Synthesis, Structure and Antileishmanial Evaluation of Endoperoxide–Pyrazole Hybrids

Leishmaniases are among the most impacting neglected tropical diseases. In attempts to repurpose antimalarial drugs or candidates, it was found that selected 1,2,4-trioxanes, 1,2,4,5-tetraoxanes, and pyrazole-containing chemotypes demonstrated activity against Leishmania parasites. This study reports the synthesis and structure of trioxolane–pyrazole (OZ1, OZ2) and tetraoxane–pyrazole (T1, T2) hybrids obtained from the reaction of 3(5)-aminopyrazole with endoperoxide-containing building blocks. Interestingly, only the endocyclic amine of 3(5)-aminopyrazole was found to act as nucleophile for amide coupling. However, the fate of the reaction was influenced by prototropic tautomerism of the pyrazole heterocycle, yielding 3- and 5-aminopyrazole containing hybrids which were characterized by different techniques, including X-ray crystallography. The compounds were evaluated for in vitro antileishmanial activity against promastigotes of L. tropica and L. infantum, and for cytotoxicity against THP-1 cells. Selected compounds were also evaluated against intramacrophage amastigote forms of L. infantum. Trioxolane–pyrazole hybrids OZ1 and OZ2 exhibited some activity against Leishmania promastigotes, while tetraoxane–pyrazole hybrids proved inactive, most likely due to solubility issues. Eight salt forms, specifically tosylate, mesylate, and hydrochloride salts, were then prepared to improve the solubility of the corresponding peroxide hybrids and were uniformly tested. Biological evaluations in promastigotes showed that the compound OZ1•HCl was the most active against both strains of Leishmania. Such finding was corroborated by the results obtained in assessments of the L. infantum amastigote susceptibility. It is noteworthy that the salt forms of the endoperoxide–pyrazole hybrids displayed a broader spectrum of action, showing activity in both strains of Leishmania. Our preliminary biological findings encourage further optimization of peroxide–pyrazole hybrids to identify a promising antileishmanial lead.


Synthesis and Structure Analysis
The novel 1,2,4-trioxolane-pyrazoles OZ1/OZ2 and 1,2,4,5-tetraoxane-pyrazoles T1/T2 were synthesized by adapting procedures described in the literature, and their complete structure elucidation was achieved by 1 H, 13 C{ 1 H}, 2D NMR, X-ray crystallography, and HRMS. The synthetic strategies adopted for both classes are depicted in Schemes 1 and 2. The detailed procedures for their preparation and characterization are described in the Materials and Methods section and Supporting Information (SI). The main difference between these syntheses resides in the method used to produce the endoperoxide pharmacophore. Formation of the 1,2,4-trioxolane ring relies on a Griesbaum coozonolysis, using adamantan-2-one O-methyloxime 1o, 4oxocyclohexanecarboxylate, and ozone to yield the expected 1,2,4-trioxolane intermediate 2o. The adamantan-2-one O-methyl-oxime 1o employed in this procedure was previously synthesized from the reaction of adamantan-2-one with O-methylhydroxylamine in pyridine and methanol (Scheme 1). The method described by Amado et al. [26] was followed for the preparation of the precursor 1,2,4,5-tetraoxane 2t, starting with the peroxyacetalization of ethyl 4-oxocyclohexanecarboxylate with 50% (w/w) aqueous hydrogen peroxide in acetonitrile in the presence of the silica sulfuric acid catalyst (SSA) to afford the crude dihydroperoxide (DHP, 1t). Cyclocondensation of crude DHP 1t with adamantan-2-one catalysed by SSA produced the desired 1,2,4,5-tetraoxane 2t (Scheme 2). Hydrolysis of trioxolane-ester 2o or tetraoxane-ester 2t with lithium hydroxide in THF/H2O afforded the corresponding carboxylic acids, trioxolane 3o or tetraoxane 3t (Schemes 1 and 2). Endoperoxide carboxylic acids (3o or 3t) were then reacted with 3aminopyrazole under an inert atmosphere in the presence of 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydrochloride (EDC•HCl) and hydroxybenzotriazole (HOBt), as coupling agents, and triethylamine, to yield a mixture of 1,2,4-trioxolane-pyrazole or 1,2,4,5-tetraoxane-pyrazole positional isomers (OZ1 and OZ2 or T1 and T2, Schemes 1 and 2). Due to conformational inversion, broad 13 C signals were observed for the dispiro rings in 1,2,4,5-tetraoxanes in contrast to the sharper signals observed for the relatively inflexible 1,2,4-trioxolanes. Generally, in the tetraoxane dispiro system, the 13 C{ 1 H} peaks expand due to conformational flipping between the dispiro-cyclohexyl rings, resulting in the broadening of the peaks and a reduction in the number of carbons associated with those respective rings' carbons. These observations are more frequent in six-membered The structures of both trioxolane and tetraoxane isomers were primarily elucidated by 1D and 2D Nuclear Magnetic Resonance (NMR), through 1 H− 1 H correlation spectroscopy (COSY), heteronuclear single-quantum coherence (HSQC), and heteronuclear multiplebond correlation (HMBC) NMR studies [Tables 1 and 2 and Figures S11-S37 (SI)]. C3 was also observed in isomer T2 ( Figure S33, SI). Notably, the data demonstrate the presence of the conjugate in its 3-aminopyrazole-linked tautomeric form, which is identifiable in the crystal phase. Theoretically, the frame of 3(5)-aminopyrazole is also susceptible to side-chain tautomerism, as it has a proton-exchangeable amine side chain [18,28], which can be observed in this situation. Due to conformational inversion, broad 13 C signals were observed for the dispiro rings in 1,2,4,5-tetraoxanes in contrast to the sharper signals observed for the relatively inflexible 1,2,4-trioxolanes. Generally, in the tetraoxane dispiro system, the 13 C{ 1 H} peaks expand due to conformational flipping between the dispiro-cyclohexyl rings, resulting in the broadening of the peaks and a reduction in the number of carbons associated with those respective rings' carbons. These observations are more frequent in six-membered ring endoperoxides, such as 1,2,4,5-tetraoxanes or 1,2,4-trioxanes [27]. Compared to the tetraoxane analogues, dispiro 1,2,4-trioxolanes exhibit much improved signal resolution and sharpness and a higher signal number in the non-aromatic region. As expected, a strong correlation coefficient was obtained in the 1 H− 1 H COSY and HMBC spectra for endoperoxide containing dispiro rings. Furthermore, the combination of the 1 H− 1 H COSY and HMBC spectra allowed for the complete identification of the pyrazole moiety's carbons, enabling a more precise identification of the multiple -NH protons. Regarding the pyrazole ring, in the 1 H− 1 H COSY spectrum, 3 J correlations between H4 with H5 were observed, together with two long-range couplings with one of the -NH proton (δ = 5.54 ppm), including a four-bond coupling ( 4 J) with H4 and a five-bond coupling ( 5 J) with H5 in both isomers OZ1/T1 ( Figures S14 and S25, SI). HMBC 3 J C-NH correlations between each -NH (δ = 5.19 and 5.54 ppm) with C4 demonstrated that both isomers OZ1 and T1 have two signals in C4, implying the presence of two non-equivalent -NH protons. No HMBC 3 J C-NH Molecules 2022, 27, 5401 5 of 23 correlation was observed between the carbonyl carbon and the -NH from the pyrazole ring, suggesting that the exocyclic amino group did not participate in the amide generation in isomers OZ1/T1 ( Figures S16 and S27, SI). These correlations indicate that a prototropic annular tautomerism [18,28] occurred during the coupling reaction, where isomers OZ1/T1 are the products from reaction of the peroxide carboxylic acids with the 5-aminopyrazole tautomer. Protons -NH are described as non-equivalent, implying a strong intramolecular hydrogen bond involving one of the exocyclic -NH protons.

X-ray Crystal Analysis
A successful preparation of the single crystals of both tetraoxane-pyrazole hybrids enabled us to further confirm the different isomers through X-ray crystallography.
The crystal of T1 used for the data collection turned out to be a non-merohedral twin corresponding to a 180° rotation around the (100) reciprocal lattice axis; the twin-law (1 0 0.860; 0 −1 0; 0 0 −1) was found using the TwinRotMax routine implemented in PLATON [29]. The proportion of the two components of the twin was found to be 59%:41% via refinement of the BASF parameter. In compound T2, examination of the residual electron density disclosed that the tetraoxane group is slightly disordered (3%) over two alternate chair conformations, which was considered in the final refinement. The same amount of disorder was observed in the analysis of the dataset collected at room temperature. Oak Ridge Thermal Ellipsoid Plots (ORTEP) of the isomer molecules of T1 and T2 are depicted in Figures 2 and 3 [30] puckering parameters Q: ~0.6 Å and pseudo-rotation angle θ < 2°. In the case of T1, the puckering amplitudes of these rings are similar to those of T2, and the conformation of the tetraoxane ring is also close to the ideal chair form, with θ =1.7(6)°. However, the cyclohexane ring in T1 features a signifi-

Isomer T1
Isomer T2 1 H, δ (ppm) 13  For isomers OZ2 and T2, a few 1 H− 1 H COSY correlations on the pyrazole ring are shown, with a 3 J correlation between H4 and H5 (Figures S19 and S31, SI). A five-bond coupling ( 5 J) correlation is observed between H5 and -NH 2 in isomer T2 ( Figure S31, SI). Regarding HMBC spectrum, a single signal 3 J C-NH in C3-NH 2 to C4 is observed, which demonstrates that both -NH protons are equivalent. A weak 2 J C-H correlation in C3-NH 2 to C3 was also observed in isomer T2 ( Figure S33, SI). Notably, the data demonstrate the presence of the conjugate in its 3-aminopyrazole-linked tautomeric form, which is identifiable in the crystal phase. Theoretically, the frame of 3(5)-aminopyrazole is also susceptible to side-chain tautomerism, as it has a proton-exchangeable amine side chain [18,28], which can be observed in this situation.

X-ray Crystal Analysis
A successful preparation of the single crystals of both tetraoxane-pyrazole hybrids enabled us to further confirm the different isomers through X-ray crystallography.
The crystal of T1 used for the data collection turned out to be a non-merohedral twin corresponding to a 180 • rotation around the (100) reciprocal lattice axis; the twinlaw (1 0 0.860; 0 −1 0; 0 0 −1) was found using the TwinRotMax routine implemented in PLATON [29]. The proportion of the two components of the twin was found to be 59%:41% via refinement of the BASF parameter. In compound T2, examination of the residual electron density disclosed that the tetraoxane group is slightly disordered (3%) over two alternate chair conformations, which was considered in the final refinement. The same amount of disorder was observed in the analysis of the dataset collected at room temperature. Oak Ridge Thermal Ellipsoid Plots (ORTEP) of the isomer molecules of T1 and T2 are depicted in Figures 2 and 3, respectively.    3,7 ]decan]-4-yl)-methanone (T2). Displacement ellipsoids are drawn at the 50% probability level. The tetraoxane group has minor disorder over two alternate chair conformations; for the sake of clarity, only the major conformation is shown.
In the crystal of T1, one of the two H atoms of the amino group forms a short intramolecular hydrogen bond with the bare O5 atom as acceptor ( Table 3). The other H atom establishes a longer intermolecular hydrogen bond with the bare N atom of the pyrazole ring of a neighbor molecule, this hydrogen bond linking the molecules in infinite chains propagating along the crystallographic c-axis ( Figure 4). In the crystal structure of T2, both H atoms of the amino group are involved in intermolecular hydrogen bonds; one of them is an N-H … N bond linking molecules across inversion centers in the unit cell; the other one is directed towards atom O5, as in the crystal of T1. In T2, the two hydrogen bonds form an infinite 2D network of molecules lying in the (100) plane ( Figure 5). In both crystals, a set of C-H … O and C-H … N short contacts can be spotted, as well as C-H … Cg interactions with the π electron cloud of the pyrazole ring, that contribute to the stabilization of the crystal structures.    3,7 ]decan]-4-yl)-methanone (T2). Displacement ellipsoids are drawn at the 50% probability level. The tetraoxane group has minor disorder over two alternate chair conformations; for the sake of clarity, only the major conformation is shown.
In the crystal of T1, one of the two H atoms of the amino group forms a short intramolecular hydrogen bond with the bare O5 atom as acceptor ( Table 3). The other H atom establishes a longer intermolecular hydrogen bond with the bare N atom of the pyrazole ring of a neighbor molecule, this hydrogen bond linking the molecules in infinite chains propagating along the crystallographic c-axis ( Figure 4). In the crystal structure of T2, both H atoms of the amino group are involved in intermolecular hydrogen bonds; one of them is an N-H … N bond linking molecules across inversion centers in the unit cell; the other one is directed towards atom O5, as in the crystal of T1. In T2, the two hydrogen bonds form an infinite 2D network of molecules lying in the (100) plane ( Figure 5). In both crystals, a set of C-H … O and C-H … N short contacts can be spotted, as well as C-H … Cg interactions with the π electron cloud of the pyrazole ring, that contribute to the stabilization of the crystal structures.  [30] puckering parameters Q:~0.6 Å and pseudo-rotation angle θ < 2 • . In the case of T1, the puckering amplitudes of these rings are similar to those of T2, and the conformation of the tetraoxane ring is also close to the ideal chair form, with θ = 1.7(6) • . However, the cyclohexane ring in T1 features a significant distortion from the ideal chair conformation, with θ = 11.3(10) • For both isomers, the substituent at C14 is at an equatorial position to the cyclohexane ring. The conformation of the adamantane substituent is close to the ideal boat-boat form of the forming eight-membered rings, with approximate C s symmetry, in the two crystals. The pyrazole ring is planar within experimental error in both isomers. The C=O group is not strictly coplanar with the pyrazole ring, with a more pronounced rotation around the single N1-C17 bond in T2, as shown by the torsion angles O5-C17-N1-N2 (I: −175.9(8) • ; II 167.90(12) • ).
In the crystal of T1, one of the two H atoms of the amino group forms a short intramolecular hydrogen bond with the bare O5 atom as acceptor ( Table 3). The other H atom establishes a longer intermolecular hydrogen bond with the bare N atom of the pyrazole ring of a neighbor molecule, this hydrogen bond linking the molecules in infinite chains propagating along the crystallographic c-axis ( Figure 4). In the crystal structure of T2, both H atoms of the amino group are involved in intermolecular hydrogen bonds; one of them is an N-H . . . N bond linking molecules across inversion centers in the unit cell; the other one is directed towards atom O5, as in the crystal of T1. In T2, the two hydrogen bonds form an infinite 2D network of molecules lying in the (100) plane ( Figure 5). In both crystals, a set of C-H . . . O and C-H . . . N short contacts can be spotted, as well as C-H . . . C g interactions with the π electron cloud of the pyrazole ring, that contribute to the stabilization of the crystal structures. Table 3. Hydrogen bonding information for the crystalline structures of isomers T1 and T2. Symmetry codes: a: x,y,−1+z; b: 1−x,2−y,1−z; c: 1−x,1/2+y,1/2−z.

In Vitro Susceptibility of Leishmania spp. to Endoperoxide-Pyrazole Hybrids
We undertook a preliminary in vitro evaluation of the antileishmanial activity of the endoperoxide-pyrazole hybrids OZ1, OZ2, T1, T2, and 3-aminopyrazole (PYR) against promastigote forms of the viscerotropic species L. infantum and the dermotropic species L. tropica. Axenic Leishmania promastigotes were used as a model for the initial compound screening, since they are a more straightforward and economical model due to their ease of culture development, requiring small amounts of the tested molecule [31]. The halfinhibitory concentration (IC 50 ) of each compound was determined upon 48 h incubation with each parasite strain ( Table 4). As control, the standard antileishmanial drug amphotericin B (AmB) was used. Compounds cytotoxicity was assessed (CC 50 ) on the monocytic THP-1 cell line upon exposure to each compound, and the ClogP and ClogS values were calculated (Table 4). Materials and methodologies applied in the biological evaluation of the compounds are described in the Section 3.

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 μm) than OZ2 (9.646 ± 2.744 μm), even slightly more pronounced than amphotericin B (6.463 ± 1.874 μm). On the other hand, both molecules

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 μm) than OZ2 (9.646 ± 2.744 μm), even slightly more pronounced than amphotericin B (6.463 ± 1.874 μm). On the other hand, both molecules

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 μm) than OZ2 (9.646 ± 2.744 μm), even slightly more

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 μm) than OZ2 (9.646 ± 2.744 μm), even slightly more

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 μm) than OZ2 (9.646 ± 2.744 μm), even slightly more

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000x amplification) in L. tropica promastigotes treated for 48 h with 300 μM of OZ1 and 200 μM of OZ2 as these concentrations were close to the IC50 values obtained for this species (219 ± 25 and 161 ± 19 μM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC50 than OZ2, compound OZ1 induced a greater re-−0.51 5 From the endoperoxide-pyrazole hybrids tested against L. tropica promastigotes, trioxolanes OZ1 and OZ2 displayed the lowest IC 50 values (219 ± 25 µM and 161 ± 19 µM, respectively). Regarding the assessments performed on L. infantum promastigotes, this strain only revealed susceptibility towards OZ2, which exhibited an IC 50 of 219 ± 44 µM, yet slightly higher than that obtained for the parent LC67 (2o) (123 ± 1 µM, Table 4). Since both ozonides have a similar value of ClogS, the lack of activity of OZ1 in L. infantum could be related to steric effects. Disparity in response to standard CL treatments or experimental drugs has become a significant burden to disease management and variations in treatment outcomes are ascribed to a number of factors, including specific characteristics and susceptibility among Leishmania species [32]; L. tropica and L. infantum are phylogenetically and genetically distinct, which justifies the appearance of different clinical forms as well as variable pharmacological susceptibility among species [33]. The calculated partition coefficient (ClogP) values proved very similar among the endoperoxides studied (1.79-2.61), suggesting that other physical properties could explain the differences in activity.
Overall, both ozonides demonstrated some antileishmanial activity, though lower than disclosed from previous in vitro studies where selected endoperoxides revealed antileishmanial potency at low micromolar (<10 µM) concentrations [5,34]. In the present study, the evaluation of compounds T1 and T2 was limited by their low solubility in M199 and RPMI medium. Calculated intrinsic aqueous solubility (ClogS) values show similar values between trioxolanes (1.20-1.38) and tetraoxanes (1.18-1.21). However, the poor solubility of dispiro 4 '-substituted tetraoxanes may be associated to their structural symmetry [35]. A decrease in molecular symmetry has been correlated to better physicochemical properties, such as absorption and solubility, resulting in enhanced systemic exposure to a drug. So it is anticipated that 3 '-substituted analogues could overcome this solubility issue [35,36]. Therefore, future optimization to 3"-substituted analogues should be considered. Likewise, conversion of the compounds into water-soluble salts, as well as encapsulation of the endoperoxide structures in delivery systems such as liposomes, could represent alternative strategies to solve the solubility issue [37,38].
Cytotoxicity tests revealed toxic effects of the hybrids OZ1 and OZ2 on THP-1 cells, with low selectivity (SI values of <5 µM). We observed that the parent 3aminopyrazole (PYR) is devoid of antileishmanial activity but exhibits similar toxicity to OZ2, indicating that the toxicity in OZ1 and OZ2 is likely due to the aminopyrazole moiety. Moreover, the efficacy of PYR in inhibiting THP-1 (202 ± 10 µM), a monocytic leukemia human-derived cell line, also indicates that it could be a promising framework in the design of compounds with antineoplastic potential.

Effects of Endoperoxide-Pyrazole Hybrids on L. tropica Promastigotes Morphology
To evaluate the effect promoted by endoperoxide-pyrazole hybrids on Leishmania morphology, which can impair its infection capacity, we performed a morphometric analysis with optical microscopy (1000× amplification) in L. tropica promastigotes treated for 48 h with 300 µM of OZ1 and 200 µM of OZ2 as these concentrations were close to the IC 50 values obtained for this species (219 ± 25 and 161 ± 19 µM, respectively). Figure 6 illustrates the measurements of L. tropica body (A) and flagellum (B) of parasites treated with both compounds and the control drug amphotericin B in comparison with non-treated parasites. Results show significant differences p < 0.001 for both compounds. Despite exhibiting a higher IC 50 than OZ2, compound OZ1 induced a greater reduction in body size (5.265 ± 1.733 µm) than OZ2 (9.646 ± 2.744 µm), even slightly more pronounced than amphotericin B (6.463 ± 1.874 µm). On the other hand, both molecules decreased the flagellum size, with effects within the same range for OZ2 (12.404 ± 3.745 µm) and OZ1 (13.561 ± 3.535 µm). In general, changes in flagellum size are less evident when compared to body size loss; nevertheless, they may limit mobility and its capacity in infecting cells.
Representative images of cellular alterations are depicted in Figure 7C-F. Non-treated cells and amphotericin B-treated cells were used as negative and positive controls, respectively ( Figure 7A,B).  Representative images of cellular alterations are depicted in Figure 7C-F. Nontreated cells and amphotericin B-treated cells were used as negative and positive controls, respectively ( Figure 7A,B).
After 48 h of incubation, non-treated promastigotes ( Figure 7A) exhibited a fusiformshaped body, well-defined intracellular organelles, and an elongated flagellum. Contrarywise, in cultures treated with OZ1 ( Figure 7C,D) and OZ2 ( Figure 7E,F), most promastigotes showed considerable damage, namely significant body size reduction with an abnormally rounded shape, cytoplasmic disintegration, apparent disruption of internal organelles, and cell shrinkage, which are the typical alterations observed upon treatment with established antileishmanial compounds [39]. Another parallel impact of these compounds is the flagellum length reduction as discussed previously, which may cause impairments in motility. As expected, all these constraints were also observed in promastigotes treated with amphotericin B ( Figure 7B).
Even though the IC50 values indicated that the ozonides have some activity, microscopic analysis revealed that L. tropica promastigotes experienced cell changes analogous to those caused by amphotericin B. These findings could support further investigations on the endoperoxide-pyrazole hybrids in the frame of leishmaniases chemotherapy, underlining the need for structure-activity studies and further optimisation.

Evaluation of the Hybrids' Stability upon Conversion to Their Salt Forms
The antileishmanial activity of the hybrids on promastigotes of L. tropica revealed the need for structural modifications to improve the antileishmanial effect of these compounds. Because solubility may play a significant role in the antileishmanial activity of this series of molecules, we converted the free amines attached to the pyrazole moiety into their salt forms. Several acids, namely p-toluenesulfonic, methanesulfonic, and hydrochlo- After 48 h of incubation, non-treated promastigotes ( Figure 7A) exhibited a fusiformshaped body, well-defined intracellular organelles, and an elongated flagellum. Contrarywise, in cultures treated with OZ1 ( Figure 7C,D) and OZ2 ( Figure 7E,F), most promastigotes showed considerable damage, namely significant body size reduction with an abnormally rounded shape, cytoplasmic disintegration, apparent disruption of internal organelles, and cell shrinkage, which are the typical alterations observed upon treatment with established antileishmanial compounds [39]. Another parallel impact of these compounds is the flagellum length reduction as discussed previously, which may cause impairments in motility. As expected, all these constraints were also observed in promastigotes treated with amphotericin B (Figure 7B).
Even though the IC 50 values indicated that the ozonides have some activity, microscopic analysis revealed that L. tropica promastigotes experienced cell changes analogous to those caused by amphotericin B. These findings could support further investigations on the endoperoxide-pyrazole hybrids in the frame of leishmaniases chemotherapy, underlining the need for structure-activity studies and further optimisation.

Evaluation of the Hybrids' Stability upon Conversion to Their Salt Forms
The antileishmanial activity of the hybrids on promastigotes of L. tropica revealed the need for structural modifications to improve the antileishmanial effect of these compounds. Because solubility may play a significant role in the antileishmanial activity of this series of molecules, we converted the free amines attached to the pyrazole moiety into their salt forms. Several acids, namely p-toluenesulfonic, methanesulfonic, and hydrochloric acid, can generate the free amines into salts with variable solubility levels. As such, the endoperoxide-pyrazole hybrids OZ1, OZ2, T1 and T2 were converted into their p-tosylate (OZ1•TsOH, OZ2•TsOH, and T1•TsOH), mesylate (OZ1•MsOH), and hydrochloride (OZ1•HCl, OZ2•HCl, T1•HCl, and T2•HCl) salts, with low to good yields (Scheme 3). Interestingly, this study has shown that the endoperoxide-pyrazole hybrids are stable even in strongly acidic conditions without decomposition of the endoperoxide ring, allowing further structure optimization. Tests of antileishmanial activity were conducted on promastigote forms of L. tropica and L. infantum, then also on intracellular amastigote forms of L. infantum to ascertain the effect of the endoperoxide-pyrazole hybrids in their salt forms. The results are depicted in Table 5. The different salts displayed different antileishmanial efficacy depending on the conjugate salt, with IC50 values ranging from 135-> 400 μM against L. tropica, and from Tests of antileishmanial activity were conducted on promastigote forms of L. tropica and L. infantum, then also on intracellular amastigote forms of L. infantum to ascertain the effect of the endoperoxide-pyrazole hybrids in their salt forms. The results are depicted in Table 5. The different salts displayed different antileishmanial efficacy depending on the conjugate salt, with IC 50 values ranging from 135-> 400 µM against L. tropica, and from 164-> 400 µM against L. infantum. Among the salt forms of endoperoxide-pyrazole hybrids tested against L. tropica promastigotes the trioxolane OZ1•HCl displayed an IC 50 value of 135 ± 36 µM, indicating that different salts of the same isomer exhibit divergent antileishmanial activities, with the OZ1•HCl salt proving to be the most active salt against L. tropica. Unlike for the results obtained for OZ1, OZ2, T1 and T2, the assessments performed on L. infantum promastigotes revealed susceptibility of this strain towards most of the different salts except for the T1 analog with TsOH and HCl. As observed for L. tropica, the trioxolane OZ1•HCl also displayed the lowest IC 50 value (164 ± 20 µM) against L. infantum. Cytotoxicity assessments revealed toxic effects of all the salt derivatives on THP-1 cells, with CC 50 values compressed in a concentration range of 331-529 µM. The improvements in solubility exhibited by the different salts enhanced the antileishmanial activity, although further structure-activity optimizations are still needed. However, it is worth noting that the salt forms of the endoperoxide-pyrazole hybrids display a broader spectrum of action, showing some effect on both strains of Leishmania. Table 5. Cytotoxic concentrations (CC 50 ) in THP-1 cell line, inhibitory concentrations (IC 50 ) in promastigote forms of L. infantum and L. tropica, and in intracellular amastigote forms of L. infantum and selective index (SI) of synthetic endoperoxide-pyrazole salts.

Compounds
Promastigotes Susceptibility IC 50 ± SEM a (µM) Cytotoxicity CC 50  Compounds that exhibited an IC 50 < 200 µM were also tested for intracellular Leishmania amastigotes susceptibility. The L. infantum strain was used as the model for this assay. According to the results of intracellular amastigote susceptibility (Tables 4 and 5), OZ1•HCl and IC0 (2t) demonstrated superior activity compared to the other compounds, with lower IC 50 values against intracellular amastigotes (87 ± 39 µM and 45 ± 28 µM) than in promastigotes (164 ± 20 µM and 128 ± 27 µM). In contrast, OZ1•TsOH, OZ1•MsOH and LC67 evidenced a diminished effect on amastigotes compared to that shown on promastigotes (Figure 8). Possible reasons could be linked to bioaccumulation and mechanism of action for this class of compounds, including passage through the host cell's cellular membrane or the formation of toxic metabolites toward Leishmania after metabolization of these compounds by host cell enzymes. As reported, the parasitophorous vacuole may serve as a barrier to direct interaction with the parasite, which was reported to eventually lead to a decrease in activity on moving from promastigote to amastigote stages [40]. Thus, it is noteworthy that OZ1•HCl and IC0 (2t) demonstrated superior activity compared to the other compounds, with slightly lower IC 50 values against intracellular amastigotes than in promastigotes. Overall, the compounds tested against the amastigote forms displayed poor selectivity since SI values are smaller than 10.

T2•HCl
266 ± 8 313 ± 40 444 ± 22 ND ND a SEM: Standard error of mean of at least three independent assays; b SEM: standard error of mean of two independent assays; c SI for amastigote forms; ND: not defined.

Chemicals
All reagents for synthesis were purchased from commercial sources and used without further purification. Analytical thin layer chromatography (TLC) was carried out using Merck (Darmstadt, Germany) TLC Silica gel 60 F254 aluminum sheets and visualized

Chemicals
All reagents for synthesis were purchased from commercial sources and used without further purification. Analytical thin layer chromatography (TLC) was carried out using Merck (Darmstadt, Germany) TLC Silica gel 60 F254 aluminum sheets and visualized under UV or by appropriate stain. P-Anisaldehyde and potassium permanganate were the most used. Column chromatography was carried out using Sigma Aldrich (Darmstadt, Germany) technical grade silica gel (pore size 60 Å, 230-400 mesh particle size, 40-63 µm particle size).

Equipment
1 H and 13 C Nuclear Magnetic Resonance (NMR) spectra were recorded using a 400 MHz Bruker (Billerica, MA, USA) instrument or using a 500 MHz JEOL system (Peabody, MA, USA) equipped with a Royal HFX probe in the deuterated solvents described in each experimental procedure. The chemical shifts (δ) are described in parts per million (ppm) downfield from an internal standard of tetramethylsilane (TMS). Melting points ( • C) were obtained on an SMP30 melting point apparatus and are uncorrected. High-Resolution Mass Spectrometry (HRMS) was recorded using the analytical service within the Centre of Marine Sciences (CCMAR, Algarve, Portugal) and was conducted on a Thermo Scientific High Resolution Mass Spectrometer (HRMS) (Waltham, MA, USA), model Orbitrap Elite, capable of MSn, n up to 10. X-ray diffraction data were collected on Bruker Kappa Apex II and D8 Quest diffractometers (Billerica, MA, USA) using graphite monochromated MoKα (λ = 0.71073 Å) radiation (see Section 3.4 for more details).
Safety. Organic peroxides are potentially hazardous compounds (inflammable and explosive) and must be handled carefully: (1) a safety shield should be used for all reactions involving H 2 O 2 ; (2) direct exposure to strong heat or light, mechanical shock, oxidizable organic materials, or transition-metal ions should be avoided.

Synthesis
General Procedure 1: Removal of the ester group. Followed procedure by Jiricek et al. with slight modifications [41]. To a solution of peroxide-ester (1 mmol) in THF:H 2 O (1:1, 5 mL), LiOH (3 mmol) was added, and the reaction mixture was stirred until completion, at room temperature. THF was removed in vacuo, diluted with water, and acidified with 1M HCl. The aqueous layer was extracted with EtOAc (3 × 30 mL), washed with brine reported [1]. 1 3,7 ]decane]-4-carboxylate (2t). Procedure adapted from Amado et al. [26]. Ethyl 4-oxocyclohexanecarboxylate (2 mL, 12.55 mmol) was dissolved in acetonitrile (38 mL), and silica sulfuric acid (SSA, 3.42 g, 2 mmol) was added to the mixture. Hydrogen peroxide 50 wt.% in H 2 O (4.09 mL, 4 mmol) was slowly added over an ice bath. Then the mixture was left to stir at room temperature until consumption of the starting material. To this mixture was added distilled water. Then the catalyst was filtered and rinsed with DCM. The filtrate was extracted with DCM (3 × 30 mL), dried over with MgSO 4 , and concentrated under reduced pressure at low temperature (30-35 • C) to obtain the gem-dihydroperoxide semi-crude, which was used immediately without further purification. The gem-dihydroperoxide semi-crude (1t) was dissolved in anhydrous DCM (75 mL) followed by addition of 2-adamantanone (5.66 g, 1.5 mmol). The mixture was cooled over an ice bath prior to addition of SSA (3.42 g, 2 mmol). The mixture was then warmed and left to stir at room temperature until consumption of the starting material. The resulting solution was then filtered, rinsed with DCM, and concentrated under reduced pressure. Purification through flash chromatography (EtOAc: n-hexane, 2.5:97.5, v/v) yielded a white solid (2.11 g, 48% yield). M.p. = 67-69 • C. Spectral data are in accordance with those reported [26]. 1

X-ray Crystallography
X-ray crystallographic studies were performed on single crystals of T1 and T2 at room temperature (292 K) for T2 and at low-temperature (150 K) for T1 and T2, using Mo Kα radiation (λ = 0.71073 Å). The room temperature data were collected in a Bruker Kappa Apex II diffractometer equipped with a 4K CCD detector; the low temperature data were measured in a Bruker D8 Quest diffractometer equipped with a PHOTON II CMOS detector and an Oxford Cryostream 700 N 2 flow cryostat. Both isomers crystallize in the monoclinic The structures of T1 and T2 were solved by direct methods using SHELXT-2018/2 [45]. Full-matrix least-squares refinements on F 2 were carried out with SHELXL-2018/3 [46] with anisotropic displacement parameters for all non-hydrogen atoms (CCDC 2151742 and 2151175 for detailed crystal data).
Hydrogen atoms were located on a difference Fourier synthesis, and their positions were refined as riding on parent atoms with an isotropic temperature constrained those of their parent atoms using SHELXL-2018/3 defaults [46], except those of the amino group that had their positions freely refined with an isotropic displacement factor constrained to 1.2× that of the parent N atom, for being involved in hydrogen bonding.

Biological Evaluation
All compounds were solubilized in dimethyl sulfoxide (DMSO, Merck, Darmstadt, Germany). Working solutions for biological studies were prepared with a maximum of 1% DMSO. Amphotericin B (Gibco) was used as control drug.
The parasite's viability was determined using 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) colorimetric test. At the end of incubation time, 20 µL per well of MTT was added (5 mg/mL in PBS), followed by 2-4 h incubation at 37 • C ± 1 • C. After a centrifugation at 1000× g (20 min. 0 • C), medium was removed, and formazan crystals were solubilized in DMSO; optical density at 595 nm was measured with Synergy TM HTX Multi Mode Microplate Reader (Dynex Technologies, Chantilly, VA, USA). GraphPad Prism V 9.0 was used to calculate IC 50 of each chemical by fitting the data as a non-linear regression with variable slope using a dose-response inhibitory model. At least three independent assays were performed.
To detect morphological alterations, at the end of the incubation period (48 h), 30 µL of the parasite's suspension were collected from different treatment concentrations of each compound to prepare smears on glass slides, fixed with methanol, and stained with 5% (v/v) Giemsa (Sigma). Stained slides were examined with bright field in an inverted microscope (Eclipse 80i Nikon, Tokyo, Japan) with 1000× amplification, and images were captured with a Nikon DS-Ri1 camera. Image J (version 1.53k) was used to measure promastigote body and flagellum size of at least 30 parasite samples.
Data were expressed as mean ± standard error of the mean (SEM), and statistical analysis was performed with one-way ANOVA (Bonferoni's multiple comparisons test) to determine whether differences between means relatively to untreated control were significant at different levels (p < 0.05, p < 0.01, p < 0.001).
Following the incubation period, 20 µL per well of MTT was added as described in Section 3.5.2. After absorbance readings, the CC 50 values of each compound were determined with GraphPad Prism V 9.0 (Dotmatics, San Diego, CA, USA) as previously described. At least three independent assays were performed.

Leishmania spp. Intracellular Amastigotes Susceptibility
The in vitro intracellular amastigote assay was performed using a THP1 cell line maintained in complete RPMI at 37 • C, 5% CO 2 . After 24 h differentiation of 5 × 10 5 cells/mL into macrophages in sterile 16-chamber LabTek slides (Thermo Scientific, Waltham, MA, USA) with 25 ng mL −1 phorbol myristic acid (PMA, Sigma), cells were washed once with warm RPMI to remove non-differentiated and non-adherent cells and further infected with 5 × 10 6 promastigotes/mL in a 10:1 parasite-macrophage ratio for another 24 h. After this period, slides were gently washed once with warm RPMI to remove non-internalised promastigotes. The compounds 2o, 2t, OZ1•HCl, OZ1•MsOH, and OZ1•TsOH at concentrations ranging 400 µM to 50 µM were added, and the slides were further incubated at 37 • C, 5% CO 2 for 72 h. After this period, cells were washed with warm PBS, fixed with methanol (Sigma) and stained with Giemsa 5% (Sigma). AmB was used as positive control, and macrophages cultivated in RPMI medium and DMSO (0.5% v/v) were used as negative control. The infection index was determined by multiplying the percentage of infected cells and the number of amastigotes per infected cell as previously described [47]. The results represent the average of two experiments. The IC 50 for each compound was calculated by fitting the data as a non-linear regression with variable slope using a dose-response inhibitory model in the GraphPad Prism V 8.0 (Dotmatics, San Diego, CA, USA) program.
Selectivity index (SI) was determined for each compound as the quotient between the cytotoxicity in cell lines (CC 50 ) and the inhibitory concentration (IC 50 ) (SI = CC 50 in mammalian cells/IC 50 L. infantum).

Conclusions
This work describes, for the first time, the synthesis and structure of 1,2,4-trioxolanepyrazole and 1,2,4,5-tetraoxane-pyrazole hybrids. Using a convergent synthetic approach, the hybrids were prepared by amide coupling of 3(5)-aminopyrazole with a carboxylic acid derivative of the endoperoxide-containing building block. The compounds were characterized in detail using 1D NMR ( 1 H and 13 C{ 1 H}), 2D NMR ( 1 H-1 H COSY, HSQC, and HMBC), X-ray crystallography, and HRMS-ESI + . The 1 H and 13 C{ 1 H} NMR spectroscopic and X-ray crystallography analyses enabled structural differentiation between 1,2,4-trioxolane and 1,2,4,5-tetraoxane isomers based on chemical shifts, signal sharpness, and number, whereas 1 H-1 H COSY and HMBC spectra allowed us to elucidate each positional isomer for both classes. X-ray crystallography unambiguously identified the structure of the two isomers and disclosed a distinct supramolecular organisation in the crystalline phase, despite both isomers crystallising in the same monoclinic space-group due to distinct hydrogen bonding patterns linking the molecules. Results reveal that only the endocyclic amine nitrogen of 3(5)-aminopyrazole participates in the amide coupling reactions with the carboxylic acidcontaining endoperoxide building blocks. The reaction affords products from reaction of both 3-and 5-amino pyrazole tautomers, revealing the impact of prototropic tautomerism in the pyrazole heterocycle on reactivity.
Preliminary studies were conducted to evaluate the antileishmanial activity of this novel chemotype. Compounds OZ1 and OZ2 displayed better activity against L. tropica than the T1 and T2 compounds, which was attributed to the poor solubility of the tetraoxane hybrids in M199 and RPMI medium. Regarding the assessments performed in L. infantum, the results demonstrated no activity of all compounds toward this strain. The 3-aminopyrazole moiety was tested against both strains to determine the role of this moiety in the antileishmanial activity of all compounds and was found inactive. These results suggest that the antileishmanial activity depends on its substituents and could mean that the incorporation of a pyrazole scaffold into the endoperoxide structure may have no synergistic effect. In the future, and upon optimization, a broad range of Leishmania species and other protozoans, namely other trypanosomatids, should be tested in order to better understand their susceptibility and evaluate the potential of the novel chemotype.
The endoperoxide-pyrazole hybrids were found to be very stable, as their salt forms could be successfully prepared without compromising the endoperoxide pharmacophore. This conversion to the salt form solved the insoluble nature of these compounds. Tests were conducted to ascertain the potential activity of these endoperoxide-pyrazole hybrids in their salt forms against promastigote and intracellular amastigote forms. Among all the salts tested against promastigotes of L. tropica and L. infantum, OZ1•HCl proved to be the best option. Our findings indicate that the conversion of the free amine in the respective salts not only improved the solubility but also amplified the spectrum of action of this chemotype.
The evaluation of L. infantum amastigotes susceptibility corroborates the findings on the antileishmanial activity in promastigotes in which OZ1•HCl was the compound with better activity of the new series. In addition, the compound with the lowest IC 50 against L. infantum amastigotes was identified as 2t, suggesting that the structure of these compounds must be simplified in order to enhance the effectiveness of new antileishmanial peroxides. Even though the issue of solubility has been addressed, it is not the primary factor influencing the antiparasitic effectiveness of this series. Considerable structure optimization was achieved with OZ1•HCl and 2t. However, the results of 1,2,4-trioxolanepyrazole and 1,2,4,5-tetraoxane-pyrazole hybrids highlight the need for further structure optimization to improve the activity of these compounds.