In Vitro Culture and Drug Sensitivity Assay of Plasmodium falciparum with Nonserum Substitute and Acute-Phase Sera

Abstract

The short-term in vitro growth of Plasmodium falciparum parasites in the asexual erythrocytic stage and the in vitro activities of eight standard antimalarial drugs were assessed and compared by using RPMI 1640 medium supplemented with 10% nonimmune human serum, 10% autologous or homologous acute-phase serum, or 0.5% Albumax I (lipid-enriched bovine serum albumin). In general, parasite growth was maximal with autologous (or homologous) serum, followed by Albumax I and nonimmune serum. The 50% inhibitory concentrations (IC₅₀s) varied widely, depending on the serum or serum substitute. The comparison of IC₅₀s between assays with autologous and nonimmune sera showed that monodesethylamodiaquine, halofantrine, pyrimethamine, and cycloguanil had similar IC₅₀s. Although the IC₅₀s of chloroquine, monodesethylamodiaquine, and dihydroartemisinin were similar with Albumax I and autologous sera, the IC₅₀s of all test compounds obtained with Albumax I differed considerably from the corresponding values obtained with nonimmune serum. Our results suggest that Albumax I and autologous and homologous sera from symptomatic, malaria-infected patients may be useful alternative sources of serum for in vitro culture of P. falciparum isolates in the field. However, autologous sera and Albumax I do not seem to be suitable for the standardization of isotopic in vitro assays for all antimalarial drugs.

Cultivation of malaria parasites is an important tool for the understanding of parasite biology, biochemistry, molecular biology, immunology, and pharmacology. One of the applications of parasite cultivation is the in vitro drug sensitivity assay, which is a major tool for the screening of potential antimalarial drugs, the monitoring of drug sensitivity, and the detection of cross-resistance patterns against Plasmodium falciparum parasites (5, 25, 34). Although several assays have been developed, the in vitro cultivation technique of the erythrocytic stages of P. falciparum remains essentially the same as that originally described by Trager and Jensen (29). In this standard technique, the following components are required: P. falciparum-infected human erythrocytes, buffered RPMI 1640 medium, and human serum. The parasitized erythrocytes are incubated in a low-oxygen atmosphere at 37°C.

The culture medium is commercially available at a relatively low cost. The sources of infected blood abound in countries where malaria is endemic. In countries where malaria is not endemic, several reference clones and strains of P. falciparum are available in research laboratories. It has generally been accepted that nonimmune human serum is required for optimal parasite growth. However, the requirement for a regular supply of nonimmune human serum entails difficulties in conducting research in most of the African continent, where malaria transmission occurs at a high level throughout the year. Nonimmune human type AB-positive serum is relatively scarce and expensive in countries where malaria is not endemic. Furthermore, it is recommended that several units of serum from different donors be pooled to reduce batch-to-batch differences in the support of parasite growth (14). Additional problems include blood type compatibility and risks associated with the handling of infectious agents.

Because of these disadvantages, numerous alternative sources of sera and nonserum substitutes have been tested in the past (1, 6, 13, 15, 19, 27, 28, 37). Although some of these substitutes have been successfully used for the continuous cultivation of laboratory-adapted P. falciparum strains and clones, they are generally less effective than human serum and have not been adopted by many research laboratories. In a recent study, Ofulla et al. (21) have identified serum albumin and lipids as the key serum components that are necessary to sustain optimal parasite growth. Further studies have shown that a commercially available lipid-enriched bovine serum albumin, Albumax I (Gibco BRL), can replace human serum for the continuous in vitro cultivation of malaria parasites, leading to its routine use in several laboratories (2, 4, 7, 8, 31).

Another alternative source of serum is malaria parasite-infected patients themselves. It has been thought that semi-immune human serum protects individuals who are continuously exposed to malaria parasites against malarial disease and therefore inhibits parasite growth (14). This assumption has not been proven experimentally through the quantitation and correlation of parasite growth inhibition, the level of acquired immunity, and the presence of antimalarial drugs in sera collected from indigenous populations. In addition, if this hypothesis were true, indigenous adults who have been exposed to the malaria parasite are expected to have acquired immune protection against further attacks of malaria. Contrary to this expectation, in many areas of endemicity in Africa, such as in Yaoundé, Cameroon, symptomatic malarial infection occurs frequently in both adults and children (24). It has also been shown that heat-inactivated serum from healthy, semi-immune African donors can support the growth of laboratory-adapted strains of parasites and fresh isolates and that acute-phase homologous serum may be useful for the continuous in vitro culture of reference strains (2, 20).

A preliminary in vitro study with lipid-enriched bovine serum albumin has reported that serum-free medium can be used instead of nonimmune serum to determine the level of drug activity (22). Autologous and homologous acute-phase sera from malaria parasite-infected patients and Albumax I have not been evaluated as alternatives for in vitro culture with fresh clinical isolates obtained from indigenous patients. In our present study, we (i) assessed the growth of clinical isolates of the malaria parasites in two different RPMI 1640 media during a single life cycle or two life cycles with nonimmune type AB-positive serum, Albumax, autologous acute-phase serum, and homologous acute-phase serum with the aim of determining the best medium for short-term culture and in vitro drug assay and (ii) compared the in vitro activities of various antimalarial compounds against fresh clinical isolates of P. falciparum using different sera or serum substitute with the aim of assessing whether these alternative sources can be used to standardize isotopic in vitro assays.

MATERIALS AND METHODS

Clinical isolates.

Thirty fresh clinical isolates were obtained from symptomatic Cameroonian patients residing in Yaoundé before treatment. Eight were children between 5 and 14 years old; 22 were adults (≥15 years old; age range, 19 to 69 years). Our previous studies have shown that populations of patients in this age range in Yaoundé present with similar clinical and laboratory features (24, 26). The following inclusion criteria were set for this study: presence of signs and symptoms of acute uncomplicated malaria, monoinfection with P. falciparum, parasitemia of >0.2%, and no history of recent antimalarial drug intake confirmed by a negative Saker-Solomons urine test (18). Since this study is part of an ongoing clinical study designed to determine the clinical efficacy of first- and second-line drugs, pregnant women and patients with signs and symptoms of severe and complicated falciparum malaria, as defined by the World Health Organization (WHO) (33), were excluded. After informed consent was obtained, 10 ml of venous blood was collected in a tube coated with anticoagulant (EDTA) and in a tube not coated with anticoagulant. The patients were treated with either amodiaquine or sulfadoxine-pyrimethamine, which are first- and second-line drugs in Cameroon, respectively, and were monitored daily until they were cured. The study was approved by the Cameroonian National Ethics Committee.

Drugs.

The following antimalarial drugs were obtained from the indicated sources: chloroquine sulfate, Rhone-Poulenc-Rorer, Antony, France; monodesethylamodiaquine, a biologically active metabolite of amodiaquine, Sapec S. A., Lugano, Switzerland; quinine hydrochloride, Sigma Chemical Co., St. Louis, Mo.; mefloquine hydrochloride, Hoffmann-La Roche, Basel, Switzerland; halofantrine hydrochloride, SmithKline Beecham, Hertfordshire, United Kingdom; dihydroartemisinin, a biologically active metabolite of artemisinin derivatives; Sapec S. A.; pyrimethamine base, Sigma Chemical Co.; and cycloguanil base, a biologically active metabolite of proguanil, Zeneca Pharma, La Defense, France. Stock solutions of chloroquine, monodesethylamodiaquine, and cycloguanil were prepared in sterile distilled water. Stock solutions of quinine, mefloquine, halofantrine, dihydroartemisinin, and pyrimethamine were prepared in methanol. Twofold serial dilutions (fourfold serial dilutions for pyrimethamine and cycloguanil) of the drugs were made in the same solvent used to prepare the stock solutions. The final concentrations ranged from 25 to 1,600 nmol/liter for chloroquine, 50 to 3,200 nmol/liter for quinine, 5 to 320 nmol/liter for monodesethylamodiaquine, 2.5 to 160 nmol/liter for mefloquine, 0.5 to 32 nmol/liter for halofantrine, 0.25 to 16 nmol/liter for dihydroartemisinin, and 0.09 to 51,200 nmol/liter for pyrimethamine and cycloguanil. Each concentration was distributed in triplicate in 96-well tissue culture plates and air dried.

In vitro assay.

Venous blood samples collected in tubes with anticoagulant were washed three times in p-aminobenzoic acid (PABA) and folic acid-free RPMI 1640 medium. Two types of RPMI 1640 medium were used to cultivate the parasites: the standard RPMI 1640 medium containing 1 mg of PABA per liter and 1 mg of folic acid per liter and PABA- and folic acid-free RPMI 1640 medium. Since PABA and folic acid compete with dihydrofolate reductase and dihydropteroate synthase inhibitors (pyrimethamine, cycloguanil, proguanil, and sulfonamides), PABA- and folic acid-free RPMI 1640 medium was used to evaluate the in vitro activities of these drugs (17, 31, 40). Both RPMI 1640 media were supplemented with HEPES (25 mmol/liter), NaHCO₃ (25 mmol/liter), and gentamicin (10 μg/ml). Infected erythrocytes were suspended in these culture media at a hematocrit of 1.5% and an initial parasitemia of between 0.2 and 0.6%. The following serum or serum substitute was added to obtain the complete media: 10% (vol/vol) nonimmune type AB-positive serum (obtained from French blood donors who resided in France and who had no previous history of malaria), 10% (vol/vol) human serum from malaria-infected patients, and concentrated Albumax I solution (10% [wt/vol]) diluted to a final concentration of 0.5%. The concentration of Albumax I used in this study was shown in a previous study (21) to be optimal for parasite growth. If the blood sample had a parasitemia of >0.6%, fresh uninfected, type A-positive erythrocytes were added to adjust the parasitemia to 0.6%. All experiments were performed within 4 h after blood collection.

The growth of 12 parasite isolates in the presence of either autologous or homologous acute-phase sera was compared in the complete standard RPMI 1640 medium. Only freshly obtained clinical isolates were used in our experiments, and the isolates were used without prior adaptation to in vitro culture. Some of the acute-phase sera were stored at −20°C for several months before use. Freshly drawn sera either were used immediately after collection or were stored at 4°C for up to 3 weeks. For these experiments, the parasitemia (range, 0.2 to 3.0%) was not adjusted; one blood sample with a parasitemia of 10%, however, was diluted with fresh uninfected type A-positive erythrocytes from a healthy donor (1:9 [vol/vol]) to obtain an initial parasitemia of 1%.

The short-term culture technique was based on the isotopic microtest of Desjardins et al. (5). Two hundred microliters of the suspension of infected erythrocytes was distributed in triplicate in 96-well tissue culture plates that were either drug-free or precoated with test compounds. The parasites were incubated at 37°C in 5% CO₂. [³H]hypoxanthine (specific activity, 16.3 Ci/mmol; 1 μCi/well; Amersham, Buckinghamshire, United Kingdom) was added after the first 18 h of incubation to assess parasite growth. The incorporation of [³H]hypoxanthine has been established as an accurate and reliable means of determining in vitro parasite growth (3, 5). After an additional 24 h of incubation (additional 48 h for experiments with PABA- and folic acid-free RPMI 1640 medium and drug assays for pyrimethamine and cycloguanil), the plates were frozen to terminate the assays. The plates were thawed to lyse the infected erythrocytes, and the contents of each well were collected on glass-fiber filter papers, washed, and dried with a cell harvester. The filter disks were transferred into scintillation tubes, and 2 ml of scintillation cocktail (Organic Counting Scintillant; Amersham) was added. The incorporation of [³H]hypoxanthine was quantitated with a liquid scintillation counter (Wallac 1410; Pharmacia, Uppsala, Sweden).

For parasite growth assays, the results either were expressed as counts per minute or were normalized to the growth of parasites in their corresponding autologous sera. The 50% inhibitory concentrations (IC₅₀s), defined as the drug concentration corresponding to 50% of the uptake of [³H]hypoxanthine measured in drug-free control wells, were determined by nonlinear regression analysis with Prism software (GraphPad Software, Inc., San Diego, Calif.). The IC₅₀s determined for isolates cultivated in medium supplemented with different sera or serum substitute were expressed as the mean IC₅₀ ratios. IC₅₀ ratios were calculated for the following: Albumax I/nonimmune serum, Albumax I/autologous acute-phase serum, and autologous acute-phase serum/nonimmune serum. If the mean IC₅₀ ratio between media supplemented with different sera or serum substitute was 0.50 to 1.50, the mean IC₅₀s were considered to be equivalent. The threshold IC₅₀s for in vitro resistance to chloroquine, monodesethylamodiaquine, quinine, mefloquine, halofantrine, cycloguanil, and pyrimethamine were estimated to be >100, >60, >800, >30, >6, >50, and >100 nmol/liter, respectively (25). The level of resistance to dihydroartemisinin is still undetermined.

RESULTS

Nine fresh clinical isolates were tested for in vitro growth in standard RPMI 1640 medium supplemented with nonimmune or autologous serum or Albumax I. All isolates adapted readily to the in vitro conditions, as shown by an adequate incorporation of tritium-labeled hypoxanthine (>3,000 cpm) during a 42-h incubation period (Fig. 1). The growth of most clinical isolates was optimal in the presence of autologous serum; the exceptions were two isolates (isolates 18/98 and 22/98) which grew better with Albumax I than with the autologous serum. One isolate (isolate 14/98) developed equally well with nonimmune serum (51,948 ± 2,530 cpm) and autologous serum (51,788 ± 989 cpm). For other isolates, in the presence of the autologous serum, the incorporation of tritium-labeled hypoxanthine was 1.4 to 5.6 times higher than that obtained with nonimmune serum. A comparison between nonimmune serum and Albumax I showed that three clinical isolates (isolates 13/98, 14/98, and 19/98) grew better (two- to eightfold better) with the nonimmune serum. For four other isolates, an opposite trend was observed, with higher (two- to sixfold) levels of hypoxanthine incorporation in the presence of Albumax I. Two isolates (isolates 6/98 and 16/98) grew almost equally well with nonimmune serum and Albumax I. Similar results were obtained with the isolates grown in PABA- and folic acid-free RPMI 1640 medium supplemented with nonimmune serum, autologous serum, or Albumax I over a 72-h incubation period.

FIG. 1.

Comparison of parasite growth over 42 h in standard RPMI 1640 medium supplemented with pooled nonimmune type AB-positive serum (open squares), Albumax I (black circles), or autologous serum (black squares) assessed by the incorporation of [³H]hypoxanthine. Results represent mean counts per minute (n = 1 to 6 triplicate tests). Bars denote standard deviations.

The assessment of parasite growth in the standard RPMI 1640 medium supplemented with different sets of acute-phase sera showed that the parasites generally grew better with autologous or homologous acute-phase sera than with pooled nonimmune sera (data not shown). The homologous sera supported the growth of fresh isolates to a widely different extent, even surpassing the growth with autologous serum for some isolates. Parasite growth was not influenced by an ABO blood type incompatibility or storage of homologous sera at 4°C or −20°C.

The in vitro drug sensitivity patterns were determined for 11 clinical isolates with nonimmune serum, autologous acute-phase serum, and Albumax I. The differences in the IC₅₀s obtained with different sources of serum or serum substitute were relatively small for chloroquine and monodesethylamodiaquine (Table 1). Nonimmune serum gave the lowest IC₅₀ for chloroquine and monodesethylamodiaquine, generally followed by autologous serum and Albumax I. Wide variations in the IC₅₀s of the other test compounds were observed. The IC₅₀s of quinine and mefloquine differed considerably (nonimmune serum < Albumax I < autologous serum), with generally greater than a 2-fold difference between nonimmune serum and Albumax and greater than a 10-fold difference between nonimmune serum and autologous serum for quinine. The differences in the IC₅₀s of mefloquine were less pronounced. The order of the IC₅₀s of halofantrine was as follows: Albumax I < nonimmune serum < autologous serum. The IC₅₀s of dihydroartemisinin, pyrimethamine, and cycloguanil generally varied according to the following order: nonimmune serum ≤ autologous serum < Albumax I.

TABLE 1.

Drug sensitivity patterns of P. falciparum isolates obtained with nonimmune or autologous sera or serum substitute (Albumax I)

Isolate	IC50 (nmol/liter)a
	Chloroquine	Amodiaquine	Quinine	Mefloquine	Halofantrine	Artemisinin	Pyrimethamine	Cycloguanil
06/98	24.6	12.5	107	7.41	1.21	1.06	1.75	1.21
	54.8	27.6	170	16.5	0.76	13.5	4.62	4.17
	45.9		1,068		5.26	9.22	1.16	1.05
09/98	24.2	13.9	47.5	17.4	0.87	2.25	2.04	1.33
	31.0	18.4	87.7	9.45	<0.5	8.71	2.08	1.92
	30.5	13.2	763	59.3	1.62	4.68	1.44	1.11
13/98	271	40.5	187	16.8	0.78	1.53	2,400	252
	410	67.4	978	20.2	<0.5	3.50	7,350	1,320
	303	46.0	1,068	40.7	1.29	1.98	1,230	152
14/98	164	28.5	325	26.1	2.77	2.04	119	8.14
	359	48.6	1,138	30.0	0.53	4.08	989	114
	330		>3,200		5.49	3.98	198	25.5
16/98	37.0		183		5.35	2.29	775	36.2
	63.9		>3,200		1.54	10.2	3,860
	76.1		>3,200		9.86	10.9	2,010
18/98	26.2	13.4	24.8	14.9	4.00	1.25
	28.9	12.5	87.2	12.7	0.52	3.18
	25.7	12.6	925	39.2	5.06	1.92
19/98	253	45.1	203	9.71	2.30	1.03
		57.9		13.4			5,300
		48.0		46.1
20/98	120	16.5	233	4.68	0.95	0.93
		45.8		14.7			0.91
		23.1		19.6
21/98	30.5		277		4.36	2.86	70.6	25.8
							1,290	414
							84.8	22.1
22/98	158	19.7	247	4.28	2.9	0.25	2.46
		32.1		11.6			123
		22.6		16.7
24/98	382			11.6			3,710	333
							49,800	1,880
							3,360	396

^aFor each drug, the IC₅₀s obtained for isolates with nonimmune human AB-positive serum (10% [vol/vol]), Albumax I (0.5% [wt/vol]), and autologous serum (10% [vol/vol]) are given in order (i.e., for each isolate, the top, middle, and bottom rows of data correspond to the IC₅₀s obtained with nonimmune human AB-positive serum, Albumax I, and autologous serum, respectively). Amodiaquine and artemisinin refer to their biologically active human metabolites monodesethylamodiaquine and dihydroartemisinin, respectively. 

The IC₅₀ ratios are summarized in Table 2. According to our criterion of an equivalent IC₅₀, defined as an IC₅₀ ratio of between 0.50 and 1.50, the mean IC₅₀s obtained with Albumax I were not equivalent to any of the IC₅₀s obtained with nonimmune serum. However, equivalent IC₅₀s were obtained with Albumax I and autologous sera for chloroquine (IC₅₀ ratio, 1.10 ± 0.17) and monodesethylamodiaquine (IC₅₀ ratio, 1.41 ± 0.33). The IC₅₀s of dihydroartemisinin obtained with Albumax I and autologous sera were also similar, but the values were widely dispersed. The mean IC₅₀s of monodesethylamodiaquine were equivalent with nonimmune and autologous sera (IC₅₀ ratio, 1.10 ± 0.17). The IC₅₀s of halofantrine, pyrimethamine, and cycloguanil were also similar, but the IC₅₀ ratios were widely dispersed.

TABLE 2.

Mean IC₅₀ ratios of antimalarial drugs obtained from nonimmune or autologous acute-phase sera or serum substitute (Albumax I) against fresh clinical isolates of P. falciparum

Antimalarial drug	IC50 ratioa
	Albumax/nonimmune serum	Albumax/autologous serum	Autologous/nonimmune serum
Chloroquine	1.67 ± 0.46	1.10 ± 0.17	1.55 ± 0.48
Monodesethylamodiaquine	1.69 ± 0.58	1.41 ± 0.33	1.10 ± 0.17
Quinineb	5.53 ± 6.01	0.44 ± 0.41	16.08 ± 11.28
Mefloquine	1.65 ± 0.93	0.45 ± 0.24	3.55 ± 0.90
Halofantrine	0.41 ± 0.23	0.20 ± 0.12	1.26 ± 1.10
Dihydroartemisinin	4.65 ± 4.07	1.45 ± 3.38	3.38 ± 2.88
Pyrimethamine	12.70 ± 16.14	6.91 ± 5.77	1.18 ± 0.74
Cycloguanil	7.62 ± 5.93	7.06 ± 6.16	1.25 ± 0.94

^aIC₅₀ ratios are mean ± standard deviation (n = 6; n = 7 for pyrimethamine) ratios of IC₅₀s obtained with media supplemented with different sera or serum substitute (Albumax I). 

^bThe quinine IC₅₀ ratios are approximate since no growth inhibition was achieved in one assay with Albumax I and two assays with autologous sera. In these assays, the highest concentration in the assay (3,200 nmol/liter) was used to calculate the estimated mean ratio. 

DISCUSSION

In previous studies, laboratory-adapted P. falciparum strains and fresh clinical isolates were successfully adapted and maintained in continuous culture with heat-inactivated semi-immune plasma or serum from African donors with no history of malarial infection in the preceding 3 weeks (2, 20). Our results further extend this observation and demonstrate that autologous and homologous acute-phase sera from malaria-infected patients also support the growth of fresh clinical isolates and often do so better than nonimmune sera. Thus, contrary to the unconfirmed assumption that local sera in areas of endemicity are unsuitable for parasite cultivation because they may contain antimalarial drugs and ill-defined immune factors (14), our results show that autologous and homologous sera from symptomatic patients may also be useful for parasite culture in the field. For most isolates, the in vitro growth attained maximal levels when the autologous acute-phase serum was added.

Clinical isolates may be readily adapted to short-term culture if the absence of antimalarial drugs is verified by a simple urine test (18). Our success rate in performing in vitro assays with nonimmune serum and fresh clinical isolates (>500 primary isolates between 1993 and 1997) exceeds 92% in Yaoundé (23, 25). The in vitro studies conducted by Oduola et al. (20) and Binh et al. (2) with semi-immune and acute-phase sera to maintain cultures of laboratory-adapted strains and clones of P. falciparum support our observation that local sera may be useful for parasite cultivation. Our data suggest that both autologous and homologous acute-phase sera support parasite development for one or two life cycles and that the patients’ own acute-phase sera may be the best source of nutrients for the corresponding P. falciparum isolate, at least for short-term cultivation or for the initiation of culture of field isolates.

The use of Albumax I as a serum substitute for a long-term continuous culture has several advantages over the use of human serum. Albumax I costs less than human serum, is compatible with any blood type, and does not have a wide batch-to-batch difference, as is the case with serum. Growth of laboratory-adapted P. falciparum strains was similar in RPMI 1640 medium supplemented with hypoxanthine (0.2 mM) and nonimmune human serum or Albumax (4). Furthermore, Albumax has been used successfully to maintain a continuous culture of several P. falciparum strains in different laboratories (2, 4, 7, 8, 31). However, Albumax may not be suitable for in vitro drug assays since the IC₅₀s of most antimalarial drugs tested in this study were consistently higher with the Albumax-supplemented medium than with the serum-supplemented medium. Using serum-free media containing 5 g of bovine albumin per liter and Cohn fraction V, Ofulla et al. (22) also found that both chloroquine and quinine IC₅₀s are, on average, 1.6 times higher than the values obtained with serum for 14 culture-adapted or fresh isolates. In their study, the ratio of amodiaquine IC₅₀s (serum-free medium versus serum-supplemented medium) was 1.1 for 11 isolates.

The underlying reason for the higher IC₅₀s with Albumax-supplemented media may be due to high levels of protein binding. In vivo, many antimalarial drugs are known to be highly bound to plasma proteins, notably, to albumin, which is the major component of plasma proteins and Albumax I. Quinine (70 to 95% protein binding), mefloquine (>98%), pyrimethamine (87%), and cycloguanil (75%), which had elevated IC₅₀s with Albumax-supplemented medium, are highly protein bound (35). Chloroquine is relatively less protein bound (50 to 70%), while monodesethylamodiaquine is highly protein bound (>90%). Protein binding alone does not explain the widely different IC₅₀s of some antimalarial drugs since the albumin concentration used in our experiments (for Albumax I, 0.5% [wt/vol], and for serum albumin, 10% [vol/vol]; the normal plasma albumin concentration is 35 to 50 g/liter) is similar in the two media.

Several possible factors may influence in the increase or decrease in the IC₅₀s. First, as suggested by Ofulla et al. (22), the differential IC₅₀s may result, at least in part, from a higher affinity of antimalarial drugs for bovine albumin than for human albumin. Second, chloroquine and amodiaquine (highly hydrophilic, weak bases) undergo marked uptake into infected and uninfected erythrocytes (11, 30), which may explain the similar IC₅₀s obtained in our study. Third, albumin binds to lipophilic compounds, such as halofantrine, by means of hydrophobic binding forces, and plasma lipoproteins, notably, triglyceride, influence the IC₅₀s of halofantrine (12). Although the protein binding properties of halofantrine are still unknown due to its low solubility in water, these properties of albumin and lipids may explain the increased transport of halofantrine into infected erythrocytes in Albumax-supplemented medium and thus the lowering of the IC₅₀s.

The standard in vitro test developed by WHO uses a mixture (9:1 [vol/vol]) of RPMI 1640 medium and the patient’s whole blood to determine the level of parasite growth in the presence of different drug concentrations (38). The results are interpreted by microscopic examination of thick blood smears. The results of the WHO in vitro test and those of in vitro assays that are based on the incorporation of tritium-labeled hypoxanthine are not comparable (36). The WHO test determines the maximal inhibitory concentration, while isotopic tests measure the IC₅₀. The former uses acute-phase plasma; the latter test is usually performed with nonimmune donor serum after washing of the infected erythrocytes. To our knowledge, no study has compared the two in vitro tests, but it is assumed that isotopic tests are more objective and accurate in determining the sensitivity levels (36). Although the standard WHO test uses whole blood, in our study autologous acute-phase sera did not yield consistent drug assay results compared with those obtained with nonimmune serum except for the results for monodesethylamodiaquine and, to a lesser extent, halofantrine, pyrimethamine, and cycloguanil. In most cases, the IC₅₀s obtained with autologous serum were increased more than two times compared with those obtained with nonimmune serum. An opposite trend would have been expected from the observation that the level of albumin in plasma is generally lower in malaria-infected patients than in healthy adults (9, 10), which should lead to a higher concentration of unbound drug available for schizontocidal action in the autologous serum. The most discordant result was observed with quinine; the IC₅₀s of quinine obtained with autologous serum were more than 10 times greater than the values obtained with nonimmune serum. The most likely explanation lies in the increased circulating concentration of the acute-phase plasma protein α-1 acid glycoprotein, which increases the level of protein binding of quinine in malaria parasite-infected patients (10, 16, 32). Furthermore, the level of protein binding of drugs is known to vary widely between patients (39). For these reasons, autologous sera are probably not useful for determination of the drug sensitivity pattern. Whether our observation extends to the WHO standard in vitro test that uses autologous plasma needs to be determined.

Our study demonstrates that autologous and homologous sera from patients with acute uncomplicated falciparum malaria are suitable for cultivation of field isolates and that their use results in a high growth rate compared with that obtained with nonimmune pooled sera and Albumax I during the first and second in vitro erythrocytic cycles. Our study further confirms the usefulness of Albumax I for the short-term cultivation of field isolates. Although autologous and homologous sera and Albumax I may serve as alternative sources of serum or serum substitute for the in vitro culture of fresh isolates in the field, at present, none of them seems to be suitable for in vitro drug sensitivity assays with the exception of autologous sera for monodesethylamodiaquine and Albumax I for halofantrine. However, if more data from a larger series of studies comparing Albumax I and nonimmune serum are accumulated, a reliable conversion factor in the form of IC₅₀ ratios can be calculated for antimalarial drugs for which the IC₅₀ ratios in this study were relatively low (chloroquine, monodesethylamodiaquine, mefloquine, and halofantrine). With such a conversion factor, threshold resistance values can be adjusted for Albumax I-supplemented medium. Other serum substitutes that provide adequate nutritional needs for the optimal growth of parasites and that do not bind strongly to antimalarial drugs may be necessary for the standardization of in vitro assays.