Corresponding Author Timothy M. E. Davis, School of Medicine and Pharmacology, Fremantle Hospital, University of Western Australia, PO Box 480, Fremantle, WA 6959, Australia. Tel.: +618 9431 3229; Fax: +618 9431 2977; E-mail: email@example.com
Objective Recent clinical studies have shown high rates of malaria treatment failure in endemic areas of Papua New Guinea (PNG), necessitating a change of treatment from chloroquine (CQ) or amodiaquine (AQ) plus sulphadoxine-pyrimethamine to the artemisinin combination therapy (ACT) artemether plus lumefantrine (LM). To facilitate the monitoring of antimalarial drug resistance in this setting, we assessed the in vitro sensitivity of Plasmodium falciparum isolates from Madang Province.
Methods A validated colorimetric lactate dehydrogenase assay was used to assess growth inhibition of 64 P. falciparum isolates in the presence of nine conventional or novel antimalarial drugs [CQ, AQ, monodesethyl-amodiaquine (DAQ), piperaquine (PQ), naphthoquine (NQ), mefloquine (MQ), LM, dihydroartemisinin and azithromycin (AZ)].
Results The geometric mean (95% confidence interval) concentration required to inhibit parasite growth by 50% (IC50) was 167 (141–197) nm for CQ, and 82% of strains were resistant (threshold 100 nm), consistent with near-fixation of the CQ resistance-associated pfcrt allele in PNG. Except for AZ [8.351 (5.418–12.871) nm], the geometric mean IC50 for the other drugs was <20 nm. There were strong associations between the IC50s of 4-aminoquinoline (CQ, AQ, DAQ and NQ), bisquinoline (PQ) and aryl aminoalcohol (MQ) compounds suggesting cross-resistance, but LM IC50 only correlated with that of MQ.
Conclusions Most PNG isolates are resistant to CQ in vitro but not to other ACT partner drugs. The non-isotopic semi-automated high-throughput nature of the Plasmodium lactate dehydrogenase assay facilitates the convenient serial assessment of local parasite sensitivity, so that emerging resistance can be identified with relative confidence at an early stage.
Sensibilité in vitro de Plasmodium falciparum aux médicaments antimalariques classiques et nouveaux en Papouasie-Nouvelle-Guinée
Objectif: De récentes études cliniques ont montré des taux élevés d’échec du traitement de la malaria dans les zones endémiques de la Papouasie-Nouvelle-Guinée (PNG), nécessitant un changement du traitement à la chloroquine (CQ) ou l’amodiaquine (AQ) plus la sulfadoxine-pyriméthamine vers la thérapie de combinaison à l’artémisinine (ACT), artéméther plus luméfantrine (LM). Afin de faciliter la surveillance de la résistance aux médicaments antimalariques dans cette région, nous avons évalué la sensibilité in vitro d’isolats de Plasmodium falciparum provenant de la province de Madang.
Méthodes: Un dosage colorimétrique validé, basé sur la lactate déshydrogénase a été utilisé pour évaluer l’inhibition de la croissance de 64 isolats de P. falciparum en présence de neuf médicaments antimalariques classiques ou nouveaux (CQ, AQ, monodesethyl-amodiaquine (DAQ), pipéraquine (PQ), naphthoquine (NQ), méfloquine (MQ), LM, dihydroartémisinine et azithromycine).
Résultats: La moyenne géométrique [intervalle de confiance 95%] de la concentration nécessaire pour inhiber de 50% la croissance du parasite (CI50) était de 167 nM [141-197] pour CQ et 82% des souches étaient résistantes (seuil de 100 nM), ce qui est cohérent avec une quasi fixation de la résistance à la CQ associée à l’allèle pfcrt en PNG. A l’exception de l’azithromycine (8351 nM [5,418-12,871]), la moyenne géométrique de CI50 pour les autres médicaments était < 20 nM. Il y avait de fortes associations entre la CI50 des composés 4-aminoquinoline (CQ, AQ, DAQ et NQ), bisquinoline (PQ) et aminoalcool aryle (MQ), suggérant une résistance croisée, mais la CI50 de LM corrélait seulement avec celle de LF.
Conclusions: La plupart des isolats de PNG sont résistants à la CQ in vitro mais pas aux autres médicaments partenaires de l’ACT. La nature semi-automatique, non-isotopiques et à haut débit du test pLDH facilite l’évaluation convenable de la sensibilité des parasites locaux afin qu’une résistance émergente puisse être identifiées avec une relative confiance à un stade précoce.
Susceptibilidad in vitro de Plasmodium falciparum frente a antimaláricos convencionales y nuevos en Papua Nueva Guinea
Objetivos: Estudios clínicos recientes han demostrado altas tasas de fallo terapéutico de la malaria en áreas endémicas de Papua New Guinea (PNG), requiriendo un cambio de tratamiento de la cloroquina (CQ) o amodiaquina (AQ) más sulfadoxina pirimetamina a terapias de combinación basadas en artemisinina (TCA) artemeter más lumefantrina (LM). Con el fin de facilitar la monitorización de la resistencia a antimaláricos en estos lugares, hemos evaluado la susceptibilidad in vitro de aislados de Plasmodium falciparum en la provincia de Madang.
Métodos: Se utilizó un ensayo colorimétrico de lactato deshidrogenasa validado para evaluar la inhibición del crecimiento de 64 aislados de P. falciparum en presencia de nueve medicamentos antimaláricos convencionales o nuevos (CQ, AQ, monodesetilamodiaquina (DAQ), piperaquina (PQ), naftoquina(NQ), mefloquina (MQ), LM, dihidroartemisinina y azitromicina).
Resultados: La media geométrica [intervalo de confianza del 95%] de la concentración requerida para inhibir el crecimiento de los parásitos en un 50% (IC50) era 167 [141-197] nM para CQ y 82% de las cepas eran resistentes (umbral 100 nM), de manera consistente con la casi fijación del alelo de pfcrt asociado a la resistencia de CQ en PNG. Excepto para la azitromicina (8,351 [5,418-12,871] nM), la media geométrica IC50 de todos los demás medicamentos era <20 nM. Había una fuerte asociación entre la IC50s de 4-aminoquinolina (CQ, AQ, DAQ y NQ), bisquinolina (PQ) y compuestos de aryl aminoalcohol (MQ) sugiriendo una resistencia cruzada, pero el IC50 de LM solo se correlacionaba con el de la MQ.
Conclusiones: La mayoría de los aislados de PNG eran resistentes a CQ in vitro pero no a otros medicamentos incluidos en terapias de combinación. La naturaleza no-isotópica, semi-automatizada y de alto rendimiento de los ensayos pLDH facilitan la evaluación en serie de la susceptibilidad de parásitos locales de manera que la aparición de resistencias podría identificarse con relativa seguridad en estadios tempranos.
The present study utilised blood samples taken in 2006 and 2007 from children aged 6 months to 10 years as part of clinical studies conducted at the Alexishafen Health Centre in Madang province and at Modilon Hospital in Madang town. In all cases, informed consent was obtained from the parents or legal guardians before recruitment and blood sampling. Scientific and ethical approval for each study was obtained from the Medical Research and Advisory Committee of the Ministry of Health of PNG. Of the 64 samples collected, 45 (70.3%) were from children recruited to a randomised trial of uncomplicated malaria (Karunajeewa et al. 2008). The remaining 18 samples (29.7%) were from pharmacokinetic studies conducted at Alexishafen or studies of severe malaria in progress at Modilon Hospital.
In vitro parasite culture
Venous blood collected into heparin-containing tubes was centrifuged, and both plasma and buffy coat were removed. The packed erythrocytes were washed twice in RPMI 1640 culture medium, and blood smears were prepared for the confirmation of P. falciparum monoinfection and quantification of parasitaemia by microscopy. Parasites were cultured using a modified candle jar method (Trager & Jensen 1976). Briefly, parasites were maintained at 37 °C in a low-oxygen atmosphere in type O human erythrocytes from a non-immune individual at 5% haematocrit in RPMI 1640 media (pH 7.3) (Gibco, Invitrogen Life Technologies, Auckland, New Zealand) supplemented with 0.5% w/v Albumax II (Gibco), 25 mm NaHCO3 (Sigma-Aldrich, St Louis, MO, USA), 37.5 mm HEPES (Sigma-Aldrich), 10 μg/ml gentamycin sulphate and 100 μg/ml neomycin sulphate (Calbiochem, Darmstadt, Germany) and 45 μg/ml hypoxanthine (Sigma-Aldrich).
Drug susceptibility assays
Stock solutions of CQ diphosphate (Sigma Chemicals, St Louis, USA), AQ dihydrochloride (Sigma), monodesethyl-amodiaquine (DAQ) (Sapec Fine Chemicals, Lugano, Switzerland), piperaquine (PQ) tetraphosphate (Yick-Vic Chemicals and Pharmaceuticals, Hong Kong), naphthoquine phosphate (NQ) (ZYF Pharm Chemical, Shanghai, China), mefloquine hydrochloride (MQ) (Sigma), LM (Novartis Pharma, Basel, Switzerland), dihydroartemisinin (DHA) (Sigma) and azithromycin (AZ) (Pfizer, NSW, Australia) were prepared in distilled water (CQ, AQ, DAQ and AZ), 0.5% w/v lactic acid (PQ), 50% v/v ethanol (NQ), 70% v/v ethanol (MQ and DHA), or in a 1:1:1 v/v/v mixture of linoleic acid, ethanol and Tween 80 (LM). Each stock solution was stored in light-proof tubes at −20 °C. On the day of assay, aliquots were thawed and further diluted in RPMI to a working standard, and further twofold serial dilutions in complete RPMI at double assay concentrations were prepared for CQ (25–1600 nm), AQ and DAQ (5–320 nm), NQ (3.13–200 nm), PQ (6.25–400 nm), MQ (0.78–200 nm), DHA (0.78–51.2 nm), LM (3.12–400 nm) and AZ (1.25–160 μm).
Drug sensitivity was assessed in triplicate in 96-well plates, with each well containing 100 μl drug-containing media and 100 μl parasitised erythrocyte suspension at 0.5–1.0% parasitaemia and a final haematocrit of 1.5%. With the exception of AZ, which was incubated for 72 h because of its slower antimalarial activity (Noedl et al. 2006), all other plates were incubated for 48 h at 37 °C. The plates were then subjected to four freeze-thaw cycles to achieve complete haemolysis. The haemolysates were kept frozen until assayed. A modification of a Plasmodium lactate dehydrogenase (pLDH) detection method was used to assess parasite growth (Makler & Hinrichs 1993; Makler et al. 1993). Briefly, 10 μl sample haemolysate was added to 200 μl of Malstat solution (Trisma base 1.21 g in 90 ml d.H2O with pH adjusted to 9.1), 200 μl Triton X-100 (Merck, Victoria, Australia), 2 g Lithium-L-lactate, 62 mg 3-acetyl-pyridine-adenine-dinucelotide (Sigma)), 10 μl nitro blue tetrazolium (10 mg/ml) (Sigma-Aldrich) and 10 μl diaphorase (10 mg/ml) (Sigma). Colour intensity was developed at room temperature for 45 min to 2 h and measured spectrophotometrically at 650 nm.
The pLDH assay was compared to the reference 3H-hypoxanthine incorporation method (Chulay et al. 1983) using the culture-adapted CQ-sensitive and CQ-resistant strains of P. falciparum 3D7, W2mef and E8B and a panel of antimalarial drugs. Briefly, we prepared two sets of CQ, MQ and DHA drug dilutions in either complete media or complete media without hypoxanthine as appropriate for the pLDH and 3H-hypoxanthine assays, respectively. To allow for between-day variability in assay performance, the drug sensitivity of each P. falciparum strain was assessed simultaneously using the two methods.
Statistical analysis were performed using graphpad prism version 4.0 (GraphPad Software, CA) and Microsoft Excel for Windows. The concentration of drug required to inhibit parasite growth by 50% (IC50) and 90% (IC90) for each antimalarial drug as measured by pLDH assay was determined by non-linear regression analysis of logarithmically transformed dose–response curves using HN NonLin v.1.1, a free tool for malaria in vitro drug sensitivity analysis (Noedl 2002). Comparisons between the IC50 values in laboratory-adapted strains obtained by pLDH and 3H-hypoxanthine incorporation assays were made using regression analysis (Bablok et al. 1988) and the Bland and Altman (1986) method. Associations between IC50 and IC90 values of drug pairs were assessed using Spearman’s rank correlation coefficient. Because of the number of comparisons, a significant P-value <0.05 was used throughout. Correlations between the concentrations of CQ, PQ and LM inhibiting 50% of parasite growth (IC50) in a subset of the present patients have been reported previously (Karunajeewa et al. 2008).
The pLDH assay was compared to the reference 3H-hypoxanthine incorporation method in cultured-adapted P. falciparum. There was a significant linear correlation between the data obtained by the two IC50 assay methods (r2 = 0.97, n =26; P =0.001), with a slope and intercept [(95% confidence intervals)] of 1.13 (1.00 to 1.25) and 7.83 (−3.3 to 18.98) nm, respectively (Figure 1). The Bland–Altman plot showed that the pLDH assay sometimes significantly underestimated the IC50 at high values and provided the least reliable estimations at IC50s >200 nm (Figure 2).
Of 125 field isolates obtained, 64 (51%) were both cultured successfully and provided valid drug sensitivity data by pLDH assay. Loss of isolates reflected mainly logistic issues with transportation, and laboratory-related problems including power supply reliability, reagent availability and bacterial contamination. Experiments involving LM, DHA and AZ were only undertaken during latter part of the study period. The mean IC50 and IC90 values for the nine antimalarial drugs are summarised in Table 1. The accepted in vitro resistance threshold for CQ (≥100 nm) is based on the in vitro response of African isolates obtained from malaria-infected, non-immune individuals taking CQ prophylaxis and semi-immune patients failing to respond to CQ treatment (Le Bras & Ringwald 1990; Cremer et al. 1995; Ringwald et al. 1996; Basco et al. 2002). This threshold was obtained using the reference 3H-hypoxanthine incorporation method but has been employed in many subsequent studies using different techniques for in vitro drug susceptibility assessment (Attlmayr et al. 2006; Kaddouri et al. 2006; Mayxay et al. 2007; Nkhoma et al. 2007). Because there was close agreement between the pLDH and 3H-hypoxanthine methods in the present study at this concentration, we also used a 100 nm cut-point, with 82% of isolates tested exhibiting an IC50 above this level. In vitro resistance thresholds for the other antimalarial drugs tested have not been established through valid correlative in vivo studies.
Table 1. In vitro susceptibilities of Plasmodium falciparum isolates against 4-aminoquinolines and other antimalarial drugs
Number of isolates tested
IC50 (nm) Geometric mean (95% CI)
IC90 (nm) Geometric mean (95% CI)
IC50, 50% inhibitory drug concentrations; IC90, 90% inhibitory drug concentrations; CI, confidence interval.
Correlations between IC50 values for the panel of antimalarial drugs are given in Table 2. Apart from LM, there were strong associations between the IC50s of 4-aminoquinoline (CQ, AQ, DAQ and NQ), bisquinoline (PQ) and aryl aminoalcohol (MQ) compounds. Although the numbers of isolates tested were low, AZ activity did not correlate significantly with that of any other drug. The artemisinin derivative DHA IC50 values showed positive associations with those of the 4-aminoquinoline and related compounds.
Table 2. Spearman’s correlation coefficients for associations between IC50 values
The number of drug pairs analysed is given in parentheses. *P <0.05, **P <0.01 and ***P <0.001.
The IC50s for AQ and its active metabolite DAQ in the present study were much lower than in previous studies from PNG (Trenholme et al. 1993; al-Yaman et al. 1996). This is likely to reflect methodological differences, because microscopic assessment of schizont maturation produces IC50 values several times higher than those derived from radioisotope incorporation (Wernsdorfer & McGregor 1988) and, by implication, from our pLDH assay. However, consistent with the present IC50 data, a recent study from neighbouring East Sepik Province found a much lower prevalence of in vitro resistance to AQ than CQ (Genton et al. 2005).
A valid in vitro resistance threshold for PQ remains to be confirmed. An IC50 < 100 nm has been used to identify sensitive strains of P. falciparum by radioisotope uptake (Deloron et al. 1985; Basco & Ringwald 2003), while Chinese investigators have reported resistant isolates with IC50 values >300 nm using the schizont maturation microtechnique (Yang et al. 1999b; Lin et al. 2005). All the present PQ IC50 values were <100 nm. Sixteen of the isolates were from children treated with DHA–PQ in our clinical trial (Karunajeewa et al. 2008), and one (IC50 19.5 nm) was a late parasitological failure. These data suggest that further in vivo–in vitro correlation studies are needed to establish a meaningful PQ resistance threshold.
MQ has not been used previously in PNG. All our isolates had MQ IC50 values below the resistance threshold of 108 nm established recently using in vivo responses, 3H-hypoxanthine uptake and molecular characteristics including P. falciparum multidrug resistance 1 (pfmdr1) gene copy number (Price et al. 2004). This threshold was also employed in a study of field isolates from Laos which were assessed using the pLDH method (Mayxay et al. 2007). Our in vitro MQ data are consistent with the results of a recent survey that reported an absence of multiple copies of pfmdr1 gene in PNG isolates (Hodel et al. 2008).
Our data are the first characterising the in vitro sensitivity of PNG isolates to LM and NQ, drugs that have both recently become available as part of ACT in PNG. A previously published resistance threshold for LM of >150 nm was based on 3H-hypoxanthine uptake studies in African isolates without in vivo correlation (Basco et al. 1998). None of our isolates had an IC50 value above this level. There are no published cut-points for NQ, but our median IC50 (10.3 nm) was less than those reported for Southern Chinese isolates using the micro-test method (mean 88.5 nm for ‘artesunate-sensitive’ and 119.4 nm for ‘artesunate-resistant’ strains) (Yang et al. 1999a). Our DHA IC50 values were all below the suggested cut-point of 10.5 nm derived from 3H-hypoxanthine uptake studies in African isolates without in vivo correlation (Pradines et al. 1998). The AZ IC50s from our study are also largely below those of Thai isolates assessed using the micro-test technique (means 13.9 μm and 29.3 μm, respectively) (Noedl et al. 2001).
The present study provides baseline data at a time when the treatment of uncomplicated malaria in PNG will change from AQ–SP or CQ–SP to artemether–LM as a result of clinical trial data (Karunajeewa et al. 2008). Ongoing assessment of in vitro sensitivity using the same techniques will facilitate the assessment of the adequacy of such treatment. Convention monitoring involves the WHO micro-test with labour-intensive visual enumeration of schizonts (al-Yaman et al. 1996; Hombhanje 1998). The colorimetric pLDH assay allows prompt semi-automated generation of parasite growth data from triplicate experiments involving multiple antimalarial drugs. The IC50 values generated correlate well with those derived using 3H-hypoxanthine incorporation, and there is no issue with disposal of radioisotopes, an important consideration in countries such as PNG. Assays based on pLDH quantification have been recently introduced for screening patient isolates against multiple antimalarial drugs in Africa and Asia (Brockman et al. 2004; Kaddouri et al. 2006; Nkhoma et al. 2007). Rational drug policy in countries such as PNG can only benefit from such convenient, high-throughput in vitro testing, especially if this is performed regularly, so that emerging resistance can be identified with relative confidence at an early stage.
We are most grateful to Dr Pascal Michon, John Taime, Manasseh Baea, Mak David, the nurses and staff at the PNG Institute of Medical Research at Yagaum, Sister Valsi Kurian and staff at the Alexishafen Health Centre, Michelle England and Dr Jane Allan from the School of Medicine and Pharmacology, University of Western Australia for technical, clinical and/or logistic assistance. This study was supported by the National Health and Medical Research Council of Australia (grant #353663) and the World Health Organisation. TMED is supported by a National Health and Medical Research Council of Australia Practitioner Fellowship.