SYBRGreen I^® (S33102) and PicoGreen^® (P7589) were purchased to Invitrogen and diluted as indicated by the manufacturer in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5) or TE plus different combinations of saponin and Triton X-100 detergents to obtain a range of concentrations from 1000 ng/mL to 5 ng/mL. Standard curves of DNA were performed by diluting bacteriophage λ DNA, provided at 100 μg/mL in the Quant-iT™ PicoGreen^® Kits (Molecular Probes™, Invitrogen) with DNase-free water. One hundred microliters of each mixture were dispensed into black 96-well microplate (Costar^®), followed by the addition of 100 μl of either PG or SGI. Plates were incubated in the dark at room temperature for 5 minutes and the fluorescence intensity was measured at 485 nm excitation and 528 nm emission using a Perkin Elmer LS-50B Luminescence Spectrophotometer. The mean background reading from the two control wells was subtracted from the test reading to give the final fluorescence measurement. Fluorescence units (arbitrary units ranging from 1 to 1,000 units) was plotted against DNA concentration.

Fluorescence measurements in the presence of haemoglobin was carried out by washing uninfected human blood in RPMI1640 medium (final 50% haematocrit) and treated with Lymphoprep™ (AXIS-SHIELD) as indicated by manufacturer to remove most of the WBC (white blood cells). Erythrocytes were resuspended in culture medium 22 to get a haematocrit of 1-5%, and lysed by freeze-thawing. Each sample was spiked with lambda DNA at a final concentration of 1 μg/mL. 100 μL of each mixture were added to 100 μL of either SGI or PG and the fluorescence signal was determined as described above. In addition, the absorption spectrum from red blood cells lysates was evaluated. All measurements were performed in triplicate from three processed samples independently

Chloroquine resistant strain Dd2 of P. falciparum (clone MRA-150, 23 was used for this study. The strain was cultivated in vitro as described previously 22. The culture media consisted of standard RPMI 1640 (Sigma-Aldrich) supplemented with 0.5% ALBUMAX I (Gibco), 100 μM hypoxanthine (Sigma-Aldrich), 25 mM HEPES (Sigma-Aldrich), 12.5 μg/mL gentamicine (Sigma-Aldrich) and 25 mM NaHCO₃ (Sigma-Aldrich). Erythrocytes were obtained from type A+ human healthy local donors and collected in tubes with citrate-phosphate-dextrose anticoagulant (VACUETTE^®), and washed as described above. Each culture was started by mixing uninfected and infected erythrocytes to achieve a haematocrit at 1% and incubated in 5% CO2 at 37°C in tissue culture flasks (Iwaki). The progress of growth in the culture was determined by microscopic counting of parasitaemia in thin blood smears with Wright's eosin methylene blue solution (Merck) using the freely available "Plasmoscore" software 24. The cultures were synchronized by incubation with sorbitol as described 25.

Synchronized rings from stock cultures were used to test 5 μM to 1 nM serial dilutions of chloroquine in 96-well culture microplates. Thus, 150 μL of parasites at 2% haematocrit and 1% parasitaemia were allowed to grow for 48-hour in 5% CO₂ at 37°C. Parasites were then centrifuged at 600 × g for 10 minutes and the re-suspended in saponin to lyse the erythrocytes. The pellet was washed to remove haemoglobin as described 20: 150 μL of each sample were dispensed in 96-well microplates, which were centrifuged at 600 × g for 10 minutes. The supernatant was removed and the cells were re-suspended in saponin (0.15%, w/v in phosphate-buffered saline (PBS) to lyse the erythrocytes and release the malaria parasites. The microplates were immediately centrifuged at 600 × g for 10 minutes, and the supernatant was removed. To eliminate all traces of haemoglobin the pellet was washed by the addition of 200 μL of PBS followed by centrifugation at 600 × g. The washing step was repeated twice to ensure complete removal of haemoglobin. Finally, pellets were re-suspended in 100 μL of PBS. 100 μL of either PG or SGI diluted in TE were added to each well. Then, the plates were incubated for 30-60 minutes in the dark. The fluorescence signal was determined as described above. Fluorescence Unit (AU) was calculated by subtracting the control negative (fluorescence from uninfected erythrocytes) 13. AU was converted to percentage (%) of survival in the fluorimetric drug susceptibility test by the following equation:

Where, Fd is the fluorescence from drug treated infected erythrocytes, Fcp is the fluorescence control positive: fluorescence from non-treated infected erythrocytes, and Fcn is fluorescence control negative: fluorescence from uninfected erythrocytes. The logarithm of the chloroquine concentration was plotted against the corresponding percentage (%) of survival, and the 50% inhibitory concentration (IC₅₀) was calculated by non-linear regression. All measurements were performed in triplicate from two processed samples independently. The curve fitting was performed using Sigmaplot^©10 from Systat Software, Inc.

A serial dilution of the synchronized cultures of P. falciparum (mature ring stage) was performed with non-parasitized erythrocytes and complete medium to yield a haematocrit of 2% and parasitaemia levels ranging from 0.05% to 15%. Haemoglobin-depleted samples were obtained as indicated above. In these assays, 100 μL of either SGI or PG were added to each microplate well and the fluorescence signal was determined. To analyse the effect of TritonX-100, a set of haemoglobin-depleted samples were mixed with SGI or PG plus TritonX-100 (2% v:v final concentration). Control samples without haemoglobin removal were analysed in parallel. In this case, the parasitized erythrocytes were lysed by using the lysis/fluorescence-mix, consisting of either SGI, 0.008% saponin and 0.08% Triton X-100, or PG with 2% Triton X-100. Fluorescence readings were plotted against parasitaemia.

Assays on whole blood were carried out by diluting a sorbitol-synchronized culture with a donor peripheral blood to achieve different levels of parasitaemia from 15% to 0.05%. Two series of dilutions were performed: the first series was prepared at 50% haematocrit and 400 μL were added to 200 μL Lymphoprep™ to remove WBC. The resulting sediment was washed and diluted with phosphate buffered saline to obtain 2% final haematocrit. The second series was directly diluted to 2% haematocrit without WBC removal. Haemoglobin was depleted both from Lymphoprep-treated and non-treated samples, and PG-DNA fluorescence was determined as above.

All measurements were performed in triplicate from three processed samples independently. The relative uncertainty (relative error) was calculated as the standard deviation divided by the mean. The limit of detection (LOD) was established as the lowest parasitaemia yielding fluorescent signal significantly different from the blank control. The limits quantification (LOQ) corresponds to the lowest parasitaemia that can be quantified with acceptable accuracy and precision. LOD and LOQ were calculated as indicated by the International Conference on Harmonization (ICH) guidelines 26 by using:

Where s_B is the standard deviation of the blank sample, m is the slope of the calibration curve and t = 3.3 is the Student's t for a 95% confidence level. Statistics and curve fitting was performed using Sigmaplot^©10 from Systat Software, Inc.

Several fluorimetric-based protocols make use of different combinations of saponin and TritonX-100 as detergents for the combined erythrocyte lysis/fluorescence detection in parasitaemia measurements. To ascertain the effect of these detergents in SGI and PG assays, DNA standard curves were prepared in the presence of different detergents currently used in lysis/fluorescent mixes. The results showed a direct relationship between DNA concentration and fluorescence (Figure 1). Higher signal was detected for PG (706 ± 33 AU per 1 μg/mL) as compared to SGI (141 ± 40 AU per 1 μg/mL) in TE. The presence of detergents changed the slope of the standard curves for both dyes. All combinations tested generated a decrease in the fluorescent signal, with the exception of PG 2% Triton X-100 which displayed a slight increment. The mixture of 0.008% saponin and 0.08% Triton X-100 appears to have a smaller effect of both dyes (<10% in SGI, <2.5% for PG), while these detergents added separately showed intermediate results.

The effect of haemoglobin on fluorescent readings was tested by mixing different amounts of lysed erythrocytes (1-5% haematocrit) with a fixed amount of λ DNA. Erythrocytes were lysed by freeze-thawing to avoid the use of detergents which, as shown above, might interfere with the fluorescence measurements.

The fluorescence emission of both SGI- and PG-DNA adducts, shown in Figure 2A, was highly quenched in presence of red cells. Control samples, containing DNA without haemoglobin, rendered fluorescence readings around 30-60 fold as compared with samples containing blood, suggesting a heavy interference of haemoglobin. Detection of fluorescence in presence of blood, even at the lowest content, was only possible by opening the optical slit of the detector to 15 nm. Under these conditions the fluorescence emission from 1 μg/mL DNA samples in 0% haematocrit was rendered off scale.

As shown in Figure 2B the absorption spectrum of red blood cell lysates revealed the existence of two peaks of maxima absorption at 485 nm and 529 nm, which overlap with the excitation/emission wavelengths of SGI and PG, and may therefore explain the quenching of haemoglobin on the fluorescent readings.

To test if the signal increase observed after depletion of haemoglobin may lead to changes in the calculated IC₅₀ values of current anti-malarial drugs, dose-response curves for chloroquine were obtained using either PG or SGI. No significant differences in the IC₅₀ values were observed for chloroquine in fluorimetric anti-malarial drug assays run in parallel (p > 0.05) (Figure 3). The IC₅₀ for chloroquine against strain Dd2 of P. falciparum was 136.2 ± 28 nM when using PG and 118 ± 15 nM for SGI. These values were similar to those reported by other authors using different techniques 152728. Chloroquine is a known DNA intercalating agent 2930, and it has been shown that it may lead to 23% PG exclusion in the range of 0.5 to 1 mM 31. The possible effect of the drug on the binding of the fluorescent dye to DNA at concentrations used in the IC50 assay was tested by comparison of the DNA-PG calibration curves in the presence of chloroquine at 1 μM and 5 μM. No significant differences in the slope or intercept were observed as compared to the control curve without drug (p > 0.05, data not shown), suggesting that chloroquine do not interfere with PG binding to DNA well above the highest concentration used.

As shown above, the choice of DNA-fluorescent dye and the removal of haemoglobin before fluorescent readings may enhance the sensitivity of the assay. In addition, the presence of 2% Triton X-100 showed an increase in the fluorescent signal of PG-DNA adducts, while other detergent formulations may reduce that signal (Figure 1). To analyse the effect of haemoglobin depletion and the addition of 2% Triton X-100 on the quantification of Plasmodium parasitaemia, the fluorescence obtained in cultures with or without haemoglobin removal was compared. While samples containing haemoglobin were lysed by adding a fluorescent mix containing detergents, haemoglobin-depleted samples were first lysed with saponin, washed and resuspended in PBS, and the fluorescent mix was added without detergents, avoiding thus any interference due to these compounds. The effect of Triton X-100 was tested only on haemoglobin-depleted samples, by mixing PG or SGI with TritonX-100 2%.

The results, shown in the Figure 4, confirm the linear correlation between the percentage of infected red blood cells and the fluorescence. As expected, PG exhibited much higher fluorescence emission than SGI (20-fold at 15% parasitaemia). SGI results showed a higher interference of TritonX-100 by decreasing both the slope (2.2 versus 1.4 AU/%parasitaemia) and the correlation of the linear response (r² > 0.99 to r² > 0.97). Best results were obtained with PG in samples devoid of haemoglobin (r² > 0.99, m = 46.6): for a 15% parasitaemia culture, fluorescent readings were in the range of 650-700 vs. 11 (arbitrary fluorescence units) in non-depleted haemoglobin. The presence of 2% TritonX-100 showed a much smaller effect, with no clear gain or loss on the signal.

The fluorescence of highly diluted samples of a P. falciparum culture using SGI or PG under these conditions (no haemoglobin, no detergent) is shown in Table 1. While SGI was not able to detect a parasite load below 0.6%, PG yielded fluorescent readings down to 0.05% and showed a 7-fold higher signal than SGI at 0.6% parasitaemia, leading to a similar reduction of the calculated LOD from 1.4% to 0.2%. Similarly, a 6-fold reduction of the LOQ was observed when using PG as compared to SGI.

Once established the PG haemoglobin/detergent-depleted method as the most appropriate for low-parasitaemia samples, its capability for detection and quantification of Plasmodium in whole blood samples, as those coming from clinical isolates, was tested. A synchronized P. falciparum culture was diluted with human peripheral blood to obtain parasitaemia ranging from 15% to 0.05% (IRBC+WBC samples). Each dilution was split in two and one of them was treated with Lymphoprep™ to remove WBC (IRBC+WBC/W). Haemoglobin was depleted from both samples and fluorescence was measured with PG. The results showed a good correlation of fluorescence versus parasitaemia (r² > 0.99) in both conditions. Data at parasitaemia in the range of 5% to 0.05% are shown in the Table 1, indicated that the method can detect down to 0.2% parasitaemia even in the presence of WBC DNA, although the high variability observed in these samples may compromise the accuracy of results. To solve this problem, an additional wash step was included in the protocol to remove WBC prior to the fluorescent reading. The Figure 5 shows a comparison of the relative uncertainty of the measurements obtained from IRBC, IRBC+WBC and IRBC+WBC/W samples at low parasitaemia. While the relative uncertainties were very low (<0.15) for parasitaemias >1.5% in all samples, only IRBC and IRBC+WBC/W samples showed relative uncertainties <0.3 at the LOD and LOQ values: Removal of WBC reduced the relative error from 0.5 to 0.12 at 0.2% parasitaemia, and from 0.2 to 0.04 at 0.6% parasitaemia. These results suggest that white cell removal should be performed to achieve the highest accuracy at the LOD and LOQ range.

The possibility of increasing the sensitivity by using higher haematocrit was also explored by measuring fluorescence from IRBC+WBC/W samples at 5% and 10% haematocrit. The results (not shown) displayed a non-linear correlation of fluorescence to parasitaemia, which could be associated to the accumulation of cell-debris after lysis interfering with the fluorescence measurement.