The anthelmintic drug praziquantel activates a schistosome transient receptor potential channel

The anthelmintic drug praziquantel (PZQ) is used to treat schistosomiasis, a neglected tropical disease that affects over 200 million people worldwide. PZQ causes Ca2+ influx and spastic paralysis of adult worms and rapid vacuolization of the worm surface. However, the mechanism of action of PZQ remains unknown even after 40 years of clinical use. Here, we demonstrate that PZQ activates a schistosome transient receptor potential (TRP) channel, christened Sm.TRPMPZQ, present in parasitic schistosomes and other PZQ-sensitive parasites. Several properties of Sm.TRPMPZQ were consistent with known effects of PZQ on schistosomes, including (i) nanomolar sensitivity to PZQ; (ii) stereoselectivity toward (R)-PZQ; (iii) mediation of sustained Ca2+ signals in response to PZQ; and (iv) a pharmacological profile that mirrors the well-known effects of PZQ on muscle contraction and tegumental disruption. We anticipate that these findings will spur development of novel therapeutic interventions to manage schistosome infections and broader interest in PZQ, which is finally unmasked as a potent flatworm TRP channel activator.

Schistosomiasis (bilharzia) is a parasitic worm infection that infects millions of people worldwide (1,2). Mature blood flukes living in the vasculature lay eggs, which become deposited in host tissues, where they trigger local inflammatory responses. Chronic infections become associated with fibrosis and obstructive disease in gastrointestinal tissues and liver (Schistosoma mansoni, Schistosoma japonicum), genitourinary disease (Schistosoma haematobium), anemia, undernutrition, and a heightened risk for other comorbidities (3). The annual disease burden has been estimated as a loss of up to 70 million disability-adjusted life years (1,2).
In 2017, ϳ100 million people (ϳ80 million school-aged children) received free preventive treatment for schistosomiasis.
This treatment depends on a drug called praziquantel (PZQ), 2 as no effective vaccine currently exists (4). The clinical formulation of PZQ is a racemate (ϮPZQ) composed of the enantiomers (R)-PZQ and (S)-PZQ. (R)-PZQ is the antischistosomal eutomer, known to cause Ca 2ϩ influx and spastic paralysis of adult worms and rapid vacuolization of the worm tegumental surface (5). (S)-PZQ is regarded as the less active distomer (6). From a therapeutic perspective, it is problematic that despite decades of clinical usage, as well as demonstration of strains with lower sensitivity to PZQ in both laboratory and field, the flatworm target(s) of PZQ remains unknown (7,8). This lack of knowledge is a longstanding roadblock for this field.
Although no binding site(s) for these enantiomers has been identified in parasitic flatworms, there has been considerable recent progress in identifying targets for (R)-PZQ and (S)-PZQ in the human host (9). (R)-PZQ is a partial agonist of the human 5-hydroxytryptamine 2B receptor (5HT 2B R (10)), and (S)-PZQ is a partial agonist of the human transient receptor potential melastatin-8 channel (hTRPM8 (11)). Whereas regulation of these host targets occurs over the micromolar range (10 -12), molecular divergence between human and flatworm ligandbinding pockets (13,14) makes it reasonable to anticipate different binding poises and affinities at a homologous schistosome target(s).
Early work on schistosomes established key pharmacological characteristics of PZQ action on parasite muscle contraction and/or 45 Ca 2ϩ uptake. These include (i) conversion of contraction from sustained to phasic in the presence of elevated Mg 2ϩ , (ii) inhibition by La 3ϩ , and (iii) insensitivity to several voltageoperated Ca 2ϩ channel (Ca v ) blockers at specific doses. We therefore examined the impact of these same manipulations on Sm.TRPM PZQ activity. First, increasing the Mg 2ϩ /Ca 2ϩ ratio to a level (75:1) that resulted in transient muscle contraction (15,16) also resulted in a transient PZQ-evoked Ca 2ϩ signal via Sm.TRPM PZQ (Fig. 2E). Second, preincubation of worms with La 3ϩ (10 mM) inhibited both PZQ-evoked 45 Ca 2ϩ accumulation and PZQ-evoked contraction (17). La 3ϩ (10 mM) also inhibited Sm.TRPM PZQ activity (Fig. 2F). Third, three Ca v blockers (methoxyverapamil, nifedipine, and nicardipine) that failed to block PZQ action on worms (17,18) also failed to inhibit PZQ-evoked Sm.TRPM PZQ activity at the same doses (Fig. 2F). Therefore, the pharmacological properties of Sm.TRPM PZQ mirror the characteristics of PZQ action on schistosome muscle.
Consistent with the homology-based search strategy, Sm. TRPM PZQ is a member of the TRP melastatin (TRPM) subfamily. Sequence analysis revealed an architecture characteristic of TRPM channels (Fig. 2G), a well-represented family within flatworm genomes (19). Features include a long N-terminal TRPM homology region (MHR) domain, followed by six predicted transmembrane (TM) domains with a pore-forming re-entry loop between TM5 and TM6, a conserved TRP helix juxtaposed to coiled-coil regions, and a cytoplasmic C-terminal enzymatic domain (Fig. 2G). This enzyme domain displayed homology with the human ADP-ribose (ADPR) pyrophosphatase NUDT9, a feature characteristic of TRPM2 channels (20 -23). TRPM2 and TRPM8 are closely related "long" TRPM channels, and Sm.TRPM PZQ displays the highest sequence iden-tity with these human TRPM variants (29.5 and 28.5% sequence identity with hTRPM2 and hTRPM8, respectively).
Analysis of flatworm genomic and transcriptomic data sets revealed the presence of Sm.TRPM PZQ homologs in other parasitic flatworms, including cestodes and flukes, known to exhibit PZQ sensitivity (Fig. S1A). To assess the broader PZQ sensitivity of schistosome TRP channels, we screened three other TRPs. First, we examined the previously characterized Sm.TRPA, which has been shown to activated by the ligands AITC and capsaicin (14). Sm.TRPA did not respond to PZQ but, as expected, did respond to the other two compounds (Fig.  S1B). Next, we focused on the schistosome TRPM subfamily, which is predicted to contain seven members (Fig. S1A). The two members most closely related to Sm.TRPM PZQ (Smp_130890 and Smp_000050) did not respond to PZQ (Fig. S1, C and D). With the caveat that there is no control for functional expression, as endogenous agonists of these TRPM channels are unknown, these data suggest that schistosome TRP (and TRPM) channels are not broadly sensitive to PZQ.
Next, to resolve the single-cell kinetics of Sm.TRPM PZQ activity, we performed confocal Ca 2ϩ imaging. In HEK cells transfected with empty vector, the addition of ϮPZQ (10 M) failed to evoke a cytoplasmic Ca 2ϩ signal (Fig. 3, A and B), although cells responded to ATP (100 M), which activated endogenous purinoceptors. In contrast, in HEK cells transiently transfected with Sm.TRPM PZQ , the addition of ϮPZQ (1 M) evoked a rapid and protracted rise in cytoplasmic Ca 2ϩ (Fig. 3, A and B). Responses were evoked by (R)-PZQ, with (S)-PZQ being ineffective at the same concentration (1 M; Fig.  3, A and B). The large and persistent increase in fluorescence evidenced little Sm.TRPM PZQ desensitization in the presence of ϮPZQ and contrasted with the smaller, transient nature of Ca 2ϩ signals evoked by ATP. This signal was triggered by Ca 2ϩ influx, as this response was seen only when Ca 2ϩ -containing medium was re-added to HEK cells initially exposed to ϮPZQ in Ca 2ϩ -free medium (Fig. S2A). Activation of Sm.TRPM PZQ by ϮPZQ was also reversible, as ϮPZQ washout resulted in a decrease of signal to baseline (Fig. S2B).
Electrophysiological analysis of Sm.TRPM PZQ was performed by measuring whole-cell currents in HEK cells expressing GFP alone or expressing GFP and Sm.TRPM PZQ . In cells expressing GFP alone, the addition of ϮPZQ (2 M) did not evoke currents (0 of 18 cells examined). In contrast, in HEK  (22)), the pre-S1 helix (shaded), the six TM-spanning helices (S1-S6) comprising the voltage sensor-like domain (VSLD; blue) and pore-forming domain (red), the TRP domain (purple), the rib and pole helices (yellow), an additional helical domain (black), and the C-terminal NUDT9H domain.
ACCELERATED COMMUNICATION: (R)-PZQ activates a schistosome TRP cells co-transfected with cDNA encoding both Sm.TRPM PZQ and GFP, the addition of ϮPZQ evoked rapidly activating inward currents in all GFP-positive cells (22 of 22 cells, holding potential of Ϫ40 mV). Characterization of current magnitude after various voltage steps, in the absence and presence of PZQ (2 M), revealed PZQ-activated Sm.TRPM PZQ -conducted large inward and outward currents with a linear I-V relationship (Fig. 3C), resembling the linear I-V relationship displayed by hTRPM2 channels (24). Based on sequence homology with another invertebrate TRPM2 channel (Nematostella vectensis ACCELERATED COMMUNICATION: (R)-PZQ activates a schistosome TRP TRPM2, Nv.TRPM2) that has been structurally and functionally characterized (25), we speculated that the substantial Ca 2ϩ permeability of Sm.TRPM PZQ (Fig. 3, B and C) is supported by the presence of a negatively charged residue in the predicted pore filter of Sm.TRPM PZQ (FGD in Fig. 3D). This closely resembles the pore filter sequence of Nv.TRPM2 (YGE in Fig.  3D), which displays substantial Ca 2ϩ permeability (25). Consistent with this idea, PZQ-evoked Ca 2ϩ signals were strongly attenuated in HEK cells expressing the mutant Sm.TRPM PZQ [D1602A] (Fig. 3E). Sm.TRPM PZQ therefore displays several characteristics consistent with the properties of TRPM2 channels.

Discussion
These data represent the first report of a flatworm target activated by PZQ. Although further experiments would be needed to confirm Sm.TRPM PZQ as the clinically relevant target in worms, our data clearly evidence Sm.TRPM PZQ as a schistosome target of PZQ.
The properties of Sm.TRPM PZQ , a TRPM2-like channel, are, however, consistent with several key facets of PZQ action on worms. These include (i) nanomolar sensitivity to PZQ (Fig. 2, C and D); (ii) stereoselectivity toward (R)-PZQ (Figs. 2 and 3); (iii) mediation of a sustained Ca 2ϩ entry in response to PZQ (Fig. 3B) that parallels the kinetics of worm contracture and tegumental disruption (15)(16)(17)26); (iv) partial blockade by Mg 2ϩ and complete inhibition by La 3ϩ , mirroring the effects of PZQ on muscle contraction and tegumental disruption (15)(16)(17)26); (v) insensitivity to specific Ca v blockers that fail to block PZQ action on worms (Fig. 2F) (16 -18); and (vi) presence of homologs in other parasitic flatworms sensitive to PZQ (Fig.  S1). Just as Sm.TRPM PZQ supports long-lasting cellular Ca 2ϩ signals (Figs. 2 and 3), human TRPM2 (hTRPM2) also exhibits long channel opening times that support substantial Ca 2ϩ influx (23,27). hTRMP2 is a well-known effector of apoptosis being responsive to reactive oxygen species through activation by H 2 O 2 and ADPR (23,28). Activation of hTRMP2 at the cell surface and within intracellular organelles causes lysosomal permeabilization and cell death (28 -30). Such regulation could underpin the deleterious actions of ϮPZQ on worm tegument crucial for the in vivo efficacy of PZQ (31,32). Therefore, there are many similarities between the properties of Sm.TRPM PZQ and the characteristics of PZQ action on schistosomes.
This discovery also prompts new questions. What are the endogenous agonists and/or environmental cues that regulate Sm.TRPM PZQ activity across the parasite life cycle? In what cell type(s) is Sm.TRPM PZQ expressed? How is Sm.TRPM PZQ activity regulated in juvenile worms known to be less sensitive to PZQ? Is Sm.TRPM PZQ activity altered in schistosome strains that show refractoriness to PZQ action? Mutagenesis demonstrates that single amino acid changes in Sm.TRPM PZQ can dramatically alter channel responses to ϮPZQ (Fig. 3E). This discovery also prioritizes analyses of TRPM PZQ homologs in other flatworms as well as all other schistosome TRPM channels to assess broader PZQ sensitivity.
Finally, we note that (R)-PZQ is a potent activator of Sm.TRPM PZQ (Fig. 2). Known regulators of hTRPM2, including the endogenous agonist ADPR (23), act over the micromolar range. This is important as hTRPM2 is an emerging clinical target for several nervous system and inflammatory disorders (23,27). Understanding the basis of (R)-PZQ affinity for Sm.TRPM PZQ and comparing regulation and gating of Sm.TRPM PZQ with recently solved TRPM structures (20 -22, 33) may reciprocally catalyze drug design at this clinically important human target.

Reagents
Enantiomers of ϮPZQ were resolved following the protocol of Woelfle et al. (34). All chemical reagents were from Sigma. Cell culture reagents were from Invitrogen. Lipofectamine 2000 was from Thermo Fisher Scientific.

Adult schistosome mobility assays
Adult schistosomes were recovered by dissection of the mesenteric vasculature in female Swiss Webster mice previously infected (ϳ49 days) with S. mansoni cercariae (NMRI strain) by the Schistosomiasis Resource Center at the Biomedical Research Institute (Rockville, MD). All animal experiments followed ethical regulations approved by the Medical College of Wisconsin institutional animal care and use committee. Harvested schistosomes were washed in RPMI 1640 supplemented with HEPES (25 mM), 5% heat-inactivated fetal bovine serum (FBS) (Gibco), and penicillin-streptomycin (100 units/ml) and incubated overnight (37°C/5% CO 2 ) in vented Petri dishes (100 ϫ 25 mm). The following day, movement assays were performed using male worms in 6-well dishes (ϳ5 individual worms/3 ml of medium per well). Video recordings were captured using a Zeiss Discovery v20 stereomicroscope with a QiCAM 12-bit cooled color CCD camera controlled by Metamorph imaging software. Recordings (1 min) of worm motility ACCELERATED COMMUNICATION: (R)-PZQ activates a schistosome TRP (4 frames/s), during the addition of various drug concentrations were analyzed as described previously (13).

Molecular cloning
For cloning of Sm.TRPM PZQ , total RNA was isolated from adult schistosome worm pairs using TRIzol and poly(A)-purified using a NucleoTrap mRNA minikit. cDNA was synthesized using the SuperScript TM III first-strand synthesis system (Invitrogen). Using the predicted sequence (Smp_246790) as a template, cDNA from transcribed sequences was amplified by PCR (LA Taq TM polymerase) and ligated into pGEM-T Easy (Promega) for sequencing. Several splice variants of Sm.TRPM PZQ were identified within both the N-terminal TRPM homology region (MHR) and cytoplasmic C-terminal domain, which will be characterized elsewhere. The sequence used here for functional analyses represents the reference sequence (2268 amino acids, Smp_246790.5).

Ca 2؉ -imaging assays
Ca 2ϩ -imaging assays were performed using a fluorescence imaging plate reader (FLIPR TETRA , Molecular Devices). HEK293 cells (naive or transfected) were seeded (50,000 cells/ well) in a black-walled clear-bottomed poly-D-lysine-coated 96-well plate (Corning) in Dulbecco's modified Eagle's medium supplemented with 10% dialyzed FBS. After 24 h, growth medium was removed, and cells were loaded with a fluorescent Ca 2ϩ indicator (Fluo-4 direct dye, Invitrogen) by incubation (100 l per well, 1 h at 37°C) in Hanks' balanced salt solution (HBSS) assay buffer containing probenecid (2.5 mM) and HEPES (20 mM). Drug dilutions were prepared in assay buffer, without probenecid and dye, in V-shaped 96-well plates (Greiner Bio-one, Frickenhausen, Germany). After loading, the Ca 2ϩ assay was performed at room temperature. Basal fluorescence was monitored for 20 s, and then 25 l of each drug was added, and the signal (raw fluorescence units) was monitored over an additional 250 s. For quantitative analyses, peak fluorescence in each well was normalized to maximum -fold increase over baseline.
For confocal Ca 2ϩ imaging, HEK cells were loaded with Fluo-4-AM (4 M) and Pluronic F127 (0.4%) for 25 min at room temperature. Cells were then washed twice with HBSS and incubated at room temperature for de-esterification (30 min). Experiments in U2OS cells (Fig. S2A and B) were done using the genetically encoded calcium indicator, GCaMP6M. Fluorescence was imaged on an Olympus IX81 microscope, and fluorescence changes ( ex ϭ 488 nM ( em ϭ 513 Ϯ 15-nm bandpass) were monitored using a Yogokawa spinning disk confocal (CSU-X-M1N) and an Andor iXon Ultra 888 EMCCD camera.
Data were expressed as a ratio (F/F 0 ) of fluorescence at any given time (F) relative to fluorescence prior to drug addition (F 0 ).

Electrophysiology
For whole-cell current recordings, HEK293 cells were transfected with a plasmid encoding GFP or co-transfected with plasmids encoding GFP and Sm.TRPM PZQ . One day later, cells were replated onto round 18-mm glass coverslips. After overnight incubation, coverslips were secured in a recording chamber over a Nikon Eclipse TE200 inverted microscope. Cells were continuously superfused (6 ml/min) with an extracellular buffer consisting of 140 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, and 10 mM glucose (pH 7.4, 310 Ϯ 3 mosM at room temperature). HEK293 cells were held at a holding voltage of Ϫ40 mV, and responses were resolved after superfusion of extracellular buffer containing ϮPZQ (2 M). Recordings were made using borosilicate pipettes (Sutter Instrument Company, Novato, CA) pulled on a Sutter micropipette puller (model P-87) to resistances of 2-5 megaohms. Patch pipettes were filled with intracellular buffer containing 135 mM KCl, 10 mM NaCl, 1 mM MgCl 2 , 1 mM EGTA, 0.2 mM Na.GTP, 2.5 mM ATP.Na 2 , and 10 mM HEPES (pH 7.20, 290 Ϯ 3 mosM at room temperature). Cell capacitance was compensated, and series resistance was kept Ͻ10 megaohms. Cells were included in analyses if the leak current stayed Ͻ200 pA. Recordings were made using an EPC10 USB amplifier (HEKA Electronics) and Patch Master software (HEKA Electronics). Patch-clamp data were analyzed using Pulse, PulseFit, or Fitmaster software (HEKA Electronics). For current-voltage measurements of HEK293 cells expressing Sm.TRPM PZQ , step potentials of 250 ms spanning the voltage range from Ϫ80 to ϩ120 mV were delivered from a holding potential of Ϫ80 mV. For I-V curves, patch pipettes were filled with intracellular buffer containing: 140 mM CsMeSO 4 , 1 mM MgCl 2 , 1 mM EGTA, 10 mM HEPES-CsOH (pH 7.2 with CsOH, 300 -310 mOsm/kg adjusted with sucrose).