Towards environmental detection of Chagas disease vectors and pathogen

Chagas disease vector control relies on prompt, accurate identification of houses infested with triatomine bugs for targeted insecticide spraying. However, most current detection methods are laborious, lack standardization, have substantial operational costs and limited sensitivity, especially when triatomine bug densities are low or highly focal. We evaluated the use of FTA cards or cotton-tipped swabs to develop a low-technology, non-invasive method of detecting environmental DNA (eDNA) from both triatomine bugs and Trypanosoma cruzi for use in household surveillance in eastern Colombia, an endemic region for Chagas disease. Study findings demonstrated that Rhodnius prolixus eDNA, collected on FTA cards, can be detected at temperatures between 21 and 32 °C, when deposited by individual, recently blood-fed nymphs. Additionally, cotton-tipped swabs are a feasible tool for field sampling of both T. cruzi and R. prolixus eDNA in infested households and may be preferable due to their lower cost. eDNA detection should not yet replace current surveillance tools, but instead be evaluated in parallel as a more sensitive, higher-throughput, lower cost alternative. eDNA collection requires virtually no skills or resources in situ and therefore has the potential to be implemented in endemic communities as part of citizen science initiatives to control Chagas disease transmission.

Chagas disease remains the most important parasitic infection in Latin America, responsible for the loss of 275,000 disability-adjusted life years (DALYs) in 2019 1 . The geographical range of the aetiological agent, Trypanosoma cruzi, extends from the southern USA to Argentinean Patagonia, where it is transmitted by triatomine bugs (Hemiptera: Reduviidae: Triatominae) to at least eight orders of domestic, synanthropic and sylvatic mammalian hosts 2,3 . Human disease occurs when infected triatomine faeces enter through intact mucosa or abraded skin, causing an initial asymptomatic or non-specific self-limiting febrile illness, followed by life-long infection and potentially fatal cardiomyopathy (30-40% of infected individuals) 4 . Without a vaccine or highly efficacious treatment options for adults 5,6 , accurate detection of triatomine-infested houses and residual insecticide spraying of domestic and peri-domestic structures are crucial to prevent new cases [7][8][9] . Domiciliary populations of triatomine bugs have been successfully controlled by spraying of pyrethroid insecticides across large parts of Latin America, contributing to reductions in their distribution from an estimated 6.28 million km 2 in the 1960s to less than 1 million km 2 today 10 . Despite these achievements, the success of contemporary Chagas disease vector control programmes is threatened by persistent peri-domestic foci 11,12 , the emergence of insecticide resistance [13][14][15] and household invasion from sylvatic triatomine bug populations [16][17][18][19][20] .
Chagas disease vector surveillance typically relies on identification of infested houses using timed-manual collections (TMCs), conducted by skilled personnel, with or without a dislodging spray 21,22 . However, this methodology is laborious and suffers from several other drawbacks, including lack of standardization, substantial operational costs and limited sensitivity, especially when infestation is highly focal and/or triatomine bug densities are very low after a recent insecticide spraying campaign [23][24][25] . Community-based bug collections or bug notifications (with or without proof of collection) performed by householders have reported similar or sometimes superior levels of sensitivity to TMCs [26][27][28] , but are prone to variability due to changes in motivation and skills of R. prolixus and T. cruzi gDNA LoD / LoQ. To estimate the analytical sensitivity, linearity and dynamic range of the two qPCR assays for eDNA detection, serial dilutions of R. prolixus and T. cruzi gDNA were tested in three independent assays, with technical triplicates per assay. Both qPCR assays produced good linearity (R 2 = 0.98 and 0.99 for R. prolixus and T. cruzi assays, respectively) and efficiencies of 90.94% and 93.07%, respectively (Fig. 1A,B). The LoD/LoQs were determined to be 1.5 × 10 -2 and 7.1 × 10 -3 copies/reaction for R. prolixus and T. cruzi assays, respectively (Fig. 1). Next, the analytical sensitivity, linearity and dynamic range of the multiplex qPCR assay was evaluated using R. prolixus and T. cruzi gDNA serially diluted in equal proportions. Similarly, the multiplex assay was highly efficient (93.07% and 92.35% for R. prolixus and T. cruzi, respectively) with good linearities (R 2 = 0.99 for detection of both R. prolixus and T. cruzi gDNA). The LoD/LoQs were determined to be 2.4 × 10 -3 and 1.19 × 10 -4 copies/reaction for R. prolixus and T. cruzi assays, respectively (Fig. 1C).
Triatomine eDNA environmental variables. The effects of altering five different environmental variables (darkness, triatomine number, temperature, feeding status and degradation at ambient temperature) on detection of R. prolixus eDNA from FTA cards were investigated (Fig. 2).
In artificial light, varying the amount of time R. prolixus 3rd/4th instars had to walk/urinate/defecate on FTA cards from 15 min to 24 h, did not substantially affect eDNA detection levels, with qPCR amplification observed at all timepoints (Fig. 3A). In general, with increasing time, Ct values decreased significantly (Fig. 3A)  www.nature.com/scientificreports/ Finally, R. prolixus physiological status also had a significant impact on eDNA detection, with significantly higher levels of sensitivity observed for recently blood-fed bugs (Fig. 3F) 60-29.79] (Mann Whitney test, p < 0.0001). Furthermore, the rate of qPCR non-amplification was higher among the unfed bug group, with three out of six DNA extractions failing to amplify during all three technical replicates; one extraction amplified once, and two extractions amplified twice.
Triatomine and T. cruzi eDNA detection from field specimens. In Colombia, sixty-three houses were sampled in Prado Veraniego (n = 30), Bogotá (non-endemic for Chagas disease), and Agualinda (n = 14) and San Isidro (n = 19), Casanare (endemic for Chagas disease), for R. prolixus/T. cruzi eDNA using cottontipped swabs. Samples were collected from a range of domestic and peridomestic locations, including from living room walls, living room ceilings, TV room walls, kitchen walls, dining room walls, stables and hen nests (Supplementary file S2). All houses in Agualinda and San Isidro were confirmed as infested with triatomine bugs by self-reporting householders, while all houses in Prado Veraniego were reported as uninfested. Only two households in San Isidro reported human T. cruzi infection.
Among 33 households in Agualinda and San Isidro with confirmed triatomine infestations, R. prolixus eDNA was detected in 20, giving an estimated sensitivity of 60.6% (Fig. 4)

Discussion
Accurate surveillance of triatomine household infestation is crucial for Chagas disease vector control. However, no gold standard vector detection method with high levels of sensitivity or specificity is currently available. Given there are several intrinsic features of triatomine bug behaviour and the lifecycle of T. cruzi which lead to deposition of pathogen and vector eDNA in infested houses, this study evaluated the use of FTA cards or cotton-tipped www.nature.com/scientificreports/ swabs for simultaneous, non-invasive parasitological and entomological surveillance in eastern Colombia, an endemic region for Chagas disease.
In the proof-of-concept laboratory experiments, R. prolixus eDNA from five 3rd/4th instar nymphs was successfully isolated and amplified from FTA cards after as little as 15 min of contact time under standard insectary conditions. This indicates that FTA cards may be able to detect transient triatomine movements and lighter triatomine infestations comprising of immature vector stages; these are features of new infestations and/or vector population recovery post insecticide spraying, and are notoriously harder to measure with conventional TMCs 26 . While altering contact time (up to 24 h), and light/dark cycles, did not significantly improve R. prolixus eDNA detection, increasing temperature from 21 to 27-32 °C and triatomine bug density from 1 to 25 bugs, were both correlated with levels of qPCR amplification. Due to the laboratory conditions being highly artificial, the sensitivity of the FTA cards may have been overestimated by placing triatomines in smaller confined spaces than their colony jars and altering their light:dark cycles, causing them to move or defecate more due to stress. R. prolixus physiological status also impacted eDNA detection, with significantly higher levels of sensitivity observed for recently blood-fed bugs, compared to those starved for 19 weeks, suggesting that the majority of detectable eDNA is likely derived from triatomine faeces/urine compared to tarsal deposition upon contact. These findings are consistent with FTA card evaluations for other vectors and pathogens of public health importance, demonstrating direct dose responses between detection levels and target organism density 41,43 . Under simulated conditions of degradation, R. prolixus eDNA was shown to be stable on FTA cards for at least two weeks at room temperature, which aligns with previous studies reporting successful pathogen isolation from cards stored for several weeks under similar conditions; published observations indicate longer-term stability can be achieved when FTA cards are stored at 4 °C 40 .
In the field specimens, R. prolixus eDNA was detected from infested houses with an estimated sensitivity of 60.6% which is comparable with reported detection levels for TMCs, householder collections and other trapping techniques 26,29,56 . Given that household triatomine presence was self-reported, one explanation for the number of false negative houses may be inaccuracies in resident reports, infestation with other local triatomine species (ten triatomine species, including R. prolixus, Triatoma maculata and Panstrongylus geniculatus, have been identified in Casanare 57 ), relative sensitivity of using cotton-tipped swabs to detect eDNA (these are not specifically optimised for DNA capture like FTA cards), or the LoD of our qPCR assay. More importantly, this technique was shown to be 100% specific in this particular field setting; false-positives in triatomine vector surveys arise www.nature.com/scientificreports/ due to taxonomic errors (i.e. mis-identifying non-triatomine reduviid nymphs or non-vector species), use of indirect proxies of vector infestation (e.g. triatomine bug faecal streaks which may be confused with those of other arthropods 58 ) or when householders report vector presence without visual confirmation 23 . In Casanare (eastern Colombia), traditional methods are still used for entomological surveillance of R. prolixus and T. cruzi. Together these observations highlight the potential for this technique to accurately identify foci of residual triatomine infestation, which are important sources of operational failure of current Chagas disease vector control programmes 59,60 . In terms of implementation, the COVID-19 pandemic increased the molecular diagnostic capabilities of several laboratories nationwide, including the Casanare department (Secretaria Departamental de Salud y Unitropico). These laboratories oversee monitoring for SARS-CoV-2 and other vector borne diseases that are endemic in this department, with the potential to support implementation of the detection tools develop herein, for accurate and feasible field surveillance of Chagas disease in the region. Interestingly, T. cruzi eDNA was amplified from 93.9% of infested houses, when only 6.06% reported human infection. We excluded possible laboratory contamination as a confounder by processing known negative controls (cotton-tipped swabs wiped on house walls in the UK) in parallel with field specimens at every analytical step. Instead, this discordance may reflect differences in relative amounts of T. cruzi and R. prolixus eDNA deposition/ differential rates of eDNA degradation between species, or the gene copy number between the two targets used for qPCR detection; in our assay the LoD for the T. cruzi satellite DNA was an order of magnitude more sensitive compared to the 12S rRNA in R. prolixus. Alternatively, these findings may be indicative of active infected vectors in these houses; parasite transmission is known to be highly precarious and inefficient, requiring an estimated 900-4000 infected contacts per case 61 and previous studies in Casanare have identified houses where 100% of collected Rhodnius are positive for T. cruzi 62 . While human infection was only self-reported in two houses, T. cruzi transmission in this area is also under-diagnosed due to substantial heterogeneity in acute symptomology 57 and other significant barriers to adequate healthcare, including lack of diagnostics, infrastructure and financial investment, and limited physician awareness 63 . Regardless, the potential presence of residual parasite eDNA in these houses requires further investigation, including confirmation of infected vectors using a second entomological surveillance method and serodiagnosis of householders. www.nature.com/scientificreports/ Findings from this study provide insights into the feasibility of using FTA cards or cotton-tipped swabs for community-level surveillance. Based on laboratory results, FTA cards could be taped to house walls for several weeks at temperatures between 21 and 32 °C and potentially detect eDNA deposited by individual, recently blood-fed R. prolixus nymphs. Field specimens further confirmed our ability to amplify R. prolixus/T. cruzi eDNA deposited on house walls in Casanare using a cheaper, lower-technology tool. By comparison to TMCs and other trapping techniques, these methodologies require virtually no skills or training and do not involve residents actively exposing themselves to potentially infectious triatomines. This surveillance strategy could be integrated into newly developed citizen science initiatives for Chagas disease, which have used social media applications and behavioural design frameworks to improve community disease awareness and reporting of house infestation [64][65][66] . In addition to field standardization, longitudinal evaluations of this methodology are needed alongside parallel TMCs and community serosurveys, to quantitatively evaluate this technique, to establish the true LoD of parasite/vector eDNA in the field using both FTA cards and cotton-tipped swabs, to determine how seasonal changes in triatomine population dynamics affect eDNA degradation and to optimise timing and logistics of eDNA wall sampling. While this study only investigated presence/absence of T. cruzi and R. prolixus, further research is warranted to assess the possibility of using recovered eDNA for community-wide blood-meal analysis, surveillance of molecular insecticide resistance and characterization of parasite and vector population genetic structures, including reinfestation dynamics after insecticide spraying campaigns 13,17,67 . These aspects are pivotal to the development of effective vector control programmes in Chagas disease endemic regions.

Conclusions
This study validated the use of FTA cards and cotton-tipped swabs for simultaneous entomological and parasitological Chagas disease surveillance. Study findings demonstrated that R. prolixus eDNA, collected on FTA cards, can be detected at temperatures between 21 and 32 °C, when deposited by individual, recently blood-fed 3rd/4th instar nymphs. Additionally, cotton-tipped swabs are feasible for field sampling of both T. cruzi and R. prolixus eDNA in situ and are arguably more preferable due to their lower cost. eDNA detection should not yet supplant current methods such as TMCs, but instead be evaluated alongside them as a more sensitive, higher-throughput, lower cost prospective alternative. eDNA collection can be implemented in local endemic communities as part of citizen science initiatives to monitor and control Chagas disease transmission. Further studies are needed to investigate the feasibility of using recovered eDNA for exploratory genetic analyses.

Triatomine colony maintenance. The insectary at the London School of Hygiene and Tropical Medicine
(LSHTM) provided the R. prolixus for this study. They were maintained at 25 °C ± 60-80% relative humidity with 12 h:12 h light:dark cycles. This colony was derived from material sent to LSHTM from Venezuela in 1927. All R. prolixus individuals used in this study were 3 rd or 4 th instar nymphs. During the study period, triatomines were blood-fed with equine blood warmed through a Hemotek feeder, used for experimentation, returned to the colony, and then re-used for experimentation once every 6 weeks.
Optimising triatomine eDNA sampling. Before altering any environmental variables, which might have affected triatomine eDNA detection, an initial experiment was conducted to determine the optimal time for eDNA sampling. One hundred and five 3rd/4th instar R. prolixus were blood-fed then immediately placed in different glass jars (dimensions: opening diameter 6.5 cm; base diameter 9.5 cm; height 14 cm) in groups of 5 under standard insectary conditions (25.1-26.2 °C, 68-74% relative humidity). For eDNA sampling, QIAcard FTA classic cards (Qiagen, UK) were used. These are cards treated with FTA, a chemical which causes cell lysis and DNA immobilisation to isolate pure DNA. These were stapled to plain A4 paper and secured over the jar openings with an elastic band (Fig. 5A). The jar was then inverted so that the R. prolixus were walking, urinating, and defecating on the card. Triatomines in the 21 jars were left on the FTA cards for 7 time points: 15 min, 30 min, 1 h, 2 h, 5 h, 12 h and 24 h (Fig. 5E). Each time point was trialled in biological triplicate, i.e. three independent jars were used per time point. After each time point, the glass jars were reverted (Fig. 5C), R. prolixus returned to the colony, and the FTA cards removed and stored at − 20 °C. eDNA extraction, amplification, and detection (described below) indicated that 24 h was the optimal time point to leave the triatomines on the FTA cards to detect their eDNA. This time point was used in subsequent experiments.
Altering triatomine eDNA environmental variables. Next a series of experiments investigating the impact of five different environmental variables (darkness, triatomine number, temperature, feeding status and degradation at ambient temperature) on triatomine eDNA detection were performed (Fig. 2). All experiments were performed in biological triplicate. Negative controls were run concurrently in all experiments by placing FTA cards over empty glass jars. Temperature and humidity were consistently controlled in all experiments, except when temperature was the independent variable under evaluation. All experiments were performed consistently in 24 h of artificial light, except when light was the independent variable under evaluation.
Darkness. After blood-feeding 105 3 rd /4 th instar R. prolixus and placing 5 in each jar, the jars were inverted onto FTA cards for one time point per group: 15 min, 30 min, 1 h, 2 h, 5 h, 12 h, and 24 h, and covered by a blackout blanket to measure the effect of darkness on the quantity of eDNA sampled (Figs. 2 and 5G). Using all time points allowed us to compare whether the triatomines, which are nocturnal, would be more active in the dark than in artificial light and therefore deposit more eDNA. At the end of each time limit, bugs were returned to the colony and the FTA cards stored at − 20 °C.  www.nature.com/scientificreports/ 2 weeks and 8 weeks to measure the effect of time at ambient conditions on eDNA degradation (Fig. 2). Ambient temperature and humidity were recorded hourly using an EL-USB-2 RH/temperature data logger (Lascar Electronics, UK). The FTA cards were then stored at − 20 °C until eDNA extraction.
Triatomine eDNA extraction from laboratory specimens. FTA cards from each different condition were cut in half and then into 1 × 2 cm strips using scissors sterilised between samples with 10% (v/v) bleach and 70% (v/v) ethanol. Two independent eDNA extractions were performed per FTA card. Strips from each half FTA card were placed into individual sterile 50 mL Falcon tubes (Fisher Scientific, UK), immersed in 5400 μL ATL buffer and 600 μL proteinase K (Qiagen, UK) and incubated overnight at 56 °C. eDNA was extracted from 6 mL of sample using Qiagen DNeasy 96 Blood and Tissue kits (Qiagen, UK), according to the manufacturer's protocol.
Triatomine eDNA detection from laboratory specimens. Triatomine eDNA was detected using qPCR to amplify a fragment of the R. prolixus 12S rRNA gene. A standard curve of Ct values for this assay was generated using a tenfold serial dilution of control R. prolixus gDNA (extracted from individual colony adults), to assess PCR efficiency. Genomic DNA concentration was determined using the Qubit 4 fluorometer Household eDNA sampling and extraction. Due to the SARS-CoV-2 pandemic, we were unable to ship FTA cards to field collaborators in Colombia. Instead, between June-September 2021, cotton-tipped swabs (Guangzhou Improve Medical Instruments Co., Ltd, China) were used to sample triatomine and T. cruzi eDNA from different surfaces in houses and peri-domestic structures in Agualinda and San Isidro, municipality of Pore, Department of Casanare, Colombia (endemic for Chagas disease) and Prado Veraniego, municipality of Bogotá (non-endemic for Chagas disease) ( Fig. 6; Supplementary file S2). Each dry swab was used to wipe one random 10 cm 2 surface area of a selected part of the house/peri-domestic structure; roofs were not sampled because in Casanare, Rhodnius are preferentially found in the walls 62 . Householders completed a short questionnaire to confirm triatomine infestation and T. cruzi infection. Field workers changed gloves between each sample and individual cotton swabs were packaged in sterile, small plastic bags filled with silica gel and shipped to the LSHTM at room temperature. In addition, a series of negative control cotton-tipped swabs, collected by sampling house walls in the United Kingdom, were prepared and processed in parallel to field specimens.
To extract eDNA, swabs were placed into individual sterile 50 mL Falcon tubes (Fisher Scientific, UK), immersed in 900 μL ATL buffer and 100 μL proteinase K (Qiagen, UK) and incubated overnight at 56 °C. eDNA was extracted from the entire sample using Qiagen DNeasy 96 Blood and Tissue kits (Qiagen, UK), according to the manufacturer's protocol.
Triatomine and T. cruzi eDNA detection from field specimens. Parasite eDNA was detected using qPCR to amplify a fragment of the T. cruzi satellite DNA 69 . A standard curve of Ct values for this assay was generated using a tenfold serial dilution of control T. cruzi gDNA (strain SMA6: TcI 70 ), to assess PCR efficiency. Genomic DNA concentration was determined using the Qubit 4 fluorometer 1X dsDNA HS assay (Invitrogen, UK).
Standard curve reactions for simultaneous detection of R. prolixus and T. cruzi gDNA were generated using a 13-fold serial dilution of equal proportions of R. prolixus and T. cruzi gDNA, to assess PCR efficiency. Genomic Data analysis. Stratagene MxPro qPCR software (Agilent Technologies, UK) was used to produce qPCR standard curves. qPCR assay limits of detection (LoD) and limits of quantification (LoQ) were determined using the "Generic qPCR LoD/LoQ calculator" 34 , implemented in R version 4.0.2 71 . All other statistical analyses were conducted in GraphPad Prism v9.2.0.
Ethics approval and consent to participate. Ethical approval for the study was obtained from the London School of Hygiene and Tropical Medicine (LSHTM; ref#25638) and the Universidad del Rosario ("Genómica, evolución y biogeografía de especies del género Rhodnius: vectores de la enfermedad de Chagas" act number 007/2016) and all study procedures were performed in accordance with relevant guidelines and regulations.

Data availability
The datasets generated and/or analysed during the current study are contained within the supplementary files.