As the pathogenicity and virulence of two of the trypanosome subspecies used were known (ILRI-2918 and ILRI-3953), available data from previous studies were used to estimate the sample size per group.^11,18,19,20 Swiss white mice of either sex (n = 108), weighing 25–30 grams, were obtained from the rodent facility at the IPR and provided with mouse feed (Mice cubes®, Unga Feeds, Nairobi, Kenya) and water ad libitum. The mice were kept in 14 cm x 30 cm x 15 cm Macrolone cages. They were maintained at an ambient temperature of between 20 and 25^°C and wood shavings (various timber mills, Embulbul, Kajiado, Kenya) were used as bedding.

Stabilate ILRI-2918 was clonally derived from Kenya Trypanosomiasis Research Institute (KETRI)-2772, which was isolated from a natural infection in a human in Alupe, Busia, Kenya and designated 2772.¹⁸ Stabilate ILRI-3953 was originally referred to as Uganda Trypanosomiasis Organisation (UTRO)-310185, which was isolated from a naturally infected human in Bugiri, Uganda in 1982. The parasite has been passaged in mice and designated ILRI-3953.^19,20 Stabilate IPR-001 was obtained from a cerebrospinal fluid sample from an infected human in Bugiri, Uganda. After three passages in outbred Swiss white mice at the IPR, the stabilate was designated IPR-001.

For the current study, all stabilates were obtained from storage in liquid nitrogen (–196 °C), thawed and diluted with phosphate saline glucose (PSG) buffer (pH 8.0), then injected intraperitoneally (0.2 ml each) into three Swiss white female mice that were immunosuppressed via irradiation with caesium chloride at 600 rads for five minutes. Upon the onset and rise of parasitaemia, the imunosuppressed mice were euthanised and their blood collected via cardiac puncture. The collected blood was then diluted to a parasitaemia of approximately 10⁴ trypanosomes/0.2 mL using PSG, after which the diluted blood was injected intraperitoneally (ip; 0.2 mL each) into the 108 mice included in the study, as described below. Six mice were used as positive controls and injected with PSG buffer only (0.2 mL ip).

The experiment was designed and executed as indicated in Table 1. After infection, experimental mice were observed daily for parasitaemia using wet blood smears obtained by tail snipping. At the same time, thin blood smears were also prepared and Giemsa stained for observation of pleomorphism via microscopy, as described by Kagira et al.¹¹ The parasitaemia was then estimated using the method described by Herbert and Lumsden.²¹

For both ILRI-2918 and IPR-001-infected mice, all surviving mice (both experimentally infected and controls) were treated with diminazene aceturate (40 mg/kg ip, Berenil®, Intervet, South Africa) at 21 days post infection (DPI) as described previously by Kagira et al.¹¹ At this stage, the parasites were assumed to have invaded the CNS. The diminazene aceturate therefore only clears the haemolymphatic trypanosome complement, because it cannot cross the blood-brain barrier in sufficient quantities to clear the CNS. Mice were then observed daily for relapse parasitaemia using wet smears obtained via tail snipping. Relapse detection sensitivity was enhanced by examining the collected blood using the Woo method.²² Upon relapse of parasitaemia, all surviving mice (both experimentally infected and controls) were treated with melarsoprol (3.6 mg/kg ip, Arsobal®, Sanofi-Aventis, Paris, France) for four days consecutively. Six mice per stabilate were euthanised at set time points (DPI 7, 14 and 21). All their major organs (heart, liver, spleen, brain, lungs and kidneys) were harvested and perfused with citrate saline buffer via the left ventricle through the hepatic portal vein until all the blood was completely drained. They were then preserved in 10% neutral buffered formalin for histopathology.

Data were managed and analysed using MS Excel (Microsoft Excel® 2007, Microsoft Corporation, Redmond, Washington, United States). The means for parasitaemia and parasite morphology were calculated using the AVERAGE function in MS Excel and variability was calculated using the standard error of the mean function in Microsoft Excel, ‘= (STDEV(A1:A2))/(SQRT(COUNT(A1:A2))’, where A1 represents the first value in the series and A2 represents the last value in the series.

All experiments were carried out at the IPR in Karen, Nairobi, Kenya. The experiments were conducted in accordance with protocols approved and authorised by the Institutional Review Committee of the Institute.

In ILRI-2918-infected mice, patency was demonstrated at five DPI (Figure 1). This was estimated at Log₁₀ 3.07. All of the patent mice were active and had smooth coats at the onset of patency. First peak parasitaemia was observed at seven DPI (Figure 1) and the mean peak parasitaemia was estimated at Log₁₀ 6.4. Secondary latency was observed at 17 DPI (Figure 1), with mean peak parasitaemia estimated at Log₁₀ 4.87. Following treatment at 21 DPI with diminazene aceturate, parasitaemia dropped below microscopically-detectable levels between 22 and 28 DPI (Figure 1). Relapse parasitaemia was observed on 50 DPI. The parasitaemia at the onset of relapse was Log₁₀ 0.6.

From 10 DPI, the mice began to develop ruffled coats, lower appetite and lower activity levels. These symptoms occurred at varying levels of parasitaemia, with some groups experiencing the symptoms during the rising phase of parasitaemia and others during the dropping phase of the first parasitaemia peak. Peri-orbital oedema also occurred during the rising and dropping phases of parasitaemia, with three mice having one or both eyes shut at all times. These symptoms persisted until the second rising phase of parasitaemia, at which time they disappeared and the mice appeared active and once again had smooth coats. At 26 and 27 DPI, hindleg paresis was observed in two mice. These mice were euthanised and their major organs collected for histopathology. At 27 DPI, after treatment with diminazene aceturate at 21 DPI, the surviving mice again had lowered activity and ruffled coats. The prognosis of these mice, however, improved steadily and the clinical symptoms slowly dissipated, disappearing by 31 DPI. At 50 DPI, all surviving mice relapsed into parasitaemia at Log₁₀ 0.6 and at 51 DPI, the mice succumbed to hindleg paresis and were subsequently euthanised.

In ILRI-3953-infected mice, patency was demonstrated from three DPI, with peak parasitaemia observed at seven DPI (Figure 2). This was estimated at Log₁₀ 8.36. All of the patent mice were active and had smooth coats at the onset of patency. Mice began to develop ruffled coats, lower activity level and paresis from seven DPI. One mouse also exhibited peri-orbital oedema at peak parasitaemia (Log₁₀ 8.4; seven DPI). As the parasitaemia rose, mice developed hindleg paresis and were found dead in their cages, with some deaths occurring at six DPI. By 10 DPI, all experimental mice had been euthanised or were found dead in their cages, necessitating a termination of the experiment on this isolate.

In IPR-001-infected mice, patency was demonstrated from three to five DPI (Figure 3). All of the patent mice were active at the onset of patency. The first peak of parasitaemia was observed at six DPI (Log₁₀ 8.36) and the second peak parasitaemia occurred at 18 DPI (Log₁₀ 8.81). Ruffled coats, lower appetite and lower activity level were observed from as early as the first parasitaemia peak. After treatment with diminazene aceturate, parasitaemia dropped below microscopically-detectable levels and clinical symptoms abated. Relapse parasitaemia occurred at 39 DPI (Log₁₀ 1.62) and peaked at 54 DPI (Log₁₀ 6.8). Following treatment with melarsoprol, parasitaemia was again observed to drop below microscopically-detectable levels until the termination of the experiment at 86 DPI.

In ILRI-2918-infected mice, survival was constant (100%) up to and after treatment with diminazene aceturate at 21 DPI (Figure 4). Upon treatment, mice developed hindleg paresis (50% of the experimental population at the time), at which time they were euthanised. Upon relapse of parasitaemia, the remaining mice succumbed to the parasite, thereby causing termination of the experiment.

In ILRI-3953-infected mice, mortality increased with the rise in parasitaemia, with mice rendered moribund as the experiment progressed. By 10 DPI, all mice were dead in their cages or had been euthanised after hindleg paresis was observed.

In IPR-001-infected mice, mortality occurred even before treatment with diminazene aceturate at 21 DPI. Approximately 3% of the experimental population succumbed to the stabilate's parasitaemia, with 97% surviving before treatment with diminazene aceturate at 21 DPI. The survival dropped further through 38% (24–31 DPI), to 35% (32–44 DPI), then 31% (45–56 DPI), before finally stabilising after treatment with melarsoprol to 27% (57–79 DPI), whereupon it remained stable until the end of the experiment. At the end of the experiment, the surviving mice were euthanised and their organs harvested for histopathology.

Examination of the Giemsa-stained thin blood smears indicated that pleomorphism occurred in two of the parasite strains used, namely, ILRI-2918 and IPR-001. In these stabilates, two trypomastigote forms – long slender and short stumpy – were observed during the course of infection. From five DPI in ILRI-2918 and three DPI in IPR-001, there was a predominance of long slender trypomastigotes (97.4% in ILRI-2918 and 83.7% in IPR-001), with few short stumpy trypomastigotes observed (2.6% in ILRI-2918 and 16.3% in IPR-001). For both strains, observation of the different morphological forms ceased at 23 DPI after treatment with diminazene aceturate at 21 DPI.

In ILRI-2918-infected mice, there was an increase in short stumpy forms (18.6%) with a decline in long slender forms (83.7%) (Figure 5) during the first parasitaemia peak at seven DPI, compared to the populations at the beginning of the observation period. At eight DPI, there was a great increase in the number of short stumpy trypomastigotes (64.6%), with a decline in the long slender forms (35.4%). At the second peak (at 17–19 DPI), there was a predominance of short stumpy forms (up to 92.4%) and a decline in long slender forms (7.6%).

In IPR-001-infected mice, during the first peak at six DPI, there was a notable increase in the number of short stumpy trypomastigotes (33.6%) and a decline in the long slender forms (66.4%) compared to the populations at the beginning of the observation period (Figure 6). In this isolate as well, at seven DPI, there was a great increase in the number of short stumpy trypomastigotes (54.4%) and a decline in the long slender forms (45.6%). At the second peak, at 18 DPI, there was a predominance of short stumpy forms (74.9%) and a decline of long slender forms (25.1%). However, the greatest increase in short stumpy forms during the IPR-001 infection occurred between 20 and 23 DPI (from 94.6% – 95.2%).

Gross anatomy observation revealed enlargement of the spleen and hyperaemia in the major organs. Histopathological observations of the brain included discrete inflammation, with the foci of meningitis characterised by progressive mononuclear cell infiltration, perivascular cuffing (Figure 7) and hyperaemia of meningeal blood vessels. These observations were seen to progress over time in all observed samples from experimental mice. Infiltration of the brain parenchyma with mononuclear cells was rare and only began to be prominent in the late stages of infection. The choroid plexus was prominent in infected mice and contained inflammatory exudates, mainly of mononuclear cells.

In all groups of experimental mice, most organs were characterised by interstitial mononuclear cell infiltration into the parenchyma and perivascular cuffing. The most affected organs were the liver, spleen, heart and lungs. There were no marked differences between ILRI-2918 and IPR-001-infected mice. The germinal centres of the spleen were active and contained many immature lymphocytes and plasma cells, especially in samples obtained at 14 and 21 DPI, although these cells were reduced in number by 28 DPI. The lung alveoli, as well as the interseptal spaces, were characterised by infiltration with inflammatory exudates and cells, including lymphocytes, macrophages and a few neutrophils. This was more prominent in samples from mice euthanised after 21 DPI. The liver indicated extensive lymphocytic infiltration coupled with necrosis of hepatocytes, especially in mice euthanised after 28 DPI.

As described in the preceding section, there were a number of limitations observed which can be summarised as follows. Firstly, the strain of mouse greatly influences the disease progression and pathology encountered. There could thus be deviations if the model is established using a different mouse strain, particularly an inbred one. Secondly, the study assumed the occurrence of CNS involvement by relapse of parasitaemia after haemolymphatic clearance with diminazene aceturate, as has been done in previous studies. However, assessment of plasma concentrations of acute-phase protein, C-reactive protein and haptoglobin could also be used as a marker for experimental infections.^16,33 Thirdly, it is difficult to interpret brain pathology in terms of relapse following treatment with trypanocidals, as they, too, have been shown to induce brain pathology. Another point of contention when using trypanocidals to stage the disease is the danger that subcurative treatment of a parasitic infection may lead to emergence of drug resistance, a position which would complicate the interpretation of efficacy data if the drug under investigation has a similar mode of action.¹¹ Finally, the parasites used in this study have not yet been characterised genetically.

In order to circumvent the preceding limitations, an initial baseline study is required for any novel mouse model utilising the described parasite isolates, incorporating more extensive measures of parasitaemia as well as parasite staging. Full genetic characterisation of the parasites used here is needed in order to ascertain the identity of the isolates during future experiments and in order to account for variations observed in subsequent studies. Because of the heterogenous nature of outbred Swiss mice, there may be significant deviations in results from mouse studies using the same isolates in the future. For this reason, studies using the more homogenous inbred mouse strains, which may not be susceptible to the isolates used here, are recommended in order to establish the late-stage HAT model.

The authors consider the experimental design of this study to be reliable and valid for the purpose of establishing a mouse model in outbred Swiss white mice. The procedures utilised in this study have been tested and validated in other studies as cited in this article.