Characterisation of melarsen-resistant Trypanosoma brucei brucei with respect to cross-resistance to other drugs and trypanothione metabolism

Molecular and Biochemical Parasitology, 53 (1992) 213-222 © 1992 Elsevier Science Publishers B.V. All fights reserved. / 0166-6851/92/$05.00

213

MOLBIO 01766

Characterisation of melarsen-resistant Trypanosoma brucei brucei with respect to cross-resistance to other drugs and trypanothione metabolism Alan H. Fairlamb, Nicola S. Carter, M a r k C u n n i n g h a m and Keith Smith Department of Medical Parasitology, London School of Hygiene and Tropical Medicine, London, UK (Received 14 January 1992; accepted 25 February 1992)

An arsenical resistant cloned line of Trypanosoma brucei brucei was derived from a parent sensitive clone by repeated selection in vivo with the pentavalent melaminophenyi arsenical, sodium melarsen. The melarsen-resistant line was tested in vivo in mice against a range of trypanocidal compounds and found to be cross-resistant to the trivalent arsenicals, melarsen oxide, melarsoprol and trimelarsen (33, 67 and 122-fold, respectively). A similar pattern of cross-resistance was found in vitro using a spectrophotometric lysis assay (> 200-fold resistance to melarsen oxide and > 20-fold resistance to both trimelarsen and melarsoprol). Both lines were equally sensitive to lysis by the lipophilic analogue phenylarsine oxide in vitro, suggesting that the melamine moiety is involved in the resistance mechanism. Although trypanothione has been reported to be the primary target for trivalent arsenical drugs [I ], levels of trypanothione and glutathione were not significantly different between the resistant and sensitive lines. Statistically significant differences were found in the levels of trypanothione reductase (50% lower in the resistant clone) and dihydrolipoamide dehydrogenase (38% higher in the resistant clone). However, the Km for trypanothione disulphide, the K, for the competitive inhibitor Mel T (the melarsen oxide adduct with trypanothione) and the pseudo-first order inactivation rates with melarsen oxide were the same for trypanothione reductase purified from both clones. The melarsen-resistant line also showed varying degrees of cross-resistance to the diamidines: stilbamidine (38-fold), berenil (31.5-fold), propamidine (5.7-fold) and pentamidine (l.5-fold). Cross-resistance correlates with the maximum interatomic distance between the amidine groups of these drugs and suggests that the diamidines and melaminophenyl arsenicals are recognised by the same transport system. Key words: Arsenical drug; Trypanothione reductase; Diamidine; Drug resistance; Drug effects; Trypanosoma brucei brucei

Introduction

The trivalent arsenical melarsoprol (Mel B, Arsobal) is still the main drug used for the treatment of late stage sleeping sickness caused by Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense. The mode of action of the melaminophenyl arsenicals remains unknown, but a primary intracellular target appears to be adduct formation [1] with the Correspondence address: Alan H. Fairlamb, Department of Medical Parasitology, London School of Hygiene and Tropical Medicine, Keppel Street, London WCIE 7HT, UK. Abbreviations: Mel T, adduct between melarsen oxide and dihydrotrypanothione.

unique dithiol, dihydrotrypanothione (NI,N 8bis(glutathionyl)spermidine) [2]. The stable adduct (Mel T) is also a potent competitive inhibitor of trypanothione reductase (EC 1.6.4.8) [1], a flavoprotein disulphide oxidoreductase that together with dihydrotrypanothione is essential for regulation of the correct intracellular redox balance and in defence against oxidant stress (see reviews in refs. 39). In order to gain further insight into the mode of action of these arsenical drugs, we have generated a melarsen-resistant cloned line from an arsenical sensitive line and report here its cross-resistance patterns to other trypanocidal compounds as well as some of its biochemical properties with particular reference to trypanothione metabolism.

214

Materials and Methods

Organisms. Bloodstream forms of an arsenical sensitive clone of T. brucei ($427 c I 18) and a sodium melarsen-resistant clone derived from this organism were obtained from the blood of Sprague-Dawley rats (200-440 g) previously infected with 10 7 organisms and purified by chromatography on DE-52 cellulose [10]. An arsenical resistant line was generated from the original clone by serial passage in mice exposed to progressively increasing sub-curative doses of sodium melarsen until cures could not be obtained with the maximum tolerated dose (160 mg kg-l). Sub-passage was continued until the growth rate of the parasites was completely unaffected by exposure to this level of drug. The melarsen-resistant line obtained was then cloned by infecting cyclophosphamide-treated mice with single trypanosomes [11] and a single clone designated $427 cRUI5 was selected for further study. Both clones were stored at - 7 0 ° C as stabilates in 10% (v/v) glycerol. Drug sensitivity assays in vivo. Outbred Tyler's Original (TO) mice (25-35 g) were infected with 104 organisms by intraperitoneal injection and 24 h later treated in groups of 5 with varying concentrations of drug (6 or more doubling dilutions). Compounds were administered in distilled water except for melarsoprol, melarsen oxide, allopurinol and nifurtimox, which were administered as suspensions in peanut oil. Unless stated otherwise, animals received a single dose of drug. Animals were inspected daily and the number of survivors recorded. The mean survival time for untreated mice was 5.56 + 0.18 (N = 45) and 5.22 + 0.16 (N = 45) days for the sensitive and resistant clones, respectively. The longest time to death following relapse was 19 days; mice surviving for >30 days were considered cured. The effective dose curing 50% of animals (EDs0) was determined using the programme DOSE (Elsevier/Biosoft). Resistance factors are expressed as the ratio of EDsos between the resistant and sensitive strains.

Drug sensitivity assays in vitro. The technique used is an adaptation of an earlier method, where cell lysis is monitored by the decrease in absorbance at 750 nm due to light scatter [12]. Purified trypanosomes were centrifuged (1000 x g, 10 min) and resuspended at 1 x 108 ml at 4°C in the following medium: 25 mM Hepes/ 120 mM NaCI/5.4 mM KC1/0.55 mM CaC12/ 0.4 mM MGSO4/5.6 mM Na2HPO4/II.1 mM glucose/5 mg 1-l phenol red, pH adjusted to 7.4 with NaOH; final osmolarity 300 mosmol kg -1 (CBSS). A small aliquot of cells was warmed to 37°C and samples added to cuvettes containing CBSS also at 37~C (I-2 x 10 7 ml-I; final volume 1 ml) and absorbance at 750 nm monitored in a thermostatted Beckman DU-70 single beam spectrophotometer fitted with an automatic 6-cell sample changer. Arsenicals (dissolved in water or dimethylformamide) were added at time zero and absorbance monitored for 30--60 min. Control samples received an equivalent amount of solvent. Biochemical assays. Trypanothione, glutathionyispermidine and glutathione were determined by ion-paired reverse phase HPLC following derivatisation with the fluorescent sulphydryl-reactive agent monobromobimane [13]. A Beckman System Gold HPLC equipped with binary solvent delivery system, autosampler, UV-167 detector coupled to a Gilson Fluoromonitor was used for the analyses. Trypanothione reductase and lipoamide dehydrogenase were assayed by standard procedures [14,15] on trypanosome extracts prepared by ultrasonication in 50 mM (K ") phosphate buffer, 2 mM EDTA, pH 7.0. Trypanothione reductase was partially purified >200-fold by ammonium sulphate precipitation and affinity chromatography on 2'5'ADP Sepharose from both clones and kinetic and inhibition constants determined exactly as previously described [1]. Inactivation kinetics with melarsen oxide were determined by preincubating aliquots of trypanothione reductase (25 mU ml- i) with 200 #M NADPH for 3 min prior to adding 50 /~M melarsen oxide. Residual activity was determined at intervals

215

by the addition of 100 pM trypanothione disulphide. Kinetic data was fitted to the Michaelis-Menten equation using simple weighting or in the case of inactivation by melarsen oxide to a single exponential decay using the non-linear regression data analysis programme E N Z F I T T E R (Elsevier/Biosoft) written by R.J. Leatherbarrow. Oxygen consumption was measured using a Clark-type oxygen electrode essentially as described previously [16].

Synthesis of Mel T. Dihydrotrypanothione was prepared from trypanothione disulphide by reduction with excess dithiothreitol [17]. After removal of excess dithiothreitol by extraction with ethyl acetate, equimolar amounts of dihydrotrypanothione (5 /~mol) and melarsen oxide (5 pmol, dissolved in dimethylformamide) were mixed, lyophilised and redissolved in 0.5 M acetic acid. The mixture was applied to a Pharmacia Mono S HR 5/5 column that had been equilibrated with the same solvent. After washing the column for 5 min at 1 ml m i n - l , Mel T was eluted by a linear gradient of solvent B (0.5 M ammonium acetate, 0-100% B over 30 min). Under these conditions Mel T eluted at 65% solvent B. After lyophilization to remove

buffer salts, the product was redissolved in 10 mM HC1 and stored at -20°C.

Reagents. Trypanothione disulphide was purchased from Bachem, monobromobimane from Calbiochem and all other biochemical reagents from either Sigma Chemical Co. or Boehringer Mannheim. The arsenical compounds, melarsoprol (Arsobal, Mel B), melarsen oxide (Mel O) and melarsonyl potassium (Mel W) were kind gifts from Specia, RhonePoulenc; pentamidine, propamidine and stilbamidine from May and Baker (now RhonePoulenc); suramin and nifurtimox from Bayer. Berenil (diminazene aceturate), allopurinol and Formycin B were purchased from Sigma. Results

Drug sensitivity in vivo. An arsenical resistant line was obtained by repeated exposure and sub-passage of the sensitive parent clone to increasing sub-curative concentrations of the pentavalent arsenical drug sodium melarsen. After cloning, the resistant line was found to be > 5.9-fold resistant to sodium melarsen and showed cross-resistance to all of the trivalent

TABLE I Cross-resistance patterns between arsenical sensitive and resistant clones of T. brucei to various trypanocidal compounds in mice Class

Compounds

ED50 (mg k g - l ) Sensitive

Arsenicals

Sodium melarsen Trimelarsen Melarsoprol Melarsen oxide

Diamidines

Stilbamidine Berenil Propamidine Pentamidine

Miscellaneous

Suramin Nifurtimox Allopurinol Formycin B

27.3 0.33 0.17 0.033

Ratio R/S Resistant > 160 40.5 11.3 1.11

> 5.9 122 67 33

0.09 0.22 1.45 0.64

3.43 6.93 8.30 0.95

38 31.5 5.7 1.5

0.50 34.5 > 128 0.51

2.86 34.0 > 128 0.42

5.8 1.0 0.8

All compounds were administered as a single intraperitoneal injection as described in Methods, except nifurtimox, (twice daily for 4 days), allopurinol (once daily for 10 days) and Formycin B (once daily for 3 days).

216

0.250

0.2£)0

. . . . . . . . .

a transient drop in parasitaemia was noted in both cases.

0.125

E

!

0.4

tU,rn

O&

}

t

I

I

I

i

I

[

I

0.4

0.2

0.1~

0.0

0.0 15

Llllllll-

I

30

TIME, mln

Fig. 1. Effectof melarsen oxide and phenylarsineoxide on lysis of melarsen-sensitiveand resistant clones of T. brucei in vitro. Lysis was determined spectrophotometricallyas described in the methods. Panels a, c and e are melarsensensitive T. brucei and panels b, d and f are melarsenresistant T. brucei. Traces A and B in each panel refer to control cells in CBSS or with the addition of 1% v/v dimethylformamide, respectively. Traces C to F in each panel contain decreasing amounts of melarsen oxide or phenylarsineoxide.(a) sensitivecells, 0.75, 0.5, 0.25, 0.1/aM melarsen oxide; (b) resistant cells, 100, 50, I0, 7.5 /IM melarsen oxide; (c) sensitive cells, 5, 2.5, 1, 0.75 ,uM phenylarsine oxide; (d) as panel (c), resistant cells; (e) sensitive cells, 100, 50, 25, 10/~M phenylarsineoxide; (f) as panel (c), resistant cells. arsenicals tested (33- to 122-fold, Table I). Similar resistance factors were obtained when ratios are calculated using the EDgo values (not shown). The melarsen-resistant line also shows varying degrees of cross-resistance to the diamidines, ranging from 1.5-fold for pentamidine to 38-fold for stilbamidine in broad agreement with the work of others [18-20]. Cross-resistance to suramin of 5.8-fold was an unexpected finding, but was confirmed in a duplicate experiment which yielded EDs0s of 0.42 and 2.86 mg kg -1, respectively for the sensitive and resistant lines (6.8-fold resistance). No cross-resistance was found for the nitrofuran, nifurtimox, or the purine analogue, formycin B. Despite prolonged treatment with allopurinol at the maximum tolerated dose (128 mg k g - ] , once daily for 10 days) no cures could be obtained with either strain, although

R e s i s t a n c e to arsenicals in vitro. Cross-resistance to various trivalent melaminophenylarsenical compounds could also be demonstrated in vitro by monitoring cell lysis by the decrease in absorbance due to light scatter at 750 nm. In the case of melarsen-sensitive cells, lysis by melarsen oxide is dose-dependent between 0.1 and 0.75 /aM (Fig. la, traces F to C). Other experiments showed that higher concentrations of melarsen oxide neither increase the rate of lysis nor abolish the time delay before lysis commences. With 0.5 or 0.75 /aM melarsen oxide, lysis is complete after 20 min (Fig. la, traces C and D), which agrees with parallel observations by phase-contrast light microscopy, where all the trypanosomes were observed to have become swollen and immotile with loss of their normal refractile appearance. No lysis was observed with pentavalent sodium melarsen (200/aM) indicating that the drug must be reduced in vivo to trivalent melarsen oxide for trypanocidal activity (not shown). Melarsen-resistant trypanosomes were completely unaffected by melarsen oxide concentrations up to the limits of solubility in CBSS (100/aM; Fig. lb) indicating that these cells are at least 200-fold resistant to this drug in this experimental system. Similar experiments with melarsoprol and trimelarsen showed that higher concentrations ( > 10/aM) or longer incubation times ( > 6 0 min) are required to achieve lysis of the sensitive line than with melarsen oxide (not shown). With the melarsen-resistant line, no lysis was observed with either 200 /aM melarsoprol or 200 /aM trimelarsen, indicating that crossresistance must be greater than 20-fold, confirming the cross-resistance observed in vivo. We were unable to transform our data into a linear form by plotting drug concentration versus the reciprocal of lysis times as reported by Yarlett et al. [21]. Previous studies have noted that crossresistance between the melaminophenyl class of arsenicals does not necessarily extend to other substituted aromatic arsenicals lacking a

217

melaminyl-group [18]. We therefore examined the effects of the lipid soluble arsenical phenylarsine oxide to see whether this was the case in our melarsen-resistant line. The melarsen-sensitive line shows dose-dependent saturable iysis (Fig. lc,e) with complete lysis occurring after 9 min exposure to concentrations in excess of 2.5 pM. As observed with melarsen oxide (Fig. la) both the rate of lysis and the delay time before lysis occurs are independent of drug concentrations above 5 /~M. The melarsen-resistant line also shows dose-dependent lysis over the same concentration range, only there is a slightly longer delay before lysis commences (6 versus 4 min) and the rate of lysis is 4 to 10-fold slower, such that complete lysis is only obtained after nearly 30 min incubation (Fig. ld,f). It was not possible to extend these studies in vivo due to its toxicity to mice.

Thiolcontent. Since trypanothione appears to be the primary target for trivalent arsenicals [1], we have examined both lines to see whether

resistance could be associated with changes in the levels of this metabolite or other lowmolecular weight thiols. Glutathione, glutathionylspermidine and trypanothione levels were not statistically different between melarsen-resistant and sensitive strains (Table II). The values for the sensitive lines are in reasonable agreement with those previously reported [13].

Trypanothione reductase and dihydrolipoamide dehydrogenase. Although thiol levels were unchanged, trypanothione reductase levels were significantly decreased by approximately 50% in the melarsen-resistant line, accompanied by a statistically significant increase of 38% in dihydrolipoamide dehydrogenase activity (Table III). Since oxygen consumption by both cell types was not statistically different, we conclude that the differences in levels of these flavoenzymes cannot be due to changes in cell volume or protein content. In view of these differences, trypanothione reductase was partially purified from both

TABLE II Thiol content of melarsen-sensitive and resistant T. brucei Thiol content (nmol (10 a cells)- t)

Glutathione b Glutathionylspermidineb Dihydrotrypanothioneb Total glutathione b'd

Sensitive

Resistant

0.578 + 0.046 < 0.1 0.731 + 0.181 2.04 + 0.33

0.517 5- 0.027 < 0.1 0.766 __+ 0.050 2.05 + 0.09

Ratio R/S

P~

0.9 1.0 1.0

NS c NS NS

aStudent's t-test. bData are means + SD of 3 separate preparations. ~Not significant by Student's t-test at P >0.1. aTotal glutathione is calculated as the sum of glutathione plus twice the concentration of dihydrotrypanothione.

TABLE IIl Levels of trypanothione reductase, dihydrolipoamide dehydrogenase and respiratory rates in arsenical-sensitive and resistant T. brucei

Trypanothione reductase b Dihydrolipoamide dehydrogenaseb Oxygen consumptionc

Units

Sensitive

Resistant

Ratio R/S

pa

mU mg -~ mU m g - l nmol 02 min-~ (10 s c e l l s ) ~

49.6 + 5.3 34.7 __+ 1.4 55.3 + 3.6

25.3 + 3.9 47.8 + 3.7 50.6 + 2.2

0.51 1.38 0.92

<0.005 <0.005 NS

aMeans of 3 separate determinations using Student's t-test; NS, not significant at P = 0.05. bMeasured on whole homogenates. CDetermined in phosphate-buffered saline containing 50 mM glucose at 25°C.

218

nR

S

HS

~ I C I ~ L

~

N

~

TR

101,

@

2.20,

4.~ 2.0~. 4,4~ 1.10,

0

•

4.0d

I

3.6~

10 1

3.2~

6.00-

2.1K

4.00-

1.6t

O.k 0.4( 020

040 060 0110 [Trupanothtone].

100

120 I~11

140

10 )

0.0(O.OO . 0.20 . .

!

. . 0.40

.0.60.

[Trypanot

0 0~'¢ I . . . . . . -0 20 0 20 0 60 I 00 rs] . 0.I,0 . . 1.00 . .

hione].

. . 1.20 ~M

. 1.40

)~ I0 10 )2

Fig. 2. Km determination for trypanothione reductases purified from arsenical sensitive and resistant lines. The curve is fitted by non-linear regression as described in the methods. The inset shows the data transformed to a Hanes-Wolf plot.

strains for further kinetic analysis. Fig. 2 shows that both trypanothione reductases obey simple Michaelis-Menten kinetics with respect to trypanothione disulphide in the presence of saturating (0.15 mM) concentrations of NADPH. The Km reported here for the sensitive line (22.5 + 2.6 #M) is considerably less than that previously reported (58/~M) [1], but would appear to be correct, since the Km value determined for pure trypanothione reductase from C. fasciculata using the same batch of trypanothione disulphide (58/IM) [22] afforded results compatible with previously published values (53 /iM [14], 51 /~M [23]). In addition, the Km values of the trypanothione reductases from sensitive and resistant lines (22.5 + 2.6 and 21.5 + 1.7 #M, respectively) are not statistically different from each other, indicating that the decreased activity found in resistant cells is not due to an alteration in the enzyme's affinity for trypanothione disulphide. Neither enzyme showed significant activity with 10mM glutathione disulphide ( < 0 . 2 % of the activity observed with 100 /~M trypanothione disulphide as substrate). Inhibition of both trypanothione reductases by Mel T was competitive with respect to trypanothione disulphide (Fig. 3). The inhibition constants for both enzymes are essentially

identical (9.9 and 10.8 /~M, respectively for trypanothione reductase from the sensitive and resistant strains) and in reasonable agreement with the previously determined value of 9/~M [1]. Trypanothione reductases like glutathione reductases contain a pair of cysteines in the disulphide-binding site that undergo oxidation and reduction during catalysis [14]. Thus, we examined whether melarsen oxide could directly inactivate the enzyme by covalent modification of the cysteine pair generated in the NADPH-reduced form of the enzyme. Trypanothione reductase from both cell types shows a time-dependent inactivation with 50 /~M melarsen oxide which follows pseudo-firstorder kinetics with inactivation rate constants (k °bsd corrected for inactivation by NADPH) of 4.71 _+ 0.34 × 1 0 - 2 m i n - ] and5.11 _+ 0.17 × 10 - 2 min- ~, respectively for trypanothione reductase from melarsen-sensitive and resistant isolates (Fig. 4). The limited amount of enzyme available precluded a determination of the second-order rate constant for inactivation for the T. brucei enzymes. However, similar inactivation kinetics with melarsen oxide are observed for C. fasciculata trypanothione reductase, including the apparent instantaneous inactivation observed on extrapolating

219

A

l/v, mU -s

100

0-3

l

,

I

,

,

J

Ki= 9-9¥M

-5

e

5

10

15

I

I

I

0

5

10

15

I I I I

ae

-1

I

II

I

Ki= 10"8 IJM

-5

1

0.S

Fig. 3. Determination of inhibitor constants for the melarsen oxide adduct with trypanothione for trypanothione reductase purified from arsenical-resistant and sensitive lines. Kis were determined by linear regression of the Dixon plots of reciprocal enzyme activity versus inhibitor concentration ([MELT]). (A) Trypanothione reductase from arsenical sensitive line. (B) trypanothione reductase from arsenical resistant line. Open circles assayed with 20/tM trypanothione disulphide and closed circles 100 /~M trypanothione disulphide.

back to zero time (M. Cunningham, K. Smith and A.H. Fairlamb, unpublished results). Although there is a slight difference in the extent of the immediate inhibition by melarsen oxide (83% versus 72% for trypanothione reductase from the resistant line, see Fig. 4), the subsequent k °b~ values are not significantly different.

Discussion

Our previous work on the mode of action of trivalent arsenicals has provided evidence for trypanothione being a primary target for these drugs, with the resulting arsenical-trypanothione complex Mel T secondarily acting as a

0

I

I

I

I

15

30

45

60

TIME, rain

Fig. 4. Kinetics of inactivation of partially purified trypanothione reductase by melarsen oxide. Open symbols, trypanothione reductase from arsenical sensitive line; closed symbols, trypanothione reductase from arsenical

resistant line. Squares, loss in enzymeactivity on incubation with NADPH; circles, loss in enzyme activity on incubation with NADPH plus 50 #M melarsenoxide. competitive inhibitor of trypanothione reductase [1]. In addition, our current findings suggest that trivalent arsenicals can also irreversibly inhibit trypanothione reductase in vitro by direct covalent modification of the active site cysteine residues. If these events are involved in the trypanocidal action of arsenicals, then one might suppose that arsenicalresistant trypanosomes would show alterations in either the primary target, trypanothione, or the secondary targets (of which there may be many) such as trypanothione reductase. The present study clearly indicates that no changes in intracellular concentrations of trypanothione or other glutathione-containing compounds has occurred, in agreement with the findings of Yarlett et al. [21]. Moreover, apart

220

tYom an unexplained decrease in trypanothione reductase levels in melarsen-resistant cells, the enzyme isolated from both sensitive and resistant strains appears to be identical with respect to their Kms for trypanothione, their sensitivities to inhibition by the competitive inhibitor Mel T and their inactivation by melarsen oxide. Thus, resistance does not appear to be due to changes in the primary target, trypanothione, or one of the possible secondary targets, trypanothione reductase. Changes in lipoic acid metabolism that may be involved in resistance to arsenicals are examined in the accompanying paper [34]. The other common mechanism of drug resistance involves a decrease in the intracellular drug levels either by decreased uptake or increased export. There are numerous reports in the early literature that normal trypanosomes readily absorb and concentrate trivalent arsenicals whereas arsenical-resistant organisms do so much less readily (see ref. 24 and references therein). Moreover, the lack of cross-resistance between different classes of arsenical drugs suggests that resistance is conferred by the non-arsenical part of the molecule. In our melarsen-resistant line this can be ascribed to the melaminyl moiety since the marked cross-resistance between all the melaminophenyl arsenicals in vivo and in vitro does not extend to the lipid soluble arsenical phenylarsine oxide. Cross-resistance between the melaminophenyl arsenicals and the diamidines has been observed previously in laboratory-derived or uncloned field isolates [18-20], although Williamson has noted that crossresistance patterns are unpredictable and may not be reciprocal [25]. In one study it was also noted that T. brucei clones selected for resistance to melarsoprol accumulated lower intracellular concentrations of the fluorescent diamidine compounds, 4',6-diamidino-2-phenyl-indole (DAPI) and Hoechst 33342 [20]. All of the above data are compatible with the hypothesis that the melaminophenyl arsenicals and the diamidines are taken up on a common transporter. Further, that the common structural feature for recognition is likely to reside in the melamine and benzamidine moieties. If

TABLE IV Maximum interatomic distances of diamidines and crossresistance factors Compound

Melarsen oxide Stilbamidine Berenil Propamidine Pentamidine

Resistance factor ~'

33.0 38.0 31.5 5.7 1.5

Interatomic distance (A,) Between amidine carbons b

Between rings ¢

N.A. 12.4 12.5 14.8 17.1

4.8 6.5 7.1 8.7 12.2

~'Resistance factor is ratio of EDs0s (mg kg 'a) for melarsen-resistant and sensitive clones. bMaximum interatomic distances wcrc measured on CPK molecular models and are in agreement with previously reported values [33]. CMeasured between centres of each benzamidine moiety in the diamidines and between benzene and melamine moieties in melarscn oxide.

so, then the relatively low cross-resistance that we have observed for pentamidine and propamidine would suggest that alteration of affinity rather than loss of a transporter must have occurred. It is therefore of interest to note that there is a strong correlation between the maximum interatomic distances between the amidine groups of the diamidines (Table IV), such that the shorter the interatomic distance the larger the resistance to that particular drug. This relationship also extends to the distance between the melaminyl and phenyl moieties of melarsen oxide and between the benzamidine moieties of the diamidines (Table IV). One could therefore envisage a structural alteration in such a transporter that would exclude 2 closely spaced bulky groups while allowing binding of more widely spaced groups such as in propamidine and pentamidine. Similar arguments could be advanced in favour of a common drug target, although it is more difficult to account for some of the observations given above. Although decreased drug accumulation due to increased export via a multi-drug transport system could account for the cross-resistance we observe, direct evidence in support of this mechanism is completely lacking for T. brucei. Indeed, the strongly ionic nature of the

221

diamidines and suramin does not accord with the hydrophobic substrate-specificity of other multi-drug transporters [26,27] and verapamil does not reverse resistance to melaminophenyl arsenicals (ref. 21; and Carter and Fairlamb, unpublished). In Leishmania mexicana, resistance to arsenite can be associated with Hcircle amplification [28] and cross-resistance to methotrexate [29]. Moreover, Leishmania tarentolae encodes several genes homologous to the mammalian P-glycoprotein pump, one of which maps within the Leishmania H-region [30]. Recent work by Callaghan and Beverley [31], has elegantly demonstrated that transfection of L. major with segments of H-region DNA containing either the P-glycoprotein gene from Leishmania major (lmpgpA) or L. tarentolae (ltpgpA) confers resistance to arsenite and trivalent antimonials, but not to pentavalent antimonials or methotrexate. The pgpA drug resistance phenotype is similar to that of heavy metal resistance in bacteria [32]. Unfortunately, resistance to trivalent aromatic arsenicals was not tested. Further work is required to elucidate the modes of drug action and mechanisms of drug resistance in African trypanosomes.

Acknowledgements We thank Dr Peter Ulrich for producing the initial strain resistant to sodium melarsen, Rhone-Poulenc, May and Baker and Bayer for the provision of trypanocidal drugs, and Christopher Keating and Alison Thomas for technical assistance. This investigation received financial support from the U N D P / W O R L D BANK/WHO Special Programme for Research and Training in Tropical Diseases and the Wellcome Trust. NSC received additional support from the Medical Research Council.

References 1 Fairlamb, A.H., Henderson, G.B. and Cerami, A. (1989) Trypanothione is the primary target for arsenical drugs against African trypanosomes. Proc. Natl. Acad. Sci. USA 86, 2607-261 I.

2 Fairlamb, A.H., Blackburn, P., Ulrich, P., Chait, B.T. and Cerami, A. (1985) Trypanothione: a novel bis(glutathionyl)spermidine cofactor for glutathione reductase in trypanosomatids. Science 227, 1485 1487. 3 Fairlamb, A.H. (1991) Trypanothione metabolism in the chemotherapy of Leishmaniasis and Trypanosomiasis. In: Molecular and Immunological Aspects of Parasitism (Wang C.C., ed.), pp. 107-121. American Association for the Advancement of Science, Washington, DC. 4 Smith, K., Mills, A., Thornton, J.M. and Fairlamb, A.H. (1991) Trypanothione metabolism as a target for drug design: molecular modelling of trypanothione reductase. In: Biochemical Protozoology (Coombs GH. and North M.J., eds.) pp. 482-492. Taylor and Francis, London. 5 Docampo, R. (1990) Sensitivity of parasites to free radical damage by antiparasitic drugs. Chem.-Biol. Interact. 73, I 27. 6 Fairlamb, A.H. (1990) Trypanothione metabolism and rational approaches to drug design. Biochem. Soc. Trans. 18, 717 720. 7 Fairlamb, A.H. (1990) Future prospects for the chemotherapy of human trypanosomiasis. Trans. R. Soc. Trop. Med. Hyg. 84, 613--617. 8 Fairlamb, A.H. (1989) Novel biochemical pathways in parasitic protozoa. Parasitol. 99S, 93-112. 9 Walsh, C.T., Bradley, M. and Nadeau, K. (1991) Molecular studies on trypanothione reductase, a target for antiparasitic drugs. Trends Biochem. Sci. 16, 305309. 10 Lanham, S.M. (1968) Separation of trypanosomes from the blood of infected rats and mice by anionexchangers. Nature 218, 1273 1274. 11 Smith, C.J., Levine, R.F. and Mansfield, J.M. (1982) Cloning of African trypanosomes in mice immunosuppressed by cyclophosphamide treatment. Am. J. Trop. Med. Hyg. 31, 1098 1102. 12 Meshnick, S.R., Blobstein, S.H., Grady, R.W. and Cerami, A. (1978) An approach to the development of new drugs for African Trypanosomiasis. J. Exp. Med. 148, 569-579. 13 Fairlamb, A.H., Henderson, G.B., Bacchi, C.J. and Cerami, A. (1987) In vivo effects of difluoromethylornithine on trypanothione and polyamine levels in bloodstream forms of Trypanosoma brucei. Mol. Biochem. Parasitol. 24, 185-191. 14 Shames, S.L., Fairlamb, A.H., Cerami, A. and Walsh, C.T. (1986) Purification and characterization of trypanothione reductase from Crithidia fasciculata, a newly discovered member of the family of disulphidecontaining flavoprotein reductases. Biochemistry 25, 3519 3526. 15 Danson, M.J., Conroy, K., McQuattie, A. and Stevenson, K.J. (1987) Dihydrolipoamide dehydrogenase from Trypanosoma brucei. Biochem. J. 243, 66 I- 665. 16 Fairlamb, A.H. and Bowman, I.B. (1980) Uptake of the trypanocidal drug suramin by bloodstream forms of Trypanosoma brucei and its effect on respiration and growth rate in vivo. Mol. Biochem. Parasitol. I, 315-333. 17 Fairlamb, A.H., Henderson, G.B. and Cerami, A. (1986) The biosynthesis of trypanothione and N Iglutathionylspermidine in Crithidia J~'ciculata. Mol. Biochem. Parasitol. 21, 247 257.

222 18 Rollo, I.M. and Williamson, J. (1951) Acquired resistance to Melarsen, tryparsamide and amidines in pathogenic trypanosomes after treatment with Melarscn alonc. Nature 167, 147-148. 19 Williamson, J. and Rollo, I.M. (1959) Drug rcsistance in trypanosomes: cross-resistance analyses. Br. J. Pharmacol. 14, 423-430. 20 Frommel, T.O. and Balbcr, A.E. (1987) Flow cytofluorimetric analysis of drug accumulation by multidrug-resistant Trypanosoma brucei hrucei and T. b. rhodesiense. Mol. Biochem. Parasitol. 26, 183 191. 21 Yarlett, N., Goldberg. B., Nathan, H.C., Garofalo, J. and Bacchi, C.J. (1991) Differential susceptibility of Trypanosoma brucei rhodesiense isolates to in vitro lysis by arsenicals. Exp. Parasitol. 72, 205 215. 22 EI-Wacr, A., Douglas, K.T., Smith, K. and Fairlamb, A.H. (1991) Synthesis of N-benzyloxycarbonyl-l_,cysteinylglycinc 3-dimethylaminopropylamide disulphide: a cheap and convenient new assay for trypanothionc reductasc. Anal. Biochem. 198, 212 216. 23 Henderson, G.B., l-airlamb, A.H., Ulrich, P. and Ccrami, A. (1987) Substrate specificity of the flavoprotein trypanothione disulfide reductase from Crithidia fasciculata. Biochemistry 26, 3023 3027. 24 Hawking, F. (1938) Analysis of the trypanocidal action of trivalent arsenicals and acriflavine. Ann. Trop. Med. Parasitol. 32, 313 331. 25 Williamson, J. (1970) Review of chemotherapeutic and chemoprophylactic agents. In: The African Trypanosomiases (Mulligan H.W., ed.), pp. 125-221. Allen and Unwin, London. 26 Gottesman. M.M. and Pastan, i. (1988) The multidrug

27

28 29

30

31 32 33

34

transporter, a double-edged sword. J. Biol. C'hem. 263, 12163-12166. Juranka, P.F., Zastawny, R.1.. and Ling, V. (1989) Pglycoprotein: Multidrug-rcsistance and a superfamily of membrane-associated transport proteins. FASEB J. 3. 2583--2592. Detkc, S., Katakura, K. and Chang, K.-P. (1989) DNA amplification in arsenite-rcsistant Leishmania. Exp. Cell Res. 180, 161 170. Katakura, K. and Chang, K.P. (1989) H DNA amplification in l.x,ishmania resistant to both arsenite and methotrexate. Mol. Biochem. Parasitol. 34, 189191. Oucllette, M., Fase-Fowler, F. and Borst, P. (1990)The amplified H circle of methotrexate-resistant Leishmania tarentolae contains a novel P-glycoprotein genc. EMBO J. 9, 1027-1033. Callaghan, H.L. and Beverley, S.M. (1991) Heavy metal resistance: a new role for P-glycoproteins in l,eishrnania. J.BioI.Chem. 266, 18427 18430. Mobley, H.L. and Rosen, B.P. (1982) Energetics of plasmid-mediated arsenate resistance in Escherichia coll. Proc. Natl. Acad. Sci. USA 79, 6119-6122. Ulrich, P. and Cerami. A. (1984) Trypanocidal 1,3arylene diketone bis(guanylhydrazone)s. Structureactivity relationships among substituted and heterocyclic analogues. J. Med. Chem. 27, 35-40. Fairlamb, A.H., Smith, K. and Hunter, K.J. (1992)The interaction of arsenical drugs with dihydrolipoamide and dihydrolipoamide dehydrogenase from arsenical resistant and sensitive strains of Trypanosoma brucei hrucei. Mol. Biochem. Parasitol. 53, 223 232.