Experimental Infection of Rhodnius prolixus (Hemiptera, Triatominae) with Mycobacterium leprae Indicates Potential for Leprosy Transmission

Leprosy is a chronic dermato-neurological disease caused by infection with Mycobacterium leprae. In 2013 almost 200,000 new cases of leprosy were detected around the world. Since the first symptoms take from years to decades to appear, the total number of asymptomatic patients is impossible to predict. Although leprosy is one of the oldest records of human disease, the mechanisms involved with its transmission and epidemiology are still not completely understood. In the present work, we experimentally investigated the hypothesis that the mosquitoes Aedes aegypti and Culex quinquefasciatus and the hemiptera Rhodnius prolixus act as leprosy vectors. By means of real-time PCR quantification of M. leprae 16SrRNA, we found that M. leprae remained viable inside the digestive tract of Rhodnius prolixus for 20 days after oral infection. In contrast, in the gut of both mosquito species tested, we were not able to detect M. leprae RNA after a similar period of time. Inside the kissing bug Rhodnius prolixus digestive tract, M. leprae was initially restricted to the anterior midgut, but gradually moved towards the hindgut, in a time course reminiscent of the life cycle of Trypanosoma cruzi, a well-known pathogen transmitted by this insect. The maintenance of M. leprae infectivity inside the digestive tract of this kissing bug is further supported by successful mice footpad inoculation with feces collected 20 days after infection. We conclude that Rhodnius prolixus defecate infective M. leprae, justifying the evaluation of the presence of M. leprae among sylvatic and domestic kissing bugs in countries endemic for leprosy.

Introduction and related to a more acute form of disease [26,27], this way of pathogen entry is still controversial in the case of leprosy [28].
In the present work, we hypothesize that kissing bugs such as Rhodnius prolixus, an insect from the family Reduviidae within the order Hemiptera (infra-order: Heteroptera), known as vectors of Chagas Disease, could also be a vector of Leprosy. We used immunofluorescence and real-time PCR of 16SrRNA to localize and quantify M. leprae Thai-53 strain along the digestive tract of the mosquitoes A. aegypti and C. quinquefasciatus and of the kissing bug R. prolixus. In contrast with mosquitoes, R. prolixus is able to maintain M. leprae viability inside its digestive tract for at least 20 days. This period is enough for the insect to start a new blood meal, defecating a large amount of live infectious bacilli. Studies are needed to evaluate the contribution of naturally infected kissing bugs to the transmission of leprosy.

Insect infection
Rhodnius prolixus, A. aegypti and C. quinquefasciatus were maintained in a biosafety insectary and fed with rabbit blood containing 10 7 /mL live M. leprae Thai-53 strain through a latex membrane (R. prolixus) or a plastic paraffin film (A. aegypti and C. quinquefasciatus) coupled to a glass chamber and maintained at 37°C with the use of a circulating water bath. The insects were allowed to feed in the dark for no more than one hour, a time interval that did not reduce M. leprae viability. After this time, insects not fully engorged were discarded. The adult kissing bugs and mosquitoes typically ingested blood meal volumes of about 250μl and 2μl respectively. After feeding, 13 groups of 5 R. prolixus each, and 10 groups of 10 A. aegypti or C. quinquefasciatus, were kept at 80% humidity and 27°C. Mosquitoes were allowed to feed ad libitum a 10% sucrose solution. All insects were kept in those conditions for 2 hours or 20 days after blood meal (ABM) until tissue dissection.

Mycobacterium leprae preparation and viability determination
Mycobacterium leprae was purified from nude mouse footpad as described [29]. Briefly, both Foxn1 nu/nu mice hind footpad were inoculated with 10 4 live M. leprae. Animals were monitored each 48 h during seven months. M. leprae infection in mice is recognized as painless, without the display of any symptoms besides footpad swelling, and for that reason analgesic protocol was not necessary. None of the animas died as a result of M. leprae infection. After seven months, animals were sacrificed through intraperitoneal administration of 15 mg/Kg and 150 mg/Kg of xylazine and ketamine (Vallée, SP, Brazil) respectively, followed by cervical dislocation. After skin and bones were removed, tissue was reduced in small pieces by scissors and digested with a solution of 170 units of colagenase type I, 2 units of dispase (Life Technologies, NY, USA), 50μ mg/mL of ampicillin (Sigma, St. Louis, USA) and 150 units of DNAse (Life Technologies, NY, USA) during 2 h at 33°C. Digested tissue was homogenized by vortex and washed, three times in water, one time in NaOH 0.1 N and one time in RPMI medium, by centrifugation at 10,000 g/5 min, and counted by acid-fast staining (Ziehl-Neelsen Kit, Becton Dickinson). For evaluation of the preparation viability, a fluorimetric staining protocol was used based on the LIVE/DEAD Bactlight Bacterial viability Kit (Life Technologies, CA, USA), performed according to the manufacturer's instructions [30]. To optimize artificial infection yield, we only used bacterial preparations with a viability of at least 85%. Freeze-thaw inactivated M. leprae was used to evidence 16Sr RNA instability inside triatomines digestive tract. Before infection, live M. leprae was frozen in dry ice and thaw at 37°C water bath five times. Although freeze-thaw does not drastically alter RNA content compared to other inactivation techniques such as irradiation or heat, M. leprae inactivation by freezing is a well characterize phenomenon [31].
To determine viable M. leprae titer inside insects digestive tract we used qPCR 16S cDNA/ DNA rate, as described elsewhere [32]. Briefly, a pool of 10 mosquitoes or R. prolixus two hours or 20 days after feeding on blood were dissected in PBS and mosquitos thorax and abdomen or R. prolixus posterior midgut, hindgut and rectum were removed under sterile conditions and transferred to FastRNA Blue tubes (MP Biomedicals), containing 1mL of TRIzol. Tissues were homogenized by vortex and M. leprae RNA and DNA were extracted after disruption in Fast Prep FP 24 homogenizer (MP Biomedicals, CA, USA) at a speed of 6.5 m/s for 45 sec. The tubes were cooled on ice for 2 min between two sections of disruption. After homogenization the tubes were cooled for 5 min and then received 200 μl of chloroform. After rapid mixing by inversion, tubes were centrifuged at 12,000 x g at 4°C for 15 min. The upper fase containing the RNA was transferred to a new tube, treated with DNA-Free kit (Ambion, Inc., Austin, USA) as specified by the manufacturer and precipitated by standard ethanol technique. Precipitated RNA was solubilized in RNAse free water and stored at -80°C until use. The DNA was purified by adding 100 μl of 10 mM Tris-EDTA (pH 8.0) and 150 μl of isoamyl alcohol and precipitated with 0.3 M sodium acetate with two volumes of cold ethanol. The DNA pellet was washed in 70% ethanol, dissolved in 30 μl of sterile distilled water and stored at -80°C until use. Qualitative RNA analysis was performed in agarose gel electrophoresis and quantification of both DNA and RNA were performed in NanoDrop One (Thermo Fisher Scientific, MA, USA). The M. leprae RNA was reverse transcribed using random primers and superscript III following manufacturer's instructions (Invitrogen, CA, USA). The levels of M. leprae 16SrRNA mRNA and DNA were determined in all tissues by real-time RT-PCR, using the primer pairs sense 5' GCA TGT CTT GTG GTG GAA AGC '3 and anti-sense 5' CAC CCC ACC AAC AAG CTG AT '3. The PCR reaction mixes were 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min monitoring SYBR Green fluorescence in an ABI StepOne Plus System (Applied Biosystems, CA, USA). In order to convert RNA/DNA Real-Time PCR signal in number of live M. leprae genomes, we added different concentrations of live M. leprae, ranging from 10 8 to 10 1 , in non infected insect tissues extracts dissected 2h and 20 days after blood meal, building fourteen curves, one for each condition. Their angular coefficient were used to determined the number of viable M. leprae in each condition.

Fluorescence microscopy
In order to observe M. leprae in triatomine feces, we infected the insects with M. leprae prestained with PKH-26 fluorophore (Sigma-Aldrich, St. Louis, USA) according to the manufacturer instructions. In all other tissues, non-fluorescent M. leprae was evidenced by immunostaining as followed. After artificial feeding, the triatomines tissues were fixed with 4% paraformaldehyde for 24 h at 4°C and cryopreserved by a two hour sequential incubations in PBS containing 15% and 30% sucrose. Soon after, the tissues were embedded in O.C.T. (Sakura Finetechnical, Tokyo, Japan), immediately frozen in liquid nitrogen and 10 μm semi thin sections were prepared using a Leica CM3050 cryostat and mounted on KCr (SO 4 ) 2 gelatinized slides.
M. leprae immunolocalization was performed using the monoclonal Anti-M. leprae lipoarabinomannan (LAM) antibody CS-35, kindly provided by Biodefense and Emerging Infections Research Resources Repository at http://www.beiresources.org/TBVTRMResearchMaterials/ tabid/1431/Default.aspx. Briefly, slides were permeabilized and blocked by 30 minutes incubation with 0.01% Triton X-100 and 10% of fetal bovine serum in PBS pH 7.2. Tissues were incubated for 2 hours with mouse IgG anti-LAM antibodies (1:50 vol/vol) and nuclei were stained by DAPI (Sigma-Aldrich, St. Louis, USA). Secondary antibodies conjugated with Alexa 633 IgG anti-mouse; (Invitrogen, CA, USA) were incubated with the samples for an additional 2 h and tissue observed in a Zeiss Axiobserver Z1 with a Colibri illumination system (Carl Zeiss, Heidenheim, Germany).

Rhodnius prolixus feces infectivity determined using the Shepard's model
In order to check triatomine feces for M. leprae infectivity, we used a model based on inoculation of M. leprae into mice footpads and counting acid-fast bacilli after 6 months, as previously described. Briefly, we infected R. prolixus with blood containing 10 7 /ml PKH-26 stained M. leprae according to manufacturer instructions, a procedure known as non-interfering with mycobacteria viability [33]. After 20 days, R. prolixus were fed with non-infected blood. Feces from 8 pools of 10 insects each were collected and pelleted by 10.000g / 5 min. The pellet was homogenized in a sterile solution of NaOH 0.1N during 5 minutes followed by centrifuging (10.000g / 5 min). The pellet was resuspended in sterile PBS and counted by fluorescence microscopy and Fite-Faraco Staining. Only samples which presented negative contaminants grow after 24h culture on blood agar and TH10 plates at 33 and 37°C (6 from 8 samples) were inoculated in animals. A total of 10 4 M. leprae were injected per BALB/c hind foot pad (ILSL, São Paulo, Brazil), using a conventional Sheppard's infection model [34]. As positive control we inoculate 10 4 M. leprae purified from nude mice foot pad. A negative control was generated by treating animals with 10mg/Kg/week rifampicin by weekly gavage during five months, one month after inoculum with 10 4 Thai-53 M. leprae purified from nude mice foot pad.
Animals were monitored each 48 h during seven months. None animal died as a result of M. leprae infection. After six months, animals were sacrificed through intraperitoneal administration of 15 mg/Kg and 150 mg/Kg of xylazine and ketamine (Vallée, SP, Brazil) respectively, followed by cervical dislocation, and foot pads were excised and macerated for bacillary counting by Fite-Faraco Staining.

Ethics Statement
Animal protocols were in agreement with the 8 th edition of the Guide of Care and Use of Laboratory Animals of the National Institutes of Health. All procedures were approved by the Animal Welfare Committee of Sagrado Coração University (São Paulo, Brazil), responsible for Instituto Lauro de Souza Lima animal care and use inspections, were all animal protocols were performed according license number 219/11.

Results
There are several reports on mosquitoes capacity to carry M. leprae from one mammal to another through biting [18], being able to maintain its cell wall morphology inside their digestive tract for until four days [19], as the only viability parameter used on that moment. To verify if M. leprae maintains its viability after ingestion by A. aegypti or C. quinquefasciatus using a reliable molecular tool, these mosquitoes were offered a blood meal containing 10 7 live M. leprae / blood mL, and 16SrRNA levels in the abdomen containing: posterior midgut, ventral diverticulum and hindgut; or thorax and head containing: proboscis, salivary glands and foregut; were measured by qPCR (Fig 1).
As we can see, viable M. leprae levels inside Aedes aegypti abdomen 20 days after blood meal (ABM) are at least two fold lower than levels observed immediately after feeding (2 hours ABM) (Fig 1A). M. leprae viability was virtually abolished inside C. quinquefasciatus digestive tract, presenting a very low positivity only in two and three groups of abdomen and thorax respectively (Fig 1B), indicating that neither A. aegypti nor C. quinquefasciatus presents, in our model, potential to transmit leprosy.
We concluded that there is a lack of an essential characteristic of a disease vector in both mosquito species regarding its interaction with M leprae, which is the ability to keep the pathogen infectious until the next blood meal.
Although there are some circumstantial evidence suggesting that armadillos could be infected by Culex mosquitoes with the St. Louis encephalitis virus in Florida (USA) [35], larger insects, such as kissing bugs, were more frequently found feeding on armadillos, which are recognized sylvatic reservoirs of Chagas Disease [24,36], transmitted by several species of triatomine bugs in Central and South Americas [37]. Rhodnius prolixus has been used over the last decades as a model to study Chagas Disease vectoring [38], due to its adaptability to artificial feeding and growth in laboratory. For these reasons, our next approach was to measure the same parameters observed in mosquitoes, in the kissing bug digestive tract compartments: anterior midgut, posterior midgut and hindgut. When Rhodnius prolixus adults were fed with blood containing M. leprae, (Fig 2) high viability was detected in anterior midgut and hindgut of the insects even 20 days after infection. During this time interval, M leprae levels dropped in posterior midgut, but increased 100-fold in the hindgut, suggesting that the bacteria were moving along with the blood meal. Our data suggests that there was no mycobacterial replication along these 20 days of blood digestion, but the bacteria remained alive, due to the fact that insects exposed to M. leprae inactivated just before ingestion by freeze-thaw was not able to sustain detectable levels of 16Sr RNA in none of the insect digestive compartments after the same period of time (S1 Fig).
In addition to PCR-based detection of live M. leprae cDNA, we performed immunolocalization of the parasite. Fig 3 demonstrates the immunohistofluorescence images from anterior midgut, posterior midgut and hindgut 2 hours and 20 days after the M. leprae infected or control blood meal. As we can see, a small number of mycobacteria could be observed after staining with anti-LAM IgG (red) inside anterior and posterior midgut epithelial cells (arrows) 20 days after infection. In fact, the majority of M. leprae signal was located in the luminal region of the posterior midgut and hindgut (asterisks), and virtually absent from the anterior midgut luminal region. The posterior midgut epithelial cells, in contrast with the anterior one, are refractory to M. leprae infection, been scarcely colonized (arrows), maintaining the large majority of the bacillary load free in the luminal region (asterisk), on its way to the hindgut (Fig 3E). This indicates that the bacillus that did not get arrested at the anterior midgut epithelial cells, were dislocated within the luminal content to the hindgut (Fig 3H) through the posterior midgut (Fig 3E). Due to the huge number of bacteria in the luminal region, and the fragility and autofluorescence of the hindgut epithelia, we were not able to determine precisely the M. leprae location within epithelial cells in this tissue. Regarding the limitation in order to immunolocalize structures inside hindgut epithelia, M. leprae location is clearly predominantly luminal in this tissue, due to the proportional area of these two compartments.
In order to observe if the hindgut high titers of M. leprae observed by PCR are defecated during blood meal, six groups of 10 R. prolixus were fed with blood containing 10 7 /mL PKH-26 stained M. leprae. After 20 days, we offered a non-infected blood meal to these insects in sterilized tubes and collected their feces (Fig 4).
M. leprae amount in feces was evaluated by counting in a fluorescence microscope. We used this material for injecting six mice with 10 4 M. leprae per BALB/c foot pad. After six months, animals were sacrificed, foot pads were macerated and acid-fast mycobacteria number estimated (Table 1). Among the animals, three demonstrated a tenfold increase and one animal presented one hundred times more bacteria than injected in the inoculum. Two animals did not demonstrate mycobacterial growth. After comparing bacillary counts of foot pads inoculated with M. leprae purified from R. prolixus feces with the positive and negative controls, where M. leprae viability is recognized to be higher than 90% and virtually absent respectively, we can see that four from six animals listed on Table 1 presented positive M. leprae infection. It is important to notice that M. leprae cell wall is a mycolic acid rich structure, able to persist intact for months mammal's tissues [39], and for that reason we were able to count dead mycobacteria in negative controls. To conclude, we observed that R. prolixus infected by M. leprae through blood meal are able to maintain the initial inoculums alive, spreading to the environment feces containing infectious M. leprae.

Discussion
Until now, the impact of all possible aspects of transmission of M. leprae is unknown [40]. We investigated here the hypothesis that arthropods could contribute to the transmission of leprosy to vertebrates. This hypothesis was elaborated based on reports regarding the contribution of several insect species to the transmission of Buruli ulcer, a dermal infection caused by M. ulcerans, including predatory water bugs Hemipteras such as belostomatids and naucorids in Africa [41,42] and mosquitos (Diptera) in Australia [43]. We therefore investigated if M. leprae could infect and maintain its viability inside the Dipteras A. aegypti, C. quinquefasciatus and the Hemiptera R. prolixus. The choice of these species was based on reports suggesting that A. aegypti [17,18] and Culex sp. mosquitoes [20,35] proboscis could be infected by M. leprae during the blood meal, mechanically transferring live mycobacteria during the next blood meal [21]. Our data that was focused in long term survival inside the digestive tract, indicates that M. leprae are not viable after 20 days inside mosquitoes digestive tract, which does not turn impossible its transmission through mechanical ways just after an incomplete blood meal in an infected mammal. Literature data estimates M. leprae viability inside mosquitoes through cell wall integrity accessed by microscopy for short periods of time from two [17] to 3 days [19]. On the other hand, longer observation periods in A. aegypti ranging from four [19] to nine days [17] did not demonstrate viable M. leprae in proboscis or posterior midguts, corroborating our data.
Different from what we observed for mosquitoes, we observed that the majority of the M. leprae inoculum was kept alive in the luminal region of the R. prolixus digestive tract, being epithelial cells scarcely infected by M. leprae. This resembles T. cruzi infection in R. prolixus, that also moves from the anterior midgut to hindgut [44]. Differently from the protozoan parasite, we did not observe significant proliferation of M. leprae in the insect gut, in spite of its capacity to maintain viability during prolonged periods. This discrepancy observed between mosquitoes and kissing bug could be attributed to the differences between their digestive tract and intestinal microflora control mechanisms. Both groups of insects are able to generate oxygen and  nitrogen reactive intermediates through prophenoloxidase system in their haemolymph [45,46]. But recently a new microflora control mechanism was described on Aedes aegypti digestive tract. This mechanism involves the formation of reactive intermediates of oxygen by epithelial cells mitochondria, which are responsible for drastically reduce the bacterial load in midgut luminal region 56h after the blood meal [47]. This mechanism could be involved in the complete abolishment of M. leprae viability during blood meal digestion in both mosquitoes analyzed on the present work. Due to a drastic reduction of its genome, suffered during its adaptation to live inside a eukaryotic cell, M. leprae presents 1116 pseudogenes in contrast with 6 in M. tuberculosis [48]. Based on this, the observation that the majority of living M. leprae maintained in the kissing bug midgut is located in the extracellular intestinal lumen contrasts with the usual idea that M. leprae is an obligatory intracellular pathogen [40]. This particular environment represented by R. prolixus digestive tract: microaerobic presenting high amounts of heme-iron and blood-proteins, pH 5,0 and 28°C, maintained M. leprae viability almost intact for 20 days. This is better than the viability observed in Middlebrook 7H12 at 4°C in the same period of time, a condition recognized as the best alternative to maintain extracellular M. leprae viability [31].
In our model we applied 10 7 M. leprae / blood mL as initial inoculum in order to successfully trace M. leprae viability in all intestinal compartments. Since an adult kissing bug can ingest 250 μl of blood and mosquitoes 2 μl, this represents an initial inoculum around 2.5x10 6 and 2x10 4 bacteria per insect respectively. Although this high number of circulating bacteria is improbable to occur in humans, it is possible in infected wild armadillos.
In order to control leprosy efficiently, observing a strong reduction in the number of new cases in the endemic countries, we need to identify other transmission factors beyond multibacillary patients. Multibacillary patients are considered as the primary source of infection of leprosy. The introduction of WHO defined chemotherapy in 1981 reduced in 90% the prevalence of the disease with a reduction in the number of new cases of 50% since then [1,3]. It therefore seems that treatment of multibacillary patients has no impact in the control of leprosy transmission, since the number of new cases around the world remained at similar levels during the last two decades [49]. Leprosy was virtually eradicated from Norway one century before the development of an efficient treatment against the disease [50] by means of industrialization, migration of the population from the countryside to the cities and deforestation of innumerous habitats of insects and small mammals [51][52][53]. New evidences point that other factors contribute to the maintenance of the disease among humans, such as water and soil contaminated with M. leprae infected amoebas [54].
The mode of Leprosy transmission by Rhodnius prolixus kissing bug hypothesized here is analogous to that described for the contribution of insect species to the transmission of Buruli ulcer, a dermal infection caused by M. ulcerans, for which transmission by predatory water bugs Hemipteras, such as belostomatids and naucorids, has been reported in Africa [41,42] and by mosquitos (Diptera) in Australia [43]. The hemipterans present potential as leprosy vectors, since kissing bugs usually feed on nine-band armadillos, the most acceptable wild reservoir of leprosy [12,24]. In fact, about 1% of nine-band armadillos population in the south of US is co-infected by M. leprae and T. cruzi [23]. Although M. leprae persistence inside R. prolixus, or even in free-living amoebae [54], is not a proof of involvement of these organisms in leprosy transmission, we hypothesize that kissing bugs, well distributed in Americas, Asia and Africa could be infected by M. leprae during blood meal in natural reservoirs such as armadillos in Americas. These infected kissing bugs could develop a role in the endemic leprosy observed in the Brazilian agricultural frontier, where we can find high incidence of leprosy, armadillos and triatomines [24].
Our data indicate that kissing bugs are able to defecate infectious M. leprae after being infected through a blood meal. The incidence of kissing bugs infected by M. leprae in highly endemic areas is currently under investigation by our group, through detection of the presence of live M. leprae in kissing bugs captured from leprosy patients domiciliary and peridomiciliary areas in Brazilian countryside.
Supporting Information S1 Fig. Inactivated M. leprae was not able to sustained detectable levels of 16SrRNA after 20 days inside digestive tract. Freeze-thaw inactivated M. leprae 16Sr RNA persistence at the different digestive compartments of adult Rhodnius prolixus: anterior midgut (spheres), posterior midgut (squares) and hindgut (triangles), just after blood meal (2h) and after total blood meal digestion (20 days). Scatter plot showing mean and SEM of four independent experiments, each point represent five insects group. ÃÃÃ means p < 0.001. Controls did not present amplification of the targets. (TIF)