Caspar-like Gene Depletion Reduces Leishmania Infection in Sand Fly Host Lutzomyia longipalpis*

Background: Caspar is a negative regulator of the inducible IMD immune signaling pathway. Results: Depletion of Caspar expression led to a reduction in Leishmania populations in sand fly gut. Conclusion: Disrupting expression of a single sand fly immune gene disrupts Leishmania development. Significance: This study suggests that IMD-activated effectors have the potential to disrupt Leishmania population development and abort infections in sand fly vector. Female phlebotomine sand flies Lutzomyia longipalpis naturally harbor populations of the medically important Leishmania infantum (syn. Leishmania chagasi) parasite in the gut, but the extent to which the parasite interacts with the immune system of the insect vector is unknown. To investigate the sand fly immune response and its interaction with the Leishmania parasite, we identified a homologue for caspar, a negative regulator of immune deficiency signaling pathway. We found that feeding antibiotics to adult female L. longipalpis resulted in an up-regulation of caspar expression relative to controls. caspar was differentially expressed when females were fed on Gram-negative and Gram-positive bacterial species. caspar expression was significantly down-regulated in females between 3 and 6 days after a blood feed containing Leishmania mexicana amastigotes. RNA interference was used to deplete caspar expression in female L. longipalpis, which were subsequently fed with Leishmania in a blood meal. Sand fly gut populations of both L. mexicana and L. infantum were significantly reduced in caspar-depleted females. The prevalence of L. infantum infection in the females fell from 85 to 45%. Our results provide the first insight into the operation of immune homeostasis in phlebotomine sand flies during the growth of bacterial and Leishmania populations in the digestive tract. We have demonstrated that the activation of the sand fly immune system, via depletion of a single gene, can lead to the abortion of Leishmania development and the disruption of transmission by the phlebotomine sand fly.

The phlebotomine sand fly Lutzomyia longipalpis is the most important vector of American visceral leishmaniasis in Central and South America. Leishmania infantum is the parasite naturally transmitted during the bite of this sand fly, although in the lab this sand fly is permissive; it allows the growth and transmission of several other Leishmania species including Leishmania mexicana. Gut microbial homeostasis is of key importance in phlebotomine sand fly transmission of Leishmania parasites. In contrast to the malaria parasite Plasmodium, the medically important Leishmania parasite resides in the gut of the insect during its entire time in the insect vector, and thousands of parasites are often observed packed inside the gut with little apparent pathogenic effect on the host. We do not know whether the insect tolerates the Leishmania parasite in the gut or whether here is immune suppression and presume that the parasite is able to resist some or all of the antibiotic activities of the sand fly.
Dipteran insects, as exemplified by Drosophila, possess an innate immunity thought to rely on two main types of effectors that act synergistically to restrict microbial growth; antimicrobial peptides and reactive oxygen species (1). Production of these antimicrobials is carefully regulated to allow for growth of commensal microbes and also limit nonspecific damage to the gut epithelia by reactive oxygen species (2). Antimicrobial peptide production is governed by the signaling immune pathways Toll, immune deficiency (IMD) 3 with JAK-STAT having an additional immune function. Epithelial immune effector induction is thought to be under the control of IMD and not Toll, probably because of the direct contact with MAMP receptors in epithelial cells (3). Repression of the NF-B antimicrobial peptide system is needed to prevent the loss of potentially beneficial bacteria, and the insect innate immune pathways are partly governed by negative regulators. The caspar gene was first described from Drosophila mutants with hyperactivated immune responses. The Caspar gene encodes a protein homologue of the human Fas-associating factor 1, which is associated with elements of the TNF/NF-B pathway and which negatively controls the IMD pathway (4).
Both phlebotomine sand flies and mosquitoes are from the order Diptera, in common with Drosophila, and belong to the same suborder Nematocera. Studies on Anopheline mosquito-Plasmodium interactions showed that Toll and IMD pathways respond to parasite infections, and these responses are specific to different Plasmodium species (5)(6)(7). Moreover, RNAi-mediated knockdown of the Caspar gene is capable of reducing Plasmodium falciparum survival in the anopheline insects (8). However, depletion of Caspar in Aedes aegypti did not affect infection by the nematode Brugia malayi that cause lymphatic filariasis (9).
The phlebotomine sand fly immune response is poorly studied, and a detailed description of the relevant pathways awaits the publication of the sand fly genome. In Phlebotomus papatasi, an anti-Gram-negative activity was detected in the insect hemolymph, as well as lysozyme-like activity against Grampositive bacteria and phenoloxidase activity (10). In L. longipalpis, it was shown that bacterial challenge can induce antibacterial peptides in the insect hemolymph (11). Additionally, cultured cells generated a nonspecific response to virus-like particles (12). Only one study has examined the sand fly immune response toward Leishmania parasites in detail. Defensin was induced in Phlebotomus duboscqi by bacterial and Leishmania challenge, with antiparasitic activity against the promastigote forms of Leishmania major (13).
In this study, we identify and describe Caspar, a putative negative immune regulator in L. longipalpis. We used the RNAi approach developed in our laboratories (12, 14 -16) to evaluate the role of Caspar in controlling both the naturally transmitted parasite L. infantum and an additional species L. mexicana that is successfully transmitted by this insect in the laboratory.

EXPERIMENTAL PROCEDURES
Ethics Statement-All procedures involving animals were approved by the ethical review committee of the University of Liverpool and performed in accordance with United Kingdom government (Home Office-approved Project License PPL40/ 2958) and European Community regulations.
Insects-All of the experiments were performed using sand flies from a laboratory colony of L. longipalpis established from sand flies caught in Jacobina (Bahia, Brazil) using standard methods described by Modi (18). The insects were fed on 70% sucrose ad libitum and blood-fed on heparinized rabbit blood when needed. The insects were maintained at 27 Ϯ 1°C, humidity of 80 -95%, and a photoperiod schedule of 12 h light/12 h dark.
Feeding Blood and Leishmania Parasites-Three-day-old female sand flies were fed through chick skin membrane on heparinized rabbit blood. Artificial infection was performed with heparinized rabbit blood containing either L. mexicana amastigotes or L. infantum early log phase promastigotes (2 ϫ 10 6 parasites/ml).
Sucrose Feeding and Antibiotic Feeding-A solution of 70% sucrose or 70% sucrose supplemented with a mixture of penicillin (200 units/ml), streptomycin (0.2 mg/ml), and amphotericin B (0.5 g/ml) (Sigma) was offered to insects on cotton wool and changed daily.
Caspar cDNA Sequence-caspar partial cDNA sequence was identified from NCBI dbESTs database. The complete caspar coding sequences was predicted from L. longipalpis genomic data available at the NCBI Sequence Read Archives Database. The next generation sequences from file SRX016813 (accession numbers SRR036068, SRR036069, SRR036071, SRR036073, SRR036074, SRR036076, SRR036079, and SRR036080) and SRX016812 (accession numbers SRR036070, SRR036072, SRR036075, SRR036077, SRR036078, and SRR036081) were originally generated in LS454 platform. The conversion from fastq to fasta file format was performed by Galaxy platform (19) and used as a database for a local search. The partial sequence of caspar was used as a query for Blast searching. The obtained reads were subsequently assembled with CAP3 program (29), and the contig sequence was resubmitted for a blastx analysis against the NCBI nr database for sequence confirmation. The complete CDS of caspar was manually reviewed based on alignments against expressed sequence tags retrieved from Gen-Bank TM and splicing sites prediction with the NNSPLICE 0.9 (20) available online at Berkeley Drosophila Genome Project website. To confirm caspar CDS prediction specific oligonucleotides were designed onto the predicted start and stop codons, and the complete ORF of Caspar was amplified by PCR from L. longipalpis cDNA prepared with total RNA extracted from L3 larvae. The amplified fragment was cloned into pGEM-T Easy (Promega) and sequenced. The identity of all cDNA sequences was assigned by similarity using BLAST search, and standard ClustalW multiple alignment tool Cladogram analysis was performed using MEGA4 software with a neighbor joining method with bootstrap 10,000 (21). Conserved domains were predicted by the MyHits (22) and Inter-ProScan (23) on-line analysis tools.
RNA Extractions and RT-PCR-Total RNA was extracted in biological triplicates from pools of three whole bodies of L. longipalpis larvae or adults using TRI reagent (Ambion). RT-PCR was performed using SuperScript III one-step RT-PCR Platinum Taq HiFi (Invitrogen) according to the manufacturer's instructions, with 10 ng of template RNA, and gene-specific primers (Caspar-F, 5Ј-CCAAAGAGGAGGCAAGAAAGA-3Ј, and Caspar-R, 5Ј-TTCCGCTTCAAGACGCATA-3Ј, using 24 cycles. Transcription of genes was normalized based on tran-scription of a housekeeping gene (AM088777, 60 S ribosomal protein L3). RT-PCR products were resolved by standardized 2% agarose 0.03 ng/ml ethidium bromide gels. Band intensity was determined by densitometric measurements using the software ImageJ 1.42q from standardized gel images captured in 70 ms with the GeneSnap system (Syngene, Cambridge, UK). Differential transcription of genes was determined by the ratio between target gene band intensity and the corresponding 60 S L3 products obtained from multiplex RT-PCR.
Double-stranded RNA Synthesis-Caspar DNA template was amplified by RT-PCR as described above with a second pair of gene specific primers added by T7 site at their 5Ј ends (T7-CasparDS-F, 5Ј-CCGTAATACGACTCACTATAGGGAAC-CCAGTGGTGATTTCCTCG-3Ј, and T7-CasparDS-R, 5Ј-CCGTAATACGACTCACTATAGGGATAGCGTTTCATC-TGCATCCATC-3Ј). The GFP DNA template was obtained from a PCR using pBluescript (Stratagene). The PCR products of Caspar and GFP were used as templates for transcription reactions with the Megascript RNAi kit (Ambion) according to the manufacturer's instructions. The purity of the dsRNA was determined by 2% agarose gel electrophoresis and quantified using a Nanodrop ND-1000 spectrophotometer, which was subsequently concentrated to 4.5 g/l using a vacuum concentrator.
Double-stranded RNA Microinjections-Double-stranded RNAs were microinjected as described (16). Recently emerged (Ͻ8 h) L. longipalpis females were injected with 32 nl of 4.5 g/l Caspar or GFP (mock-injected group) dsRNA solution in RNase free water. After injections, the flies were kept in cages (Ͼ90% relative humidity) and supplied with 70% (w/v) sucrose solution ad libitum. The insects were subsequently infected with L. mexicana or L. infantum (2 ϫ 10 6 parasites ml Ϫ1 ) 3 days after microinjections. Additional control groups were sand flies without injections and flies mock injected (pricked) with a sterile needle.
Verification of Caspar Depletion-The depletion of the Caspar gene in sand flies injected with Caspar dsRNA, GFP dsRNAs, or uninjected was evaluated by RT-PCR using pools of three sand flies from each group collected in triplicate 3 and 4 days after microinjections.
Bacterial Quantification in Caspar-depleted Insects-Detection of bacterial DNA by quantitative PCR was performed with an Opticon 2 (Mj Research) using the QuantiTect SYBR Green PCR kit (Qiagen). Total DNA was extracted from pools of 10 whole female sand flies using the DNeasy blood and tissue kit (Qiagen), either from sugar-fed flies (20% sucrose) or from flies injected with Caspar or GFP dsRNA and collected at 3 and 4 days after the treatment. Each sample was diluted to a final concentration of 50 ng of DNA l Ϫ1 and added to the quantitative PCR. The quantitative PCR was performed in triplicate in a total volume of 25 l with 0.3 M of each universal bacteriaspecific primers as previously described (24). Asaia sp. DNA was amplified from bacterial lysates using universal bacterial primers 27F and 1492R (25) and used as standard for determining bacterial copy numbers.
Parasite Population in Caspar-depleted Insects-The midguts of dsRNA microinjected insects were dissected 5 days after infection and homogenized in 50 l of 0.15 M NaCl. The number of Leishmania parasites present in the midgut of sand flies injected with Caspar dsRNA, mock injected, and not injected was estimated using a hemocytometer after fixation in 2% (w/v) formaldehyde in PBS.
Statistical Analysis-All of the experiments were analyzed for statistical significance using the software SPSS v17.0. The Kruskall Wallis test was used for multiple comparisons, and the Mann-Whitney test was used for pair wise comparisons. p Ͻ 0.05 was considered statistically significant.

RESULTS
Caspar cDNA Sequence Identification-The Caspar cDNA sequence (AM093416 L. longipalpis (26)) predicted translation showed a putative amino acid sequence of 676 amino acids (supplemental Fig. S1) with an estimated molecular mass of 76 kDa and pI of 5.31 containing a UAS-FAF1 like domain (ProSite access PS50033) (residues 375-510) and an UBX ubiquitin-like domain (ProSite access PS50033) (residues 596 -673) (Fig. 1). The sand fly Caspar-like amino acid sequence shared 43 and 46% identity with Drosophila melanogaster and A. gambiae sequences, respectively. Multiple alignment of this sequence with FAS-like sequences from other organisms (supplemental Fig. S2) was used to construct a cladogram, in which the L. longipalpis Caspar-like sequence was shown to be closely related to the three other Culicidae (Fig. 1). A broader clade containing all of the nematoceran sequences grouped with Caspar from Drosophila to form a clade with all dipteran sequences. Caspar sequences from coleopterans, homopterans, and hymenopterans clustered together but were placed more distantly from all dipteran sequences.
Caspar Expression Profile-Semiquantitative RT-PCR was performed with samples obtained from developmental stages and adults of L. longipalpis. Expression of the genes was noted in all stages. Generally there was no clear pattern in relative transcription of Caspar in L3, L4, pupae, and newly emerged females ( Fig. 2A). The response of the sand flies toward microbial infection of the gut was initially investigated by treating the females with a mixture of penicillin, streptomycin, and amphotericin via the sugar meal. There was a trend of an increase in Caspar when compared with non-antibiotic-treated insects that was statistically significant at 2 days post-sugar feeding (Fig. 2B).
RT-PCR was also performed (Fig. 3) with sand flies collected at 1, 2, 3, and 6 days after feeding on rabbit blood or a blood meal containing L. mexicana amastigotes (80% of sand flies were estimated to be infected after 5 days). A significant decrease in Caspar expression was recorded for infected insects compared with controls sampled 3 and 6 days after initial infection (Fig. 3). Caspar expression was subsequently examined after introducing other microorganisms into the sand fly gut by feeding live bacteria in 20% sucrose solution relative to controls fed on 20% sucrose only (Fig. 4). E. coli and M. luteus were included as standard type species. S. marcescens was used because of its potential as an insect pathogen that has been found in sand flies (27), and Ochrobactrum is found at a low level in the sand fly colony and has been found in other sand fly colonies (28). P. agglomerans is a commonly isolated commensal/symbiont from many species of insects including sand flies (27,29). Expression was significantly increased in the first 24 h after feeding E. coli relative to the control but did not differ thereafter (Fig. 4A). The M. luteus initiated an early caspar upregulation after 24 h followed by down-regulation after 3 and 4 days (Fig. 4B). Expression after feeding Ochrobactrum sp., P. agglomerans, and S. marcescens showed similar overall patterns of expression (Fig. 4, C-E). Caspar was significantly reduced in insects fed Ochrobactrum sp. or P. agglomerans 2 and 3 days after feeding, but Caspar expression was also increased in adults fed after 5 days on Ochrobactrum sp. Caspar expression was significantly up-regulated 4 days after feeding with the S. marcescens (Fig. 4E).
Confirmation of Caspar Silencing-The longevity of Caspar depletion after dsRNA microinjection was investigated by sampling relative Caspar expression in uninjected, GFP dsRNAinjected, and Caspar dsRNA-injected sand flies 3 and 4 days after microinjections; the time points were selected to correspond to the time between initial sand fly infection and the time of Leishmania population estimation (Fig. 5). After microinjections with Caspar dsRNA, Caspar relative expression was ϳ3-fold in comparison with uninjected controls and ϳ2-fold in comparison with mock injected controls at 3 days after injection ( Fig. 5; p Ͻ 0.05).
Leishmania Survival in Caspar-depleted Insects-Caspar transcripts were silenced using RNAi (60% depletion after 3 days) and females fed L. mexicana amastigotes in a blood meal 3 days after the dsRNA microinjections. Infections in Caspardepleted sand flies initiated with L. mexicana amastigotes showed a significant decrease in the size of the parasite population surviving 5 days after initial feeding in the blood meal (Fig. 6A, dsGFP versus dsCaspar p Ͻ 0.0001).
Two additional controls (no injections and mock injection/ pricking the cuticle) were included in our RNAi experiments to FIGURE 2. Caspar expression. A, immature developmental stages of L. longipalpis. Semiquantitative RT-PCR was performed on L3, L4-f (feeding), and L4 (nonfeeding) larval stages, pupae, and newly emerged females (NE). The error bars represent the standard errors for three biological replicates consisting of pools of three whole bodies. B, feeding antibiotics changes Caspar expression in female L. longipalpis. The adults were fed a solution of 70% (w/v) sucrose supplemented with antibiotics (penicillin 200 units/ml; streptomycin 0.2 mg/ml; amphotericin B 0.5 g/ml; open bars) compared with controls fed 70% sucrose solution (gray bars) changed daily. Semiquantitative RT-PCR was performed on newly emerged females (NE) and females collected on days 1-3 (1d, 2d, and 3d, respectively). The error bars represent the standard errors for three biological replicates consisting of pools of three whole bodies per sample. *, p Ͻ 0.05. account for the possibility that injections alone may affect immune gene transcription. It was interesting to note that merely pricking the cuticle was sufficient to cause a significant reduction ( Fig. 6A; uninjected versus pricked p Ͻ 0.0242) in the population of parasites in the gut. This result suggests that general systemic immune activation prior to feeding the sand flies may influence the outcome of a gut parasite infection in these insects. The final control found no significant difference in parasite population size between pricked insects and dsGFP-injected sand flies, confirming that the double-stranded RNA did not promote nonspecific anti-leishmanial effects. Caspar-depleted sand flies possessed significantly lower numbers of parasites compared with all three controls.
The experiment was repeated using the natural host-parasite combination of L. longipalpis-L. infantum. In this experiment, early log phase promastigote parasite forms were used to initiate infections, and the percentage of infected sand flies was significantly reduced from 85% in dsGFP-injected controls to 45% in dsCaspar-injected (60% depletion) insects ( Fig. 6D; p Ͻ 0.0001), and there was a significant reduction in parasite numbers in Caspar-depleted insects compared with the dsGFP-injected control ( Fig. 6C; p Ͻ 0.044).
Promastigote initiated infections produce populations with more rapid growth kinetics compared with amastigote initiated infections; there is no lag phase in Leishmania population growth. Early log phase promastigotes will provide a vigorous, fast growing parasite population in the gut. This also means that the potentially sensitive transformation phase from amastigote to promastigote is not present, and sand fly host-initiated anti-leishmanial action is therefore occurring against the log phase promastigote forms and not confined to the transformation stage. The mean parasite populations present in the L. infantum infections were less than 25% that of the L. mexicana infections (Fig. 6, A and C); this was not a consequence of the use of promastigote-initiated infections, because Gossage et al. (17) used amastigote-initiated infections and reported more vigorous growth of L. mexicana in vivo compared with L. infantum. Therefore L. infantum amastigote-initiated infections may show even more significant reductions in parasite infection after Caspar depletion.
Bacterial Quantification in Caspar-depleted Insects-Total bacterial population in the female sand flies was measured 3 and 4 days after microinjections for Caspar silencing (Fig. 7). After 3 days there was no difference as measured by bacterial 16 S gene copy number. Although there was a trend toward an increase in bacterial population in Caspar-silenced sand flies after 4 days, there were no significant differences between dsCaspar-injected insects and dsGFP-injected (n ϭ 6, pools of 10 insects/replicate; p Ͼ 0.05).

DISCUSSION
By inference from our knowledge of immune pathways in mosquitoes and Drosophila, we suggest that Caspar regulates the IMD signaling pathway in sand flies. Caspar depletion would then lead to IMD-related effector up-regulation leading to as yet unspecified, anti-leishmanial activities in the gut. This suggests that IMD-activated effectors have the potential to disrupt Leishmania population development and abort infections in the midgut lumen of the sand fly. Caspar depletion led to a fall from 85 to 45% of sand flies infected with L. infantum. This is the first time that depletion of a single immune gene has been shown to lead to abortion of Leishmania development in sand flies.
The phlebotomine sand fly has evolved to deal with a diverse microbiota throughout its life and is exposed to high titers of MAMPs. The larvae live on decomposing plant and animal material, whereas the adults feed on plant sap (30) and presumably will acquire phyllosphere-associated microbiota. Female sand flies are also telmophagic; their blood feeding is preceded by the creation of a wound and formation of a blood pool, which is contaminated with animal skin microbiota. To control the equilibrium of potentially beneficial microorganisms with harmful microorganisms acquired by sand flies, they must activate and maintain a basal immune response. From the perspective of the sand fly, the medically important Leishmania parasite is just one of many occupants of the gut; even in areas considered to harbor Leishmania epidemics, the percentage of female vectors harboring Leishmania may be as low as 7% (31).
Very little is known about the signaling cascades initiated after exposure to MAMPs in these sand flies, but the collection of putative immune genes described so far (26,(32)(33)(34) suggests that Lutzomyia is similar to Anopheles in the way it is regulating NF-B activation of signaling cascades. We identified one sequence in the L. longipalpis cDNA libraries (26) that was similar to the Drosophila Caspar gene sharing higher similarity to the A. gambiae Caspar sequence. The sand fly amino acid sequence contains the conserved UAS and UBX domains as described in Drosophila (4) and in the equivalent human FAF1 (35).
Considering that Caspar is thought to be a negative regulator of the innate immune pathways, the overall profile of expression of these genes appeared unchanged throughout the juvenile developmental stages of the sand fly. The development of the gut microbiota in the larval stages is unknown, but gut microbial homeostasis will be important for an appropriate response toward pathogens and the prevention of a constant immune activation by harmless or beneficial food-associated microorganisms. Feeding sucrose solution to adult sand flies during laboratory rearing is a replacement for a carbohydrate-rich plant meal. L. longipalpis exhibited lower Caspar transcription in sucrosefed females when compared with sucrose with antibiotic-fed females. Bacteria are commonly associated with the gut of healthy phlebotomine sand flies in our laboratory colony, 4 and the transcription profiles confirm that all three immune signaling pathways are responding and functional in phlebotomine sand flies. Transcription of the putative negative regulator of the immune response was investigated during the 6 days following L. mexicana infections in comparison with a blood-fed control group. This time period covers the whole development process of the promastigote population from transformation within the blood meal during digestion through to the induction of mammalian infective metacyclic promastigotes (36).
Caspar transcription did not change during the first 2 days but was significantly down-regulated in infected insects between 3 and 6 days after Leishmania entered the gut; this corresponds to the period when the parasite escapes the blood meal inside the peritrophic matrix and starts to colonize the midgut lumen. Caspar may be down-regulated as a belated immune response by the sand fly host toward the large developing population of Leishmania parasites.
A more detailed study of Caspar expression was conducted during the 5 days after feeding on sugar meals with five different bacteria. There were significant changes in the relative expression but no clear pattern in transcription related to the different types of bacteria. This may reflect the differences in the ability of the bacteria to grow within the insect gut, as well as the type of bacteria used. There was a trend for down-regulation of Caspar in the first 3 days after feeding S. marcescens, but it was not significant as might be expected in response to being fed the potential pathogen. Feeding S. marcescens to Drosophila activates a local but not systemic IMD-dependent response in the midgut with Diptericin transcription (37). Another consideration is that preformed antimicrobial effectors might be present for fast release in response to sudden detection of MAMPs and that transcription is for replenishment (6).
We hypothesized that if the IMD pathway was potentially important in gut epithelial immunity, then early activation of the IMD pathway by RNAi-mediated depletion of Caspar might lead to abrogation of the Leishmania gut infection. Infection experiments with Caspar-depleted insects verified the hypothesis, showing significant effects on both L. mexicana and L. infantum infections. The results demonstrate the robustness of the effect of Caspar depletion on Leishmania infection in L. longipalpis. The growth of L. mexicana, which generally produce parasite populations more than four times larger than that of the naturally occurring L. infantum, was significantly impeded. The experiment also showed that merely pricking the cuticle with a sterile glass needle was sufficient to significantly reduce the promastigote population in the gut compared with uninjected controls. This control has not been included in most previous RNAi studies on insect immunity that used the injection route to administer the dsRNA. This confirms our previous finding for silencing a sand fly trypsin gene where injections   with the control dsRNA caused a reduction in the Leishmania population (15). The potential for systemic effects of a nonspecific immune activation on the development of a parasite population needs to be considered. Experiments with tsetse flies showed that injection of bacteria was sufficient to elicit a systemic immune response that resulted in a fall in trypanosome gut infection prevalence (38). However, there was no effect of sterile injection on infection prevalence in tsetse flies. Further studies on the nature of this nonspecific anti-leishmanial response are needed to establish whether this is a systemic immune-related response.
Although attempts were made to minimize bacterial populations in the sand fly gut by aseptic technique, the influence of even low levels of insect gut bacteria in the establishment of gut-colonizing Leishmania cannot be ignored (29,40). The effect of gut microbiota of anopheline mosquitoes on Plasmodium development in the gut suggests a tripartite interplay between insect-bacteria-parasite with a combination of indirect bacterial mediated immune activation (41) and direct bacterial effects on the malaria parasite (42). Quantification of bac-teria in Caspar-depleted insects suggests no difference in the low level of total bacterial populations associated with the insects at the time of feeding the Leishmania parasite in the blood meal. However, the possibility that Caspar-depleted insects contained a different composition of bacteria that influence Leishmania development cannot be ruled out.
The role of Caspar and the IMD pathway appears to be pivotal in malaria-mosquito interactions. Caspar depletion by RNAi was shown to reduce P. falciparum infections in all of the major malaria mosquito vectors (8). Plasmodium sp. infections were assessed by estimating oocyst numbers, so in mosquitoes the immune activation may be operating either on the parasite in the gut or as it traverses the midgut epithelium prior to oocyst formation on hemolymph side of the gut. This differs from the current study because the Leishmania parasites are not thought to penetrate through the gut epithelia, and any immune effectors must pass into the gut lumen after Caspar depletion.
The degree to which sand flies invoke an immune response against Leishmania infections in the natural situation is not Females were fed via a chick skin membrane on blood containing L. mexicana amastigotes or L. infantum promastigotes. The insects were injected via the thorax with 144 ng of dsRNA in sterile water 3 days before blood feeding, and parasite populations were recorded 5 days after blood feeding. Further controls included insects without injections, insects pricked in the thorax with an empty glass microneedle, and insects injected with GFP dsRNA. The results are the products of three separate experiments (A and B) or two separate experiments (C and D). The horizontal bars are the standard errors of the mean. The asterisk in D denotes significant difference using chi-squared analysis between GFP dsRNA-and Caspar dsRNA-injected flies.
known, but in our experimental lab system, 80 -100% infections are commonly achieved, particularly with L. mexicana. The metabolic costs of inducing an insect vector immune response are high (43), and this will be balanced against the potential pathogenicity of the parasite. Leishmania are thought to have limited effects on L. longipalpis longevity and fecundity (39), and this corroborates our data suggesting that there might be a limited immune response during infections.
This study of the Caspar gene provides a basis to further investigate the immune signaling cascades and their role in Leishmania survival and development in the vector. The results suggest that Leishmania gut infections may invoke a "late" IMD-mediated immune response, but this is ineffective in preventing the development of a transmissible parasite population. The discovery that Caspar silencing can abort Leishmania transmission in the sand fly demonstrates the importance of the innate immune pathways as targets for developing new strategies to disrupt sand fly vector competence. One such scenario would be a Leishmania gut infection-inducible promoter that switches on the IMD pathway leading to the loss of sand fly vector competence.