It is almost an axiom that if exposure to an infectious agent results in a rapid cure that induces strong long-lasting immunity against a second exposure, then an effective vaccine is a feasible proposition (think measles or smallpox). Conversely, if an infection follows a chronic time course with little evidence for protection against further exposure, then developing a vaccine will be an onerous task. Schistosomes surely fall into this second category. The problem is compounded by, and indeed possibly related to, the fact that helminths do not multiply in the human host. The elicited immune response depends on the frequency with which infective larvae are encountered and the size of the burden acquired. There is anecdotal evidence from travellers that adult S. mansoni can persist for years, even decades, in infected humans ^[1][2][1,2], while a study of S. haematobium in The Gambia estimated the mean worm lifespan as 3.4 years [3]. The existence of a protective immune response in humans was initially suggested by the shape of the age–intensity curve in endemic communities. This curve characteristically shows a rise in intensity during the first two decades of life, followed by a decline in adults to very low levels. Was this an immunity to reinfection, a reduction in exposure in older individuals, or both? [4]. Treatment and reinfection studies revealed that children rapidly became reinfected but adults did not, leading to the idea that protective immunity took years to develop. This hypothesis was counteracted by observations on naïve populations migrating into endemic areas.

Very early in these human studies, a strong correlation between IgE response and a 22.6 kDa protein allergen was established in all three main human schistosome species (see [5]). The protein, which possesses two EF-hand calcium domains, was localised by immunocytochemistry to the tegument, gastrodermis, and nephridial canals [6]. Two similar proteins with Mw of 21.7 and 20.8 kDa were also identified. Although Sm22.6 is not confined to the tegument, these three proteins were later renamed as TAL (Tegument-Allergen-Like) constituents of a 13-member family in S. mansoni [5]. A similar family has recently been characterised in S. haematobium [7]. Inherent problems with Sm22.6, the principal allergen, are its situation within the tegument cytoplasm of adult worms and lack of expression in the penetrating cercaria or early schistosomulum. A consensus appears to be emerging about the potential role of IgE in protective immunity in humans. Stimulation of IgE by released allergens is posited to result from the death of adult worms due to senility; chemotherapy may hasten the process.

2. The Radiation-Attenuated (RA) Vaccine Model

Summaries of research on the model can be found in detailed reviews by James ^[8][20], Coulson ^[9][21], Hewitson et al. ^[10][22], and Bickle ^[11][23]. Most information about the dynamics and mechanism of the model has come from mouse experiments. Cercariae are attenuated by various forms of ionising radiation (gamma rays, X-rays), while exposure to radiation over a short period is crucial. With too little radiation, attenuated parasites reach and develop in the portal system, leading to misinterpretation of results ^[12][24]. With too much, they are confined to the skin, failing to elicit protection ^[13][25]. A crucial feature of the model is that it allows a large dose of larvae, which would normally be fatal, to be applied (typically 500 for a mouse, 1000 for a rat, and 9000 for a primate like the baboon). In studies with gamma rays from a Co⁶⁰ or X-ray source, 200 Gy (20 krad) was optimal. This allowed vaccinating parasites to exit the mouse skin, some via draining lymph nodes, and reach the lungs, where they persisted and recruited CD4+ memory effector cells to arm that organ ^[14][26]. Detailed information is not available for primate hosts, which must be treated as a “black box” in terms of mechanism. Unfortunately, the laboratory mouse has intrinsic flaws as a vaccine test bed ^[15][27]. Briefly, the arguments are as follows. The low level of maturation of penetrating cercariae (~32% for S. mansoni) is a major limitation of the model since 68/100 parasites fail to mature in naïve mice due to natural causes. The pulmonary capillary bed presents a particular obstacle to intravascular migration en route to the portal system. The fragility of pulmonary capillaries and their susceptibility to cytokine-induced vascular leak syndrome result in schistosomula bursting into the alveoli from which they possess only limited capacity to re-enter tissues. A single exposure of attenuated schistosomula arms the lungs by recruiting persistent memory/effector CD4+ Th cells to the pulmonary parenchyma. The interval between the last vaccination and challenge has conventionally been set at 5 weeks, precisely to allow inflammatory processes to subside before the primed mouse encounters a normal challenge. The arrival of a normal challenge parasite in the lungs rapidly triggers a cellular effector focus, which blocks its onward migration, increasing the probability of deflection into an alveolus. It must be emphasised that the trapping of migrating schistosomula in the lungs, associated with inflammatory foci, is not in itself a lethal killing event. An ultrastructural study of such challenge schistosomula found minimal evidence for cytological damage ^[16][28]. The problem with most single antigen vaccine protocols in the laboratory mouse, compared to the attenuated vaccine, is the short interval, sometimes only 10 or 14 days between the last antigen boost and cercarial challenge of test and control animals. This means challenge parasites will reach the lungs when both activated T cells and cytokine levels resulting from the preceding vaccination are maximal in the circulation. It was suggested ^[15][27] that “protection” in this situation was the result of physiological effects on pulmonary blood vessels, increasing the proportion of parasites that enter the alveoli. This hypothesis explains why internal antigens, which are unlikely to interact with the immune response in living schistosomula, plus a variety of heterologous proteins, can reduce the level of maturation in a non-antigen-specific way. One aspect of the RA vaccine model that has received recent attention is the identification of the glycan and protein targets of protective antibodies using microarrays printed on glass slides ^[17][18][33,53]. The glycan array comprises a library of naturally occurring schistosome-derived N-, O- and glycosphingolipid (GSL) glycans isolated from different life stages. Mature schistosome eggs are a potent source of highly antigenic glycans, but vaccination with viable schistosome eggs does not elicit protection in different animal models ^[19][54]. This observation led to the smokescreen hypothesis suggesting that antigenic glycans deflect immune responses away from critical and susceptible protein epitopes ^[20][55]. However, the presence of eggs does not compromise the efficacy of the RA cercaria vaccine in baboons ^[21][56]. Many larval glycan epitopes are shared with eggs, but perhaps anti-larval antibody responses differ in some respects from those induced by eggs. It is notable that high titres of IgG against multifucosylated glycan epitopes were elicited by the RA vaccine in the absence of eggs ^[18][53]. A systems biology approach has been used to identify the immune system components and pathways that correlate with the induction of protective immunity by vaccination for a number of infectious agents (e.g., yellow fever and influenza ^[22][23][24][57,58,59]). Farias et al. ^[25][60] recently used this approach to investigate the impact of the RA vaccine in mice. They found that the upregulation of hemostasis pathways after vaccination may contribute to parasite blockade in the lungs. It was notable that single exposure to attenuated parasites revealed early establishment of a Th1 bias and prominent encoding chemokines and their receptors, indicating an enhanced capacity for inflammation, again potentially augmenting the inhibition of intravascular migration. Increasing the number of vaccinations from one to three did not proportionally elevate protection, and there was a clear shift towards antibody-mediated effectors while elements of the early Th1 bias remained. Notable features after three vaccinations were markers of cytotoxicity (including IL-6 and NK cells), together with growth factors and their receptors (FGFR/VEGF/EGF) and the apoptosis pathway. Indeed, there was evidence for the development of anergy after three vaccinations, reinforced by the limited responses detected in 3× vaccinated animals after challenge. It was inferred that persistence of a Th1 response puts a limit on the expression of antibody-mediated mechanisms. This tendency to anergy may explain the failure of multiple vaccine doses to drive protection towards sterile immunity. These observations hark back to the “happy valley” hypothesis of 1999 ^[26][61], fleshing it out with details of potential underlying mechanisms. The contention was that schistosomes thrive best in the intermediate zone between Th1 and Th2 extremes. The attenuated schistosomulum is a large and complex vaccine “capsule” that interacts with the host skin, draining lymph nodes, immune elements in the circulation such as PBMC, and the pulmonary vascular network. It is probably not simply a matter of identifying antigenic targets. Vaccine strategies that seek to replicate the mechanisms of the RA vaccine using antigenic constructs may need a subtle formulation.

3. Self-Cure in the Laboratory Rat (Rattus norvegicus)

Interest in self-cure by Fisher rats from S. mansoni infection was stimulated by the detailed study conducted by Phillips et al. in 1975 ^[27][62]. Worm numbers in the portal system rose to a peak at 28 days and then declined to very low levels by 6–8 weeks. Cure was not absolute, as a few worms remained even at 6 and 12 months (1% and 0.5% of applied cercariae, respectively). RNASeq analysis of gene expression in such survivors would be informative about a worm’s ability to resist immune attack. This study ^[27][62] was performed not long after the basis of immediate hypersensitivity (allergy) was established by the identification of IgE class immunoglobulins and their ability to sensitise mast cells ^[28][63]. It is important to remember that IgE is essentially a tissue antibody produced locally by plasma cells to bind mast cells and basophils via their Fc epsilon receptors. It is present in the circulation at high levels after infection with intestinal nematodes like Nippostrongylus, and its association with expulsion of the parasite load was established ^[29][64], including degranulation of mucosal mast cells and increased permeability of the gut wall. A high level of total IgE production in helminth-infected rats, due to potentiation of a previous sensitisation to different non-parasite antigens, was also documented ^[30][65]. This potentiation of IgE responses to bystander antigens is a powerful argument against potential allergens as components of worm vaccines. Inevitably, the production of IgE by schistosome-infected rats was also reported ^[31][66]. The profile of total IgE rose to a maximum at 42 days and remained high long after most worms had been eliminated. The habitat of schistosomes is not like that of Nippostrongylus in the gut lumen attached to the mucosal surface. Prior to maturity and mating, worms are located in the hepatic portal distributaries of the liver. If IgE is the specific mediator of worm killing, how does it operate? It was established that pronounced hepatic mastocytosis occurred in the rat, concomitant with parasite elimination ^[32][67]. The majority of recruited hepatic mast cells contained rat mast cell protease II, a useful indicator of mast cell degranulation, which was released into the bloodstream during the period of parasite elimination. This was quite different to the situation in schistosome-infected mice in which intraepithelial mastocytosis in the gut wall occurred later, in response to egg deposition. Furthermore, these intraepithelial cells were typical mucosal mast cells containing mast cell protease I.  Collectively, it was difficult to see how these numerous observations on self-cure in the rat could inform schistosome vaccine development—the involvement of IgE, allergens, and mast cells was not promising. However, after a gap of 20 years, a completely new angle on S. japonicum worm expulsion was provided by experiments with inducible nitric oxide (NO) synthase knockout (iNOS−/−) rats ^[33][70]. Following infection, iNOS−/− rats showed 4-fold greater worm burden and 22-fold greater egg deposition in the liver compared to wild-type rats. Female fecundity was 5-fold higher, and worms recovered at seven weeks were significantly larger than those from wild-type rats. The reproductive organs of both male and female schistosomes from iNOS−/− rats had a more normal appearance and their uteri were filled with eggs. The authors concluded that in wild-type rats, nitric oxide synthesised by iNOS inhibited parasite growth, reproductive organ development, egg production, and viability, achieved by interfering with mitochondrial function. Later experiments suggested that the situation was more complicated ^[34][71]. There was a significant decrease in S. japonicum-elicited Th2/Th1 responses and cytokine and chemokine-producing capability in infected iNOS−/− rats. Is it just an accident of evolution that Rattus norvegicus has high levels of endogenous nitric oxide that impose immune pressure on pre-adult worms, resulting in their death? Was the hepatic mastocytosis described in Fisher rats ^[32][67] absent in iNOS−/− Sprague–Dawley rats? Is there a message here about the mechanism of adult worm elimination in rhesus macaques?

4. Self-Cure in the Rhesus Macaque (Macaca Mulatta)

The features of the model were established in the 1960s in experiments using small numbers of animals, so they were not amenable to statistical analysis (e.g., ^[35][36][72,73]). Briefly, primary exposure to between 100 and 1600 cercariae resulted in patent infection, judged by faecal egg excretion from five weeks. However, starting around 10 weeks, egg output declined to very low levels. Following challenge with 2000 cercariae at 21 weeks, all animals were completely protected compared to a control. A lower primary dose of 25 to 100 cercariae did not adequately prime the animals, but both primary and challenge parasites were eventually eliminated. Smithers and Terry ^[36][73] believed that the persistence of primary worms was necessary for the maintenance of immunity to challenge and coined the term “concomitant immunity,” although it is unclear whether they ever perfused their animals to find out. However, worm transfer experiments inferred the presence of host molecules on the adult worm surface as potentially being an immune evasion mechanism. After an interval of 40 years, scholars revisited the rhesus macaque model using both S. mansoni ^[37][75] and S. japonicum ^[38][76] infections in groups of six animals, with strikingly similar outcomes between the two schistosome species. Large numbers of adult worms were established after exposure to 1000 cercariae, with maturation estimated at 44–65% of penetrants in macaques perfused at 8 weeks in an independent experiment ^[39][77]. Oviposition began at week 5, but from ∼8–10 weeks, egg excretion in faeces tailed off. Scholars were able to estimate worm burden by assaying a circulating antigen (CAA) released from the worm gut as a surrogate. A decline in circulatory CAA levels indicated a cessation of blood feeding. At the 18-week perfusion time, the six macaques exposed to S. mansoni fell into three distinct groups with low, medium, and high worm burdens, which were inversely correlated with antibody titres. Strikingly, many recovered worms were pallid and emaciated; they appeared to be starving to death. In the subsequent experiment with S. japonicum, the six macaques could not be stratified on the basis of worm burden at 22 weeks or on antibody titre. However, scholars were able to document the severe size reduction of worms, especially females, compared to worms from mice, with shrinkage of ovary and testes, and fewer eggs in the uterus and sperm in the seminal vesicle. These morphological changes in both schistosome species were attributed to immunological pressure, with the conclusion that it took several weeks for worms to die. This was clearly not the same as the rapid worm death in rats that occurred over a period of days. Adult worms acquire nutrients across both their surface and via the gut, the latter being more important in females with their greater needs owing to egg production ^[40][78]. The female’s response to cessation of blood feeding appears to be the gradual resorption of non-essential internal organs, leading to shrinkage of the body, while the shrinkage of males is less marked ^[37][75]. Ultimately, both sexes succumb to the pressure. What are the targets of the immune response in self-curing macaques? Characterisation of responses after self-cure from S. japonicum implicated the schistosome esophagus as a target for macaque antibodies ^[38][76]. Investigation of the morphology and secretions of the anterior and posterior esophageal glands (using RNA-Seq, in situ hybridization, and immunocytochemistry) identified >40 gland products comprising numerous MEGs, lysosomal hydrolases, and some potentially cytotoxic products involved in the initial processing of blood ^[41][80]. Evidence gathered by electron microscopy and immunocytochemistry indicated that several products served as in vivo targets for antibodies. With this in mind, scholars designed small 15-mer peptide arrays covering the amino acid sequences from 32 esophageal gland proteins. Scholars used these to screen the reactivity of 22-week sera from self-curing rhesus macaques versus serum from permissive mice and rabbit host with a chronic infection. ^[41][80]. Immunodominant regions were evident across species, notably MEGs 4.1, 4.2, 11, and 12, aspartyl protease, and a Tetraspanin 1 loop, while responses to MEGs 8.1C and 8.2C were largely confined to macaques. As proof of principle, three synthetic genes encoding key targets were designed, and one was expressed as a recombinant protein. When used to vaccinate rabbits, it elicited higher antibody titres to the majority of reactive regions than those elicited after prolonged infection. This clearly demonstrated the feasibility of simultaneously priming an animal against the reactive regions of multiple target proteins.

5. The Message for Vaccine Development

5.1. Schistosomes Are Hard to Kill In Vivo

• There is minimal evidence for killing of invading larvae in the skin in any model system, attractive as that might be. If slow development of immunity in humans is based on IgE responses, then it is emphatically not a route to a vaccine.
• After arming of the mouse (and rat) lungs by CD4+ T cells caused by a single dose of the RA vaccine, protection depends on blocked migration with some larval deflection into the alveoli. It is possible that the antibody-mediated protection displayed by IFNgR KO mice ^[17][33] works in the same way as the early elimination of challenge parasites observed in fast responder macaques ^[42][79]. Certainly, the larvae are not killed in the RA vaccinated mouse, but their progress is blocked.
• Antibody-mediated killing of adult worms by self-curing rhesus macaques is a protracted process of several weeks, involving immune pressure on multiple targets and differing specificities between individual animals due to MHC restriction of antigen presentation.
• Only in the rat self-cure process is there rapid and acute elimination of pre-adults. Soluble host-derived mediators (NO, histamine, serotonin, proteases) are the most likely agents, but the requirement for IgE and mast cells is not amenable to vaccine technology. However, the extended lifespan of worms in iNOS rats suggests a weak point that effector responses generating NO might usefully exploit.

5.2. What Are the Targets?

• The primary criterion is that the target(s) must be accessible to immune effectors in living parasites. Among current/recent human vaccine candidates, glutathione-S-transferases and fatty acid-binding protein Sm14 are located in the cytosol. The tetraspanin TSP-2 loops are likely exposed and accessible on the tegument surface. Smp80 calpain, while it lacks a signal peptide, is located at the tegument surface, and results from baboon vaccine experiments indicate it is accessible to antibodies ^[43][82].
• In recent years, transcriptomic and proteomic studies have identified a significant number of tegument surface constituents and components of esophageal gland and gastrodermal secretions that are largely untested in vaccine experiments (as many as 50 or 60). In parenthesis, there have been some very uncritical proteomic analyses of worm fractions and these issues have been dealt with in recent reviews ^[44][45][74,83].
• One unusual feature of the exposed proteins, revealed by peptide array analysis, is their often very low immunogenicity. Indeed, a bioinformatic analysis of MEG and VAL protein evolution across schistosome species suggests that they have been selected for immunological silence ^[46][84].
• Protection mediated by multiple targets seems more probable than a single magic bullet antigen. The feasibility of multi-epitope constructs has been demonstrated and these are now at the design/implementation stage for protection experiments.

5.3. What Does a Vaccine Need to Achieve?

• Negotiating the pulmonary vascular bed presents an obstacle in mice and rats that might be exploitable, and immunity conferred by arming of the lungs with memory/effector T cells appears to be reasonably persistent. Use of modified BCG as a vehicle, incorporating a multi-epitope construct, provides a potential way of achieving this by vaccination. Immunisation of mice with the rBCG-LTAK63 vaccine was recently shown to induce a persistent increase in memory and effector T cell numbers in lymph nodes and the lungs for at least 6 months after administration, which correlated with increased protection against Mycobacterium tuberculosis ^[47][85].
• Self-cure in the rhesus macaque offers the most as a paradigm for a human vaccine, but a strategy of eliciting persistent high titres and a rapid recall response from memory cells is paramount. The development of adjuvants that can accomplish this is an active field with the introduction of products like the synthetic glucopyranosyl lipid A (GLA) agonist of toll-like receptor-4 (TLR-4) ^[48][86]. This adjuvant has been extensively used in numerous studies with the Smp80 vaccine (e.g., ^[49][87]).
• A recent rhesus macaque experiment ^[42][79] indicated that a recall response in fast responder animals could be detected by one week after challenge. This was conducted with 700 cercariae, representing a biomass of ~25 μg protein. The developing larvae begin blood feeding and releasing esophageal and gastrodermal secretions in μg amounts from around day 8 ^[45][83], clearly sufficient to trigger a memory response. In the real world of a community living in an endemic region, most encounters will be with very small numbers of infective larvae on a sporadic basis.