodium berghei (Sun et al. 2012). Better understanding of effective immune responses against the parasite will help to elucidate this. However, TvMIF was found to increase T. vaginalis survival by inhibiting apoptosis in parasites under nutrient stress (Chen et al. 2018), indicating that TvMIF could play an indispensable role in T. vaginalis survival during infection, potentially overshadowing any negative role associated with enhanced immune-recognition. T. vaginalis is also thought to secrete a homolog of Leukotriene B4 (Nam et al. 2012), which has activity in neutrophil recruitment. It is not yet clear what function T. vaginalis-derived LTB4 has in infection.

EVs secreted by T. vaginalis (discussed in Sect. 4.2) likely also contribute to immune subversion. While exosomes secreted by T. vaginalis were shown to induce inflammatory cytokine release from human cells (Twu et al. 2013) indicating that they can mobilize immunity against T. vaginalis, several reports have also shown evidence that T. vaginalis exosomes could be immuno-suppressive. Twu and colleagues demonstrated that pre-treatment with T. vaginalis exosomes reduced subsequent IL-8 responses to the parasite itself in vitro (Twu et al. 2013), indicating that exosomes could “prime” more distal parts of the vaginal mucosa to be tolerant toward the parasite in vivo. Moreover, another group found that pre-treatment of mice with T. vaginalis exosomes reduced IL-17 secretion and gross inflammation scores during infection, indicating that exosomes could promote reduced inflammation towards the parasite (Olmos-Ortiz et al. 2017). Moreover, two reports found that exosomes or secretory products from T. vaginalis increased IL-10 production (Olmos-Ortiz et al. 2017; Song et al. 2015), and another report found IL-10 induction from bovine cells with the bovine vaginal trichomonad Tritrichomonas foetus (Vilela and Benchimol 2013). Since IL-10 is an anti-inflammatory cytokine promoting tolerance to microbes, T. vaginalis exosome induction of IL-10 production could be another mechanism by which the parasite evades immune clearance. T. vaginalis exosome generation could therefore be a mechanism that the parasite uses to dampen host immunity, allowing it to establish infection similar to the secretion of virulence factors and vesicles by bacteria (Kuehn and Kesty 2005; Kulp and Kuehn 2010). It is not clear whether anti-inflammatory properties of exosomes could also have beneficial anti-inflammatory roles for the host, or whether they merely thwart effective clearance of the parasite.

T. vaginalis can modify the local environment aiding its ability to attach to and lyse cells. In parallel, the immune response is also mobilized to control infection. The spatial and temporal nature of these processes and how they influence one another now remains to be further explored in vivo. Furthermore, during host colonization T. vaginalis also has important interactions with the microbes residing in the urogenital tract, as described next.

5 Part 5: T. vaginalis as a Member of a Rich Microbial Ecosystem

A real view of host-parasite interactions, particularly for extracellular pathogens, would be incomplete without including the other microbes indigenous to the site of infection and colonization. Interestingly, in addition to members of the urogenital microbiome, T. vaginalis also carries its own microbial symbionts (bacteria and viruses). We will first discuss the bacterial and viral partners of T. vaginalis, that often hitch-a-ride during infection and are undoubtedly important contributors to infection biology. We then look further outward, considering T. vaginalis’ interaction with members of the vaginal microbiome and how the parasite may participate in driving dysbiosis and/or exploit this microbial shift  for its own benefit.

5.1 Symbionts of T. vaginalis

A distinctive and peculiar feature of T. vaginalis is its unique ability, among human pathogenic protists, to establish symbiotic relationships with eubacteria belonging to the Mycoplasma genus (Dessi et al. 2019), namely Mycoplasma hominis and Candidatus Mycoplasma girerdii.

In the complex field of interspecies relationships found in the microbial world, two paradigmic protists for symbiotic relationships with bacteria and viruses have emerged: the obligate human parasite T. vaginalis and the free-living amoeba (FLA) Acanthamoeba spp. While the potential role of Acanthamoeba as an environmental reservoir and a “biological gym” for intracellular bacterial human pathogens has been the subject of intense investigations (Balczun and Scheid 2017), research on the impact of T. vaginalis' symbionts on its biology is still unexplored in many aspects, despite the first description of the biological association between T. vaginalis and Mycoplasma hominis over 20 years ago (Rappelli et al. 1998). This is the first reported case of a biological association between two human microbial pathogens with an obligate parasitic lifestyle, which are able to cause independent diseases in the same anatomical region. A considerable number of studies ever since, have investigated the T. vaginalis and M. hominis association rates in clinical isolates from different geographical locations, highlighting a remarkable variability (5% to 89%) (R Fichorova et al. 2017).

5.1.1 Mycoplasma hominis

M. hominis is a human sexually transmitted bacterium belonging to the group of Mollicutes, a Class of atypical bacteria with a small genome size, lacking a cell wall and several metabolic pathways. These unique characteristics are reflected by the strong dependency of mycoplasmas on host cell environment: indeed, these microorganisms depend metabolically on a host cell to survive and reproduce, thus they have evolved an obligate parasitic lifestyle. The complete genome sequences of several Mollicutes revealed the genetic basis of such dependence. Mycoplasmas lack genes necessary for the biosynthesis of amino acids and nucleic acid precursors, and are thus unable to synthesize essential macromolecules. M. hominis utilizes arginine as an energy source through the Arginine dihydrolase (ADH) pathway, which leads to ATP production via substrate-level phosphorylation (Razin et al. 1998).

Genital infections by M. hominis are associated with bacterial vaginosis and with a number of symptoms, but the pathogenicity mechanisms of this bacterium are still unclear. M. hominis can be isolated from the genital tract of both symptomatic and asymptomatic individuals and it is considered a commensal microorganism of the genital tract that can act as a pathogen in certain circumstances. Nevertheless, there is evidence that M. hominis may play an important etiologic role not only in genital tract diseases of both men and women, but also in extragenital infections. The mechanisms by which M. hominis exerts its pathogenic effects still need to be elucidated in detail: only a limited number of proteins have been associated with virulence. P80, P50/vaa and OppA were shown to play a role in adhesion to host cells (Henrich et al. 1996; Hopfe et al. 2004; Kitzerow et al. 1999). Interactions of M. hominis with host immunity has also received attention: mycoplasmal lipoproteins are able to induce IL-23 production by dendritic cells (Goret et al. 2016), and a specific predicted surface lipoprotein (MHO_0730) was suggested to have a role in evasion from innate immune response, through the disruption of Neutrophil Extracellular Traps (NETs), as discussed in Sect. 4.3.3 (Cacciotto et al. 2019).

A first characterization of T. vaginalis: M. hominis symbiotic association showed how the bacterium is able to reside and multiply within trichomonad cells (Fig. 5a) (Dessi et al. 2005), thus gaining a niche which may allow protection from immune responses and antibiotic treatments. T. vaginalis could also transmit the mycoplasmal infection to a human cell line, and to naturally M. hominis-free trichomonad isolates (Rappelli et al. 2001). Additionally, Thu and colleagues investigated a potential mechanism by which T. vaginalis may function as a protective niche to transport virulent M. hominis to human cell cultures (Thi Trung Thu et al. 2018). They treated Mycoplasma-infected trichomonad cells with metronidazole, which led to T. vaginalis death and subsequent release of viable bacteria from T. vaginalis, which in-turn infected cultured human cells. A similar process might also occur in vivo. Together, these findings suggest a role for T. vaginalis to serve as a “Trojan horse” for M. hominis.

Fig. 5
figure 5

T. vaginalis association with Mycoplasma hominis. (a) Double-immunofluorescence micrographs depicting the cellular localization of M. hominis infecting T. vaginalis. Panels A to C represent the same area of a protozoan monolayer. (A) FITC fluorescence showing extracellular and intracellular mycoplasmas.(B) Rhodamine fluorescence showing mycoplasmas that are extracellularly located. (C) Superimposed images of panels A and B indicating the localization of extracellular (red) and intracellular (green) mycoplasmas. (b) SEM image of T. vaginalis harboring M. hominis in contact with human vaginal epithelial cells (hVECs). Notice the damage to the vaginal epithelial cells comparing uninfected hVECs control (A) and after 1 hr of interaction with T. vaginalis and its associated endosymbiont M. hominis (B, C). Scale bar=20μm. Figure (a) was adapted from Copyright © American Society for Microbiology, [Dessi et al. Infect Immun, 73(2), 2005, 1180-6, DOI 10.1128/IAI.73.2.1180-1186.2005]. Figure (b) was adapted by permission from the publisher: Springer Nature, Eur J Clin Microbiol Infect Dis, https://springerlink.bibliotecabuap.elogim.com/article/10.1007%2Fs10096-007-0422-1, Trichomonas vaginalis harboring Mycoplasma hominis increases cytopathogenicity in vitro, Dessi et al. Copyright 2008

As described previously, the current model of trichomoniasis pathogenesis is driven both by the direct cytopathic effect exerted by T. vaginalis through cytoadhesion and expression of cytotoxic proteins, and by the host inflammatory response which can contribute to mucosal tissue damage (Mercer and Johnson 2018). The association of T. vaginalis with M. hominis affects both aspects of T. vaginalis immunopathogenesis. Vancini and collaborators showed that M. hominis-infected T. vaginalis displayed an increased ability to induce epithelial cellular damage in vitro, compared to uninfected trichomonads (Vancini et al. 2008). Similarly, researchers found that M. hominis upregulates T. vaginalis haemolytic activity in vitro (Margarita et al. 2016). Lysis of erythrocytes by T. vaginalis occurs in vivo, and the heamolytic activity and the consequent acquisition of nutrients, especially iron, have been suggested to be responsible  for the exacerbation of symptoms observed during and following menstruation (Lehker et al. 1990). However, no increased lysis of leukocytes was observed when T. vaginalis harbored M. hominis (Mercer et al. 2016), indicating that the ability of M. hominis to influence T. vaginalis’ host-cell killing behavior may vary based on the host cell type or cell death pathway triggered (discussed in Sect. 4.1.3). Regarding the inflammatory potential of M. hominis however, the microbial symbiosis induced an increase in inflammatory cytokines produced by host immune cells, as discussed in Sect. 4.3.1 (Fiori et al. 2013; Mercer et al. 2016). Thus M. hominis could play a role in the increased risk of HIV acquisition and cervical cancer progression associated with trichomoniasis, since both are likely linked to inflammation (McClelland et al. 2007; Sutcliffe et al. 2012; Yap et al. 1995). Furthermore, T. vaginalis and M. hominis are both associated with an increased risk of adverse pregnancy outcomes, which are also associated with increased inflammation (Gomez-Lopez et al. 2010; Pararas et al. 2006; Petrin et al. 1998). Therefore, we hypothesize that the upregulation of the host inflammatory response resulting from protist-mycoplasma symbiosis might have an impact on clinically important morbidity associated with trichomoniasis.

Besides the contribution of the T. vaginalis-M. hominis symbiotic relationship to the infection biology of both microorganisms, the possible effect of M. hominis on the metabolism and physiology of the protozoan has also been investigated. T. vaginalis and M. hominis are both equipped with the ADH metabolic pathway, which uses arginine as an energy source for ATP production. While this pathway contributes up to 10% of the ATP in T. vaginalis (Yarlett et al. 1996), the energy metabolism of M. hominis entirely relies on ADH (Pereyre et al. 2009). Morada and colleagues investigated the arginine metabolism in M. hominis-infected T. vaginalis, and described how the production of ornithine and putrescine, two key intermediates in this biochemical pathway, are increased 16 and 3 fold respectively, compared to Mycoplasma-free trichomonad cells (Morada et al. 2010). This phenomenon is accompanied by an increase in arginine consumption and has two important consequences. Firstly, M. hominis increases growth rates of T. vaginalis, in an arginine-metabolism dependant manner (Margarita et al. 2016). Secondly, the upregulation of arginine scavenging may contribute to the survival strategies of these pathogens in the host. Arginine is the substrate for the production of nitric oxide, a known compound with microbicidal activity produced by host macrophages (Dillon et al. 2002). Thus reduced nitric oxide levels may negatively affect host defense, and the metabolic synergy between T. vaginalis and M. hominis may therefore confer a fitness advantge during infection, to both microbes.

5.1.2 Candidatus Mycoplasma girerdii

A 16S rRNA-based metagenomic analysis of the vaginal microbiota in T. vaginalis-infected and uninfected patients (Martin et al. 2013), led the authors to serendipitously identify a potential new symbiont: a previously unknown Mycoplasma species strictly associated with T. vaginalis infection. Initially dubbed Mnola, this unknown Mycoplasma was further characterized through the reconstruction of its genome (Fettweis et al. 2014). The authors proposed the name “Candidatus Mycoplasma girerdii” for this microorganism, still to be officially recognized as a new species. Ca. M. girerdii is still a non-culturable bacterium, and only a limited number of studies have taken into account its presence in patients (Costello et al. 2017; Ioannidis et al. 2017; Masha et al. 2018). Despite the detection of putative virulence genes in the genome sequence (Fettweis et al. 2014), the hypothetical influence of Ca. M. girerdii on T. vaginalis pathogenesis is still unknown.

5.1.3 Trichomonasvirus

In addition to the symbiotic relationships with M. hominis and Ca. M. girerdii, T. vaginalis can be infected by a dsRNA virus, Trichomonasvirus. T. vaginalis is not unique, among parasitic protozoa, in harboring viruses (Wang and Wang 1991). The presence of non-fragmented dsRNA of viral origin in T. vaginalis isolates was first established by Alice and C.C. Wang back in 1986 (Wang and Wang 1986). Since then, many studies focused on the extent of this viral infection in T. vaginalis clinical isolates, and on the molecular characterization of Trichomonasvirus (Goodman et al. 2011). T. vaginalis isolates are commonly infected by four different viral species (Trichomonasvirus I–IV), which can concurrently be found in the same T. vaginalis isolate. The rates of detection of Trichomonasvirus show a high degree of variability globally, ranging from 17.4% in a study conducted in Iran (Heidary et al. 2013) to 81.9% detected in South Africa (Weber et al. 2003). The reasons for the extent of this variability could be due to the different geographical settings, to differences in the detection techniques used, or the small population sizes subjected to analysis. We predict that until now, the potential impact of the presence of Trichomonasvirus in T. vaginalis isolates on human pathology has been underestimated. Indeed, work in this specific field is sparse and open to interpretation. As described in depth previously in this chapter, adhesion to the mucosal surface is an initial critical step in T. vaginalis virulence. Fraga and colleagues, observed that Trichomonasvirus-harboring T. vaginalis strains isolated in Cuba showed increased adherence levels in vitro and pathogenicity in patients, evaluated as severity of symptoms (Fraga et al. 2012). This observation could be explained by the presence of the virus itself or, alternatively, by a predisposition to viral infection by more inherently pathogenic strains. The most significant impact of Trichomonasvirus on T. vaginalis pathogenic processes is likely on the host immune response, as discussed in Sect. 4.3.1.

The discovery of microbial symbionts co-existing with T. vaginalis opens many exciting questions focused on a more nuanced understanding of infection biology in the context of the metabolic and immunological contributions that the symbionts make to the parasitic unit. In many ways, infection with strains harbouring any or all of these symbionts, could be considered a polymicrobial infection, or a potential evolutionary unit—“holobiont” (O’Malley 2017). However, many questions that these complexities raise remain to be answered. These include: (1) What are the specific mechanisms by which M. hominis/Ca. M. girerdii/Trichomonasvirus influence immunopathogenesis? (2) Do the symbionts have an impact on clinical therapy for trichomoniasis? (3) What are the evolutionary advantages and implications for all the actors involved in this polymicrobial consortium which could at least partially explain the strict relationships observed? (4) Are these symbiotic relationships still in infancy, from an evolutionary point of view?

5.2 T. vaginalis and the Host Microbiome

In addition to carrying its own symbionts, T. vaginalis colonizes two distictive sites of the human body that are naturally populated by commensal microorganisms (i.e. the host microbiome). Therefore, it is highly expected that the interaction of T. vaginalis with these indigenous microorganisms might be influential to the outcomes of infection and disease. The urinary and reproductive tracts, sites of T. vaginalis infection, are very distinctive anatomical and physiological environments. Their mucosal surfaces are colonized by niche-specialized communities of microorganisms that display specificity to a location within these tracts and to the sex. Even deep organs of the urogenital tract in men and women, once thought to be sterile, have now been found to harbor a specific microbiota (Bao et al. 2017; Javurek et al. 2016; Moreno and Simon 2019). The microbial density in both urinary and reproductive tracts tends to decrease when ascending from the lower to the higher region of the tracts; possibly helping preserve essential functions like reproduction (Moreno and Simon 2019). The lower urogenital tract (i.e. habitat of T. vaginalis) in particular, is densely colonized by commensal microorganisms.

No other site of the urogenital tract is as densely colonized by microorganisms as the ectocervix and vagina, with up to 107–108 bacteria per ml recovered from cervicovaginal secretions of women at reproductive age (Redondo-Lopez et al. 1990). Like no other female mammal, humans have developed a complex relationship with their cervicovaginal microbiome (CVM) that has a direct impact on their reproductive health (Anahtar et al. 2018; Younes et al. 2018). The higher incidence of trichomoniasis in women than men can be attributed to a combination of factors in men, including insensitive diagnosis and presumptive treatment, milder symptoms, and shorter duration of infection (Kissinger 2015). For these reasons, research on the interaction of T. vaginalis with the microbiota has been particularly focused on the CVM.

5.2.1 Mutualism Between the Human Cervicovaginal Microbiome and Host

The histophysiology of the cervicovaginal mucosa and the CVM change with age (Younes et al. 2018). The mucosal lining, consisting of a stratified squamous non-keratinized epithelium, reaches its maximum thickness at reproductive age. While the microbial community of the pre-menarchal (before first menstruation) and menopausal vagina may resemble the one of the skin or gut, the composition and dynamics of the CVM at reproductive age is unique. Specifically, reproductive-age CVM is rich in lactobacilli in most women, with almost exclusive predominance of a single vaginal species of Lactobacillus. In an ethnically-restricted cohort of asymptomatic women of reproductive age, ~75% were found to harbour a lactobacilli-rich CVM, which was classified into community state types (CSTs) depending on the predominant species of Lactobacillus: CST-I for L. crispatus, CST-II for L. gasseri, CST-III for L. iners and CST-V for L. jensenii, while CST-IV which was characterized by low abundance or absence of lactobacilli but diversified bacterial species (Ravel et al. 2011). However, it is important to note that on average, the microbiome literature presents a biased representation of ethnic and racial groups with emphasis on Caucasians, African-Americans and Hispanics.

A compelling mutualistic relationship between the human host and vaginal lactobacilli develops during reproductive age. Estrogen induces proliferation of the cervicovaginal cells and glycogen production, while progesterone promotes cell lysis and glycogen release. The consumption of glycogen by lactobacilli leads to the natural acidification of the vagina, due to release of high amounts of lactic acid. In addition to acidification and secretion of other anti-microbial substances, lactobacilli support the host by enhancing epithelial barrier function and stimulating innate immune responses (Anahtar et al. 2018; Pekmezovic et al. 2019; Younes et al. 2018). Strikingly, this is a human-specific phenomenon, not seen even among our closest primate relatives, and is driven by the natural fluctuation of hormones across the menstrual cycle (Younes et al. 2018). The vaginal microbiomes of nonhuman primates display species specificity following phylogeny; in spite of this, not even our closest primate relatives exhibit a vaginal microbiome with Lactobacillus dominance (Yildirim et al. 2014a). Longer gestation periods and a smaller pelvic oulet in humans than other primates may have acted as evolutionary drivers of this recent friendship between humans and their Lactobacillus-dominant vaginal microbiota (Yildirim et al. 2014a). A microbially acidified vagina provides better pregnancy outcomes, in addition to protecting both mother and fetus against infections.

5.2.2 Correlates of Typical Cervicovaginal Microbiome with Eubiosis and Dysbiosis

Despite the dominance of lactobacilli, a considerable fraction of asymptomatic, childbearing age women (~25%) harbour a species-diversified CVM (CST-IV); which virtually excludes lactobacilli and instead is composed of mostly anaerobic bacteria with the predominance of species commonly found in bacterial vaginosis (BV). These bacterial species include Gardnerella vaginalis, Prevotella, Megasphera, Snethia and Atopobium vaginae (Ravel et al. 2011). Lactobacilli are keystone commensals contributing to eubiosis of the vagina (i.e. the homeostasis of the vaginal biome). Lactobacilli are certainly known to confer protection and optimal immune-physiological function of the vagina; however they do not always necessarily define vaginal health (Amabebe and Anumba 2018), due to several considerations. Firstly, the lactobacilli-low CST-IV and BV are more common among Black and Hispanic women than Caucasians for example. Secondly, the CVM of an individual is dynamic; its composition varies temporarily with the menstrual cycle and sexual activity, among many other factors. Finally, not all lactobacilli provide host protection to the same level. The protective role of Lactobacillus iners, for instance, is actually questionable. A L. iners-dominated vagina (CST-III) is less stable, more prone to transition and more associated with dysbiosis than CST-I and CST–V (L. crispatus and L. jensenii, respectively) (Amabebe and Anumba 2018). A L. gasseri-dominated vagina (CST-II) is the most stable of all, rarely transitioning to dysbiosis (Boris and Barbes 2000; Gajer et al. 2012). While most vaginal lactobacilli absolutely contribute to eubiosis of the vagina, the hallmarks of a ‘healthy vagina’ cannot be established without a real understanding of gene-environment interactions (Amabebe and Anumba 2018).

5.2.3 Correlates of Cervicovaginal Microbiome with Trichomoniasis

Clinically, trichomoniasis is often associated with vaginal dysbiosis, characterized by a reduction of vaginal lactobacilli, which is thought to contribute to some of the symptoms of the infection. Additionally, patients frequently display an intermediate Nugent score, indicating diversified bacterial morphotypes (Petrin et al. 1998). However, only recently did cross-sectional surveys using metagenomics show a clear association of trichomoniasis with CST-IV. One study has identified that 8 out of 11 cases of asymptomatic T. vaginalis infections, among 396 women of childbearing age, were accompanied by the CST-IV microbiome (Brotman et al. 2012). In a separate study, samples stratified by Nugent score were examined (Martin et al. 2013). As expected, women with intermediate Nugent score were more likely to have T. vaginalis infection. Importantly, T. vaginalis infected women with an intermediate Nugent score carried a CVM that closely mirrors that of BV (Martin et al. 2013). This finding reinforces the early epidemiological finding that women with BV have higher chances to acquire trichomoniasis (Rathod et al. 2011). Together, these two studies (Brotman et al. 2012; Martin et al. 2013) have brought out compelling evidence that either a specific CVM facilitates T. vaginalis infection or that the parasite itself could possibly shape the microbiota towards CST-IV, perhaps favoring its survival or transmissibility. Although these studies agree upon a strong association between trichomoniasis and a species-diversified CVM, they should be interpreted with a note of caution. For example, women of Black and Hispanic ethnicities were largely represented in both studies. In addition to limited sample sizes, neither study took into consideration the temporal stability of the T. vaginalis-association with the microbiome. Ideally, longitudinal studies from variable ethnicities will be necessary to confirm the association of T. vaginalis with the CST-IV type of CMV, and to determine the drivers of cause-and-effect (i.e. whether T. vaginalis drives CST-IV, or CST-IV predisposes to T. vaginalis infection) (Fig. 6a).

Fig. 6
figure 6

T. vaginalis interaction with the cervicovaginal microbiome (CVM). (a) Interplay of T. vaginalis and the CVM. (Top panel) T. vaginalis infection may lead to the loss of predominant Lactobacillus species (turquoise bacteria) in the cervicovaginal microbiome (CVM) and promote increased vaginal colonization by a mixed and largely anaerobic bacterial population, similar to that associated with bacterial vaginosis (BV-CVM, mixed color bacteria). (Bottom panel) Alternatively, a diverse and largely anaerobic bacterial population that is part of the BV-CVM may help facilitate T. vaginalis colonization. (b) Competitive and synergistic interactions of T. vaginalis with microbial communities of the CVM. Protective lactobacilli and T. vaginalis are competitors in the vagina. The parasite inhibits the growth of lactobacilli and the bacteria inhibit T. vaginalis adhesion to the epithelium. On the other hand, a dysbiotic BV-CVM (anaerobes) promotes T. vaginalis adhesion to pathogenic biofilm and host cells. BV bacteria are pathobionts of T. vaginalis; enhancing host cell and epithelial damage. Additionally, T. vaginalis and BV-CVM impact the host immune response synergistically. Killing of T. vaginalis by host immune defense (e.g. trogocytosis) or by the action of the 5-nitromidazole drug is likely to release pathogenic M. hominis; which can invade vaginal cells. Although the drug might eliminate the parasite, it does not help restore microbiome disturbances. Figure A created with BioRender.com. Image B was reprinted from Trends in Parasitology, de Aquino et al. Trichomonas vaginalis, Parasite of the Month, Copyright (2020) with permission from Elsevier

Nonetheless, the correlation of trichomoniasis with a specific type of microbiota, which excludes lactobacilli and resembles BV, is at least intriguing. This possible unique inter-domain relationship (protozoa and bacteria) in the vagina evokes basic principles of microbial ecology (Coyte et al. 2015). It suggests that T. vaginalis and indigenous vaginal bacteria are competitors in this specific ecological niche. If a cause-and-effect can be established between microbiota and T. vaginalis, then niche competition or tradeoffs (Bauer et al. 2018; Limberger and Wickham 2011; Smith et al. 2018; Yawata et al. 2014) between T. vaginalis and vaginal commensals might be critical for infection and potentially decisive to its outcomes. Lactobacilli-mediated host defense may impact the evolution of T. vaginalis virulence (Ford et al. 2016). Competitive exclusion, where protozoa wins over lactobacilli, could potentially drive vaginal dysbiosis. Meanwhile, competition-colonization tradeoffs could explain a preferential coexistence of T. vaginalis with the species-diversified CST-IV. Applying this knowledge to the development of novel treatments will be possible, if we understand the underlying principles and mechanisms that support these microbial interactions in the vagina. Just recently, vaginal microbiome transplantation was shown to be relatively successful for women suffering from intractable and recurrent BV (Lev-Sagie et al. 2019).

5.2.4 The Interplay of Vaginal Bacteria, T. vaginalis Virulence, and Host Response

The interactions of T. vaginalis with the microbiota are certainly influential to disease and have important consequences to women’s sexual health. Because the vaginal microbiome composition appears to be largely human specific (as detailed in Sect. 5.2.1) (Vrbanac et al. 2018; Yildirim et al. 2014b), replicating a T. vaginalis infection together with its associated microbiome will be highly challenging, if possible at all, with any animal model. However, investigating how T. vaginalis modulates endogenous CVMs in different animals may help reveal common effects exerted by T. vaginalis on specific bacterial species.

Currently, to investigate the influence of the CVM on T. vaginalis virulence and host responses, researchers have applied tissue culture models of microbial colonization and infection. These models use immortalized ectocervical or vaginal cells that maintain the expression of immune and physiological markers from their tissues of origin (Fichorova et al. 1997). Ex-vivo models of the human cervicovaginal mucosa, e.g. the commercially available EpiVaginal cultures (MatTek Corporation), represent a further step towards replicating the physiological complexity of this tissue. These cultures are established from primary epithelial cells conditioned to form highly differentiated, polarized and multi-layered cervicovaginal tissue resembling the structure and physiology of the native tissue. In addition to being supplemented with immune cells, human microbiomes can be transplanted to, or defined microbial communities can be established on these 3D multilayer cultures [reviewed in Herbst-Kralovetz et al. 2016]. Tissue damage and immunotoxicity in response to commensals and pathogens can be evaluated on these organotypic systems; which are also quite valuable to assess biocompatibility of microbicides (Trifonova et al. 2006).

Although no artificial system can recapitulate the entire physiological complexity of the cervicovaginal mucosa, in vitro and ex vivo culture systems have been very useful to establish models for microbial colonization of the vagina (Fichorova et al. 2011) with particular emphasis on BV [reviewed in Herbst-Kralovetz et al. 2016]. Importantly, the use of culture systems is now providing evidence that the correlation between trichomoniasis and CST-IV is not merely coincidental. The interaction of T. vaginalis with the vaginal commensals has profound effects on parasite virulence and host responses (Fichorova et al. 2013; Hinderfeld et al. 2019; Hinderfeld and Simoes-Barbosa 2019; Phukan et al. 2013, 2018). Protective lactobacilli, particularly the most stable vaginal species L. gasseri, are very inhibitory to the adhesion of T. vaginalis to human ectocervical cells (Ect-1) (Phukan et al. 2013), compared to other bacteria. L. gasseri can preclude parasites from binding to, or displace already bound parasites from host cells during short incubation periods (30–60 min), even against very cytoadherent strains of T. vaginalis. This inhibitory phenotype depends on the physical contact between bacteria and parasite, with participation of bacterial surface proteins that promote aggregation. A specific L. gasseri aggregation-promoting factor was identified to participate in this inhibition (Phukan et al. 2018). T. vaginalis, on the other hand, can reduce the colonization of Ect-1 by lactobacilli but not BV-associated bacteria after long incubation periods (>24 h) (Fichorova et al. 2013). Although there is much more to be understood, these findings support the hypothesis that T. vaginalis and protective lactobacilli are natural competitors in the vagina.

Contrary to protective lactobacilli, BV bacteria were found to enhance the adhesion of T. vaginalis to Ect-1 cells very significantly (Hinderfeld and Simoes-Barbosa 2019) (Fig. 6b). T. vaginalis G3 (the genome strain) is a very lowly cytoadherent strain (Brooks et al. 2013; Lustig et al. 2013). However, in the presence of BV bacteria, G3 adheres to Ect-1 as strongly as the highly cytoadherent T. vaginalis strain B7RC2. In addition, BV bacteria were found to delay the binding of T. vaginalis to mucins, the major component of the host-protective mucus barrier. A notorious activity of BV bacteria is mucin deglycosylation, particularly by removal of sialic acid via sialidases (Howe et al. 1999; Lewis et al. 2013; Wiggins et al. 2001). Sialic acid is a major moiety responsible for the rheological properties of mucins (Ridley and Thornton 2018; Wiggins et al. 2001), thus BV bacteria can make mucous less viscous or rigid. T. vaginalis is capable of proteolytic degradation of mucins (Lehker and Sweeney 1999); however T. vaginalis sialidase is enzymatically incapable of removing alpha 2-6 linked sialic acid from mucins (Padilla-Vaca and Anaya-Velazquez 1997). Therefore, BV bacteria and T. vaginalis may have complementary activities, acting cooperatively on the modification of host mucins at the vaginal mucosal interface. Interestingly, prokaryotic and eukaryotic microbes that live in association with animal host mucosa were found to have a common set of proteins carrying a zinc-metallopeptidase-like motif (M60-like/PF13402 domain) with an extracellular signature and, in some cases, glycan-binding domains (Nakjang et al. 2012). T. vaginalis was found to have 25 genes encoding proteins with the PF13402 domain and a bacterial enzyme containing this domain was shown to cause proteolytic degradation of mucins in vitro (Nakjang et al. 2012). Thus, T. vaginalis has a large uncharacterized repertoire of putative mucinases. It will be important to understand how BV and T. vaginalis act together on the modification of the host mucins, as this is the “first line of defense” serving as the outermost physical barrier of the mucosa.

Although experimental support is still lacking, cooperative modification and degradation of mucins between BV bacteria and T. vaginalis could provide advantages to both microorganisms. This cooperative activity could support microbial nutrition from alternative sugars and amino acids released from the degraded mucin. Sialidases from BV bacteria may be important to initiate the process of breaching this protective mucus layer. This is because BV sialidases are likely to soften the gel layer structure, potentially granting T. vaginalis access to bind and proteolytically degrade mucins. Together, these activities would help both bacteria and parasite break through the mucus and reach the underlying epithelium. Mucinase and sialidase activities of BV bacteria were found to be implicated in the pathogenesis of preterm labor (Howe et al. 1999; McGregor et al. 1994). A CST-IV colonized mucus was found to facilitate the mobility of HIV virions, as compared to a L. crispatus (CST-I) colonized mucus (Hoang et al. 2020). Hence, the cooperative activities of BV bacteria and T. vaginalis on mucins might also be a contributor to increased HIV infection susceptibility and preterm labor observed during trichomoniasis (Kissinger 2015).

It is also possible that BV bacteria could further enhance T. vaginalis adherence to the cervicovaginal epithelium in vivo, due to the biofilm forming-properties of BV bacteria. This polymicrobial biofilm, initiated by G. vaginalis, is a major clinical feature of BV that makes treatment more difficult, and is implicated in higher rates of recurrence (Machado and Cerca 2015; Swidsinski et al. 2005). A minimal in vitro biofilm produced by G. vaginalis was shown to be an excellent substrate for T. vaginalis adhesion (Hinderfeld and Simoes-Barbosa 2019) (Fig. 6b). It will be important to understand whether BV-derived biofilms provide a niche for metabolic exchanges between bacteria and T. vaginalis, and additionally contribute to drug tolerance, as has been observed between BV bacteria themselves (Machado and Cerca 2015). Additionally, this interaction within the biofilm may be advantageous for both bacteria and protozoa. As a parallel example, Giardia lamblia (an extracellular protozoan parasite of the human gut mucosa) is capable of modifying ex vivo gut biofilms promoting the release of specific bacteria (Beatty et al. 2017). This feature might explain dysbiosis and post-infectious complications associated with giardiasis (Allain et al. 2017). If T. vaginalis can similarly modify bacterial biofilms, this processs may help disperse and seed specific bacteria to different areas of the ectocervix and vagina, contributing to dysbiosis and potentially driving changes of the microbiome to a CST-IV.

T. vaginalis evolution alongside the CVM has likely also given opportunities for genetic exchange. Interestingly, one family of genes in T. vaginalis, the NlpC/P60 genes, were recently shown to be inherited by horizontal gene transfer (HGT) from bacteria (Pinheiro et al. 2018). NlpC/P60 peptidases were originally identified as D, L-endopeptidases of peptidoglycan (PG), the major polymer of bacterial cell walls. Despite phylogenetic signal being insufficient to identify the bacterial ancestor, researchers predicted two possible events of HGT followed by gene duplications that led to a family of nine NlpC/P60 genes in the T. vaginalis genome (Pinheiro et al. 2018). The characterization of two of these T. vaginalis NlpC/P60 enzymes (TvNlpC/P60) have shown that they are bonafide PG hydrolases (Pinheiro et al. 2018) with a classical papain-like fold and catalytic triad, in addition to two bacterial SH3 domains giving a unique tri-dimensional arrangement (Pinheiro et al. 2018). These enzymes are fully capable of cleaving the model PG of Escherichia coli, between D-isoGlu and m-DAP residues and with preference for mature PG (Pinheiro et al. 2018). The TvNlpC/P60 enzymes were detectable on the surface of the parasite, and also secreted (Pinheiro et al. 2018). Their presence was also identified in the T. vaginalis secretome (Stafkova et al. 2018). Interestingly, TvNlpC/P60 expression was upregulated upon T. vaginalis contact with bacteria, and exogenous expression of the enzymes in T. vaginalis led to a dramatic capability of controlling the E. coli population in mixed cultures (Pinheiro et al. 2018). Therefore, these genes, once ‘stolen’ from bacteria, are apparently used by the parasite against them. Although this was the first time that a specific factor from T. vaginalis was shown to contribute directly towards controlling bacteria populations, at least in vitro, many questions remain. It is unclear if other factors (human- or parasite-derived) may support the potential role of TvNlpC/P60 in controlling bacterial populations. For instance, vaginal secretions are known to have the most abundant levels of lysozyme of all mucosal sites (Bard et al. 2003). It is also unclear how these T. vaginalis enzymes reach the PG of bacterial cells and whether there is bacterial target specificity. Finally, these enzymes may possibly provide other advantages to the parasite such as nutrition and/or immune modulation, as PG fragments are known to be ligands of pattern recognition receptors (Chu and Mazmanian 2013; Wolf and Underhill 2018).

Finally, evidence supports that BV bacteria and T. vaginalis synergistically impact the host immune response to T. vaginalis (RN Fichorova et al. 2013). Some BV bacteria (particularly G. vaginalis and A. vaginae) act in concert with T. vaginalis to amplify pro-inflammatory responses, promoting secretion of chemokines such as IL-8 and RANTES; effects not seen with lactobacilli controls (RN Fichorova et al. 2013). Increases in IL-8 are predicted to attract more neutrophils to the infection site, which although they are players killing T. vaginalis, also highly promote inflammation. Increases in RANTES, which can recruit CCR5+ leukocytes to the infection site, are linked to increased HIV susceptibility, as discussed in Sect. 4.3 (Kissinger 2015). Furthermore, G. vaginalis and A. vaginae were found to significantly enhance T.vaginalis-mediated disruption of paracellular permeability, in a cervicovaginal epithelial model (Hinderfeld et al. 2019). The microbial-driven increase in permeability was not due to cytotoxicity, but was accompanied by dysregulation of the tight junctions of the epithelium. The major component occludin was virtually excluded from the junctional complex, possibly due to high levels of phosphatase activity, as detected in this study (Hinderfeld et al. 2019). Increase of paracellular permeability of the ectocervical epithelium allows bacteria, viruses and toxins to reach the underlying tissue. Dysfunction of the cervicovaginal epithelium induces cervical remodelling, the ripening process of the cervix necessary for delivery of the foetus (Yellon 2017). Once enhanced by BV bacteria and T. vaginalis synergistically, early onset of cervical remodelling in pregnant women with trichomoniasis might explain the association of these infections with preterm birth (Kissinger 2015).

As with many other parasites, T. vaginalis research has traditionally focused on host and parasite in isolation. As detailed above, recent progress has been achieved towards understanding trichomoniasis in the context of the triad host: parasite: microbiota. In the near future, it will be important to decipher the mechanisms and the genetic basis of microbial competitions and trade-offs that either support vaginal eubiosis or drive dysbiosis during T. vaginalis infection. Once mechanisms are  better understood, a new avenue for interventions against trichomoniasis - that consider restoring the normal function of the vaginal biome - might be envisaged.

6 Part 6: Treatment and Vaccination Prospects

T. vaginalis has been historically diagnosed through wet mount (observations of trichomonads microscopically), but several higher sensitivity and more rapid PCR tests are now also available, which can detect T. vaginalis in vaginal/cervical swabs, and urine (Van Gerwen and Muzny 2019). However, many cases of T. vaginalis go undiagnosed as testing for it is largely symptom-driven and due to its prevalence in resource-poor settings. Knowledge of T. vaginalis biology and metabolism has enabled the use of good treatment options that can clear most symptomatic infections. Still, alternative therapies that consider the complexities in the host-pathogen interface discussed above, including those that consider the “microbiomes” of both host and parasite, may help to inform novel therapeutic options that may be useful during pregnancy, or in the face of rising drug resistence. A protective vaccine is also a very attractive prospect for this highly prevalent human infection, which disproportionately affects those with limited access to healthcare.

6.1 5′ Nitroimidazole Drugs

If diagnosed, most infections can be cleared with treatment by the 5-nitroimidazole drugs metronidazole (Mz) and tinidazole (Van Gerwen and Muzny 2019). Metronidazole is the standard first line of treatment, and tinidazole is used less commonly due to its higher cost (Muzny et al. 2020). The 5-nitroimidazoles are prodrugs, that need to be activated inside T. vaginalis after passive entry into the parasite. Specifically, these classes of drugs are activated inside the hydrogenosome via redox enzymes and/or flavin reductases that reduce the prodrugs (Hrdy et al. 2005). Thus, the parasite’s anaerobic metabolism confers susceptibility (Leitsch 2016). Resistance to Mz is currently estimated to be as high as 5% in the US (Leitsch et al. 2012). Alarmingly, in Papau New Guinea, where T. vaginalis is endemic (prevalence of 46% and 33% have been reported from different regions), drug resistance was found to be 17.4% (Upcroft et al. 2009). Resistance to tinidazole, often used against Mz-resistant cases, is also emerging, and present in some Mz-resistant strains (Leitsch et al. 2012). Despite the increasing levels of Mz resistance, the T. vaginalis mechanisms that have evolved to resist Mz are currently under-characterized. Bradic and colleagues recently identified 72 Single-Nucleotide Polymorphisms (SNPs) associated with moderate or high Mz resistance; proposing them as a panel of biomarkers that can be used to advance personalized treatment for trichomoniasis (Bradic et al. 2017). Transcriptomic analysis of Mz-resistant strains further provided a global view of molecular pathways involved in Mz resistance, highlighting a differential expression of genes that could be involved in drug activation, drug efflux, and detoxification (Bradic et al. 2017). However, the identification of specific targets of Mz would be useful in developing alternative drugs to target these essential proteins.

6.2 Alternative Treatments

Secindazole, another 5-nitroimidazole that is used against the protozoan Dientamoeba fragilis, shows 96% efficacy against Mz-resistant T. vaginalis strains tested, indicating that this could be efficacious against hard-to-treat infections (Ghosh et al. 2018). Since the T. vaginalis and mammalian proteasomes are divergent, the 20S proteasome was also recently shown to be a promising new drug target (O’Donoghue et al. 2019). Other proposed treatments include boric acid suppository (Van Gerwen and Muzny 2019), ceragenins, and topical mimics of antimicrobial peptides which can kill Mz-resistant strains in vitro (Polat et al. 2011), as well as natural products such as the Manilkara Rufula plant extract, which killed T. vaginalis but not host cells in vitro (de Brum Vieira et al. 2017a, b). As recently reviewed, there is a phletora of natural compounds with potential to lead to alternative treatments for trichomoniasis (Friedman et al. 2020). Many of theses natural products have been shown to kill T. vaginalis (and other trichomonads) or display some form of anti-trichomonad activity, whether in vitro, in animals, or in humans, and have sometimes been shown to enhace the efficacy of Mz. It is undeniable that new more efficacious treatment strategies against T. vaginalis are needed.

6.3 Vaccines

The high prevalence of T. vaginalis and the increasing reports of strains refractory to treatment make a preventative vaccine an attractive goal. However, several confounding factors currently complicate vaccine development. Firstly, a more thorough understanding of the mechanisms of effective immune clearance, and memory to the parasite is necessary to define the best immunogens and adjuvants. Establishing a mouse model of infection (or any other animal) would aid in better understanding of in vivo pathogenesis, and for testing immunogens and adjuvants. At this stage, the immune response to T. vaginalis remains under-characterized, and a mouse model that is able to retain adequate titers of infection to allow in vivo studies of pathogenesis and immune responses remains challenging, possibly due to differences in murine and human vaginal pH, microbial content, and availability of appropriate adhesin receptors. The most successful murine models of T. vaginalis have used estrogen and dexamethasone pre-treatment in order to promote T. vaginalis colonization (Cobo et al. 2011), which are both immunosuppressive, making these models problematic for studies of immune function. Furthermore, the required inoculum to establish any infection is extremely high (one million parasites), and only 20–70% of mice inoculated in these studies become infected, indicating a failure of most parasites inoculated to survive and colonize the mouse vagina, or rapid clearance of parasites (Cobo et al. 2011; Smith and Garber 2015). Despite these challenges, mice vaccinated with a whole cell killed T. vaginalis vaccine containing alum adjuvant were shown to be infected by vaginal inoculation at lower rates than unvaccinated controls, indicating promise for continued T. vaginalis vaccine development efforts (Smith and Garber 2015).

A vaccine does exist against a related vaginal trichomonad, Tritrichomonas foetus, which infects cattle (Palomares et al. 2017). However, while this vaccine, called TrichGuard can reduce the incidence of spontaneous abortion and infertility on endemic farms, it does not confer robust immunological memory, as it requires re-administration and does not completely prevent infection (Edmondson et al. 2017). TrichGuard uses whole-cell killed trichomonad and oil adjuvant (Palomares et al. 2017), and was shown to induce vaginal and uterine antibodies against the major surface antigen of the parasite in its native conformation (Palomares et al. 2017), indicating that whole- killed or purified surface antigen of T. vaginalis might be good immunogens to test in a T. vaginalis vaccine. Furthermore, in a mouse model utilizing intraperitoneal injection of T. vaginalis and death of mice as a marker of T. vaginalis pathogenesis, immunization with recombinant T. vaginalis alpha- actinin conferred enhanced survival and higher cytokine levels (Xie et al. 2017). It is not clear whether these findings are applicable to vaginal infection or would provide protection against infection acquisition. In addition to prophylactic vaccines, it may also be interesting to think about therapeutic vaccines that could confer enhanced control against adverse consequences during pregnancy, since Mz is not currently efficacious in decreasing pre-term births for T. vaginalis-infected pregnant women (Klebanoff et al. 2001). Overall, more research into alternative therapies and vaccination strategies against this highly prevalent parasite is needed.

7 Part 7: Conclusions and Perspectives

As described previously, Trichomonas vaginalis has unique cellular and molecular features, engaging with myriad host and microbial cells in different ways. In addition to host immunity and parasite evasion, endosymbionts of T. vaginalis and the accompanying indigenous microbiome contribute to pathogenesis of trichomoniasis (Fig. 6b). The discovery of extracellular vesicles that T. vaginalis produces, and the observation of trogocytosis in host-parasite interactions also suggest new exciting modes of cell interaction and communication that blur the boundaries of cells and their membranes. While the publication of the T. vaginalis genome in 2007 and subsequent omics research have aided these investigations considerably, the poor annotation of the published genome and the lack of genome sequences for more strains of the parasite have hindered additional progress. Clinical isolates of the parasite vary considerably in their in vitro behaviors towards host cells, and differing inherent virulence among strains very likely contributes to the varying clinical outcomes observed. We therefore advocate for the development of more datasets on different clinical isolates, including sequencing of genomes and transcriptomes. Additional functional studies comparing different strains would also bolster our understanding of the contribution of strain type to outcomes, and a greater attention to strain identity is important when comparing results from different studies. On this note, functional characterization of virulence factors and other molecular players will be greatly bolstered by improved genetic tools, such as the CRISPR/Cas9 system, recently adapted for use in T. vaginalis (Janssen et al. 2018).

A substantial hurdle for studying T. vaginalis pathogenesis is the lack of a suitable animal model whereby infections can be naturally established and maintained chronically. Such an animal model would allow scientists to test hypotheses or translate their findings in the laboratory to a real infection situation, teasing out key molecular players that contribute to pathogenesis in vivo. Researchers would be able to study the efficacy of innate and adative immune responses as well as immune-evasive behaviors of the parasite in vivo. The contribution of microbial interactions (symbionts and microbiome) to the pathogenesis of trichomoniasis could then be characterized in the context of the host response in  a more realistic scenario. Undoubtedly, these animal models would allow more translatable research that could aid the development of immunotherapy interventions or vaccine design.

However, popular murine models for infection are likely to pose significant limitations for studying certain aspects of trichomoniasis. Firstly, humans and rodents do not share all immune receptors and effector molecules. For example, IL-8, the predominant cytokine released in T. vaginalis-stimulated cells and the major chemokine thought to mobilize neutrophils against T. vaginalis in the human system, does not exist in mice, and polarizing cytokine requirements for Th17 responses, which are suspected to be important during trichomoniasis, differ slightly in mice versus humans (de Jong et al. 2010). Secondly, the physiology of women’s reproductive tract comprises a mutualistic host-microbiome relationship (Yildirim et al. 2014a; Younes et al. 2018) that is unique across the animal kingdom. Although a ‘humanized’ immune system in mice is a viable option (Li and Di Santo 2019; Walsh et al. 2017) and has been utilized with other parasites successfully (Tyagi et al. 2018), humanizing the CVM of mice stably may not be practical or possible. Despite this, encouraging results have been achieved with BV models in mice. Recently, transient cervicovaginal colonization of mice with BV-bacteria was able to recapitulate some typical features of BV (Gilbert et al. 2019). In pregnant mice, this microbial colonization was accompanied by the expression of molecular, immune and cellular markers, and physiological changes typical of cervical remodeling (Sierra et al. 2018). When triggered prematurely, cervical remodelling will lead to preterm birth. Although very speculative, a partial or transient reconstruction of the human CVM in mice may possibly facilitate T. vaginalis infection.

The investigation of trichomoniasis in males is on the horizon. A recent study by Jang et al. has developed a rat model of T. vaginalis infecton using a combination of multiple T. vaginalis infections and T. vaginalis excretory-secretory products (Jang et al. 2019). This study found increased levels of CCL2 and mast cell infiltration in the prostate of T. vaginalis-infected rats. Future studies with this model will aid dissection of the inflammatory steps that promote T. vaginalis-induced prostatitis.

T. vaginalis treatment has been solely dependent on 5′ nitromidazoles ever since the development of metronidazole in 1959 (Cosar and Julou 1959). While T. vaginalis’ anaerobic metabolism makes this drug quite efficient and specific, the rise in metronidazole resistance has signalled the need to develop novel and alternative treatment approaches for this highly prevalent human infection. We envisage that new interventions should be guided by a holistic understanding of the parasite cellular and molecular biology and its interactions with host and microbiome. A proper understanding of this disease requires attention to the biology of T. vaginalis and microbial ecology. It is important to identify key players in parasite metabolism, growth, and virulence; but it is also critical to understand its interaction with other cells including from the host, endosymbionts and commensals. Therefore, further studies on the the interactions with the endosymbionts (M. hominis, M. girerdii, and Trichomonasvirus) and microbiome are needed.

T. vaginalis frequently carries pathogenic Mycoplasma. In addition, trichomoniasis is often accompanied by a dysbiotic microbiome typical of BV. Metronidazole therapy does not eliminate Mycoplasma and the majority of treated women will experience either recurrent BV or an abnormal microbiome (Bradshaw et al. 2006). Thus, novel interventions should aim to eliminate T. vaginalis and Mycoplasma concomitantly in order to restore the healthy functional status of the immune- and microbial biomes. If this intervention strategy can be developed in the future, it will boost protection to re-infections and mitigate co-morbidities that are clinically relevant and commonly associated with trichomoniasis.