A lineage-specific rapid diagnostic test (Chagas Sero K-SeT) identifies Brazilian Trypanosoma cruzi II/V/VI reservoir hosts among diverse mammalian orders

Trypanosoma cruzi, the protozoan agent of Chagas disease in the Americas, is comprised of six genetic lineages (TcI-TcVI) and a possible seventh (TcBat, related to TcI). Identification of T. cruzi lineages infecting reservoir mammalian species is fundamental to resolving transmission cycles. However, this is hindered by the limited sensitivity and technical complexity of parasite isolation and genotyping. An alternative approach is serology using T. cruzi lineage-specific epitopes, such as those of the trypomastigote small surface antigen (TSSA). For surveillance of T. cruzi lineage infections in mammal species from diverse Brazilian regions, we apply a novel rapid diagnostic test (RDT, Chagas Sero K-SeT), which incorporates the TSSA peptide epitope specific to TcII/V/VI (TSSApep-II/V/VI) and Protein G detection of antibodies. Chagas Sero K-SeT RDT results with sera from experimentally infected mice, from tamarin primates (Leontopithecus spp.) and from canines (Canis familiaris) were concordant with corresponding TSSApep-II/V/VI ELISAs. The Chagas Sero K-Set detected TcII/V/VI infections in Leontopithecus spp. from the Atlantic forest (n = 46), in C. familiaris (n = 16) and Thrichomys laurentius (n = 2) from Caatinga biome and Chiroptera (n = 1) from Acre, Amazonia. The Chagas Sero K-SeT RDT is directly applicable to TcII/V/VI-specific serological surveillance of T. cruzi infection in several different mammalian Orders. It can replace ELISAs and provides efficient, point-of-sampling, low-cost detection of TcII/V/VI infections, with at least equivalent sensitivity, although some mammals may be difficult to trap, and, not unexpectedly, Chagas Sero K-SeT could not recognise feline IgG. Knowledge of sylvatic hosts of T. cruzi can be expanded, new reservoir species discovered, and the ecology of transmission cycles clarified, particularly with adaptation to further mammalian Orders.

Introduction commercial collaborator (Coris BioConcept) provided no external funding or salary support. The work undertaken by NM was funded by the Sir Halley Stewart Trust (www.sirhalleystewart.org. uk). The views expressed within this article are those of the authors and not necessarily those of the Trust. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript Competing interests: We affirm that there is no competing interest; there is nothing that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of this manuscript. The commercial collaborator provided expertise in design of the rapid diagnostic tests (RDTs) and produced them to the senior authors' requirements; the RDTs are available for research purposes and are not patented. The commercial affiliation (Coris BioConcept) does not alter our adherence to PLOS ONE policies on sharing data and materials.
Here, we apply Chagas Sero K-SeT to sera from experimental murine infections, and to rapid serological surveillance for TcII/V/VI infections among a range of mammalian Orders and biomes in Brazil.

Ethics approval
Production of mouse sera adhered to the European 3Rs policy of Refinement, Reduction and Replacement (99/167/EG: Council decision of 25/1/99), took place in authorised animal facilities by licensed staff in agreement with the European Directive 86/609/EEC, and with review and approvals under UK Home office regulations [Animals (Scientific Procedures) Act 1986; project licence number 70/6997 to the London School of Hygiene and Tropical Medicine]. All Brazilian samples were obtained following the guidelines of the Animal Ethics Committee (CEUA) of the Oswaldo Cruz Institute/FIOCRUZ and all procedures followed protocols approved by the FIOCRUZ Committee of Bioethics (license LW 81/12).

Origins of animal sera
Experimental infections. Mus musculus (strain CD1) were inoculated with 10 6 organisms from stationary phase cultures containing infective metacyclic trypomastigotes of known biological clones of T. cruzi representing the lineages TcII (MHOM/BR/00/Y), TcIII (MDAS/PY/ 00/Arma18) and TcIV (MAOT/BO/00/10R26). Serum was collected at approximately 10 months post-inoculation, and serology for infection was detected using T. cruzi lysate as described below in section 'Experimental murine sera' Naturally infected mammals. Sera were archived samples collected in Brazil from naturally infected mammals as part of the ongoing field research programmes of author AMJ and collaborators. Collection sites encompass a range of geographical locations, mammalian Orders, species and biomes, as shown in Fig 1, map derived from www.simplemappr.net [21].

Lineage-specific TSSApep ELISA
Replica assays were performed simultaneously in duplicate plates. TSSA peptide (TSSApep) ELISAs were performed with the four lineage-specific peptides shown in Table 1. Mean values of optical density (OD) were calculated from the duplicate plates; cut-off values were mean negative serum values + 3 SD, with at least two reference negatives in every ELISA plate. Positive/negative controls for each sylvatic mammal type were used according to the results of previous serology by immunofluorescence.
Canine sera. ELISA plates were coated directly with 1 μg/100 μL/well TSSApep in coating buffer overnight, without avidin, because we observed that some dog sera bound non-specifically to avidin. After washing and blocking steps as described above, 100 μL/well of 1:200 dilution of canine serum in PBS/T/M was added and incubated, for 1 h at 37˚C. After washing (x 6), 100 μL/well of 1:10 000 dilution of Peroxidase-AffiniPure Rabbit Anti-Dog IgG (H+L) antibody (polyclonal; Jackson Immunoresearch, USA; Cat. no. 304-035-003, Lot no. 105408; Antibody Registry AB_2339344) was added, for 1 h at 37˚C. ELISA development was by Biomanghuinos TMB as described above, or alternatively by TMB ELISA substrate (ab171523: Abcam, UK). The dilutions of sera and peroxidase-conjugated antibody were previously optimised by titration.
Comparison of Protein A and Protein G peroxidase conjugates. In comparative pilot assays, Protein A-HRP (Southern Biotech, USA: Cat. no. 7300-05) and Protein G-HRP (Fisher Scientific, UK; Cat. no. 11899150) were used at 1:1000 dilution in ELISA with sera from armadillos, cats, rodents and dogs. In these assays, TSSApep-II/V/VI (directly coated) and lysate were used separately as antigens. Reactions were developed with OPD.

Chagas Sero K-SeT rapid diagnostic test (RDT)
This novel, low cost, lateral flow immunochromatographic rapid test was developed with Coris BioConcept, and employs TSSApep-II/V/VI as the coated antigen, and colloidal goldlabelled Protein G as the detection molecule for specific IgG, as previously described [19,20]. Following application of the serum sample on the sample zone (wide green line, Fig 2) then buffer in the buffer window, reactions were considered positive if at 15 minutes incubation a band was observed at the antigen line, in conjunction with an integral test validation line.

Statistical analysis
Significance of concordance between RDT results and ELISA (Kappa test) was calculated using GraphPad (GraphPad Software, San Diego, California, USA).

Lineage-specific TSSApep ELISA
TSSApep ELISA for experimental murine, and naturally infected primate, dog and cat are described below in relation to the Chagas Sero K-SeT RDT. TSSApep ELISA data could not be obtained using the available conjugated secondary antibodies for other mammals, whether mammal-specific (bat, opossum, coati) or Protein G conjugates (armadillo).
Concordance of lineage-specific ELISA and Chagas Sero K-SeT RDT. Fig 2 shows representative examples of results with sera from experimental murine infections, and natural infections of rodents (Thrichomys laurentius), canines (Canis familiaris) and primates (Leontopithecus chrysomelas), assayed by both TSSApep ELISA and Chagas Sero K-SeT RDT.
In experimental murine infections and natural infections of Primata (Callitrichidae) and Carnivora (Canidae), TSSApep-II/V/VI ELISA and Chagas Sero K-SeT RDT results were concordant; P < 0.0001 for primates, for which most data were available ( Table 2). Three of the primates TcII seropositive by ELISA were also TcV/VI ELISA positive, indicating TcV/VI infection or both TcII and TcV/VI infection. In agreement with previous ELISAs [24], serum from experimental murine infection with TcII was Chagas Sero K-SeT positive, whereas serum from TcIII and TcIV infections were negative by this RDT (Fig 2). Thus, not only did these two lineage-specific assays perform equally well, the colloidal gold-labelled Protein G in Chagas Sero K-SeT had the capacity to recognise specific IgG in diverse mammalian Orders (Rodentia, Carnivora, Primata).

Prevalence of TcII/V/VI infections detected by Chagas Sero K-SeT RDT
Among the primates, the Chagas Sero K-SeT showed that 19 of 68 (29%) L. rosalia in Rio de Janeiro State and 27 of 35 (77%) of L. chrysomelas in the northern Brazilian State of Bahia were seropositive, indicating a high prevalence of TcII/V/VI infections, for which the natural hosts were previously poorly known. For L. rosalia and L. chrysomelas, 8 and 23 of the same samples, respectively, were previously tested by TSSApep ELISA [18]. In all but one case the Chagas Sero K-SeT RDT result accorded with the ELISAs; the non-matching case was positive by the RDT but negative by ELISA; Kappa test = 0.84 (95% confidence intervals: 0.64-1.00) (S1 Appendix).
Among the dogs, 16 of 57 (28%) of C. familiaris from the Caatinga biome of Ceará State were seropositive by Chagas Sero K-SeT.  (Fig 2) from the Caatinga biome in Piauí State were Chagas Sero K-SeT seropositive and thus carried a TcII and/or TcV/VI infection (Table 2).
ELISA using the anti-cat IgG secondary antibody detected two TSSApep-II/V/VI seropositive cats. Chagas Sero K-SeT detected no infections in felines.
The Protein G used in Chagas Sero K-SeT did recognise nearly a quarter or the dogs tested by this RDT; the Protein G-HRP used in ELISA failed to recognise dog IgG. The Protein A-HRP accorded with the result with specific anti-cat secondary antibody, and also recognised anti-T. cruzi lysate IgG in two armadillos.

Discussion
Recent publications have reviewed the complexities relating to Chagas disease and T. cruzi molecular epidemiology [3,25]. Regarding the latter point, the application of serology based on lineage-specific antigens has great potential for resolving the cryptic ecological cycles and the discovery of novel reservoir hosts of this parasite. Applications of T. cruzi lineage-specific serology to naturally-infected animals are with dogs [17,26], primates [18] and sympatric dogs, cats, and armadillo [20]. We have applied Chagas Sero K-SeT to exploit the capacity of Protein G to recognise TSSApep-II/V/VI specific IgG from a range of mammalian Orders. Previous reports of RDTs for T. cruzi serology of animal reservoirs have used InBios Trypanosoma cruzi-Detect-Canine [27] or Chagas StatPak [28], with no lineage-specific diagnosis. We demonstrate the versatility of the Chagas Sero K-SeT to recognise TSSApep-II/V/VI specific IgG in experimental murine T. cruzi infections, and in natural infections across several Orders of Brazilian mammals, namely Primata, Carnivora (canine), Rodentia and Chiroptera.
The Chagas Sero K-SeT with Leontopithecus spp. sera demonstrated the capacity of this RDT for lineage-specific serology of primates, previously achieved only with ELISA using conjugated anti-human IgG [18]. Previous reports [29,30] hypothesised that primates could be reservoir hosts of TcII; however, searches for TcII infections in Brazilian Amazonian primates only yielded TcI and TcIV. Our results confirm the high prevalence of TcII/V/VI infections in Brazilian Leontopithecus primates in the Atlantic Forest region of eastern Brazil, not in the Amazonian forest [29,30]. Beyond Brazil, TcII was recently isolated by xenodiagnosis of a free-living capuchin (Sapajus cay) from eastern Paraguay [31], and T. cruzi genotypes compatible with TcII, TcV and TcVI were reported in howler monkeys (Alouatta caraya) in northern Argentina [32]. The occurrence of these lineages in non-contiguous areas of South America and their broad range of host genera suggest that the primate ecological cycles of these lineages are far from fully elucidated.
Rocha et al [13] genotyped TcII from Brazilian dogs, as single or mixed TcI-TcII infections in Minas Gerais State, southern Brazil. Previous reports [17,26] have used a recombinant E. coli-produced TSSA fusion protein from CL Brener strain (TcVI; a clone from parental strain CL) in ELISA with naturally infected dogs from northern Argentina; however, the recombinant included sequences that are shared with other lineages. Chagas Sero K-SeT identified TcII/V/VI infections in dogs from Ceará state, in north-eastern Brazil, a region previously known to be highly endemic for TcII in domestic transmission cycles, although intradomiciliary transmission of T. cruzi has now been controlled. However, Lima et al [14] also reported TcII/V/VI and TcV/VI respectively in two dogs from Pará state, north of the Amazon Basin.
Domestic cats, which can acquire infection by eating triatomines or infected mice, are hosts of T. cruzi, as known from the early discovery of Chagas disease [33,34], and Rocha et al [13] reported TcI from a wild ocelot (Leopardus pardalis) in Minas Gerais State. The two feline samples TSSApep-II/V/VI seropositive by ELISA were negative by Chagas Sero K-SeT as expected, because Protein G does not bind feline IgG [35]. We found here that Protein A could recognise feline IgG, and its further application for serological surveillance is warranted.
Regarding rodents, we show that the TSSApep-II/V/VI ELISA with anti-mouse IgG can be replaced by Chagas Sero K-SeT with Protein G. Thrichomys laurentius is a sylvatic host of T. cruzi [36,37] in Ceará state, northeast Brazil. The TSSApep-II/V/VI seropositives identified here by Chagas Sero K-SeT were from the neighbouring state of Piauí. Similarly, a single bat specimen was positive by Chagas Sero K-Set; T. cruzi lineages have been reported from Brazilian bats, by genotyping [38,39].
Armadillos, especially D. novemcinctus, have been reported as natural hosts of TcIII throughout South America [11,40,41], thus the absence of TcII/V/VI infections from all 49 armadillos was not surprising. Although Chagas Sero K-SeT was shown to be able to detect infection in armadillo [20], positive test line intensity was weak; Protein A may be more appropriate than Protein G for binding of armadillo IgG, as others have reported [42], and which we observed here. No armadillo-specific conjugate was available for ELISA; this also encourages further deployment of Protein A for these important reservoirs hosts. Opossums have been identified as common hosts of TcI but rarely reported as hosts of TcII [11,43].
Diagnosis of T. cruzi infection has not hitherto been lineage-specific. The search for a TcIspecific antigen applicable to animals and humans remains an important current research goal, and may elucidate the reported differences in serological and immune responses in regions endemic for different lineages [44][45][46].
As recently described [9,10], the association of T. cruzi lineages with hosts, biomes or habitats is complex and not fixed. Infecting lineage composition may fluctuate within mammalian populations over time if there is varying exposure and transmission efficiency. Mixed infections, such as TcI and TcII, may occur in single mammals, and are not currently detectable by lineage-specific serology; nevertheless, due to sustained infection antibodies to TcII should not be lost. The Chagas Sero K-SeT RDT described here does not make redundant T. cruzi genotyping, or advanced analytical techniques such as flow cytometry [47]; however, by detecting lineage-specific host IgG, it can provide epidemiological information when direct parasite genotyping is hampered or subject to confounding biases.

Conclusion
The biological and geographical range of mammals from which specific anti-TSSApep-II/V/ VI IgG has been identified by Chagas Sero K-SeT RDT demonstrates its great potential in identifying novel reservoir hosts and elucidating ecological cycles of lineage transmission. We have shown here that this rapid, easy to use and interpret, Protein G-based RDT can replace ELISA in T. cruzi TSSApep-II/V/VI lineage-specific serology of a range of mammalian Orders; it does not require isolation, culture or direct genotyping of parasites, nor is it subject to confounders of those approaches. The combined use of Protein A and Protein G (or Protein A/G) as detection molecules for IgG may increase further the range of mammalian Orders to which lineage-specific serology can be applied.