Drug combinations as effective anti-leishmanials against drug resistant Leishmania mexicana

Abstract

Leishmania is a parasite that causes the disease leishmaniasis, and 700 000 to 1 million new cases occur each year. There are few drugs that treat the disease and drug resistance in the parasite limits the clinical utility of existing drugs. One way to combat drug resistance is to use combination therapy rather than monotherapy. In this study we have compared the effect of single and combination treatments with four different compounds, i.e. alkylphosphocholine analogues APC12 and APC14, miltefosine (MIL), ketoconazole (KTZ), and amphotericin B (AmpB), on the survival of Leishmania mexicana wild-type promastigotes and a cell line derived from the WT with induced resistance to APC12 (C12Rx). The combination treatment with APC14 and APC16 had a synergistic effect in killing the WT while the combination treatment with KTZ and APC12 or APC14 or APC12 and APC14 had a synergistic effect against C12Rx. More than 90% killing efficiency was obtained using APC12 alone at >1 mg ml⁻¹ against the C12Rx strain; however, combinations with APC14 produced a similar killing efficiency using APC12 at 0.063–0.25 mg ml⁻¹ and APC14 at 0.003–0.5 mg ml⁻¹. These results show that combination therapy can negate induced drug resistance in L. mexicana and that the use of this type of screening system could accelerate the development of drug combinations for clinical use.

Synergistic and antagonist drug interactions of drug combinations against Leishmania drug sensitive and resistant cell lines.

Introduction

Leishmaniasis a disease caused by infection with the protozoan parasite Leishmania, and 700 000 to 1 million new cases occur each year.¹ There are three clinical forms of leishmaniasis, cutaneous leishmaniasis, visceral leishmaniasis (VL) and mucocutaneous leishmaniasis. Development of preventive strategies such as vaccines and vector control are challenging. Drug overuse and misuse in treatment centres have selected a plethora of resistant parasites, and the net consequences are, existing drugs having limited clinical efficacy.² Subversion of the therapeutic anti-leishmanial activity by the parasite is due to the parasite's genomic and metabolic plasticity and gene switching, regulated by epigenetic and post-translational modifications.^3–5 The nature of drug resistance genes varies, and drug combinations administered simultaneously and sequentially have been shown to circumvent acquired resistance.⁶ For this reason, drug combinations are often used for treatment for drug resistant conditions, for example cancer therapy.^7–9 However, identification of what drugs or drug regimen should be used is often done on an ad hoc basis rather than having a cohesive strategy to identify optimal drug combinations and treatment regimens. Joint treatment with paromomycin (PMM) and miltefosine (MIL) can delay the selection of PMM resistance^10,11 in L. donovani, and joint treatment with pentavalent antimonials with PMM and liposomal AMB is recommended for VL in some countries.¹² For widespread application of this approach, elaborate hypothesis-free, high-throughput screening assays of all possible physiological dose-ratio matrixes in pair-wise cytotoxicity assays with distinctive endpoints are required. This approach aids the mapping out of synergistic, additive, and antagonistic interaction profiles with mathematical models based upon the median-effect principle, combination index theorem or Loewe additivity and Bliss models, aided by online software such as Combenefit. The output from these models can inform the identification of molecules that target compensatory pathways used by drug resistant cells.^7,13–15

Many molecular targets have been described in Leishmania, and the biosynthetic pathway for its main membrane sterol ergosterol, synthesised from leucine,^16,17e.g. sterol 14α-demethylase (14α-DM; EC1.14.13.70), has been described as essential for parasite development.^18–21 Further, changes in the composition and proportion of sterols in most drug-resistant parasites, even with molecular targets unrelated to sterol biosynthesis, point towards this pathway having an essential role in developing drug resistance^22–28 particularly in the biosynthesis of membrane phospholipids of Leishmania parasites.^29–31 Thus we hypothesise that the link between the selection of resistant parasites and metabolic and genetic plasticity suggests that inhibitors that simultaneously target these biosynthetic pathways could be effective anti-leishmanials, particularly against drug resistant Leishmania parasites. To this end, we have developed a systematic proof-of-concept study to evaluate the cytotoxicity of two-drug combinations that target sterols and phospholipids of Leishmania mexicana wild-type (WT) parasites and a cell line with laboratory induced resistance to the MIL analogue APC12.³²

Results and discussion

C12Rx promastigotes have altered sterol composition and are sensitive to sterol biosynthetic inhibitors

It well known that drug-resistant Leishmania-parasites have the proteins involved in sterol biosynthesis, their intermediates and end products altered well known that drug-resistant Leishmania-parasites have the proteins involved in sterol biosynthesis, their intermediates and end products altered.^{22–26,28,30,33,34} For example, sterol 14α-DM in Leishmania amazonensis and sterol C-22 desaturase (EC 1.14.19.41) in L. donovani are upregulated in parasites resistant to ketoconazole (KTZ)²⁶ and amphotericin B (AmpB),²³ whilst stigmasterol is a potential biomarker for AmpB resistance in L. donovani.²² Therefore the sterol profile of WT and a daughter strain C12Rx, which had laboratory induced resistance to the alkylphosphocholine analogue, APC12,³² was compared using logarithmic (log) and stationary growth phase promastigotes using gas chromatography-mass spectrometry (GC-MS). Five sterols, namely squalene, zymosterol, ergosterol, brassicasterol and cholesterol, were identified in the samples based on retention time, molecular weights, and electron impact mass spectra. The abundance of squalene and brassicasterol in the WT was low at log and stationary phases, respectively, relative to cholesterol, zymosterol and ergosterol (Fig. 1, dark gray bar). Squalene accounted for 0.01% of the total sterol in log stage promastigotes but increased to 0.19% in stationary phase promastigotes. Ergosterol and zymosterol were the major sterol in log stage parasites (ergosterol 62% and zymosterol 22%, Fig. 1a, dark gray bar) and continued to be the major constituents in stationary phase promastigotes (ergosterol 68% and zymosterol 17%; Fig. 1b, dark bar and Table S1†). Cholesterol which cannot be synthesised by Leishmania but is salvaged from their serum-supplemented (10% v/v) growth media, used for in vitro growth,³⁵ increased in abundance (Fig. 1, dark gray bar), between the log and stationary growth phase L. mexicana WT, constituting 15% and 17% of the total sterol, respectively. In the APC12 resistant strain, C12Rx, ergosterol continued to be the major sterol present in log and stationary stage parasites, and was significantly reduced when compared to that in the WT (p < 0.01).

Fig. 1. Sterol abundance in L. mexicana promastigotes. Sterols were extracted from 109 WT (dark gray bar) or C12Rx (light gray bar) promastigotes, using log (3 days in in vitro cultures from a starting cell density of 5 × 10⁵ cells; top panel) or stationary phase promastigotes (6 days in in vitro culture, bottom panel). The concentration of sterols present was determined by comparing the area under the curve in chromatograms of the 5 sterols with the relevant sterol standard run in the same GC-MS assay. Data are mean ± SD of sterols from three replicates from pooled cultures.

The C12Rx strain took up significantly more cholesterol at the stationary growth phase compared to the WT (p < 0.01), perhaps, to compensate for the loss of ergosterol, since cholesterol can only be derived from the culture medium.^36,37 Similar experiments in complete and spent medium from the WT and C12Rx produced inconsistent results, probably reflecting the detection limit of the assay. The brassicasterol abundance was significantly different between the log and stationary phases (Fig. 1, light gray bar, p < 0.01), constituting 5% and 1%, respectively. Interestingly, the brassicasterol to ergosterol conversion rate was significantly higher in the C12Rx strain compared to that in the WT for log phase promastigotes (WT, 3.11%; C12Rx, 6.92%) but almost equal for stationary phase parasites (WT, 2.71%; C12Rx, 2.48%), suggesting an altered sterol 17β-hydroxylsterol dehydrogenase (EC 1.1.1.51) activity.³⁵ The abundance of all 5 sterols was significantly higher in stationary than log phase cells for both strains (p < 0.01) being 1.7-fold and 1.7-fold higher in the WT (log, 2.12 × 10⁹ ± 5.99 × 10⁷; stationary, 6.46 × 10⁹ ± 6.28 × 10⁷) than in the C12Rx (log, 1.25 × 10⁹ ± 4.78 × 10⁷; stationary, 3.93 × 10⁹ ± 8.10 × 10⁷) strain at log and stationary growth phases, respectively (Table S1†). Overall, these data showed that the APC12 resistance was associated with altered sterol profiles. This reduced sterol level in the C12Rx strain suggested that C12Rx was heavily reliant on its reduced intracellular sterols for survival, so we used sterol biosynthetic inhibitors (SBIs) to test the druggability of enzymes in this pathway.

Activity of sterol biosynthetic inhibitors against L. mexicana

The importance of the enzyme sterol 14α-demethylase in parasite survival was determined by determining the sensitivity of the WT and C12Rx L. mexicana parasites to ketoconazole, KTZ, a specific inhibitor of this enzyme.³⁸ The APC12 strain was more sensitive to the KTZ treatment as it had significantly lower IC₅₀ and IC₉₀ values (p < 0.05, mean IC₉₀ ± SD, WT, IC₉₀ 60.36 ± 4.21 μg ml⁻¹, C12Rx, 4.40 ± 0.48 μg ml⁻¹; mean IC₅₀ values are shown in Table 1). Amphotericin B, AmpB, binds to and removes ergosterol from membranes and induces cell death in Leishmania parasites by altering the integrity of their cell membrane.³⁹ Treatment with AmpB was significantly more effective against the APC12 resistant strain compared to the WT (p < 0.05, mean IC₉₀ ± SD, WT 9480 ± 474.28 μg ml⁻¹, C12Rx 5196 ± 24.23 μg ml⁻¹; mean IC₅₀ values are shown in Table 1). Although the C12Rx strain had a significantly higher endogenous cholesterol level at the stationary phase of parasite growth compared to the WT, this did have a protective role.³⁶ As the cholesterol uptake has been deemed protective, our results for C12Rx indicated that the higher drug-induced death than that in the WT occurred whilst the parasites were in the log phase of growth and the cholesterol level was reduced (compare Table 1 and Fig. 1).

Mean IC₅₀ of SBIs against L. mexicana strains. WT and C12Rx promastigotes (5 × 10⁶/ml, n = 3 independent treatment in triplicate) at the log phase were cultured in the presence of medium alone (control) or different concentrations of ketoconazole (KTZ) or amphotericin B (AmpB) for 3 days at 25 °C. The effect of treatment on parasite survival was determined by determining the mean suppression in parasite growth for each experimental value compared to the mean control value. These data were then used to determine the mean IC₅₀ value using the Grafit software. The resistant index RI₅₀ was the quotient of the mean IC₅₀ values of the two strains with C12Rx and the WT being the numerator and denominator, respectively. The p value compares the IC₅₀ data for the WT vs. C12Rx strain.

SBIs	IC50 μM (μg ml−1)		Resistant index (RI50)	p value
	WT	C12Rx
KTZ	37.86 ± 0.62	9.03 ± 0.50	0.24	p < 0.05
	(20.12 ± 6.65)	(4.8 ± 1.97)
AmpB	3419.62 ± 52.96	1899.19 ± 99.50	0.56	p < 0.05
	(3160 ± 232.12)	(1755 ± 13.28)

Activity of alkylphosphocholines against L. mexicana

Previous studies have shown that (a) the composition of phospholipids is different between miltefosine-, MIL-resistant and WT parasites for L. donovani,² and (b) the degree of cytotoxicity of APCs against L. mexicana was dependent on their alkyl chain length; with APC12 being more active than APC16 against naïve parasites, WT and C12Rx.³² In this study we determined the activity of a 14 alkyl carbon chain, APC14, against both WT and C12Rx L. mexicana promastigotes to determine if cross-resistance occurred (Table 2). The dose response curves for the three APCs for WT and C12Rx are presented in Fig. 2. The resistant indices obtained from the IC₅₀ values of the APCs indicated that cross resistance did occur within the series but its effect was limited for APCs with longer alkyl carbon chains (Table 2). Death from APCs is caused by solubilisation of the parasite's membrane including Leishmania, leading to the formation of individual or aggregated molecules called micelles;³⁰ the latter are formed above a threshold concentration called the critical micellar concentration (CMC). For example, MIL (APC16) at concentrations below, at, or above the CMC has distinctive properties, for example, promoting fluidisation of model membranes, loss of vesicle integrity and formation of mixed aggregates.^30,37 Extrapolation of this to the APCs used in this study showed a structure–activity relationship, indicating that all the APCs tested were micellar-independent against the WT (CMC > IC₅₀; Table 2) but micellar formation was required for APC14 and APC16 against the C12Rx strain (IC₅₀ > CMC; Table 2). These results suggest that the membrane phospholipid metabolism or composition was altered in the C12Rx strain compared to those in the WT parent. Alteration of fatty acid and sterol metabolism has been described previously in miltefosine-resistant Leishmania donovani promastigotes.³⁰

The activity of APCs against L. mexicana. WT or C12Rx promastigotes (5 × 10⁶/ml, n = 3 independent treatment in triplicate) were cultured in the presence of medium alone (control) or different concentrations of the alkylphosphocholines APC12, APC14 and APC16 for 3 days at 25 °C. The effect of treatment on parasite survival was determined by determining the mean suppression in parasite growth for each experimental value compared to the mean control value. These data were then used to determine the mean IC₅₀ value using the Grafit software. The resistant index RI₅₀ was the quotient of the mean IC₅₀ values of the two strains with C12Rx and the WT being the numerator and denominator, respectively; CMC, critical micellar concentration.

Compound	IC50 μM (μg ml−1)		p value	RI50	CMC (μM)
	WT	C12Rx
APC12	0.19 ± 001b	189.58 ± 10.14b	p < 0.05	952.00	1000a
	(0.07 ± 0.03)	(66.64 ± 12.26)
APC14	0.72 ± 0.02b	130.96 ± 6.12a	p < 0.05	184.07	120a
	(0.27 ± 0.05)	(49.70 ± 7.23)
APC16	4.32 ± 0.09xb	70.92 ± 6.78a	p < 0.05	16.42	10a
	(1.76 ± 0.20)	(28.90 ± 9.97)

^aMicellar. See Anatrace product description sheet for respective APCs.

^bNo micellar requirement.

Fig. 2. Dose response curves of alkylphosphocholine analogues against promastigotes of Leishmania mexicana. WT (open squares) or C12Rx (closed squares) promastigotes (5 × 10⁶) were incubated for 72 h at 26 °C with different APC analogues, named APC12 (top panel), APC14 (middle panel) and APC16 (bottom panel) at various concentrations (WT, 2.44 × 10⁻⁵–0.025 μg ml⁻¹; C12Rx, 0.003 mg ml⁻¹ to 4 mg ml⁻¹). Cell viability was assessed using the luciferase assay at a wavelength/bandwidth of 545/40 nm. Data are presented as mean ± standard deviation from three independent assays. Grafit® software was used to produce the dose response curve and values from the three experiments were used to calculate the mean IC₅₀ values.

Activity of mixed APCs against L. mexicana

Combination therapies are now a standard treatment for infectious diseases such as tuberculosis, HIV/AIDS and those caused by drug resistant parasites.^40–43 The success of this approach is due to the differences in the intracellular distribution drug profiles, their pharmacokinetics,⁴⁴ pharmacodynamics⁴⁴ and the probability that the genes that confer the parasite's resistance belong to the same molecular pathway, or if from different pathways, are not genetically linked.⁶ We therefore assessed the ability of different APC combinations to inhibit the survival of L. mexicana. APC12 mixed with APC14 (Fig. 3a) or APC16 (Fig. 3b) interacted predominantly antagonistically against the WT parasites. In contrast, APC14 and APC16 had significant synergistic interactions (Fig. 3c) suggesting that identification of synergistic combination ratios required a systematic approach and could not be predicted. Similar studies using C12Rx showed that APC14 and APC16 interacted antagonistically (Fig. 3f) whereas, APC12 and APC16 and APC12 and APC14 acted antagonistically (Fig. 3e) and synergistically (Fig. 3d), respectively. These results confirmed the findings from the WT, i.e. the physical properties of the individual APCs were not a key determinant to decipher their interaction profiles. Biophysical properties such as surface tension, osmotic pressure, solubility, packing on membranes and ease of forming micelles of other tailed compounds will also have an impact. For example, a mixture of cetyl trimethylammonium bromide (CTAB) and alkyltrimethylammonium bromide (CnTAB) formed novel molecular species different from their individual counterparts.^45–47 Interestingly, the treatment with APC12 and other APCs which interact synergistically negated the inherent resistance of the C12Rx strain (Fig. 3d).

Fig. 3. Contour maps showing the levels of compound interaction based on anti-leishmanial activities between APCs, mapped-out using the Loewe model. Combenefit® software was used to produce an interaction profile for WT (left) and C12Rx (right) treated with APC12 and APC14 (a), APC12 and APC16 (b), and APC14 and APC16 (c). The APC drug concentrations ranged from 2.44 × 10⁻⁵ to 0.025 μg ml⁻¹ for the WT strain and 0.003–4.000 mg ml⁻¹ for the C12Rx strain. Synergistic (blue), additive (green to yellow) and antagonistic (red) interactions were noted.

The effect of combination treatment with APCs and KTZ and AmpB was also investigated. KTZ mixed with APC12 or APC16 gave a synergistic effect (Fig. 4a and c) but this did not occur for APC14 (Fig. 4b). Joint treatment with individual APCs and AmpB had an antagonistic effect (Fig. 4d–f). Previous studies have shown that treatment with MIL (APC16) and AmpB produces pharmacokinetic antagonism as AmpB dissociates into monomers and forms mixed aggregates which cannot penetrate through membranes.⁴⁶ Although our studies indicate that the APC–AmpB mixture in our assays contained molecular species which were impermeable to the membrane of Leishmania promastigotes, liposomal AmpB and miltefosine have been successfully used as an effective treatment for VL.^48–50 However, in this case entrapment of AmpB into vesicles meant that the vesicle protected the drug from the external environment and targeted the drug for uptake by infected macrophages.^51,52

Fig. 4. Contour map of the interaction based on anti-leishmanial activities of APCs and sterol biosynthetic inhibitors, KTX (left) and AmpB (right), based on the Loewe model. The Combenefit® software produced an interaction profile for C12Rx treated with APC12 (a and d), APC14 (b and e) or APC16 (c and f) and KTZ (a–c) or AmpB (d–f). The drug concentrations for APCs used in assays were 0.003–4 mg ml⁻¹; KTZ, 6.10 × 10⁻⁵–0.0625 mg ml⁻¹; AmpB, 0.003–3.125 mg ml⁻¹. Synergistic (blue), additive (green to yellow) and antagonistic (red) interactions were noted.

Conclusion

Treatment with drug combinations that have a synergistic effect can increase the efficacy of a drug compared to monotherapy and can be an approach that can be used to extend the clinical life of an existing drug.⁵³ Our phenotypic repurposing screening has allowed the identification of new synergistic APC and SBI drug combinations for the treatment of L. mexicana, and has given a greater understanding of metabolic changes associated with APC resistance in the C12Rx strain. This approach can be used with other Leishmania species and extended to include the intracellular amastigote stage, which is the stage treated in the clinic. The use of suitably labelled parasites e.g. luciferase-expressing parasites, would allow higher throughput screening⁵⁴ and help in drug repositioning studies using parasites with different inherent drug sensitivities, so that any identified regimens are active against drug resistant parasites.

Experimental section

Strain and cultures

Transgenic Leishmania mexicana promastigotes (5 × 10⁶ cells per ml) of strain MYNC/BZ/62/M379 expressing the firefly luciferase gene and sensitive to MIL APC12 with a 12 alkyl carbon chain called APC12 were designated as WT; a related strain, C12Rx, resistant to 80 μg ml⁻¹ APC12, was selected under controlled conditions by a stepwise progressive increase of APC12 as described by Alotaibi, et al., 2019.³² Both were cultured in complete modified Eagle's medium (M199 supplemented with 10% (v/v) heat inactivated foetal calf serum) at 25 °C. The transgenic line cultures were further supplemented with hygromycin B in order to retain the extrachromosomal luciferase gene.

Gas chromatography mass spectroscopy (GC/MS) of Leishmania spp. sterols

The samples used for this extraction were log phase (day 3) and stationary phase (day 6) cells (10⁹ cells per ml) pooled from several in vitro cultures to achieve the desired cell density with sterol abundance above the baseline of the GC-MS. Each pooled sample from the experiment was split into three replicates containing 10⁹ cells and the sterols present were extracted. Sterols were extracted using a sterol extraction kit, which was done as detailed in the manufacturer's instruction (Sigma Chemical Co Ltd, Poole, UK). The extracted sterols were resolved and analysed by GC/MS on a Thermo Scientific TRS-MS (5% (v/v) polar) controlled with the Thermos Xcalibur software using the method described by Xu et al.³⁶ Electron ionization mass spectroscopy of major Leishmania sterols was performed at 70 eV. The retention time and total ion mass spectra compared with known standards in the NIST mass spectral library were used for sterol identification. Data are presented as mean ± SD for three separate experiments. The conversion of ergosterol to brassicasterol was calculated as the percentage of brassicasterol formed compared to the mass of brassicasterol + ergosterol extracted from the cells. Fold changes in total sterols were the quotient of the total sterol abundance in the WT divided by that in C12Rx.

Compound efficacy studies

The anti-leishmanial activity of APCs against L. mexicana luciferase-expressing promastigotes cultured in complete M199 was determined by adding 100 μl of the appropriate parasite line (5 × 10⁵ cells per ml) to the wells of a 96 well plate and adding 100 μl complete M199 alone (control) or 100 μl medium containing the relevant compounds (APC, 0.01–6.25 μg ml⁻¹ for WT and 0.01–4 mg ml⁻¹ for C12Rx, n = 3/treatment), KTZ (6.25–0.06 μg ml⁻¹ for WT and C12Rx) or AmpB (3.125–0.0003 mg ml⁻¹ for WT and C12Rx). The samples were incubated for 72 h at 25 °C, long enough for nutrients in medium to be non-limiting in the no drug controls. After that, 20 μl luciferin solution was added to the appropriate wells of the 96 well plate (1 μg ml⁻¹ in medium without FCS), and the amount of light emitted per well was measured using a luminometer (Biotek Synergy HT, relative light units) at a wavelength and bandwidth of 440/40 nm. The effect of drug treatment on parasite survival was determined by calculating the mean suppression in the light emitted from the drug treated sample compared to the mean control value. The suppression data were then used to calculate the IC₅₀ or IC₉₀ value using Grafit software (version 5.0). The resistant index (RI₅₀) was the quotient of the mean IC₅₀ values of the two strains with C12Rx and the WT being the numerator and denominator, respectively. Drug interaction profiles were determined using published methods using Combenefit®.⁵⁵ Data are mean ± SD for three independent experiments carried out in triplicate.

Conflicts of interest

There are no conflicts to declare.