Sleeping sickness, or Human African Tropanosomiasis (HAT), is a disease caused by two species of trypanosomes, Trypanosoma brucei rhodesiense and Trypanosoma brucei gambiense. The disease occurs in 36 countries in sub Saharan Africa. In 2006, there was an estimated cumulative rate of 50000-70000 cases worldwide 1. The disease occurs in two stages, the first whereby trypanosomes multiply in the blood and lymphatic system, and includes symptoms such as headaches, fever, joints pain and itching. In the second stage, trypanosomes cross the blood brain barrier to the central nervous system and neurological symptoms often become apparent. Without treatment of the infection, the disease becomes fatal.

HAT chemotherapy relies upon a limited number of drugs, which are not effective against all stages or trypanosome species, have associated side effects or impractical administration regimes. Two of these drugs, suramin and pentamidine, are only effective against the first stage of disease, and are selective against T.b.rhodesiense and T.b.gambiense, respectively. Melarsoprol, which is effective against the second stage of the disease, causes encephalopathy in 5-10% of treated patients 2. Eflornithine, also a second stage specific drug, is less toxic however is not effective against T.b.rhodesiense 3. Suramin, pentamidine, melarsoprol and eflornithine all require intravenous injection for treatment which is not practical considering the poor medical facilities in most of the disease endemic areas. There also has been an increase in melarsoprol-refractory sleeping sickness cases which suggests possible emergence of melarsoprol-resistant trypanosomes 45.

The strong need for new, less toxic HAT drugs, which are particularly effective against the second stage of the disease has resulted in a variety of approaches to identify new treatments which are not uncommon approaches within anti-parasitic drug discovery. These options include combinations of existing HAT drugs (such as eflornithine and melarsoprol; 6), new indications for existing drugs for unrelated disorders and improvements to a number of known drugs and compound classes 7. High throughput screening (HTS) of whole parasites against large compound libraries has also been undertaken in the search for new therapeutic candidates 8. The application of HTS in drug discovery has been explored in neglected disease area for some time, as this approach has the potential of identifying new drugs with novel modes of action. Thus, HTS remains an important consideration for HAT drug discovery programmes.

HTS of specific targets have been reported in the literature for T.brucei species. Assays in 384-well format have been developed to detect trypanothione reductase activity 910, however these do not involve the whole parasite. There have been few reports in the literature for HTS methods for whole cell screening for HAT. A viability assay which has been extensively reported in the literature for all T. brucei species and which utilises the dye Alamar Blue™ has been used a 96-well format by a number of authors 11121314. We have recently developed a 384-well Alamar Blue assay for T.b.brucei whole cell viability estimation 15. As an alternative to the Alamar Blue assay, a luciferase viability assay has recently been published for T.b.brucei whole cell screening in a 96-well format 16. This paper reports the use of a bioluminescent ATP detection assay for HTS for a screening set of 2000 compounds. The same author has also reported the use of this assay for a smaller study of protease inhibitors against T.b.brucei 17. It would be of benefit, in terms of time and cost considerations, to develop such a viability assay for T.b.brucei whole cells in a 384-well format for HTS. T.b.brucei has been routinely used in screening for initial identification of anti-trypanosomal compounds for potential anti-HAT drugs 7. T.b.brucei causes trypanosomiasis in cattle 18 and is non-infective to humans 19. For identification of lead compounds following HTS, actives are subsequently screened against human infective HAT strains for confirmation of activity.

The luciferase based viability assay described here utilizes firefly luciferase to detect intracellular ATP levels in viable cells. The reaction involves the mono-oxygenation of luciferin catalysed by luciferase in the presence of ATP, magnesium and molecular oxygen, the reaction yielding a luminescent signal. This assay has previously been used for determining viability of various cell types, and numerous kits are commercially available from a number of different suppliers. These kits, such as CellTiter- Glo^®, are quick and efficient, with only 10 minutes required for signal development. In comparison, the Alamar Blue assay has a 6-24 hour incubation period for T.b.brucei cell viability assays reported in a 96-well format 5111213. Although many published 96 well assay formats employ Alamar Blue development at 5% CO₂ and 37°C inclusive of a total assay incubation time, we recently found these conditions to reduce the growth of T.b.brucei during 384 well plate assay optimisation 15. We used a 24 hour incubation at room temperature with Alamar Blue following 72 hours of cell growth, for further signal development, as cell growth was reduced under these conditions (40 hours doubling time). We also found assay sensitivity was not significantly affected by this extra incubation. In this case, turn around would be faster for the luciferase assay as the development of signal is more rapid.

Here we report the development of a 384-well whole cell luciferase based viability assay for the detection of the viability of T.b.brucei whole cells. This assay was developed to HTS standards by incorporating the following statistical parameters: Z' factor (Z'), % coefficient of variation (% CV) and the signal to background assay window to reflect reproducibility. The Z' is a measurement of the reproducibility of an assay as described by Zhang and others 20. This value is an indicator of the quality of an assay by analysing the variation of samples of positive and negative controls (no test compounds), and application of the use of this parameter to define the reproducibility of assays has been reported extensively 21. An ideal assay would have a Z' of 1. A Z' of 0 would define an assay that could not be applied to screening due to poor reproducibility 22 and therefore would require further optimisation. The % CV denotes variation of an assay signal relative to the mean. It can be applied to one control variable rather than requiring both negative and positive controls. Generally, a Z' of greater than 0.5 with a % CV of less than 10 is acceptable for HTS. A % CV of greater than 17% indicates that the Z' will not reach 0.5 23 and the closer to 0 the % CV is, the more reproducible, and therefore suitable an assay will be for HTS.

Optimisation of the established assay into 384-well format incorporated analysis of assay variables such as cell concentration, sensitivity to compound solvent, compound dilution medium and the IC₅₀ values of a panel of reference compounds reported in the literature. The reference compounds used in this study consisted of the known drugs, pentamidine, and diminazene and a general kinase inhibitor, staurosporine. Diminazene and pentamidine have reported IC₅₀ values from various publications which employ the Alamar Blue reagent for viability detection 111213. Diminazene is a registered veterinary drug used against T.b.brucei. Staurosporine has reported activity against T.b.brucei 16, as well as other trypanosome species, such as Trypanosoma cruzi 24 and Leishmania species 25.

The luciferase based ATP detection assay described here satisfies HTS assay criteria with a Z' of greater than 0.5. The assay is comparable with respect to known reference compound activities as those reported in the literature for the Alamar Blue assay in format 111213, as well as the reported luciferase based assay in a 96-well format 16. The parameters that were explored and optimised to provide a HTS assay in 384-well format are discussed in detail.

A number of assay formats have been used to screen trypanosome species including ³H-hypoxanthine incorporation, nuclear staining techniques, flow cytofluorometry and measurement of culture pH 27. The Alamar Blue assay was developed to improve associated assay variables such as long incubations, costly reagents and poor assay reproducibility/signal 14 and the assay has been extensively reported in the literature 511121314. More recently, a luminescence assay to detect intracellular ATP has been used in 96-well format 1617. Such an assay could be an alternative assay for HTS to the commonly used Alamar Blue assay. When comparing the feasibility of using these assays one point is the necessity to add Alamar Blue under sterile conditions, which may impact on potential robotic manipulations. Some HTS robotic systems may not be under sterile conditions and therefore the luciferase assay may be favoured to decrease manual work associated with the end point addition. However, the luciferase assay requires plate shaking for cell lysis and not all HTS systems are capable of this step. Both assays are, however, commercially available and this allows access to these reagents for many laboratories for comparative development, dependant upon the capability of the laboratory equipment at hand.

We have described a luciferase based ATP detection assay for estimation of T.b.brucei viability in a 384-well plate for a detection of a small volume of trypanosomes (55 μL) compared with previously employed 96-well protocols which utilised a sample volume of 100 μL 1617. The assay described here is reproducible, conforms to HTS standards with a Z' of 0.65 (Figure 6B), and has similar IC₅₀ values for known T.b.brucei reference compounds as previously reported Alamar Blue assays, a luminescent viability assay (Table 3) and also to a recently reported Alamar Blue assay in 384-well format 15.

Variables that were explored to determine the best conditions to use in the 384-well assay included cell concentration, volume of detection agent, DMSO sensitivity and the compound diluent. The concentration of the cell inoculum to provide optimal cell growth and detection under our experimental conditions was found to be 500 cells/mL. For cell growth, an inoculum of 2000 cells/mL produced a maximum cell number of 3 × 10⁶ cells/mL following 72 hours incubation. In comparison, an inoculum of 3000 cells/mL resulted in a decrease in the final cell number (Figure 4). This would be most likely attributed to the fact that the culture is in decline, with the occurrence of observed cell death. We have previously seen this decline in cultures exceeding over an approximated density of 3 × 10⁶ cells/mL 15. This density maximum density had also previously been determined from other studies for other T.brucei bloodstream species to be 3 × 10⁶cells/mL 28. The doubling time reported here was calculated to be on average 6.8 hours which is also similar to previously reported doubling times for other T. brucei bloodstream species in HMI-9 medium 28.

A cell inoculum of 2000 cells/mL produced the maximum cell number (Figure 4, just over 3 × 10⁶ cells/mL) and this cell concentration was synonymous with a slight decrease in the luciferase signal in comparison to 1 or 2 × 10⁶ cells/mL with cells directly transferred from a flask (Figures 2 and 3). The signal was also reduced in comparison to cell inocula of 1000 and 500 cells/mL grown in a plate (Figure 5). As cell number was at maximum from cells grown from an inoculum of 2000 cells/mL, this signal decrease could be a function of cells reaching a stationary phase of growth and therefore the ATP production decreasing. This phase related decline in ATP production has been observed in other cell lines 293031 and is synonymous with a high cell density in fibroblast cell culture 32. Cell density has been shown to regulate trypanosome number in infections by differentiation induced by stumpy induction factor at high concentrations 33, and it could be suggested that other metabolites may be produced at higher cell densities which effect cultures in monomorphic forms. As cells detected directly from a flask culture exhibited a decrease in CPM at 3 × 10⁶ cells/mL in comparison to 2 × 10⁶ cells/mL (diluted from a 3 × 10⁶ cells/mL culture density; Figure 2), there may have also been interference from ATPases released from the maximum concentration of lysed cells, reducing the signal. This signal interference was not a function of the maximum ATP that could be detected in the well, as this signal plateaued at much higher CPM (Figure 1).

Titrating down to lower densities of trypanosomes revealed a linear relationship between cell number and signal between 1 × 10³ to 1 × 10⁶ cells/mL trypanosomes from cells directly transferred from a flask grown culture (Figure 3B). Mackey and colleagues 16 found a linear relationship between trypanosome concentration and luciferase signal to 6 × 10⁴ cells/well. As the inoculum for this assay was 1 × 10⁵ cells/mL, with a doubling time estimated at 6.8 hours for example, the final assay concentration of cells following 48 hours incubation would have reached over maximum, at 3 × 10⁶ cells/mL, or 3 × 10⁵ cells/well, so it can not be determined if a similar effect was seen in this assay with the concentrations examined for a linear trend.

With plate inoculated cultures, the cell number that produced the greatest signal window was 500 cells/mL (Figure 5). This inoculum resulted in approximately 1.7 × 10⁶ cells/mL following 72 hours growth in plate (Figure 4). This suggested a slightly faster doubling time in the plate, as forecast by the doubling time equation this cell inoculum should be under 1 × 10⁶ cells/mL. However, at this low cell inoculum, there is not a linear trend in growth as we have previously observed 15, and therefore the doubling time would not be expected to be constant. Relating this concentration to the linearity of cell number versus CPM from direct cell transfer, suggested that this cell concentration was slightly above the linear range of the assay, which was at a maximum of 1 × 10⁶ cells/mL (Figure 3B). Taking this into consideration, it was determined that reduction of the cell number to less than 1 × 10⁶ cells/mL final in the assay, or an inoculum of 250 cells/mL of trypanosomes, could be applied for HTS. The DMSO tolerability for 250 compared with 500 cells/mL was explored at both a 0.417% and 0.208% final DMSO concentration. The assay incorporating a 500 cells/mL inoculum at 0.208% DMSO resulted in the least variability, reflected in a Z' of 0.64 (Figure 6B). However, a 250 cells/mL inoculum resulted in a Z' of less than 0.5 for both DMSO concentrations and therefore was not acceptable for HTS application. The DMSO concentration would need to be further reduced to allow this inoculum to be applied, which would bring about potential library concerns as samples would have to be diluted further before screening, therefore potentially reducing the sensitivity of the assay at this cell density. Mackey and colleagues 16 employed a DMSO concentration of 1% at the stage of cell addition in a 96-well luciferase assay. The 384-well assay would have an estimated cell number at the time of addition of DMSO of around 345 cells/well (estimating a doubling time of 6.8 hours) in comparison to an estimated 1 × 10⁴ cells/well in the 96-well format which could explain the differences in the necessity for a lower DMSO requirement for 384-well screening using the HTS conditions as outlined here.

Table 3 shows reference compound activities for both reported Alamar Blue assays and a reported luciferase based ATP detection assay in 96-well format. We found that a 500 cell/mL inoculum resulted in IC₅₀ concentrations that were comparable to those found in the literature from dosing compounds with differing trypanosome inocula. From Table 1, increasing the cell density decreased the sensitivity to both diminazene and staurosporine, when making comparisons of inoculums of 500 cells/mL with 2000 cells/mL. There was a 3.7 and 1.8 times reduction in the IC₅₀values of diminazene and staurosporine, respectively. A 3.3 times reduction in activity was also revealed for pentamidine (results not shown). It is not possible to directly compare the Alamar Blue assay formats to the assay reported here due to the detection reagents used, incubation time, media differences and cell inoculation variables between assays. However, a similar sensitivity was observed with results obtained by Lanteri and collegues 11, with an IC₅₀ value for pentamidine of 10 nM and diminazene 90 nM reported, compared to those observed, pentamidine 11.91 nM and diminazene 60.88 nM. The sensitivity to staurosporine observed using a 500 cells/mL inoculum (an IC₅₀ of 1.94 nM) was comparable to the 96-well format previously reported 16 (IC₅₀ of 7 nM). As there may have been some interference with a higher cell density used by these authors in comparison to the protocol reported here, and as staurosporine has a high Hill Slope (> 3 over all cell doses), further experiments would further determine sensitivities between these assay formats with other known anti-trypanosomals.

Importantly, the assay inoculum employed in the HTS assay developed fits HTS statistical criteria The assay is within the linear range of the detection of the reader (Figure 1), and approximately 0.7 × 10⁶ cells/mL from the linear range of detection of cells estimated in the assay system (Figure 3B). The assay at the defined cell inoculum of 500 cells/mL has sensitivity to known compounds similar to published reference compound activity in other assay formats (Table 3).

The potential diluent used for the addition of compounds and/or controls to the assay was explored, as pre-dilution of DMSO, due to DMSO sensitivity, was necessary. During a HTS campaign it is not uncommon for the screening library to be pre-diluted into the assay plate prior to HTS commencing. Exclusion of FCS from such a dilution medium was considered as this may reduce the risk of certain contaminating microorganisms 34, considering that there is a brief exposure of plates to the atmosphere whilst the dilution is carried out. FCS has also been found to influence cellular toxicity due to components binding to inhibitor compounds 35 and therefore dilutions and storage in a medium with FCS may potentially influence compound activity. Thus, FCS was not included in the diluents used here. The use of water as a diluent increased the IC₅₀ value of the reference compounds by 7.6, 2.3 and 2.4 times, respectively for pentamidine, diminazene and staurosporine in comparison to medium without FCS (Table 2). There was also a decrease in the cell number by 1.5 times with the water addition (results not shown). This could have been as a result of a slight difference in the buffering capacity of the system or an osmotic effect of water on the cells, potentially causing cell lysis. For the optimal 384-well assay performance, HMI-9 medium with exclusion of the FCS component was the diluent of choice, as it did not significantly change the IC₅₀ value of the reference compounds.

We have reported here the successful development and optimisation of a luciferase based cell viability assay in 384-well format for use in HTS of T.b.brucei BS427, which exhibits benefits compared with previously published assay methodologies. Notably the assay is reproducible, fits statistical parameters, with a similar sensitivity to known drugs in comparison with previously published results and is cheaper than the currently used 96-well assay format for HTS. The assay described here is approximately three times less expensive per well than the 96-well luciferase assay currently reported in the literature. Importantly, the assay is biologically relevant, and has been shown statistically to perform well, adhering to the Z' criteria imposed (cut off of 0.5, but closest to 1 as possible). Additions of assay reagents and cells were made with automated liquid handlers, allowing for rapid additions for HTS and handling of a large plate numbers.

To our knowledge this is the first report of a 384-well cell viability assay for T.b.brucei utilising cellular ATP estimation, with a luciferase based detection agent. A future considerations is the comparison of the luciferase based 384-well assay to the Alamar Blue based 384 well assay 15 for HTS application with a larger panel of compounds.

The authors declare that they have no competing interests.

MLS carried out the experimental work and drafted the manuscript. MLS participated in the design of the study and performed data analysis. VMA conceived the study, participated in its design and coordination and contributed to the manuscript preparation. All authors read and approved the final manuscript.