Antitrypanosomal and antileishmanial activity of prenyl-1,2,3-triazoles

A series of prenyl 1,2,3-triazoles were prepared from isoprenyl azides and different alkynes. Most of the compounds were active against T. cruzi and L. donovani.

Abstract

A series of prenyl 1,2,3-triazoles were prepared from isoprenyl azides and different alkynes. The dipolar cycloaddition reaction provided exclusively primary azide products as regioisomeric mixtures that were separated by column chromatography and then fully characterized. Most of the compounds displayed antiparasitic activity against Trypanosoma cruzi and Leishmania donovani. The most active compounds were assayed as potential TcCYP51 inhibitors.

Trypanosomatids are protist parasites which cause a series of devastating illnesses that affect millions of people worldwide. In the Americas, these kinetoplastids cause Chagas disease (Trypanosoma cruzi) and different forms of leishmaniasis (Leishmania spp.). These tropical diseases are considered neglected as they incisively reduce human potential, keeping people in poverty. They are especially harmful to vulnerable populations, particularly among children or immunosuppressed people.1

Neglect in the public sector arises when government agendas are pushed aside and the required resources of the already limited health budgets of most of the countries where these diseases are endemic are limited. The immediate consequence is an inadequate and inefficient chemotherapeutic arsenal to fight these diseases, which has remained virtually unchanged in the last half century. In addition, the appearance of strains resistant to commercial drugs presents an increasing threat.1

The main chemotherapy for the treatment of leishmaniasis is stibogluconate (Pentostam®), which is a highly toxic antimony derivative. Other drugs used when patients do not respond to antimony-derived drugs are amphotericin B, which has a relatively low toxicity but has a high cost, and pentamidine, which is even more toxic than those derived from antimony (Fig. 1) and must be supplied under controlled supervision. The only drugs available for the treatment of Chagas disease are nifurtimox and benznidazol (Fig. 1),2 which are effective for the treatment of the disease in its acute phase, with a parasitological cure in more than 80% of the treated patients. Their effectiveness varies depending on the geographical region, probably due to the difference in susceptibility of the strains of T. cruzi found in those areas. However, there is less than 20% chance of parasitological cure in patients who are in the chronic phase of the disease.3 Although these drugs have several side effects, they are better tolerated by children than by adults.

Fig. 1. Chemotherapies known for leishmaniasis and Chagas disease.

Multidisciplinary collaboration between different actors is the key to addressing the problem of neglected diseases. Governments, non-profit organizations or alliances, and the scientific community have been working together to establish a feasible drug development pipeline for diseases caused by parasites, including kinetoplastids.4–8

Isoprenes are very interesting structures that can be abundantly found in nature as well as in many bioactive compounds in which they are present. Isoprenyl 1,2,3-triazoles have shown promising and varied biological activities (Fig. 2) as antimicrobials,9 antibiofilms,10,11 antioxidants,9 and antiparasitics.12 Additionally they also displayed activity as geranylgeranyl transferase II,13 geranylgeranyl diphosphate synthase,14,15 farnesyltransferase, Ras and Rab geranylgeranyl transferase inhibitors. In the same way, they are used as precursors for the construction of analogues and metabolites, for example, FPP,16 glycolipids17 and pharmacological drugs.18

Fig. 2. Bioactive prenyl 1,2,3-triazoles.

Based on precedents where complex prenyl 1,2,3-triazolyl steroids12 have shown promising antitrypanosomal activities, we wanted to explore simplified versions of these compounds. Looking for new chemical entities with antitrypanosomal activity, we wanted to investigate if simple prenylated 1,2,3-triazoles could display antiparasitic activity. Therefore, we proposed a series of 1,2,3-triazoles that could be prepared by the CuAAC of prenyl azides and different alkynes, also looking to obtain isomerically pure compounds. In this current article, we report the synthesis of a collection of 1,2,3-triazoles to explore the structural requirements for activity. Data are reported for all the compounds against the causative organisms of visceral leishmaniasis and Chagas disease, and structure–activity relationships are discussed.

Synthesis

Prenyl azides, the key intermediates, have been previously prepared as mixtures of regioisomers.12,19 They interconvert rapidly at room temperature due to [3,3]-sigmatropic Winstein's rearrangement (Fig. 3).20,21

Fig. 3. [3,3]-Sigmatropic rearrangement of geranyl azide.

The required prenyl azides were prepared from the corresponding prenyl alcohols following Thompson's reaction (Fig. 4).22

Fig. 4. Prenyl azides synthesis.

The azides were obtained in 8 hours with an average yield of 85%. NMR studies confirmed that geranyl and farnesyl azides were a mixture of primary (E,Z) and tertiary isomers in a 5 : 3 : 1 ratio, respectively, while prenyl azide was in a primary : tertiary ratio of 9 : 1.

Once the intermediates were synthesized, Sharpless' conditions were used to prepare 1,4-disubstituted 1,2,3-triazoles, using CuSO₄ as the copper source and sodium ascorbate as the reductant.23 Five terminal alkynes were used as synthetic counterparts of three allylic azide mixtures. The dipolarophiles were selected by using cheminformatic tools considering three factors: Lipinski's rule of five,24 the hydrophobic nature, and the physicochemical profiles of active reported isoprenyl triazoles. Based on these factors, commercially available phenylacetylene, methyl propiolate, 1-pentyne, 1-heptyne and 1-decyne were selected. The combination of the three mixtures of prenyl azides and five alkynes allowed the efficient synthesis of 25 compounds with an average yield of 83% in an 8 hour reaction at room temperature (Table 1). In all cases, a mixture of regiomeric primary azides was obtained. The lack of reactivity of the tertiary azides in the 1,3-dipolar cycloaddition was not unexpected because it has been described by Fokin and Sharpless in 2005.25 They showed that, under dynamic equilibrium, only primary azides react, while the mixture interconverts to restore the initial composition.

Table 1. Synthesized isoprenyl-1,2,3-triazoles library.

R =	COOMe a		Phenyl a		Propyl a		Pentyl a		Octyl a
n = 0; prenyl	IT-1	83%	IT-2	93%	IT-3	74%	IT-4	83%	IT-5	80%
n = 1; Z-geranyl	IT-6 (1.0)	86%	IT-7 (1.0)	75%	IT-8 (1.0)	87%	IT-9 (1.0)	85%	IT-10 (1.0)	82%
E-Geranyl	IT-11 (2.0)		IT-12 (2.0)		IT-13 (2.0)		IT-14 (2.0)		IT-15 (2.0)
n = 2; Z,E-farnesyl	IT-16 (1.0)	72%	IT-17 (1.0)	81%	IT-18 (1.0)	83%	IT-19 (1.0)	89%	IT-20 (1.0)	69%
E,E-Farnesyl	IT-21 (1.9)		IT-22 (1.9)		IT-23 (1.9)		IT-24 (1.9)		IT-25 (1.9)

^aRegioisomer relationship is specified in parentheses.

After that, contrary to the literature,9–15 a careful chromatographic procedure allowed the separation of the regioisomeric mixture when it was present (IT-6 to IT-25). This approach becomes a powerful tool to develop new regioisomerically pure molecular entities with potential pharmacological action.

Chemically pure isomers of two standards were assigned unambiguously and reliably by nOe experiments. Analysis of the composition of the isomers by GC-MS reveals that geranyl triazoles have a Z/E relationship of 33 : 67, and farnesyl triazoles of 35 : 65. Compared with the Z/E ratio of the azide, there has not been a significant change in the regiomeric composition after the reaction. This might indicate that the dipolar cycloaddition has a lower relative rate than Winstein's rearrangement.

Cheminformatics analysis

In chemogenomics, cheminformatics tools and molecular similarity methods are frequently used. Analysis of the structural similarity, using the ChemMine Tools platform,26 of our chemical library reveals that it has a high degree of similarity, particularly in the area defined in Fig. 5.

Fig. 5. Analysis of molecular similarity of the chemical library. The high similarity area is delimited.

Additionally, inferential analysis of the potential mechanisms using in silico ADME-Tox gave substantial information for the feasible and pharmacotherapeutic use of the chemical library. On the one hand, the toxicology analysis conducted using the OSIRIS Property Explorer platform27 indicates that none of the products prepared are potentially mutagenic, irritant, teratogenic or toxic for sexual reproduction.

On the other hand, the most relevant parameters (molecular weight, polar surface area and log P) of the collection allow us to predict the potential of the compounds as leads and possible administration routes, and they complemented the structure–activity studies. Fig. 6 shows the physicochemical profile and distribution of the prepared compounds based on molecular weight, polar surface area and octanol/water partition coefficient.

Fig. 6. Distribution of physicochemical parameters of the chemical library.

None of the prepared compounds contain more than 5 hydrogen bond donors or more than 10 oxygen and/or nitrogen atoms, and all of them have molecular weights below 500 Da. Also, in all cases, the total polar surface area is less than 100 A³. These parameters augured good oral and passive absorption in the body and that the compounds may also potentially penetrate the blood–brain barrier. The Hammett equation and Hansch's principle allowed the use of descriptors like log P for lipophilicity, considering that for a homologous series of compounds, generally parabolic relationships between partition coefficients and biological activity are obtained. In our case, the distribution of the log P values of the 25 compounds fluctuates between 1 and 9.

Biological activity

The in vitro activity in two of the trypanosomatid pathogens, the etiological agents of Chagas disease (T. cruzi epimastigotes) and visceral leishmaniasis (L. donovani promastigotes), was studied. Table 2 displays the activities in both parasites and the physicochemical and structural parameters.

Table 2. Physicochemical parameters and activities of the synthesized structures.

Entry	Compound	Regiochemistry	R =	TPSA	Log P	MW	Volume	T. cruzi IC50 (μM)	L. donovani IC50 (μM)
1	IT-1	—	COOMe	57.01	1.30	195.22	182.968	>100	>100
2	IT-2	—	Ph	30.71	2.82	213.28	209.847	>100	>100
3	IT-3	—	C3H7	30.71	2.47	179.10	188.603	>100	>100
4	IT-4	—	C5H11	30.71	3.46	207.20	222.207	>100	>100
5	IT-5	—	C8H17	30.71	4.92	249.20	272.612	27	>100
6	IT-6	Z	COOMe	57.01	3.20	263.15	260.549	87	>100
7	IT-7	Z	Ph	30.71	4.91	281.20	287.428	39	49
8	IT-8	Z	C3H7	30.71	4.05	247.38	266.184	44	44
9	IT-9	Z	C5H11	30.71	5.45	275.20	299.788	36	40
10	IT-10	Z	C8H17	30.71	6.92	317.52	350.193	34	16
11	IT-11	E	COOMe	57.01	3.20	263.15	260.549	87	>100
12	IT-12	E	Ph	30.71	4.91	281.28	287.428	39	49
13	IT-13	E	C3H7	30.71	4.50	247.39	266.184	31	72
14	IT-14	E	C5H11	30.71	5.45	275.20	299.788	51	94
15	IT-15	E	C8H17	30.71	6.92	317.52	350.193	27	57
16	IT-16	Z,E	COOMe	57.01	5.25	331.15	338.130	15	11
17	IT-17	Z,E	Ph	30.71	6.95	349.28	365.010	32	49
18	IT-18	Z,E	C3H7	30.71	6.50	315.39	343.766	17	33
19	IT-19	Z,E	C5H11	30.71	7.50	343.20	377.369	18	11
20	IT-20	Z,E	C8H17	30.71	8.76	385.52	427.775	>100	52
21	IT-21	E,E	COOMe	57.01	5.25	331.15	338.130	9.0	31
22	IT-22	E,E	Ph	30.71	6.95	349.28	365.010	34	48
23	IT-23	E,E	C3H7	30.71	6.50	315.39	343.766	18	51
24	IT-24	E,E	C5H11	30.71	7.50	343.20	377.369	29	44
25	IT-25	E,E	C8H17	30.71	8.76	385.52	427.775	>100	36
26	Benznidazole	—	—	92.75	0.78	260.25	224.99	10	—
27	Pentamidine	—	—	118.22	1.49	340.43	324.60	—	6.2
28	Amphotericin B	—	—	319.61	–2.49	924.09	865.48	—	0.35

None of the prepared triazoles show cytotoxicity in VERO cells at a maximum concentration of 20 μM. From the analysis of the in vitro activity of the compounds in both parasites, those containing prenyl in their structure are inactive at concentrations lower than 100 μM in both parasites, excluding IT-5 (entry 5) in T. cruzi. This compound has the highest log P, PM and volume values in this subfamily. The triazoles that are decorated with a neryl group (Z) in their structures (entries 6 to 10) have moderate IC₅₀ values in both parasites, except for IT-10 which contains octyl (entry 10), with an IC₅₀ of 16 μM in L. donovani. This compound also has the highest values in terms of the physicochemical properties of this subfamily. In T. cruzi, these analogs were active when they lack oxygenated functional groups in their structure. However, there is no difference in their activities (between 34 and 44 μM).

When the triazoles contain a geranyl substituent (E; entries 11 to 15), a significant loss of antileishmanial activity in this subfamily is observed. Conversely, the largest compound, IT-15, is the best candidate in the family of monoterpenyl triazoles against T. cruzi, with an IC₅₀ of 27 μM. Interestingly, it appears that a change in the regiochemistry of the proximal double bond to the triazole ring modulates the activity towards one parasite, revealing the importance of chromatographic separation of the products.

Globally in monoterpenyl triazoles, if oxygenated groups (IT-6 and IT-11) are present, compounds lose their activity in T. cruzi and L. donovani.

The last two subfamilies are the compounds that have the best activities in both parasites. IT-16 (Z,E-farnesyl, R = COOMe) and IT-19 (Z,E-farnesyl, R = pentyl) are the best candidates as leishmanicidal agents with an IC₅₀ of 11 μM, which is double the IC₅₀ of pentamidine (entry 27), and are thirty times less active than amphotericin B (entry 28). In contrast, the best compound in the entire collection against T. cruzi is IT-21 (E,E-farnesyl, R = COOMe) with an IC₅₀ of 9 μM, being slightly more active than benznidazole (entry 26). The subtle change in the regiochemistry of the first isoprene unit governs the antiparasitic activity.

Compounds IT-20 and IT-25 are inactive in T. cruzi at concentrations lower than 100 μM and moderately active in L. donovani. These two analogs have the highest MW, volume and log P in the entire collection, suggesting that above a threshold of values (log P > 8.76, volume > 427 ų and PM > 385 Da), these structures begin to lose activity.

In conclusion, the two best compounds against L. donovani (IT-16 and IT-19) and the best compound against T. cruzi (IT-21) have similarities and comparable physicochemical parameters. The log P values of these three compounds are found to be between 5.25 and 7.50, the molecular weights between 331 and 343, and the volumes between 338 and 377 ų.

T. cruzi CYP51 inhibition assay

Molecules with volumes lower than 260 ų or greater than 400 ų lose activity in both parasites. Similarly, molecules with log P below 5 or above 8 result in inefficient trypanosomacidal activity (Fig. S1, ESI‡). We described this as an “island of activity” (IC₅₀ < 40 μM) delimited by these three selected physicochemical parameters.

In general, structures such as imidazoles or 1,2,4-triazoles present in fluconazole, ketoconazole or itraconazole (Fig. 7) may act as inhibitors of sterol 14α-demethylase (CYP51). CYP51 is the cytochrome P450 enzyme that is widely distributed throughout different biological kingdoms. It is found in animals, plants, fungi, yeasts, protozoa and bacteria28 and is considered the oldest member of the P450 superfamily.29 In all cases, CYP51 catalyzes the same three-step reaction that removes the 14α-methyl group from the cyclized sterol precursors, wherein each step requires an oxygen molecule and NADPH. During the first step of the CYP51 reaction, the sterol 14α-methyl group is converted into the alcohol and subsequently into the aldehyde, and is finally removed as formic acid. The CYP51 reaction is a required step upon the biosynthesis of sterols.

Fig. 7. A. Mechanism of CYP51. B. Known CYP51 inhibitors. C. Isoprenyltriazoles tested in TcCYP51.

Trypanosomatids are sensitive to CYP51's inhibitors because they synthesize ergosterol and cannot use host cholesterol for their cellular membranes. Blocking this enzyme also alters the structure of various organelles and decreases the total level of sterols in the parasite. Furthermore, accumulation of 14α-methylated sterols in trypanosomatids produces cytostatic and cytotoxic effects.30

The inhibitory effect of selected hits of the “island of activity” (IT-16, IT-18, IT-19, IT-21, IT-23 and IT-24) was determined on T. cruzi CYP51. The methodology employed involves the expression and purification of T. cruzi CYP51 and the use of cytochrome P450 reductase as an electron donor.31

The inhibitory potency of the compounds was expressed as the percentage of enzymatic activity of T. cruzi CYP51 in the presence of the tested compounds at a fixed concentration (100 μM, a 100-fold molar excess over the enzyme) at two time periods (5 and 60 minutes).

The results are shown in Table 3. Overall, the inhibitory effect is rather moderate, because, as we have shown previously, under these experimental conditions the most potent azoles completely abolish T. cruzi CYP51 activity at a 1 : 1 enzyme/inhibitor molar ratio.32 Amongst all tested analogs, the strongest effects were produced by IT-19 and IT-24. Interestingly, the inhibition is persistent over time, suggesting that the compounds might have a relatively low initial binding affinity to the enzyme but also a limited tendency to be replaced by the substrate during the CYP51 reaction.32

Table 3. Antiparasitic activity and CYP51 inhibition of selected analogs.

Compound	Antiparasitic activity		T. cruzi CYP51
	T. cruzi IC50 (μM)	L. donovani IC50 (μM)	5 min % in h	60 min % in h
IT-16	15	11	26	59
IT-21	9	31	35	63
IT-18	17	33	14	49
IT-23	18	51	27	56
IT-19	18	11	7	39
IT-24	29	44	9	33

Thus, there would be a contribution to their antiparasitic activity by inhibiting CYP51; these compounds must also possess one or more alternative targets. Having shown the inhibitory action of the compounds, the binding mode of these novel chemical entities was determined. The interaction between the CYP51 enzyme and its sterol substrate produces a blue shift in the Soret band (from 417 to 394 nm) caused by the expulsion of a water molecule from the coordination sphere of the iron in the heme. This spectral response (type 1) reflects the transition of the iron from the hexa-coordinated low-spin state to the penta-coordinated high-spin state. In contrast, the direct coordination of a basic atom to the heme iron causes a red shift in the Soret band (from 417 to 421–424 nm), called a type 2 response, which is characteristic of azoles (Fig. 8).

Fig. 8. Shift in the T. cruzi CYP51 Soret band in response to the binding of compound IT-19. Top – absolute spectra, bottom – type 1-like difference spectra upon titration. The CYP51 concentration was 2 μM, the IT-19 titration range 1–6 μM, and the titration step 1 μM. The apparent K_d = 7.6 ± 1.4 μM.

Interestingly, our analogs produced a blue shift in the Soret band maximum of the enzyme (a modified type 1 spectral response, Fig. 8), indicating that their binding mode is different from the binding mode of antifungal azoles. This means that they bind in the CYP51 active site, but do not coordinate directly to the heme iron. The shift in the Soret band suggests that the binding of IT-19 might be changing the position of the heme coordinated water molecule, pushing it away from the iron and thus weakening the Fe–O coordination bond.33

Analysis of the crystal structure of the CYP51 in complex with its substrate analog34 shows that it possesses three distinguishable regions: a hydrophobic arm (isoprenyl motif) that is immersed in the deepest portion of a single hydrophobic cavity, characteristic of the CYP51 family.35 This deep cavity allows the enzyme to interact efficiently with the substrate and to maintain its correct placement in the active site during the three steps of catalysis. In the steroid fraction, the oxygen atom is in an area near the substrate entrance channel and at a 3.5 Å distance from the main chain oxygen of methionine 358, with which it interacts through a hydrogen bridge. Finally, the 14α-methylenecyclopropyl group of the sterol is located perpendicularly to the heme plane and forms contacts with leucine 356 and threonine 295.

The binding mode of IT-19 might consist of the interaction of the isoprenyl chain with the hydrophobic cavity deep in the enzyme, the interaction of the triazole ring with the catalytic water molecule above the heme plane (the blue shift in the Soret band), and finally positioning of the alkyl chain along the substrate access channel.

Conclusions

The prenyl triazole motif is an interesting structure from which to prepare new bioactive molecular entities. Thus, we prepared a library of 25 compounds including a cheminformatics design. When a mixture of regioisomers was obtained, the products were carefully resolved by chromatography. The complete collection was tested against the etiological agents of Chagas disease (T. cruzi) and visceral leishmaniasis (L. donovani). An island of high activity was defined by means of their physicochemical parameters and similarity in structures. Compounds IT-16 and IT-19 are the best candidates as leishmanicidal agents and IT-21 as antichagasic chemotherapy. Finally, combining the results of biological activities with the physicochemical parameters allows us to propose the SAR detailed in Fig. 9. In addition, the inhibition of T. cruzi CYP51 by some of the most active analogs demonstrated that the compound binding to the enzyme in a different way compared to the azoles (modified type 1) might be a contributing factor to their antiprotozoan activity. Their lack of toxicity (predicted in silico and experimentally determined) and their structural and synthetic simplicity prompted us to continue working on this structure as a promising lead to trypanocidal drugs.

Fig. 9. Analysis of the structure/activity of the chemical library of isoprenyl triazoles.