DRUG RESISTANCE IN SCHISTOSOMES

In any discussion of drug resistance in schistosomes, it is necessary to define clearly what drug resistance is and to distinguish it from other explanations of sub-optimal drug activity (see discussion in Day and Botros, Reference Day, Botros, Maule and Marks2006). As opposed to tolerance, which represents an innate lack of susceptibility that is not in response to prior drug exposure, resistance is a heritable increase in the frequency of individuals in a population able to tolerate doses of a compound following exposure of that population to the drug (Prichard et al. Reference Prichard, Hall, Kelly, Martin and Donald1980; Coles and Kinoti, Reference Coles and Kinoti1997). Drug resistance therefore depends on the selective pressure of drug exposure, and is heritable. Thus, to distinguish properly between resistance and native tolerance, there needs to be some knowledge of the endogenous susceptibility of a particular population prior to drug administration; typically, such information is not available for schistosomiasis treatment programmes. Indeed, historically, it has been unusual to find any regular monitoring of susceptibility for these programmes. To complicate matters further, as noted above, there is already a significant background failure rate for a drug such as PZQ, as well as a range of factors other than resistance that can increase the incidence of drug failure (e.g. compromised health and immunocompetency of the host), as enumerated in Day and Botros (Reference Day, Botros, Maule and Marks2006). One of the factors suggested to account for persistence of infections following PZQ treatment is the reduced susceptibility of juvenile parasites to the drug (Cioli and Pica-Mattoccia, Reference Cioli and Pica-Mattoccia2003). Recent infections will contain a significant portion of PZQ-refractory juvenile worms, leading to less than optimal cure rates, and population genetic evidence from Brazil supports this idea. Thus, worms that persist following PZQ treatment were shown to have genotypes that do not differ significantly from susceptible worms, and therefore do not appear to represent a sub-population selected for PZQ resistance (Blanton et al. Reference Blanton, Blank, Costa, Carmo, Reis, Silva, Barbosa, Test and Reis2011). Despite these caveats, there are nonetheless several reports of possible PZQ resistance in schistosomes, both in the laboratory and in the field (reviewed by Fallon et al. Reference Fallon, Tao, Ismail and Bennett1996; Cioli, Reference Cioli2000; Day and Botros, Reference Day, Botros, Maule and Marks2006; Doenhoff et al. Reference Doenhoff, Cioli and Utzinger2008; Wang et al. Reference Wang, Wang and Liang2012b), including those using newer strategies for experimentally inducing and screening for PZQ-resistant schistosomes.

EXPERIMENTALLY-INDUCED PZQ RESISTANCE

Fallon and Doenhoff (Reference Fallon and Doenhoff1994) exploited an approach similar to that used to induce oxamniquine/hycanthone resistance in the laboratory (Cioli et al. Reference Cioli, Pica-Mattoccia and Archer1993) to select for resistance to PZQ in S. mansoni. Thus, sub-curative, but increasing PZQ doses were administered to S. mansoni-infected mice over seven passages through the life cycle (a separate group of worms was also selected for resistance to oxamniquine in this study). By the seventh life cycle passage, this PZQ drug pressure produced a population of schistosomes in which 93% of the worms survived a PZQ dose that killed 89% of control, unselected worms. Interestingly, PZQ- and oxamniquine-resistant worms showed no cross-resistance to the other drug, indicating that resistance to the two drugs arises via different mechanisms. More recently, experimentally-induced PZQ resistance has been reported in Schistosoma japonicum using a similar approach (Liang et al. Reference Liang, Li, Dai, Wang, Qu, Tao, Xing, Li, Qian and Wei2011).

A notable recent advance in obtaining PZQ-resistant schistosomes uses drug selection on the asexual stages of the life cycle in the snail host (Couto et al. Reference Couto, Coelho, Araujo, Kusel, Katz, Jannotti-Passos and Mattos2011). The technique derives from observations showing that treating S. mansoni-infected Biomphalaria glabrata snails with 1000 mg/kg PZQ interrupts almost 90% of cercarial shedding (Mattos et al. Reference Mattos, Pereira, Jannotti-Passos, Kusel and Coelho2007). Based on that finding, Couto et al. (Reference Couto, Coelho, Araujo, Kusel, Katz, Jannotti-Passos and Mattos2011) used successive treatments of B. glabrata infected with S. mansoni (LE strain) with the far lower dose of 100 mg/kg PZQ to select for cercariae that, upon infection of mice, developed into adult worms with a significantly reduced susceptibility to PZQ. The ED50 of PZQ for these LE-PZQ worms in mice was approximately five-fold higher than that for the parental LE strain (362 mg/kg for LE-PZQ vs 68 mg/kg for LE). Following PZQ, LE-PZQ worms were also less contracted than LE worms and, as assayed by fluorescent probes, showed less severe tegumental damage and, unlike LE worms, appeared to retain a functional excretory system (Couto et al. Reference Couto, Coelho, Araujo, Kusel, Katz and Mattos2010). The ability to use this approach to select for drug resistance at the snail stage is far less costly and labour intensive than previous strategies of applying drug pressure through multiple intramammalian-stage passages. It holds the promise of much more readily providing new drug-resistant isolates that will be useful for studying the mechanisms of PZQ resistance and perhaps lend insights into the mode of action of PZQ.

FIELD ISOLATES

There have been several reports of schistosome field isolates exhibiting reduced PZQ susceptibility. Many of these reports have been thoroughly reviewed and discussed by others, and will therefore be described only briefly here. More recent reports will also be included.

In Northern Senegal, lower than expected cure rates were initially reported in the 1990s for S. mansoni infections treated with PZQ. Alarmingly, cure rates were reported to be as low as 18% (Gryseels et al. Reference Gryseels, Stelma, Talla, van Dam, Polman, Sow, Diaw, Sturrock, Doehring-Schwerdtfeger, Kardorff, Decam, Niang and Deelder1994; Stelma et al. Reference Stelma, Talla, Sow, Kongs, Niang, Polman, Deelder and Gryseels1995). Subsequent follow-up studies and analysis of the data (Fallon, Reference Fallon1998; Gryseels et al. Reference Gryseels, Mbaye, De Vlas, Stelma, Guisse, Van Lieshout, Faye, Diop, Ly, Tchuem-Tchuente, Engels and Polman2001; Danso-Appiah and De Vlas, Reference Danso-Appiah and De Vlas2002) suggested that some portion (though not all) of this drug failure could be attributed to factors other than drug resistance, including high-intensity infection, rapid reinfection and transmission, presence of PZQ-refractory juvenile worms, variations in methodology for analysis of efficacy, and perhaps native tolerance of these schistosomes. Interestingly, in patients relocated to urban areas (which do not have ongoing transmission), cure rates rose to near-normal levels (Gryseels et al. Reference Gryseels, Mbaye, De Vlas, Stelma, Guisse, Van Lieshout, Faye, Diop, Ly, Tchuem-Tchuente, Engels and Polman2001). Furthermore, despite the reduced cure rates in the area of interest, PZQ treatment nonetheless dramatically lowered the infection intensity and curbed morbidity in treated individuals (reviewed by Fallon, Reference Fallon1998). On the other hand, worms from the Senegalese isolate exhibiting reduced cure rates are also less susceptible to PZQ when grown in experimentally-infected mice (Fallon, Reference Fallon1995; Fallon et al. Reference Fallon, Mubarak, Fookes, Niang, Butterworth, Sturrock and Doenhoff1997), suggesting loss of PZQ sensitivity is an endogenous trait of the worms themselves. Furthermore, in both human and mouse infections, these worms were susceptible to oxamniquine, which, like PZQ, is not effective against immature worms, casting some doubt on arguments suggesting low cure rates in this area may reflect large numbers of PZQ-refractory juvenile worms due to high rates of transmission (Fallon et al. Reference Fallon, Mubarak, Fookes, Niang, Butterworth, Sturrock and Doenhoff1997; Stelma et al. Reference Stelma, Sall, Daff, Sow, Niang and Gryseels1997).

Another site for intense study of potential PZQ resistance has been in Egypt, where schistosomes were isolated from several S. mansoni-infected patients (1·6% of those screened) who continued to pass viable eggs following three successive doses of PZQ (Ismail et al. Reference Ismail, Metwally, Farghaly, Bruce, Tao and Bennett1996). The schistosomes isolated from these patients were subsequently propagated in mice, where they showed 3–5-fold lower sensitivity to PZQ, as measured by ED50 (Ismail et al. Reference Ismail, Metwally, Farghaly, Bruce, Tao and Bennett1996, Reference Ismail, Botros, Metwally, William, Farghally, Tao, Day and Bennett1999). Tests of known responses of worms to PZQ in vitro (i.e. in the absence of any confounding host factors) showed that at least some of these isolates were less susceptible to the drug (Ismail et al. Reference Ismail, Botros, Metwally, William, Farghally, Tao, Day and Bennett1999; William et al. Reference William, Botros, Ismail, Farghally, Day and Bennett2001a). Indeed, in vitro measures of PZQ susceptibility correlated well in some cases with ED50 determinations in murine infections (William and Botros, Reference William and Botros2004), further indicating that factors in the worms themselves were responsible for the reduced PZQ susceptibility of these isolates. Interestingly, approximately half of the isolates tested retained their lower response to PZQ even after multiple passages through the life cycle in the absence of drug pressure, while others reverted. Indeed, application of drug pressure does not appear to be required for maintenance of the PZQ insusceptibility trait (Sabra and Botros, Reference Sabra and Botros2008), and ED50 differences, though reproducible, are relatively small (2–3-fold), and certainly not indicative of ‘super-resistant’ worms (Cioli and Pica-Mattoccia, Reference Cioli, Pica-Mattoccia, Secor and Colley2005). On the other hand, those isolates that did retain the trait often exhibited evidence of compromised biological fitness such as reduced cercarial production by infected snails (William et al. Reference William, Sabra, Ramzy, Mousa, Demerdash, Bennett, Day and Botros2001b). This observation, as well as those of others (Liang et al. Reference Liang, Coles, Dai, Zhu and Doenhoff2001) suggest that there are costs to schistosomes associated with lessening PZQ susceptibility. These costs may serve to limit the spread of PZQ resistance. Indeed, 10 years after the initial Egyptian studies, the same villages in Egypt in which the original PZQ failures were followed up and revealed no evidence of uncured patients despite a decade of drug pressure (Botros et al. Reference Botros, Sayed, Amer, El-Ghannam, Bennett and Day2005). The presence of large refugia in endemic areas may also limit the spread of resistance; indeed, PZQ-refractory immature schistosomes may act as a refugia (Webster et al. Reference Webster, Gower and Norton2008).

Further evidence for isolates showing PZQ insusceptibility has been found in Kenya. Researchers used an in vitro assay on miracidia hatched from eggs excreted by S. mansoni-infected Kenyan car washers to screen for S. mansoni exhibiting decreased susceptibility to PZQ (Melman et al. Reference Melman, Steinauer, Cunningham, Kubatko, Mwangi, Wynn, Mutuku, Karanja, Colley, Black, Secor, Mkoji and Loker2009). Different patients produced eggs that hatched into miracidia with variable PZQ sensitivity (as measured by miracidial killing); miracidia from previously-treated patients showed significantly lower sensitivity to the drug. Further characterization of an isolate from a patient who was never fully cured by PZQ (KCW) revealed that adult worms derived from these eggs were less sensitive to PZQ, both in vivo, in murine infections and in vitro, as assayed by schistosome length. Interestingly, the reduced susceptibility of one sub-isolate of KCW was heritable and persisted through at least 6 life cycle passages in the absence of drug pressure. However, a second KCW sub-isolate had reverted to a PZQ-susceptible state when retested after 8 generations. This now-susceptible sub-isolate survived; the sub-isolate that retained PZQ tolerance eventually perished (Melman et al. Reference Melman, Steinauer, Cunningham, Kubatko, Mwangi, Wynn, Mutuku, Karanja, Colley, Black, Secor, Mkoji and Loker2009). Thus, as with the Egyptian isolates, there appears to be variability in the stability of this trait, as well as a biological cost associated with PZQ insusceptibility.

There have also been attempts to assess the status of PZQ resistance in other species of schistosomes that infect humans (S. japonicum, S. haematobium). China has for many years relied on PZQ-based chemotherapy in its programme against S. japonicum infections. Wang et al. (Reference Wang, Wang and Liang2012b) recently reviewed several studies that monitored different endemic areas of China for evidence of PZQ insusceptibility. These studies found little if any evidence for emerging PZQ resistance, and suggest that, despite decades of intense chemotherapy, PZQ continues to be effective in treating schistosomiasis japonicum in China (Yu et al. 2001; Wang et al. Reference Wang, Dai, Li, Shen and Liang2010, Reference Wang, Dai, Li, Shen and Liang2012a; Seto et al. Reference Seto, Wong, Lu and Zhong2011). Isolated incidents of failure of PZQ to cure S. haematobium infections have been reported, including a notable case in which PZQ failed to cure Brazilian soldiers returning from Africa (Silva et al. Reference Silva, Thiengo, Conceição, Rey, Lenzi, Perreira Filho and Ribeiro2005), though there is currently no evidence for heritable resistance (Herwaldt et al. Reference Herwaldt, Tao, van Pelt, Tsang and Bruce1995; Alonso et al. Reference Alonso, Munoz, Gascon, Valls and Corachan2006).

Evidence for drug failure has also been found in other trematodes. Of particular interest is the liver fluke Fasciola hepatica. Though not particularly susceptible to PZQ, F. hepatica can be treated quite effectively with other compounds such as the benzimidazoles, which target β-tubulin, and are most frequently used as anti-nematodals. The benzimidazole triclabendazole (TCBZ) is effective against both immature and mature flukes (Boray et al. Reference Boray, Crowfoot, Strong, Allison, Schellembaum, von Orelli and Sarasin1983) and has seen widespread use since its introduction. Recent reports have suggested the localized emergence of TCBZ-resistant fluke isolates (Moll et al. Reference Moll, Gaasenbeek, Vellema and Borgsteede2000; Fairweather, Reference Fairweather2011), and work described below has focused on defining the underlying source of resistance in one of these TCBZ-resistant isolates.

MECHANISMS OF RESISTANCE

Resistance to a single class of drugs can arise via several mechanisms. The most obvious is target modification. For example, benzimidazoles such as albendazole act to inhibit microtubule polymerization; in nematodes and fungi, resistance has been mapped to a F200Y point mutation in β-tubulin (Kwa et al. Reference Kwa, Veenstra, Dijk and Roos1995). Similarly, simultaneous point mutations in three glutamate-gated chloride channel α-type subunits in Caenorhabditis elegans confer resistance (∼4000-fold) to the antiparasitic drug ivermectin (IVM) in these worms (Dent et al. Reference Dent, McHardy, Vassilatis and Avery2000). Interestingly, mutations in the Dyf (dye filling defective) class of genes, which appear to affect cuticle permeability, produce moderate IVM resistance (2–5-fold) and act additively to increase resistance of the channel mutations. This effect speaks to yet another mechanism for generation of resistance, namely heritable alterations that reduce drug availability or activity. These changes can be in uptake/permeability, activation/metabolism or drug efflux.

One of the more instructive cases of such non-target-dependent development of resistance comes from studies on schistosome resistance to oxamniquine (reviewed by Cioli et al. Reference Cioli, Pica-Mattoccia and Archer1995). As noted above, oxaminiquine is highly effective against S. mansoni, but lacks activity against other human schistosomes such as S. haematobium and S. japonicum (hycanthone, a related antischistosomal compound, is active against S. mansoni and S. haematobium, but not S. japonicum). In a series of elegant and challenging experiments using genetic crosses of drug-sensitive and drug-resistant schistosomes, Donato Cioli and colleagues showed that oxamniquine/hycanthone resistance in these worms was controlled by a single autosomal recessive gene. They also showed that the antischistosomal activity of the drug requires biotransformation to an active form by a parasite sulfotransferase. When activated, the drug is thought to act as an alkylating agent of schistosome DNA and other macromolecules, interfering with nucleic acid synthesis (Cioli and Pica-Mattoccia, Reference Cioli and Pica-Mattoccia1984; Cioli et al. Reference Cioli, Pica-Mattoccia and Moroni1992, Reference Cioli, Pica-Mattoccia and Archer1993). The drug is inactive against schistosome species that lack this sulfotransferase activity and drug resistance can arise when this activity is lost in species that normally express it (Pica-Mattoccia et al. Reference Pica-Mattoccia, Archer and Cioli1992, Reference Pica-Mattoccia, Novi and Cioli1997). More recent similar genetic studies on worms showing reduced sensitivity to PZQ suggest either dominant (Liang et al. Reference Liang, Dai, Zhu, Coles and Doenhoff2003) or partially dominant (Pica-Mattoccia et al. Reference Pica-Mattoccia, Doenhoff, Valle, Basso, Troiani, Liberti, Festucci, Guidi and Cioli2009) inheritance of the trait. Other examples of non-target-based mechanisms involved in anthelmintic drug action and development of resistance have recently been reviewed (Cvilink et al. Reference Cvilink, Lamka and Skalova2009; James et al. Reference James, Hudson and Davey2009).

With regard to PZQ failure, the fact that the PZQ target has not been rigorously defined makes the search for differences more problematic. However, no clear changes in candidate targets have been found to date. Thus, voltage-gated Ca²⁺ (Ca_v) channel β subunits have been implicated in PZQ action (Greenberg, Reference Greenberg2005; Nogi et al. Reference Nogi, Zhang, Chan and Marchant2009), but an examination of Ca_v channel β subunits in different isolates showing reduced PZQ susceptibility revealed no meaningful sequence differences or changes in expression levels (Valle et al. Reference Valle, Troiani, Festucci, Pica-Mattoccia, Liberti, Wolstenholme, Francklow, Doenhoff and Cioli2003; Kohn and Greenberg, unpublished data). On the other hand, reducing Ca_v channel subunit levels in the planarian Dugesia japonica confers resistance to these free-living platyhelminths against PZQ-elicited dramatic disruptions of normal regeneration patterns (Nogi et al. Reference Nogi, Zhang, Chan and Marchant2009; Zhang et al. Reference Zhang, Chan, Nogi and Marchant2011). The relationship between PZQ effects on planarian regeneration vs. its antischistosomal activity is not clear. Interestingly, however, pre-treating worms with the actin depolymerizing agent cytochalasin D renders S. mansoni refractory to PZQ, suggesting that changes in cytoskeletal dynamics can alter susceptibility to PZQ (Pica-Mattoccia et al. Reference Pica-Mattoccia, Valle, Basso, Troiani, Vigorosi, Liberti, Festucci and Cioli2007).

There are also several non-target-based changes that could alter PZQ effectiveness. For example, as noted above, juvenile schistosomes are refractory to PZQ. Additionally, adult female schistosomes, though still PZQ-sensitive, are more tolerant of the drug than adult males (Pica-Mattoccia and Cioli, Reference Pica-Mattoccia and Cioli2004). Thus, changes in worm maturation rates (Fallon et al. Reference Fallon, Mubarak, Fookes, Niang, Butterworth, Sturrock and Doenhoff1997) or sex ratios could influence the effectiveness of PZQ. Since PZQ-induced killing of S. mansoni within the mammalian host appears to be immune dependent (Brindley and Sher, Reference Brindley and Sher1987; Doenhoff et al. Reference Doenhoff, Sabah, Fletcher, Webbe and Bain1987; Brindley, Reference Brindley1994), another possibility is that loss or modulation of schistosome antigens that become exposed following PZQ treatment could lead to reduced antischistosomal activity. Interestingly, recent evidence indicates that two other platyhelminths (the trematode Dicrocoelium dendriticum and the cestode Hymenolepis nana) are not capable of enzymatically metabolizing PZQ (Vokřál et al. Reference Vokřál, Jirásko, Jedličková, Bártíková, Skálová, Lamka, Holčapek and Szotáková2012). Acknowledging the caveat that schistosomes may differ from these other platyhelminths, these results nonetheless suggest that development of more efficient PZQ metabolism by the parasite is not a particularly likely scenario for acquisition of PZQ resistance.

Molecular differences associated with reduced PZQ susceptibility in schistosomes would provide useful markers to monitor emergence of resistance in drug administration programmes. They could also serve as entrées into understanding how resistance develops and provide insights into the mechanism of drug action. There have been a handful of attempts to define such molecular correlates of PZQ resistance. For example, subtractive PCR and cloning of differentially-expressed RNAs revealed higher levels of an RNA encoding subunit 1 of mitochondrial cytochrome c-oxidase (SCOX1) in schistosomes selected for reduced PZQ susceptibility (Pereira et al. Reference Pereira, Fallon, Cornette, Capron, Doenhoff and Pierce1998). Analysis by semi-quantitative RT-PCR confirmed that the SCOX1 RNA was expressed at 5–10-fold higher levels in the resistant worms than in a PZQ-sensitive strain. Interestingly, no differences were found in expression of RNAs encoding SMDR2, a schistosome multidrug transporter (see below), nor NADH dehydrogenase subunit 5, another mitochondrial gene. Surprisingly, however, the enzymatic activity of cytochrome c-oxidase showed an expression pattern opposite to that found for the SCOX1 RNA. Thus, cytochrome c-oxidase activity in resistant worms was approximately 4-fold lower than in the PZQ-susceptible worms, an unexpected result given the 5–10-fold higher levels of SCOX1 RNA found in the resistant worms.

Another group (Tsai et al. Reference Tsai, Marx, Ismail and Tao2000) used random amplified polymorphic DNA (RAPD) PCR to test for markers of PZQ resistance. They found that an Egyptian isolate with reduced PZQ susceptibility (SO5) had 2 major differences in banding pattern from several PZQ-sensitive strains from the same endemic area of Egypt. Whether this difference can serve as a marker is unclear, as are any potential functional implications.

One of the more common mechanisms for development of drug resistance is through increased drug efflux, often mediated by multidrug transporters. Multidrug transporters underlie multidrug resistance (MDR), a phenomenon in which resistance to a single drug is accompanied by unexpected cross-resistance to several structurally unrelated compounds. Multidrug transporters have broad substrate specificity and actively remove xenobiotics and toxic compounds, including drugs, from cells and tissues, though non-transport-related MDR can also occur (Pommier et al. Reference Pommier, Pourquier, Urasaki, Wu and Laco1999, Reference Pommier, Sordet, Antony, Hayward and Kohn2004). Genes for multidrug transporters are found in all living cells (Blackmore et al. Reference Blackmore, McNaughton and Van Veen2001), and are classified into five basic families (Paulsen, Reference Paulsen2003; Higgins, Reference Higgins2007). The crystal structure of at least one representative of each of these families has been solved (van Veen, Reference van Veen2010). Broadly speaking, these different transporter types fit into one of two major classes, the primary-active transporters and the secondary-active transporters (Ventner et al. Reference Ventner, Shahi, Balakrishnan, Velamakanni, Bapna, Woebking and van Veen2005). The primary-active transporters couple translocation of substrate directly to the hydrolysis of ATP, while transport in the secondary-active transporters utilizes chemiosmotic energy derived from the electrochemical gradient of proton/sodium ions across the cytoplasmic membrane.

Members of the ATP-binding cassette (ABC) superfamily of transporters are primary-active transporters that comprise one of the largest groups of transmembrane proteins found in living cells (Dassa and Bouige, Reference Dassa and Bouige2001; Borst and Elferink, Reference Borst and Elferink2002). ABC transporters are found in organisms from all living kingdoms. They bind and hydrolyze ATP and use the resultant energy to translocate compounds across the membrane. ABC importers transport compounds into the cell, and are found in prokaryotes; ABC exporters are efflux transporters found in both prokaryotes and eukaryotes (Dassa and Bouige, Reference Dassa and Bouige2001; Saier and Paulsen, Reference Saier and Paulsen2001).

All ABC transporters share at least one highly conserved ATPase domain containing the Walker_A and Walker_B motifs typically found in ATPases as well as a specific signature motif. Full ABC transporters contain two of these cytoplasmic ATP-binding cassettes that alternate with two membrane-spanning domains; half transporters contain one of each of these structural features (Ambudkar et al. Reference Ambudkar, Kimchi-Sarfaty, Sauna and Gottesman2003; Szakacs et al. Reference Szakacs, Paterson, Ludwig, Booth-Genthe and Gottesman2006). Vertebrates have on the order of 50 ABC transporter genes that define seven distinct sub-families (designated ABCA to ABCG) based on phylogenetic analysis (Dean et al. Reference Dean, Rzhetsky and Allikmets2001; Dean and Annilo, Reference Dean and Annilo2005). Subsets of these ABC transporters are associated with MDR (Szakacs et al. Reference Szakacs, Paterson, Ludwig, Booth-Genthe and Gottesman2006). The S. mansoni genome appears to contain approximately 25 genes for ABC transporters, including several potentially involved in MDR (Kasinathan and Greenberg, Reference Kasinathan and Greenberg2012).

Largely because of its role in MDR in cancer chemotherapy, P-glycoprotein (Pgp; ABCB1) is the most thoroughly studied of the eukaryotic multidrug transporters, and mammalian and C. elegans Pgp have recently been crystallized and their structures solved (Aller et al. Reference Aller, Yu, Ward, Weng, Chittaboina, Zhuo, Harrell, Trinh, Zhang, Urbatsch and Chang2009; Jin et al. Reference Jin, Oldham, Zhang and Chen2012). MDR is linked to gene amplification, overexpression or mutation of Pgp or other multidrug transporters, resulting in increased drug efflux (reviewed by Borst and Elferink, Reference Borst and Elferink2002; Ambudkar et al. Reference Ambudkar, Kimchi-Sarfaty, Sauna and Gottesman2003; Szakacs et al. Reference Szakacs, Paterson, Ludwig, Booth-Genthe and Gottesman2006). In addition to Pgp, known ABC proteins involved in MDR include the multidrug resistance-associated proteins (MRPs; ABCCs), breast cancer resistance protein (BCRP; ABCG2), as well as others (Szakacs et al. Reference Szakacs, Paterson, Ludwig, Booth-Genthe and Gottesman2006).

Pgp and other multidrug transporters such as MRP1 transport a broad spectrum of compounds including several anticancer and other drugs (Kartner et al. Reference Kartner, Riordan and Ling1983; Higgins, Reference Higgins2007). Though the substrate specificities of the transporters show some overlap, there are clear preferences. Thus, Pgp shows selectivity for neutral and cationic hydrophobic compounds, while MRP1 preferentially transports organic anions, drugs and other compounds such as glutathione and other biotransformed conjugates, and signaling molecules such as the immunomodulator leukotriene C4 (reviewed by Ambudkar et al. Reference Ambudkar, Kimchi-Sarfaty, Sauna and Gottesman2003; Gimenez-Bonafe et al. Reference Gimenez-Bonafe, Guillen Canovas, Ambrosio, Tortosa, Perez-Tomas and Colabufo2008). Members of the ABC transporter family also show selectivity for biologically significant compounds such as lipids, steroids, cyclic nucleotides and peptides, indicative of their important roles in cellular and organismal physiology (Mizutani et al. Reference Mizutani, Masuda, Nakai, Furumiya, Togawa, Nakamura, Kawai, Nakahira, Shinkai and Takahashi2008; van de Ven et al. Reference van de Ven, Oerlemans, van der Heijden, Scheffer, de Gruijl, Jansen and Scheper2009). In addition to the broad selection of substrates that interact with these transporters, there are also a host of inhibitors that can reverse MDR by blocking multidrug transporter-mediated drug efflux (reviewed by Gimenez-Bonafe et al. Reference Gimenez-Bonafe, Guillen Canovas, Ambrosio, Tortosa, Perez-Tomas and Colabufo2008). Many of these inhibitors are inexpensive and safe compounds in wide clinical use (e.g. verapamil). Indeed, PZQ is an inhibitor of both mammalian and S. mansoni Pgp (Hayeshi et al. Reference Hayeshi, Masimirembwa, Mukanganyama and Ungell2006; Kasinathan et al. Reference Kasinathan, Goronga, Messerli, Webb and Greenberg2010a).

MULTIDRUG TRANSPORTERS IN SCHISTOSOMES AND OTHER TREMATODES

Could changes in multidrug transporter expression or structure be contributing to drug resistance in schistosomes? There are precedents in the literature for such an association, as ABC multidrug transporters such as Pgp have been implicated in drug resistance in other parasites, including parasitic helminths (reviewed by Kerboeuf et al. Reference Kerboeuf, Blackhall, Kaminsky and von Samson-Himmelstjerna2003; Jones and George, Reference Jones and George2005; James et al. Reference James, Hudson and Davey2009; Leprohon et al. Reference Leprohon, Legare and Ouellette2011; Lespine et al. Reference Lespine, Menez, Bourguinat and Prichard2012). For example, the macrocyclic lactone ivermectin is an anthelmintic that is both a substrate and inhibitor of Pgp. Indeed, the excellent safety profile of ivermectin is due in large part to Pgp in the blood-brain barrier excluding the drug from the host central nervous system; defects in Pgp in the blood-brain barrier result in hypersensitivity to ivermectin neurotoxicity (Schinkel et al. Reference Schinkel, Smit, van Tellingen, Beijnen, Wagenaar, van Deemter, Mol, van der Valk, Robanus-Maandag, te Riele, Berns and Borst1994; Mealey et al. Reference Mealey, Bentjen, Gay and Cantor2001). Ivermectin also likely interacts with nematode multidrug transporters, and resistance to it and other macrocyclic lactones is associated with changes in Pgp alleles or expression levels. Notably, several studies show that co-administration of MDR reversing agents (e.g. Pgp inhibitors such as verapamil) can increase the efficacy of macrocyclic lactones in drug-resistant and -sensitive nematodes (Xu et al. Reference Xu, Molento, Blackhall, Ribeiro, Beech and Prichard1998; Molento and Prichard, Reference Molento and Prichard1999; Bartley et al. Reference Bartley, McAllister, Bartley, Dupuy, Menez, Alvinerie, Jackson and Lespine2009; Tompkins et al. Reference Tompkins, Stitt, Morrissette and Ardelli2011; Ardelli and Prichard, Reference Ardelli and Prichard2013).

There have also been reports suggesting a role for ABC multidrug transporters in other platyhelminths. As noted above, certain F. hepatica isolates exhibit reduced susceptibility to TCBZ. Recent preliminary evidence indicates that an amino acid substitution in a critical region of F. hepatica Pgp is associated with TCBZ resistance (Wilkinson et al. Reference Wilkinson, Law, Hoey, Fairweather, Brennan and Trudgett2012). Work on another liver fluke, F. gigantica, provided evidence for expression of four ABC multidrug transporters in this worm, with expression of two of them increased in the presence of TCBZ in isolated fluke cells. Furthermore, efflux of rhodamine from these fluke cells could be inhibited by a MDR reversing agent (Kumkate et al. Reference Kumkate, Chunchob and Janvilisri2008).

Work on schistosome Pgp and other multidrug transporters essentially began in 1994, when cDNAs encoding Pgp (SMDR2) and an ABC half transporter (SMDR1) were cloned and sequenced (Bosch et al. Reference Bosch, Wang, Tao and Shoemaker1994). Two different oxamniquine/hycanthone-resistant isolates showed no evidence for amplification or overexpression of SMDR2. As noted above, the subsequent availability of the sequenced genome revealed ∼25 ABC transporter-like sequences in S. mansoni, including other Pgp-like genes and representatives of other ABC transporter sub-families (Kasinathan and Greenberg, Reference Kasinathan and Greenberg2012).

Fluorescent substrates of mammalian Pgp and MRP have been used to localize these substrates to the excretory system of schistosomes (Sato et al. Reference Sato, Kusel and Thornhill2002, Reference Sato, Kusel and Thornhill2004). PZQ dramatically disrupts the distribution of the Pgp substrate in PZQ-susceptible worms (Kusel et al. Reference Kusel, Oliveira, Todd, Ronketti, Lima, Mattos, Reis, Coelho, Thornhill and Ribeiro2006; Oliveira et al. Reference Oliveira, Kusel, Ribeiro and Coelho2006), but not in the recently-derived S. mansoni isolate selected at the snail stage for reduced PZQ susceptibility (Couto et al. Reference Couto, Coelho, Araujo, Kusel, Katz and Mattos2010). These results suggest a role for ABC multidrug transporters in schistosome excretory activity, and may be providing hints of a role for Pgp and other ABC transporters in PZQ resistance.

There are further indications that multidrug transporters may be involved in modulating levels of PZQ susceptibility in schistosomes. PZQ is both an inhibitor and a substrate of recombinant SMDR2 (Kasinathan et al. Reference Kasinathan, Goronga, Messerli, Webb and Greenberg2010a), and chronic exposure of worms to sub-lethal concentrations of PZQ results in up-regulation of SMDR2 and schistosome MRP1 (SmMRP1), and changes the distribution of anti-Pgp immunoreactivity in the worm (Messerli et al. Reference Messerli, Kasinathan, Morgan, Spranger and Greenberg2009; Kasinathan et al. Reference Kasinathan, Morgan and Greenberg2010b). Importantly, higher levels of schistosome SMDR2 and SmMRP1 are associated with reduced PZQ susceptibility (Messerli et al. Reference Messerli, Kasinathan, Morgan, Spranger and Greenberg2009; Kasinathan et al. Reference Kasinathan, Morgan and Greenberg2010b). Indeed, EE2, an Egyptian isolate with reduced PZQ susceptibility, expresses dramatically higher levels of SMDR2 RNA and protein (Messerli et al. Reference Messerli, Kasinathan, Morgan, Spranger and Greenberg2009); interestingly, SmMRP1 does not appear to be expressed at a higher level in EE2 (Kasinathan et al. Reference Kasinathan, Morgan and Greenberg2010b). At this juncture, only an association between reduced drug susceptibility and ABC multidrug transporters has been demonstrated. Further pharmacological and genetic experiments will be required to establish a causal relationship. However, we have recently used both of these approaches to implicate S. mansoni ABC multidrug transporters in schistosome reproduction (Kasinathan et al. Reference Kasinathan, Morgan and Greenberg2011).

NEW POST-GENOMIC APPROACHES TO DEFINE THE MECHANISMS OF DRUG RESISTANCE IN SCHISTOSOMES

For a variety of reasons, most notably that they are obligate parasites, schistosomes are notoriously difficult systems for experimental analysis. In recent years, however, tools that are feasible in other organisms have been, or are being, adapted to schistosomes. These new approaches, some of which are listed below, hold the promise of providing major advances in our knowledge about schistosome biology and physiology, including the underlying basis for anti-schistosomal drug action and drug resistance.

GENE MAPPING

As noted, traditional genetic experiments provided important insights into the mechanism of oxamniquine/hycanthone drug resistance and mode of action (Cioli et al. Reference Cioli, Pica-Mattoccia and Archer1993). Genetic crosses showed that oxamniquine/hycanthone resistance is linked to inheritance of a single autosomal recessive gene, while reduced PZQ susceptibility appears to be either a dominant (Liang et al. Reference Liang, Dai, Zhu, Coles and Doenhoff2003) or partially dominant (Pica-Mattoccia et al. Reference Pica-Mattoccia, Doenhoff, Valle, Basso, Troiani, Liberti, Festucci, Guidi and Cioli2009) trait. Newer, post-genomic approaches have the potential for much greater power. By combining the ability to conduct genetic crosses with the availability of a high resolution genetic linkage map (Criscione et al. Reference Criscione, Valentim, Hirai, LoVerde and Anderson2009) and an increasingly well-assembled and annotated genome sequence for S. mansoni (Protasio et al. Reference Protasio, Tsai, Babbage, Nichol, Hunt, Aslett, De Silva, Velarde, Anderson, Clark, Davidson, Dillon, Holroyd, LoVerde, Lloyd, McQuillan, Oliveira, Otto, Parker-Manuel, Quail, Wilson, Zerlotini, Dunne and Berriman2012), linkage mapping holds the promise to become a feasible approach for locating genes that underlie drug resistance. A particular appeal of this approach is that it surveys all of the genome, and is therefore not based on preconceived ideas about mechanism or possible candidate genes.

TRANSCRIPTOMICS AND OTHER ’-omics’

Examination of global changes in gene expression following drug treatment or between isolates showing differential drug susceptibility may provide an entrée into identification of genes underlying drug action or resistance. Microarray studies of schistosomes following low-dose PZQ treatment ex vivo have revealed molecular pathways that appear to be activated by PZQ, including expression of multidrug transporters and Ca²⁺ regulatory-, stress-, and apoptosis-related pathways (Aragon et al. Reference Aragon, Imani, Blackburn, Cupit, Melman, Goronga, Webb, Loker and Cunningham2009; Hines-Kay et al. Reference Hines-Kay, Cupit, Sanchez, Rosenberg, Hanelt and Cunningham2012). Comparison of the gene expression patterns of PZQ-refractory juvenile and PZQ-susceptible adult schistosomes following sub-lethal PZQ showed that the juvenile worms exhibited a greater transcriptomic flexibility that may allow them to respond to and survive exposure to PZQ.

Although microarray studies can be enlightening, the greater power of next-generation sequencing technologies such as RNAseq, along with robust bioinformatics algorithms (Cantacessi et al. Reference Cantacessi, Campbell, Jex, Young, Hall, Ranganathan and Gasser2012), may provide a higher-resolution analysis of meaningful gene expression changes in schistosomes following exposure to PZQ, as well as new drug targets. Furthermore, use of these approaches on worms treated with PZQ within the host should provide a better approximation of real-world responses to treatment. Examination of differences in microRNAs, which could produce reduced drug susceptibility by repressing expression of the drug target (or associated co-factors), could provide novel pathways to development of resistance (Devaney et al. Reference Devaney, Winter and Britton2010). Other ’-omics’ analyses of parasite responses to drug treatment should of course be explored; proteomic, glycomic, lipidomic and metabolomic approaches will likely also prove enlightening. Powerful ‘chemogenomics’ and comparative genomics approaches have the potential to provide new drug targets and perhaps insights into drug resistance (Caffrey et al. Reference Caffrey, Rohwer, Oellien, Marhofer, Braschi, Oliveira, McKerrow and Selzer2009; Swain et al. Reference Swain, Larkin, Caffrey, Davies, Loukas, Skelly and Hoffmann2011). Finally, in order to obtain a better understanding of mechanisms underlying resistance, these same tools should be brought to bear to compare drug-sensitive worms vs isolates with reduced susceptibility.

RNA INTERFERENCE (RNAI)

RNAi is gene silencing (or suppression) triggered by exogenous double-stranded (ds) RNA. Knockdown of genes using RNAi has proven to be an especially powerful molecular tool for analysis of a variety of gene functions. Though many parasitic nematodes appear to be refractory to RNAi (Britton et al. Reference Britton, Samarasinghe and Knox2012; Selkirk et al. Reference Selkirk, Huang, Knox and Britton2012), the methodology has proven relatively robust in schistosomes, and has provided important insights into schistosome biology (reviewed by Krautz-Peterson et al. Reference Krautz-Peterson, Bhardwaj, Faghiri, Tararam and Skelly2010). Both small interfering ds RNAs (siRNAs) and larger dsRNAs are effective. Analysis of phenotypes following knockdown of potential drug targets or of components of pathways hypothesized to modulate drug activity could provide important clues regarding drug action and mechanisms underlying resistance. However, there are several caveats for the use of this approach in schistosomes. For example, there can be tremendous variability in knockdown efficiency depending on the target (Mourao et al. Reference Mourao, Dinguirard, Franco and Yoshino2009; Stefanic et al. Reference Stefanic, Dvorak, Horn, Braschi, Sojka, Ruelas, Suzuki, Lim, Hopkins, McKerrow and Caffrey2010). To a first approximation, targets expressed in the intestine and on the tegument appear to be most amenable to knockdown. Off-target effects, in which dsRNA directs knockdown of transcripts other than those intended, have been reported in other systems and are an ever-present concern (Sioud, Reference Sioud2011). Knockdown is furthermore often partial even when successful, with perhaps 30–50% of transcripts remaining; a 50% level of expression might be sufficient to maintain function and mask any detectable phenotypes. Redundancy of genes can also often be a confounding factor. For example, as noted above, there are several predicted Pgp-like genes in S. mansoni, and suppression of one might be compensated for by the others, again masking a phenotype. Defining phenotypes other than those that are obvious (death, paralysis, egg production) can be challenging, even ex vivo, and particularly in vivo, within the host. Finally, it is currently technically difficult to perform these types of experiments in vivo, within the mammalian host. As knockdown of genes in infectious cercariae has not been reported, experiments are instead typically done by performing knockdown in schistosomules produced in vitro from cercariae. These schistosomules are then injected into the host, which is a far less efficient means of infection than using cercariae. Though this approach has been successful (see, for example, Bhardwaj et al. Reference Bhardwaj, Krautz-Peterson, Da'dara, Tzipori and Skelly2011), there is low and variable recovery of the adults that develop from these injected schistosomules, confounding data analysis. Reports of knockdown of parasite genes by injecting the infected host with siRNA have appeared (Pereira et al. Reference Pereira, Pascoal, Marchesini, Maia, Magalhaes, Zanotti-Magalhaes and Lopes-Cendes2008; Cheng et al. Reference Cheng, Fu, Lin, Shi, Zhou, Jin and Cai2009; Yang et al. Reference Yang, Jin, Liu, Shi, Cao, Liu, Shi, Li and Lin2012), though it remains to be seen if this approach will be incorporated more generally in the field.

TRANSGENESIS

Although RNAi is an extraordinarily powerful tool for analyzing gene function, a more complete armamentarium for functional genomics in schistosomes will require the availability of somatic and germline transgenesis. In conjunction with the completion of draft genomes for all three major human schistosome species and many years of concerted efforts, relatively efficient gene insertion and knockout strategies may be coming to fruition for schistosomes (Tchoubrieva and Kalinna, Reference Tchoubrieva and Kalinna2010; Beckmann and Grevelding, Reference Beckmann and Grevelding2012; Suttiprapa et al. Reference Suttiprapa, Rinaldi and Brindley2012). Of particular note is an exciting recent report of murine leukaemia virus-mediated germ-line transgenesis and insertional mutagenesis in S. mansoni (Rinaldi et al. Reference Rinaldi, Eckert, Tsai, Suttiprapa, Kines, Tort, Mann, Turner, Berriman and Brindley2012a). Other technical advances have also been steadily appearing, including antibiotic selection of transgenic worms (Rinaldi et al. Reference Rinaldi, Suttiprapa, Tort, Folley, Skinner and Brindley2012b) and vector-mediated RNAi (Tchoubrieva et al. Reference Tchoubrieva, Ong, Pike, Brindley and Kalinna2010; Duvoisin et al. Reference Duvoisin, Ayuk, Rinaldi, Suttiprapa, Mann, Lee, Harris and Brindley2012). These types of advances offer the promise of more feasible transgenesis in schistosomes, with more widespread adoption, and hold the promise of new strategies for studies of drug action and resistance. A model for the type of power that could be brought to bear on these questions was recently provided in studies using transgenesis of C. elegans resistant to the anthelmintic emodepside. These C. elegans were transformed with genes from other nematodes or mammals to investigate the role of Ca²⁺-activated potassium (SLO-1) channels in the selectivity and mode of action of this drug (Crisford et al. Reference Crisford, Murray, O'Connor, Edwards, Kruger, Welz, von Samson-Himmelstjerna, Harder, Walker and Holden-Dye2011; Welz et al. Reference Welz, Kruger, Schniederjans, Miltsch, Krucken, Guest, Holden-Dye, Harder and von Samson-Himmelstjerna2011).

OTHER ADVANCES

One of the challenges of working with schistosomes is that they are obligate parasites and, currently, only one developmental stage (schistosomules) can be archived by freezing for re-establishment of the life cycle (Cooper et al. Reference Cooper, Lewis and File-Emperador1989). Though primary S. mansoni cells such as muscle fibres have been used to study drug and neurotransmitter action (Novozhilova et al. Reference Novozhilova, Kimber, Qian, McVeigh, Robertson, Zamanian, Maule and Day2010), the establishment of cell lines in schistosomes would be a huge boon for the field, and could provide high-throughput screening opportunities to investigate drug targets and test hypotheses regarding mechanisms of resistance. To date, no such cell lines have been successfully established, but newer approaches for immortalization of cells are promising (Quack et al. Reference Quack, Wippersteg and Grevelding2010), and advances in the study and culture of multipotent cells (neoblasts) from Schistosomes and other platyhelminths (Eisenhoffer et al. Reference Eisenhoffer, Kang and Sanchez Alvarado2008; Brehm, Reference Brehm2010; Collins, et al. Reference Collins, Wang, Lambrus, Tharp, Iver and Newmark2013) may eventually prove adaptable to schistosomes.

High- and medium-throughput systems for screening schistosome phenotypes such as paralysis or contraction will also be useful. These include video analysis systems (Caffrey et al. Reference Caffrey, Rohwer, Oellien, Marhofer, Braschi, Oliveira, McKerrow and Selzer2009) and a novel adoption of the xCelligence (Roche) system for measuring electrical impedance across micro-electrodes interdigitated on the bottom of tissue culture plates (Smout et al. Reference Smout, Kotze, McCarthy and Loukas2010). New and powerful markers that have been developed for S. mansoni organs and developmental stages (Collins et al. Reference Collins, King, Cogswell, Williams and Newmark2011) should also aid in these types of analyses.

CONCLUDING REMARKS

The prospect of schistosomes developing widespread resistance to PZQ is an alarming one, and understanding mechanisms by which resistance might emerge could provide markers for monitoring response to mass treatment programmes, as well as possible strategies for reversing drug resistance. It is somewhat heartening that worms exhibiting apparent PZQ resistance often appear to be compromised, and furthermore, exhibit relatively small (2–5-fold) changes in drug susceptibility, as measured by ED50s. Indeed, typical treatment with PZQ uses an ED90 dose, which should be sufficient to eliminate worms showing 2–3-fold reductions in susceptibility (Cioli and Pica-Mattoccia, Reference Cioli, Pica-Mattoccia, Secor and Colley2005). Nonetheless, complacency is clearly unwarranted, as in both Egypt and Kenya (Ismail et al. Reference Ismail, Botros, Metwally, William, Farghally, Tao, Day and Bennett1999; Melman et al. Reference Melman, Steinauer, Cunningham, Kubatko, Mwangi, Wynn, Mutuku, Karanja, Colley, Black, Secor, Mkoji and Loker2009), isolates were derived from patients who simply were not cured by such doses. These patients continued to pass eggs (presumably carrying a PZQ-tolerant trait), suggesting that clinically-significant treatment failures potentially indicative of emerging resistance can occur in the field. One hopes that some of the newer technologies described in this review may lead to development of tools to deal with such resistance before disastrous consequences ensue.