Caught in a trap: DNA contamination in tsetse xenomonitoring can lead to over-estimates of Trypanosoma brucei infection

Abstract

Background: Tsetse flies (Glossina sp.) are vectors of Trypanosoma brucei subspecies that cause human African trypanosomiasis (HAT). Capturing and screening tsetse is critical for HAT surveillance. Classically, tsetse have been microscopically analysed to identify trypanosomes, but this is increasingly replaced with molecular xenomonitoring. Nonetheless, sensitive T. brucei-detection assays, such as TBR-PCR, are vulnerable to DNA cross-contamination. This may occur at capture, when often multiple live tsetse are retained temporarily in the cage of a trap. This study set out to determine whether infected tsetse can contaminate naïve tsetse with T. brucei DNA via faeces when co-housed.

Methodology/principle findings: Insectary-reared teneral G. morsitans morsitans were fed an infectious T. b. brucei-spiked bloodmeal. At 19 days post-infection, infected and naïve tsetse were caged together in the following ratios: (T1) 9:3, (T2) 6:6 (T3) 1:11 and a control (C0) 0:12 in triplicate. Following 24-hour incubation, DNA was extracted from each fly and screened for parasite DNA presence using PCR and qPCR. All insectary-reared infected flies were positive for T. brucei DNA using TBR-qPCR. However, naïve tsetse also tested positive. Even at a ratio of 1 infected to 11 naïve flies, 91% of naïve tsetse gave positive TBR-qPCR results. Furthermore, the quantity of T. brucei DNA detected in naïve tsetse was significantly correlated with cage infection ratio. With evidence of cross-contamination, field-caught tsetse from Tanzania were then assessed using the same screening protocol. End-point TBR-PCR predicted a sample population prevalence of 24.8%. Using qPCR and Cq cut-offs optimised on insectary-reared flies, we estimated that prevalence was 0.5% (95% confidence interval [0.36, 0.73]).

Conclusions/significance: Our results show that infected tsetse can contaminate naïve flies with T. brucei DNA when co-caged, and that the level of contamination can be extensive. Whilst simple PCR may overestimate infection prevalence, quantitative PCR offers a means of eliminating false positives.

Fig 1. A flow diagram depicting basic experimental framework for the trap experiments.

Figure created using biorender.com (www.biorender.com [accessed 01/02/24]).

Fig 2

An example of an Nzi trap with trap cage at the apex (A) and detail of a Glossina sp. captured within the trap cage (B).

Fig 3

Box-and-whisker plots showing Cq value data from T. brucei (A) multi-copy target TBR-qPCR screening and (B) single-copy target PLC-qPCR screening of infected flies (IF) and naïve (UF) across four trap types (T1-T3, C0). C0 A+B were placed within close proximity (< 1 metre) of experiments (T1-3), C0 C was placed in a separate room. This was to test localised airborne DNA contamination. Crosses represent the mean Cq values. Grey bars display proportion of samples recording amplification using respective qPCR assays.

Fig 4. Plots displaying Cq values for field-caught flies.

(A) shows TBR-qPCR Cq values (circular, black symbol) for all field flies where DNA was available (n = 640). (B) shows PLC-qPCR Cq values (triangular, red symbol) for a subset of field flies with TBR-qPCR Cq <22.13 and where DNA was available (n = 45). (C) shows comparison of TBR-qPCR Cq values from female (circular symbol, n = 428) and male (diamond symbol, n = 212) in field-caught flies. There was no significant difference in median TBR-qPCR Cq values from females (median = 26.22) and males (median = 25.97, p = 0.5336). For all plots (A, B, C) grey boxplot shows median and 1–99% percentiles, error bars display range. The red dotted horizontal lines represent the Cq cut offs of 22.13 for TBR (A, C) and 25.36 for PLC (B).

Fig 5. Catches where >95% of trapped flies were collected and screened and total catch >1 (n = 19). Arranged in order of proportion of TBR +ve flies (L-R, largest to smallest).

(A) shows TBR-qPCR Cq values (circular, black symbol) for each fly sample in each catch. (B) shows PLC-qPCR Cq values (triangular, red symbol) for each fly sample in each catch that also had a TBR-qPCR Cq value <22.13 and had DNA available. Grey bars (right axes) represent the proportion (%) of flies in each catch testing TBR-PCR positive. The red dotted horizontal lines represent the Cq cut offs of 22.13 for TBR (A) and 25.36 for PLC (B).