Viruses of protozoan parasites and viral therapy: Is the time now right?

The flagellated protozoon Trichomonas vaginalis is responsible for 170 million cases/year of trichomoniasis, the most common non-viral sexually transmitted infection worldwide [39]. Although asymptomatic in males, symptoms of trichomoniasis in women may vary from asymptomatic to severe vaginitis, eventually with pregnancy and postpartum complications [40]. The Trichomonas vaginalis virus, TVV, belongs to the genus Trichomonasvirus and was the first virus from a protozoan parasite to be described and characterised in the 1980s, and the first for which the full-length genome sequence was reported [41, 42]. Four main phylogenetically distinct viral species, 1 to 4, have been described, with TVV1 closer to TVV2 and TVV3 to TVV4 [43]. Diversity exists within each TVV species, with, for example, translation of the CP/RdRp fusion protein of TVV1 involving a -2 ribosomal frameshift unlike TVV2-4 [44, 45]. Co-infection of a single T. vaginalis with different TVV strains has been reported [43]. The TVV infection rate among T. vaginalis strains from different geographic origins ranges from 40 to 100%, with TVV1 being the most commonly detected [46]. TVV seems to be transmitted only vertically, although some studies suggest a correlation between specific genetic polymorphisms and the entry and multiplication of TVV [42, 44, 47]. The presence of TVV influences negatively the growth rate of T. vaginalis in vitro if compared to uninfected protozoan isolates [48] and there is also evidence for lytic effects of TVV on T. vaginalis [49]. Almost 50 proteins are expressed differentially between TVV-infected and uninfected isolates, including metabolic enzymes, heat shock proteins (down-regulated in TVV-positive strains), and ribosomal proteins (up-regulated in TVV-positive strains) [50]. Indeed, infection with TVV increases both cytoplasmic and surface expression of the p270 protein, the major immunogenic protein of T. vaginalis, in a phosphorylation-dependent fashion [51]. Similarly, TVV infection can modulate the quantitative and qualitative expression of the protozoan cysteine proteases [52]. Since cysteine proteases are involved in modulating T. vaginalis cyto-adherence to human host cells and in degradation of basement membrane, human cellular molecules, and secretory IgAs, the viral endosymbiont seems to influence and modulate protozoan virulence [52]. A correlation between TVV symbiosis with T. vaginalis isolates and the severity of clinical symptoms of trichomoniasis in humans is emerging; while different papers report a positive association between TVV infection and the exacerbation of trichomoniasis symptoms, other authors have shown the absence of any correlation [53]. However, T. vaginalis and its virus appear to have a clear role in the subversion of the innate immune response and inflammation in the human host [54, 55]. Although TVV is unable to infect and replicate in human cells [56], its presence can modulate the pro-inflammatory response in the human host, amplifying the innate response, and thus exacerbating clinical symptoms and the severity of disease. The viral dsRNA and TVV particles can be sensed by receptors exposed on vaginal cells, triggering NF-κB activation via endosomal TLR3/TRIF-dependent pathways and leading to expression of Interferon Type 1 genes [54]. This release of viral dsRNA may be favoured by the presence of wide channels in the virion [45]. Although the percentage of clinical T. vaginalis isolates resistant to 5-nitroimidazole treatment is increasing [57], and although a correlation with the presence of TVV has been postulated, this is still debated and poorly understood [53, 58, 59]. Paradoxically, in the case of infection caused by T. vaginalis carrying TVV, failure of anti-protozoan therapy with metronidazole in order to prevent preterm delivery in pregnant women results in an exacerbated inflammatory response explained by the increased release of virions and dsRNA as a result of parasite killing [45, 60]. The co-existence of dsRNA virus may also disturb the equilibrium of the mucosal microbiome, contributing to its modification in the vagina; infection with TVV-positive T. vaginalis isolates indeed promotes vaginal colonization by pathogenic bacteria associated with bacterial vaginosis while decreasing the adherence to the vaginal epithelium of the major vaginal microflora that dominate during eubiosis [42, 54].

The flagellate protozoan parasite G. duodenalis (syn. G. lamblia, G. intestinalis) infects the upper intestine of humans and other mammals causing giardiasis, a zoonotic diarrhoeal disease. The parasite has a global distribution with 250–300 million symptomatic human infections reported annually, and its impact is more pronounced in the developing world and under poor socioeconomic conditions [61]. Although self-limiting, Giardia infections can become chronic and predispose individuals to other chronic gastrointestinal disorders such as irritable bowel syndrome (IBS) [62]. Soon after the discovery of TVV, the Giardia lamblia virus (GLV) was described [63] which is included in the genus Giardiavirus together with the virus isolated from the fungus Gigaspora margarita [37]. The complete viral genome from three isolates is available [64, 65], providing evidence of minimal sequence variability. GLV shows some unique characteristics among the Totiviridae: (1) CP translation is driven by an unusual Internal Ribosomal Entry Site (IRES), spanning the 5′-UTR (untranslated region) and the initial portion of the CP coding region [66]; (2) the GLV particle is the largest, most robust and thermo-stable among Totiviridae allowing extracellular horizontal transmission of the virus [67]. Susceptibility of G. duodenalis to GLV infection has been demonstrated both by parasite transfection with GLV ssRNA and incubation with purified GLV virions, with infected parasites being maintained stably over time [68, 69]. GLV virus entry is thought to occur via receptor-mediated endocytosis since it can be prevented by specific blocking agents [70] and not all Giardia isolates are susceptible to GLV infection [69]. GLV was shown to be unable to infect other species of protozoan parasites [59] or to induce cytopathic effects in kidney or intestinal mammalian cell lines [71, 72]. GLV has been detected, by PCR or gel electrophoresis, in more than 30% of G. duodenalis isolates belonging to different genetic groups (assemblages), both host-specific and zoonotic [73‐76]. Different cytopathic effects, including growth arrest and parasite lysis, have been observed on naïve G. duodenalis isolates when first infected with different GLVs purified from naturally infected G. duodenalis isolates [63, 73]. However, the reported effects could be explained by differences in either (1) the amount of virus administered, (2) the properties, such as replication efficiency, of each virus isolate or (3) variation in parasite susceptibility to viral infection [73]. No robust evidence associates GLV infection with resistance to metronidazole, the first-line antigiardial drug [77].

Different species of the trypanosomatid protozoan Leishmania are the causative agents of cutaneous (CL), visceral (VL) or mucocutaneous (MCL) and disseminated (DCL) leishmaniasis, afflicting > 12 million people worldwide, with 1.2 million new cases/year. The parasites are transmitted by phlebotomine sand flies [78, 79]. Within Leishmania spp., the Leishmania RNA viruses (LRVs) are isolated mainly from the Leishmania Viannia (V.) subgenus from South America, and designated LRV1 [80, 81], whereas from the old-world Leishmania subgenus (i.e. L. aethiopica, L. major and L. tropica) were named LRV2 [82, 83]. LRV1 and LRV2 genomes differ slightly [84] with limited sequence homology of ca. 40% at the protein level. LRV transmission from one parasite to another must occur vertically during cell proliferation, as there is no evidence of horizontal transmission. Nevertheless, a recent study suggested that LRV particles are also present in parasite exosomes which might implicate horizontal transmission [85]. Although a systematic review with meta-analysis reported a possible estimated prevalence of 26%, this value is difficult to be estimated since Leishmania parasites have the tendency to lose their LRVs endosymbiont in culture [86]. Viral loss possibly relates to a burden for parasite replication that in natural conditions is counterbalanced by the selective advantage provided by LRV within the sand fly vector, in the mammalian host or in both. Supporting this hypothesis LRVs are found mainly in Leishmania spp. equipped with RNA interference machinery (the system allowing the cell to actively recognize and degrade non-self dsRNA, likely to be mainly of viral origin) suggesting the importance of the parasite’s anti-viral defense in finely balancing viral replication rate [87]. Differently from TVV, the presence of LRV does not affect parasite growth and there is no evidence of modulation of parasite gene expression [88]. Despite LRV has been isolated from active and healing lesions and scars [89‐91], thus far no experimental system has allowed long-term maintenance of LRV in Leishmania following experimental transfer (N. Fasel personal communication). The presence of LRV is reported to significantly promote disease relapses in humans infected with L. braziliensis, L. guyanensis or L. naiffi and receiving antimony or pentamidine treatment [92]. MCL, characterized by the dissemination of the infection to secondary sites with a high inflammatory component, has been associated with the presence of LRV in the cytoplasm of the parasites [93]. Another striking effect of LRV is its impact on treatment failure, while the viral dsRNA is also suspected to participate in difficulties in treating HIV/L. braziliensis/LRV co-infected patients [94]. How LRV impacts drug resistance and relapse is not yet known but thanks to the development of different animal models a number of mechanisms underlying disease progression have been identified. Similar to TVV, the pathology-exacerbating role of LRV relies on the subversion of the host innate immune response, as was shown for L. guyanensis or L. aethiopica infection, where the virus exacerbates the disease by inducing hyper-inflammation and increasing the parasite burden as well as lesion size [95, 96]. These phenotypes depend on Toll-like receptor (TLR) 3-mediated recognition of viral dsRNA, which leads to the production of Type 1 Interferon (IFN-I), and pro-inflammatory cytokines and chemokines [97]. Furthermore, LRV promotes the survival of the mammalian cells infected by Leishmania by phosphorylating the AKT1 pro-survival kinase, and favors parasite dissemination via the induction of IL-17 production [93, 98].

Noteworthy, a tri-segmented linear negative-stranded RNA virus (termed, LmarLBV1) with characteristics of leishbunyaviruses (an unassigned bunyaviridae-like group of viruses) was recently discovered in Leishmania martiniquensis, a protozoan transmitted by biting midges and responsible for severe visceral disease in humans [99]. The presence of LmarLBV1 slightly increases the in vitro infectivity of L. martiniquensis on primary murine macrophages [99]. Compared to Totiviridae, Bunyaviridae have enveloped and spherical virions of around 100 nm suggesting that they could be shed by L. martiniquensis cells, allowing the virus to interact with the immune system of the human host [99].

Viruses have been found in other protozoan parasites responsible for serious illness in humans, including the apicomplexan Cryptosporidium spp. and Plasmodium vivax. However, information on the effect of viral infection on these parasites is fragmentary.

Different species of Cryptosporidium (i.e. C. parvum, C. hominis, C felis and C. meleagridis) are pathogenic for humans and other vertebrates, and are responsible for cryptosporidiosis, a severe diarrhoeal disease causing death in young children especially in developing countries [100]. A bi-segmented dsRNA, termed Cryspovirus, of the family Partitiviridae and commonly associated with plants and fungi, was originally detected in C. parvum isolates, and later in other Cryptosporidium spp.[101, 102]. The dsRNA segments 1 (1.8 Kb) and 2 (1.4 Kb), likely uncapped and not polyadenylated, encode the RdRp and CP, respectively. The isometric virion is composed of 120 subunits (T = 1) of the 37 kDa CP, the smallest capsid protein known among Partitiviridae with each viral genome segment encapsidated separately [102]. As with other partitiviruses, CspV1 is thought to be transmitted intracellularly, being present in the environmentally resistant oocyst stage, although recent observation suggests that CspV1 is also released into the medium early in the parasite infection of the human host cell [103]. Based on dsRNA2 sequence comparison, a greater divergence exists between CSpV1 and viruses from C. hominis, C. felis and C. meleagridis, suggesting that the virus might have a certain degree of host-specificity and, therefore, the existence of more than one species in the genus Cryspovirus [104]. Infection by CSpV1 is commonly detected in field and clinical isolates of C. parvum [105]. A correlation between CSpV1 dsRNA2 levels and C. parvum fecundity has been reported in culture models, where higher level of the viral symbiont are associated with greater parasite multiplication [106]. However, no information on infection in an animal model is available.

Six Plasmodium spp. (P. falciparum, P. vivax, P. malariae, P. ovale curtisi, P. ovale wallikeri, and P. knowlesi) infect humans, transmitted by the bite of an infected female Anopheles mosquito, and are responsible for malaria, the most important PPD affecting humans, with an incidence of more than 200 million case/year and 400 thousand deaths/year [107]. A bi-segmented narna-like ssRNA (+) virus has been very recently isolated only from P. vivax and named Matryoshka (after the Russian dolls) RNA virus 1 (MaRNAV-1) [33], showing high homology to the monopartite, linear, positive sense (ss)RNA narnaviruses found in fungi, plants and other protists. The trypanosomatid Leptomonas seymouri, transmitted by the sandfly vector, as is Leishmania donovani, can also harbour a symbiotic narna-like virus (LepsyNLV1) [108]. It is noteworthy that severe cases of visceral leishmaniasis in India have been associated to co-infection with by L. donovani and a narna-like positive L. seymouri [109].

To maximize any parasitopathic activity of a viral symbiont, the selectivity of the virus toward the right parasite target must be high. This seems to not be an issue, at least for GLV, TVV and LRV viruses [56, 63]. Moreover, replication of the virus should be maximal at the most appropriate parasite stage infecting the human (or animal) host. For those protozoan parasites with relatively simple life cycles, such has Giardia, Cryptosporidium and Trichomonas, with parasite multiplication occurring predominantly in one site (e.g. gastrointestinal epithelium for Giardia and Cryptosporidium, vaginal epithelium for Trichomonas) within the final human (or animal) host, the endosymbiont virus is likely to replicate maximally where the parasite is also multiplying (e.g. trophozoite for Giardia and Trichomonas, amastigote for Leishmania, merozoite for Cryptosporidium) (Fig. 1). The picture is likely to be more complex when different replication stages occur during the parasite life cycle, as for Leishmania and Plasmodium, for instance, and in more than one host (i.e. intermediate and definitive hosts) in addition to multiple sites (tissues or organs) in the same host (Fig. 1). Thus, viruses that multiply optimally in one phase of the life cycle might, nevertheless, be used profitably for infection control in another. These speculations remain to be tested. The extent to which the viruses might be used alone or synergistically also remains to be assessed.

Such use of wild-type viral symbionts seems, at the moment, to be applicable only to Giardia (at least for some parasite isolates) where parasite overload with GLV causes cell growth arrest, and likely cell death [63, 73]. Of course, for this kind of strategy some key aspects must be taken into account, such as: (1) an efficient cell factory for high titre virus production; (2) pharmaceutical grade purification of the viral particles for treatment use, and (3) a successful delivery system for the administration of the viral dose to the patient. Problems associated with human (or animal) immune response to the viral administration must of course be evaluated in advance. It is evident that TVV and LRV might be disadvantaged in this specific issue due to their adverse impact on the human immune system [54, 55, 95].

An alternative to using infectious viruses to attack parasites, is their use as VLPs to deliver toxic antiparasitic agents to the parasite host cells. VLPs consist of the assembled viral coat protein subunits that provide empty viral shells which can be loaded with the desired material. The outer surface can also be decorated with targeting moieties, either genetically or chemically. VLPs based on plant viruses have been exploited particularly for this application as they can be produced in large quantities, are stable and can often be readily disassembled and reassembled in vitro to allow the incorporation of drug molecules [136‐138].

To date, much of the research on the use of VLPs has concentrated on delivering drugs for anti-cancer therapy [139‐142]. However, there has been increasing interest in using VLPs to deliver agricultural pesticides to combat plant-parasitic nematodes [143, 144]. The encapsulation of nematicides within VLPs increases their mobility and persistence, although the effect varies with the nature of the VLP [143, 144]. These recent results indicate that drug delivery via VLPs may be an effective way of treating parasites other than those of plants. Of particular concern in anti-parasitic drug therapy in humans and animals is, for example, the low bioavailability (e.g. poor bioavailability of benzimidazoles) or the off-target effect (e.g. nitroimidazoles also targeting commensal anaerobic bacteria in the intestine) of certain drugs [62, 146] Use of VLP delivery could indeed help to overcome such undesired limitations thus improving treatment quality and increasing drug efficacy. Indeed, the viral particles would be depleted of their nucleic acid content and, in the case of TTV and LRV, this would avoid adverse human host response to the viral dsRNA release [54, 95]. Among the obstacles in production of VLP for therapeutic purposes are, in addition to those for preparation of wild-type virus, the need to express at high yield the sole protein(s) constituting the capsid and guaranteeing the correct assembly of the viral particles, likely in an appropriate heterologous biological system, in addition to efficiently loading the antiparasitic agent.

Several technical approaches may be contemplated for the delivery of nucleic acid sequences that will interrupt parasite physiology or kill the parasite. The introduction of antisense sequences to block translation of key genes has been discussed for modulating bacterial physiology but is thought to be less effective and not as easy as originally thought [147], with tRNA processing with external guide sequences as an example [148]. For instance, morpholinos (DNA oligomers modified with methylenemorpholine rings) have been used successfully for transient knock-down of essential genes in Giardia and Cryptosporidium [149, 150]. Other approaches could be used to target genes directly with key function(s) for parasite survival. More recently, the CRISPR-Cas system has been used increasingly for making targeted mutations being successful also in Giardia, Plasmodium, Trichomonas and Leishmania [151‐154]. The system uses guide RNAs to direct a DNA nuclease complex to a target sequence by complementarity with the guide DNA. This has been modified by using dCas9 which is a double nuclease active-site mutant that binds to the DNA target selectively, in this case leading not to cleavage but to inhibition of RNA polymerase initiation and elongation [155]. However, antisense molecules and gene-editing systems do not readily cross cell membranes limiting their use in in vitro systems where techniques, such as electroporation, have been developed to introduce the molecules into cells. VLPs could, indeed, be very useful for their delivery in vivo. An example of such use of the viral symbiont is represented by engineered GLV to express hammerhead ribozymes [156]. The use of catalytic ribozymes [157] to inactivate target genes is a good example but it may be dependent on the relative transcription rates of the genes involved and may also be affected by poor turnover rates under physiological conditions. In Giardia, this approach has been successful to knock down specific genes and also provided evidence for the possibility of combining gene silencing by virus-mediated hammerhead ribozymes with chemotherapy [156]. However, this approach always requires co-infection with a replication-competent helper virus.

Further strategies to engineered VLP of viral symbionts of parasites could, of course, be explored to improve gene delivery to different parasites, as exemplified by the use of engineered bacteriophage T4 expressing antigens facilitating specific cell targeting [120].