Comparison of collection methods for Phlebotomus argentipes sand flies to use in a molecular xenomonitoring system for the surveillance of visceral leishmaniasis

Abstract

Background: The kala-azar elimination programme has resulted in a significant reduction in visceral leishmaniasis (VL) cases across the Indian Subcontinent. To detect any resurgence of transmission, a sensitive cost-effective surveillance system is required. Molecular xenomonitoring (MX), detection of pathogen DNA/RNA in vectors, provides a proxy of human infection in the lymphatic filariasis elimination programme. To determine whether MX can be used for VL surveillance in a low transmission setting, large numbers of the sand fly vector Phlebotomus argentipes are required. This study will determine the best method for capturing P. argentipes females for MX.

Methodology/principal findings: The field study was performed in two programmatic and two non-programmatic villages in Bihar, India. A total of 48 households (12/village) were recruited. Centers for Disease Control and Prevention light traps (CDC-LTs) were compared with Improved Prokopack (PKP) and mechanical vacuum aspirators (MVA) using standardised methods. Four 12x12 Latin squares, 576 collections, were attempted (12/house, 144/village,192/method). Molecular analyses of collections were conducted to confirm identification of P. argentipes and to detect human and Leishmania DNA. Operational factors, such as time burden, acceptance to householders and RNA preservation, were also considered. A total of 562 collections (97.7%) were completed with 6,809 sand flies captured. Females comprised 49.0% of captures, of which 1,934 (57.9%) were identified as P. argentipes. CDC-LTs collected 4.04 times more P. argentipes females than MVA and 3.62 times more than PKP (p<0.0001 for each). Of 21,735 mosquitoes in the same collections, no significant differences between collection methods were observed. CDC-LTs took less time to install and collect than to perform aspirations and their greater yield compensated for increased sorting time. No significant differences in Leishmania RNA detection and quantitation between methods were observed in experimentally infected sand flies maintained in conditions simulating field conditions. CDC-LTs were favoured by householders.

Conclusions/significance: CDC-LTs are the most useful collection tool of those tested for MX surveillance since they collected higher numbers of P. argentipes females without compromising mosquito captures or the preservation of RNA. However, capture rates are still low.

Fig 1. Location of study villages and Rajendra Memorial Research Institute of Medical Sciences in Bihar, India.

Base layer available at: World Topo Map (MapServer) (arcgisonline.com). Map generated using Epi Info 7 software.

Fig 2. Specimens collected by sex and collection location over the study period.

(A) displays sand fly data, (B) displays mosquito data.

Fig 3. Mean female sand flies collected by method and village.

Error bars indicate the 95% confidence interval of the mean.

Fig 4. Mean female mosquitoes collected by method and village.

Error bars indicate the 95% confidence interval of the mean.

Fig 5. Mean sand flies collected per trap-night by collection round for the first three collection rounds (with 95% CI).

Fig 6. Box plots showing the number of P. argentipes per collection event by method of collection.

Medians indicated by horizontal line (at 0 for MVA and PKP) and outliers by black dots.

Fig 7. RT-PCR amplification curves and corresponding parasite numbers for Leishmania donovani DD8 in sand flies 8 days post-infection with high (3000 amastigotes/sandfly) or low (<100 amastigotes/sandfly) doses of parasites.

Flies were killed and stored under simulated conditions for Prokopack aspirator (PKP: orange) or CDC-LT collection (CDC: blue) and kept in a cool bag for 7 hours. (A, C, D) T0 amplification curves for infected sand flies (black), processed immediately after killing (A: High dose infection; C & D: Low dose infections), including a positive control of 1x10⁶ cultured Log phase L. donovani promastigotes (magenta) and a non-template negative control of DNA-free water (grey). (B, E) T3 amplification curves for high (B) and low (E) infected sand flies stored under ‘worst case scenario’ field conditions of 40°C for 30 min followed by 7 hours at 4°C (PKP), or 40°C for 16 hours followed by 7 hours at 4°C (CDC) before processing. (F, G) Parasite numbers per sand fly for T0 and T3 High (F) and low (G) dose infected sand flies, extrapolated from a standard curve. Intermediary storage conditions T1 and T2 (see Table 1) were tested and parasite numbers are included in F and G; the corresponding amplification curves are in S1 Fig. Data shown are representative of two independent experiments. Asterisks indicate values that are statistically significant (**P < 0.005; *P < 0.05; ^nsP ≥ 0.05) between the indicated groups using a two-sided unpaired Mann Whitney U t-test.