Glutathione Derivatives Active against Trypanosoma brucei rhodesiense and T.brucei brucei In Vitro

Diesters based on N-benzyloxycarbonyl-S-(2,4-dinitrophenyl)GSH (CBzGSDNP) containing linear alcohols 3 to 9, branched alcohols10 to 20, or heteroatom linear alcohols 21 to 25, were investigatedas in vitro inhibitors of pathogenic parasites. Diesters 3 to25 were better inhibitors of Trypanosoma brucei rhodesiensethan of T. brucei brucei and had low cytotoxicities. The mostactive compound had a 50% effective dose (ED₅₀) of 0.2 µM.A quantitative structure activity regression equation relatingthe log (1/ED₅₀) versus the hydrophobicity parameter (log P),Taft's steric parameter (E_s), molecular weight (MW), and theWienI descriptor (W) was determined, and the species differencewas found to be related to membrane penetration and steric effects.

INTRODUCTION

Top

Introduction

Trypanosomiasis and leishmaniasis are parasitic diseases thatcause severe infections in humans and domestic animals in thetropics. These infections pose a serious health problem to thecountries in tropical regions, in terms of the suffering theyinflict and the effects on their economies (22). Chemotherapyfor treatment of these diseases is inadequate, because manytreatments have poor clinical efficacy, produce side effects,or are toxic, especially in the late chronic stages, which inevitablyleads to death. Parasites display a variety of unique metabolicreactions not present in other microorganisms and eukaryotes.One general approach to the development of novel antiparasiticdrugs is to identify key differences in metabolism between thehost and pathogen and use them in the design of selective toxicagents. Thiol metabolism in the trypanosomatids (15) is characterizedby dependence on the hexapeptide, trypanothione [N¹, N⁸-bis-(glutathionyl)spermidine][T(SH)₂] (Fig. 1),an antioxidant that replaces glutathione (GSH),which is the major antioxidant of eukaryotic cells. The importanceof trypanothione as an antiprotozoal drug target is highlightedby the fact that existing trypanocidal drugs, notably the arsenicals(e.g., melarsoprol) (3) and difluoromethylornithine (23), maywork by interfering with the metabolism or synthesis of thishexapeptide. The central roles of trypanothione in trypanosomiasisand leishmaniasis make thiol-dependent enzymes potential targetsfor the development of chemotherapeutic drugs (15). Enzymesinvestigated to date include glutathionylspermidine synthetase,by using substrate analogues (6, 9), and trypanothione reductase,by using substrate analogues (12), subversive substrates (19),irreversible inhibitors (5), and phenothiazine derivatives (1).However, many of these compounds have proven to be inactiveagainst trypanosomes in vitro (6, 9). By using a lead-directedapproach to identify potential antiprotozoal compounds, we recentlyreported the identification of several GSH derivatives activein vitro against trypanosomiasis and leishmaniasis (10). A structure-functionstudy of S-bromobenzylglutathione derivatives identified theantiparasitic activity to be exclusively associated with N,S-blocked GSH diester derivatives (10), with the nature of theN and S groups contributing to the compound's activity and abilityto differentiate between the parasite and host (10). GSH diesters,due to their ease of membrane penetration and hydrolysis bynonspecific esterases to free acids, have been proposed as chemicaldelivery systems for transport of GSH into cells (16, 20). Thenature of the ester group in the case of Trypanosoma bruceibrucei control membrane penetration and susceptibility to hydrolysis(8) and the hydrophobicity parameter (log P) and Taft stericparameter (E_s) are an index of these two factors. To identifydifferences in the antiprotozoal activity of GSH diesters againstTrypanosoma brucei rhodesiense and T. brucei brucei, we haveundertaken a comparative study of the dependence of hydrophobicity(log P) and E_s on inhibitory activity.

View larger version (13K):
[in this window]
[in a new window]
FIG. 1. Structure of trypanothione [T(SH)₂].

MATERIALS AND METHODS

Top

Materials and Methods

GSH diester derivatives. Diester derivatives 3 to 25 (Fig. 2)were prepared as previouslydescribed (8, 10) with either S-(2,4-dinitrophenyl) GSH (GSDNP)or N-benzyloxycarbonyl-S-(2,4-dinitrophenyl) GSH (CbzGSDNP)as the starting material suspended in the appropriate alcohol,to which was added 2 to 3 mol eq of thionyl chloride.

View larger version (28K):
[in this window]
[in a new window]
FIG. 2. Structures of GSH derivatives.

QSAR. For quantitative structure activity regression (QSAR), SciQSAR,a module of Alchemy 2000 (Tripos), was used for data analysis.This program searched Alchemy 2000 for low-energy molecularconformations and used the information to calculate severalmolecular descriptors (MW, W, D, etc). The WienI (W) descriptoris a topological parameter (21) whose value is larger for extendedmolecules than compact ones. The log P (log K_OW) descriptorwas calculated for each of the molecules (8) by using the interactivedemo program established at the Environmental Service Centre,Syracuse Research Corporation, in which structures were enteredin SMILES notation (limited to 100 characters). The programestimated the octagonal/water partition coefficient (log P)by using an atom/fragment contribution method. The Taft stericparameter (E_s), is based on values determined experimentallyfor a series of menthol esters, of which values were reportedfor 13 of the 17 esters used, while the remainder were calculatedby regression analysis or extrapolation (8).

The creation of a regression equation by using these descriptorsand their statistical regression analysis was undertaken byusing the LINEST curve-fitting routine provided with Excel (MicrosoftCorporation). The program calculated the coefficients of theequation: the multiple correlation coefficient (R²) and Fisherstatistic value (F).

Evaluation of the parasitic activity of GSH derivatives in vitro: parasites. T. brucei brucei (S427) and T. brucei rhodesiense (STIB900)bloodstream-form trypomastigotes were maintained in HMI-18 medium(11, 13) with 20% heat-inactivated fetal calf serum (HIFCS)(Harlan Sera-Lab., Crawley, United Kingdom) at 37°C in a5% CO₂-air mixture.

All compounds were tested in triplicate in a threefold dilutionseries from a top concentration of 30 µM. Parasites werediluted to 2 x 10⁵/ml and added in equal volumes to the testcompounds in 96-well, flat-bottom Microtest III tissue cultureplates (Becton Dickinson and Company, Paramus, N.J.). Appropriatecontrols with pentamidine isethionate (Aventis, Sussex, UnitedKingdom) as the positive control (50% effective dose [ED₅₀],0.33 nM) were set up in parallel. Plates were maintained for3 days at 37°C in a 5% CO₂-air mixture. Compound activitywas determined on day 3 with the Alamar blue assay (18) forT. brucei rhodesiense and a tetrazolium salt colorimetric assay(11) for T. brucei brucei.

Evaluation of cytotoxicity of GSH derivatives in vitro. Cytotoxicity testing with KB cells (human oral pharyngeal carcinomaline) was evaluated with the Alamar blue assay (18). Cell cultureswere grown in RPMI medium with 10% HIFCS at 37°C in a 5%CO₂-air mixture in a humidified incubator. Plates were incubatedwith compound for 3 days at 37°C in a 5% CO₂-humidifiedair mixture prior to determination of activity.

K562 (human chronic myelogenous leukemia line) and A2780 (humanovarian carcinoma line) cytotoxicity studies were undertakenby the Christie Patterson Institute for Cancer Research, Manchester,United Kingdom. The cytotoxicity of drugs was evaluated by usingthe semiautomated MMT assay developed by the National CancerInstitute, Frederick, Md., based on the original work of Mosmann(17). Cell cultures were grown in RPMI medium with 10% FCS at37°C in a 5% CO₂-air mixture in a humidified incubator.Plates were incubated with compound for 5 days at 37°C ina 5% CO₂-humidified air mixture prior to determination of activityby the use of a tetrazolium salt colorimetric assay (11) onday 5.

RESULTS AND DISCUSSION

Top

Results and Discussion

The diesters investigated in this study were based on N-benzyloxycarbonyl-S-2,4-dinitrophenylglutathione(CbzGSDNP) and were prepared with either linear alcohols 3 to9, branched alcohols 10 to 20, or heteroatom (-F, -Cl, -OMe)-containinglinear alcohols 21 to 25.

Diester compounds 5 to 7, 10 to 19, 23, and 24 (Table 1) showedsignificant inhibitory activity against T. brucei rhodesiense(>1 µM). Of the linear GSH diester derivatives (Table1), compound 6, the butyl diester (log P, $~$ 5.06), proved themost active against T. brucei rhodesiense, with an ED₅₀ of 0.42µM. The ED₅₀ against T. brucei brucei was 1.07 µM.In the same series, the most active compound against T. bruceibrucei proved to be compound 7, the pentyl diester (log P =6.04), with an ED₅₀ of 0.38 µM, while for the heteroatom-substituteddiesters 21 to 25, the most active linear diester against T.brucei rhodesiense (ED₅₀, $~$ 0.47 µM) was the chlorobutyldiester 24, as well as the chloroethyl diester 22 for T. bruceibrucei. However, the toxicity of these electrophilic substituteddiesters to KB cells was high, as seen by the small relativetoxicity values that showed less than unity (Table 1) observedfor compounds 23 to 25. In the branched series, compounds 10to 19 displayed the highest activity against T. brucei rhodesiense—between0.18 and 0.65 µM. The antiparasitic activity of diestercompounds 3 to 25 was, in most cases, better for T. brucei rhodesiensethan for T. brucei brucei. The exceptions were compound 22,which displayed a 16-fold-higher activity against T. bruceibrucei than T. brucei rhodesiense, and compound 10, which displayeda 55-fold-higher activity against T. brucei rhodesiense thanT. brucei brucei (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Activity of linear, branched, and heteroatom GSH diester derivatives against T. brucei rhodesiense and T. brucei brucei^a

A plot of log (1/ED₅₀) versus log P with the data in Table 1indicated a parabolic dependence between these parameters forT. brucei rhodesiense based on regression analysis, with anoptimum value of log P in the range 4.5 to 5.5 compared to 5.0to 6.0 for T. brucei brucei (8). This value is in agreementwith those for compounds possessing good membrane penetration,with log P values in the range $~$ 4 to 7 (2, 4). The optimum valuefor log P (log P^o) of 5.03 determined from the parabolic curvewas lower than the value of 5.8 determined for T. brucei brucei.Similarly, a plot of log (1/ED₅₀) versus E_s showed optimum valuesfor E_s of -0.84 for T. brucei rhodesiense and -0.70 for T. bruceibrucei. This result indicates that some degree of branchingon the diester is beneficial to membrane penetration, but beyondthe optimum value (E_s = -0.7 or -0.84), the benefits decrease.Examination of the E_s values reported in Table 1 indicates thatbranched diesters with values in the optimum range (compounds10, 17, and 18), are branched on the carbon directly attachedto the alcohol.

The parabolic dependence of log (1/ED₅₀) versus log P and E_sfor T. brucei rhodesiense as found for T. brucei brucei indicatesthat, in both cases, the activity of these GSH diester derivativesis primarily controlled by their ability to enter parasiticcells, as previously suggested (8). The best QSAR equation derived,using a set of 17 compounds, excluding the outliers 11, 14,and 15, was equation 1 with the multiple correlation coefficientR² = 0.86 and Fisher statistic value F = 14.0:

(1)
Equation 1 indicates that the therapeutic activityof these compounds decreases with increasing values of log Pand E_s, the latter a result of the overall negative characterof these parameters. Log P is positive in character (Table 1),but its coefficient is negative (see equation 1), while E_s valuesare predominantly negative (Table 1). These two parameters renderthe overall equation negative in character and so decrease thepositive value of log (1/ED₅₀). For high therapeutic activity,log P, E_s, and their corresponding coefficients have to be small.Comparison of equation 1 with that identified for T. bruceibrucei (equation 2) (8) shows that log P^o, E_s, and their associatedcoefficients make an overall smaller negative contribution tothe equation, thus resulting in a higher degree of therapeuticactivity for these compounds against T. brucei rhodesiense thanagainst T. brucei brucei, as observed (Table 1):

(2)
T. brucei brucei has been used as model systemfor testing of drugs against African sleeping sickness, withsome variability in sensitivity between different T. bruceibrucei strains due to drug resistance (14). The differencesin inhibitory activity observed between T. brucei rhodesienseand T. brucei brucei in this study can be accounted for fromthe QSAR analysis by differences in mechanism (drug penetrationor de-esterification) of these prodrugs, which may reflect differencesin coat and lipid proteins among species rather than biochemistry.In view of these results, we have altered our testing regimeand now use T. brucei rhodesiense instead of T. brucei bruceito test these compounds.

The pharmacological values of the compounds identified in thisstudy, 6, 7, 10 to 18, and 24, are dependent on their specificity.The relative toxicity values given in Table 1, expressed asthe ratio of the activity of KB cells to that of T. brucei rhodesiensecells, are a suitable indicator of this: 7 < 16 < 10 <12 < 14 < 11. The significant activity of several of thesecompounds combined with their relatively low toxicity to KBcells indicates the potential of GSH derivatives for use inthe treatment of trypanosomiasis infections.

GSH diesters have been previously proposed as therapeutic agentsfor use in cancer chemotherapy targeted against the glyoxalasesystem, with branched diesters showing the highest activityagainst HL60 cells (median growth-inhibitory concentration [GC₅₀] ${approx}$ 4.2 µM) (20). The toxicity data presented in Table 1indicate that linear diesters 23 to 25 are the most active compoundsagainst KB cells (ED₅₀ ${approx}$ 0.38 to 0.70 µM), while both lineardiesters 7 and 24 and branched diesters 12, 16, 18, and 19 arethe most active against A2780 cells (ED₅₀ ${approx}$ 1.5 to 5.0 µM),and only the branched diester 13 is significantly active againstK562 cells (ED₅₀ ${approx}$ 4.8 µM). In the case of compounds 23and 24, the nature of the diester groups appears to contributedirectly to the high toxicity of these compounds against KBcells. The electrophilic nature of the ester groups suggeststhat these GSH diesters either function as alkylating agentsor act as carriers (prodrugs) of these toxic reagents into cells,prior to cleavage and release by nonspecific esterases. Thelatter explanation seems the most plausible, based on a systematicstudy of some 90 GSH derivatives (7), of which only 4 were foundwith comparable activity against KB cells, and their toxicitycould be associated with the nature of the groups attached tothe S-site of GSH. This result indicates an alternative strategyfor the design of anticancer compounds based on GSH with highin vitro activities.

ACKNOWLEDGMENTS

We thank J. A. Ballantine, D. E. Games, and B. Stein of theEPSRC Mass Spectrometry Service Centre, Swansea, United Kingdom,for FABMS, HRFABMS, and ESIMS measurements. H. Kendrick andS. L. Croft of the LSHTM and the WHO are thanked for antiprotozoaltesting of compounds against T. brucei rhodesiense, and theChristie Patterson Institute for Cancer Research, Manchester,United Kingdom, is thanked for toxicity testing of compounds.

FOOTNOTES

* Corresponding author. Mailing address: Department of Chemistry & Materials, The Manchester Metropolitan University, Faculty of Science and Engineering, John Dalton Building, Chester St., Manchester M1 5GD, United Kingdom. Phone: 00 44 161 2471416. Fax: 00 44 161 2476357. E-mail: C.DSilva{at}mmu.ac.uk.

REFERENCES

Top