The Anthelmintic Activity of Praziquantel Analogs Correlates with Structure–Activity Relationships at TRPM_PZQ Orthologs

Abstract

The anthelmintic drug praziquantel remains a key clinical therapy for treating various diseases caused by parasitic flatworms. The parasite target of praziquantel has remained undefined despite longstanding usage in the clinic, although a candidate ion channel target, named TRPM_PZQ, has recently been identified. Intriguingly, certain praziquantel derivatives show different activities against different parasites: for example, some praziquantel analogs are considerably more active against cestodes than against schistosomes. Here we interrogate whether the different activities of praziquantel analogs against different parasites are also reflected by unique structure–activity relationships at the TRPM_PZQ channels found in these different organisms. To do this, several praziquantel analogs were synthesized and functionally profiled against schistosome and cestode TRPM_PZQ channels. Data demonstrate that structure–activity relationships are closely mirrored between parasites and their TRPM_PZQ orthologs, providing further support for TRPM_PZQ as the therapeutically relevant target of praziquantel.

Exactly 40 years ago, a highly influential review on the anthelmintic activity of praziquantel (PZQ) was published by Peter Andrews and Herbert Thomas (both at Bayer AG) and Rolf Pohlke and Jürgen Seubert (both at E. Merck KG).¹ That work detailed the discovery of the anthelmintic activity of PZQ, derivatization of the scaffold, the drug’s pharmacokinetic and safety profile, and the broad efficacy of this new therapeutic agent against a range of parasitic flatworms. The summary of data interrogating the activity of different pyrazino[2,1-a]isoquinoline derivatives against a representative trematode (Schistosoma mansoni) and cestode model (Hymenolepis nana) established both the “tightness” of the pharmacophore that underpins the efficacy of PZQ and a “structure–activity” fingerprint for the action of this drug that has long served as a reference standard for the field.

Following decades of clinical usage of PZQ for treatment of various diseases caused by parasitic flatworms,²⁻⁵ a candidate target was recently identified⁶ in Schistosoma mansoni. This target is a transient receptor potential (TRP) ion channel of the melastatin subfamily, named Sm.TRPM_PZQ,⁶⁻⁸ that mirrors the structure–activity relationship (SAR) of PZQ derivatives⁷ as described by Andrews et al.¹ TRPM_PZQ is a large nonselective cation channel unique to flatworms.^6,8−10 Activation of Sm.TRPM_PZQ is thought to elicit excitotoxicity through membrane depolarization, spastic contraction, and surface damage to the parasite, which then catalyzes immunological clearance from infected hosts.¹¹ Additional evidence from genetic association studies,¹² pharmacological screening,¹³ and functional profiling of TRPM_PZQ orthologs^9,10 add support for TRPM_PZQ serving as the therapeutically relevant parasite target of PZQ.

The work of Andrews et al., however, provides an additional opportunity to interrogate the candidacy of TRPM_PZQ.¹ From the extensive list of PZQ derivatives described by Andrews et al.,¹ a small number of analogs displayed good activity against cestodes but not against schistosomes. These analogs of PZQ encompassed replacement of the cyclohexyl group by 3-pyridyl ((±)-2, (R)-2, and (S)-2), 4-nitrophenyl (3), 4-N-methylaniline (4), 4-N,N-dimethylaniline (5), 4-aniline (6), cis-4-aminocyclohexyl (7) and -cyclopropyl (8) groups (Chart 1). Do such analogs, which exhibit differential antiparasitic activity versus cestodes and schistosomes, display differential activity at schistosome and cestode TRPM_PZQ? If the specific SAR of PZQ analogs against these different parasites mimics the SAR at the corresponding TRPM_PZQ channel, these data would provide further support for TRPM_PZQ as the clinically relevant in vivo target of PZQ. To tackle this question, we synthesized eight PZQ analogs from Andrews et al.¹ that exhibited divergent bioactivity against schistosomes and cestodes and profiled them against schistosome TRPM_PZQ and two cestode TRPM_PZQ representatives.¹⁰

Chart 1. Structures of the PZQ Analogs Studied in This Work.

Table 1 (columns 2–4) reproduces data from Andrews et al.¹ that scored the activity of PZQ derivatives against Schistosoma mansoni and Hymenolepis nana. Data were collected against S. mansoniin vitro and in vivo using a mouse model and also against H. nana using an in vivo mouse model.¹ Results were previously scored using broad potency ranges, graded “+++”, “++”, “+”, or “0” (see Table 1 legend). While there are caveats in the interpretation of these data, it is evident that the selected analogs displayed appreciable activity in the cestode model (column 4) but lower or minimal activity when tested against schistosomes (columns 2 and 3). In contrast, (±)-PZQ (1) showed equivalent activity in phenotypic grading across all of the bioassays, with preferential stereoselectivity toward the R enantiomer (Table 1).

Table 1. Comparison of Antischistosomal and Anticestodal Activities of PZQ Derivatives with Their Activities at the Corresponding TRPM_PZQ Orthologs^a.

	previous studies
	S. mansoni		H. nana	this work: EC50 (μM)
compound	in vitro	in vivo	in vivo	Sm.TRPMPZQ	Eg.TRPMPZQ	Mc.TRPMPZQ
1 ((±)-PZQ)	+++	+++	+++	0.65 ± 0.057	0.10 ± 0.009	0.12 ± 0.016
((R)-PZQ)b	+++	+++	+++	0.28 ± 0.03	0.05 ± 0.01	0.08 ± 0.003
((S)-PZQ)b	++	+	++	28 ± 2.8	0.78 ± 0.13	1.23 ± 0.17
(±)-2	+	+	+++	inactive	2.3 ± 1.1	5.3 ± 1.3
(R)-2	0	++	+++	>100	1.4 ± 0.11	0.98 ± 0.088
(S)-2	0	0	++	inactive	22 ± 2.3	16 ± 2.2
3	0	+	++	inactive	inactive	inactive
4	0	++	+++	inactive	>100	>100
5	0	++	+++	inactive	inactive	inactive
6	++	++	+++	12 ± 2.6	0.96 ± 0.14	0.80 ± 0.22
7	0	0	+	inactive	>100	>100
8	0	+	++	inactive	53 ± 16	65 ± 14

^aAll presented in vitro and in vivo data on organisms (columns 2–4) were taken from ref (1). For in vitro studies using S. mansoni, “+++” indicates a maximal effect ≤3.2 μM, “++” indicates a maximal effect ≤320 μM, “+” indicates a less than maximal effect at ≤320 μM, and “0” indicates no effect at ≤320 μM. For in vivo studies using S. mansoni, “+++” indicates a complete reduction of worms at 50 mg/kg dosing (×5), “++” indicates complete reduction of worms at 500 mg/kg dosing (×5), “+” indicates less than 90% worm reduction at 500 mg/kg dosing (×5), and “0” indicates no effect at 500 mg/kg dosing (×5). For in vivo studies using H. nana, “+++” indicates a complete reduction of worms at 25 mg/kg dosing (×1), “++” indicates complete reduction of worms at 500 mg/kg dosing (×1), “+” indicates less than 90% worm reduction at 500 mg/kg dosing (×1), and “0” indicates no effect at 500 mg/kg dosing (×1). Columns 5–7 tabulate the EC₅₀ values for analog activation of trematode and cestode TRPM_PZQ orthologs in vitro. Data are shown as mean ± SEM for n ≥ 3 independent transfections. Eg.TRPM_PZQ = Echinococcus granulosus TRPM_PZQ; Mc.TRPM_PZQ = Mesocestoides corti TRPM_PZQ.

^bEC₅₀ values on the channel are reported in ref (10) and are provided for completeness.

Based on Andrews’ data,¹ we resynthesized compounds 2–8. These analogs were then tested for activity at Sm.TRPM_PZQ and two representative cestode TRPM_PZQ orthologs (Echinococcus granulosus TRPM_PZQ (Eg.TRPM_PZQ) and Mesocestoides corti TRPM_PZQ (Mc.TRPM_PZQ)) that have been successfully heterologously expressed.¹⁰ This was done using a fluorescence-based reporter assay to measure changes in cytosolic Ca²⁺ in HEK293 cells transiently expressing the individual TRPM_PZQ ion channels. Results from these assays are presented in Figure 1, and all EC₅₀ values are tabulated in Table 1 (columns 5–7).

Figure 1.

Functional profiling of PZQ analogs against different TRPM_PZQ channels. Shown are concentration–response relationships for Sm.TRPM_PZQ (blue circles), Eg.TRPM_PZQ (red circles), and Mc.TRPM_PZQ (orange circles) in response to increasing concentrations of PZQ analogs. Responses to molecules in untransfected HEK293 cells are shown as controls (red diamonds).

Data for (±)-PZQ (1) are shown in Figure 1A. All three TRPM_PZQ channels, Sm.TRPM_PZQ (EC₅₀ = 645 ± 57 nM), Eg.TRPM_PZQ (EC₅₀ = 104 ± 9 nM), and Mc.TRPM_PZQ (EC₅₀= 112 ± 16 nM), were potently activated by 1. The cestode TRPM_PZQ channels displayed ∼6-fold higher sensitivity to PZQ, consistent with previous reports and the known sensitivity of many cestode species to PZQ.^6,7,10Figure 1B–D shows results for the 3-pyridyl PZQ analogs. The racemate, (±)-2, showed little activity at Sm.TRPM_PZQ but displayed low-micromolar potency at both cestode channels (EC₅₀ for Eg.TRPM_PZQ = 2.3 ± 1.1 μM, EC₅₀ for Mc.TRPM_PZQ = 5.3 ± 1.3 μM; Figure 1B). Consistent with the previously detailed activity of PZQ enantiomers in vivo(14) that is also mirrored at TRPM_PZQ⁶ (Table 1, entries 2 and 3), the R enantiomer (R)-2 was more potent than the S enantiomer (S)-2 at the cestode channels (Figure 1C vs Figure 1D). All activity at Sm.TRPM_PZQ was attributed to enantiomer (R)-2 (Figure 1C). Activation of both cestode channels, with negligible activity at the schistosome channel, was consistent with the Andrews et al. grading classification (Table 1).¹

A series of substituted phenyl derivatives were then profiled. In previous phenotypic assays, these analogs were active against H. nana and S. mansoniin vivo but were inactive against S. mansoniin vitro (Table 1).¹ Consistent with these prior observations, analogs 3–5 showed no activity at Sm.TRPM_PZQ (Figure 1E–G), and the activity of these analogs at cyclophyllidean cestode TRPM_PZQ orthologs was also low. As previously proposed,¹ the in vivo activity of compounds 3–5 is likely caused by dealkylative metabolism to aniline 6, and subsequent synthesis and profiling of 6 confirmed this (Figure 1H). Compound 6 activated Sm.TRPM_PZQ (EC₅₀ = 11.6 ± 2.6 μM), Eg.TRPM_PZQ (EC₅₀ = 957 ± 141 nM), and Mc.TRPM_PZQ (EC₅₀ = 796 ± 224 nM), and displayed an ∼8–10-fold increase in potency at cestode TRPM_PZQ compared to schistosome TRPM_PZQ. This was similar to the ∼6-fold increase in potency for (±)-PZQ (1) at cestode versus schistosome TRPM_PZQ.

Finally, we profiled modifications of the cyclohexyl group of PZQ. Compound 7, a 4′-aminocyclohexyl derivative, lacked activity at Sm.TRPM_PZQ and was only weakly active at cestode TRPM_PZQ orthologs (Figure 1I), corresponding to the weak in vivo activity against H. nana previously reported (Table 1).¹ Finally, cyclopropyl analog 8 activated both cestode TRPM_PZQ representatives at concentrations >10 μM (EC₅₀ = 53 ± 16 μM for Eg.TRPM_PZQ, EC₅₀ = 65 ± 14 μM for Mc.TRPM_PZQ; Figure 1J). Little activity was observed at Sm.TRPM_PZQ. These target-based data are again consistent with the phenotypic observations of Andrews et al. (Table 1).¹

Overall, from the profiled PZQ analogs, only a single analog (compound 6) was sufficiently active at Sm.TRPM_PZQ to derive an EC₅₀ value, while five analogs displayed activity at the cestode channels (Table 1). Analog 6 was not one of the eight PZQ analogs selected based on the differential potency between cestodes and schistosomes but was synthesized to explain the in vivo activity of the other analogs. Therefore, the different potencies of these analogs against cestodes and schistosomes, seen in phenotypic the data of Andrews et al.¹ 40 years ago, was mirrored by the same differential potency of these analogs in target-based assays at the different TRPM_PZQ channels. Some caveats are however appropriate.

First, the original data did not report the activity of the PZQ derivatives against cestodes ex vivo (in vitro), so in selecting these analogs, there was no direct comparator for the action of all of these analogs between schistosomes and cestodes. Therefore, two analogs (R)-2 and (S)-2 were tested on Echinococcus multilocularis protoscoleces for comparison with PZQ (Figure 2A–D). When compared with the vehicle control, treatment with (±)-PZQ (1 μM) caused a sustained contraction of the protoscoleces (Figure 2A,B). When protoscoleces were treated with the 3-pyridyl enantiomers (R)-2 (Figure 2C) and (S)-2 (Figure 2D), a similar contraction was observed, with (R)-2 being more potent than (S)-2. To determine IC₅₀ values, concentration–response curves were obtained at multiple time points (Figure 2E–G). Low concentrations of PZQ and the 3-pyridyl analogs stimulated motility at the early time points (Figure 2E). After 24 h, the IC₅₀ for (R)-2 was ∼3 μM, and the IC₅₀ for (S)-2 was ∼30 μM (Figure 2G). Prolonged treatment with (±)-PZQ proved toxic after 24 h, and therefore, a more realistic IC₅₀ value (∼100 nM) was calculated at 12 h postincubation (Figure 2F). This mirrors the EC₅₀ at a cestode TRPM_PZQ of 100 nM (Table 1). These data for activity against cestodes ex vivo are again consistent with the potencies of the molecules at cestode TRPM_PZQ (Table 1).

Figure 2.

E. multilocularis protoscoleces after treatment with (A) DMSO, (B) (±)-PZQ (1 μM), (C) (R)-2 (30 μM), and (D) (S)-2 (100 μM). (E–G) Concentration–response motility graphs of E. multilocularis protosoleces after treatment with (±)-PZQ (blue circles), (R)-2 (red circles), and (S)-2 (purple circles) after (E) 0.5 h, (F) 12 h, and (G) 24 h. Data are plotted as % motility vs vehicle (DMSO) control (mean ± SEM).

Second, the cestode TRPM_PZQ and motility assays derive from different cyclophyllidean cestodes (E. granulosus, M. corti, and E. multilocularis) than the model (H. nana) used by Andrews et al.¹ However, we note that the amino acid residues lining the PZQ binding pocket of TRPM_PZQ are identical across all cyclophyllidean cestode TRPM_PZQ orthologs examined to date, including H. nana TRPM_PZQ (Figure S1).¹⁰ This is consistent with the similarity of the functional data from Eg.TRPM_PZQ and Mc.TRPM_PZQ.

Considering the structures of the analogs profiled here, it is evident that the cestode TRPM_PZQ binding pocket is more tolerant to substitutions of the cyclohexyl moiety of PZQ than is the schistosome TRPM_PZQ binding pocket. Aniline analog 6 and pyridyl analog (R)-2 show submicromolar potency, and even the cyclopropyl analog 8, which displayed no activity at Sm.TRPM_PZQ at 100 μM, was clearly active at the cestode TRPM_PZQ orthologs, consistent with the differential activity seen by Andrews et al. (Table 1).¹

This increased tolerability to modifications of the cyclohexyl group of PZQ, a key part of the pharmacophore at Sm.TRPM_PZQ,⁷ may provide opportunity to accommodate other cyclohexane ring modifications—notably, more metabolically stable PZQ derivatives—within the cestode TRPM_PZQ binding pocket.^15,16 This could potentially enhance the in vivo efficacy of these analogs for treating cestode infections, and these data therefore highlight an opportunity to design drugs that selectively target cestode TRPM_PZQ. Whether the absolute potency of such analogs can be further improved over PZQ to yield better treatments for cestode species less sensitive to PZQ (e.g., noncyclophyllidean cestodes^10,17) or for cestode life cycle stages that are hard to treat will require further work and a better understanding of the molecular basis by which cestode-selective analogs engage the TRPM_PZQ binding pocket. Such understanding will be aided by the recent mapping of the PZQ binding pocket in TRPM_PZQ orthologs in different parasitic flatworms and a capacity to model these interactions.^7,10 Of likely relevance are two natural amino acid variants—a histidine residue in the S1 transmembrane helix and a serine residue in the S4/S5 linker—that are different between the PZQ binding pocket of trematode and cyclophyllidean cestode TRPM_PZQ (compare Figure 3A and Figure 3B).¹⁰ This natural variation within the TRPM_PZQ binding pocket provides a possible molecular explanation underpinning the differential SAR of the PZQ analogs. Natural variation in the binding pocket has previously been shown to render Fasciola spp. TRPM_PZQ insensitive to PZQ.^7,10

Figure 3.

(A) Predicted binding pose of (R)-PZQ (white) in Sm.TRPM_PZQ. (B) Visualization of (R)-2 (white) adjacent to residues showing variation in Eg.TRPM_PZQ, based on (A). Residues showing variation between the channels are highlighted in TM1 (Asn 1388/His 1231, teal) and TM4 (T1518/S1361, green).

The activity of (R)-2 at cestode TRPM_PZQversusSm.TRPM_PZQ is noteworthy in the context of this natural variation. The transmembrane helix 1 (S1) variation occurs at a position in close proximity to the pyridyl nitrogen (Figure 3B), and it is conceivable that there is an electrostatic interaction between the histidine residue in cestode TRPM_PZQ (e.g., H1231 in Eg.TRPM_PZQ) and the pyridine ring that is absent with the uncharged asparagine residue in schistosome TRPM_PZQ (e.g., N1388 in Sm.TRPM_PZQ). Interactions between PZQ and this S1 residue are important for PZQ activation of TRPM_PZQ across species.^7,10

In summary, functional profiling of various PZQ derivatives on parasitic flatworms and at their respective TRPM_PZQ orthologs shows that “the glove fits”: the SAR between different parasites and different parasite TRPM_PZQ orthologs matches well. These data provide additional support for TRPM_PZQ serving as the relevant therapeutic target of PZQ in parasitic flatworms.

Glossary

Abbreviations

PZQ

praziquantel

TRP

transient receptor potential

Sm.TRPM_PZQ

Schistosoma mansoni transient receptor potential praziquantel

S. mansoni

Schistosoma mansoni

H. nana

Hymenolepis nana

Eg.TRPM_PZQ

Echinococcus granulosus transient receptor potential praziquantel

Mc.TRPM_PZQ

Mesocestoides corti transient receptor potential praziquantel

E. multilocularis

Echinococcus multilocularis

E. granulosus

Echinococcus granulosus

M. corti

Mesocestoides corti

DMEM

Dulbecco’s modified Eagle’s medium

FBS

fetal bovine serum

FLIPR

fluorescence imaging plate reader

HBSS

Hanks’ balanced salt solution

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00350.

• Sequence alignment of the PZQ binding pocket residues in cyclophyllidean cestodes; materials and methods, synthetic procedures, characterization data, and ¹H/¹³C NMR spectra for all previously uncharacterized molecules (PDF)

Author Contributions

D.J.S. and J.L.H. performed chemical synthesis; S.-K.P. and D.J.S. performed pharmacological assays; M.K. performed E. multilocularis motility experiments, which were analyzed by M.K. and B.L.-S.; C.M.R. performed bioinformatic studies; D.M. provided chemicals to support syntheses and together with T.S. and J.S.M. supervised data evaluation and discussion; J.S.M. wrote the initial draft of the manuscript and supervised this project with B.L.-S; all of the authors worked on revisions and approved the final version of the manuscript.

This work was supported by National Institutes of Health (NIH) Grant R01-AI145871 (J.S.M.). D.J.S. acknowledges support from the NIH (T32 HL134643) and the MCW Cardiovascular Center’s A.O. Smith Fellowship Scholars Program. B.L.-S. acknowledges funding from the Swiss National Science Foundation (192072).

The authors declare no competing financial interest.