Evaluation of a Fluorescence-Based Method for Antibabesial Drug Screening

Abstract

In vitro evaluation of chemotherapeutic agents against Babesia and Theileria parasites has become routine, and the effectiveness of these chemicals is usually determined by comparing the parasitemia dynamics of untreated and treated parasites. Although microscopy is widely used to calculate parasitemia, several disadvantages are associated with this technique. The present study evaluated a fluorescence-based method using SYBR green I stain (SG I) to screen antibabesial agents in in vitro cultures of Babesia bovis. The linearity between relative fluorescence units (RFU) and parasitemia was found to be well correlated with a 0.9944 goodness-of-fit (r²) value. Subsequently, 50% inhibitory concentration (IC₅₀) values were calculated for 3 antiprotozoan agents, diminazene aceturate, nimbolide, and gedunin, by this method. For diminazene aceturate and nimbolide, the IC₅₀s determined by the fluorescence-based method (408 nM and 8.13 μM, respectively) and microscopy (400.3 nM and 9.4 μM, respectively) were in agreement. Furthermore, the IC₅₀ of gedunin determined by the fluorescence-based method (19 μM) was similar to the recently described microscopy-based value (21.7 μM) for B. bovis. Additionally, the Z′ factor (0.80 to 0.90), signal-to-noise (S/N) ratio (44.15 to 87.64), coefficient of variation at the maximum signal (%CVmax) (0.50 to 2.85), and coefficient of variation at the minimum signal (%CVmin) (1.23 to 2.21) calculated for the fluorescence method using diminazene aceturate were comparable to those previously determined in malaria research for this assay. These findings suggest that the fluorescence-based method might be useful for antibabesial drug screening and may have potential to be developed into a high-throughput screening (HTS) assay.

INTRODUCTION

Babesiosis, which is caused by Babesia parasites, has a deleterious effect on the health status of cattle (1). Early diagnosis and treatment with effective antiprotozoan agents are essential for the fast recovery of affected animals (2, 3). The development of drug resistance to and toxic side effects of the drugs currently available for treating babesiosis indicate the importance of introducing new alternative treatment options in veterinary practice (4). Potential novel antibabesial agents are usually evaluated in vitro before clinical trials are conducted. At present, the growth-inhibiting effects of the drug candidates are determined based on parasitemia dynamics, which usually involves microscopic examination of Giemsa-stained thin blood smears of treated and untreated parasites (5, 6). However, microscopy that requires good-quality smears is a time-consuming technique and is unsuitable for mass screening of potential drug candidates. In addition, significant differences in parasitemia estimated by different personnel may be found.

Incorporation of radioactive substances, such as [³H]hypoxanthine, [³H]ethanolamine, and [³H]isoleucine, has been a common method in in vitro antimalarial drug screening (7, 8). This method is slowly being replaced by a fluorescence-based method because the former is costly, requires numerous procedures, and is constrained by problems with disposing of radioactive materials (9). Recently, a fluorescence-based method using SYBR green I (SG I), which binds to the double-stranded DNA of parasites, was employed for antimalarial drug screening (10). For a technique to be eligible for use as a high-throughput screening (HTS) assay, the value of the Z′ factor, which is a statistical parameter that represents the separation of the distributions between the positive and negative controls in drug screenings, should be more than 0.5 (11). Past studies related to antimalarial drug evaluations which demonstrated Z′-factor values ranging from 0.73 to 0.95 suggested that the fluorescence-based method might be developed into an HTS assay (9). In the present study, we evaluated the fluorescence method using SG I to monitor the in vitro growth dynamics of Babesia bovis. Subsequently, this method was used to screen three antiprotozoan agents, diminazene aceturate (2), gedunin (12), and nimbolide (13), against B. bovis. In addition, the possibility of developing a fluorescence-based HTS assay for antibabesial drug screening was assessed based on the Z′ factor and other relevant parameters.

MATERIALS AND METHODS

Reagents.

SG I nucleic acid stain purchased from Lonza was stored at −20°C as a 10,000× stock solution and freshly thawed before use. A lysis buffer consisting of Tris (20 mM; pH 7.5), EDTA (5 mM), saponin (0.008% [wt/vol]), and Triton X-100 (0.08% [vol/vol]) was prepared in advance and stored at 4°C. Diminazene aceturate (Novartis, Japan), gedunin (Tocris Bioscience, United Kingdom), and nimbolide (BioVision) were prepared as 10 mM, 50 mM, and 50 mM stock solutions, respectively.

In vitro cultivation of Babesia bovis.

B. bovis (Texas strain) was maintained in bovine red blood cells (RBCs) by using a microaerophilic, stationary-phase culture system as described previously (14). Briefly, medium 199 supplemented with 40% bovine serum, 60 U/ml of penicillin G, 60 μg/ml of streptomycin, and 0.15 μg/ml of amphotericin B (Sigma-Aldrich) was used for the in vitro cultivation of B. bovis. The culture plates, which contain the medium, parasite-infected red blood cells (iRBCs), and normal bovine RBCs, were incubated at 37°C in an atmosphere of 5% CO₂, 5% O₂, and 90% N₂. The medium was replaced every day with fresh medium.

Linearity assessment.

To assess the linearity between fluorescence values and parasitemia, B. bovis-infected RBCs (iRBCs) were serially diluted with uninfected RBCs to adjust the parasitemias to a range of 0.25% to 8% (9). Uninfected RBCs were used as a control. A thin blood smear was also prepared from each dilution and stained with Giemsa to confirm the parasitemia by microscopy. In a 96-well plate, 100 μl of a lysis buffer containing 2× SG I was added to 100 μl of each dilution of the iRBCs (prepared in M119 medium with 10% hematocrit [HCT]) in triplicate and incubated in a dark place at room temperature for 1 h. Subsequently, the fluorescence values were determined using a fluorescence plate reader (Fluoroskan Ascent; Thermo Labsystems) at 485-nm and 518-nm excitation and emission wavelengths, respectively. Gain values were set to 100. The fluorescence (after subtraction of the background fluorescence for nonparasitized RBCs) and parasitemia values were plotted and analyzed by linear regression.

Antibabesial drug screening by the fluorescence-based method.

The experiment was conducted by using 96-well culture plates. Two hundred microliters of medium, medium with indicated concentrations of drugs, and 10% hematocrit with 1% B. bovis iRBCs or uninfected RBCs as a blank control was loaded into each well in triplicate. Initially, growth inhibition assays (15) were conducted at between 1 and 4 days after drug addition using diminazene aceturate (0.01, 0.1, 0.2, 0.5, 1, 5, and 10 μM) to evaluate whether the drug-induced growth inhibition is reflected by fluorescence signals in a dose-dependent manner. Each concentration was tested in triplicate, and four plates were used in quadruplicate for each experiment. Each day, 100 μl of culture from each well in a single plate was processed as described for linearity assessment to determine the fluorescence values. For the remaining plates on each day, the medium, solvents, or medium with the indicated drug concentrations was replaced with a fresh version. The experiment was repeated three times, and the mean fluorescence values were then plotted (after subtraction of the background fluorescence for nonparasitized RBCs) against the logarithm of drug concentrations.

Subsequently, 50% inhibitory concentration (IC₅₀) values were calculated for diminazene aceturate and nimbolide using both the fluorescence-based method and microscopy. However, the IC₅₀ of gedunin was calculated using only the fluorescence-based method. In vitro growth inhibition assays were conducted as described previously (15). While 0.01, 0.1, 0.25, 0.5, 1, 2.5, 5, 10, and 15 μM concentrations were used for diminazene aceturate, 0.005, 0.01, 0.1, 0.25, 0.5, 1, 5, 10, 25, 50, 100, 200, and 400 μM nimbolide and gedunin were used. Each day, the medium, solvents, or medium with the indicated drug concentrations was replaced. Fluorescence values were determined on the fourth day of the experiments. In addition, for diminazene aceturate and nimbolide, thin blood smears were prepared on the fourth day, stained with Giemsa, and observed under a microscope to calculate the parasitemia. Each experiment was repeated three times. The mean fluorescence values were then plotted (after subtraction of the background fluorescence for nonparasitized RBCs) against the logarithm of drug concentrations to monitor the rate of inhibition. The IC₅₀s were calculated using GraphPad Prism ver.5 (GraphPad Software).

Determination of statistical parameters for HTS assays.

An experiment was designed to assess whether the fluorescence-based technique can be used to develop a high-throughput screening (HTS) assay as described previously (16). Briefly, nine samples with 10% hematocrit (HCT) and 1% parasitemia (B. bovis) were prepared using medium (positive control, 100% growth) or medium with a supralethal dose (20 μM) of diminazene aceturate (negative control, 0% growth). The cultures were incubated under appropriate conditions, and the medium or medium containing diminazene aceturate was replaced daily. Fluorescence values were determined after the fourth day, and a previously proposed formula $[Z\mathit{\prime }=1-\frac{(3{\sigma }_{c+}+3{\sigma }_{c-})}{|{\mu }_{c+}-{\mu }_{c-}|}]$ was used to determine the Z′ factor (11). The Z′ factor was calculated using control data, where σ_c+ and σ_c− were the standard deviations of the positive and negative controls, respectively. The denominator represents the absolute value of the difference between the means of the positive and negative controls. In addition, the signal-to-noise (S/N) ratio, coefficient of variation at the maximum signal (%CVmax, positive control), and coefficient of variation at the minimum signal (%CVmin, negative control) were calculated.

Statistical analysis.

The mean IC₅₀s determined by the fluorescence method and microscopy for each drug tested were analyzed by an unpaired t test using the GraphPad software available online. P < 0.05 was considered to be statistically significant.

RESULTS

In the assessment of linearity, a positive correlation was found between the fluorescence value and parasitemia with a 0.9944 goodness of fit (r²) (Fig. 1). As a next step, we investigated whether the drug-induced growth inhibition could be correlated with the fluorescence signals emitted. The results confirmed that the emission of fluorescence signals inversely correlated with the concentration of diminazene aceturate (Fig. 2A). Subsequently, the fluorescence method and microscopy were used to determine the IC₅₀s of two antiprotozoan agents, diminazene aceturate and gedunin. For diminazene aceturate, the IC₅₀s determined by fluorescence (408 ± 11.04 nM) and microscopy (400.3 ± 17.7 nM) were in agreement (P > 0.05) (Fig. 2B and C and Table 1). Similarly, although the IC₅₀s determined for nimbolide by fluorescence (8.13 ± 1 μM) were slightly lower than those determined by microscopy (9.4 ± 0.85 μM) (Fig. 3), the difference was not statistically supported. Furthermore, the IC₅₀ of gedunin was calculated using only the fluorescence-based method (Fig. 4). The IC₅₀ (19 ± 5.1 μM) was comparable to a previously published microscopy-based value (21.72 ± 6 μM) (17), and the difference between these values was not statistically significant.

FIG 1.

Assessment of linearity between fluorescence readings and percent parasitemia of B. bovis-infected RBCs. Each value is presented as the mean of triplicate measurements after subtraction of the background fluorescence for nonparasitized RBCs.

FIG 2.

Fluorescence-based monitoring of diminazene aceturate-induced growth inhibition of B. bovis. (A) Fluorescence-based monitoring of drug-induced growth inhibition during 4 days of treatment. (B) Correlation between diminazene aceturate concentrations and RFU on the fourth day of treatment. (C) Growth inhibition estimated by the fluorescence-based method (black) and microscopy (gray). Each value represents the mean ± standard deviation (SD) of the results determined with triplicate wells after subtraction of the background fluorescence for nonparasitized RBCs.

TABLE 1.

IC₅₀s of antibabesial drugs evaluated in the present study

Drugs	IC50a		P valueb
	Fluorescence-based method	Microscopy
Diminazene aceturate	408 ± 11.04 nM	400.3 ± 17.7 nM	0.5574
Nimbolide	8.13 ± 1 μM	9.4 ± 0.85 μM	0.1690
Gedunin	19 ± 5.1 μM	21.7 ± 6 μMc	0.6590

^aIC₅₀s for each drug were calculated based on the levels of growth inhibition determined by microscopy and the fluorescence-based method in three separate experiments.

^bThe differences between the microscopy- and fluorescence-based IC₅₀s were not statically significant for all three drugs analyzed.

^cData are from reference 17.

FIG 3.

Fluorescence-based monitoring of nimbolide-induced growth inhibition of B. bovis on the fourth day of treatment. (A) Correlation between nimbolide concentrations and RFU. (B) Growth inhibition estimated by the fluorescence method (black) and microscopy (gray). Each value represents the mean ± SD of the results determined with triplicate wells after subtraction of the background fluorescence for nonparasitized RBCs.

FIG 4.

Fluorescence-based monitoring of gedunin-induced growth inhibition of B. bovis on the fourth day of treatment. (A) Correlation between the gedunin concentrations and RFU. (B) Growth inhibition estimated by the fluorescence-based method. Each value represents the mean ± SD of the results determined with triplicate wells after subtraction of the background fluorescence for nonparasitized RBCs.

The Z′ factor for the fluorescence-based method was calculated to range from 0.80 to 0.90. In addition, the S/N ratios ranged from 44.15 to 87.64, values which are 10- to 20-fold higher than those calculated for malaria parasites (16). On the other hand, the ranges for the %CVmax and %CVmin values calculated in the present study were 0.5 to 2.85 and 1.23 to 2.21, respectively.

DISCUSSION

The present report describes the potential application of the fluorescence-based method for antibabesial drug screening using B. bovis, which is thought to be the most virulent Babesia parasite to cause clinical babesiosis among cattle (18). In preliminary experiments, linearity between the fluorescence signals emitted from SG I-treated B. bovis parasites and the microscopy-based parasitemia was observed with high r² values. As a next step, we confirmed the negative correlation between drug-induced growth inhibition and fluorescence values. Consequently, the fluorescence method was further evaluated using antiprotozoan agents to assess the potential application of this assay in antibabesial drug screening.

The IC₅₀s calculated by both the fluorescence method and microscopy were in agreement for diminazene aceturate and nimbolide, suggesting the potential use of the former method to analyze the efficacy of antibabesial agents in vitro. Several disadvantages associated with microscopy in determining the IC₅₀s can be overcome by the fluorescence method. For instance, the microscopy-based results might provide inconsistent IC₅₀s, as the accuracy of these values can be influenced by the individuals who determine the parasitemia. However, the fluorescence values remain the same despite being measured by different researchers. The IC₅₀ for gedunin determined by the fluorescence method was comparable to a previously published microscopy-based value (17), confirming that the fluorescence method can generate IC₅₀s similar to previously published microscopy-based IC₅₀s.

Since we knew that the fluorescence method might be useful in drug screening, the method was further evaluated to determine whether the technique could be used to develop an HTS assay. A high Z′-factor value obtained in the present study suggests that the fluorescence-based assay is of excellent quality, since the value is more than 0.5 (11). The values also agreed with the Z′-factor values (0.73 to 0.95) obtained in malaria research (9). The S/N ratio of the fluorescence method was higher than that determined for its malaria counterpart, indicating that more-reliable signals can be obtained in Babesia research. The low %CVmax value obtained in the present study suggests that the fluorescence-based method might be capable of providing reproducible results (16). The %CVmin value being lower that of Plasmodium parasites might have been due to the relatively higher background signals (16). The use of RBCs with relatively higher HCT might explain the differences between the Plasmodium and Babesia drug-screening assays in the %CVmin values. Although it would obviously be advantageous to adopt 2% HCT, which was recommended for P. falciparum, the present study used 10% HCT, as it is the standard in antibabesial drug screening (6, 19). In the current study, although we employed the previously described protocols to perform the experiments (9), an updated version of the methodology is now available (http://www.wwarn.org/sites/default/files/INV08_PFalciparumDrugSensitivity.pdf) for Plasmodium parasites (20). There are significant differences between the older and newer versions in the composition of lysis buffer, the concentration of SG I, and the duration of incubation before reading the fluorescence signals. Therefore, an optimization of the fluorescence method for antibabesial drug screening is now a priority in Babesia research to improve the quality of the assay. Although we have successfully evaluated the fluorescence assay using 3 different antibabesial agents, large numbers of drugs should essentially be analyzed to determine the suitability of this method for an HTS assay. In addition, evaluation of the fluorescence method for in vitro drug screening against other Babesia and Theileria parasites, such as B. bigemina, B. caballi, and T. equi, is very important, as the present study focused only on B. bovis. Nevertheless, the present findings, in general, suggest that the fluorescence-based method might have potential to be developed into an HTS assay for antibabesial drug screening. Clearly, further evaluation and optimization are required before this technique can be used for large-scale antibabesial drug screening.